ci: run the OpenBSD --use-tcp test step at -j2

The OpenBSD job runs inside a nested VM. At -j8 the --use-tcp run starts
many concurrent loopback daemons, and under that resource pressure the
daemon connection handshake occasionally loses a timing race and one test
hangs to the 300s runner timeout. It is an environment artifact, not an
rsync defect: the daemon handshake writes-then-reads with unbuffered early
I/O (no flush/mutual-wait deadlock), the indefinite wait is the documented
no-timeout daemon behaviour, and it does not reproduce off OpenBSD even with
the full suite pinned to a single CPU at -j8.

Drop just this job's --use-tcp parallelism to -j2 so the nested VM stops
over-subscribing; the pipe `make check` and every other platform are
unchanged.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

.gitattributes

@@ -1 +1,8 @@
 * text=auto eol=lf
+
+# The rsync-web/ subdirectory holds the project website source content
+# (mirrors what gets pushed to https://rsync.samba.org).  Exclude it from
+# `git archive` output so the release source tarball produced by
+# packaging/release.py step_7_tarball does not bloat with HTML the
+# tarball doesn't need.
+/rsync-web/ export-ignore

.github/workflows/almalinux-8-build.yml

@@ -0,0 +1,82 @@
+name: Test rsync on AlmaLinux 8
+
+# Older-LTS coverage on the Fedora/RHEL family to help with backporting
+# security fixes. AlmaLinux 8 is the RHEL 8 rebuild and is the oldest
+# active LTS in this family (RHEL 8 full support runs to 2029).
+# GitHub Actions has no native runner for this family, so the job runs
+# inside an almalinux:8 container hosted on ubuntu-latest.
+
+on:
+  push:
+    branches: [ master ]
+    paths-ignore:
+      - '.github/workflows/*.yml'
+      - '!.github/workflows/almalinux-8-build.yml'
+  pull_request:
+    branches: [ master ]
+    paths-ignore:
+      - '.github/workflows/*.yml'
+      - '!.github/workflows/almalinux-8-build.yml'
+  schedule:
+    - cron: '42 8 * * *'
+
+jobs:
+  test:
+    runs-on: ubuntu-latest
+    container:
+      image: almalinux:8
+    name: Test rsync on AlmaLinux 8
+    steps:
+    - name: install git
+      # actions/checkout needs git in the container before the checkout step.
+      run: dnf -y install git
+    - uses: actions/checkout@v4
+      with:
+        fetch-depth: 0
+    - name: prep
+      # PowerTools is needed for libzstd-devel etc; xxhash and lz4 dev
+      # headers live in EPEL on RHEL 8. The default python3 on RHEL 8
+      # is 3.6, which is too old for runtests.py (uses capture_output=
+      # / text= introduced in 3.7), so install python39 and point
+      # /usr/bin/python3 at it.
+      run: |
+        dnf -y install epel-release
+        dnf config-manager --set-enabled powertools
+        dnf -y install gcc gcc-c++ make autoconf automake m4 \
+            python39 python39-pip diffutils \
+            openssl openssl-devel \
+            attr libattr-devel acl libacl-devel \
+            zstd libzstd-devel \
+            lz4 lz4-devel \
+            xxhash xxhash-devel
+        alternatives --set python3 /usr/bin/python3.9
+        pip3 install commonmark
+    - name: configure
+      run: ./configure --with-rrsync
+    - name: make
+      run: make
+    - name: info
+      run: ./rsync --version
+    - name: check
+      # In the container we already run as root, so no sudo. The
+      # crtimes-not-supported skip matches the other Linux jobs;
+      # daemon-chroot-acl and proxy-response-line-too-long skip because
+      # the default (secure) transport opens no listening socket.
+      run: RSYNC_EXPECT_SKIPPED=crtimes,daemon-access-ip,daemon-chroot-acl,proxy-response-line-too-long make check
+    - name: check (TCP daemon transport)
+      # Second run exercising the real loopback-TCP daemon path.
+      run: ./runtests.py --rsync-bin=`pwd`/rsync --use-tcp -j 8
+    - name: ssl file list
+      run: ./rsync-ssl --no-motd download.samba.org::rsyncftp/ || true
+    - name: save artifact
+      uses: actions/upload-artifact@v4
+      with:
+        name: almalinux-8-bin
+        path: |
+          rsync
+          rsync-ssl
+          rsync.1
+          rsync-ssl.1
+          rsyncd.conf.5
+          rrsync.1
+          rrsync

.github/workflows/android-static-build.yml

@@ -0,0 +1,120 @@
+name: Build static rsync for Android
+
+# Cross-compiles statically-linked rsync binaries with the Android NDK,
+# suitable for dropping onto a phone (adb push / Termux) with no shared
+# libraries. arm64-v8a covers all modern phones; armeabi-v7a covers older
+# 32-bit devices. The binaries are uploaded as workflow artifacts.
+#
+# These are cross-compiled, so the test suite can't run here; we sanity
+# check that each binary is the right architecture, is static, and that
+# it executes (`--version`) under qemu-user.
+
+on:
+  push:
+    branches: [ master ]
+    paths-ignore:
+      - '.github/workflows/*.yml'
+      - '!.github/workflows/android-static-build.yml'
+  pull_request:
+    branches: [ master ]
+    paths-ignore:
+      - '.github/workflows/*.yml'
+      - '!.github/workflows/android-static-build.yml'
+  schedule:
+    - cron: '42 8 * * *'
+  workflow_dispatch:
+
+env:
+  # Minimum supported API level. 24 (Android 7.0) runs on every modern
+  # phone while keeping broad reach; bump if you need newer Bionic APIs.
+  ANDROID_API: 24
+
+jobs:
+  build:
+    runs-on: ubuntu-latest
+    name: ${{ matrix.abi }}
+    strategy:
+      fail-fast: false
+      matrix:
+        include:
+          - abi: arm64-v8a       # modern phones
+            triple: aarch64-linux-android
+            qemu: qemu-aarch64-static
+          - abi: armeabi-v7a     # older 32-bit phones
+            triple: armv7a-linux-androideabi
+            qemu: qemu-arm-static
+    steps:
+      - uses: actions/checkout@v4
+        with:
+          fetch-depth: 0
+
+      - name: Install build prerequisites
+        run: sudo apt-get update && sudo apt-get install -y autoconf automake gawk qemu-user-static
+
+      - name: Configure and build (${{ matrix.abi }})
+        shell: bash
+        run: |
+          set -euo pipefail
+          NDK="${ANDROID_NDK_LATEST_HOME:-$ANDROID_NDK_ROOT}"
+          TC="$NDK/toolchains/llvm/prebuilt/linux-x86_64/bin"
+          export CC="$TC/${{ matrix.triple }}${ANDROID_API}-clang"
+          export AR="$TC/llvm-ar" RANLIB="$TC/llvm-ranlib" STRIP="$TC/llvm-strip"
+          export CFLAGS="-O2" LDFLAGS="-static"
+
+          # Bionic doesn't declare lchmod()/lutimes() until API 36, but the
+          # symbols link, so configure mis-detects them -- force them off so
+          # rsync uses its fallbacks. The other cache vars restore values
+          # that configure can't probe when cross-compiling (Android runs a
+          # normal Linux kernel, so these match the native Linux result).
+          export ac_cv_func_lchmod=no ac_cv_func_lutimes=no \
+                 rsync_cv_HAVE_SOCKETPAIR=yes \
+                 rsync_cv_MKNOD_CREATES_FIFOS=yes \
+                 rsync_cv_MKNOD_CREATES_SOCKETS=yes
+
+          # Self-contained build: drop optional external libraries so the
+          # static binary needs nothing at runtime. rsync keeps md5/md4
+          # checksums and its bundled zlib.
+          ./configure --host=${{ matrix.triple }} --build=x86_64-pc-linux-gnu \
+            --enable-ipv6 \
+            --disable-zstd --disable-lz4 --disable-xxhash --disable-openssl \
+            --disable-iconv --disable-iconv-open \
+            --disable-acl-support --disable-xattr-support \
+            --disable-md2man --disable-roll-simd \
+            --with-included-popt --with-included-zlib
+
+          # Generate the awk-built headers serially first so the parallel
+          # build can't race on proto.h <- daemon-parm.h.
+          make proto.h
+          make -j"$(nproc)" rsync
+          "$STRIP" rsync
+
+      - name: Verify binary
+        shell: bash
+        run: |
+          set -euo pipefail
+          file rsync
+          # Gate: must be a statically-linked executable (no interpreter).
+          file rsync | grep -q "statically linked"
+          if file rsync | grep -q "dynamically linked"; then
+            echo "ERROR: binary is not static" >&2; exit 1
+          fi
+          # Best-effort: confirm it actually runs under qemu-user.
+          ${{ matrix.qemu }} ./rsync --version | head -3 || \
+            echo "WARNING: qemu smoke test did not run cleanly (check on a real device)"
+
+      - name: Package
+        shell: bash
+        run: |
+          set -euo pipefail
+          VER=$(sed -n 's/.*RSYNC_VERSION "\([^"]*\)".*/\1/p' version.h)
+          out="rsync-${VER}-android-${{ matrix.abi }}"
+          mkdir -p dist
+          cp rsync "dist/$out"
+          ( cd dist && sha256sum "$out" > "$out.sha256" )
+          echo "ARTIFACT_NAME=rsync-android-${{ matrix.abi }}" >>"$GITHUB_ENV"
+
+      - name: Upload artifact
+        uses: actions/upload-artifact@v4
+        with:
+          name: ${{ env.ARTIFACT_NAME }}
+          path: dist/

.github/workflows/coverage.yml

@@ -0,0 +1,71 @@
+name: Coverage (Ubuntu)
+
+on:
+  push:
+    branches: [ master ]
+    paths-ignore:
+      - '.github/workflows/*.yml'
+      - '!.github/workflows/coverage.yml'
+  pull_request:
+    branches: [ master ]
+    paths-ignore:
+      - '.github/workflows/*.yml'
+      - '!.github/workflows/coverage.yml'
+  schedule:
+    - cron: '42 9 * * *'
+  workflow_dispatch:
+
+jobs:
+  coverage:
+    runs-on: ubuntu-latest
+    name: gcov coverage
+    steps:
+    - uses: actions/checkout@v4
+      with:
+        fetch-depth: 0
+    - name: prep
+      run: |
+        sudo apt-get update
+        sudo apt-get install -y acl libacl1-dev attr libattr1-dev liblz4-dev libzstd-dev libxxhash-dev python3-cmarkgfm openssl gcovr
+        echo "/usr/local/bin" >>$GITHUB_PATH
+    - name: configure
+      run: ./configure --enable-coverage --with-rrsync
+    - name: make
+      run: make
+    - name: info
+      run: rsync --version
+    # Two coverage runs: the default pipe transport, then a second pass over a
+    # real loopback rsyncd (--use-tcp) which also exercises the require_tcp-only
+    # tests. gcovr's --print-summary line/branch/decision totals go to the step
+    # log (and the job summary below), so the numbers are visible in CI.
+    # `make coverage` exits with the suite's status, so a regression fails CI.
+    - name: coverage (pipe transport)
+      run: |
+        set -o pipefail
+        sudo make coverage 2>&1 | tee cov-pipe.log
+    - name: coverage (TCP transport)
+      run: |
+        set -o pipefail
+        sudo make coverage-tcp 2>&1 | tee cov-tcp.log
+    - name: coverage summary
+      if: always()
+      run: |
+        {
+          echo "## gcov coverage"
+          echo "### Pipe transport (\`make coverage\`)"
+          echo '```'
+          grep -E '^(lines|functions|branches|decisions):' cov-pipe.log || echo '(no summary -- see step log)'
+          echo '```'
+          echo "### TCP transport (\`make coverage-tcp\`)"
+          echo '```'
+          grep -E '^(lines|functions|branches|decisions):' cov-tcp.log || echo '(no summary -- see step log)'
+          echo '```'
+        } >> "$GITHUB_STEP_SUMMARY"
+    - name: upload HTML reports
+      if: always()
+      uses: actions/upload-artifact@v4
+      with:
+        name: coverage-html
+        path: |
+          coverage
+          coverage-tcp

.github/workflows/cygwin-build.yml

@@ -39,7 +39,16 @@ jobs:
     - name: info
       run: bash -c '/usr/local/bin/rsync --version'
     - name: check
-      run: bash -c 'RSYNC_EXPECT_SKIPPED=acls-default,acls,chown,devices,dir-sgid,open-noatime,protected-regular,simd-checksum make check'
+      # chown-fake / devices-fake / xattrs / xattrs-hlink now RUN on Cygwin
+      # (rsyncfns.py drives xattrs via getfattr/setfattr from the `attr`
+      # package installed above), verified on a real Cygwin host. The real
+      # chown/devices tests still skip (need root/mknod), as do the
+      # RESOLVE_BENEATH symlink-race tests.
+      run: bash -c 'RSYNC_EXPECT_SKIPPED=acls-default,acls-depth,acls,bare-do-open-symlink-race,chdir-symlink-race,chown,daemon-access-ip,daemon-chroot-acl,devices,dir-sgid,open-noatime,protected-regular,proxy-response-line-too-long,sender-flist-symlink-leak,simd-checksum,symlink-dirlink-basis make check'
+    - name: check (TCP daemon transport)
+      # Second run with daemon tests over a real loopback rsyncd; the default
+      # 'make check' above uses the secure stdio-pipe transport.
+      run: bash -c './runtests.py --rsync-bin=`pwd`/rsync.exe --use-tcp -j 8'
     - name: ssl file list
       run: bash -c 'PATH="/usr/local/bin:$PATH" rsync-ssl --no-motd download.samba.org::rsyncftp/ || true'
     - name: save artifact

.github/workflows/freebsd-build.yml

@@ -34,6 +34,8 @@ jobs:
           ./configure --with-rrsync -disable-zstd --disable-md2man --disable-xxhash --disable-lz4
           make
           ./rsync --version
+          make check
+          ./runtests.py --rsync-bin=`pwd`/rsync --use-tcp -j 8
           ./rsync-ssl --no-motd download.samba.org::rsyncftp/ || true
     - name: save artifact
       uses: actions/upload-artifact@v4

.github/workflows/macos-build.yml

@@ -41,7 +41,14 @@ jobs:
     - name: info
       run: rsync --version
     - name: check
-      run: sudo RSYNC_EXPECT_SKIPPED=acls-default,chmod-temp-dir,chown-fake,devices-fake,dir-sgid,open-noatime,protected-regular,simd-checksum,xattrs-hlink,xattrs make check
+      # chown-fake / devices-fake / xattrs / xattrs-hlink now RUN on macOS
+      # (rsyncfns.py drives xattrs via the `xattr` command), verified on a
+      # real macOS host, so they're no longer in the skip set.
+      run: sudo RSYNC_EXPECT_SKIPPED=acls-default,acls-depth,chmod-temp-dir,daemon-access-ip,daemon-chroot-acl,dir-sgid,open-noatime,preallocate,protected-regular,proxy-response-line-too-long,simd-checksum,sparse make check
+    - name: check (TCP daemon transport)
+      # Second run with daemon tests over a real loopback rsyncd; the default
+      # 'make check' above uses the secure stdio-pipe transport.
+      run: sudo ./runtests.py --rsync-bin=`pwd`/rsync --use-tcp -j 8
     - name: ssl file list
       run: rsync-ssl --no-motd download.samba.org::rsyncftp/ || true
     - name: save artifact

.github/workflows/netbsd-build.yml

@@ -0,0 +1,52 @@
+name: Test rsync on NetBSD
+
+on:
+  push:
+    branches: [ master ]
+    paths-ignore:
+      - '.github/workflows/*.yml'
+      - '!.github/workflows/netbsd-build.yml'
+  pull_request:
+    branches: [ master ]
+    paths-ignore:
+      - '.github/workflows/*.yml'
+      - '!.github/workflows/netbsd-build.yml'
+  schedule:
+    - cron: '42 8 * * *'
+
+jobs:
+  test:
+    runs-on: ubuntu-latest
+    name: Test rsync on NetBSD
+    steps:
+    - uses: actions/checkout@v4
+      with:
+        fetch-depth: 0
+    - name: Test in NetBSD VM
+      id: test
+      uses: vmactions/netbsd-vm@v1
+      with:
+        usesh: true
+        prepare: |
+          PATH=/usr/sbin:$PATH pkg_add autoconf automake python312
+          ln -sf /usr/pkg/bin/python3.12 /usr/pkg/bin/python3
+        run: |
+          uname -a
+          ./configure --with-rrsync --disable-zstd --disable-md2man --disable-xxhash --disable-lz4
+          make
+          ./rsync --version
+          make check
+          ./runtests.py --rsync-bin=`pwd`/rsync --use-tcp -j 8
+          ./rsync-ssl --no-motd download.samba.org::rsyncftp/ || true
+    - name: save artifact
+      uses: actions/upload-artifact@v4
+      with:
+        name: netbsd-bin
+        path: |
+          rsync
+          rsync-ssl
+          rsync.1
+          rsync-ssl.1
+          rsyncd.conf.5
+          rrsync.1
+          rrsync

.github/workflows/openbsd-build.yml

@@ -0,0 +1,61 @@
+name: Test rsync on OpenBSD
+
+on:
+  push:
+    branches: [ master ]
+    paths-ignore:
+      - '.github/workflows/*.yml'
+      - '!.github/workflows/openbsd-build.yml'
+  pull_request:
+    branches: [ master ]
+    paths-ignore:
+      - '.github/workflows/*.yml'
+      - '!.github/workflows/openbsd-build.yml'
+  schedule:
+    - cron: '42 8 * * *'
+
+jobs:
+  test:
+    runs-on: ubuntu-latest
+    name: Test rsync on OpenBSD
+    steps:
+    - uses: actions/checkout@v4
+      with:
+        fetch-depth: 0
+    - name: Test in OpenBSD VM
+      id: test
+      uses: vmactions/openbsd-vm@v1
+      with:
+        usesh: true
+        prepare: |
+          pkg_add -I bash autoconf%2.71 automake%1.16
+        run: |
+          uname -a
+          export AUTOCONF_VERSION=2.71
+          export AUTOMAKE_VERSION=1.16
+          ./configure --with-rrsync --disable-zstd --disable-md2man --disable-xxhash --disable-lz4
+          make
+          ./rsync --version
+          make check
+          # The --use-tcp daemon tests run at -j2 here (vs -j8 elsewhere): this
+          # job runs inside a nested VM, and at -j8 the many concurrent loopback
+          # daemons occasionally lose a connection-handshake timing race under
+          # that resource pressure, hanging one test to the 300s timeout. It is
+          # an environment artifact, not an rsync bug (the handshake is
+          # deadlock-free and unreproducible elsewhere, even pinned to 1 CPU at
+          # -j8); -j2 keeps the VM from over-subscribing. The pipe `make check`
+          # above stays at the default parallelism.
+          ./runtests.py --rsync-bin=`pwd`/rsync --use-tcp -j 2
+          ./rsync-ssl --no-motd download.samba.org::rsyncftp/ || true
+    - name: save artifact
+      uses: actions/upload-artifact@v4
+      with:
+        name: openbsd-bin
+        path: |
+          rsync
+          rsync-ssl
+          rsync.1
+          rsync-ssl.1
+          rsyncd.conf.5
+          rrsync.1
+          rrsync

.github/workflows/solaris-build.yml

@@ -34,6 +34,8 @@ jobs:
           ./configure --with-rrsync -disable-zstd --disable-md2man --disable-xxhash --disable-lz4
           make
           ./rsync --version
+          make check
+          ./runtests.py --rsync-bin=`pwd`/rsync --use-tcp -j 8
           ./rsync-ssl --no-motd download.samba.org::rsyncftp/ || true
     - name: save artifact
       uses: actions/upload-artifact@v4

.github/workflows/ubuntu-22.04-build.yml

@@ -0,0 +1,64 @@
+name: Test rsync on Ubuntu 22.04
+
+# Older-LTS coverage to help with backporting security fixes. ubuntu-22.04
+# is currently the oldest GitHub Actions runner image (20.04 was retired
+# in April 2025).
+
+on:
+  push:
+    branches: [ master ]
+    paths-ignore:
+      - '.github/workflows/*.yml'
+      - '!.github/workflows/ubuntu-22.04-build.yml'
+  pull_request:
+    branches: [ master ]
+    paths-ignore:
+      - '.github/workflows/*.yml'
+      - '!.github/workflows/ubuntu-22.04-build.yml'
+  schedule:
+    - cron: '42 8 * * *'
+
+jobs:
+  test:
+    runs-on: ubuntu-22.04
+    name: Test rsync on Ubuntu 22.04
+    steps:
+    - uses: actions/checkout@v4
+      with:
+        fetch-depth: 0
+    - name: prep
+      run: |
+        sudo apt-get install acl libacl1-dev attr libattr1-dev liblz4-dev libzstd-dev libxxhash-dev python3-cmarkgfm openssl
+        echo "/usr/local/bin" >>$GITHUB_PATH
+    - name: configure
+      run: ./configure --with-rrsync
+    - name: make
+      run: make
+    - name: install
+      run: sudo make install
+    - name: info
+      run: rsync --version
+    - name: check
+      run: sudo RSYNC_EXPECT_SKIPPED=crtimes,daemon-access-ip,daemon-chroot-acl,proxy-response-line-too-long make check
+    - name: check30
+      run: sudo RSYNC_EXPECT_SKIPPED=crtimes,daemon-access-ip,daemon-chroot-acl,proxy-response-line-too-long make check30
+    - name: check29
+      run: sudo RSYNC_EXPECT_SKIPPED=crtimes,daemon-access-ip,daemon-chroot-acl,proxy-response-line-too-long make check29
+    - name: check (TCP daemon transport)
+      # Second run with daemon tests over a real loopback rsyncd; the default
+      # 'make check' above uses the secure stdio-pipe transport.
+      run: sudo ./runtests.py --rsync-bin=`pwd`/rsync --use-tcp -j 8
+    - name: ssl file list
+      run: rsync-ssl --no-motd download.samba.org::rsyncftp/ || true
+    - name: save artifact
+      uses: actions/upload-artifact@v4
+      with:
+        name: ubuntu-22.04-bin
+        path: |
+          rsync
+          rsync-ssl
+          rsync.1
+          rsync-ssl.1
+          rsyncd.conf.5
+          rrsync.1
+          rrsync

.github/workflows/ubuntu-build.yml

@@ -35,11 +35,17 @@ jobs:
     - name: info
       run: rsync --version
     - name: check
-      run: sudo RSYNC_EXPECT_SKIPPED=crtimes make check
+      run: sudo RSYNC_EXPECT_SKIPPED=crtimes,daemon-access-ip,daemon-chroot-acl,proxy-response-line-too-long make check
     - name: check30
-      run: sudo RSYNC_EXPECT_SKIPPED=crtimes make check30
+      run: sudo RSYNC_EXPECT_SKIPPED=crtimes,daemon-access-ip,daemon-chroot-acl,proxy-response-line-too-long make check30
     - name: check29
-      run: sudo RSYNC_EXPECT_SKIPPED=crtimes make check29
+      run: sudo RSYNC_EXPECT_SKIPPED=crtimes,daemon-access-ip,daemon-chroot-acl,proxy-response-line-too-long make check29
+    - name: check (TCP daemon transport)
+      # Second run with daemon tests over a real loopback rsyncd. The default
+      # 'make check' above uses the secure stdio-pipe transport (no listening
+      # sockets); this run exercises the real TCP accept/auth path. Skip-set
+      # is env-dependent here (chroot-acl), so leave RSYNC_EXPECT_SKIPPED unset.
+      run: sudo ./runtests.py --rsync-bin=`pwd`/rsync --use-tcp -j 8
     - name: ssl file list
       run: rsync-ssl --no-motd download.samba.org::rsyncftp/ || true
     - name: save artifact

INSTALL.md

@@ -7,6 +7,22 @@ option to use if you want to just skip that feature.  What follows are various
 support libraries that you may want to install to build rsync with the maximum
 features (the impatient can skip down to the package summary):
 
+## Ubuntu users: skip the build, use the PPA
+
+If you are on a currently supported Ubuntu series (jammy 22.04 LTS, noble
+24.04 LTS, questing 25.10, resolute 26.04 LTS) and just want the latest
+upstream rsync, the rsync project maintains a Launchpad PPA that tracks
+stable releases:
+
+>     sudo add-apt-repository ppa:rsyncproject/rsync
+>     sudo apt update && sudo apt install rsync
+
+See [the PPA page][ppa] for current build status across architectures.
+
+[ppa]: https://launchpad.net/~rsyncproject/+archive/ubuntu/rsync
+
+The rest of this document covers building from source.
+
 ## The basic setup
 
 You need to have a C compiler installed and optionally a C++ compiler in order

Makefile.in

@@ -57,13 +57,14 @@ TLS_OBJ = tls.o syscall.o util2.o t_stub.o lib/compat.o lib/snprintf.o lib/perms
 
 # Programs we must have to run the test cases
 CHECK_PROGS = rsync$(EXEEXT) tls$(EXEEXT) getgroups$(EXEEXT) getfsdev$(EXEEXT) \
-	testrun$(EXEEXT) trimslash$(EXEEXT) t_unsafe$(EXEEXT) wildtest$(EXEEXT) \
+	testrun$(EXEEXT) trimslash$(EXEEXT) t_unsafe$(EXEEXT) t_chmod_secure$(EXEEXT) \
-	simdtest$(EXEEXT)
+	t_secure_relpath$(EXEEXT) wildtest$(EXEEXT) simdtest$(EXEEXT)
 
-CHECK_SYMLINKS = testsuite/chown-fake.test testsuite/devices-fake.test testsuite/xattrs-hlink.test
+CHECK_SYMLINKS = testsuite/chown-fake_test.py testsuite/devices-fake_test.py \
+	testsuite/xattrs-hlink_test.py testsuite/exclude-lsh_test.py
 
 # Objects for CHECK_PROGS to clean
-CHECK_OBJS=tls.o testrun.o getgroups.o getfsdev.o t_stub.o t_unsafe.o trimslash.o wildtest.o
+CHECK_OBJS=tls.o testrun.o getgroups.o getfsdev.o t_stub.o t_unsafe.o t_chmod_secure.o t_secure_relpath.o trimslash.o wildtest.o
 
 # note that the -I. is needed to handle config.h when using VPATH
 .c.o:
@@ -179,6 +180,14 @@ T_UNSAFE_OBJ = t_unsafe.o syscall.o util1.o util2.o t_stub.o lib/compat.o lib/sn
 t_unsafe$(EXEEXT): $(T_UNSAFE_OBJ)
 	$(CC) $(CFLAGS) $(LDFLAGS) -o $@ $(T_UNSAFE_OBJ) $(LIBS)
 
+T_CHMOD_SECURE_OBJ = t_chmod_secure.o syscall.o util1.o util2.o t_stub.o lib/compat.o lib/snprintf.o lib/wildmatch.o lib/permstring.o
+t_chmod_secure$(EXEEXT): $(T_CHMOD_SECURE_OBJ)
+	$(CC) $(CFLAGS) $(LDFLAGS) -o $@ $(T_CHMOD_SECURE_OBJ) $(LIBS)
+
+T_SECURE_RELPATH_OBJ = t_secure_relpath.o syscall.o util1.o util2.o t_stub.o lib/compat.o lib/snprintf.o lib/wildmatch.o lib/permstring.o
+t_secure_relpath$(EXEEXT): $(T_SECURE_RELPATH_OBJ)
+	$(CC) $(CFLAGS) $(LDFLAGS) -o $@ $(T_SECURE_RELPATH_OBJ) $(LIBS)
+
 .PHONY: conf
 conf: configure.sh config.h.in
 
@@ -271,6 +280,8 @@ clean: cleantests
 	rm -f *~ $(OBJS) $(CHECK_PROGS) $(CHECK_OBJS) $(CHECK_SYMLINKS) @MAKE_RRSYNC@ \
 		git-version.h rounding rounding.h *.old rsync*.1 rsync*.5 @MAKE_RRSYNC_1@ \
 		*.html daemon-parm.h help-*.h default-*.h proto.h proto.h-tstamp
+	rm -f *.gcno *.gcda lib/*.gcno lib/*.gcda zlib/*.gcno zlib/*.gcda popt/*.gcno popt/*.gcda
+	rm -rf coverage coverage-tcp coverage-all coverage-fallback
 
 .PHONY: cleantests
 cleantests:
@@ -311,17 +322,109 @@ test: check
 # catch Bash-isms earlier even if we're running on GNU.  Of course, we
 # might lose in the future where POSIX diverges from old sh.
 
+# `make check` runs tests in parallel by default. Override with
+# `make check CHECK_J=1` (serial) or any other value.
+CHECK_J = 8
+
+# Parallelism for `make coverage`. Defaults to the same as CHECK_J: the
+# coverage build sets -fprofile-update=atomic (atomic in-memory counters) and
+# gcc's libgcov serializes the per-source .gcda read-modify-write merge with a
+# file lock, so concurrent rsync processes (incl. the forked sender/generator/
+# receiver) accumulate exactly -- verified by a count-linearity check (a hot
+# line accumulates identically at -j1 and -P16). Override with
+# `make coverage COVERAGE_J=1` if your libgcov does not lock .gcda merges.
+COVERAGE_J = $(CHECK_J)
+
+# Output directory and extra runtests.py flags for `make coverage`. The
+# `coverage-tcp` target reuses the coverage recipe with --use-tcp (real
+# loopback rsyncd, which exercises the TCP accept/auth path and the
+# require_tcp-only tests) and a separate output directory.
+COVERAGE_DIR = coverage
+COVERAGE_RUNFLAGS =
+
+# Bundled third-party code that rsync ships but does not own; excluded from the
+# coverage report so the percentages reflect rsync's own source. zlib/ and popt/
+# are wholly vendored; the named lib/ files are PostgreSQL (getaddrinfo) and ISC
+# (inet_ntop/inet_pton) / standalone (getpass) imports. The other lib/*.c
+# (md5, mdfour, wildmatch, permstring, pool_alloc, snprintf, sysacls, sysxattrs,
+# compat) are rsync's own and stay in the report.
+COVERAGE_EXCLUDE = -e '(^|/)zlib/' -e '(^|/)popt/' \
+	-e '(^|/)lib/(getaddrinfo|getpass|inet_ntop|inet_pton)\.'
+
 .PHONY: check
 check: all $(CHECK_PROGS) $(CHECK_SYMLINKS)
-	$(srcdir)/runtests.py --rsync-bin=`pwd`/rsync$(EXEEXT)
+	$(srcdir)/runtests.py --rsync-bin=`pwd`/rsync$(EXEEXT) -j $(CHECK_J)
 
 .PHONY: check29
 check29: all $(CHECK_PROGS) $(CHECK_SYMLINKS)
-	$(srcdir)/runtests.py --rsync-bin=`pwd`/rsync$(EXEEXT) --protocol=29
+	$(srcdir)/runtests.py --rsync-bin=`pwd`/rsync$(EXEEXT) -j $(CHECK_J) --protocol=29
 
 .PHONY: check30
 check30: all $(CHECK_PROGS) $(CHECK_SYMLINKS)
-	$(srcdir)/runtests.py --rsync-bin=`pwd`/rsync$(EXEEXT) --protocol=30
+	$(srcdir)/runtests.py --rsync-bin=`pwd`/rsync$(EXEEXT) -j $(CHECK_J) --protocol=30
+
+# Whole-suite gcov coverage report (HTML, with branch + decision coverage).
+# Requires a build configured with --enable-coverage and the `gcovr` tool
+# (pip install gcovr). Runs the suite in parallel (COVERAGE_J, default CHECK_J):
+# this is safe because the coverage build uses -fprofile-update=atomic and
+# libgcov locks the per-source .gcda during its merge, so concurrent rsync
+# processes accumulate exactly (see COVERAGE_J above). Use COVERAGE_J=1 if your
+# toolchain's libgcov does not lock .gcda merges.
+.PHONY: coverage
+coverage: all $(CHECK_PROGS) $(CHECK_SYMLINKS)
+	@case '$(CFLAGS)' in *--coverage*) ;; \
+	  *) echo "*** not a coverage build; reconfigure with --enable-coverage"; exit 1 ;; esac
+	@command -v gcovr >/dev/null 2>&1 || { echo "*** gcovr not found (pip install gcovr)"; exit 1; }
+	find . -name '*.gcda' -delete
+	@rc=0; $(srcdir)/runtests.py --rsync-bin=`pwd`/rsync$(EXEEXT) -j $(COVERAGE_J) $(COVERAGE_RUNFLAGS) || rc=$$?; \
+	  rm -rf $(COVERAGE_DIR) && mkdir -p $(COVERAGE_DIR); \
+	  gcovr --root $(srcdir) $(COVERAGE_EXCLUDE) --decisions --print-summary \
+	      --html-details -o $(COVERAGE_DIR)/index.html . || exit $$?; \
+	  echo "Coverage report written to $(COVERAGE_DIR)/index.html"; \
+	  if test $$rc != 0; then \
+	    echo "*** test suite FAILED (status $$rc) -- coverage report still written above"; \
+	  fi; \
+	  exit $$rc
+
+# Same as `make coverage` but with the daemon tests run over a real loopback
+# rsyncd (--use-tcp), into a separate report directory.
+.PHONY: coverage-tcp
+coverage-tcp:
+	$(MAKE) coverage COVERAGE_RUNFLAGS=--use-tcp COVERAGE_DIR=coverage-tcp
+
+# Comprehensive single report: run the suite under several configurations,
+# accumulating into the shared .gcda counters (NOT cleared between runs), then
+# emit one merged, rsync-scoped report. Covers the default (pipe) transport, the
+# protocol-29/30 compat branches, and the real-TCP daemon path (which also runs
+# the require_tcp-only tests). Run under sudo to additionally cover root-only
+# paths (devices, chown, use-chroot, protected-regular). Local target -- CI uses
+# the plain `coverage`/`coverage-tcp` targets.
+.PHONY: coverage-all
+coverage-all: all $(CHECK_PROGS) $(CHECK_SYMLINKS)
+	@case '$(CFLAGS)' in *--coverage*) ;; \
+	  *) echo "*** not a coverage build; reconfigure with --enable-coverage"; exit 1 ;; esac
+	@command -v gcovr >/dev/null 2>&1 || { echo "*** gcovr not found (pip install gcovr)"; exit 1; }
+	find . -name '*.gcda' -delete
+	@rc=0; \
+	  for cfg in '' '--protocol=30' '--protocol=29' '--use-tcp'; do \
+	    echo "===== coverage-all: runtests.py $$cfg ====="; \
+	    $(srcdir)/runtests.py --rsync-bin=`pwd`/rsync$(EXEEXT) -j $(COVERAGE_J) $$cfg || rc=$$?; \
+	  done; \
+	  rm -rf coverage-all && mkdir -p coverage-all; \
+	  gcovr --root $(srcdir) $(COVERAGE_EXCLUDE) --decisions --print-summary \
+	      --html-details -o coverage-all/index.html . || exit $$?; \
+	  echo "Merged coverage report written to coverage-all/index.html"; \
+	  if test $$rc != 0; then \
+	    echo "*** some suite runs FAILED (status $$rc) -- report still written above"; \
+	  fi; \
+	  exit $$rc
+
+# Coverage for the portable (non-openat2) resolver tier. Requires a SEPARATE
+# build configured with --enable-coverage --disable-openat2: its .gcno differ
+# from the openat2 build, so this report cannot be merged with the others.
+.PHONY: coverage-fallback
+coverage-fallback:
+	$(MAKE) coverage COVERAGE_DIR=coverage-fallback
 
 wildtest.o: wildtest.c t_stub.o lib/wildmatch.c rsync.h config.h
 wildtest$(EXEEXT): wildtest.o lib/compat.o lib/snprintf.o @BUILD_POPT@
@@ -335,14 +438,17 @@ simdtest$(EXEEXT): simd-checksum-x86_64.cpp $(HEADERS)
 	    touch $@; \
 	fi
 
-testsuite/chown-fake.test:
+testsuite/chown-fake_test.py:
-	ln -s chown.test $(srcdir)/testsuite/chown-fake.test
+	ln -s chown_test.py $(srcdir)/testsuite/chown-fake_test.py
 
-testsuite/devices-fake.test:
+testsuite/devices-fake_test.py:
-	ln -s devices.test $(srcdir)/testsuite/devices-fake.test
+	ln -s devices_test.py $(srcdir)/testsuite/devices-fake_test.py
 
-testsuite/xattrs-hlink.test:
+testsuite/xattrs-hlink_test.py:
-	ln -s xattrs.test $(srcdir)/testsuite/xattrs-hlink.test
+	ln -s xattrs_test.py $(srcdir)/testsuite/xattrs-hlink_test.py
+
+testsuite/exclude-lsh_test.py:
+	ln -s exclude_test.py $(srcdir)/testsuite/exclude-lsh_test.py
 
 # This does *not* depend on building or installing: you can use it to
 # check a version installed from a binary or some other source tree,
@@ -350,7 +456,7 @@ testsuite/xattrs-hlink.test:
 
 .PHONY: installcheck
 installcheck: $(CHECK_PROGS) $(CHECK_SYMLINKS)
-	$(srcdir)/runtests.py --rsync-bin="$(bindir)/rsync$(EXEEXT)" --srcdir="$(srcdir)" --tooldir=`pwd`
+	$(srcdir)/runtests.py --rsync-bin="$(bindir)/rsync$(EXEEXT)" --srcdir="$(srcdir)" --tooldir=`pwd` -j $(CHECK_J)
 
 # TODO: Add 'dist' target; need to know which files will be included
 

NEWS.md

@@ -1,3 +1,192 @@
+# NEWS for rsync 3.4.3 (20 May 2026)
+
+## Changes in this version:
+
+### SECURITY FIXES:
+
+Six CVEs are fixed in this release.  All six are assigned by
+VulnCheck as CNA.  Affected versions are 3.4.2 and earlier in every
+case.  Three of the six (CVE-2026-29518, CVE-2026-43617,
+CVE-2026-43619) require non-default daemon configuration to reach:
+the first and third need `use chroot = no` for a module, the second
+needs `daemon chroot = ...` set in rsyncd.conf.  Two (CVE-2026-43618,
+CVE-2026-43620) are reachable from a normal pull or a normal
+authenticated daemon connection.  The sixth (CVE-2026-45232) is
+reachable only when `RSYNC_PROXY` is set and the proxy (or a MITM)
+returns a pathological response.  Many thanks to the external
+researchers who reported these issues.
+
+- CVE-2026-29518 (CVSS v4.0 7.3, HIGH): TOCTOU symlink race condition
+  allowing local privilege escalation in daemon mode without chroot.
+  An rsync daemon configured with "use chroot = no" was exposed to a
+  time-of-check / time-of-use race on parent path components: a local
+  attacker with write access to a module could replace a parent
+  directory component with a symlink between the receiver's check and
+  its open(), redirecting reads (basis-file disclosure) and writes
+  (file overwrite) outside the module.  Default "use chroot = yes" is
+  not exposed.  `secure_relative_open()` (added in 3.4.0 for
+  CVE-2024-12086) was previously unused in the daemon-no-chroot
+  case; the fix enables it there and reroutes the sender's
+  read-path opens through it.  Reported by Nullx3D (Batuhan Sancak),
+  Damien Neil and Michael Stapelberg.
+
+- CVE-2026-43617 (CVSS v3.1 4.8, MEDIUM): Hostname/ACL bypass on an
+  rsync daemon configured with `daemon chroot = /X` in rsyncd.conf
+  when the chroot tree lacks DNS resolution support.  The
+  reverse-DNS lookup of the connecting client was performed *after*
+  the daemon chroot had been entered; if /X did not contain the
+  libc resolver fixtures (`/etc/resolv.conf`, `/etc/nsswitch.conf`,
+  `/etc/hosts`, NSS service modules) the lookup failed and the
+  connecting hostname was set to "UNKNOWN", causing hostname-based
+  deny rules to silently fail open.  IP-based ACLs are unaffected.
+  The per-module `use chroot` setting is unrelated to this issue.
+  The fix performs the lookup before entering the daemon chroot.
+  Reported by MegaManSec.
+
+- CVE-2026-43618 (CVSS v3.1 8.1, HIGH): Integer overflow in the
+  compressed-token decoder enabling remote memory disclosure to an
+  authenticated daemon peer.  The receiver accumulated a 32-bit
+  signed counter without overflow checking; a malicious sender could
+  trigger an overflow that, with careful manipulation, leaked process
+  memory contents to the attacker -- environment variables,
+  passwords, heap and library pointers -- significantly weakening
+  ASLR.  The fix bounds the counter and adds wire-input validation in
+  several adjacent places (defence-in-depth).  Workaround for older
+  releases: `refuse options = compress` in rsyncd.conf.  Reported by
+  Omar Elsayed.
+
+- CVE-2026-43619 (CVSS v3.1 6.3, MEDIUM): Symlink races on path-based
+  system calls in "use chroot = no" daemon mode (generalisation of
+  CVE-2026-29518).  Earlier fixes for symlink races on the receiver's
+  open() call missed the same race class on every other path-based
+  system call: chmod, lchown, utimes, rename, unlink, mkdir, symlink,
+  mknod, link, rmdir and lstat.  The fix routes each affected
+  path-based syscall through a parent dirfd opened under
+  RESOLVE_BENEATH-equivalent kernel-enforced confinement (openat2 on
+  Linux 5.6+, O_RESOLVE_BENEATH on FreeBSD 13+ and macOS 15+,
+  per-component O_NOFOLLOW walk elsewhere).  Default "use chroot =
+  yes" is not exposed.  Reported by Andrew Tridgell as a follow-on
+  audit of CVE-2026-29518.
+
+- CVE-2026-43620 (CVSS v3.1 6.5, MEDIUM): Out-of-bounds read in the
+  receiver's recv_files() enabling remote denial-of-service of any
+  client pulling from a malicious server (incomplete fix of commit
+  797e17f).  The earlier parent_ndx<0 guard added to send_files() was
+  not applied to the visually-identical block in recv_files().  A
+  malicious rsync server can drive any connecting client into a
+  deterministic SIGSEGV by setting CF_INC_RECURSE in the
+  compatibility flags and sending a crafted file list and transfer
+  record.  inc_recurse is the protocol-30+ default, so no special
+  options are required on the victim.  Workaround for older
+  releases: `--no-inc-recursive` on the client.  Reported by Pratham
+  Gupta.
+
+- CVE-2026-45232 (CVSS v3.1 3.1, LOW): Off-by-one out-of-bounds stack
+  write in the rsync client's HTTP CONNECT proxy handler
+  (`establish_proxy_connection()` in `socket.c`).  After issuing the
+  CONNECT request, rsync read the proxy's first response line one
+  byte at a time into a 1024-byte stack buffer with the bound
+  `cp < &buffer[sizeof buffer - 1]`.  If the proxy (or a MITM in
+  front of it) returned 1023+ bytes on that first line without a
+  newline terminator, `cp` exited the loop pointing at a buffer slot
+  the loop never wrote, leaving `*cp` holding stale stack data from
+  the earlier `snprintf()` of the outgoing CONNECT request.  The
+  post-loop logic then wrote a single `\0` one byte past the end of
+  the buffer on the stack.  Reach is client-side only, and only when
+  `RSYNC_PROXY` is set so rsync tunnels an `rsync://` connection
+  through an HTTP CONNECT proxy.  The written byte is always `\0`
+  and the offset is fixed by the buffer size, not attacker-chosen,
+  so this is not an arbitrary-write primitive: practical impact is
+  corruption of one adjacent stack byte and possible later
+  misbehaviour or crash.  The fix detects the "buffer filled without
+  finding `\n`" case explicitly by position and refuses the response
+  with "proxy response line too long".  Reported by Aisle Research
+  via Michal Ruprich (rsync-3.4.1-2.el10 QE).
+
+In addition to the six CVE fixes, this release adds defence-in-depth
+hardening on several adjacent paths: bounded wire-supplied counts and
+lengths in flist/io/acls/xattrs, a guard against length underflow in
+cumulative `snprintf()` callers, a parent block-index bounds check on
+the receiver, a NULL check in `read_delay_line()`, a lower ceiling on
+`MAX_WIRE_DEL_STAT` to avoid signed-int overflow in the
+`read_del_stats()` accumulator, rejection of hyphen-prefixed
+remote-shell hostnames (defence-in-depth against argv-injection in
+tooling that forwards untrusted input into the hostspec position;
+reported by Aisle Research via Michal Ruprich), and a NULL-check on
+`localtime_r()` in `timestring()` to keep a malicious server from
+crashing the client by advertising a file with an out-of-range
+modtime.
+
+### BUG FIXES:
+
+- Fixed a regression introduced by the 3.4.0 secure_relative_open()
+  CVE fix where legitimate directory symlinks on the receiver side
+  (e.g. when using `-K` / `--copy-dirlinks`) caused "failed
+  verification -- update discarded" errors on delta transfers. The
+  old code rejected every symlink in the path with a per-component
+  `O_NOFOLLOW` walk; the receiver now uses kernel-enforced "stay
+  below dirfd" path resolution where available. Fixes #715.
+
+### PORTABILITY / BUILD:
+
+- secure_relative_open() now uses `openat2(RESOLVE_BENEATH |
+  RESOLVE_NO_MAGICLINKS)` on Linux 5.6+, and `openat()` with
+  `O_RESOLVE_BENEATH` on FreeBSD 13+ and macOS 15+ (Sequoia) /
+  iOS 18+. The kernel rejects ".." escapes, absolute symlinks, and
+  symlinks whose target lies outside the starting directory, while
+  still following symlinks that resolve within it -- the same
+  trade-off that fixes the issue #715 regression without weakening
+  the original CVE protection. Other platforms (Solaris, OpenBSD,
+  NetBSD, Cygwin) retain the previous per-component `O_NOFOLLOW`
+  walk; on those platforms the issue #715 regression remains
+  visible.
+
+- testsuite/xattrs: ignore `SUNWattr_*` in the Solaris `xls`
+  helper.
+
+### DEVELOPER RELATED:
+
+- Added testsuite/symlink-dirlink-basis.test (taken from PR #864
+  by Samuel Henrique) covering the issue #715 regression and
+  several edge cases (`--backup`, `--inplace`, `--partial-dir`
+  with protocol < 29, top-level files). The test skips on
+  platforms without a RESOLVE_BENEATH equivalent.
+
+- Added regression tests for the new security fixes:
+  `chmod-symlink-race.test`, `chdir-symlink-race.test`,
+  `bare-do-open-symlink-race.test`, `alt-dest-symlink-race.test`,
+  `copy-dest-source-symlink.test`, `sender-flist-symlink-leak.test`,
+  `secure-relpath-validation.test`, `daemon-chroot-acl.test` and
+  `daemon-refuse-compress.test`. The symlink-race tests skip on
+  Cygwin, Solaris, OpenBSD and NetBSD (no RESOLVE_BENEATH
+  equivalent on those platforms).
+
+- runtests.py now errors early with a clear message when any of
+  the test helper programs (`tls`, `trimslash`, `t_unsafe`,
+  `t_chmod_secure`, `t_secure_relpath`, `wildtest`, `getgroups`,
+  `getfsdev`) are missing, instead of letting many tests fail with
+  confusing "not found" errors.
+
+- Added OpenBSD and NetBSD CI jobs that run `make check` on those
+  platforms.
+
+- Added Ubuntu 22.04 and AlmaLinux 8 CI workflows so future
+  backports to the two mainstream LTS families build and test on
+  the same CI surface as trunk.
+
+- testsuite/protected-regular.test now runs unprivileged via
+  `unshare` with user-namespace UID mapping, falling back to skip
+  if `unshare`/`uidmap` is not available; previously it required
+  real root.
+
+- Added `symlink-dirlink-basis` to the Cygwin CI's expected-skipped
+  list.
+
+- Removed the old release system (replaced by the new release
+  script in 3.4.2).
+
+------------------------------------------------------------------------------
+
 # NEWS for rsync 3.4.2 (28 Apr 2026)
 
 ## Changes in this version:
@@ -4980,6 +5169,7 @@ to develop and test fixes.
 
 | RELEASE DATE | VER.   | DATE OF COMMIT\* | PROTOCOL    |
 |--------------|--------|------------------|-------------|
+| 20 May 2026  | 3.4.3  |                  | 32          |
 | 28 Apr 2026  | 3.4.2  |                  | 32          |
 | 16 Jan 2025  | 3.4.1  |                  | 32          |
 | 15 Jan 2025  | 3.4.0  | 15 Jan 2025      | 32          |

README.md

@@ -93,6 +93,15 @@ details.
 [3]: https://rsync.samba.org/lists.html
 
 
+DISCORD
+-------
+
+There is also an rsync [Discord server][d] for real-time chat about rsync
+and its development.
+
+[d]: https://discord.gg/Avfvy9zhdp
+
+
 BUG REPORTS
 -----------
 

acls.c

@@ -697,7 +697,7 @@ static uint32 recv_acl_access(int f, uchar *name_follows_ptr)
 static uchar recv_ida_entries(int f, ida_entries *ent)
 {
 	uchar computed_mask_bits = 0;
-	int i, count = read_varint(f);
+	int i, count = read_varint_bounded(f, 0, MAX_WIRE_ACL_COUNT, "ACL count");
 
 	ent->idas = count ? new_array(id_access, count) : NULL;
 	ent->count = count;

backup.c

@@ -39,7 +39,7 @@ static int validate_backup_dir(void)
 {
 	STRUCT_STAT st;
 
-	if (do_lstat(backup_dir_buf, &st) < 0) {
+	if (do_lstat_at(backup_dir_buf, &st) < 0) {
 		if (errno == ENOENT)
 			return 0;
 		rsyserr(FERROR, errno, "backup lstat %s failed", backup_dir_buf);
@@ -98,7 +98,7 @@ static BOOL copy_valid_path(const char *fname)
 	for ( ; b; name = b + 1, b = strchr(name, '/')) {
 		*b = '\0';
 
-		while (do_mkdir(backup_dir_buf, ACCESSPERMS) < 0) {
+		while (do_mkdir_at(backup_dir_buf, ACCESSPERMS) < 0) {
 			if (errno == EEXIST) {
 				val = validate_backup_dir();
 				if (val > 0)
@@ -197,7 +197,7 @@ static inline int link_or_rename(const char *from, const char *to,
 		if (IS_SPECIAL(stp->st_mode) || IS_DEVICE(stp->st_mode))
 			return 0; /* Use copy code. */
 #endif
-		if (do_link(from, to) == 0) {
+		if (do_link_at(from, to) == 0) {
 			if (DEBUG_GTE(BACKUP, 1))
 				rprintf(FINFO, "make_backup: HLINK %s successful.\n", from);
 			return 2;
@@ -207,7 +207,7 @@ static inline int link_or_rename(const char *from, const char *to,
 			return 0;
 	}
 #endif
-	if (do_rename(from, to) == 0) {
+	if (do_rename_at(from, to) == 0) {
 		if (stp->st_nlink > 1 && !S_ISDIR(stp->st_mode)) {
 			/* If someone has hard-linked the file into the backup
 			 * dir, rename() might return success but do nothing! */
@@ -246,7 +246,7 @@ int make_backup(const char *fname, BOOL prefer_rename)
 		goto success;
 	if (errno == EEXIST || errno == EISDIR) {
 		STRUCT_STAT bakst;
-		if (do_lstat(buf, &bakst) == 0) {
+		if (do_lstat_at(buf, &bakst) == 0) {
 			int flags = get_del_for_flag(bakst.st_mode) | DEL_FOR_BACKUP | DEL_RECURSE;
 			if (delete_item(buf, bakst.st_mode, flags) != 0)
 				return 0;
@@ -277,7 +277,7 @@ int make_backup(const char *fname, BOOL prefer_rename)
 	/* Check to see if this is a device file, or link */
 	if ((am_root && preserve_devices && IS_DEVICE(file->mode))
 	 || (preserve_specials && IS_SPECIAL(file->mode))) {
-		if (do_mknod(buf, file->mode, sx.st.st_rdev) < 0)
+		if (do_mknod_at(buf, file->mode, sx.st.st_rdev) < 0)
 			rsyserr(FERROR, errno, "mknod %s failed", full_fname(buf));
 		else if (DEBUG_GTE(BACKUP, 1))
 			rprintf(FINFO, "make_backup: DEVICE %s successful.\n", fname);
@@ -294,7 +294,7 @@ int make_backup(const char *fname, BOOL prefer_rename)
 			}
 			ret = 2;
 		} else {
-			if (do_symlink(sl, buf) < 0)
+			if (do_symlink_at(sl, buf) < 0)
 				rsyserr(FERROR, errno, "link %s -> \"%s\"", full_fname(buf), sl);
 			else if (DEBUG_GTE(BACKUP, 1))
 				rprintf(FINFO, "make_backup: SYMLINK %s successful.\n", fname);

cleanup.c

@@ -198,7 +198,7 @@ NORETURN void _exit_cleanup(int code, const char *file, int line)
 		switch_step++;
 
 		if (cleanup_fname)
-			do_unlink(cleanup_fname);
+			do_unlink_at(cleanup_fname);
 		if (exit_code)
 			kill_all(SIGUSR1);
 		if (cleanup_pid && cleanup_pid == getpid()) {
@@ -269,8 +269,16 @@ NORETURN void _exit_cleanup(int code, const char *file, int line)
 		break;
 	}
 
-	if (called_from_signal_handler)
+	if (called_from_signal_handler) {
+#ifdef GCOV_COVERAGE
+		/* _exit() bypasses the gcov atexit flush; rsync's generator (and
+		 * other processes) normally finish via the signal handler, so
+		 * without this they would write no .gcda. Harmless otherwise. */
+		extern void __gcov_dump(void);
+		__gcov_dump();
+#endif
 		_exit(exit_code);
+	}
 	exit(exit_code);
 }
 

clientserver.c

@@ -30,6 +30,7 @@ extern int list_only;
 extern int am_sender;
 extern int am_server;
 extern int am_daemon;
+extern int am_chrooted;
 extern int am_root;
 extern int msgs2stderr;
 extern int rsync_port;
@@ -38,6 +39,7 @@ extern int ignore_errors;
 extern int preserve_xattrs;
 extern int kluge_around_eof;
 extern int munge_symlinks;
+extern int use_secure_symlinks;
 extern int open_noatime;
 extern int sanitize_paths;
 extern int numeric_ids;
@@ -983,6 +985,7 @@ static int rsync_module(int f_in, int f_out, int i, const char *addr, const char
 			io_printf(f_out, "@ERROR: chroot failed\n");
 			return -1;
 		}
+		am_chrooted = 1;
 		module_chdir = module_dir;
 	}
 
@@ -1005,6 +1008,15 @@ static int rsync_module(int f_in, int f_out, int i, const char *addr, const char
 		}
 	}
 
+	/* Enable secure symlink handling for any non-chrooted daemon module.
+	 * This prevents TOCTOU race attacks where an attacker could switch a
+	 * directory to a symlink between path validation and file open.
+	 * Match the gate used by the do_*_at() wrappers in syscall.c
+	 * (am_daemon && !am_chrooted) -- the protection has nothing to do
+	 * with symlink munging, so a module configured with
+	 * "munge symlinks = false" must still get the secure-open path. */
+	use_secure_symlinks = am_daemon && !am_chrooted;
+
 	if (gid_list.count) {
 		gid_t *gid_array = gid_list.items;
 		if (setgid(gid_array[0])) {
@@ -1300,6 +1312,28 @@ int start_daemon(int f_in, int f_out)
 	if (lp_proxy_protocol() && !read_proxy_protocol_header(f_in))
 		return -1;
 
+	/* Do reverse DNS lookup before chroot/setuid. The result is cached,
+	 * so the later client_name() call will use this cached value. This
+	 * ensures hostname-based ACLs work even when DNS is unavailable
+	 * after chroot.
+	 *
+	 * "reverse lookup" can be set globally OR per-module, so we also
+	 * scan each module: a deployment with "reverse lookup = no" in the
+	 * global section but "reverse lookup = yes" in a specific module
+	 * still triggers a post-chroot lookup at access-check time
+	 * (rsync_module() in this file), which would also fail in the
+	 * chroot and turn hostname-based deny rules into silent bypasses. */
+	{
+		int need_reverse = lp_reverse_lookup(-1);
+		int j, num_modules = lp_num_modules();
+		for (j = 0; !need_reverse && j < num_modules; j++) {
+			if (lp_reverse_lookup(j))
+				need_reverse = 1;
+		}
+		if (need_reverse)
+			(void)client_name(client_addr(f_in));
+	}
+
 	p = lp_daemon_chroot();
 	if (*p) {
 		log_init(0); /* Make use we've initialized syslog before chrooting. */
@@ -1308,6 +1342,19 @@ int start_daemon(int f_in, int f_out)
 			rsyserr(FLOG, errno, "daemon chroot(\"%s\") failed", p);
 			return -1;
 		}
+		/* Deliberately do NOT set am_chrooted here.  am_chrooted
+		 * gates the per-module symlink-race defenses
+		 * (secure_relative_open() and the do_*_at() wrappers in
+		 * syscall.c) and means "the kernel is enforcing path
+		 * confinement at the module boundary".  The daemon chroot
+		 * confines path resolution to the daemon-chroot directory,
+		 * not to any individual module path -- modules sharing the
+		 * daemon chroot are still distinguishable filesystem
+		 * subtrees and a sender-controlled symlink in module A
+		 * could redirect a syscall to module B (or to other files
+		 * inside the daemon chroot) without the per-module
+		 * defenses.  Leave am_chrooted=0 here so secure_relative_open()
+		 * still fires for "use chroot = no" modules. */
 		if (chdir("/") < 0) {
 			rsyserr(FLOG, errno, "daemon chdir(\"/\") failed");
 			return -1;

configure.ac

@@ -82,6 +82,32 @@ if test x"$enable_profile" = x"yes"; then
 	CFLAGS="$CFLAGS -pg"
 fi
 
+dnl Coverage build (gcov) for `make coverage`. NOTE: --enable-profile above is
+dnl gprof (-pg) and is NOT coverage. -O0 keeps branch coverage meaningful;
+dnl -fprofile-update=atomic keeps the shared .gcda counters correct while the
+dnl suite runs many rsync processes in parallel.
+AC_ARG_ENABLE(coverage,
+	AS_HELP_STRING([--enable-coverage],[build with gcov instrumentation for `make coverage`]))
+if test x"$enable_coverage" = x"yes"; then
+	CFLAGS="$CFLAGS --coverage -fprofile-update=atomic -O0"
+	CXXFLAGS="$CXXFLAGS --coverage -fprofile-update=atomic -O0"
+	LDFLAGS="$LDFLAGS --coverage"
+	AC_DEFINE([GCOV_COVERAGE], 1,
+		[Flush gcov counters at exit_cleanup: rsync's children exit via _exit(), which bypasses the gcov atexit handler, so without this no .gcda is written for the receiver/generator/daemon-worker processes.])
+fi
+
+dnl openat2(RESOLVE_BENEATH) is used on Linux 5.6+ for the secure resolver.
+dnl --disable-openat2 forces the portable per-component O_NOFOLLOW fallback to
+dnl run as the primary resolver on ordinary Linux, so that tier is exercised
+dnl (and coverage-counted) without needing a pre-5.6 kernel. Behaviour-neutral
+dnl by default (the knob only REMOVES a tier when explicitly disabled).
+AC_ARG_ENABLE(openat2,
+	AS_HELP_STRING([--disable-openat2],[do not use Linux openat2(RESOLVE_BENEATH); force the portable resolver (for exercising the fallback tier)]))
+if test x"$enable_openat2" != x"no"; then
+	AC_DEFINE([HAVE_OPENAT2], 1,
+		[Define to use Linux openat2(RESOLVE_BENEATH) in secure_relative_open where available.])
+fi
+
 AC_MSG_CHECKING([if md2man can create manpages])
 if test x"$ac_cv_path_PYTHON3" = x; then
     AC_MSG_RESULT(no - python3 not found)

delete.c

@@ -98,7 +98,7 @@ static enum delret delete_dir_contents(char *fname, uint16 flags)
 
 		strlcpy(p, fp->basename, remainder);
 		if (!(fp->mode & S_IWUSR) && !am_root && fp->flags & FLAG_OWNED_BY_US)
-			do_chmod(fname, fp->mode | S_IWUSR);
+			do_chmod_at(fname, fp->mode | S_IWUSR);
 		/* Save stack by recursing to ourself directly. */
 		if (S_ISDIR(fp->mode)) {
 			if (delete_dir_contents(fname, flags | DEL_RECURSE) != DR_SUCCESS)
@@ -139,7 +139,7 @@ enum delret delete_item(char *fbuf, uint16 mode, uint16 flags)
 	}
 
 	if (flags & DEL_NO_UID_WRITE)
-		do_chmod(fbuf, mode | S_IWUSR);
+		do_chmod_at(fbuf, mode | S_IWUSR);
 
 	if (S_ISDIR(mode) && !(flags & DEL_DIR_IS_EMPTY)) {
 		/* This only happens on the first call to delete_item() since
@@ -160,7 +160,7 @@ enum delret delete_item(char *fbuf, uint16 mode, uint16 flags)
 
 	if (S_ISDIR(mode)) {
 		what = "rmdir";
-		ok = do_rmdir(fbuf) == 0;
+		ok = do_rmdir_at(fbuf) == 0;
 	} else {
 		if (make_backups > 0 && !(flags & DEL_FOR_BACKUP) && (backup_dir || !is_backup_file(fbuf))) {
 			what = "make_backup";

flist.c

@@ -840,9 +840,9 @@ static struct file_struct *recv_file_entry(int f, struct file_list *flist, int x
 	}
 	if (xflags & XMIT_MOD_NSEC)
 #ifndef CAN_SET_NSEC
-		(void)read_varint(f);
+		(void)read_varint_bounded(f, 0, MAX_WIRE_NSEC, "modtime_nsec");
 #else
-		modtime_nsec = read_varint(f);
+		modtime_nsec = read_varint_bounded(f, 0, MAX_WIRE_NSEC, "modtime_nsec");
 	else
 		modtime_nsec = 0;
 #endif
@@ -861,8 +861,19 @@ static struct file_struct *recv_file_entry(int f, struct file_list *flist, int x
 #endif
 	}
 #endif
-	if (!(xflags & XMIT_SAME_MODE))
+	if (!(xflags & XMIT_SAME_MODE)) {
 		mode = from_wire_mode(read_int(f));
+		/* Reject modes whose type bits are not one of the standard
+		 * file types; otherwise garbage mode values propagate through
+		 * the file-type checks below unpredictably. */
+		if (!S_ISREG(mode) && !S_ISDIR(mode) && !S_ISLNK(mode)
+		 && !S_ISCHR(mode) && !S_ISBLK(mode)
+		 && !S_ISFIFO(mode) && !S_ISSOCK(mode)) {
+			rprintf(FERROR, "invalid file mode 0%o for %s [%s]\n",
+				(unsigned)mode, lastname, who_am_i());
+			exit_cleanup(RERR_PROTOCOL);
+		}
+	}
 	if (atimes_ndx && !S_ISDIR(mode) && !(xflags & XMIT_SAME_ATIME)) {
 		atime = read_varlong(f, 4);
 #if SIZEOF_TIME_T < SIZEOF_INT64

generator.c

@@ -229,11 +229,13 @@ static int read_delay_line(char *buf, int *flags_p)
 		*flags_p = 0;
 
 	if (sscanf(bp, "%x ", &mode) != 1) {
-	  invalid_data:
+		goto invalid_data;
-		rprintf(FERROR, "ERROR: invalid data in delete-delay file.\n");
-		return -1;
 	}
-	past_space = strchr(bp, ' ') + 1;
+	past_space = strchr(bp, ' ');
+	if (!past_space) {
+		goto invalid_data;
+	}
+	past_space++;
 	len = j - read_pos - (past_space - bp) + 1; /* count the '\0' */
 	read_pos = j + 1;
 
@@ -247,6 +249,10 @@ static int read_delay_line(char *buf, int *flags_p)
 	memcpy(buf, past_space, len);
 
 	return mode;
+
+invalid_data:
+	rprintf(FERROR, "ERROR: invalid data in delete-delay file.\n");
+	return -1;
 }
 
 static void do_delayed_deletions(char *delbuf)
@@ -984,7 +990,7 @@ static int try_dests_reg(struct file_struct *file, char *fname, int ndx,
 		if (find_exact_for_existing) {
 			if (alt_dest_type == LINK_DEST && real_st.st_dev == sxp->st.st_dev && real_st.st_ino == sxp->st.st_ino)
 				return -1;
-			if (do_unlink(fname) < 0 && errno != ENOENT)
+			if (do_unlink_at(fname) < 0 && errno != ENOENT)
 				goto got_nothing_for_ya;
 		}
 #ifdef SUPPORT_HARD_LINKS
@@ -1112,7 +1118,7 @@ static int try_dests_non(struct file_struct *file, char *fname, int ndx,
 		 && !IS_SPECIAL(file->mode) && !IS_DEVICE(file->mode)
 #endif
 		 && !S_ISDIR(file->mode)) {
-			if (do_link(cmpbuf, fname) < 0) {
+			if (do_link_at(cmpbuf, fname) < 0) {
 				rsyserr(FERROR_XFER, errno,
 					"failed to hard-link %s with %s",
 					cmpbuf, fname);
@@ -1315,7 +1321,7 @@ static void recv_generator(char *fname, struct file_struct *file, int ndx,
 				}
 			}
 			if (relative_paths && !implied_dirs && file->mode != 0
-			 && do_stat(dn, &sx.st) < 0) {
+			 && do_stat_at(dn, &sx.st) < 0) {
 				if (dry_run)
 					goto parent_is_dry_missing;
 				if (make_path(fname, MKP_DROP_NAME | MKP_SKIP_SLASH) < 0) {
@@ -1427,7 +1433,7 @@ static void recv_generator(char *fname, struct file_struct *file, int ndx,
 			 && (stype == FT_DIR
 			  || delete_item(fname, sx.st.st_mode, del_opts | DEL_FOR_DIR) != 0))
 				goto cleanup; /* Any errors get reported later. */
-			if (do_mkdir(fname, (file->mode|added_perms) & 0700) == 0)
+			if (do_mkdir_at(fname, (file->mode|added_perms) & 0700) == 0)
 				file->flags |= FLAG_DIR_CREATED;
 			goto cleanup;
 		}
@@ -1469,10 +1475,10 @@ static void recv_generator(char *fname, struct file_struct *file, int ndx,
 			itemize(fnamecmp, file, ndx, statret, &sx,
 				statret ? ITEM_LOCAL_CHANGE : 0, 0, NULL);
 		}
-		if (real_ret != 0 && do_mkdir(fname,file->mode|added_perms) < 0 && errno != EEXIST) {
+		if (real_ret != 0 && do_mkdir_at(fname,file->mode|added_perms) < 0 && errno != EEXIST) {
 			if (!relative_paths || errno != ENOENT
 			 || make_path(fname, MKP_DROP_NAME | MKP_SKIP_SLASH) < 0
-			 || (do_mkdir(fname, file->mode|added_perms) < 0 && errno != EEXIST)) {
+			 || (do_mkdir_at(fname, file->mode|added_perms) < 0 && errno != EEXIST)) {
 				rsyserr(FERROR_XFER, errno,
 					"recv_generator: mkdir %s failed",
 					full_fname(fname));
@@ -1499,7 +1505,7 @@ static void recv_generator(char *fname, struct file_struct *file, int ndx,
 #ifdef HAVE_CHMOD
 		if (!am_root && (file->mode & S_IRWXU) != S_IRWXU && dir_tweaking) {
 			mode_t mode = file->mode | S_IRWXU;
-			if (do_chmod(fname, mode) < 0) {
+			if (do_chmod_at(fname, mode) < 0) {
 				rsyserr(FERROR_XFER, errno,
 					"failed to modify permissions on %s",
 					full_fname(fname));
@@ -1808,7 +1814,7 @@ static void recv_generator(char *fname, struct file_struct *file, int ndx,
 		;
 	else if (quick_check_ok(FT_REG, fnamecmp, file, &sx.st)) {
 		if (partialptr) {
-			do_unlink(partialptr);
+			do_unlink_at(partialptr);
 			handle_partial_dir(partialptr, PDIR_DELETE);
 		}
 		set_file_attrs(fname, file, &sx, NULL, maybe_ATTRS_REPORT | maybe_ATTRS_ACCURATE_TIME);
@@ -1896,7 +1902,7 @@ static void recv_generator(char *fname, struct file_struct *file, int ndx,
 			back_file = NULL;
 			goto cleanup;
 		}
-		if ((f_copy = do_open(backupptr, O_WRONLY | O_CREAT | O_TRUNC | O_EXCL, 0600)) < 0) {
+		if ((f_copy = do_open_at(backupptr, O_WRONLY | O_CREAT | O_TRUNC | O_EXCL, 0600)) < 0) {
 			rsyserr(FERROR_XFER, errno, "open %s", full_fname(backupptr));
 			unmake_file(back_file);
 			back_file = NULL;
@@ -2016,7 +2022,7 @@ int atomic_create(struct file_struct *file, char *fname, const char *slnk, const
 
 	if (slnk) {
 #ifdef SUPPORT_LINKS
-		if (do_symlink(slnk, create_name) < 0) {
+		if (do_symlink_at(slnk, create_name) < 0) {
 			rsyserr(FERROR_XFER, errno, "symlink %s -> \"%s\" failed",
 				full_fname(create_name), slnk);
 			return 0;
@@ -2032,7 +2038,7 @@ int atomic_create(struct file_struct *file, char *fname, const char *slnk, const
 		return 0;
 #endif
 	} else {
-		if (do_mknod(create_name, file->mode, rdev) < 0) {
+		if (do_mknod_at(create_name, file->mode, rdev) < 0) {
 			rsyserr(FERROR_XFER, errno, "mknod %s failed",
 				full_fname(create_name));
 			return 0;
@@ -2040,14 +2046,14 @@ int atomic_create(struct file_struct *file, char *fname, const char *slnk, const
 	}
 
 	if (!skip_atomic) {
-		if (do_rename(tmpname, fname) < 0) {
+		if (do_rename_at(tmpname, fname) < 0) {
 			char *full_tmpname = strdup(full_fname(tmpname));
 			if (full_tmpname == NULL)
 				out_of_memory("atomic_create");
 			rsyserr(FERROR_XFER, errno, "rename %s -> \"%s\" failed",
 				full_tmpname, full_fname(fname));
 			free(full_tmpname);
-			do_unlink(tmpname);
+			do_unlink_at(tmpname);
 			return 0;
 		}
 	}
@@ -2111,7 +2117,7 @@ static void touch_up_dirs(struct file_list *flist, int ndx)
 			continue;
 		fname = f_name(file, NULL);
 		if (fix_dir_perms)
-			do_chmod(fname, file->mode);
+			do_chmod_at(fname, file->mode);
 		if (need_retouch_dir_times) {
 			STRUCT_STAT st;
 			if (link_stat(fname, &st, 0) == 0 && mtime_differs(&st, file)) {
@@ -2146,6 +2152,8 @@ void check_for_finished_files(int itemizing, enum logcode code, int check_redo)
 			if (send_failed)
 				ndx = get_hlink_num();
 			flist = flist_for_ndx(ndx, "check_for_finished_files.1");
+			if (ndx < flist->ndx_start)
+				exit_cleanup(RERR_PROTOCOL);
 			file = flist->files[ndx - flist->ndx_start];
 			assert(file->flags & FLAG_HLINKED);
 			if (send_failed)
@@ -2174,6 +2182,8 @@ void check_for_finished_files(int itemizing, enum logcode code, int check_redo)
 
 			flist = cur_flist;
 			cur_flist = flist_for_ndx(ndx, "check_for_finished_files.2");
+			if (ndx < cur_flist->ndx_start)
+				exit_cleanup(RERR_PROTOCOL);
 
 			file = cur_flist->files[ndx - cur_flist->ndx_start];
 			if (solo_file)

hlink.c

@@ -454,7 +454,7 @@ int hard_link_check(struct file_struct *file, int ndx, char *fname,
 int hard_link_one(struct file_struct *file, const char *fname,
 		  const char *oldname, int terse)
 {
-	if (do_link(oldname, fname) < 0) {
+	if (do_link_at(oldname, fname) < 0) {
 		enum logcode code;
 		if (terse) {
 			if (!INFO_GTE(NAME, 1))

io.c

@@ -1090,6 +1090,9 @@ static void got_flist_entry_status(enum festatus status, int ndx)
 {
 	struct file_list *flist = flist_for_ndx(ndx, "got_flist_entry_status");
 
+	if (ndx < flist->ndx_start)
+		exit_cleanup(RERR_PROTOCOL);
+
 	if (remove_source_files) {
 		active_filecnt--;
 		active_bytecnt -= F_LENGTH(flist->files[ndx - flist->ndx_start]);
@@ -1865,6 +1868,45 @@ int64 read_varlong(int f, uchar min_bytes)
 	return u.x;
 }
 
+/* Read an int32 and verify lo <= v <= hi. On out-of-range, abort with a
+ * protocol error naming "what". The bound is co-located with the read so it
+ * cannot be forgotten by a downstream user. */
+int32 read_int_bounded(int f, int32 lo, int32 hi, const char *what)
+{
+	int32 v = read_int(f);
+	if (v < lo || v > hi) {
+		rprintf(FERROR, "wire value %s out of range: %ld not in [%ld,%ld] [%s]\n",
+			what, (long)v, (long)lo, (long)hi, who_am_i());
+		exit_cleanup(RERR_PROTOCOL);
+	}
+	return v;
+}
+
+/* As read_int_bounded but for varint-encoded values. */
+int32 read_varint_bounded(int f, int32 lo, int32 hi, const char *what)
+{
+	int32 v = read_varint(f);
+	if (v < lo || v > hi) {
+		rprintf(FERROR, "wire value %s out of range: %ld not in [%ld,%ld] [%s]\n",
+			what, (long)v, (long)lo, (long)hi, who_am_i());
+		exit_cleanup(RERR_PROTOCOL);
+	}
+	return v;
+}
+
+/* Read a varint that will be used as a size_t. Rejects negative values
+ * (which would wrap to ~SIZE_MAX) and values exceeding the supplied max. */
+size_t read_varint_size(int f, size_t max, const char *what)
+{
+	int32 v = read_varint(f);
+	if (v < 0 || (size_t)v > max) {
+		rprintf(FERROR, "wire size %s out of range: %ld > %lu [%s]\n",
+			what, (long)v, (unsigned long)max, who_am_i());
+		exit_cleanup(RERR_PROTOCOL);
+	}
+	return (size_t)v;
+}
+
 int64 read_longint(int f)
 {
 #if SIZEOF_INT64 >= 8
@@ -1971,6 +2013,21 @@ void read_sum_head(int f, struct sum_struct *sum)
 			(long)sum->count, who_am_i());
 		exit_cleanup(RERR_PROTOCOL);
 	}
+	/* Guard against integer overflow in downstream allocations sized by
+	 * count*element_size. my_alloc uses divide-not-multiply so it is
+	 * already wraparound-safe, but checking here gives a clearer error
+	 * and also covers the (size_t)count * xfer_sum_len arithmetic that
+	 * is performed *before* reaching my_alloc. */
+	if (xfer_sum_len > 0 && (size_t)sum->count > SIZE_MAX / (size_t)xfer_sum_len) {
+		rprintf(FERROR, "Invalid checksum count %ld (too large) [%s]\n",
+			(long)sum->count, who_am_i());
+		exit_cleanup(RERR_PROTOCOL);
+	}
+	if ((size_t)sum->count > SIZE_MAX / sizeof(struct sum_buf)) {
+		rprintf(FERROR, "Invalid checksum count %ld (sum_buf overflow) [%s]\n",
+			(long)sum->count, who_am_i());
+		exit_cleanup(RERR_PROTOCOL);
+	}
 	sum->blength = read_int(f);
 	if (sum->blength < 0 || sum->blength > max_blength) {
 		rprintf(FERROR, "Invalid block length %ld [%s]\n",

log.c

@@ -456,11 +456,17 @@ void rsyserr(enum logcode code, int errcode, const char *format, ...)
 	char buf[BIGPATHBUFLEN];
 	size_t len;
 
+	/* snprintf returns the would-have-been length on truncation, so
+	 * each cumulative call must be guarded; if not, sizeof buf - len
+	 * can underflow when promoted to size_t and the next call writes
+	 * past the buffer. */
 	len = snprintf(buf, sizeof buf, RSYNC_NAME ": [%s] ", who_am_i());
 
-	va_start(ap, format);
+	if (len < sizeof buf) {
-	len += vsnprintf(buf + len, sizeof buf - len, format, ap);
+		va_start(ap, format);
-	va_end(ap);
+		len += vsnprintf(buf + len, sizeof buf - len, format, ap);
+		va_end(ap);
+	}
 
 	if (len < sizeof buf) {
 		len += snprintf(buf + len, sizeof buf - len,

main.c

@@ -239,11 +239,11 @@ void write_del_stats(int f)
 
 void read_del_stats(int f)
 {
-	stats.deleted_files = read_varint(f);
+	stats.deleted_files = read_varint_bounded(f, 0, MAX_WIRE_DEL_STAT, "deleted_files");
-	stats.deleted_files += stats.deleted_dirs = read_varint(f);
+	stats.deleted_files += stats.deleted_dirs = read_varint_bounded(f, 0, MAX_WIRE_DEL_STAT, "deleted_dirs");
-	stats.deleted_files += stats.deleted_symlinks = read_varint(f);
+	stats.deleted_files += stats.deleted_symlinks = read_varint_bounded(f, 0, MAX_WIRE_DEL_STAT, "deleted_symlinks");
-	stats.deleted_files += stats.deleted_devices = read_varint(f);
+	stats.deleted_files += stats.deleted_devices = read_varint_bounded(f, 0, MAX_WIRE_DEL_STAT, "deleted_devices");
-	stats.deleted_files += stats.deleted_specials = read_varint(f);
+	stats.deleted_files += stats.deleted_specials = read_varint_bounded(f, 0, MAX_WIRE_DEL_STAT, "deleted_specials");
 }
 
 static void become_copy_as_user()
@@ -394,9 +394,18 @@ static void output_itemized_counts(const char *prefix, int *counts)
 		counts[0] -= counts[1] + counts[2] + counts[3] + counts[4];
 		for (j = 0; j < 5; j++) {
 			if (counts[j]) {
+				/* snprintf can return more than its size arg
+				 * on truncation; keep len <= sizeof buf - 2 so
+				 * the closing ')' and trailing NUL always
+				 * have room and the next iteration's
+				 * sizeof buf - len - 2 cannot underflow. */
+				if (len >= (int)sizeof buf - 2)
+					break;
 				len += snprintf(buf+len, sizeof buf - len - 2,
 					"%s%s: %s",
 					pre, labels[j], comma_num(counts[j]));
+				if (len > (int)sizeof buf - 2)
+					len = (int)sizeof buf - 2;
 				pre = ", ";
 			}
 		}
@@ -1559,6 +1568,10 @@ static int start_client(int argc, char *argv[])
 			shell_user = shell_machine;
 			shell_machine = p+1;
 		}
+		if (*shell_machine == '-') {
+			rprintf(FERROR, "Invalid remote host: hostnames may not start with '-'.\n");
+			exit_cleanup(RERR_SYNTAX);
+		}
 	}
 
 	if (DEBUG_GTE(CMD, 2)) {
@@ -1605,6 +1618,11 @@ static void sigusr2_handler(UNUSED(int val))
 	if (!am_server)
 		output_summary();
 	close_all();
+#ifdef GCOV_COVERAGE
+	/* The receiver child is killed here via SIGUSR2 and exits with _exit(),
+	 * bypassing the gcov atexit flush; without this it writes no .gcda. */
+	{ extern void __gcov_dump(void); __gcov_dump(); }
+#endif
 	if (got_xfer_error)
 		_exit(RERR_PARTIAL);
 	_exit(0);

options.c

@@ -114,11 +114,20 @@ int mkpath_dest_arg = 0;
 int allow_inc_recurse = 1;
 int xfer_dirs = -1;
 int am_daemon = 0;
+/* Set after a successful per-module chroot ("use chroot = yes") in
+ * clientserver.c. NOT set for the daemon-level "daemon chroot = /X"
+ * chroot: that confines path resolution to /X, but module paths
+ * /X/modA, /X/modB, etc. are not chroot boundaries, so the per-module
+ * symlink-race defenses (secure_relative_open() / do_*_at() in
+ * syscall.c, gated by `am_daemon && !am_chrooted`) must still fire
+ * even when the daemon is inside a daemon chroot. */
+int am_chrooted = 0;
 int connect_timeout = 0;
 int keep_partial = 0;
 int safe_symlinks = 0;
 int copy_unsafe_links = 0;
 int munge_symlinks = 0;
+int use_secure_symlinks = 0;
 int size_only = 0;
 int daemon_bwlimit = 0;
 int bwlimit = 0;

packaging/branch-from-patch

@@ -1,174 +0,0 @@
-#!/usr/bin/env -S python3 -B
-
-# This script turns one or more diff files in the patches dir (which is
-# expected to be a checkout of the rsync-patches git repo) into a branch
-# in the main rsync git checkout. This allows the applied patch to be
-# merged with the latest rsync changes and tested.  To update the diff
-# with the resulting changes, see the patch-update script.
-
-import os, sys, re, argparse, glob
-
-sys.path = ['packaging'] + sys.path
-
-from pkglib import *
-
-def main():
-    global created, info, local_branch
-
-    cur_branch, args.base_branch = check_git_state(args.base_branch, not args.skip_check, args.patches_dir)
-
-    local_branch = get_patch_branches(args.base_branch)
-
-    if args.delete_local_branches:
-        for name in sorted(local_branch):
-            branch = f"patch/{args.base_branch}/{name}"
-            cmd_chk(['git', 'branch', '-D', branch])
-        local_branch = set()
-
-    if args.add_missing:
-        for fn in sorted(glob.glob(f"{args.patches_dir}/*.diff")):
-            name = re.sub(r'\.diff$', '', re.sub(r'.+/', '', fn))
-            if name not in local_branch and fn not in args.patch_files:
-                args.patch_files.append(fn)
-
-    if not args.patch_files:
-        return
-
-    for fn in args.patch_files:
-        if not fn.endswith('.diff'):
-            die(f"Filename is not a .diff file: {fn}")
-        if not os.path.isfile(fn):
-            die(f"File not found: {fn}")
-
-    scanned = set()
-    info = { }
-
-    patch_list = [ ]
-    for fn in args.patch_files:
-        m = re.match(r'^(?P<dir>.*?)(?P<name>[^/]+)\.diff$', fn)
-        patch = argparse.Namespace(**m.groupdict())
-        if patch.name in scanned:
-            continue
-        patch.fn = fn
-
-        lines = [ ]
-        commit_hash = None
-        with open(patch.fn, 'r', encoding='utf-8') as fh:
-            for line in fh:
-                m = re.match(r'^based-on: (\S+)', line)
-                if m:
-                    commit_hash = m[1]
-                    break
-                if (re.match(r'^index .*\.\..* \d', line)
-                 or re.match(r'^diff --git ', line)
-                 or re.match(r'^--- (old|a)/', line)):
-                    break
-                lines.append(re.sub(r'\s*\Z', "\n", line, 1))
-        info_txt = ''.join(lines).strip() + "\n"
-        lines = None
-
-        parent = args.base_branch
-        patches = re.findall(r'patch -p1 <%s/(\S+)\.diff' % args.patches_dir, info_txt)
-        if patches:
-            last = patches.pop()
-            if last != patch.name:
-                warn(f"No identity patch line in {patch.fn}")
-                patches.append(last)
-            if patches:
-                parent = patches.pop()
-                if parent not in scanned:
-                    diff_fn = patch.dir + parent + '.diff'
-                    if not os.path.isfile(diff_fn):
-                        die(f"Failed to find parent of {patch.fn}: {parent}")
-                    # Add parent to args.patch_files so that we will look for the
-                    # parent's parent.  Any duplicates will be ignored.
-                    args.patch_files.append(diff_fn)
-        else:
-            warn(f"No patch lines found in {patch.fn}")
-
-        info[patch.name] = [ parent, info_txt, commit_hash ]
-
-        patch_list.append(patch)
-
-    created = set()
-    for patch in patch_list:
-        create_branch(patch)
-
-    cmd_chk(['git', 'checkout', args.base_branch])
-
-
-def create_branch(patch):
-    if patch.name in created:
-        return
-    created.add(patch.name)
-
-    parent, info_txt, commit_hash = info[patch.name]
-    parent = argparse.Namespace(dir=patch.dir, name=parent, fn=patch.dir + parent + '.diff')
-
-    if parent.name == args.base_branch:
-        parent_branch = commit_hash if commit_hash else args.base_branch
-    else:
-        create_branch(parent)
-        parent_branch = '/'.join(['patch', args.base_branch, parent.name])
-
-    branch = '/'.join(['patch', args.base_branch, patch.name])
-    print("\n" + '=' * 64)
-    print(f"Processing {branch} ({parent_branch})")
-
-    if patch.name in local_branch:
-        cmd_chk(['git', 'branch', '-D', branch])
-
-    cmd_chk(['git', 'checkout', '-b', branch, parent_branch])
-
-    info_fn = 'PATCH.' + patch.name
-    with open(info_fn, 'w', encoding='utf-8') as fh:
-        fh.write(info_txt)
-    cmd_chk(['git', 'add', info_fn])
-
-    with open(patch.fn, 'r', encoding='utf-8') as fh:
-        patch_txt = fh.read()
-
-    cmd_run('patch -p1'.split(), input=patch_txt)
-
-    for fn in glob.glob('*.orig') + glob.glob('*/*.orig'):
-        os.unlink(fn)
-
-    pos = 0
-    new_file_re = re.compile(r'\nnew file mode (?P<mode>\d+)\s+--- /dev/null\s+\+\+\+ b/(?P<fn>.+)')
-    while True:
-        m = new_file_re.search(patch_txt, pos)
-        if not m:
-            break
-        os.chmod(m['fn'], int(m['mode'], 8))
-        cmd_chk(['git', 'add', m['fn']])
-        pos = m.end()
-
-    while True:
-        cmd_chk('git status'.split())
-        ans = input('Press Enter to commit, Ctrl-C to abort, or type a wild-name to add a new file: ')
-        if ans == '':
-            break
-        cmd_chk("git add " + ans, shell=True)
-
-    while True:
-        s = cmd_run(['git', 'commit', '-a', '-m', f"Creating branch from {patch.name}.diff."])
-        if not s.returncode:
-            break
-        s = cmd_run([os.environ.get('SHELL', '/bin/sh')])
-        if s.returncode:
-            die('Aborting due to shell error code')
-
-
-if __name__ == '__main__':
-    parser = argparse.ArgumentParser(description="Create a git patch branch from an rsync patch file.", add_help=False)
-    parser.add_argument('--branch', '-b', dest='base_branch', metavar='BASE_BRANCH', default='master', help="The branch the patch is based on. Default: master.")
-    parser.add_argument('--add-missing', '-a', action='store_true', help="Add a branch for every patches/*.diff that doesn't have a branch.")
-    parser.add_argument('--skip-check', action='store_true', help="Skip the check that ensures starting with a clean branch.")
-    parser.add_argument('--delete', dest='delete_local_branches', action='store_true', help="Delete all the local patch/BASE/* branches, not just the ones that are being recreated.")
-    parser.add_argument('--patches-dir', '-p', metavar='DIR', default='patches', help="Override the location of the rsync-patches dir. Default: patches.")
-    parser.add_argument('patch_files', metavar='patches/DIFF_FILE', nargs='*', help="Specify what patch diff files to process. Default: all of them.")
-    parser.add_argument("--help", "-h", action="help", help="Output this help message and exit.")
-    args = parser.parse_args()
-    main()
-
-# vim: sw=4 et ft=python

packaging/ftp.filt

@@ -0,0 +1,2 @@
+- /generated-files/
+- /binaries/

packaging/lsb/rsync.spec

@@ -1,6 +1,6 @@
 Summary: A fast, versatile, remote (and local) file-copying tool
 Name: rsync
-Version: 3.4.2
+Version: 3.4.3
 %define fullversion %{version}
 Release: 1
 %define srcdir src
@@ -79,5 +79,5 @@ rm -rf $RPM_BUILD_ROOT
 %dir /etc/rsync-ssl/certs
 
 %changelog
-* Tue Apr 28 2026 Rsync Project <rsync.project@gmail.com>
+* Wed May 20 2026 Rsync Project <rsync.project@gmail.com>
-Released 3.4.2.
+Released 3.4.3.

packaging/patch-update

@@ -1,244 +0,0 @@
-#!/usr/bin/env -S python3 -B
-
-# This script is used to turn one or more of the "patch/BASE/*" branches
-# into one or more diffs in the "patches" directory.  Pass the option
-# --gen if you want generated files in the diffs.  Pass the name of
-# one or more diffs if you want to just update a subset of all the
-# diffs.
-
-import os, sys, re, argparse, time, shutil
-
-sys.path = ['packaging'] + sys.path
-
-from pkglib import *
-
-MAKE_GEN_CMDS = [
-        './prepare-source'.split(),
-        'cd build && if test -f config.status ; then ./config.status ; else ../configure ; fi',
-        'make -C build gen'.split(),
-        ]
-TMP_DIR = "patches.gen"
-
-os.environ['GIT_MERGE_AUTOEDIT'] = 'no'
-
-def main():
-    global master_commit, parent_patch, description, completed, last_touch
-
-    if not os.path.isdir(args.patches_dir):
-        die(f'No "{args.patches_dir}" directory was found.')
-    if not os.path.isdir('.git'):
-        die('No ".git" directory present in the current dir.')
-
-    starting_branch, args.base_branch = check_git_state(args.base_branch, not args.skip_check, args.patches_dir)
-
-    master_commit = latest_git_hash(args.base_branch)
-
-    if cmd_txt_chk(['packaging/prep-auto-dir']).out == '':
-        die('You must setup an auto-build-save dir to use this script.')
-
-    if args.gen:
-        if os.path.lexists(TMP_DIR):
-            die(f'"{TMP_DIR}" must not exist in the current directory.')
-        gen_files = get_gen_files()
-        os.mkdir(TMP_DIR, 0o700)
-        for cmd in MAKE_GEN_CMDS:
-            cmd_chk(cmd)
-        cmd_chk(['rsync', '-a', *gen_files, f'{TMP_DIR}/master/'])
-
-    last_touch = int(time.time())
-
-    # Start by finding all patches so that we can load all possible parents.
-    patches = sorted(list(get_patch_branches(args.base_branch)))
-
-    parent_patch = { }
-    description = { }
-
-    for patch in patches:
-        branch = f"patch/{args.base_branch}/{patch}"
-        desc = ''
-        proc = cmd_pipe(['git', 'diff', '-U1000', f"{args.base_branch}...{branch}", '--', f"PATCH.{patch}"])
-        in_diff = False
-        for line in proc.stdout:
-            if in_diff:
-                if not re.match(r'^[ +]', line):
-                    continue
-                line = line[1:]
-                m = re.search(r'patch -p1 <patches/(\S+)\.diff', line)
-                if m and m[1] != patch:
-                    parpat = parent_patch[patch] = m[1]
-                    if not parpat in patches:
-                        die(f"Parent of {patch} is not a local branch: {parpat}")
-                desc += line
-            elif re.match(r'^@@ ', line):
-                in_diff = True
-        description[patch] = desc
-        proc.communicate()
-
-    if args.patch_files: # Limit the list of patches to actually process
-        valid_patches = patches
-        patches = [ ]
-        for fn in args.patch_files:
-            name = re.sub(r'\.diff$', '', re.sub(r'.+/', '', fn))
-            if name not in valid_patches:
-                die(f"Local branch not available for patch: {name}")
-            patches.append(name)
-
-    completed = set()
-
-    for patch in patches:
-        if patch in completed:
-            continue
-        if not update_patch(patch):
-            break
-
-    if args.gen:
-        shutil.rmtree(TMP_DIR)
-
-    while last_touch >= int(time.time()):
-        time.sleep(1)
-    cmd_chk(['git', 'checkout', starting_branch])
-    cmd_chk(['packaging/prep-auto-dir'], discard='output')
-
-
-def update_patch(patch):
-    global last_touch
-
-    completed.add(patch) # Mark it as completed early to short-circuit any (bogus) dependency loops.
-
-    parent = parent_patch.get(patch, None)
-    if parent:
-        if parent not in completed:
-            if not update_patch(parent):
-                return 0
-        based_on = parent = f"patch/{args.base_branch}/{parent}"
-    else:
-        parent = args.base_branch
-        based_on = master_commit
-
-    print(f"======== {patch} ========")
-
-    while args.gen and last_touch >= int(time.time()):
-        time.sleep(1)
-
-    branch = f"patch/{args.base_branch}/{patch}"
-    s = cmd_run(['git', 'checkout', branch])
-    if s.returncode != 0:
-        return 0
-
-    s = cmd_run(['git', 'merge', based_on])
-    ok = s.returncode == 0
-    skip_shell = False
-    if not ok or args.cmd or args.make or args.shell:
-        cmd_chk(['packaging/prep-auto-dir'], discard='output')
-    if not ok:
-        print(f'"git merge {based_on}" incomplete -- please fix.')
-        if not run_a_shell(parent, patch):
-            return 0
-        if not args.make and not args.cmd:
-            skip_shell = True
-    if args.make:
-        if cmd_run(['packaging/smart-make']).returncode != 0:
-            if not run_a_shell(parent, patch):
-                return 0
-            if not args.cmd:
-                skip_shell = True
-    if args.cmd:
-        if cmd_run(args.cmd).returncode != 0:
-            if not run_a_shell(parent, patch):
-                return 0
-            skip_shell = True
-    if args.shell and not skip_shell:
-        if not run_a_shell(parent, patch):
-            return 0
-
-    with open(f"{args.patches_dir}/{patch}.diff", 'w', encoding='utf-8') as fh:
-        fh.write(description[patch])
-        fh.write(f"\nbased-on: {based_on}\n")
-
-        if args.gen:
-            gen_files = get_gen_files()
-            for cmd in MAKE_GEN_CMDS:
-                cmd_chk(cmd)
-            cmd_chk(['rsync', '-a', *gen_files, f"{TMP_DIR}/{patch}/"])
-        else:
-            gen_files = [ ]
-        last_touch = int(time.time())
-
-        proc = cmd_pipe(['git', 'diff', based_on])
-        skipping = False
-        for line in proc.stdout:
-            if skipping:
-                if not re.match(r'^diff --git a/', line):
-                    continue
-                skipping = False
-            elif re.match(r'^diff --git a/PATCH', line):
-                skipping = True
-                continue
-            if not re.match(r'^index ', line):
-                fh.write(line)
-        proc.communicate()
-
-        if args.gen:
-            e_tmp_dir = re.escape(TMP_DIR)
-            diff_re  = re.compile(r'^(diff -Nurp) %s/[^/]+/(.*?) %s/[^/]+/(.*)' % (e_tmp_dir, e_tmp_dir))
-            minus_re = re.compile(r'^\-\-\- %s/[^/]+/([^\t]+)\t.*' % e_tmp_dir)
-            plus_re  = re.compile(r'^\+\+\+ %s/[^/]+/([^\t]+)\t.*' % e_tmp_dir)
-
-            if parent == args.base_branch:
-                parent_dir = 'master'
-            else:
-                m = re.search(r'([^/]+)$', parent)
-                parent_dir = m[1]
-
-            proc = cmd_pipe(['diff', '-Nurp', f"{TMP_DIR}/{parent_dir}", f"{TMP_DIR}/{patch}"])
-            for line in proc.stdout:
-                line = diff_re.sub(r'\1 a/\2 b/\3', line)
-                line = minus_re.sub(r'--- a/\1', line)
-                line =  plus_re.sub(r'+++ b/\1', line)
-                fh.write(line)
-            proc.communicate()
-
-    return 1
-
-
-def run_a_shell(parent, patch):
-    m = re.search(r'([^/]+)$', parent)
-    parent_dir = m[1]
-    os.environ['PS1'] = f"[{parent_dir}] {patch}: "
-
-    while True:
-        s = cmd_run([os.environ.get('SHELL', '/bin/sh')])
-        if s.returncode != 0:
-            ans = input("Abort? [n/y] ")
-            if re.match(r'^y', ans, flags=re.I):
-                return False
-            continue
-        cur_branch, is_clean, status_txt = check_git_status(0)
-        if is_clean:
-            break
-        print(status_txt, end='')
-
-    cmd_run('rm -f build/*.o build/*/*.o')
-
-    return True
-
-
-if __name__ == '__main__':
-    parser = argparse.ArgumentParser(description="Turn a git branch back into a diff files in the patches dir.", add_help=False)
-    parser.add_argument('--branch', '-b', dest='base_branch', metavar='BASE_BRANCH', default='master', help="The branch the patch is based on. Default: master.")
-    parser.add_argument('--skip-check', action='store_true', help="Skip the check that ensures starting with a clean branch.")
-    parser.add_argument('--make', '-m', action='store_true', help="Run the smart-make script in every patch branch.")
-    parser.add_argument('--cmd', '-c', help="Run a command in every patch branch.")
-    parser.add_argument('--shell', '-s', action='store_true', help="Launch a shell for every patch/BASE/* branch updated, not just when a conflict occurs.")
-    parser.add_argument('--gen', metavar='DIR', nargs='?', const='', help='Include generated files. Optional DIR value overrides the default of using the "patches" dir.')
-    parser.add_argument('--patches-dir', '-p', metavar='DIR', default='patches', help="Override the location of the rsync-patches dir. Default: patches.")
-    parser.add_argument('patch_files', metavar='patches/DIFF_FILE', nargs='*', help="Specify what patch diff files to process. Default: all of them.")
-    parser.add_argument("--help", "-h", action="help", help="Output this help message and exit.")
-    args = parser.parse_args()
-    if args.gen == '':
-        args.gen = args.patches_dir
-    elif args.gen is not None:
-        args.patches_dir = args.gen
-    main()
-
-# vim: sw=4 et ft=python

packaging/release-rsync

@@ -1,385 +0,0 @@
-#!/usr/bin/env -S python3 -B
-
-# This script expects the directory ~/samba-rsync-ftp to exist and to be a
-# copy of the /home/ftp/pub/rsync dir on samba.org.  When the script is done,
-# the git repository in the current directory will be updated, and the local
-# ~/samba-rsync-ftp dir will be ready to be rsynced to samba.org. See the
-# script samba-rsync for an easy way to initialize the local ftp copy and to
-# thereafter update the remote files from your local copy.
-
-# This script also expects to be able to gpg sign the resulting tar files
-# using your default gpg key. Make sure that the html download.html file
-# has a link to the relevant keys that are authorized to sign the tar files
-# and also make sure that the following commands work as expected:
-#
-#   touch TeMp
-#   gpg --sign TeMp
-#   gpg --verify TeMp.gpg
-#   gpg --sign TeMp
-#   rm TeMp*
-#
-# The second time you sign the file it should NOT prompt you for your password
-# (unless the timeout period has passed).  It will prompt about overriding the
-# existing TeMp.gpg file, though.
-
-import os, sys, re, argparse, glob, shutil, signal
-from datetime import datetime
-from getpass import getpass
-
-sys.path = ['packaging'] + sys.path
-
-from pkglib import *
-
-os.environ['LESS'] = 'mqeiXR'; # Make sure that -F is turned off and -R is turned on.
-dest = os.environ['HOME'] + '/samba-rsync-ftp'
-ORIGINAL_PATH = os.environ['PATH']
-
-def main():
-    if not os.path.isfile('packaging/release-rsync'):
-        die('You must run this script from the top of your rsync checkout.')
-
-    now = datetime.now().astimezone() # Requires python 3.6 or later
-    cl_today = now.strftime('* %a %b %d %Y')
-    year = now.strftime('%Y')
-    ztoday = now.strftime('%d %b %Y')
-    today = ztoday.lstrip('0')
-
-    # The MAINTAINER_TZ_OFFSET is a float number of hours vs UTC. It can start with '-' but not '+'.
-    tz_now = now.strftime('%z')
-    tz_num = tz_now[0:1].replace('+', '') + str(float(tz_now[1:3]) + float(tz_now[3:]) / 60)
-
-    curdir = os.getcwd()
-
-    signal.signal(signal.SIGINT, signal_handler)
-
-    if cmd_txt_chk(['packaging/prep-auto-dir']).out == '':
-        die('You must setup an auto-build-save dir to use this script.');
-
-    auto_dir, gen_files = get_gen_files(True)
-    gen_pathnames = [ os.path.join(auto_dir, fn) for fn in gen_files ]
-
-    dash_line = '=' * 74
-
-    print(f"""\
-{dash_line}
-== This will release a new version of rsync onto an unsuspecting world. ==
-{dash_line}
-""")
-
-    with open('build/rsync.1') as fh:
-        for line in fh:
-            if line.startswith(r'.\" prefix='):
-                doc_prefix = line.split('=')[1].strip()
-                if doc_prefix != '/usr':
-                    warn(f"*** The documentation was built with prefix {doc_prefix} instead of /usr ***")
-                    die("*** Read the md2man script for a way to override this. ***")
-                break
-            if line.startswith('.P'):
-                die("Failed to find the prefix comment at the start of the rsync.1 manpage.")
-
-    if not os.path.isdir(dest):
-        die(dest, "dest does not exist")
-    if not os.path.isdir('.git'):
-        die("There is no .git dir in the current directory.")
-    if os.path.lexists('a'):
-        die('"a" must not exist in the current directory.')
-    if os.path.lexists('b'):
-        die('"b" must not exist in the current directory.')
-
-    curversion = get_rsync_version()
-
-    # All version values are strings!
-    lastversion, last_protocol_version, pdate = get_NEWS_version_info()
-    protocol_version, subprotocol_version = get_protocol_versions()
-
-    version = curversion
-    m = re.search(r'pre(\d+)', version)
-    if m:
-        version = re.sub(r'pre\d+', 'pre' + str(int(m[1]) + 1), version)
-    else:
-        version = version.replace('dev', 'pre1')
-
-    ans = input(f"Please enter the version number of this release: [{version}] ")
-    if ans == '.':
-        version = re.sub(r'pre\d+', '', version)
-    elif ans != '':
-        version = ans
-    if not re.match(r'^[\d.]+(pre\d+)?$', version):
-        die(f'Invalid version: "{version}"')
-
-    v_ver = 'v' + version
-    rsync_ver = 'rsync-' + version
-
-    if os.path.lexists(rsync_ver):
-        die(f'"{rsync_ver}" must not exist in the current directory.')
-
-    out = cmd_txt_chk(['git', 'tag', '-l', v_ver]).out
-    if out != '':
-        print(f"Tag {v_ver} already exists.")
-        ans = input("\nDelete tag or quit? [Q/del] ")
-        if not re.match(r'^del', ans, flags=re.I):
-            die("Aborted")
-        cmd_chk(['git', 'tag', '-d', v_ver])
-
-    version = re.sub(r'[-.]*pre[-.]*', 'pre', version)
-    if 'pre' in version and not curversion.endswith('dev'):
-        lastversion = curversion
-
-    ans = input(f"Enter the previous version to produce a patch against: [{lastversion}] ")
-    if ans != '':
-        lastversion = ans
-    lastversion = re.sub(r'[-.]*pre[-.]*', 'pre', lastversion)
-
-    rsync_lastver = 'rsync-' + lastversion
-    if os.path.lexists(rsync_lastver):
-        die(f'"{rsync_lastver}" must not exist in the current directory.')
-
-    m = re.search(r'(pre\d+)', version)
-    pre = m[1] if m else ''
-
-    release = '0.1' if pre else '1'
-    ans = input(f"Please enter the RPM release number of this release: [{release}] ")
-    if ans != '':
-        release = ans
-    if pre:
-        release += '.' + pre
-
-    finalversion = re.sub(r'pre\d+', '', version)
-    proto_changed = protocol_version != last_protocol_version
-    if proto_changed:
-        if finalversion in pdate:
-            proto_change_date = pdate[finalversion]
-        else:
-            while True:
-                ans = input("On what date did the protocol change to {protocol_version} get checked in? (dd Mmm yyyy) ")
-                if re.match(r'^\d\d \w\w\w \d\d\d\d$', ans):
-                    break
-            proto_change_date = ans
-    else:
-        proto_change_date = ' ' * 11
-
-    if 'pre' in lastversion:
-        if not pre:
-            die("You should not diff a release version against a pre-release version.")
-        srcdir = srcdiffdir = lastsrcdir = 'src-previews'
-        skipping = ' ** SKIPPING **'
-    elif pre:
-        srcdir = srcdiffdir = 'src-previews'
-        lastsrcdir = 'src'
-        skipping = ' ** SKIPPING **'
-    else:
-        srcdir = lastsrcdir = 'src'
-        srcdiffdir = 'src-diffs'
-        skipping = ''
-
-    print(f"""
-{dash_line}
-version is "{version}"
-lastversion is "{lastversion}"
-dest is "{dest}"
-curdir is "{curdir}"
-srcdir is "{srcdir}"
-srcdiffdir is "{srcdiffdir}"
-lastsrcdir is "{lastsrcdir}"
-release is "{release}"
-
-About to:
-    - tweak SUBPROTOCOL_VERSION in rsync.h, if needed
-    - tweak the version in version.h and the spec files
-    - tweak NEWS.md to ensure header values are correct
-    - generate configure.sh, config.h.in, and proto.h
-    - page through the differences
-""")
-    ans = input("<Press Enter to continue> ")
-
-    specvars = {
-        'Version:': finalversion,
-        'Release:': release,
-        '%define fullversion': f'%{{version}}{pre}',
-        'Released': version + '.',
-        '%define srcdir': srcdir,
-        }
-
-    tweak_files = 'version.h rsync.h'.split()
-    tweak_files += glob.glob('packaging/*.spec')
-    tweak_files += glob.glob('packaging/*/*.spec')
-
-    for fn in tweak_files:
-        with open(fn, 'r', encoding='utf-8') as fh:
-            old_txt = txt = fh.read()
-        if fn == 'version.h':
-            x_re = re.compile(r'^(#define RSYNC_VERSION).*', re.M)
-            msg = f"Unable to update RSYNC_VERSION in {fn}"
-            txt = replace_or_die(x_re, r'\1 "%s"' % version, txt, msg)
-            x_re = re.compile(r'^(#define MAINTAINER_TZ_OFFSET).*', re.M)
-            msg = f"Unable to update MAINTAINER_TZ_OFFSET in {fn}"
-            txt = replace_or_die(x_re, r'\1 ' + tz_num, txt, msg)
-        elif '.spec' in fn:
-            for var, val in specvars.items():
-                x_re = re.compile(r'^%s .*' % re.escape(var), re.M)
-                txt = replace_or_die(x_re, var + ' ' + val, txt, f"Unable to update {var} in {fn}")
-            x_re = re.compile(r'^\* \w\w\w \w\w\w \d\d \d\d\d\d (.*)', re.M)
-            txt = replace_or_die(x_re, r'%s \1' % cl_today, txt, f"Unable to update ChangeLog header in {fn}")
-        elif fn == 'rsync.h':
-            x_re = re.compile(r'(#define\s+SUBPROTOCOL_VERSION)\s+(\d+)')
-            repl = lambda m: m[1] + ' ' + ('0' if not pre or not proto_changed else '1' if m[2] == '0' else m[2])
-            txt = replace_or_die(x_re, repl, txt, f"Unable to find SUBPROTOCOL_VERSION define in {fn}")
-        elif fn == 'NEWS.md':
-            efv = re.escape(finalversion)
-            x_re = re.compile(r'^# NEWS for rsync %s \(UNRELEASED\)\s+## Changes in this version:\n' % efv
-                    + r'(\n### PROTOCOL NUMBER:\s+- The protocol number was changed to \d+\.\n)?')
-            rel_day = 'UNRELEASED' if pre else today
-            repl = (f'# NEWS for rsync {finalversion} ({rel_day})\n\n'
-                + '## Changes in this version:\n')
-            if proto_changed:
-                repl += f'\n### PROTOCOL NUMBER:\n\n - The protocol number was changed to {protocol_version}.\n'
-            good_top = re.sub(r'\(.*?\)', '(UNRELEASED)', repl, 1)
-            msg = f"The top lines of {fn} are not in the right format.  It should be:\n" + good_top
-            txt = replace_or_die(x_re, repl, txt, msg)
-            x_re = re.compile(r'^(\| )(\S{2} \S{3} \d{4})(\s+\|\s+%s\s+\| ).{11}(\s+\| )\S{2}(\s+\|+)$' % efv, re.M)
-            repl = lambda m: m[1] + (m[2] if pre else ztoday) + m[3] + proto_change_date + m[4] + protocol_version + m[5]
-            txt = replace_or_die(x_re, repl, txt, f'Unable to find "| ?? ??? {year} | {finalversion} | ... |" line in {fn}')
-        else:
-            die(f"Unrecognized file in tweak_files: {fn}")
-
-        if txt != old_txt:
-            print(f"Updating {fn}")
-            with open(fn, 'w', encoding='utf-8') as fh:
-                fh.write(txt)
-
-    cmd_chk(['packaging/year-tweak'])
-
-    print(dash_line)
-    cmd_run("git diff".split())
-
-    srctar_name = f"{rsync_ver}.tar.gz"
-    diff_name = f"{rsync_lastver}-{version}.diffs.gz"
-    srctar_file = os.path.join(dest, srcdir, srctar_name)
-    pattar_file = os.path.join(dest, srcdir, pattar_name)
-    diff_file = os.path.join(dest, srcdiffdir, diff_name)
-    lasttar_file = os.path.join(dest, lastsrcdir, rsync_lastver + '.tar.gz')
-
-    print(f"""\
-{dash_line}
-
-About to:
-    - git commit all changes
-    - run a full build, ensuring that the manpages & configure.sh are up-to-date
-    - merge the {args.master_branch} branch into the patch/{args.master_branch}/* branches
-""")
-    ans = input("<Press Enter OR 'y' to continue> ")
-
-    s = cmd_run(['git', 'commit', '-a', '-m', f'Preparing for release of {version} [buildall]'])
-    if s.returncode:
-        die('Aborting')
-
-    cmd_chk('touch configure.ac && packaging/smart-make && make gen')
-
-    print('Creating any missing patch branches.')
-    s = cmd_run(f'packaging/branch-from-patch --branch={args.master_branch} --add-missing')
-    if s.returncode:
-        die('Aborting')
-
-    if re.match(r'^y', ans, re.I):
-        print(f'\nRunning smart-make on all "patch/{args.master_branch}/*" branches ...')
-        cmd_run(f"packaging/patch-update --branch={args.master_branch} --skip-check --make")
-
-    print(f"""\
-{dash_line}
-
-About to:
-    - create signed tag for this release: {v_ver}
-    - create release diffs, "{diff_name}"
-    - create release tar, "{srctar_name}"
-    - update top-level README.md, NEWS.md, TODO, and ChangeLog
-    - update top-level rsync*.html manpages
-    - gpg-sign the release files
-    - update hard-linked top-level release files{skipping}
-""")
-    ans = input("<Press Enter to continue> ")
-
-    # TODO: is there a better way to ensure that our passphrase is in the agent?
-    cmd_run("touch TeMp; gpg --sign TeMp; rm TeMp*")
-
-    out = cmd_txt(f"git tag -s -m 'Version {version}.' {v_ver}", capture='combined').out
-    print(out, end='')
-    if 'bad passphrase' in out or 'failed' in out:
-        die('Aborting')
-
-    os.environ['PATH'] = ORIGINAL_PATH
-
-    # Extract the generated files from the old tar.
-    tweaked_gen_files = [ os.path.join(rsync_lastver, fn) for fn in gen_files ]
-    cmd_run(['tar', 'xzf', lasttar_file, *tweaked_gen_files])
-    os.rename(rsync_lastver, 'a')
-
-    print(f"Creating {diff_file} ...")
-    cmd_chk(['rsync', '-a', *gen_pathnames, 'b/'])
-
-    sed_script = r's:^((---|\+\+\+) [ab]/[^\t]+)\t.*:\1:' # CAUTION: must not contain any single quotes!
-    cmd_chk(f"(git diff v{lastversion} {v_ver} -- ':!.github'; diff -upN a b | sed -r '{sed_script}') | gzip -9 >{diff_file}")
-    shutil.rmtree('a')
-    os.rename('b', rsync_ver)
-
-    print(f"Creating {srctar_file} ...")
-    cmd_chk(f"git archive --format=tar --prefix={rsync_ver}/ {v_ver} | tar xf -")
-    cmd_chk(f"support/git-set-file-times --quiet --prefix={rsync_ver}/")
-    cmd_chk(['fakeroot', 'tar', 'czf', srctar_file, '--exclude=.github', rsync_ver])
-    shutil.rmtree(rsync_ver)
-
-    print(f"Updating the other files in {dest} ...")
-    md_files = 'README.md NEWS.md INSTALL.md'.split()
-    html_files = [ fn for fn in gen_pathnames if fn.endswith('.html') ]
-    cmd_chk(['rsync', '-a', *md_files, *html_files, dest])
-    cmd_chk(["./md-convert", "--dest", dest, *md_files])
-
-    cmd_chk(f"git log --name-status | gzip -9 >{dest}/ChangeLog.gz")
-
-    for fn in (srctar_file, pattar_file, diff_file):
-        asc_fn = fn + '.asc'
-        if os.path.lexists(asc_fn):
-            os.unlink(asc_fn)
-        res = cmd_run(['gpg', '--batch', '-ba', fn])
-        if res.returncode != 0 and res.returncode != 2:
-            die("gpg signing failed")
-
-    if not pre:
-        for find in f'{dest}/rsync-*.gz {dest}/rsync-*.asc {dest}/src-previews/rsync-*diffs.gz*'.split():
-            for fn in glob.glob(find):
-                os.unlink(fn)
-        top_link = [
-                srctar_file, f"{srctar_file}.asc",
-                pattar_file, f"{pattar_file}.asc",
-                diff_file, f"{diff_file}.asc",
-                ]
-        for fn in top_link:
-            os.link(fn, re.sub(r'/src(-\w+)?/', '/', fn))
-
-    print(f"""\
-{dash_line}
-
-Local changes are done.  When you're satisfied, push the git repository
-and rsync the release files.  Remember to announce the release on *BOTH*
-rsync-announce@lists.samba.org and rsync@lists.samba.org (and the web)!
-""")
-
-
-def replace_or_die(regex, repl, txt, die_msg):
-    m = regex.search(txt)
-    if not m:
-        die(die_msg)
-    return regex.sub(repl, txt, 1)
-
-
-def signal_handler(sig, frame):
-    die("\nAborting due to SIGINT.")
-
-
-if __name__ == '__main__':
-    parser = argparse.ArgumentParser(description="Prepare a new release of rsync in the git repo & ftp dir.", add_help=False)
-    parser.add_argument('--branch', '-b', dest='master_branch', default='master', help="The branch to release. Default: master.")
-    parser.add_argument("--help", "-h", action="help", help="Output this help message and exit.")
-    args = parser.parse_args()
-    main()
-
-# vim: sw=4 et ft=python

packaging/release.py

@@ -8,7 +8,7 @@
 # the rsync git checkout):
 #
 #   ../release/rsync-ftp/       mirror of samba.org:/home/ftp/pub/rsync
-#   ../release/rsync-html/      git checkout of rsync-web (the html site)
+#   ../release/rsync-html/      release-time snapshot of the html site
 #   ../release/work/            scratch space for tarball / diff staging
 #   ../release/release-state.json   info shared between steps
 #
@@ -35,10 +35,11 @@ HTML_DIR    = os.path.join(RELEASE_DIR, 'rsync-html')
 WORK_DIR    = os.path.join(RELEASE_DIR, 'work')
 STATE_FILE  = os.path.join(RELEASE_DIR, 'release-state.json')
 
-# Local rsync-web checkout (sibling of rsync-git) is the source-of-truth for
+# The rsync-web/ subdirectory in the rsync source tree is the source-of-truth
-# the git-tracked html content.  The maintainer pulls/commits/pushes there;
+# for the git-tracked html content.  step-1-fetch snapshots it into HTML_DIR
-# step-1-fetch just snapshots it into HTML_DIR for the release flow.
+# for the release flow, where it can be edited or augmented with server-side
-HTML_SRC = os.path.realpath('../rsync-web')
+# content before step-11-push-html sends it to samba.org.
+HTML_SRC = os.path.realpath('rsync-web')
 
 FTP_REMOTE_PATH  = '/home/ftp/pub/rsync'
 HTML_REMOTE_PATH = '/home/httpd/html/rsync'
@@ -60,7 +61,7 @@ GEN_FILES = [
 # ---------- Step registry ----------
 
 STEPS = [
-    ('step-1-fetch',       'mirror ../release/rsync-ftp from samba.org and snapshot ../release/rsync-html from ../rsync-web'),
+    ('step-1-fetch',       'mirror ../release/rsync-ftp from samba.org and snapshot ../release/rsync-html from rsync-web/'),
     ('step-2-prepare',     'gather release info interactively and write release-state.json'),
     ('step-3-tweak',       'update version.h, rsync.h, NEWS.md, and packaging/*.spec'),
     ('step-4-build',       'run smart-make + make gen'),
@@ -136,27 +137,29 @@ def step_1_fetch(args):
     section(f"Fetching ftp dir into {FTP_DIR}")
     if not os.path.isdir(FTP_DIR):
         os.makedirs(FTP_DIR)
-    # The .filt file lives in the ftp dir on the server; mirror down using the
+    # packaging/ftp.filt is the authoritative copy of the .filt filter file
-    # transmitted filter, falling back to no filter on the very first pull.
+    # that controls which subtrees rsync excludes from the FTP mirror.
+    # Seed FTP_DIR/.filt from it so the bundled version is what step-1's
+    # rsync uses here, and so step-10-push-ftp propagates it back to the
+    # server.  --exclude=/.filt below stops the server's copy from
+    # overwriting our bundled one on the way down.
     filt = os.path.join(FTP_DIR, '.filt')
-    if os.path.exists(filt):
+    bundled_filt = os.path.realpath('packaging/ftp.filt')
-        opts = ['-aivOHP', f'-f:_{filt}']
+    if not os.path.isfile(bundled_filt):
-    else:
+        die(f"{bundled_filt} not found; cannot seed .filt for the FTP pull.")
-        opts = ['-aivOHP']
+    shutil.copyfile(bundled_filt, filt)
-    cmd_chk(['rsync', *opts, f'{host}:{FTP_REMOTE_PATH}/', f'{FTP_DIR}/'])
+    cmd_chk(['rsync', '-aivOHP', f'-f:_{filt}', '--exclude=/.filt',
+             f'{host}:{FTP_REMOTE_PATH}/', f'{FTP_DIR}/'])
 
     section(f"Snapshotting html dir from {HTML_SRC} into {HTML_DIR}")
     if not os.path.isdir(HTML_SRC):
-        die(f"{HTML_SRC} not found.  Clone the rsync-web repo there first.")
+        die(f"{HTML_SRC} not found.  This should be the in-tree rsync-web/ "
-    if not os.path.isdir(os.path.join(HTML_SRC, '.git')):
+            f"subdirectory; something is wrong with your checkout.")
-        die(f"{HTML_SRC} exists but is not a git checkout.")
-    print(f"(Make sure {HTML_SRC} is up to date — this script does not 'git pull' for you.)")
     os.makedirs(HTML_DIR, exist_ok=True)
-    cmd_chk(['rsync', '-aiv', '--exclude=/.git',
+    cmd_chk(['rsync', '-aiv', f'{HTML_SRC}/', f'{HTML_DIR}/'])
-             f'{HTML_SRC}/', f'{HTML_DIR}/'])
 
-    # Then mirror non-git html content from the server (mirroring samba-rsync's
+    # Then mirror non-git html content from the server, skipping files that
-    # behavior: skip files that the html git already provides).
+    # the html git already provides (driven by the 'filt' file in HTML_DIR).
     filt = os.path.join(HTML_DIR, 'filt')
     if os.path.exists(filt):
         tmp_filt = os.path.join(HTML_DIR, 'tmp-filt')
@@ -631,9 +634,8 @@ If you have a 'samba' remote configured (git.samba.org:/data/git/rsync.git):
     git push samba {master_branch}
     git push samba {v_ver}
 
-Then upload the tarball + .asc to the GitHub release for {v_ver}, run
+Then upload the tarball + .asc to the GitHub release for {v_ver},
-packaging/send-news (when convenient), and announce on rsync-announce@,
+and announce on rsync-announce@, rsync@, and Discord.
-rsync@, and Discord.
 """)
 
 

packaging/samba-rsync

@@ -1,124 +0,0 @@
-#!/bin/bash
-# This script makes it easy to update the ftp & html directories on the samba.org server.
-# It expects the 2 *_DEST directories to contain updated files that need to be sent to
-# the remote server. If these directories don't exist yet, they will be copied from the
-# remote server (while also making the html dir a git checkout).
-
-FTP_SRC="$HOME/samba-rsync-ftp"
-HTML_SRC="$HOME/samba-rsync-html"
-
-FTP_DEST="/home/ftp/pub/rsync"
-HTML_DEST="/home/httpd/html/rsync"
-
-HTML_GIT='git.samba.org:/data/git/rsync-web.git'
-
-export RSYNC_PARTIAL_DIR=''
-
-case "$RSYNC_SAMBA_HOST" in
-    *.samba.org) ;;
-    *)
-	echo "You must set RSYNC_SAMBA_HOST in your environment to the samba hostname to use." >&2
-	exit 1
-	;;
-esac
-
-MODE=''
-REVERSE=''
-while (( $# )); do
-    case "$1" in
-	-R|--reverse) REVERSE=yes ;;
-	f|ftp) MODE=ftp ;;
-	h|html) MODE=html ;;
-	-h|--help)
-	    echo "Usage: [-R] [f|ftp|h|html]"
-	    echo "-R --reverse  Copy the files from the server to the local host."
-	    echo "              The default is to update the remote files."
-	    echo "-h --help     Output this help message."
-	    echo " "
-	    echo "The script will prompt if ftp or html is not specified on the command line."
-	    echo "Only one category can be copied at a time. When pulling html files, a git"
-	    echo "checkout will be either created or updated prior to the rsync copy."
-	    exit
-	    ;;
-	*)
-	    echo "Invalid option: $1" >&2
-	    exit 1
-	    ;;
-    esac
-    shift
-done
-
-while [ ! "$MODE" ]; do
-    if [ "$REVERSE" = yes ]; then
-	DIRECTION=FROM
-    else
-	DIRECTION=TO
-    fi
-    echo -n "Copy which files $DIRECTION the server? ftp or html? "
-    read ans
-    case "$ans" in
-	f*) MODE=ftp ;;
-	h*) MODE=html ;;
-	'') exit 1 ;;
-	*) echo "You must answer f or h to copy the ftp or html data." ;;
-    esac
-done
-
-if [ "$MODE" = ftp ]; then
-    SRC_DIR="$FTP_SRC"
-    DEST_DIR="$FTP_DEST"
-    FILT=".filt"
-else
-    SRC_DIR="$HTML_SRC"
-    DEST_DIR="$HTML_DEST"
-    FILT="filt"
-fi
-
-function do_rsync {
-    rsync --dry-run "${@}" | grep -v 'is uptodate$'
-    echo ''
-    echo -n "Run without --dry-run? [n] "
-    read ans
-    case "$ans" in
-	y*) rsync "${@}" | grep -v 'is uptodate$' ;;
-    esac
-}
-
-if [ -d "$SRC_DIR" ]; then
-    REVERSE_RSYNC=do_rsync
-else
-    echo "The directory $SRC_DIR does not exist yet."
-    echo -n "Do you want to create it? [n] "
-    read ans
-    case "$ans" in
-	y*) ;;
-	*) exit 1 ;;
-    esac
-    REVERSE=yes
-    REVERSE_RSYNC=rsync
-fi
-
-if [ "$REVERSE" = yes ]; then
-    OPTS='-aivOHP'
-    TMP_FILT="$SRC_DIR/tmp-filt"
-    echo "Copying files from $RSYNC_SAMBA_HOST to $SRC_DIR ..."
-    if [ "$MODE" = html ]; then
-	if [ $REVERSE_RSYNC = rsync ]; then
-	    git clone "$HTML_GIT" "$SRC_DIR" || exit 1
-	else
-	    cd "$SRC_DIR" || exit 1
-	    git pull || exit 1
-	fi
-	sed -n -e 's/[-P]/H/p' "$SRC_DIR/$FILT" >"$TMP_FILT"
-	OPTS="${OPTS}f._$TMP_FILT"
-    else
-	OPTS="${OPTS}f:_$FILT"
-    fi
-    $REVERSE_RSYNC "$OPTS" "$RSYNC_SAMBA_HOST:$DEST_DIR/" "$SRC_DIR/"
-    rm -f "$TMP_FILT"
-    exit
-fi
-
-cd "$SRC_DIR" || exit 1
-echo "Copying files from $SRC_DIR to $RSYNC_SAMBA_HOST ..."
-do_rsync -aivOHP --chown=:rsync --del -f._$FILT . "$RSYNC_SAMBA_HOST:$DEST_DIR/"

packaging/send-news

@@ -1,33 +0,0 @@
-#!/bin/bash -e
-
-# This script expects the ~/src/rsync directory to contain the rsync
-# source that has been updated. It also expects the auto-build-save
-# directory to have been created prior to the running of configure so
-# that each branch has its own build directory underneath. This supports
-# the maintainer workflow for the rsync-patches files maintenace.
-
-FTP_SRC="$HOME/samba-rsync-ftp"
-FTP_DEST="/home/ftp/pub/rsync"
-MD_FILES="README.md INSTALL.md NEWS.md"
-
-case "$RSYNC_SAMBA_HOST" in
-    *.samba.org) ;;
-    *)
-	echo "You must set RSYNC_SAMBA_HOST in your environment to the samba hostname to use." >&2
-	exit 1
-	;;
-esac
-
-if [ ! -d "$FTP_SRC" ]; then
-    packaging/samba-rsync ftp # Ask to initialize the local ftp dir
-fi
-
-cd ~/src/rsync
-
-make man
-./md-convert --dest="$FTP_SRC" $MD_FILES
-rsync -aiic $MD_FILES auto-build-save/master/*.?.html "$FTP_SRC"
-
-cd "$FTP_SRC"
-
-rsync -aiic README.* INSTALL.* NEWS.* *.?.html "$RSYNC_SAMBA_HOST:$FTP_DEST/"

receiver.c

@@ -70,6 +70,7 @@ extern int fuzzy_basis;
 
 extern struct name_num_item *xfer_sum_nni;
 extern int xfer_sum_len;
+extern int use_secure_symlinks;
 
 static struct bitbag *delayed_bits = NULL;
 static int phase = 0, redoing = 0;
@@ -82,6 +83,44 @@ static int updating_basis_or_equiv;
 #define MAX_UNIQUE_NUMBER 999999
 #define MAX_UNIQUE_LOOP 100
 
+/* Open a basis/output path that may legitimately be an operator-trusted
+ * ABSOLUTE path -- e.g. an absolute --partial-dir ("a directory reserved for
+ * partial-dir work") or --backup-dir. secure_relative_open() deliberately
+ * rejects an absolute relpath, so feeding it the whole absolute partialptr
+ * (with a NULL basedir) returns EINVAL: the basis fd is then -1, no basis is
+ * mapped, and receive_data() omits every matched block from the whole-file
+ * verification checksum -> a spurious "failed verification" that strands the
+ * (correct) data in the partial-dir forever.
+ *
+ * The operator's directory is trusted; only the leaf basename is peer-supplied.
+ * So when basedir is NULL and relpath is absolute, split it into its directory
+ * (trusted) and leaf and confine just the leaf -- exactly how secure_relative_
+ * open already trusts an absolute basedir while O_NOFOLLOW-confining the leaf.
+ * Anything else is a straight pass-through that preserves the strict contract. */
+static int secure_basis_open(const char *basedir, const char *relpath, int flags, mode_t mode)
+{
+	if (!basedir && relpath && *relpath == '/') {
+		const char *slash = strrchr(relpath, '/');
+		const char *leaf = slash + 1;
+		char dirbuf[MAXPATHLEN];
+		const char *dir;
+		if (slash == relpath)
+			dir = "/";
+		else {
+			size_t dlen = slash - relpath;
+			if (dlen >= sizeof dirbuf) {
+				errno = ENAMETOOLONG;
+				return -1;
+			}
+			memcpy(dirbuf, relpath, dlen);
+			dirbuf[dlen] = '\0';
+			dir = dirbuf;
+		}
+		return secure_relative_open(dir, leaf, flags, mode);
+	}
+	return secure_relative_open(basedir, relpath, flags, mode);
+}
+
 /* get_tmpname() - create a tmp filename for a given filename
  *
  * If a tmpdir is defined, use that as the directory to put it in.  Otherwise,
@@ -214,7 +253,12 @@ int open_tmpfile(char *fnametmp, const char *fname, struct file_struct *file)
 	 * access to ensure that there is no race condition.  They will be
 	 * correctly updated after the right owner and group info is set.
 	 * (Thanks to snabb@epipe.fi for pointing this out.) */
-	fd = do_mkstemp(fnametmp, (file->mode|added_perms) & INITACCESSPERMS);
+	/* When use_secure_symlinks is on (non-chroot daemon with munge_symlinks),
+	 * use secure_mkstemp to prevent symlink race attacks on parent directories. */
+	if (use_secure_symlinks)
+		fd = secure_mkstemp(fnametmp, (file->mode|added_perms) & INITACCESSPERMS);
+	else
+		fd = do_mkstemp(fnametmp, (file->mode|added_perms) & INITACCESSPERMS);
 
 #if 0
 	/* In most cases parent directories will already exist because their
@@ -312,7 +356,12 @@ static int receive_data(int f_in, char *fname_r, int fd_r, OFF_T size_r,
 		}
 	}
 
-	while ((i = recv_token(f_in, &data)) != 0) {
+	while (1) {
+		data = NULL;
+		i = recv_token(f_in, &data);
+		if (i == 0)
+			break;
+
 		if (INFO_GTE(PROGRESS, 1))
 			show_progress(offset, total_size);
 
@@ -320,6 +369,10 @@ static int receive_data(int f_in, char *fname_r, int fd_r, OFF_T size_r,
 			maybe_send_keepalive(time(NULL), MSK_ALLOW_FLUSH | MSK_ACTIVE_RECEIVER);
 
 		if (i > 0) {
+			if (!data) {
+				rprintf(FERROR, "Invalid literal token with no data [%s]\n", who_am_i());
+				exit_cleanup(RERR_PROTOCOL);
+			}
 			if (DEBUG_GTE(DELTASUM, 3)) {
 				rprintf(FINFO,"data recv %d at %s\n",
 					i, big_num(offset));
@@ -337,6 +390,11 @@ static int receive_data(int f_in, char *fname_r, int fd_r, OFF_T size_r,
 		}
 
 		i = -(i+1);
+		if (i < 0 || i >= sum.count) {
+			rprintf(FERROR, "Invalid block index %d (count=%ld) [%s]\n",
+				i, (long)sum.count, who_am_i());
+			exit_cleanup(RERR_PROTOCOL);
+		}
 		offset2 = i * (OFF_T)sum.blength;
 		len = sum.blength;
 		if (i == (int)sum.count-1 && sum.remainder != 0)
@@ -344,6 +402,18 @@ static int receive_data(int f_in, char *fname_r, int fd_r, OFF_T size_r,
 
 		stats.matched_data += len;
 
+		/* A block match can only be honored if we actually mapped the
+		 * basis. If we didn't (basis open failed), the sender should
+		 * never have been told a basis existed -- treat it as a protocol
+		 * inconsistency rather than silently omitting these bytes from
+		 * the verification checksum (which yields a spurious failure) or
+		 * leaving a hole in the output. */
+		if (!mapbuf) {
+			rprintf(FERROR, "got a block match with no basis file for %s [%s]\n",
+				full_fname(fname), who_am_i());
+			exit_cleanup(RERR_PROTOCOL);
+		}
+
 		if (DEBUG_GTE(DELTASUM, 3)) {
 			rprintf(FINFO,
 				"chunk[%d] of size %ld at %s offset=%s%s\n",
@@ -436,7 +506,7 @@ static void handle_delayed_updates(char *local_name)
 			}
 			/* We don't use robust_rename() here because the
 			 * partial-dir must be on the same drive. */
-			if (do_rename(partialptr, fname) < 0) {
+			if (do_rename_at(partialptr, fname) < 0) {
 				rsyserr(FERROR_XFER, errno,
 					"rename failed for %s (from %s)",
 					full_fname(fname), partialptr);
@@ -452,7 +522,10 @@ static void handle_delayed_updates(char *local_name)
 static void no_batched_update(int ndx, BOOL is_redo)
 {
 	struct file_list *flist = flist_for_ndx(ndx, "no_batched_update");
-	struct file_struct *file = flist->files[ndx - flist->ndx_start];
+	struct file_struct *file;
+	if (ndx < flist->ndx_start)
+		exit_cleanup(RERR_PROTOCOL);
+	file = flist->files[ndx - flist->ndx_start];
 
 	rprintf(FERROR_XFER, "(No batched update for%s \"%s\")\n",
 		is_redo ? " resend of" : "", f_name(file, NULL));
@@ -589,6 +662,8 @@ int recv_files(int f_in, int f_out, char *local_name)
 
 		if (ndx - cur_flist->ndx_start >= 0)
 			file = cur_flist->files[ndx - cur_flist->ndx_start];
+		else if (cur_flist->parent_ndx < 0)
+			exit_cleanup(RERR_PROTOCOL);
 		else
 			file = dir_flist->files[cur_flist->parent_ndx];
 		fname = local_name ? local_name : f_name(file, fbuf);
@@ -768,8 +843,9 @@ int recv_files(int f_in, int f_out, char *local_name)
 				fnamecmp = fname;
 		}
 
-		/* open the file */
+		/* open the file (secure_basis_open tolerates an operator-trusted
-		fd1 = secure_relative_open(basedir, fnamecmp, O_RDONLY, 0);
+		 * absolute fnamecmp, e.g. an absolute --partial-dir basis) */
+		fd1 = secure_basis_open(basedir, fnamecmp, O_RDONLY, 0);
 
 		if (fd1 == -1 && protocol_version < 29) {
 			if (fnamecmp != fname) {
@@ -854,11 +930,21 @@ int recv_files(int f_in, int f_out, char *local_name)
 		/* We now check to see if we are writing the file "inplace" */
 		if (inplace || one_inplace)  {
 			fnametmp = one_inplace ? partialptr : fname;
-			fd2 = do_open(fnametmp, O_WRONLY|O_CREAT, 0600);
+			/* When use_secure_symlinks is on (non-chroot daemon),
+			 * use secure open to prevent symlink race attacks where an
+			 * attacker could switch a directory to a symlink between
+			 * path validation and file open. */
+			if (use_secure_symlinks)
+				fd2 = secure_basis_open(NULL, fnametmp, O_WRONLY|O_CREAT, 0600);
+			else
+				fd2 = do_open(fnametmp, O_WRONLY|O_CREAT, 0600);
 #ifdef linux
 			if (fd2 == -1 && errno == EACCES) {
 				/* Maybe the error was due to protected_regular setting? */
-				fd2 = do_open(fname, O_WRONLY, 0600);
+				if (use_secure_symlinks)
+					fd2 = secure_relative_open(NULL, fname, O_WRONLY, 0600);
+				else
+					fd2 = do_open(fname, O_WRONLY, 0600);
 			}
 #endif
 			if (fd2 == -1) {
@@ -910,7 +996,7 @@ int recv_files(int f_in, int f_out, char *local_name)
 				recv_ok = -1;
 			else if (fnamecmp == partialptr) {
 				if (!one_inplace)
-					do_unlink(partialptr);
+					do_unlink_at(partialptr);
 				handle_partial_dir(partialptr, PDIR_DELETE);
 			}
 		} else if (keep_partial && partialptr && (!one_inplace || delay_updates)) {
@@ -919,7 +1005,7 @@ int recv_files(int f_in, int f_out, char *local_name)
 					"Unable to create partial-dir for %s -- discarding %s.\n",
 					local_name ? local_name : f_name(file, NULL),
 					recv_ok ? "completed file" : "partial file");
-				do_unlink(fnametmp);
+				do_unlink_at(fnametmp);
 				recv_ok = -1;
 			} else if (!finish_transfer(partialptr, fnametmp, fnamecmp, NULL,
 						    file, recv_ok, !partial_dir))
@@ -930,7 +1016,7 @@ int recv_files(int f_in, int f_out, char *local_name)
 			} else
 				partialptr = NULL;
 		} else if (!one_inplace)
-			do_unlink(fnametmp);
+			do_unlink_at(fnametmp);
 
 		cleanup_disable();
 

rsync-web/.gitignore

@@ -0,0 +1,11 @@
+/.xvpics
+/doxygen/
+/tech_report/IMG_PARAMS.dir
+/tech_report/IMG_PARAMS.pag
+/upload
+/netware
+/pre-change
+/index.html-*
+/rsync-and-debian/rsync-and-debian.html
+/rsync-and-debian/rsync-and-debian.ps
+/badge.svg

rsync-web/COPYING.html

@@ -0,0 +1,674 @@
+                    GNU GENERAL PUBLIC LICENSE
+                       Version 3, 29 June 2007
+
+ Copyright (C) 2007 Free Software Foundation, Inc. &lt;<a href="https://fsf.org/">https://fsf.org/</a>&gt;
+ Everyone is permitted to copy and distribute verbatim copies
+ of this license document, but changing it is not allowed.
+
+                            Preamble
+
+  The GNU General Public License is a free, copyleft license for
+software and other kinds of works.
+
+  The licenses for most software and other practical works are designed
+to take away your freedom to share and change the works.  By contrast,
+the GNU General Public License is intended to guarantee your freedom to
+share and change all versions of a program--to make sure it remains free
+software for all its users.  We, the Free Software Foundation, use the
+GNU General Public License for most of our software; it applies also to
+any other work released this way by its authors.  You can apply it to
+your programs, too.
+
+  When we speak of free software, we are referring to freedom, not
+price.  Our General Public Licenses are designed to make sure that you
+have the freedom to distribute copies of free software (and charge for
+them if you wish), that you receive source code or can get it if you
+want it, that you can change the software or use pieces of it in new
+free programs, and that you know you can do these things.
+
+  To protect your rights, we need to prevent others from denying you
+these rights or asking you to surrender the rights.  Therefore, you have
+certain responsibilities if you distribute copies of the software, or if
+you modify it: responsibilities to respect the freedom of others.
+
+  For example, if you distribute copies of such a program, whether
+gratis or for a fee, you must pass on to the recipients the same
+freedoms that you received.  You must make sure that they, too, receive
+or can get the source code.  And you must show them these terms so they
+know their rights.
+
+  Developers that use the GNU GPL protect your rights with two steps:
+(1) assert copyright on the software, and (2) offer you this License
+giving you legal permission to copy, distribute and/or modify it.
+
+  For the developers' and authors' protection, the GPL clearly explains
+that there is no warranty for this free software.  For both users' and
+authors' sake, the GPL requires that modified versions be marked as
+changed, so that their problems will not be attributed erroneously to
+authors of previous versions.
+
+  Some devices are designed to deny users access to install or run
+modified versions of the software inside them, although the manufacturer
+can do so.  This is fundamentally incompatible with the aim of
+protecting users' freedom to change the software.  The systematic
+pattern of such abuse occurs in the area of products for individuals to
+use, which is precisely where it is most unacceptable.  Therefore, we
+have designed this version of the GPL to prohibit the practice for those
+products.  If such problems arise substantially in other domains, we
+stand ready to extend this provision to those domains in future versions
+of the GPL, as needed to protect the freedom of users.
+
+  Finally, every program is threatened constantly by software patents.
+States should not allow patents to restrict development and use of
+software on general-purpose computers, but in those that do, we wish to
+avoid the special danger that patents applied to a free program could
+make it effectively proprietary.  To prevent this, the GPL assures that
+patents cannot be used to render the program non-free.
+
+  The precise terms and conditions for copying, distribution and
+modification follow.
+
+                       TERMS AND CONDITIONS
+
+  0. Definitions.
+
+  "This License" refers to version 3 of the GNU General Public License.
+
+  "Copyright" also means copyright-like laws that apply to other kinds of
+works, such as semiconductor masks.
+
+  "The Program" refers to any copyrightable work licensed under this
+License.  Each licensee is addressed as "you".  "Licensees" and
+"recipients" may be individuals or organizations.
+
+  To "modify" a work means to copy from or adapt all or part of the work
+in a fashion requiring copyright permission, other than the making of an
+exact copy.  The resulting work is called a "modified version" of the
+earlier work or a work "based on" the earlier work.
+
+  A "covered work" means either the unmodified Program or a work based
+on the Program.
+
+  To "propagate" a work means to do anything with it that, without
+permission, would make you directly or secondarily liable for
+infringement under applicable copyright law, except executing it on a
+computer or modifying a private copy.  Propagation includes copying,
+distribution (with or without modification), making available to the
+public, and in some countries other activities as well.
+
+  To "convey" a work means any kind of propagation that enables other
+parties to make or receive copies.  Mere interaction with a user through
+a computer network, with no transfer of a copy, is not conveying.
+
+  An interactive user interface displays "Appropriate Legal Notices"
+to the extent that it includes a convenient and prominently visible
+feature that (1) displays an appropriate copyright notice, and (2)
+tells the user that there is no warranty for the work (except to the
+extent that warranties are provided), that licensees may convey the
+work under this License, and how to view a copy of this License.  If
+the interface presents a list of user commands or options, such as a
+menu, a prominent item in the list meets this criterion.
+
+  1. Source Code.
+
+  The "source code" for a work means the preferred form of the work
+for making modifications to it.  "Object code" means any non-source
+form of a work.
+
+  A "Standard Interface" means an interface that either is an official
+standard defined by a recognized standards body, or, in the case of
+interfaces specified for a particular programming language, one that
+is widely used among developers working in that language.
+
+  The "System Libraries" of an executable work include anything, other
+than the work as a whole, that (a) is included in the normal form of
+packaging a Major Component, but which is not part of that Major
+Component, and (b) serves only to enable use of the work with that
+Major Component, or to implement a Standard Interface for which an
+implementation is available to the public in source code form.  A
+"Major Component", in this context, means a major essential component
+(kernel, window system, and so on) of the specific operating system
+(if any) on which the executable work runs, or a compiler used to
+produce the work, or an object code interpreter used to run it.
+
+  The "Corresponding Source" for a work in object code form means all
+the source code needed to generate, install, and (for an executable
+work) run the object code and to modify the work, including scripts to
+control those activities.  However, it does not include the work's
+System Libraries, or general-purpose tools or generally available free
+programs which are used unmodified in performing those activities but
+which are not part of the work.  For example, Corresponding Source
+includes interface definition files associated with source files for
+the work, and the source code for shared libraries and dynamically
+linked subprograms that the work is specifically designed to require,
+such as by intimate data communication or control flow between those
+subprograms and other parts of the work.
+
+  The Corresponding Source need not include anything that users
+can regenerate automatically from other parts of the Corresponding
+Source.
+
+  The Corresponding Source for a work in source code form is that
+same work.
+
+  2. Basic Permissions.
+
+  All rights granted under this License are granted for the term of
+copyright on the Program, and are irrevocable provided the stated
+conditions are met.  This License explicitly affirms your unlimited
+permission to run the unmodified Program.  The output from running a
+covered work is covered by this License only if the output, given its
+content, constitutes a covered work.  This License acknowledges your
+rights of fair use or other equivalent, as provided by copyright law.
+
+  You may make, run and propagate covered works that you do not
+convey, without conditions so long as your license otherwise remains
+in force.  You may convey covered works to others for the sole purpose
+of having them make modifications exclusively for you, or provide you
+with facilities for running those works, provided that you comply with
+the terms of this License in conveying all material for which you do
+not control copyright.  Those thus making or running the covered works
+for you must do so exclusively on your behalf, under your direction
+and control, on terms that prohibit them from making any copies of
+your copyrighted material outside their relationship with you.
+
+  Conveying under any other circumstances is permitted solely under
+the conditions stated below.  Sublicensing is not allowed; section 10
+makes it unnecessary.
+
+  3. Protecting Users' Legal Rights From Anti-Circumvention Law.
+
+  No covered work shall be deemed part of an effective technological
+measure under any applicable law fulfilling obligations under article
+11 of the WIPO copyright treaty adopted on 20 December 1996, or
+similar laws prohibiting or restricting circumvention of such
+measures.
+
+  When you convey a covered work, you waive any legal power to forbid
+circumvention of technological measures to the extent such circumvention
+is effected by exercising rights under this License with respect to
+the covered work, and you disclaim any intention to limit operation or
+modification of the work as a means of enforcing, against the work's
+users, your or third parties' legal rights to forbid circumvention of
+technological measures.
+
+  4. Conveying Verbatim Copies.
+
+  You may convey verbatim copies of the Program's source code as you
+receive it, in any medium, provided that you conspicuously and
+appropriately publish on each copy an appropriate copyright notice;
+keep intact all notices stating that this License and any
+non-permissive terms added in accord with section 7 apply to the code;
+keep intact all notices of the absence of any warranty; and give all
+recipients a copy of this License along with the Program.
+
+  You may charge any price or no price for each copy that you convey,
+and you may offer support or warranty protection for a fee.
+
+  5. Conveying Modified Source Versions.
+
+  You may convey a work based on the Program, or the modifications to
+produce it from the Program, in the form of source code under the
+terms of section 4, provided that you also meet all of these conditions:
+
+    a) The work must carry prominent notices stating that you modified
+    it, and giving a relevant date.
+
+    b) The work must carry prominent notices stating that it is
+    released under this License and any conditions added under section
+    7.  This requirement modifies the requirement in section 4 to
+    "keep intact all notices".
+
+    c) You must license the entire work, as a whole, under this
+    License to anyone who comes into possession of a copy.  This
+    License will therefore apply, along with any applicable section 7
+    additional terms, to the whole of the work, and all its parts,
+    regardless of how they are packaged.  This License gives no
+    permission to license the work in any other way, but it does not
+    invalidate such permission if you have separately received it.
+
+    d) If the work has interactive user interfaces, each must display
+    Appropriate Legal Notices; however, if the Program has interactive
+    interfaces that do not display Appropriate Legal Notices, your
+    work need not make them do so.
+
+  A compilation of a covered work with other separate and independent
+works, which are not by their nature extensions of the covered work,
+and which are not combined with it such as to form a larger program,
+in or on a volume of a storage or distribution medium, is called an
+"aggregate" if the compilation and its resulting copyright are not
+used to limit the access or legal rights of the compilation's users
+beyond what the individual works permit.  Inclusion of a covered work
+in an aggregate does not cause this License to apply to the other
+parts of the aggregate.
+
+  6. Conveying Non-Source Forms.
+
+  You may convey a covered work in object code form under the terms
+of sections 4 and 5, provided that you also convey the
+machine-readable Corresponding Source under the terms of this License,
+in one of these ways:
+
+    a) Convey the object code in, or embodied in, a physical product
+    (including a physical distribution medium), accompanied by the
+    Corresponding Source fixed on a durable physical medium
+    customarily used for software interchange.
+
+    b) Convey the object code in, or embodied in, a physical product
+    (including a physical distribution medium), accompanied by a
+    written offer, valid for at least three years and valid for as
+    long as you offer spare parts or customer support for that product
+    model, to give anyone who possesses the object code either (1) a
+    copy of the Corresponding Source for all the software in the
+    product that is covered by this License, on a durable physical
+    medium customarily used for software interchange, for a price no
+    more than your reasonable cost of physically performing this
+    conveying of source, or (2) access to copy the
+    Corresponding Source from a network server at no charge.
+
+    c) Convey individual copies of the object code with a copy of the
+    written offer to provide the Corresponding Source.  This
+    alternative is allowed only occasionally and noncommercially, and
+    only if you received the object code with such an offer, in accord
+    with subsection 6b.
+
+    d) Convey the object code by offering access from a designated
+    place (gratis or for a charge), and offer equivalent access to the
+    Corresponding Source in the same way through the same place at no
+    further charge.  You need not require recipients to copy the
+    Corresponding Source along with the object code.  If the place to
+    copy the object code is a network server, the Corresponding Source
+    may be on a different server (operated by you or a third party)
+    that supports equivalent copying facilities, provided you maintain
+    clear directions next to the object code saying where to find the
+    Corresponding Source.  Regardless of what server hosts the
+    Corresponding Source, you remain obligated to ensure that it is
+    available for as long as needed to satisfy these requirements.
+
+    e) Convey the object code using peer-to-peer transmission, provided
+    you inform other peers where the object code and Corresponding
+    Source of the work are being offered to the general public at no
+    charge under subsection 6d.
+
+  A separable portion of the object code, whose source code is excluded
+from the Corresponding Source as a System Library, need not be
+included in conveying the object code work.
+
+  A "User Product" is either (1) a "consumer product", which means any
+tangible personal property which is normally used for personal, family,
+or household purposes, or (2) anything designed or sold for incorporation
+into a dwelling.  In determining whether a product is a consumer product,
+doubtful cases shall be resolved in favor of coverage.  For a particular
+product received by a particular user, "normally used" refers to a
+typical or common use of that class of product, regardless of the status
+of the particular user or of the way in which the particular user
+actually uses, or expects or is expected to use, the product.  A product
+is a consumer product regardless of whether the product has substantial
+commercial, industrial or non-consumer uses, unless such uses represent
+the only significant mode of use of the product.
+
+  "Installation Information" for a User Product means any methods,
+procedures, authorization keys, or other information required to install
+and execute modified versions of a covered work in that User Product from
+a modified version of its Corresponding Source.  The information must
+suffice to ensure that the continued functioning of the modified object
+code is in no case prevented or interfered with solely because
+modification has been made.
+
+  If you convey an object code work under this section in, or with, or
+specifically for use in, a User Product, and the conveying occurs as
+part of a transaction in which the right of possession and use of the
+User Product is transferred to the recipient in perpetuity or for a
+fixed term (regardless of how the transaction is characterized), the
+Corresponding Source conveyed under this section must be accompanied
+by the Installation Information.  But this requirement does not apply
+if neither you nor any third party retains the ability to install
+modified object code on the User Product (for example, the work has
+been installed in ROM).
+
+  The requirement to provide Installation Information does not include a
+requirement to continue to provide support service, warranty, or updates
+for a work that has been modified or installed by the recipient, or for
+the User Product in which it has been modified or installed.  Access to a
+network may be denied when the modification itself materially and
+adversely affects the operation of the network or violates the rules and
+protocols for communication across the network.
+
+  Corresponding Source conveyed, and Installation Information provided,
+in accord with this section must be in a format that is publicly
+documented (and with an implementation available to the public in
+source code form), and must require no special password or key for
+unpacking, reading or copying.
+
+  7. Additional Terms.
+
+  "Additional permissions" are terms that supplement the terms of this
+License by making exceptions from one or more of its conditions.
+Additional permissions that are applicable to the entire Program shall
+be treated as though they were included in this License, to the extent
+that they are valid under applicable law.  If additional permissions
+apply only to part of the Program, that part may be used separately
+under those permissions, but the entire Program remains governed by
+this License without regard to the additional permissions.
+
+  When you convey a copy of a covered work, you may at your option
+remove any additional permissions from that copy, or from any part of
+it.  (Additional permissions may be written to require their own
+removal in certain cases when you modify the work.)  You may place
+additional permissions on material, added by you to a covered work,
+for which you have or can give appropriate copyright permission.
+
+  Notwithstanding any other provision of this License, for material you
+add to a covered work, you may (if authorized by the copyright holders of
+that material) supplement the terms of this License with terms:
+
+    a) Disclaiming warranty or limiting liability differently from the
+    terms of sections 15 and 16 of this License; or
+
+    b) Requiring preservation of specified reasonable legal notices or
+    author attributions in that material or in the Appropriate Legal
+    Notices displayed by works containing it; or
+
+    c) Prohibiting misrepresentation of the origin of that material, or
+    requiring that modified versions of such material be marked in
+    reasonable ways as different from the original version; or
+
+    d) Limiting the use for publicity purposes of names of licensors or
+    authors of the material; or
+
+    e) Declining to grant rights under trademark law for use of some
+    trade names, trademarks, or service marks; or
+
+    f) Requiring indemnification of licensors and authors of that
+    material by anyone who conveys the material (or modified versions of
+    it) with contractual assumptions of liability to the recipient, for
+    any liability that these contractual assumptions directly impose on
+    those licensors and authors.
+
+  All other non-permissive additional terms are considered "further
+restrictions" within the meaning of section 10.  If the Program as you
+received it, or any part of it, contains a notice stating that it is
+governed by this License along with a term that is a further
+restriction, you may remove that term.  If a license document contains
+a further restriction but permits relicensing or conveying under this
+License, you may add to a covered work material governed by the terms
+of that license document, provided that the further restriction does
+not survive such relicensing or conveying.
+
+  If you add terms to a covered work in accord with this section, you
+must place, in the relevant source files, a statement of the
+additional terms that apply to those files, or a notice indicating
+where to find the applicable terms.
+
+  Additional terms, permissive or non-permissive, may be stated in the
+form of a separately written license, or stated as exceptions;
+the above requirements apply either way.
+
+  8. Termination.
+
+  You may not propagate or modify a covered work except as expressly
+provided under this License.  Any attempt otherwise to propagate or
+modify it is void, and will automatically terminate your rights under
+this License (including any patent licenses granted under the third
+paragraph of section 11).
+
+  However, if you cease all violation of this License, then your
+license from a particular copyright holder is reinstated (a)
+provisionally, unless and until the copyright holder explicitly and
+finally terminates your license, and (b) permanently, if the copyright
+holder fails to notify you of the violation by some reasonable means
+prior to 60 days after the cessation.
+
+  Moreover, your license from a particular copyright holder is
+reinstated permanently if the copyright holder notifies you of the
+violation by some reasonable means, this is the first time you have
+received notice of violation of this License (for any work) from that
+copyright holder, and you cure the violation prior to 30 days after
+your receipt of the notice.
+
+  Termination of your rights under this section does not terminate the
+licenses of parties who have received copies or rights from you under
+this License.  If your rights have been terminated and not permanently
+reinstated, you do not qualify to receive new licenses for the same
+material under section 10.
+
+  9. Acceptance Not Required for Having Copies.
+
+  You are not required to accept this License in order to receive or
+run a copy of the Program.  Ancillary propagation of a covered work
+occurring solely as a consequence of using peer-to-peer transmission
+to receive a copy likewise does not require acceptance.  However,
+nothing other than this License grants you permission to propagate or
+modify any covered work.  These actions infringe copyright if you do
+not accept this License.  Therefore, by modifying or propagating a
+covered work, you indicate your acceptance of this License to do so.
+
+  10. Automatic Licensing of Downstream Recipients.
+
+  Each time you convey a covered work, the recipient automatically
+receives a license from the original licensors, to run, modify and
+propagate that work, subject to this License.  You are not responsible
+for enforcing compliance by third parties with this License.
+
+  An "entity transaction" is a transaction transferring control of an
+organization, or substantially all assets of one, or subdividing an
+organization, or merging organizations.  If propagation of a covered
+work results from an entity transaction, each party to that
+transaction who receives a copy of the work also receives whatever
+licenses to the work the party's predecessor in interest had or could
+give under the previous paragraph, plus a right to possession of the
+Corresponding Source of the work from the predecessor in interest, if
+the predecessor has it or can get it with reasonable efforts.
+
+  You may not impose any further restrictions on the exercise of the
+rights granted or affirmed under this License.  For example, you may
+not impose a license fee, royalty, or other charge for exercise of
+rights granted under this License, and you may not initiate litigation
+(including a cross-claim or counterclaim in a lawsuit) alleging that
+any patent claim is infringed by making, using, selling, offering for
+sale, or importing the Program or any portion of it.
+
+  11. Patents.
+
+  A "contributor" is a copyright holder who authorizes use under this
+License of the Program or a work on which the Program is based.  The
+work thus licensed is called the contributor's "contributor version".
+
+  A contributor's "essential patent claims" are all patent claims
+owned or controlled by the contributor, whether already acquired or
+hereafter acquired, that would be infringed by some manner, permitted
+by this License, of making, using, or selling its contributor version,
+but do not include claims that would be infringed only as a
+consequence of further modification of the contributor version.  For
+purposes of this definition, "control" includes the right to grant
+patent sublicenses in a manner consistent with the requirements of
+this License.
+
+  Each contributor grants you a non-exclusive, worldwide, royalty-free
+patent license under the contributor's essential patent claims, to
+make, use, sell, offer for sale, import and otherwise run, modify and
+propagate the contents of its contributor version.
+
+  In the following three paragraphs, a "patent license" is any express
+agreement or commitment, however denominated, not to enforce a patent
+(such as an express permission to practice a patent or covenant not to
+sue for patent infringement).  To "grant" such a patent license to a
+party means to make such an agreement or commitment not to enforce a
+patent against the party.
+
+  If you convey a covered work, knowingly relying on a patent license,
+and the Corresponding Source of the work is not available for anyone
+to copy, free of charge and under the terms of this License, through a
+publicly available network server or other readily accessible means,
+then you must either (1) cause the Corresponding Source to be so
+available, or (2) arrange to deprive yourself of the benefit of the
+patent license for this particular work, or (3) arrange, in a manner
+consistent with the requirements of this License, to extend the patent
+license to downstream recipients.  "Knowingly relying" means you have
+actual knowledge that, but for the patent license, your conveying the
+covered work in a country, or your recipient's use of the covered work
+in a country, would infringe one or more identifiable patents in that
+country that you have reason to believe are valid.
+
+  If, pursuant to or in connection with a single transaction or
+arrangement, you convey, or propagate by procuring conveyance of, a
+covered work, and grant a patent license to some of the parties
+receiving the covered work authorizing them to use, propagate, modify
+or convey a specific copy of the covered work, then the patent license
+you grant is automatically extended to all recipients of the covered
+work and works based on it.
+
+  A patent license is "discriminatory" if it does not include within
+the scope of its coverage, prohibits the exercise of, or is
+conditioned on the non-exercise of one or more of the rights that are
+specifically granted under this License.  You may not convey a covered
+work if you are a party to an arrangement with a third party that is
+in the business of distributing software, under which you make payment
+to the third party based on the extent of your activity of conveying
+the work, and under which the third party grants, to any of the
+parties who would receive the covered work from you, a discriminatory
+patent license (a) in connection with copies of the covered work
+conveyed by you (or copies made from those copies), or (b) primarily
+for and in connection with specific products or compilations that
+contain the covered work, unless you entered into that arrangement,
+or that patent license was granted, prior to 28 March 2007.
+
+  Nothing in this License shall be construed as excluding or limiting
+any implied license or other defenses to infringement that may
+otherwise be available to you under applicable patent law.
+
+  12. No Surrender of Others' Freedom.
+
+  If conditions are imposed on you (whether by court order, agreement or
+otherwise) that contradict the conditions of this License, they do not
+excuse you from the conditions of this License.  If you cannot convey a
+covered work so as to satisfy simultaneously your obligations under this
+License and any other pertinent obligations, then as a consequence you may
+not convey it at all.  For example, if you agree to terms that obligate you
+to collect a royalty for further conveying from those to whom you convey
+the Program, the only way you could satisfy both those terms and this
+License would be to refrain entirely from conveying the Program.
+
+  13. Use with the GNU Affero General Public License.
+
+  Notwithstanding any other provision of this License, you have
+permission to link or combine any covered work with a work licensed
+under version 3 of the GNU Affero General Public License into a single
+combined work, and to convey the resulting work.  The terms of this
+License will continue to apply to the part which is the covered work,
+but the special requirements of the GNU Affero General Public License,
+section 13, concerning interaction through a network will apply to the
+combination as such.
+
+  14. Revised Versions of this License.
+
+  The Free Software Foundation may publish revised and/or new versions of
+the GNU General Public License from time to time.  Such new versions will
+be similar in spirit to the present version, but may differ in detail to
+address new problems or concerns.
+
+  Each version is given a distinguishing version number.  If the
+Program specifies that a certain numbered version of the GNU General
+Public License "or any later version" applies to it, you have the
+option of following the terms and conditions either of that numbered
+version or of any later version published by the Free Software
+Foundation.  If the Program does not specify a version number of the
+GNU General Public License, you may choose any version ever published
+by the Free Software Foundation.
+
+  If the Program specifies that a proxy can decide which future
+versions of the GNU General Public License can be used, that proxy's
+public statement of acceptance of a version permanently authorizes you
+to choose that version for the Program.
+
+  Later license versions may give you additional or different
+permissions.  However, no additional obligations are imposed on any
+author or copyright holder as a result of your choosing to follow a
+later version.
+
+  15. Disclaimer of Warranty.
+
+  THERE IS NO WARRANTY FOR THE PROGRAM, TO THE EXTENT PERMITTED BY
+APPLICABLE LAW.  EXCEPT WHEN OTHERWISE STATED IN WRITING THE COPYRIGHT
+HOLDERS AND/OR OTHER PARTIES PROVIDE THE PROGRAM "AS IS" WITHOUT WARRANTY
+OF ANY KIND, EITHER EXPRESSED OR IMPLIED, INCLUDING, BUT NOT LIMITED TO,
+THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+PURPOSE.  THE ENTIRE RISK AS TO THE QUALITY AND PERFORMANCE OF THE PROGRAM
+IS WITH YOU.  SHOULD THE PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF
+ALL NECESSARY SERVICING, REPAIR OR CORRECTION.
+
+  16. Limitation of Liability.
+
+  IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING
+WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MODIFIES AND/OR CONVEYS
+THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES, INCLUDING ANY
+GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING OUT OF THE
+USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED TO LOSS OF
+DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY YOU OR THIRD
+PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER PROGRAMS),
+EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE POSSIBILITY OF
+SUCH DAMAGES.
+
+  17. Interpretation of Sections 15 and 16.
+
+  If the disclaimer of warranty and limitation of liability provided
+above cannot be given local legal effect according to their terms,
+reviewing courts shall apply local law that most closely approximates
+an absolute waiver of all civil liability in connection with the
+Program, unless a warranty or assumption of liability accompanies a
+copy of the Program in return for a fee.
+
+                     END OF TERMS AND CONDITIONS
+
+            How to Apply These Terms to Your New Programs
+
+  If you develop a new program, and you want it to be of the greatest
+possible use to the public, the best way to achieve this is to make it
+free software which everyone can redistribute and change under these terms.
+
+  To do so, attach the following notices to the program.  It is safest
+to attach them to the start of each source file to most effectively
+state the exclusion of warranty; and each file should have at least
+the "copyright" line and a pointer to where the full notice is found.
+
+    &lt;one line to give the program's name and a brief idea of what it does.&gt;
+    Copyright (C) &lt;year&gt;  &lt;name of author&gt;
+
+    This program is free software: you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation, either version 3 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License
+    along with this program.  If not, see &lt;<a href="https://www.gnu.org/licenses/">https://www.gnu.org/licenses/</a>&gt;.
+
+Also add information on how to contact you by electronic and paper mail.
+
+  If the program does terminal interaction, make it output a short
+notice like this when it starts in an interactive mode:
+
+    &lt;program&gt;  Copyright (C) &lt;year&gt;  &lt;name of author&gt;
+    This program comes with ABSOLUTELY NO WARRANTY; for details type `show w'.
+    This is free software, and you are welcome to redistribute it
+    under certain conditions; type `show c' for details.
+
+The hypothetical commands `show w' and `show c' should show the appropriate
+parts of the General Public License.  Of course, your program's commands
+might be different; for a GUI interface, you would use an "about box".
+
+  You should also get your employer (if you work as a programmer) or school,
+if any, to sign a "copyright disclaimer" for the program, if necessary.
+For more information on this, and how to apply and follow the GNU GPL, see
+&lt;<a href="https://www.gnu.org/licenses/">https://www.gnu.org/licenses/</a>&gt;.
+
+  The GNU General Public License does not permit incorporating your program
+into proprietary programs.  If your program is a subroutine library, you
+may consider it more useful to permit linking proprietary applications with
+the library.  If this is what you want to do, use the GNU Lesser General
+Public License instead of this License.  But first, please read
+&lt;<a href="https://www.gnu.org/philosophy/why-not-lgpl.html">https://www.gnu.org/philosophy/why-not-lgpl.html</a>&gt;.

rsync-web/COPYING2.html

@@ -0,0 +1,340 @@
+		    GNU GENERAL PUBLIC LICENSE
+		       Version 2, June 1991
+
+ Copyright (C) 1989, 1991 Free Software Foundation, Inc.
+     51 Franklin Street - Fifth Floor, Boston, MA 02110-1301, USA.
+ Everyone is permitted to copy and distribute verbatim copies
+ of this license document, but changing it is not allowed.
+
+			    Preamble
+
+  The licenses for most software are designed to take away your
+freedom to share and change it.  By contrast, the GNU General Public
+License is intended to guarantee your freedom to share and change free
+software--to make sure the software is free for all its users.  This
+General Public License applies to most of the Free Software
+Foundation's software and to any other program whose authors commit to
+using it.  (Some other Free Software Foundation software is covered by
+the GNU Library General Public License instead.)  You can apply it to
+your programs, too.
+
+  When we speak of free software, we are referring to freedom, not
+price.  Our General Public Licenses are designed to make sure that you
+have the freedom to distribute copies of free software (and charge for
+this service if you wish), that you receive source code or can get it
+if you want it, that you can change the software or use pieces of it
+in new free programs; and that you know you can do these things.
+
+  To protect your rights, we need to make restrictions that forbid
+anyone to deny you these rights or to ask you to surrender the rights.
+These restrictions translate to certain responsibilities for you if you
+distribute copies of the software, or if you modify it.
+
+  For example, if you distribute copies of such a program, whether
+gratis or for a fee, you must give the recipients all the rights that
+you have.  You must make sure that they, too, receive or can get the
+source code.  And you must show them these terms so they know their
+rights.
+
+  We protect your rights with two steps: (1) copyright the software, and
+(2) offer you this license which gives you legal permission to copy,
+distribute and/or modify the software.
+
+  Also, for each author's protection and ours, we want to make certain
+that everyone understands that there is no warranty for this free
+software.  If the software is modified by someone else and passed on, we
+want its recipients to know that what they have is not the original, so
+that any problems introduced by others will not reflect on the original
+authors' reputations.
+
+  Finally, any free program is threatened constantly by software
+patents.  We wish to avoid the danger that redistributors of a free
+program will individually obtain patent licenses, in effect making the
+program proprietary.  To prevent this, we have made it clear that any
+patent must be licensed for everyone's free use or not licensed at all.
+
+  The precise terms and conditions for copying, distribution and
+modification follow.
+<hr>
+		    GNU GENERAL PUBLIC LICENSE
+   TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION
+
+  0. This License applies to any program or other work which contains
+a notice placed by the copyright holder saying it may be distributed
+under the terms of this General Public License.  The "Program", below,
+refers to any such program or work, and a "work based on the Program"
+means either the Program or any derivative work under copyright law:
+that is to say, a work containing the Program or a portion of it,
+either verbatim or with modifications and/or translated into another
+language.  (Hereinafter, translation is included without limitation in
+the term "modification".)  Each licensee is addressed as "you".
+
+Activities other than copying, distribution and modification are not
+covered by this License; they are outside its scope.  The act of
+running the Program is not restricted, and the output from the Program
+is covered only if its contents constitute a work based on the
+Program (independent of having been made by running the Program).
+Whether that is true depends on what the Program does.
+
+  1. You may copy and distribute verbatim copies of the Program's
+source code as you receive it, in any medium, provided that you
+conspicuously and appropriately publish on each copy an appropriate
+copyright notice and disclaimer of warranty; keep intact all the
+notices that refer to this License and to the absence of any warranty;
+and give any other recipients of the Program a copy of this License
+along with the Program.
+
+You may charge a fee for the physical act of transferring a copy, and
+you may at your option offer warranty protection in exchange for a fee.
+
+  2. You may modify your copy or copies of the Program or any portion
+of it, thus forming a work based on the Program, and copy and
+distribute such modifications or work under the terms of Section 1
+above, provided that you also meet all of these conditions:
+
+    a) You must cause the modified files to carry prominent notices
+    stating that you changed the files and the date of any change.
+
+    b) You must cause any work that you distribute or publish, that in
+    whole or in part contains or is derived from the Program or any
+    part thereof, to be licensed as a whole at no charge to all third
+    parties under the terms of this License.
+
+    c) If the modified program normally reads commands interactively
+    when run, you must cause it, when started running for such
+    interactive use in the most ordinary way, to print or display an
+    announcement including an appropriate copyright notice and a
+    notice that there is no warranty (or else, saying that you provide
+    a warranty) and that users may redistribute the program under
+    these conditions, and telling the user how to view a copy of this
+    License.  (Exception: if the Program itself is interactive but
+    does not normally print such an announcement, your work based on
+    the Program is not required to print an announcement.)
+<hr>
+These requirements apply to the modified work as a whole.  If
+identifiable sections of that work are not derived from the Program,
+and can be reasonably considered independent and separate works in
+themselves, then this License, and its terms, do not apply to those
+sections when you distribute them as separate works.  But when you
+distribute the same sections as part of a whole which is a work based
+on the Program, the distribution of the whole must be on the terms of
+this License, whose permissions for other licensees extend to the
+entire whole, and thus to each and every part regardless of who wrote it.
+
+Thus, it is not the intent of this section to claim rights or contest
+your rights to work written entirely by you; rather, the intent is to
+exercise the right to control the distribution of derivative or
+collective works based on the Program.
+
+In addition, mere aggregation of another work not based on the Program
+with the Program (or with a work based on the Program) on a volume of
+a storage or distribution medium does not bring the other work under
+the scope of this License.
+
+  3. You may copy and distribute the Program (or a work based on it,
+under Section 2) in object code or executable form under the terms of
+Sections 1 and 2 above provided that you also do one of the following:
+
+    a) Accompany it with the complete corresponding machine-readable
+    source code, which must be distributed under the terms of Sections
+    1 and 2 above on a medium customarily used for software interchange; or,
+
+    b) Accompany it with a written offer, valid for at least three
+    years, to give any third party, for a charge no more than your
+    cost of physically performing source distribution, a complete
+    machine-readable copy of the corresponding source code, to be
+    distributed under the terms of Sections 1 and 2 above on a medium
+    customarily used for software interchange; or,
+
+    c) Accompany it with the information you received as to the offer
+    to distribute corresponding source code.  (This alternative is
+    allowed only for noncommercial distribution and only if you
+    received the program in object code or executable form with such
+    an offer, in accord with Subsection b above.)
+
+The source code for a work means the preferred form of the work for
+making modifications to it.  For an executable work, complete source
+code means all the source code for all modules it contains, plus any
+associated interface definition files, plus the scripts used to
+control compilation and installation of the executable.  However, as a
+special exception, the source code distributed need not include
+anything that is normally distributed (in either source or binary
+form) with the major components (compiler, kernel, and so on) of the
+operating system on which the executable runs, unless that component
+itself accompanies the executable.
+
+If distribution of executable or object code is made by offering
+access to copy from a designated place, then offering equivalent
+access to copy the source code from the same place counts as
+distribution of the source code, even though third parties are not
+compelled to copy the source along with the object code.
+<hr>
+  4. You may not copy, modify, sublicense, or distribute the Program
+except as expressly provided under this License.  Any attempt
+otherwise to copy, modify, sublicense or distribute the Program is
+void, and will automatically terminate your rights under this License.
+However, parties who have received copies, or rights, from you under
+this License will not have their licenses terminated so long as such
+parties remain in full compliance.
+
+  5. You are not required to accept this License, since you have not
+signed it.  However, nothing else grants you permission to modify or
+distribute the Program or its derivative works.  These actions are
+prohibited by law if you do not accept this License.  Therefore, by
+modifying or distributing the Program (or any work based on the
+Program), you indicate your acceptance of this License to do so, and
+all its terms and conditions for copying, distributing or modifying
+the Program or works based on it.
+
+  6. Each time you redistribute the Program (or any work based on the
+Program), the recipient automatically receives a license from the
+original licensor to copy, distribute or modify the Program subject to
+these terms and conditions.  You may not impose any further
+restrictions on the recipients' exercise of the rights granted herein.
+You are not responsible for enforcing compliance by third parties to
+this License.
+
+  7. If, as a consequence of a court judgment or allegation of patent
+infringement or for any other reason (not limited to patent issues),
+conditions are imposed on you (whether by court order, agreement or
+otherwise) that contradict the conditions of this License, they do not
+excuse you from the conditions of this License.  If you cannot
+distribute so as to satisfy simultaneously your obligations under this
+License and any other pertinent obligations, then as a consequence you
+may not distribute the Program at all.  For example, if a patent
+license would not permit royalty-free redistribution of the Program by
+all those who receive copies directly or indirectly through you, then
+the only way you could satisfy both it and this License would be to
+refrain entirely from distribution of the Program.
+
+If any portion of this section is held invalid or unenforceable under
+any particular circumstance, the balance of the section is intended to
+apply and the section as a whole is intended to apply in other
+circumstances.
+
+It is not the purpose of this section to induce you to infringe any
+patents or other property right claims or to contest validity of any
+such claims; this section has the sole purpose of protecting the
+integrity of the free software distribution system, which is
+implemented by public license practices.  Many people have made
+generous contributions to the wide range of software distributed
+through that system in reliance on consistent application of that
+system; it is up to the author/donor to decide if he or she is willing
+to distribute software through any other system and a licensee cannot
+impose that choice.
+
+This section is intended to make thoroughly clear what is believed to
+be a consequence of the rest of this License.
+<hr>
+  8. If the distribution and/or use of the Program is restricted in
+certain countries either by patents or by copyrighted interfaces, the
+original copyright holder who places the Program under this License
+may add an explicit geographical distribution limitation excluding
+those countries, so that distribution is permitted only in or among
+countries not thus excluded.  In such case, this License incorporates
+the limitation as if written in the body of this License.
+
+  9. The Free Software Foundation may publish revised and/or new versions
+of the General Public License from time to time.  Such new versions will
+be similar in spirit to the present version, but may differ in detail to
+address new problems or concerns.
+
+Each version is given a distinguishing version number.  If the Program
+specifies a version number of this License which applies to it and "any
+later version", you have the option of following the terms and conditions
+either of that version or of any later version published by the Free
+Software Foundation.  If the Program does not specify a version number of
+this License, you may choose any version ever published by the Free Software
+Foundation.
+
+  10. If you wish to incorporate parts of the Program into other free
+programs whose distribution conditions are different, write to the author
+to ask for permission.  For software which is copyrighted by the Free
+Software Foundation, write to the Free Software Foundation; we sometimes
+make exceptions for this.  Our decision will be guided by the two goals
+of preserving the free status of all derivatives of our free software and
+of promoting the sharing and reuse of software generally.
+
+			    NO WARRANTY
+
+  11. BECAUSE THE PROGRAM IS LICENSED FREE OF CHARGE, THERE IS NO WARRANTY
+FOR THE PROGRAM, TO THE EXTENT PERMITTED BY APPLICABLE LAW.  EXCEPT WHEN
+OTHERWISE STATED IN WRITING THE COPYRIGHT HOLDERS AND/OR OTHER PARTIES
+PROVIDE THE PROGRAM "AS IS" WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESSED
+OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
+MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE.  THE ENTIRE RISK AS
+TO THE QUALITY AND PERFORMANCE OF THE PROGRAM IS WITH YOU.  SHOULD THE
+PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF ALL NECESSARY SERVICING,
+REPAIR OR CORRECTION.
+
+  12. IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING
+WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MAY MODIFY AND/OR
+REDISTRIBUTE THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES,
+INCLUDING ANY GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING
+OUT OF THE USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED
+TO LOSS OF DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY
+YOU OR THIRD PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER
+PROGRAMS), EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE
+POSSIBILITY OF SUCH DAMAGES.
+
+		     END OF TERMS AND CONDITIONS
+<hr>
+	    How to Apply These Terms to Your New Programs
+
+  If you develop a new program, and you want it to be of the greatest
+possible use to the public, the best way to achieve this is to make it
+free software which everyone can redistribute and change under these terms.
+
+  To do so, attach the following notices to the program.  It is safest
+to attach them to the start of each source file to most effectively
+convey the exclusion of warranty; and each file should have at least
+the "copyright" line and a pointer to where the full notice is found.
+
+    &lt;one line to give the program's name and a brief idea of what it does.&gt;
+    Copyright (C) &lt;year&gt;  &lt;name of author&gt;
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License
+    along with this program; if not, write to the Free Software
+    Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+
+
+Also add information on how to contact you by electronic and paper mail.
+
+If the program is interactive, make it output a short notice like this
+when it starts in an interactive mode:
+
+    Gnomovision version 69, Copyright (C) year  name of author
+    Gnomovision comes with ABSOLUTELY NO WARRANTY; for details type `show w'.
+    This is free software, and you are welcome to redistribute it
+    under certain conditions; type `show c' for details.
+
+The hypothetical commands `show w' and `show c' should show the appropriate
+parts of the General Public License.  Of course, the commands you use may
+be called something other than `show w' and `show c'; they could even be
+mouse-clicks or menu items--whatever suits your program.
+
+You should also get your employer (if you work as a programmer) or your
+school, if any, to sign a "copyright disclaimer" for the program, if
+necessary.  Here is a sample; alter the names:
+
+  Yoyodyne, Inc., hereby disclaims all copyright interest in the program
+  `Gnomovision' (which makes passes at compilers) written by James Hacker.
+
+  &lt;signature of Ty Coon&gt;, 1 April 1989
+  Ty Coon, President of Vice
+
+This General Public License does not permit incorporating your program into
+proprietary programs.  If your program is a subroutine library, you may
+consider it more useful to permit linking proprietary applications with the
+library.  If this is what you want to do, use the GNU Library General
+Public License instead of this License.

rsync-web/FAQ.html

@@ -0,0 +1,284 @@
+<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
+<HTML>
+<HEAD>
+<TITLE>rsync FAQ</TITLE>
+</HEAD>
+<!--#include virtual="header.html" -->
+
+<H2 align="center">Frequently Asked Questions</H2>
+
+<table><tr valign=top><td><ul>
+<li><a href="#1">the transfer fails to finish</a><br>
+<li><a href="#2">rsync recopies the same files</a><br>
+<li><a href="#3">is your shell clean</a><br>
+<li><a href="#4">memory usage</a><br>
+<li><a href="#5">out of memory</a><br>
+<li><a href="#6">rsync through a firewall</a><br>
+<li><a href="#7">rsync and cron</a><br>
+</ul></td><td>&nbsp;&nbsp;&nbsp;&nbsp;</td><td><ul>
+<li><a href="#8">rsync: Command not found</a><br>
+<li><a href="#9">spaces in filenames</a><br>
+<li><a href="#10">ignore "vanished files" warning</a><br>
+<li><a href="#11">read-only file system</a><br>
+<li><a href="#12">multiplexing overflow 101:7104843</a><br>
+<li><a href="#13">inflate (token) returned -5</a><br>
+</ul></td></tr></table>
+
+<hr>
+<h3><a name=1>the transfer fails to finish</a></h3>
+
+<p>If you get an error like one of these:
+
+<pre>rsync: error writing 4 unbuffered bytes - exiting: Broken pipe
+rsync error: error in rsync protocol data stream (code 12) at io.c(463)
+</pre>
+
+<p>or
+
+<pre>rsync: connection unexpectedly closed (24 bytes read so far)
+rsync error: error in rsync protocol data stream (code 12) at io.c(342)
+</pre>
+
+<p>please read the <a href="issues.html">issues and debugging page</a>
+for details on how you can try to figure out what is going wrong.
+
+<hr>
+<h3><a name=2>rsync recopies the same files</a></h3>
+
+<p>Some people occasionally report that rsync copies too many files when
+they expect it to copy only a few. In most cases the explanation is
+that you forgot to include the --times (-t) option in the original copy,
+so rsync is forced to (efficiently) transfer every file that differs in
+its modified time to discover what data (if any) has changed.
+
+<p>Another common cause involves sending files to an Microsoft filesystem:
+if the file's modified time is an odd value but the receiving filesystem
+can only store even values, then rsync will re-transfer too many files.
+You can avoid this by specifying the --modify-window=1 option.
+
+<p>Yet another periodic case can happen when daylight-savings time
+changes if your OS+filesystem saves file times in local time instead of
+UTC.  For a full explanation of this and some suggestions on how to
+avoid them problem, see <a href="daylight-savings.html">this document</a>.
+
+<p>Something else that can trip up rsync is a filesystem changeing the
+filename behind the scenes.  This can happen when a filesystem changes
+an all-uppercase name into lowercase, or when it decomposes UTF-8 behind
+your back.
+
+<blockquote>
+
+<p>An example of the latter can occur with HFS+ on Mac OS X:  if you
+copy a directory with a file that has a UTF-8 character sequence in it,
+say a 2-byte umlaut-u (\0303\0274), the file will get that character
+stored by the filesystem using 3 bytes (\0165\0314\0210), and rsync will
+not know that these differing filenames are the same file (it will, in
+fact, remove a prior copy of the file if --delete is enabled, and then
+recreate it).
+
+<p>You can avoid a charset problem by passing an appropriate --iconv
+option to rsync that tells it what character-set the source files are,
+and what character-set the destination files get stored in.  For
+instance, the above Mac OS X problem would be dealt with by using
+--iconv=UTF-8,UTF8-MAC (UTF8-MAC is a pseudo-charset recognized by Mac
+OS X iconv in which all characters are decomposed).
+
+</blockquote>
+
+<p>If you think that rsync is copying too many files, look at the
+itemized output (-i) to see why rsync is doing the update (e.g. the 't'
+flag indicates that the time differs, or all pluses indicates that rsync
+thinks the file doesn't exist).  You can also look at the stats produced
+with -v and see if rsync is really sending all the data.  See also the
+--checksum (-c) option for one way to avoid the extra copying of files
+that don't have synchronized modified times (but keep in mind that the
+-c option eats lots of disk I/O, and can be rather slow).
+
+<hr>
+<h3><a name=3>is your shell clean</a></h3>
+
+<p>The "is your shell clean" message and the "protocol mismatch" message
+are usually caused by having some sort of program in your .cshrc, .profile,
+.bashrc or equivalent file that writes a message every time you connect
+using a remote-shell program (such as ssh or rsh).  Data written in this
+way corrupts the rsync data stream. rsync detects this at startup and
+produces those error messages.  However, if you are using rsync-daemon
+syntax (host::path or rsync://) without using a remote-shell program (no
+--rsh or -e option), there is not remote-shell program involved, and the
+problem is probably caused by an error on the daemon side (so check the
+daemon logs).
+
+<p>A good way to test if your remote-shell connection is clean is to try
+something like this (use ssh or rsh, as appropriate):
+
+<blockquote><pre>ssh remotesystem /bin/true &gt; test.dat</pre></blockquote>
+
+<p>That should create a file called test.dat with nothing in it. If
+test.dat is not of zero length then your shell is not clean.  Look at the
+contents of test.dat to see what was sent. Look at all the startup files on
+remotesystem to try and find the problem.
+
+<hr>
+<h3><a name=4>memory usage</a></h3>
+
+<p>Rsync versions before 3.0.0 always build the entire list of files to be
+transferred at the beginning and hold it in memory for the entire run.  Rsync
+needs about 100 bytes to store all the relevant information for one file,
+so (for example) a run with 800,000 files would consume about 80M of
+memory.  -H and --delete increase the memory usage further.
+
+<p>Version 3.0.0 slightly reduced the memory used per file by not storing fields
+not needed for a particular file.  It also introduced an incremental recursion
+mode that builds the file list in chunks and holds each chunk in memory only as
+long as it is needed.  This mode dramatically reduces memory usage, but it
+only works provided that both sides are 3.0.0 or newer and certain options that
+rsync currently can't handle in this mode are not being used.
+
+<hr>
+<h3><a name=5>out of memory</a></h3>
+
+<p>The usual reason for "out of memory" when running rsync is that you are
+transferring a _very_ large number of files.  The size of the files doesn't
+matter, only the total number of files.  If memory is a problem, first try to
+use the incremental recursion mode: upgrade both sides to rsync 3.0.0 or
+newer and avoid options that disable incremental recursion (e.g., use
+<tt>--delete-delay</tt> instead of <tt>--delete-after</tt>).  If this is not
+possible, you can break the rsync run into smaller chunks operating on
+individual subdirectories using <tt>--relative</tt> and/or exclude rules.
+
+<hr>
+<h3><a name=6>rsync through a firewall</a></h3>
+
+<p>If you have a setup where there is no way to directly connect two
+systems for an rsync transfer, there are several ways to get a firewall
+system to act as an intermediary in the transfer.  You'll find full details
+on the <a href="firewall.html">firewall methods</a> page.
+
+<hr>
+<h3><a name=7>rsync and cron</a></h3>
+
+<p>On some systems (notably SunOS4) cron supplies what looks like a socket
+to rsync, so rsync thinks that stdin is a socket. This means that if you
+start rsync with the --daemon switch from a cron job you end up rsync
+thinking it has been started from inetd. The fix is simple&mdash;just
+redirect stdin from /dev/null in your cron job.
+
+<hr>
+<h3><a name=8>rsync: Command not found</a></h3>
+
+<p>This error is produced when the remote shell is unable to locate the rsync
+binary in your path. There are 3 possible solutions:
+
+<ol>
+
+<li>install rsync in a "standard" location that is in your remote path. 
+
+<li>modify your .cshrc, .bashrc etc on the remote system to include the path
+that rsync is in
+
+<li>use the --rsync-path option to explicitly specify the path on the
+remote system where rsync is installed
+
+</ol>
+
+<p>You may echo find the command:
+
+<blockquote><pre>ssh host 'echo $PATH'</pre></blockquote>
+
+<p>for determining what your remote path is.
+
+<hr>
+<h3><a name=9>spaces in filenames</a></h3>
+
+<p>Can rsync copy files with spaces in them?
+
+<p>Short answer: Yes, rsync can handle filenames with spaces.
+
+<p>Long answer: 
+
+<p>Rsync handles spaces just like any other unix command line application.
+Within the code spaces are treated just like any other character so a
+filename with a space is no different from a filename with any other
+character in it.
+
+<p>The problem of spaces is in the argv processing done to interpret the
+command line.  As with any other unix application you have to escape spaces
+in some way on the command line or they will be used to separate arguments. 
+
+<p>It is slightly trickier in rsync (and other remote-copy programs like
+scp) because rsync sends a command line to the remote system to launch the
+peer copy of rsync (this assumes that we're not talking about daemon mode,
+which is not affected by this problem because no remote shell is involved
+in the reception of the filenames).  The command line is interpreted by the
+remote shell and thus the spaces need to arrive on the remote system
+escaped so that the shell doesn't split such filenames into multiple
+arguments.
+
+<p>For example:
+
+<blockquote><pre>rsync -av host:'a long filename' /tmp/</pre></blockquote>
+
+<p>This is usually a request for rsync to copy 3 files from the remote
+system, "a", "long", and "filename" (the only exception to this is for a
+system running a shell that does not word-split arguments in its commands,
+and that is exceedingly rare).  If you wanted to request a single file with
+spaces, you need to get some kind of space-quoting characters to the remote
+shell that is running the remote rsync command.  The following commands
+should all work:
+
+<blockquote><pre>rsync -av host:'"a long filename"' /tmp/
+rsync -av host:'a\ long\ filename' /tmp/
+rsync -av host:a\\\ long\\\ filename /tmp/</pre></blockquote>
+
+<p>You might also like to use a '?' in place of a space as long as there
+are no other matching filenames than the one with spaces (since '?' matches
+any character):
+
+<blockquote><pre>rsync -av host:a?long?filename /tmp/</pre></blockquote>
+
+<p>As long as you know that the remote filenames on the command line
+are interpreted by the remote shell then it all works fine. 
+
+<hr>
+<h3><a name=10>ignore "vanished files" warning</a></h3>
+
+<p>Some folks would like to ignore the "vanished files" warning, which
+manifests as an exit-code 24.  The easiest way to do this is to create
+a shell script wrapper.  For instance, name this something like
+"rsync-no24":
+
+<blockquote><pre>#!/bin/sh
+rsync "$@"
+e=$?
+if test $e = 24; then
+    exit 0
+fi
+exit $e</pre></blockquote>
+
+<hr>
+<h3><a name=11>read-only file system</a></h3>
+
+<p>If you get "Read-only file system" as an error when sending to a rsync
+daemon then you probably forgot to set "read only = no" for that module.
+
+<hr>
+<h3><a name=12>multiplexing overflow 101:7104843</a></h3>
+
+<p>This mysterious error, or the similar "invalid message 101:7104843", can
+happen if one of the rsync processes is killed for some reason and a message
+beginning with the four characters "Kill" gets inserted into the protocol
+stream as a result.  To solve the problem, you'll need to figure out why rsync
+is being killed.
+
+<hr>
+<h3><a name=13>inflate (token) returned -5</a></h3>
+
+This error means that rsync failed to handle an expected error from the
+compression code for a file that happened to be transferred with a block size
+of 32816 bytes.  You can avoid this issue for the affected file by transferring
+it with a manually-set block size (e.g. --block-size=33000), or by upgrading
+the receiving side to rsync 3.0.7.
+
+<hr>
+
+<!--#include virtual="footer.html" -->

rsync-web/GPL.html

@@ -0,0 +1,16 @@
+<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
+<HTML>
+<HEAD>
+<TITLE>rsync's license</TITLE>
+</HEAD>
+<!--#include virtual="header.html" -->
+
+Beginning with 3.0.0, rsync is available under the <b>GNU General Public
+License version 3</b>. <i>(Older releases were available under the
+<a href="GPL2.html">GPL version 2</a>.)</i>
+
+<pre><small>
+<!--#include virtual="COPYING.html" -->
+</small></pre>
+
+<!--#include virtual="footer.html" -->

rsync-web/GPL2.html

@@ -0,0 +1,15 @@
+<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
+<HTML>
+<HEAD>
+<TITLE>rsync's license</TITLE>
+</HEAD>
+<!--#include virtual="header.html" -->
+
+Releases <b>prior</b> to 3.0.0 were released under the
+<b>GNU General Public License version 2</b>:
+
+<pre><small>
+<!--#include virtual="COPYING2.html" -->
+</small></pre>
+
+<!--#include virtual="footer.html" -->

rsync-web/backup.txt

@@ -0,0 +1,36 @@
+#!/bin/sh
+
+# This script does personal backups to a rsync backup server. You will end up
+# with a 7 day rotating incremental backup. The incrementals will go
+# into subdirectories named after the day of the week, and the current
+# full backup goes into a directory called "current"
+# tridge@linuxcare.com
+
+# directory to backup
+BDIR=/home/$USER
+
+# excludes file - this contains a wildcard pattern per line of files to exclude
+EXCLUDES=$HOME/cron/excludes
+
+# the name of the backup machine
+BSERVER=owl
+
+# your password on the backup server
+export RSYNC_PASSWORD=XXXXXX
+
+
+########################################################################
+
+BACKUPDIR=`date +%A`
+OPTS="--force --ignore-errors --delete-excluded --exclude-from=$EXCLUDES 
+      --delete --backup --backup-dir=/$BACKUPDIR -a"
+
+export PATH=$PATH:/bin:/usr/bin:/usr/local/bin
+
+# the following line clears the last weeks incremental directory
+[ -d $HOME/emptydir ] || mkdir $HOME/emptydir
+rsync --delete -a $HOME/emptydir/ $BSERVER::$USER/$BACKUPDIR/
+rmdir $HOME/emptydir
+
+# now the actual transfer
+rsync $OPTS $BDIR $BSERVER::$USER/current

rsync-web/bin/badge-update

@@ -0,0 +1,4 @@
+#!/bin/sh
+rm -f badge.svg
+wget https://github.com/RsyncProject/rsync/workflows/build/badge.svg
+rsync -aiic --inplace --remove-source-files badge.svg $SAMBA_HOST:/home/httpd/html/rsync/

rsync-web/bin/upload

@@ -0,0 +1,7 @@
+#!/bin/sh
+if [ -d rsync-and-debian ]; then
+    rsync -aviOHFFc --del -f._filt . $SAMBA_HOST:/home/httpd/html/rsync/ "${@}"
+else
+    echo "Run this from the root of the html hierarchy."
+    exit 1
+fi

rsync-web/bug-tracking.html

@@ -0,0 +1,69 @@
+<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
+<HTML>
+<HEAD>
+<TITLE>rsync bug-tracking</TITLE>
+</HEAD>
+<!--#include virtual="header.html" -->
+
+<H2 align="center">rsync bug-tracking</H2>
+
+<p> Please use this checklist combined with the help on the
+<a href="issues.html">issues and debugging</a> page before
+reporting a bug.  Thanks!
+
+<ul>
+
+<li> If you're not using the latest released version, please upgrade before
+reporting a bug.
+
+<li> If you're using the latest released version, consult the
+<a href="https://github.com/RsyncProject/rsync/blob/master/NEWS.md">NEWS file from the git repository</a> to see if what
+you're seeing has already been handled in the version under development.
+
+<li> It is also helpful to search the bug reports at the
+<a href="https://github.com/RsyncProject/rsync/issues?q=is%3Aissue">GitHub issues tracker</a>
+to see if the problem is already known.
+
+<li> See also the <a href="issues.html">issues and debugging</a> page to
+help you figure out if what you're seeing is a known bug and perhaps to
+help diagnose what is going wrong.
+
+<li> Discuss the bug on the
+<a href="https://lists.samba.org/mailman/listinfo/rsync">rsync mailing list</a>
+(which is at <tt>rsync@lists.samba.org</tt>) to help you figure out if what
+you're seeing is really a bug or a mistake.
+
+<li>There are several patches for features that are under consideration that
+can be found in the <a href="https://github.com/RsyncProject/rsync-patches">rsync-patches repo</a>.
+
+<li> If you haven't already done so, please take a couple of minutes to read Simon Tatham's
+<a href="https://www.chiark.greenend.org.uk/~sgtatham/bugs.html">advice on how to report bugs</a>.
+
+</ul>
+
+<p> To report a bug or make suggestions, use one of these methods:
+
+<ul>
+
+<li> The mailing list (mentioned above) is a good resource for discussing
+bugs and suggesting new features.  It accepts patches (typically as MIME
+attachments), but for fixes is often easier to attach a patch to an
+appropriate GitHub issue report or use a pull request.  Note that there is
+no mandate to use pull requests for patches, as that can be a pretty high
+bar of git know-how that not everyone needs to be familiar with.
+
+<li> If you'd like to see a bug-report or feature-request get officially noted,
+<a href="https://github.com/RsyncProject/rsync/issues">create an issue on GitHub</a>
+(this does require that you have created a GitHub account).  If you want to
+stay abreast of what's going on with the issues, make use of the GitHub
+subscriptions to pick and choose what kind of notifications you want to
+receive (e.g. just a single issue, all issues, all rsync activity, etc.).
+
+<li>For security issues please send email
+  to <a href="mailto:rsync.project@gmail.com">rsync.project@gmail.com</a>
+  with details of the issue
+</ul>
+
+<p> Thanks for helping out!
+
+<!--#include virtual="footer.html" -->

rsync-web/bugtracking.html

@@ -0,0 +1,7 @@
+<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
+<HTML>
+<HEAD>
+<TITLE>rsync bug-tracking</TITLE>
+</HEAD>
+
+<meta http-equiv="refresh" content="0;URL='https://rsync.samba.org/bug-tracking.html'" />

rsync-web/bugzilla.html

@@ -0,0 +1,7 @@
+<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
+<HTML>
+<HEAD>
+<TITLE>rsync bug-tracking</TITLE>
+</HEAD>
+
+<meta http-equiv="refresh" content="0;URL='https://rsync.samba.org/bug-tracking.html'" />

rsync-web/convert-gpl

@@ -0,0 +1,11 @@
+#!/usr/bin/perl
+use strict;
+
+while (<>) {
+    s/&/&amp;/g;
+    s/</&lt;/g;
+    s/>/&gt;/g;
+    s//<hr>/;
+    s#(&lt;)(https?:.*?)(&gt;)#$1<a href="$2">$2</a>$3#g;
+    print $_;
+}

rsync-web/daylight-savings.html

@@ -0,0 +1,183 @@
+<HTML><HEAD>
+<TITLE>DST change and date comparisons</TITLE>
+</HEAD><BODY>
+
+<p><small><i>J.W. Schultz wrote the following text in a
+<a href="http://www.cygwin.com/ml/cygwin/2003-10/msg00995.html">post</a>
+to the cygwin mailing list.&nbsp; It has been slightly edited and beautified
+for inclusion here.</i></small>
+
+<h1>How the DST Change can adversely affect FAT filesystems</h1>
+
+<p>
+The vernal and autumnal transitions to and from daylight-savings
+time have important implications
+for those with Microsoft systems and use utilities that
+compare file timestamps on different filesystem types or
+with filesystems on other operating systems that can lead to
+a problem in how the file's date is handled.
+This problem lies in the way FAT filesystems stores
+timestamps and how Windows converts between local time and
+UTC. 
+
+<h2>Background</h2>
+
+<p>
+In UNIX and UNIX-like systems (such as Linux) file timestamps
+are stored in UTC (universal time) and are only converted to
+local-time by user-space programs for display purposes.&nbsp; At
+the system call level all time values are in UTC and
+utilities that compare timestamps do so in UTC.&nbsp; Also, the
+standard UTC-&gt;local and local-&gt;UTC conversion functions are
+aware of DST and conversions reflect this so that if a
+timestamp was recorded during ST it will be converted using
+the ST offset even when the current system time is DST.
+
+<p>
+In Windows things are not so simple.&nbsp; Windows operates in
+local-time.&nbsp; Timestamps in the various FAT derived
+filesystems are stored in local-time.&nbsp; Timestamps in NTFS
+filesystems are stored in UTC.&nbsp; This inconsistency is
+further complicated by the fact that the conversion routines
+used are not DST aware.&nbsp; Instead of being DST aware the
+system has a fixed offset to convert between local-time and
+UTC regardless of the date in the timestamp.&nbsp; This fixed
+offset is calculated at boot time and only changed when
+systems transition to or from DST.&nbsp; As a result the apparent
+modification time of a file on NTFS as reported in a windows
+utility will change by one hour when reported in local-time
+and FAT based files when reported in UTC.
+
+<p>
+The difficulty that this produces is that any utilities that
+compare timestamps between FAT and NTFS filesystems or
+between Windows and other platforms will view files that
+have not changed as having a different modified time.&nbsp; Among other things
+this will affect rsync, rdiff, unison, wget, and make.&nbsp; However,
+for the purposes of this document, we will only discuss rsync.
+
+<p>
+With the reduced cost of hard disks many newer backup
+systems are using hard disk based storage and take advantage
+of timestamp comparison to detect file changes for the sake
+of efficiency.&nbsp; Rsync is probably premier in this role and
+is used by a fair number of free and even commercial backup
+systems as well as being the basis for many home-brew backup
+solutions.
+
+<p>
+With rsync and similar systems the effect of this is that
+every file will appear to have been changed.&nbsp; The result is
+any space savings associated with linking (--link-dest) or
+with decremental backup approaches (--compare-dest and
+--backup-dir) will be defeated.&nbsp; Perhaps worse, because
+every file will appear to have changed the time required to
+do a backup or a non-backup rsync will be much longer than
+normal.&nbsp; In some cases backups that normally complete in
+less than one hour can take several days.
+
+<p>
+So what can be done about it?  Several things, there are
+ways to merely mitigate the problem, to correct it and finally
+to prevent the problem entirely.
+
+<h2>Mitigation</h2>
+
+<p>
+Rsync has a --modify-window option.&nbsp; Many of you already use
+--modify-window=1 to cope with the fact that windows often
+stores timestamps with a two second resolution.&nbsp; Using a
+--modify-window=3601 will cause rsync to ignore timestamp
+differences of up to one hour.
+
+<p>
+This if often not particularly dangerous because a file would have to
+be changed, synced and changed again without changing size
+within a single hour and have no subsequent changes for this copy
+to miss a file change.&nbsp; However, for some systems any such risk is
+unacceptable, so other solutions are needed.
+
+<h2>Correcting the Timestamps</h2>
+
+<p>
+There are two ways to correct.
+
+<p>For the first run after the time change, you can run rsync with the
+--checksum option in order to ensure that only files that have a changed
+checksum get transferred, and update the modified time on all unchanged
+files.&nbsp; This option has the drawback that it increases disk I/O by
+a large amount on both the sending and receiving side, slowing down the
+copy.
+
+<p>
+The other way to correct things is to change the timestamps
+on the files on the backup server before doing a copy after a
+time change has occurred.
+
+<p>
+Included here is an example perl script that will change the
+timestamps of files in a list on standard-input.&nbsp; Whether
+you use a positive or negative shift will depend on which
+end you decide to adjust.
+
+<p>
+This is an example of how to use the script:
+
+<blockquote><pre>
+touch -d '01:00 13-apr-03' /tmp/cmpfile
+find . -type f ! -newer /tmp/cmpfile | shifttime.pl 3600
+</pre></blockquote>
+
+<hr>
+<pre>
+#!/usr/bin/perl
+use strict;
+
+my $offset = shift() + 0;
+die &quot;Usage: $0 OFFSET_SECONDS\n&quot; unless $offset;
+
+while (&lt;&gt;) {
+    chomp;
+    my $mtime = (stat $_)[9];
+    next unless $mtime;
+    $mtime += $offset;
+    utime $mtime, $mtime, $_;
+}
+</pre>
+<hr>
+
+<h2>Prevention</h2>
+
+<p>
+To prevent the problem in the first place you need to
+prevent changing to DST.&nbsp; This can be done by either running
+the windows system in UTC, by disabling DST and changing
+the system time manually twice each year, or by avoiding the use
+of the FAT filesystem (perhaps by switching to a different OS).
+
+
+<h2>Notes and References</h2>
+
+<p>
+Here are some references that Wayne Piekarski collected
+while researching this problem.&nbsp; They contain a lot of
+information about the ways that Windows deals with
+timestamps on NTFS and FAT filesystems.
+
+<p>
+<a  href="http://optics.ph.unimelb.edu.au/help/rsync/rsync_pc1.html#gotchas">http://optics.ph.unimelb.edu.au/help/rsync/rsync_pc1.html#gotchas</a></br>
+<a  href="http://list-archive.xemacs.org/xemacs-nt/199911/msg00130.html">http://list-archive.xemacs.org/xemacs-nt/199911/msg00130.html</a></br>
+<a  href="http://p2p.wrox.com/archive/c_plus_plus_programming/2001-06/53.asp">http://p2p.wrox.com/archive/c_plus_plus_programming/2001-06/53.asp</a></br>
+<a  href="http://www.codeproject.com/datetime/dstbugs.asp">http://www.codeproject.com/datetime/dstbugs.asp</a></br>
+<a  href="http://support.microsoft.com/default.aspx?scid=kb;[LN];158588">http://support.microsoft.com/default.aspx?scid=kb;[LN];158588</a>
+
+<p>
+I wish to thank Wayne Piekarski for having copiled the
+references and also supplying some additional insights.
+
+<p>
+Permission is granted without reservation reprint and
+distribute this in whole and in part to any interested
+parties.
+
+<p>&mdash;J.W. Schultz

rsync-web/doc-resources.html

@@ -0,0 +1,15 @@
+<!-- This file gets included into the resources and documentation pages -->
+
+<li> A nice tutorial on <a href="https://linuxize.com/post/how-to-setup-passwordless-ssh-login/">setting up ssh to avoid password prompts</a>.
+
+<li> Karsten Thygesen has written a doc on how to setup
+<A HREF="http://dslab.lzu.edu.cn:8080/members/wangbj/wangbaojun/howtos/rsync-mirror-HOWTO/">anonymous rsync servers</A>.
+
+<li> Anthony Wesley has written a doc on
+<A HREF="/win95.txt">how to build rsync for Windows95</A>.
+
+<li> Mike McHenry has written up some info on how to get <a href="nt.html">rsync working under NT</a>.
+
+<li> Michael Holve has written a very useful <a href="http://everythinglinux.org/rsync/">rsync tutorial</a>.
+
+<li> Daniel Teklu has a detailed HOWTO on <a href="http://www.netbits.us/docs/stunnel_rsync.html">rsync and stunnel</a>.

rsync-web/documentation.html

@@ -0,0 +1,54 @@
+<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
+<HTML>
+<HEAD>
+<TITLE>rsync documentation</TITLE>
+</HEAD>
+<!--#include virtual="header.html" -->
+
+<H2 align="center">documentation</H2>
+
+<ul>
+
+<li> An html version of the <a href="https://download.samba.org/pub/rsync/rsync.1"><b>rsync</b>(1) manpage</a>.
+
+<li> An html version of the
+<a href="https://download.samba.org/pub/rsync/rsync-ssl.1"><b>rsync-ssl</b>(1) manpage</a>.
+
+<li> An html version of the
+<a href="https://download.samba.org/pub/rsync/rsyncd.conf.5"><b>rsyncd.conf</b>(5) manpage</a>.
+
+<li> An html version of the
+<a href="https://download.samba.org/pub/rsync/rrsync.1"><b>rrsync</b>(1) manpage</a>
+
+<li> The <a href="FAQ.html">FAQ</a> (frequently asked questions list).
+
+<!--#include virtual="doc-resources.html" -->
+
+<!--
+<li>Cross-referenced
+<a href="doxygen/head/files.html">rsync source code</a> produced by
+<a href="http://doxygen.org/">Doxygen</a>.
+-->
+
+<li> A html version of the original
+<a href="tech_report/">rsync technical report</a>.
+
+<li> A copy of Andrew Tridgell's
+<a href="http://samba.org/~tridge/phd_thesis.pdf">PhD thesis</a>
+(which includes three chapters on rsync).
+
+<li> A nice page on <a href="how-rsync-works.html">how rsync works</a>.
+
+<li> A copy of John Langford's thesis on
+<a href="http://www-2.cs.cmu.edu/~jcl/research/mrsync/mrsync.ps">Multiround rsync</a>,
+which is not used in rsync, but is interesting none-the-less.
+
+<li>A <a href="http://www.devshed.com/c/a/Administration/File-Synchronization-With-Rsync/">DevShed tutorial on rsync</a>.
+
+<li> <a href="rsync-and-debian/">Notes on rsync and Debian mirrors</a>.
+
+</ul>
+
+<p> <i>(Some of the above items are also listed on the <a href="resources.html">rsync resources page</a>.</i>
+
+<!--#include virtual="footer.html" -->

rsync-web/download.html

@@ -0,0 +1,127 @@
+<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
+<html>
+<head>
+<TITLE>rsync download</TITLE>
+</head>
+<!--#include virtual="header.html" -->
+
+<h2 align="center">rsync download</h2>
+
+<div style="float: right">
+<a href="https://github.com/RsyncProject/rsync/actions">
+<img src="badge.svg">
+</a></div>
+
+<h2>Source-code releases</h2>
+
+<p>You can grab the latest source code and other related files in a variety of ways:
+
+<ul>
+
+<li><p>The latest version is linked on the <a href="https://rsync.samba.org/">main page</a>.
+
+<li><p>A directory listing of these latest files and various historical release and diff files
+are available via <a href="https://download.samba.org/pub/rsync/">this web page</a> and
+via <i>anonymous SSL rsync</i> using this command:
+<p><code><small>rsync-ssl rsync://download.samba.org/rsyncftp/</small></code></p>
+
+<li><p>You can also get .zip and .tar.gz versions of the various git repo's release
+tags via the <a href="https://github.com/RsyncProject/rsync/tags">rsync GitHub tags page</a>
+and the associated patches via the
+<a href="https://github.com/RsyncProject/rsync-patches/tags">rsync-patches GitHub tags page</a>.
+Keep in mind that these git-derived files do NOT come with the extra generated files that are included
+in the official release tar files.
+
+<li><p>You can browse the very latest source files, clone the source using git, or download a .zip file of the latest
+master branch from <a href="https://github.com/RsyncProject/rsync">rsync's GitHub page</a>.
+
+<li><p>The <a href="https://git.samba.org/?p=rsync.git">Samba git repo</a> is also available,
+though it might lag behind the GitHub repo every now and then.
+
+</ul>
+
+Once you have the source, read the <a
+href="https://download.samba.org/pub/rsync/INSTALL">INSTALL.md</a> file for
+details on some development libraries that you will need to build it.
+
+<h2>The GPG Signing Key</h2>
+
+The GPG signing key that is used to sign the release files is available from the public pgp key-server
+network.  If you have automatic key-fetching enabled, just running a normal
+"gpg&nbsp;--verify" will grab my key automatically.
+Or, feel free to grab <a href="https://opencoder.net/WayneDavison.key">the gpp
+  key for Wayne Davison</a> manually.<p>
+
+From 3.4.0 and later releases will be signed by Andrew
+Tridgell. Please fetch the key for andrew@tridgell.net from https://keys.openpgp.org/
+
+<h2>Binaries</h2>
+
+<p>Precompiled binaries are available in most modern OS distributions, so
+you should first check if you can install an rsync package via your
+standard package-install tools for your OS.
+
+<h3>Ubuntu</h3>
+
+<p>The rsync project maintains a Launchpad PPA that tracks upstream stable
+releases for the currently supported Ubuntu series (jammy 22.04 LTS,
+noble 24.04 LTS, questing 25.10, resolute 26.04 LTS).  This is the
+fastest way to get the latest upstream rsync on Ubuntu without waiting
+for the distro to update its packaged version:
+
+<p><code><small>sudo add-apt-repository ppa:rsyncproject/rsync<br>
+sudo apt update &amp;&amp; sudo apt install rsync</small></code>
+
+<p>See the <a href="https://launchpad.net/~rsyncproject/+archive/ubuntu/rsync">PPA page on Launchpad</a>
+for build status across architectures.
+
+<p>The <a href="https://github.com/RsyncProject/rsync/actions">GitHub Actions
+page</a> has build events that each generate a few binary artifact zip files
+(just click through via the build's title to see them).  The actions page is
+also linked via the various green build-status icons on the web pages here.
+These builds use the newest libraries, such as xxhash checksums and zstd
+compression, and are dynamically linked, so you may need to install some
+official library packages for your distribution.  If you're curious how the
+build was done, you can look at the build rules in the "Workflow file" tab.
+See the <a href="https://download.samba.org/pub/rsync/INSTALL">INSTALL.md</a>
+file for some package name hints, though you can use the non-devel versions of
+the various lib packages and ignore the gcc/autoconf/awk packages.
+
+<p>There are also packages available from some <b>3rd-parties</b> (note that we
+cannot vouch for 3rd parties, so use a source that you trust):
+
+<ul>
+
+<li><p><a href="https://www.cygwin.com/">Cygwin</a> is a Posix runtime for MS
+Windows that includes rsync among their many packages.</p></li>
+
+<li><p><a href="https://www.itefix.net/cwrsync">cwRsync</a> is a native
+packaging of rsync for MS Windows (they appear to only provide paid releases,
+though).</p></li>
+
+</ul>
+
+<h2>Git source vs Release files</h2>
+
+The release tar files come with a few generated files that are not checked in to git.
+These mainly include the man pages and the configure related files.  To make use of
+the git-derived files you will need autoconf, autoheader, and a version of python3
+that has the commonmark lib (OR cmarkgfm).  If you have trouble with setting up the
+those required files, you can try running "./prepare-source fetchgen" to grab the
+very latest generated files that were created from the latest commit into the master
+branch.
+
+<p> <b>Note:</b> Since the source repository is a work in progress it may, at
+times, not compile or fail in various ways, though it is usually pretty good.
+
+<h2>Source repository patches</h2>
+
+<p>There are also various patch files in the "rsync-patches" repository that
+represent either some work-in-progress features or features that are considered
+to be a little too fringe-interest for the main release.  See the github link
+above for how to look around at what is available, or snag a release tar file.
+The maintainer like to put the rsync-patches dir into his rsync checkout as a
+directory named "patches" and has some helper scripts for how to use local git
+branches to test and update the diffs.
+
+<!--#include virtual="footer.html" -->

rsync-web/examples.html

@@ -0,0 +1,54 @@
+<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
+<HTML>
+<HEAD>
+<TITLE>rsync examples</TITLE>
+</HEAD>
+<!--#include virtual="header.html" -->
+
+<H2 align="center">rsync examples</H2>
+
+If you have an interesting example of how you use rsync then please
+submit it to the 
+<A HREF="mailto:rsync-bugs@samba.org">rsync-bugs@samba.org</A>
+for inclusion on this page.
+
+<h2>backup to a central backup server with 7 day incremental</h2>
+
+<pre><small>
+<!--#include virtual="backup.txt" -->
+</small></pre>
+
+<H2>backup to a spare disk</H2>
+
+<pre><small>
+<!--#include virtual="horus.txt" -->
+</small></pre>
+
+<H2>mirroring vger CVS tree</H2>
+
+<pre><small>
+<!--#include virtual="vger.txt" -->
+</small></pre>
+
+<H2>automated backup at home</H2>
+
+<pre><small>
+<!--#include virtual="susan.txt" -->
+</small></pre>
+
+<H2>Fancy footwork with remote file lists</H2>
+
+<pre><small>
+One little known feature of rsync is the fact that when run over a
+remote shell (such as rsh or ssh) you can give any shell command as
+the remote file list. The shell command is expanded by your remote
+shell before rsync is called. For example, see if you can work out
+what this does:
+
+	rsync -avR remote:'`find /home -name "*.[ch]"`' /tmp/
+
+note that that is backquotes enclosed by quotes (some browsers don't
+show that correctly).
+</small></pre>
+
+<!--#include virtual="footer.html" -->

rsync-web/features.html

@@ -0,0 +1,28 @@
+<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
+<HTML>
+<HEAD>
+<TITLE>rsync features</TITLE>
+</HEAD>
+<!--#include virtual="header.html" -->
+
+<H2 align="center">rsync features</H2>
+
+rsync is a file transfer program for Unix systems. rsync uses the
+"rsync algorithm" which provides a very fast method for bringing
+remote files into sync. It does this by sending just the differences
+in the files across the link, without requiring that both sets of
+files are present at one of the ends of the link beforehand. <p>
+
+Some features of rsync include
+
+<ul>
+<li> can update whole directory trees and filesystems
+<li> optionally preserves symbolic links, hard links, file ownership,
+permissions, devices and times
+<li> requires no special privileges to install
+<li> internal pipelining reduces latency for multiple files
+<li> can use rsh, ssh or direct sockets as the transport
+<li> supports <A HREF="http://dslab.lzu.edu.cn:8080/members/wangbj/wangbaojun/howtos/rsync-mirror-HOWTO/rsync-mirroring02.html">anonymous rsync</A> which is ideal for mirroring
+</ul>
+
+<!--#include virtual="footer.html" -->

rsync-web/filt

@@ -0,0 +1,13 @@
+P /badge.svg
+H /.git/
+H /.gitignore
+H /bin/
+H /upload/
+P /USE_THE_GIT_rsync-web_REPO_NOW
+H /filt
+H *.swp
+- /netware/
+- /pre-change
+- /rsync-and-debian/
+- /index.html-*
+- *~

rsync-web/find.html

@@ -0,0 +1,13 @@
+<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
+<HTML>
+<HEAD>
+<TITLE>rsync search</TITLE>
+</HEAD>
+<!--#include virtual="header.html" -->
+
+<H2 align="center">Search the rsync web pages</H2>
+
+<p>Searching samba.org will find results in the main web pages and in the
+archived mailing-list pages too.
+
+<!--#include virtual="footer.html" -->

rsync-web/firewall.html

@@ -0,0 +1,189 @@
+<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
+<HTML>
+<HEAD>
+<TITLE>rsync firewall tunneling</TITLE>
+</HEAD>
+<!--#include virtual="header.html" -->
+
+<H2 align="center">Using rsync through a firewall</H2>
+
+<p>If you have a setup where there is no way to directly connect two
+systems for an rsync transfer, there are several ways to get a firewall
+system to act as an intermediary in the transfer.
+
+<p>This first method should work for any remote-shell (e.g. ssh, rsh, etc).
+The other methods are all targeted at ssh (which has a lot of flexibility
+in making a tunneled connection possible).
+
+<h4>Method 1 -- should work with any remote-shell</h4>
+
+<p>Use your remote shell (e.g. ssh) to access the middle system and have it
+use a remote shell to hop over to the actual target system.
+
+<p>To effect this extra hop, you'll need to make sure that the remote-shell
+connection from the middle system to the target system does not involve any
+tty-based user interaction (such as prompting for a password) because there
+is no way for the middle system to access the local user's tty.
+
+<p>One way that should work for all remote-shell programs is to enable host-based
+authentication, which would allow all connections from the middle system to
+the target system to succeed (when the username remains the same).
+However, this may not be a desirable setup.
+
+<p>A better method that works with ssh (and is very safe) is to setup an ssh
+key (see the ssh-keygen manpage) and ensure that ssh-agent forwarding is turned
+on in your ssh client config (e.g. "ForwardAgent&nbsp;yes").  You would put
+the public version of your key onto the middle and target systems (in the
+~/.ssh/authorized_keys file), and the private key on your local system (which
+I recommend you encrypt).  With this setup, a series of ssh connections that
+starts from the system where your private key is available will auto-authorize
+(after a pass-phrase prompt on the first system if your key is encrypted).
+See also ssh-agent for a way to keep a public key unlocked in memory for an
+extended time, and the keychain project for a way to manage a system-wide,
+per-user ssh-agent.
+
+<p>To test that you have things setup right, first test a series of
+remote-shell connections outside of rsync.  A command such as the following
+should work without issuing multiple prompts (use the appropriate remote-shell
+and the substitute the real hostnames for "middle" and "target", of course):
+
+<blockquote><pre>ssh middle ssh target uptime</pre></blockquote>
+
+<p>If you get a password/passphrase prompt to get into the first ("middle")
+system that's fine, but the extra hop needs to occur without any extra user
+interaction.
+
+<p>Once that's done, you can do an rsync copy like this:
+
+<blockquote><pre>rsync -av -e "ssh middle ssh" target:/src/ /dest/</pre></blockquote>
+
+<h4>Method 2 -- requires ssh and nc (netcat)</h4>
+
+<p>Assuming you're using ssh as your remote shell, you can configure ssh to
+use a proxy command to get to the remote host you're interested in reaching.
+Doing this will allow the multi-hop connection to work with rsync, even if
+both hosts prompt for a password -- this is because both ssh connections
+originate from the localhost, and thus both instances of ssh have access to
+the local tty to use for an out-of-band password prompt.
+
+<p>Here is an example config for your ~/.ssh/config file (substitute "target",
+"target_user", and "middle" as appropriate):
+
+<blockquote><pre>Host target
+  ProxyCommand nohup ssh middle nc -w1 %h %p
+  User target_user
+</pre></blockquote>
+
+<p>This proxy setup uses "ssh" to login to the firewall system ("middle") and
+uses "nc" (netcat) to connect to the target host ("target") using the default
+port number.  The use of "nohup" silences a warning at the end of the run, and
+the "-w1" option tells nc to shut down when the connection closes.
+
+<p>With this done, you can run a normal-looking rsync command to "target" that
+will run the proxy command to get through the firewall system:
+
+<blockquote><pre>rsync -av /src/ target:/dest/</pre></blockquote>
+
+<h4>Method 3 -- an alternate ssh method for those without nc (netcat)</h4>
+
+<p>Assuming you're using ssh as your remote shell, you can configure ssh to
+forward a local port through your middle system to the ssh port (22) on the
+target system.  This method does not require the use of nc (it uses only
+ssh to effect the extra hop), but otherwise it is similar to, but slightly
+less convenient than, method 2.
+
+<p>The first thing we need is an ssh configuration that will allow us to
+connect to the forwarded port as if we were connecting to the target
+system, and we need ssh to know what we're doing so that it doesn't
+complain about the host keys being wrong.  We can do this by adding this
+section to your ~/.ssh/config file (substitute "target" and "target_user"
+as appropriate, but leave "localhost" unchanged):
+
+<blockquote><pre>Host target
+  HostName localhost
+  Port 2222
+  HostKeyAlias target
+  User target_user
+</pre></blockquote>
+
+<p>Next, we need to enable the port forwarding:
+
+<blockquote><pre>ssh -fN -l middle_user -L 2222:target:22 middle</pre></blockquote>
+
+<p>What this does is cause a connection to port 2222 on the local system to
+get tunneled to the middle system and then turn into a connection to the
+target system's port 22.  The -N option tells ssh not to start a shell on
+the remote system, which works with modern ssh versions (you can run a
+sleep command if -N doesn't work).  The -f option tells ssh to put the
+command in the background after any password/passphrase prompts.
+
+<p>With this done, you could run a normal-looking rsync command to "target"
+that would use a connection to port 2222 on localhost automatically:
+
+<blockquote><pre>rsync -av target:/src/ /dest/</pre></blockquote>
+
+<p><b>Note:</b> starting an ssh tunnel allows anyone on the source system
+to connect to the localhost port 2222, not just you, but they'd still need
+to be able to login to the target system using their own credentials.
+
+<h4>Method 4 -- for using rsync in daemon-mode (requires nc)</h4>
+
+<p>Install and configure an rsync daemon on the target and use ssh and nc
+to send the socket data to the remote host.
+
+<blockquote><pre>RSYNC_CONNECT_PROG='ssh -l middle_user middle nc %H 873' \
+    rsync daemonuser@target::module/src/ /dest/</pre></blockquote>
+
+<p>(You can also export that variable into your environment if you want to
+perform a series of daemon rsync commands through the same middle host.)
+
+<p>This command takes advantage of the RSYNC_CONNECT_PROG environment
+variable, which tells rsync to pipe its socket data to an external program
+in place of making a direct socket connection.  The command specifed above
+uses ssh to run the nc (netcat) command on the middle host, which forwards
+all socket data to port 873 on the target host (%H).  The "%H" will be
+substituted with the target host from the rsync command as long as you're
+running rsync 3.0.0 or newer (you'll need to replace %H with the actual
+target hostname for earlier rsync versions, which makes the hostname
+specified in the rsync command superfluous).
+
+<h4>Method 5 -- for using rsync in daemon-mode (for those without nc)</h4>
+
+<p>Install and configure an rsync daemon on the target and use an ssh
+tunnel to reach the rsync sever.  This is similar to method 3, but it
+tunnels the daemon port for those that prefer to use an rsync daemon.
+
+<p>(Note that this method also works to tunnel a socket connection
+directly to a destination system if you use "localhost" as the target
+hostname and your destination system's name as the middle hostname.)
+
+<p>Installing the rsync daemon is beyond the scope of this document, but
+see the rsyncd.conf manpage for more information.  Keep in mind that you
+don't need to be root to run an rsync daemon as long as you don't use a
+protected port.
+
+<p>Once your rsync daemon is up and running, you build an ssh tunnel
+through your middle system like this:
+
+<blockquote><pre>ssh -fN -l middle_user -L 8873:target:873 middle</pre></blockquote>
+
+<p>What this does is cause a connection to port 8873 on the local system to
+turn into a connection from the middle system to the target system on port
+873.  (Port 873 is the normal port for an rsync daemon.) The -N option
+tells ssh not to start a shell on the remote system, which works with
+modern ssh versions (you can run a sleep command if -N doesn't work).  The
+-f option tells ssh to put the command in the background after any
+password/passphrase prompts.
+
+<p>Now when an rsync command is executed with a daemon-mode command-line
+syntax to the local system, the conversation is directed to the target
+system.  For example:
+
+<blockquote><pre>rsync -av --port 8873 localhost::module/src/ dest/
+rsync -av rsync://localhost:8873/module/src/ dest/</pre></blockquote>
+
+<p><b>Note:</b> starting an ssh tunnel allows anyone on the source system
+to connect to the localhost port 8873, not just you, so you may want to
+enable username/password restrictions on your rsync daemon.
+
+<!--#include virtual="footer.html" -->

rsync-web/footer.html

@@ -0,0 +1,19 @@
+</td>
+</tr>
+
+  <TR ALIGN="center">
+    <TD><BR><a name="search"></a><img src="bar1.jpg" WIDTH="493" HEIGHT="26" BORDER="0" alt="=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=">
+
+<!-- SiteSearch Google -->
+<form action="https://www.google.com/cse" id="cse-search-box"><div>
+<input type="hidden" name="cx" value="partner-pub-1444957896811922:vxjk7n-bst5" />
+<input type="hidden" name="ie" value="ISO-8859-1" />
+<input type="text" name="q" size="31" />
+<input type="submit" name="sa" value="Search" />
+</div></form>
+<script type="text/javascript" src="https://www.google.com/cse/brand?form=cse-search-box&amp;lang=en"></script>
+<!-- SiteSearch Google -->
+
+  </TD>
+  </TR>
+</TABLE>

rsync-web/header.html

@@ -0,0 +1,40 @@
+<BODY BGCOLOR="#ffffff" TEXT="#000000"
+style="margin-top: 0">
+<TABLE BORDER=0 WIDTH="640" ALIGN="CENTER">
+  <tr VALIGN="middle">
+    <td ALIGN="left">
+	<ul>
+      <li><small><a href=".">home</a></small>
+      <li><small><a href="FAQ.html">FAQ</a></small>
+      <li><small><a href="resources.html">resources</a></small>
+      <li><small><a href="features.html">features</a></small>
+      <li><small><a href="examples.html">examples</a></small>
+      <li><small><a href="lists.html">mailing&nbsp;lists</a></small>
+	</ul>
+    </td>
+    <td align="center">
+      <a href="."><img src="newrsynclogo.jpg" border="0" alt="rsync"></a>
+    </td>
+    <td align="left">
+      <ul>
+      <li><small><a href="bug-tracking.html">bug-tracking</a></small>
+      <li><small><a href="issues.html">current&nbsp;issues and&nbsp;debugging</a></small>
+      <li><small><a href="download.html">download</a></small>
+      <li><small><a href="documentation.html">documentation</a></small>
+      <li><small><a href="find.html">search</a></small>
+      </ul>
+    </td>
+  </tr>
+
+  <TR ALIGN="center">
+    <TD COLSPAN="3">
+    <img src="bar1.jpg" WIDTH="493" HEIGHT="26"
+    BORDER="0"
+    alt="=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=">
+    </TD>
+  </TR>
+</TABLE>
+
+<TABLE BORDER=0 WIDTH="640" ALIGN="CENTER">
+  <tr VALIGN="middle">
+    <td ALIGN="left">

rsync-web/horus.txt

@@ -0,0 +1,28 @@
+I do local backups on several of my machines using rsync. I have an
+extra disk installed that can hold all the contents of the main
+disk. I then have a nightly cron job that backs up the main disk to
+the backup. This is the script I use on one of those machines.
+
+    #!/bin/sh
+
+    export PATH=/usr/local/bin:/usr/bin:/bin
+
+    LIST="rootfs usr data data2"
+
+    for d in $LIST; do
+	mount /backup/$d
+	rsync -ax --exclude fstab --delete /$d/ /backup/$d/
+	umount /backup/$d
+    done
+
+    DAY=`date "+%A"`
+    
+    rsync -a --delete /usr/local/apache /data2/backups/$DAY
+    rsync -a --delete /data/solid /data2/backups/$DAY
+
+   
+
+The first part does the backup on the spare disk. The second part
+backs up the critical parts to daily directories.  I also backup the
+critical parts using a rsync over ssh to a remote machine.
+

rsync-web/how-rsync-works.html

@@ -0,0 +1,350 @@
+<html>
+<head>
+<title>How Rsync Works</title>
+</head>
+<body>
+<h1 align="center">How Rsync Works<br>A Practical Overview</h1>
+<h2 align="center">Foreword</h2>
+<p>
+The original
+<a href="http://rsync.samba.org/tech_report/">Rsync technical report</a>
+and
+Andrew Tridgell's
+<a href="http://samba.org/%7Etridge/phd_thesis.pdf">Phd thesis (pdf)</a>
+Are both excellent documents for understanding the
+theoretical mathematics and some of the mechanics of the rsync algorithm.
+Unfortunately they are more about the theory than the
+implementation of the rsync utility (hereafter referred to as
+Rsync).
+<p>
+In this document I hope to describe...
+<ul>
+	<li>A non-mathematical overview of the rsync algorithm.
+	<li>How that algorithm is implemented in the rsync utility.
+	<li>The protocol, in general terms, used by the rsync utility.
+	<li>The identifiable roles the rsync processes play.
+</ul>
+<p>
+This document be able to serve as a guide for programmers
+needing something of an entr&eacute; into the source code but the
+primary purpose is to give the reader a foundation from
+which he may understand
+<ul>
+	<li>Why rsync behaves as it does.
+	<li>The limitations of rsync.
+	<li>Why a requested feature is unsuited to the code-base.
+</ul>
+<p>
+This document describes in general terms the construction
+and behaviour of Rsync.  In some cases details and exceptions
+that would contribute to specific accuracy have
+been sacrificed for the sake meeting the broader goals.
+<h2 align="center">Processes and Roles</h2>
+<p>
+When we talk about Rsync we use specific terms to refer to
+various processes and their roles in the task performed by
+the utility.  For effective communication it is important that we
+all be speaking the same language; likewise it is important
+that we mean the same things when we use certain terms in a
+given context.  On the rsync mailing list there is often
+some confusion with regards to role and processes.  For
+these reasons I will define a few terms
+used in the role and process contexts that will be used henceforth.
+
+<table cellspacing="20"><tr valign="top">
+<td>client
+</td><td>role
+</td><td>
+	The client initiates the synchronisation.
+</td></tr><tr valign="top">
+<td>server
+</td><td>role
+</td><td>
+	The remote rsync process or system to which the
+	client connects either within a local transfer, via
+	a remote shell or via a network socket.
+	<p>
+	This is a general term and should not be confused with the daemon.
+</td></tr><tr valign="top">
+<td>
+</td><td>
+</td><td bgcolor="#dddddd">
+	Once the connection between the client and server is established
+	the distinction between them is superseded by the
+	sender and receiver roles.
+</td></tr><tr valign="top">
+<td>daemon
+</td><td>Role and process
+</td><td>
+	An Rsync process that awaits connections from
+	clients.  On a certain platform this would be called a
+	service.
+</td></tr><tr valign="top">
+<td>remote&nbsp;shell
+</td><td>role and set of processes
+</td><td>
+	One or more processes that provide connectivity
+	between an Rsync client and an Rsync server on a
+	remote system.
+</td></tr><tr valign="top">
+<td>sender
+</td><td>role and process
+</td><td>
+	The Rsync process that has access to the source
+	files being synchronised.
+</td></tr><tr valign="top">
+<td>receiver
+</td><td>role and process
+</td><td>
+	As a role the receiver is the destination system.
+	As a process the receiver is the process that
+	receives update data and writes it to disk.
+</td></tr><tr valign="top">
+<td>generator
+</td><td>process
+</td><td>
+	The generator process identifies changed files and
+	manages the file level logic.
+</td></tr></table>
+<p>
+<h2 align="center">Process Startup</h2>
+<p>
+When an Rsync client is started it will first establish a
+connection with a server process.  This connection may be
+through pipes or over a network socket.
+<p>
+When Rsync communicates with a remote non-daemon server via
+a remote shell the startup method is to fork the remote
+shell which will start an Rsync server on the remote system.
+Both the Rsync client and server are communicating via pipes
+through the remote shell.  As far as the rsync processes are
+concerned there is no network.
+In this mode the rsync options for the server process are
+passed on the command-line that is used to start the remote
+shell.
+<p>
+When Rsync is communicating with a daemon it is
+communicating directly with a network socket.  This is the
+only sort of Rsync communication that could be called
+network aware.
+In this mode the rsync options must be sent over the socket, as
+described below.
+<p>
+At the very start of the communication between the client
+and the server, they each send the maximum protocol version
+they support to the other side.
+Each side then uses the minimum value as the the protocol
+level for the transfer.
+If this is a daemon-mode connection, rsync options are sent
+from the client to the server.  Then, the exclude list is
+transmitted.  From this point onward the
+client-server relationship is relevant only with regards
+to error and log message delivery.
+<p>
+Local Rsync jobs (when the source and destination are both on locally
+mounted filesystems) are done exactly like a push.  The
+client, which becomes the sender, forks a server process to
+fulfill the receiver role.  The client/sender and
+server/receiver communicate with each other over pipes.
+<h2 align="center">The File List</h2>
+The file list includes not only the pathnames but also
+ownership, mode, permissions, size and modtime.
+If the --checksum option has been specified it also includes
+the file checksums.
+<p>
+The first thing that happens once the startup has completed
+is that the sender will create the file list.
+While it is being built, each entry is transmitted to the
+receiving side in a network-optimised way.
+<p>
+When this is done, each side sorts the file list lexicographically by path
+relative to the base directory of the transfer.
+(The exact sorting algorithm varies depending on what protocol
+version is in effect for the transfer.)
+Once that has happened all references to files
+will be done by their index in the file list.
+<p>
+If necessary the sender follows the file list with id&rarr;name
+tables for users and groups which the receiver will use to
+do a id&rarr;name&rarr;id translation for every file in the file
+list.
+<p>
+After the file list has been received by the receiver, it
+will fork to become the generator and receiver pair
+completing the pipeline.
+<h2 align="center">The Pipeline</h2>
+Rsync is heavily pipelined.  This means that it is a set of
+processes that communicate in a (largely) unidirectional
+way.  Once the file list has been shared the pipeline
+behaves like this:
+<blockquote>
+	generator &rarr; sender &rarr; receiver
+</blockquote>
+<p>
+The output of the generator is input for the sender and the
+output of the sender is input for the receiver.
+Each process runs independently and is delayed only when the
+pipelines stall or when waiting for disk I/O or CPU resources.
+<h2 align="center">The Generator</h2>
+<p>
+The generator process compares the file list with its local
+directory tree.  Prior to beginning its primary function, if
+--delete has been specified, it will first identify local
+files not on the sender and delete them on the receiver.
+<p>
+The generator will then start walking the file list.  Each
+file will be checked to see if it can be skipped.  In the
+most common mode of operation files are not skipped if the
+modification time or size differs.  If --checksum was
+specified a file-level checksum will be created and
+compared.  Directories, device nodes and symlinks are not
+skipped.  Missing directories will be created.  
+<p>
+If a file is not to be skipped, any existing version on the
+receiving side becomes the "basis file" for the transfer, and is
+used as a data source that will help to eliminate matching data
+from having to be sent by the sender.  To effect this remote
+matching of data, block checksums are created for the basis file
+and sent to the sender immediately following the file's index
+number.
+An empty block checksum set is sent for new files and if
+--whole-file was specified.
+<p>
+The block size and, in later versions, the size of the
+block checksum are calculated on a per file basis according
+to the size of that file.
+<h2 align="center">The Sender</h2>
+The sender process reads the file index numbers and associated
+block checksum sets one at a time from the generator.
+<p>
+For each file id the generator sends it will store the
+block checksums and build a hash index of them for rapid lookup.
+<p>
+Then the local file is read and a checksum is
+generated for the block beginning with the first byte of the
+local file.  This block checksum is looked for in the
+set that was sent by the generator, and if no match is found,
+the non-matching byte will be appended to the non-matching data
+and the block starting at the next byte will be compared.
+This is what
+is referred to as the &ldquo;rolling checksum&rdquo;
+<p>
+If a block checksum match is found it is considered a
+matching block and any accumulated non-matching data will be
+sent to the receiver followed by the offset and length in
+the receiver's file of the matching block and the block
+checksum generator will be advanced to the next byte after
+the matching block.
+<p>
+Matching blocks can be identified in this way even if
+the blocks are reordered or at different offsets.  
+This process is the very heart of the rsync algorithm.
+<p>
+In this way, the sender will give the receiver instructions for
+how to reconstruct the source file into a new destination file.
+These instructions detail all the matching data that can be
+copied from the basis file (if one exists for the transfe),
+and includes any raw data that was not available locally.
+At the end of each file's processing a whole-file
+checksum is sent and the sender proceeds with the next
+file.
+<p>
+Generating the rolling checksums and searching for matches
+in the checksum set sent by the generator require a good
+deal of CPU power.  Of all the rsync processes it is the
+sender that is the most CPU intensive.
+<h2 align="center">The Receiver</h2>
+<p>
+The receiver will read from the sender data for each file
+identified by the file index number.  It will open the local
+file (called the basis) and will create a temporary file.
+<p>
+The receiver will expect to read non-matched data and/or to match
+records all in sequence for the final file contents.  When
+non-matched data is read it will be written to the
+temp-file.  When a block match record is received the
+receiver will seek to the block offset in the basis file
+and copy the block to the temp-file.  In this way the
+temp-file is built from beginning to end.
+<p>
+The file's checksum is generated as the temp-file is built.
+At the end of the file, this checksum is compared with the
+file checksum from the sender.  If the file checksums do not
+match the temp-file is deleted.  If the file fails once it
+will be reprocessed in a second phase, and if it fails twice
+an error is reported.
+<p>
+After the temp-file has been completed, its ownership and
+permissions and modification time are set.  It is then
+renamed to replace the basis file.
+<p>
+Copying data from the basis file to the temp-file make the
+receiver the most disk intensive of all the rsync processes.
+Small files may still be in disk cache mitigating this but
+for large files the cache may thrash as the generator has
+moved on to other files and there is further latency caused
+by the sender.  As
+data is read possibly at random from one file and written to
+another, if the working set is larger than the disk cache,
+then what is called a seek storm can occur, further
+hurting performance.
+<h2 align="center">The Daemon</h2>
+The daemon process, like many daemons, forks for every
+connection.  On startup, it parses the rsyncd.conf file
+to determine what modules exist and to set the global options.
+<p>
+When a connection is received for a defined module the
+daemon forks a new child process to handle the connection.
+That child process then reads the rsyncd.conf file to set
+the options for the requested module, which may chroot to the
+module path and may drop setuid and setgid for the
+process.  After that it will behave just like any other
+rsync server process adopting either a sender or receiver
+role.
+<h2 align="center">The Rsync Protocol</h2>
+<p>
+A well-designed communications protocol has a number of
+characteristics.
+<ul>
+	<li>Everything is sent in well defined packets with
+		a header and an optional body or data payload.
+	<li>In each packet's header a type and or command
+		specified.
+	<li>Each packet has a definite length.
+</ul>
+<p>
+In addition to these characteristics, protocols have varying degrees of
+statefulness, inter-packet independence, human readability,
+and the ability to reestablish a disconnected session.
+<p>
+Rsync's protocol has none of these good characteristics.  The data is
+transferred as an unbroken stream of bytes.  With the
+exception of the unmatched file-data, there are no length
+specifiers nor counts.  Instead the meaning of each byte is
+dependent on its context as defined by the protocol level.
+<p>
+As an example, when the sender is sending  the file list it
+simply sends each file list entry and terminates the list
+with a null byte.  Within the file list entries, a bitfield
+indicates which fields of the structure to expect and those
+that are variable length strings are simply null terminated.
+The generator sending file numbers and block checksum sets
+works the same way.
+<p>
+This method of communication works quite well on reliable
+connections and it certainly has less data overhead than the
+formal protocols.  It unfortunately makes the protocol
+extremely difficult to document, debug or extend.
+Each version of the protocol will have subtle differences on
+the wire that can only be anticipated by knowing the exact
+protocol version.
+<h2 align="center">notes</h2>
+This document is a work in progress.  The author expects
+that it has some glaring oversights and some portions that may be
+more confusing than enlightening for some readers.  It is
+hoped that this could evolve into a useful reference.
+<p>
+Specific suggestions for improvement are welcome, as would be a
+complete rewrite.
+</body>
+</html>

rsync-web/index.html

@@ -0,0 +1,528 @@
+<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
+<HTML>
+<HEAD>
+<TITLE>rsync</TITLE>
+<style>
+.security { color: red; }
+h3 { margin-bottom: 0px; }
+.date { color: #D25A0B; }
+</style>
+</HEAD>
+<!--#include virtual="header.html" -->
+
+<H2 align="center">Welcome to the rsync web pages</H2>
+
+rsync is an <A HREF="https://www.opensource.org/">open source</A>
+utility that provides fast incremental file transfer. rsync is freely
+available under the <A HREF="GPL.html">GNU General Public
+License</A> and is currently being maintained by
+<a href="email:<tridge@samba.org>">Andrew Tridgell</a>.
+
+<p>A full changelog of all the releases, including upcoming releases, is in the
+<a href="https://download.samba.org/pub/rsync/NEWS">NEWS file</a>.
+
+<div style="float: right">
+<a href="https://github.com/RsyncProject/rsync/actions">
+<img src="badge.svg">
+</a></div>
+
+<h3>Rsync version 3.4.3 released</h3>
+<i class=date>May 20th, 2026</i>
+
+<p>Rsync version 3.4.3 has been released. This is a major security release.
+
+<p>See the <a href="https://download.samba.org/pub/rsync/NEWS#3.4.3">3.4.3 NEWS</a> for a detailed changelog.
+The latest manpages are also available for:<ul>
+<li><a href="https://download.samba.org/pub/rsync/rsync.1"><b>rsync</b>(1)</a>
+<li><a href="https://download.samba.org/pub/rsync/rsync-ssl.1"><b>rsync-ssl</b>(1)</a>
+<li><a href="https://download.samba.org/pub/rsync/rsyncd.conf.5"><b>rsyncd.conf</b>(5)</a>
+<li><a href="https://download.samba.org/pub/rsync/rrsync.1"><b>rrsync</b>(1)</a>
+</ul>
+
+<p>The source tar is available here:
+<b><a href="https://download.samba.org/pub/rsync/src/rsync-3.4.3.tar.gz">rsync-3.4.3.tar.gz</a>
+(<a href="https://download.samba.org/pub/rsync/src/rsync-3.4.3.tar.gz.asc">signature</a>)</b>,
+and the diffs from version 3.4.2 are available here:
+<b><a href="https://download.samba.org/pub/rsync/src-diffs/rsync-3.4.2-3.4.3.diffs.gz">rsync-3.4.2-3.4.3.diffs.gz</a>
+(<a href="https://download.samba.org/pub/rsync/src-diffs/rsync-3.4.2-3.4.3.diffs.gz.asc">signature</a>)</b>.
+
+<h3>Rsync version 3.4.2 released</h3>
+<i class=date>April 28th, 2026</i>
+
+<p>Rsync version 3.4.2 has been released.
+
+<p>See the <a href="https://download.samba.org/pub/rsync/NEWS#3.4.2">3.4.2 NEWS</a> for a detailed changelog.
+The latest manpages are also available for:<ul>
+<li><a href="https://download.samba.org/pub/rsync/rsync.1"><b>rsync</b>(1)</a>
+<li><a href="https://download.samba.org/pub/rsync/rsync-ssl.1"><b>rsync-ssl</b>(1)</a>
+<li><a href="https://download.samba.org/pub/rsync/rsyncd.conf.5"><b>rsyncd.conf</b>(5)</a>
+<li><a href="https://download.samba.org/pub/rsync/rrsync.1"><b>rrsync</b>(1)</a>
+</ul>
+
+<p>The source tar is available here:
+<b><a href="https://download.samba.org/pub/rsync/src/rsync-3.4.2.tar.gz">rsync-3.4.2.tar.gz</a>
+(<a href="https://download.samba.org/pub/rsync/src/rsync-3.4.2.tar.gz.asc">signature</a>)</b>,
+and the diffs from version 3.4.1 are available here:
+<b><a href="https://download.samba.org/pub/rsync/src-diffs/rsync-3.4.1-3.4.2.diffs.gz">rsync-3.4.1-3.4.2.diffs.gz</a>
+(<a href="https://download.samba.org/pub/rsync/src-diffs/rsync-3.4.1-3.4.2.diffs.gz.asc">signature</a>)</b>.
+
+<h3>Rsync version 3.4.1 released</h3>
+<i class=date>January 15th, 2025</i>
+
+<p>Rsync version 3.4.1 has been released.
+This is a fix for some regressions in 3.4.0
+
+<p>See the <a href="https://download.samba.org/pub/rsync/NEWS#3.4.1">3.4.1 NEWS</a> for a detailed changelog.
+The latest manpages are also available for:<ul>
+<li><a href="https://download.samba.org/pub/rsync/rsync.1"><b>rsync</b>(1)</a>
+<li><a href="https://download.samba.org/pub/rsync/rsync-ssl.1"><b>rsync-ssl</b>(1)</a>
+<li><a href="https://download.samba.org/pub/rsync/rsyncd.conf.5"><b>rsyncd.conf</b>(5)</a>
+<li><a href="https://download.samba.org/pub/rsync/rrsync.1"><b>rrsync</b>(1)</a>
+</ul>
+
+<p>The source tar is available here:
+<b><a href="https://download.samba.org/pub/rsync/src/rsync-3.4.1.tar.gz">rsync-3.4.1.tar.gz</a>
+(<a href="https://download.samba.org/pub/rsync/src/rsync-3.4.1.tar.gz.asc">signature</a>)</b>,
+with a tar file of the "rsync-patches" repository released in a separate file:
+<b><a href="https://download.samba.org/pub/rsync/src/rsync-patches-3.4.1.tar.gz">rsync-patches-3.4.1.tar.gz</a>
+(<a href="https://download.samba.org/pub/rsync/src/rsync-patches-3.4.1.tar.gz.asc">signature</a>)</b>,
+and the diffs from version 3.4.0 are available here:
+<b><a href="https://download.samba.org/pub/rsync/src-diffs/rsync-3.4.0-3.4.1.diffs.gz">rsync-3.4.0-3.4.1.diffs.gz</a>
+(<a href="https://download.samba.org/pub/rsync/src-diffs/rsync-3.4.0-3.4.1.diffs.gz.asc">signature</a>)</b>.
+
+
+<h3>Rsync version 3.4.0 released</h3>
+<i class=date>January 14th, 2025</i>
+
+<p>Rsync version 3.4.0 has been released.
+This is a security release, fixing several important security vulnerabilities.
+
+<p>See the <a href="https://download.samba.org/pub/rsync/NEWS#3.4.0">3.4.0 NEWS</a> for a detailed changelog.
+
+<p>The source tar is available here:
+<b><a href="https://download.samba.org/pub/rsync/src/rsync-3.4.0.tar.gz">rsync-3.4.0.tar.gz</a>
+(<a href="https://download.samba.org/pub/rsync/src/rsync-3.4.0.tar.gz.asc">signature</a>)</b>,
+with a tar file of the "rsync-patches" repository released in a separate file:
+<b><a href="https://download.samba.org/pub/rsync/src/rsync-patches-3.4.0.tar.gz">rsync-patches-3.4.0.tar.gz</a>
+(<a href="https://download.samba.org/pub/rsync/src/rsync-patches-3.4.0.tar.gz.asc">signature</a>)</b>,
+and the diffs from version 3.3.0 are available here:
+<b><a href="https://download.samba.org/pub/rsync/src-diffs/rsync-3.3.0-3.4.0.diffs.gz">rsync-3.3.0-3.4.0.diffs.gz</a>
+(<a href="https://download.samba.org/pub/rsync/src-diffs/rsync-3.3.0-3.4.0.diffs.gz.asc">signature</a>)</b>.
+
+<h3>Rsync version 3.3.0 released</h3>
+<i class=date>April 6th, 2024</i>
+
+<p>Rsync version 3.3.0 has been released.
+This is a bug-fix release.
+
+<p>See the <a href="https://download.samba.org/pub/rsync/NEWS#3.3.0">3.3.0 NEWS</a> for a detailed changelog.
+
+<p>The source tar is available here:
+<b><a href="https://download.samba.org/pub/rsync/src/rsync-3.3.0.tar.gz">rsync-3.3.0.tar.gz</a>
+(<a href="https://download.samba.org/pub/rsync/src/rsync-3.3.0.tar.gz.asc">signature</a>)</b>,
+with a tar file of the "rsync-patches" repository released in a separate file:
+<b><a href="https://download.samba.org/pub/rsync/src/rsync-patches-3.3.0.tar.gz">rsync-patches-3.3.0.tar.gz</a>
+(<a href="https://download.samba.org/pub/rsync/src/rsync-patches-3.3.0.tar.gz.asc">signature</a>)</b>,
+and the diffs from version 3.2.2 are available here:
+<b><a href="https://download.samba.org/pub/rsync/src-diffs/rsync-3.2.7-3.3.0.diffs.gz">rsync-3.2.7-3.3.0.diffs.gz</a>
+(<a href="https://download.samba.org/pub/rsync/src-diffs/rsync-3.2.7-3.3.0.diffs.gz.asc">signature</a>)</b>.
+
+<p><hr>
+<h3>Rsync version 3.2.7 released</h3>
+<i class=date>October 20th, 2022</i>
+
+<p>Rsync version 3.2.7 has been released.
+This has some new features &amp; fixes, including various bug fixes for arg validation &amp; filter-rule validation.
+
+<p>See the <a href="https://download.samba.org/pub/rsync/NEWS#3.2.7">3.2.7 NEWS</a> for a detailed changelog.
+
+<p>The source tar is available here:
+<b><a href="https://download.samba.org/pub/rsync/src/rsync-3.2.7.tar.gz">rsync-3.2.7.tar.gz</a>
+(<a href="https://download.samba.org/pub/rsync/src/rsync-3.2.7.tar.gz.asc">signature</a>)</b>,
+with a tar file of the "rsync-patches" repository released in a separate file:
+<b><a href="https://download.samba.org/pub/rsync/src/rsync-patches-3.2.7.tar.gz">rsync-patches-3.2.7.tar.gz</a>
+(<a href="https://download.samba.org/pub/rsync/src/rsync-patches-3.2.7.tar.gz.asc">signature</a>)</b>,
+and the diffs from version 3.2.2 are available here:
+<b><a href="https://download.samba.org/pub/rsync/src-diffs/rsync-3.2.6-3.2.7.diffs.gz">rsync-3.2.6-3.2.7.diffs.gz</a>
+(<a href="https://download.samba.org/pub/rsync/src-diffs/rsync-3.2.6-3.2.7.diffs.gz.asc">signature</a>)</b>.
+
+<p><hr>
+<h3>Rsync version 3.2.6 released</h3>
+<i class=date>September 9th, 2022</i>
+
+<p>Rsync version 3.2.6 has been released.
+This is a bug-fix release.
+
+<p>See the <a href="https://download.samba.org/pub/rsync/NEWS#3.2.6">3.2.6 NEWS</a> for a detailed changelog.
+
+<p>The source tar is available here:
+<b><a href="https://download.samba.org/pub/rsync/src/rsync-3.2.6.tar.gz">rsync-3.2.6.tar.gz</a>
+(<a href="https://download.samba.org/pub/rsync/src/rsync-3.2.6.tar.gz.asc">signature</a>)</b>,
+with a tar file of the "rsync-patches" repository released in a separate file:
+<b><a href="https://download.samba.org/pub/rsync/src/rsync-patches-3.2.6.tar.gz">rsync-patches-3.2.6.tar.gz</a>
+(<a href="https://download.samba.org/pub/rsync/src/rsync-patches-3.2.6.tar.gz.asc">signature</a>)</b>,
+and the diffs from version 3.2.2 are available here:
+<b><a href="https://download.samba.org/pub/rsync/src-diffs/rsync-3.2.5-3.2.6.diffs.gz">rsync-3.2.5-3.2.6.diffs.gz</a>
+(<a href="https://download.samba.org/pub/rsync/src-diffs/rsync-3.2.5-3.2.6.diffs.gz.asc">signature</a>)</b>.
+
+<p><hr>
+<h3>Rsync version 3.2.5 released</h3>
+<i class=date>August 14th, 2022</i>
+
+<p>Rsync version 3.2.5 has been released.
+This is a bug-fix and <a href="security.html" class=security>security</a> release.
+
+<p>See the <a href="https://download.samba.org/pub/rsync/NEWS#3.2.5">3.2.5 NEWS</a> for a detailed changelog.
+
+<p><hr>
+<h3>Rsync version 3.2.4 released</h3>
+<i class=date>April 15th, 2022</i>
+
+<p>Rsync version 3.2.4 has been released.
+Another typical release with both bug fixes and some enhancements. It also contains a
+<a href="security.html#s3_2_4" class=security>security fix</a>
+for the bundled zlib 1.2.8, which may or may not be used in your particular build configuration.
+
+<p>See the <a href="https://download.samba.org/pub/rsync/NEWS#3.2.4">3.2.4 NEWS</a>
+for a detailed changelog.
+
+<p><hr>
+<h3>Rsync version 3.2.3 released</h3>
+<i class=date>August 6th, 2020</i>
+
+<p>Rsync version 3.2.3 has been released.
+It contains a smattering of bug fixes and various enhancements.
+
+<p>See the <a href="https://download.samba.org/pub/rsync/NEWS#3.2.3">3.2.3 NEWS</a>
+for a detailed changelog.
+
+<p><hr>
+<h3>Rsync version 3.2.2 released</h3>
+<i class=date>July 4th, 2020</i>
+
+<p>Rsync version 3.2.2 has been released.
+This is a few more portability fixes, some improvements to the newest features, and some
+other simple changes.  Hopefully this will be the last of these recent touch-up releases.
+
+<p>See the <a href="https://download.samba.org/pub/rsync/NEWS#3.2.2">3.2.2 NEWS</a>
+for a detailed changelog.
+
+<p><hr>
+<h3>Rsync version 3.2.1 released</h3>
+<i class=date>June 22th, 2020</i>
+
+<p>Rsync version 3.2.1 has been released.
+This is mainly a few fixes for some release issues and portability problems.
+There's also a couple new features, just for good measure.
+
+<p>See the <a href="https://download.samba.org/pub/rsync/NEWS#3.2.1">3.2.1 NEWS</a>
+for a detailed changelog.
+
+<p><hr>
+<h3>Rsync version 3.2.0 released</h3>
+<i class=date>June 19th, 2020</i>
+
+<p>Rsync version 3.2.0 has been released.
+This release has a good number of a few new features and various bug fixes.
+
+<p>See the <a href="https://download.samba.org/pub/rsync/NEWS#3.2.0">3.2.0 NEWS</a>
+for a detailed changelog.
+
+<p><hr>
+<h3>Rsync version 3.1.3 released</h3>
+<i class=date>January 28th, 2018</i>
+
+<p>Rsync version 3.1.3 has been released.
+This release has a
+<a href="security.html#s3_1_3" class=security>couple security fixes</a>,
+a few new features, and various bug fixes.
+
+<p>See the <a href="https://download.samba.org/pub/rsync/NEWS#3.1.3">3.1.3 NEWS</a>
+for a detailed changelog.
+
+<p><hr>
+<h3>Rsync version 3.1.2 released</h3>
+<i class=date>December 21st, 2015</i>
+
+<p>Rsync version 3.1.2 has been released.  This is a bug-fix release.
+It includes a <a href="security.html#s3_1_2" class=security>security fix</a>
+for a transfer from a sender that you don't fully trust.
+
+<p>See the <a href="https://download.samba.org/pub/rsync/NEWS#3.1.2">3.1.2 NEWS</a>
+for a detailed changelog.
+
+<p><hr>
+<h3>Rsync version 3.1.1 released</h3>
+<i class=date>June 22nd, 2014</i>
+
+<p>Rsync version 3.1.1 has been released.  This is a bug-fix release.
+
+<p>See the <a href="https://download.samba.org/pub/rsync/NEWS#3.1.1">3.1.1 NEWS</a>
+for a detailed changelog.
+
+<p><hr>
+<h3>Rsync version 3.1.0 released</h3>
+<i class=date>September 28th, 2013</i>
+
+<p>Rsync version 3.1.0 has been released.  This is a
+feature release that improves performance, provides several new options, and
+fixes a few bugs along the way.
+
+<p>See the <a href="https://download.samba.org/pub/rsync/NEWS#3.1.0">3.1.0 NEWS</a>
+for a detailed changelog.
+
+<p><hr>
+<h3>Rsync version 3.0.9 released</h3>
+<i class=date>September 23th, 2011</i>
+
+<p>Rsync version 3.0.9 has been released.  This is a bug-fix release.
+
+<p>See the <a href="https://download.samba.org/pub/rsync/NEWS#3.0.9">3.0.9 NEWS</a>
+for a detailed changelog.
+
+<p><hr>
+<h3>Rsync version 3.0.8 released</h3>
+<i class=date>March 26th, 2011</i>
+
+<p>Rsync version 3.0.8 has been released.  This is a bug-fix release.
+
+<p>See the <a href="https://download.samba.org/pub/rsync/NEWS#3.0.8">3.0.8 NEWS</a>
+for a detailed changelog.
+
+<p><hr>
+<h3>Rsync version 3.0.7 released</h3>
+<i class=date>December 31th, 2009</i>
+
+<p>Rsync version 3.0.7 has been released.  This is a bug-fix release.
+
+<p>See the <a href="https://download.samba.org/pub/rsync/NEWS#3.0.7">3.0.7 NEWS</a>
+for a detailed changelog.
+
+<p><hr>
+<h3>Rsync version 3.0.6 released</h3>
+<i class=date>May 8th, 2009</i>
+
+<p>Rsync version 3.0.6 has been released.  This is a bug-fix release.
+
+<p>See the <a href="https://download.samba.org/pub/rsync/NEWS#3.0.6">3.0.6 NEWS</a>
+for a detailed changelog.
+
+<p><hr>
+<h3>Rsync version 3.0.5 released</h3>
+<i class=date>December 28th, 2008</i>
+
+<p>Rsync version 3.0.5 has been released.  This is another bug-fix release.
+
+<p>See the <a href="https://download.samba.org/pub/rsync/NEWS#3.0.5">3.0.5 NEWS</a>
+for a detailed changelog.
+
+<p><hr>
+<h3>Rsync version 3.0.4 released</h3>
+<i class=date>September 6th, 2008</i>
+
+<p>Rsync version 3.0.4 has been released.  This is a bug-fix release with the
+only enhancement being the adding of a way to interact with an
+overly-restrictive server that refuses rsync's behind-the-scenes use of the -e
+option.
+
+<p>See the <a href="https://download.samba.org/pub/rsync/NEWS#3.0.4">3.0.4 NEWS</a>
+for a detailed changelog.
+
+<p><hr>
+<h3>Rsync version 3.0.3 released</h3>
+<i class=date>June 29th, 2008</i>
+
+<p>Rsync version 3.0.3 has been released.  This is a bug-fix release that has
+no new features (though it does have one new script in the support directory).
+
+<p>See the <a href="https://download.samba.org/pub/rsync/NEWS#3.0.3">3.0.3 NEWS</a>
+for a detailed changelog.
+
+<p><hr>
+<h3>Rsync version 3.0.2 released</h3>
+<i class=date>April 8th, 2008</i>
+
+<p>Rsync version 3.0.2 has been released.  This is a
+<a href="security.html#s3_0_2" class=security>security release</a>
+that fixes a potential buffer-overflow issue.
+
+<p>See the <a href="https://download.samba.org/pub/rsync/NEWS#3.0.2">3.0.2 NEWS</a>
+for a detailed changelog.
+
+<p><hr>
+<h3>Rsync version 3.0.1 released</h3>
+<i class=date>April 3rd, 2008</i>
+
+<p>Rsync version 3.0.1 has been released.  This is a bug-fix release, which also
+includes fixes/improvements for several issues in the daemon-exclude code.
+
+<p>See the <a href="https://download.samba.org/pub/rsync/NEWS#3.0.1">3.0.1 NEWS</a>
+for a detailed changelog.
+
+<p><hr>
+<h3>Rsync version 3.0.0 released</h3>
+<i class=date>March 1st, 2008</i>
+
+<p>Rsync version 3.0.0 is finally here!  This is a feature release that
+also includes quite a few bug fixes.
+
+<p>The 3.0.0 version number is such a large bump up from 2.6.9 due to the
+addition of an
+incremental recursion scan (which helps a lot with large transfers) and the
+official arrival of several other new features, including ACL support, extended
+attribute support, filename character-set conversion, etc.
+
+<p>See the <a href="https://download.samba.org/pub/rsync/NEWS#3.0.0">3.0.0 NEWS</a>
+for a detailed changelog.
+
+<p><hr>
+<h3>Rsync version 2.6.9 released</h3>
+<i class=date>November 6th, 2006</i>
+
+<p>Rsync version 2.6.9 has been released.  This is primarily a bug-fix
+release with a few minor new features.
+
+<p>See the <a href="https://download.samba.org/pub/rsync/NEWS#2.6.9">2.6.9 NEWS</a>
+for a detailed changelog.
+
+<p><hr>
+<h3>Rsync version 2.6.8 released</h3>
+<i class=date>April 22th, 2006</i>
+
+<p>Rsync version 2.6.8 has been released.  This is a bug-fix release that
+primarily addresses an exclude problem that affected the --relative option,
+but also includes a <a href="security.html#s2_6_8" class=security>security fix</a> for
+the xattrs.diff patch (which is not an
+official part of rsync, but some packagers include it in their release).
+
+<p>See the <a href="https://download.samba.org/pub/rsync/NEWS#2.6.8">2.6.8 NEWS</a>
+for a detailed changelog.
+
+<p><hr>
+<h3>Rsync 2.6.7 released</h3>
+<i class=date>March 11th, 2006</i>
+
+<p>Rsync version 2.6.7 has been released.  This release has both several new
+features and the usual accompaniment of bug fixes.
+
+<p>See the <a href="https://download.samba.org/pub/rsync/NEWS#2.6.7">2.6.7 NEWS</a>
+for a detailed changelog.
+
+<p><hr>
+<h3>Rsync 2.6.6 released</h3>
+<i class=date>July 28th, 2005</i>
+
+<p>Rsync version 2.6.6 has been released.  This release is a bug-fix release
+which contains a <a href="security.html#s2_6_6" class=security>security fix</a>
+to handle a null-pointer bug that turned up in rsync's version of zlib
+1.1.4 (this is not the recent zlib 1.2.2 security fix, which did not
+affect rsync) and to squash a few other minor bugs.  To deal with the
+zlib issue, rsync has been upgraded to include zlib 1.2.3.
+
+<p>See the <a href="https://download.samba.org/pub/rsync/NEWS#2.6.6">2.6.6 NEWS</a>
+for a detailed changelog.
+
+<p><hr>
+<h3>Rsync 2.6.5 released</h3>
+<i class=date>June 1st, 2005</i>
+
+<p>Rsync version 2.6.5 has been released.  This release is primarily a bug-fix
+release to squash some annoying problems that made it into the (feature-filled)
+release of 2.6.4, plus a few minor enhancements.
+
+<p>See the <a href="https://download.samba.org/pub/rsync/NEWS#2.6.5">2.6.5 NEWS</a>
+for a detailed changelog.
+
+<p><hr>
+<h3>Rsync 2.6.4 released</h3>
+<i class=date>March 30th, 2005</i>
+
+<p>Rsync version 2.6.4 has been released.  This release combines quite a
+few new features, some improved delete efficiency, and the usual array of
+bug fixes.
+
+<p>See the <a href="https://download.samba.org/pub/rsync/NEWS#2.6.4">2.6.4 NEWS</a>
+for a detailed changelog.
+
+<p><hr>
+<h3>Rsync 2.6.3 released</h3>
+<i class=date>September 30th, 2004</i>
+
+<p>Rsync version 2.6.3 has been released.  It contains several new features
+and quite a few bug fixes, including a <a href="security.html#s2_6_3" class=security>security
+fix</a> for a patch-sanitizing bug in the daemon code.
+
+<p>See the <a href="https://download.samba.org/pub/rsync/NEWS#2.6.3">2.6.3 NEWS</a>
+for a detailed changelog.
+
+<p><hr>
+<h3>Rsync 2.6.2 released</h3>
+<i class=date>April 30th, 2004</i>
+
+<p>Rsync version 2.6.2 has been released.  It is a bug-fix release that mainly
+fixes <b>a bug with the --relative option (-R) in 2.6.1</b>
+that could cause files to be transferred incorrectly.  This only affected a
+source right at the root of the filesystem, such as "/" or "/*" (if you
+first "cd /" and then copy from ".", it would not tickle the bug).
+
+<p>See the <a href="https://download.samba.org/pub/rsync/NEWS#2.6.2">2.6.2 NEWS</a>
+for a detailed changelog.
+
+<p><hr>
+<h3>Rsync 2.6.1 released</h3>
+<i class=date>April 26th, 2004</i>
+
+<p>Rsync version 2.6.1 has been released.  It is primarily a performance
+release that requires less memory to run, makes fewer write calls to the socket
+(lowering the system CPU time), does less string copying (lowering the user CPU
+time), and also reduces the amount of data that is transmitted over the wire.
+There have also been quite a few bug fixes, including a
+<a href="security.html#s2_6_1" class=security>security fix</a> for a daemon problem when chroot
+is not enabled.  See the
+<a href="https://download.samba.org/pub/rsync/NEWS#2.6.1">2.6.1 NEWS</a>
+for a detailed changelog.
+
+<p><hr>
+<h3>One Cygwin hang-problem resolved</h3>
+
+<p>The problem with rsync hanging at the end of the transfer on
+<a href="https://www.cygwin.com/">Cygwin</a> had been previously traced to a
+signal-handling bug in their compatibility DLL.  This bug appears to now be
+fixed in DLL version 1.5.7-1, and Cygwin users are reporting that upgrading the
+DLL removes the hang-at-end-of-transfer problem for their existing rsync executable.
+(Note that this doesn't solve a hang that some folks see in the middle of a
+transfer -- using daemon mode instead of ssh can work around that one.)
+
+<p><hr>
+<h3>Rsync 2.6.0 released</h3>
+<i class=date>January 1st, 2004</i>
+
+<P> Two important things to note in the new release:
+
+<ol>
+
+<li>The default remote shell is now "ssh" unless you tell configure you want to
+make something else the default.
+
+<li>Some bug fixes in the include/exclude code, while making things work
+properly, have resulted in some user-visible changes for certain wildcard
+strings.  Read the BUG FIXES section in the
+<a href="https://download.samba.org/pub/rsync/NEWS#2.6.0">2.6.0 NEWS</a>
+to see if any of these changes apply to you.
+(Most people should be unaffected.)
+
+</ol>
+
+<p>One other item of note is that the oft-requested option "--files-from" is now
+available.  This option lets you specify a list of files to transfer, and can
+be much more efficient than a recursive descent using include/exclude
+statements (if you know in advance what files you want to transfer).  The list
+of files can come from either side of the connection, so it is possible for a
+server to provide the file-list that lets someone grab a server-specified set of
+files, for example.  See the <a href="https://download.samba.org/pub/rsync/rsync.1">rsync man page</a>
+for more details.
+
+<p>See the <a href="https://download.samba.org/pub/rsync/NEWS#2.6.0">2.6.0 NEWS</a>
+for a detailed changelog.
+
+<!--#include virtual="footer.html" -->

rsync-web/issues.html

@@ -0,0 +1,210 @@
+<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
+<HTML>
+<HEAD>
+<TITLE>rsync current issues</TITLE>
+<style>
+code {
+  font-family: monospace;
+  font-weight: bold;
+  white-space: pre;
+}
+pre code {
+  display: block;
+  font-weight: normal;
+}
+blockquote pre code {
+  background: #f1f1f1;
+}
+</style>
+</HEAD>
+<!--#include virtual="header.html" -->
+
+<H2 align="center">current issues and debugging</H2>
+
+<ol>
+
+<li><p><b>Q:</b>
+
+Rsync appears hung -- what should I do?
+
+<p><b>A:</b>
+
+When experiencing a hang or freeze <b>please</b> gather the following
+information before killing the rsync process:
+
+<ul>
+
+<li> The state of the send/receive queues shown with netstat on the two ends.
+
+<li> The system call that each of the 3 processes is stuck in (use truss on
+solaris, strace on Linux, etc.).
+
+</ul>
+
+<p>Try telling rsync on both sides of the connection to send messages to
+stderr, which might make the failure message visible.  i.e., use:
+
+<blockquote<pre><code>
+--msgs2stderr -M--msgs2stderr
+</code></pre></blockquote>
+
+<p>That alone might get rsync to stop hanging.  Also, if you're using more than
+one <code>--verbose</code> (<code>-v</code>) option then I have 2 simple words
+for you: stop it.  If you need more info on what rsync is changing, using the
+<code>--itemize-changes</code> option (<code>-i</code>) and repeat it if you
+need to see unchanged files. This is a much better way to go that doesn't fill
+up the communication pipeline with a large quanity of debug messages.
+
+<p>See the "rsync-debug" script below for an example of how to grab strace
+information from the remote rsync process(es).  If you need help, send email to
+the mailing list. 
+
+<li><p><b>Q:</b>
+
+Why does my chrooted rsync daemon crash when doing an LDAP lookup for a user or
+group?
+
+<p><b>A:</b>
+
+There is a bug in some LDAP libraries (e.g. Fedora Core 3) where it crashes
+when someone looks up a name from inside a chrooted process (one that does not
+contain copies of the libraries to perform the lookup).  This is a bug that the
+LDAP libraries will need to fix, and is out of rsync's hands.  You can work
+around the problem by using the <code>--numeric-ids</code> option, turning
+chroot off, or getting rid of LDAP lookups.
+
+<li><p><b>Q:</b>
+
+Why does my transfer die with something like the following error?
+
+<blockquote><pre><code>
+rsync: error writing 4 unbuffered bytes - exiting: Broken pipe
+rsync error: error in rsync protocol data stream (code 12) at io.c(463)
+</code></pre></blockquote>
+
+<p>or
+
+<blockquote><pre><code>
+rsync: connection unexpectedly closed (24 bytes read so far)
+rsync error: error in rsync protocol data stream (code 12) at io.c(342)
+</code></pre></blockquote>
+
+<p><b>A:</b>
+
+This error tells you that the local rsync was trying to talk to the remote
+rsync, but the connection to that rsync is now gone.  The thing you must
+figure out is <b>why</b>, and that can involve some investigative work.
+
+<p>It is a good idea use the <code>--msgs2stderr</code> options mentioned at
+the top of this page to get rsync to output any errors it encounters to stderr
+instead of trying to write them down the failing pipeline.
+
+<p>If the connection is via ssh (or other remote-shell command) then you should
+run some tests to make sure that you can actually run the remote rsync and that
+your shell isn't injecting extraneous output into the rsync stream.  For instance,
+try running these two commands using whatever HOST (and user) options you need:
+
+<blockquote><pre><code>
+echo hi | ssh HOST cat
+ssh HOST rsync --version
+</code></pre></blockquote>
+
+<p>The first command should output just the string "hi" and nothing else. The
+second command should successfully start the remote rsync and report its version.
+
+<p>If the remote rsync is a daemon, your first step should be to look at the
+daemon's log file to see if it logged an error explaining why it aborted the
+transfer.  Also double-check to ensure that the log file is setup right, as a
+wrong "log file" setting in your rsyncd.conf file can also cause this problem.
+You could also halt the daemon and run it interactively using the
+<code>--no-detach</code> and <code>--msgs2stderr</code> options and look for
+errors while someone tries the rsync copy in another window.
+
+<p>As for the cause of the remote rsync going away, there are several
+common issues that people run into:
+
+<ul>
+
+<li>The destination disk is full (remember that you need at least the
+size of the largest file that needs to be updated available in free
+disk space for the transfer to succeed).
+
+<li>An idle connection caused a router or remote-shell server to close
+the connection.
+
+<li>A network error caused the connection to be dropped.
+
+<li>The remote rsync executable wasn't found.
+
+<li>Your remote-shell setup isn't working right or isn't "clean"
+(i.e. it is sending spurious text to rsync).
+
+</ul>
+
+<p>If you think the problem might be an idle connection getting closed, you
+might be able to work around the problem by using a <code>--timeout</code>
+option (newer rsyncs send keep-alive messages during lulls).  You can also
+configure ssh to send keep-alive messages when using Protocol 2 (look for
+KeepAlive, ServerAliveInterval, ClientAliveInterval, ServerAliveCountMax, and
+ClientAliveCountMax).  You can also avoid some lulls by switching from
+<code>--delete</code> (aka <code>--delete-before</code>) to <code>--del</code>
+(aka <code>--delete-during</code>).
+
+<p>If you can't figure out why the failure happened, there are steps
+you can take to debug the situation.  One way is to create a shell
+script on the remote system such as
+<a href="rsync-debug">this one named "rsync-debug"</a>.
+You would use the script like this:
+
+<blockquote><pre><code>
+rsync -av --rsync-path=/some/path/rsync-debug HOST:SOURCE DEST
+rsync -av --rsync-path=/some/path/rsync-debug SOURCE HOST:DEST
+</code></pre></blockquote>
+
+<p>This script enables core dumps and also logs all the OS system calls
+that lead up to the failure to a file in the /tmp dir.  You can use the
+resulting files to help figure out why the remote rsync failed.
+
+<p>If you are rsyncing directly to an rsync daemon (without using a
+remote-shell transport), the above script won't have
+any effect.  Instead, halt the current daemon and run a debug version
+with core-dumps enabled and (if desired) using a
+system-call tracing utility such as <i>strace</i>, <i>truss</i>, or
+<i>tusc</i>.  For strace, you would do it like this (the -f option
+tells strace to follow the child processes too):
+
+<blockquote><pre><code>
+ulimit -c unlimited
+strace -f -t -s 1024 -o /tmp/rsync-$$.out rsync --daemon --no-detach
+</code></pre></blockquote>
+
+<p>Then, use a separate window to actually run the failing transfer, after
+which you can kill the debug rsync daemon (pressing Ctrl-C should do it).
+
+<p>If you are using rsync under inetd, I'd suggest temporarily disabling
+that and using the above daemon approach to debug what is going on.
+
+<li><p><b>Q:</b>
+
+Why does my connection to an rsync daemon (using the "::" syntax)
+fail immediately with an error like the following?
+
+<blockquote><pre><code>
+rsync: connection unexpectedly closed (24 bytes read so far)
+rsync error: error in rsync protocol data stream (code 12) at io.c(342)
+</code></pre></blockquote>
+
+<p><b>A:</b>
+
+Older rsync daemons (before 2.6.3) were unable to return errors that were
+generated during the option-parsing phase of the transfer.  Look in the
+logfile on the server to see if an error was reported, such as a "refused"
+option, an option that the server rsync doesn't support (e.g.  perhaps
+links are not supported by the server), or some other failure (such as
+trying to send data to a read-only module).  Upgrading the version of rsync
+that is running as a daemon to at least 2.6.3 will allow these errors to
+get returned to all rsync clients, old or new alike.
+
+</ol>
+
+<!--#include virtual="footer.html" -->

rsync-web/lists.html

@@ -0,0 +1,74 @@
+<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
+<HTML>
+<HEAD>
+<TITLE>rsync mailing lists</TITLE>
+</HEAD>
+<!--#include virtual="header.html" -->
+
+<H2 align="center">rsync mailing lists</H2>
+
+    <p>
+      There are three mailing lists for rsync.  All are optionally
+      available in digest mode and also through web archives.
+    </p>
+
+    <p>
+      Since these lists can generate a lot of traffic we suggest that
+      you should not subscribe from a web-mail account such as Yahoo!
+      or Hotmail, because your mailbox is likely to overflow.
+      Instead, please read directly from the archives.
+    </p>
+
+    <ul>
+      <li>
+	<b>rsync</b> is the main mailing lists for developers and
+	users.  It sees between zero and twenty messages per day.
+	This is a good place to send questions about rsync, but please
+	<a
+	href="https://www.catb.org/~esr/faqs/smart-questions.html">read
+	this before posting</a>.
+
+	[<a href="https://www.mail-archive.com/rsync@lists.samba.org/">archive 1</a>,
+	<a href="https://lists.samba.org/archive/rsync/">archive 2</a>,
+	<a href="https://lists.samba.org/mailman/listinfo/rsync">subscriptions</a>]
+      </li>
+
+      
+      <li>
+	<b>rsync-announce</b> carries only messages from the
+	maintainer annoucing new releases, which happen at most a few
+	times per month.
+
+	[<a href="https://www.mail-archive.com/rsync-announce@lists.samba.org/">archive 1</a>,
+	<a href="https://lists.samba.org/archive/rsync-announce/">archive 2</a>,
+	<a href="https://lists.samba.org/mailman/listinfo/rsync-announce">subscriptions</a>]
+      </li>
+
+      
+      <li>
+	<b>rsync-cvs</b> carries messages are automatically generated
+	whenever a developer changes the source code, which can happen
+	many times per day.  If you subscribe, you should probably
+	choose digest mode.
+
+	[<a href="https://www.mail-archive.com/rsync-cvs@lists.samba.org/maillist.html">archive 1</a>,
+	<a href="https://lists.samba.org/archive/rsync-cvs/">archive 2</a>,
+	<a href="https://lists.samba.org/mailman/listinfo/rsync-cvs">subscriptions</a>]
+      </li>
+    </ul>
+
+    <p>
+      Please report problems with the lists to the <tt>postmaster</tt>
+      at <tt>samba.org</tt>, but note that you can control your own
+      subscription using the web interface.
+    </p>
+
+    <H2 align="center">rsync Discord server</H2>
+
+    <p>
+      If you prefer real-time chat, there is also an rsync
+      <a href="https://discord.gg/Avfvy9zhdp">Discord server</a> for
+      discussion about rsync and its development.
+    </p>
+
+<!--#include virtual="footer.html" -->

rsync-web/nt.html

@@ -0,0 +1,85 @@
+<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
+<HTML>
+<HEAD>
+<TITLE>rsync on NT</TITLE>
+</HEAD>
+<!--#include virtual="header.html" -->
+
+<H2 align="center">rsync on NT</H2>
+
+<pre>
+From:   "Mike McHenry" <mmchen@minn.net>
+Subject: Rsync 2.3.1 WinNT binaries and instructions available
+Date:   Fri, 15 Oct 1999 02:53:30 +1000
+
+Hello all,
+
+I have created Windows NT binaries for rsync 2.3.1 and have decided to make
+them available for others to download. These binaries have been tested on
+Windows NT Server 4.0 SP5 and WILL run in daemon mode if you follow my
+instructions below. I make no guarantees about these binaries, they have
+however been working for me for weeks on several NT machines.
+
+Binaries at ftp://ftp.minn.net/usr/mmchen/
+
+Instructions for running in daemon mode:
+
+1. You will need two files, rsync.exe and cygwin1.dll. Place rsync.exe
+anywhere you like (I chose c:\program files\rsync\rsync.exe) and put
+cygwin1.dll in c:\winnt\system32
+
+2. You will need a program from the NT Server resource kit called
+srvany.exe. This program allows you to run any executable as a service. If
+you simply install the entire service pack it will be located in c:\ntreskit
+
+3. Create a service for rsync by typing the following:
+  instsrv Rsync "C:\ntreskit\srvany.exe"
+
+4. You should now have a new service called Rsync and you can verify by
+looking in Start->Control Panel->Services DON'T START IT YET!
+
+5. If you want to run rsync in daemon mode you will need a configuration
+file. Here is the one I use, call it rsyncd.conf and place it in the same
+directory as rsync (C:\Program files\rsync\rsyncd.conf)
+  use chroot = false
+  strict modes = false
+  hosts allow = *
+
+  [backup]
+     path = /
+     read only = yes
+     list = no
+
+This example configuration will make one big anonymous anonymous rsync area
+available, I use this to backup my NT machines from a central Unix machine.
+This configuration might not be ideal for you, change to suit your tastes.
+The first two lines are important for rsync to work on Windows NT however.
+
+6. You are going to need to hack some keys in the registry to make it work.
+Don't do this unless you are comfortable with the changes! Run regedit32 and
+add the following keys and values (quotation marks ARE IMPORTANT):
+  HKEY_LOCAL_MACHINE->SYSTEM->CurrentControlSet->Services->Rsync
+    Edit->Add Key->  Key Name: Paramaters
+    Edit->Add Value->  Value Name: AppDirectory  Value: "C:\programfiles\rsync"
+    Edit->Add Value->  Value Name: Application  Value: "C:\programfiles\rsync\rsync.exe"
+    Edit->Add Value->  Value Name: AppParamters  Value: --config="C:\programfiles\rsync\rsyncd.conf" --daemon
+
+7. That's it, you should be able to start and stop the rsync service at will
+using the Services Control Panel. When running with the above configuration
+you should be able to test by attempting to telnet to port 873 from a remote
+machine.
+  telnet rsync.server.com 873 (replacing rsync.server.com with your own
+server's address)
+You should get a connection to the rsync daemon running on your NT box.
+
+8. If you have problems you are on your own, sorry, I have enough to do :) I
+would suggest triple-checking your spelling on EVERYTHING (filenames,
+configs, reg keys). If you have any comments or suggestions I would be happy
+to hear them at mmchen@minn.net.
+
+Mike McHenry
+ Systems Administrator
+ MinnNet Communications, Inc.
+</pre>
+
+<!--#include virtual="footer.html" -->

rsync-web/nt.txt

@@ -0,0 +1,74 @@
+From:   "Mike McHenry" <mmchen@minn.net>
+Subject: Rsync 2.3.1 WinNT binaries and instructions available
+Date:   Fri, 15 Oct 1999 02:53:30 +1000
+
+
+Hello all,
+
+I have created Windows NT binaries for rsync 2.3.1 and have decided to make
+them available for others to download. These binaries have been tested on
+Windows NT Server 4.0 SP5 and WILL run in daemon mode if you follow my
+instructions below. I make no guarantees about these binaries, they have
+however been working for me for weeks on several NT machines.
+
+Binaries at ftp://ftp.minn.net/usr/mmchen/
+
+Instructions for running in daemon mode:
+
+1. You will need two files, rsync.exe and cygwin1.dll. Place rsync.exe
+anywhere you like (I chose c:\program files\rsync\rsync.exe) and put
+cygwin1.dll in c:\winnt\system32
+
+2. You will need a program from the NT Server resource kit called
+srvany.exe. This program allows you to run any executable as a service. If
+you simply install the entire service pack it will be located in c:\ntreskit
+
+3. Create a service for rsync by typing the following:
+  instsrv Rsync "C:\ntreskit\srvany.exe"
+
+4. You should now have a new service called Rsync and you can verify by
+looking in Start->Control Panel->Services DON'T START IT YET!
+
+5. If you want to run rsync in daemon mode you will need a configuration
+file. Here is the one I use, call it rsyncd.conf and place it in the same
+directory as rsync (C:\Program files\rsync\rsyncd.conf)
+  use chroot = false
+  strict modes = false
+  hosts allow = *
+
+  [backup]
+     path = /
+     read only = yes
+     list = no
+
+This example configuration will make one big anonymous anonymous rsync area
+available, I use this to backup my NT machines from a central Unix machine.
+This configuration might not be ideal for you, change to suit your tastes.
+The first two lines are important for rsync to work on Windows NT however.
+
+6. You are going to need to hack some keys in the registry to make it work.
+Don't do this unless you are comfortable with the changes! Run regedit32 and
+add the following keys and values (quotation marks ARE IMPORTANT):
+  HKEY_LOCAL_MACHINE->SYSTEM->CurrentControlSet->Services->Rsync
+    Edit->Add Key->  Key Name: Paramaters
+    Edit->Add Value->  Value Name: AppDirectory  Value: "C:\programfiles\rsync"
+    Edit->Add Value->  Value Name: Application  Value: "C:\programfiles\rsync\rsync.exe"
+    Edit->Add Value->  Value Name: AppParamters  Value: --config="C:\programfiles\rsync\rsyncd.conf" --daemon
+
+7. That's it, you should be able to start and stop the rsync service at will
+using the Services Control Panel. When running with the above configuration
+you should be able to test by attempting to telnet to port 873 from a remote
+machine.
+  telnet rsync.server.com 873 (replacing rsync.server.com with your own
+server's address)
+You should get a connection to the rsync daemon running on your NT box.
+
+8. If you have problems you are on your own, sorry, I have enough to do :) I
+would suggest triple-checking your spelling on EVERYTHING (filenames,
+configs, reg keys). If you have any comments or suggestions I would be happy
+to hear them at mmchen@minn.net.
+
+Mike McHenry
+ Systems Administrator
+ MinnNet Communications, Inc.
+

rsync-web/resources.html

@@ -0,0 +1,127 @@
+<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
+<HTML>
+<HEAD>
+<TITLE>rsync resources</TITLE>
+</HEAD>
+<!--#include virtual="header.html" -->
+
+<H2 align="center">rsync resources</H2>
+
+Please <a href="lists.html">let us know</a> if you have any rsync-related
+documents to add to this list:
+
+<ul> 
+
+<li>Be sure to search for the latest rsync info to get up-to-the-minute
+results.  You can use the search box at the top of the page for either
+web searching or project searching (they are done via Google).
+
+<li>
+2002-05-15: rsync is not official GNU software, but we try to
+work more or less in accordance with their <a
+href="http://www.gnu.org/prep/maintain_toc.html">Guidelines for
+Maintaining GNU Software</a>.
+
+<li> 2002-04-10: A new tutorial on using rsync to create a system of <a
+href="http://www.mikerubel.org/computers/rsync_snapshots/">rotating
+backups</a>, by Mike Rubel.
+
+<li>If you still don't know what rsync is, then take a look at the
+<A HREF="https://download.samba.org/pub/rsync/README">README</A>.
+
+<li>There is now a python script that implements
+<a href="https://download.samba.org/pub/unpacked/rsync/support/atomic-rsync">an
+atomic update</a> of the received files at the end of the transfer (when pulling).
+
+<li> Brian Elliott Finley has put together a great Linux install system based
+on rsync. You you read about it at <a href="http://thefinleys.com/SystemImager/">http://thefinleys.com/SystemImager/</a>
+
+<li><a href="http://www.dirvish.com/">Dirvish</a> is a fast, disk based,
+rotating network backup system that was originally written by JW Schultz.
+
+<li><a href="http://backuppc.sourceforge.net/">BackupPC</a>: a backup
+system using rsync.  Hard-links all identical files (even between multiple
+runs and multiple backup sources), compresses the files, provides an easy
+interface to find and restore files, etc.
+
+<li><a href="https://github.com/CharlesMAtkinson/bung">Bung</a>: (BackUp Next
+Generation) backs up files, MariaDB, OpenLDAP, postgres, etc. via an extensible
+templates system with git support.  The rsync-based "rolling full" backup is
+easy to browse and restore from using everyday tools.
+
+<li><a href="http://hacks.dlux.hu/drsync/">drsync</a>: a wrapper for rsync
+that remembers file sets between invocations so that a 2-way synchronization
+of two systems is possible.
+
+<li><a href="http://rsyncbackup.erlang.no/">rsyncbackup</a>: a helper
+script that uses config files to setup multiple backup scenarios and
+invokes rsync (or rsyncX on macOS).
+
+<li>Users who use the new character-set conversion option of rsync (--iconv)
+may want to check into the <a href="http://www.j3e.de/linux/convmv/man/">convmv</a>
+package that lets you convert the names of already-transferred files into a
+new characterset (for when you want to change or normalize the characterset
+of a hierarchy of files).
+
+<li>For those wanting to use launchd to run an rsync daemon (e.g. Mac
+OS X Tiger users), Glen Scott provides the necessary
+<a href="http://www.designsolution.co.uk/resources/rsync/">rsync.plist</a>
+file.
+
+<li>For the developer wanting to work on a branched rsync version based on
+one of the diffs in the patches dir, you may want to check into Matt's
+<a href="http://www.kepreon.com/~matt/myrsync/index.html#patchsync">patchsync</a>
+script.
+
+<!--#include virtual="doc-resources.html" -->
+
+<li>There are a few choices for making rsync work with OS X's resource forks.
+One is the official apple patch found on their opendarwin site, such as
+<a href="http://darwinsource.opendarwin.org/10.4/rsync-20">this one</a>
+(I've heard patch inefficiently transfers the entire resource fork information
+for every file on every transfer.)  Another choice is to use a third-party
+adapted rsync, such as
+<a href="http://archive.macosxlabs.org/rsyncx/rsyncx.html">rsyncx</a> or a
+<a href="http://www.quesera.com/reynhout/misc/rsync+hfsmode">rsync+hfsmode
+patch</a> by D Andrew Reynhout.  For the future, I would like to see an rsync
+that supports ACLs and Posix xattrs adapted to interact with resource forks in
+a seamless way (if that's possible).
+
+<li>Piero Orsoni wrote a GTK-based GUI for rsync called
+<a href="http://www.opbyte.it/grsync/">grsync</a>.
+
+<li>Those interested in using an rsync daemon over SSL may be interested in
+<a href="http://dozzie.jarowit.net/trac/wiki/RsyncSSL">this wiki page</a>
+that outlines a way to use a modern, simplified stunnel setup.
+
+<li>Thomas Roessler has written an rsync wrapper for
+<a href="ftp://riemann.iam.uni-bonn.de/pub/users/roessler/cvslock/">efficient,
+safe CVS mirroring</a>.
+
+<li>Rsync is distributed with the
+<a href="https://download.samba.org/pub/unpacked/rsync/support/rrsync">rrsync python script</a>
+that lets you restrict the rsync commands that can be run via ssh.  (This is
+a enhanced and reworked version of Joe Smith's original perl version.)
+
+<li><a href="mailto:LEakin@Nostrum.COM">Lee Eakin</a> has written a <a href="rsync_wrapper.pl">perl wrapper for rsync</a>.
+
+<li>A wire-compatible <a href="http://search.cpan.org/~cbarratt/">rsync implementation in perl</a>.
+
+<li>A <a href="http://www.srehttp.org/apps/rxrsync/">REXX implementation of rsync</a>.
+
+<li>An initial version of a <a href="http://www.kolosy.com/wordpress/?p=8">rewrite of rsync for .Net</a>.
+
+<li>You might want to check out an encryption program that is being developed
+to produce more rsync-friendly output:
+<a href="http://rsyncrypto.lingnu.com/">rsyncrypto</a>.
+
+<li>If you need a 2-way synchronization because both ends of the transfer may
+be changing files, you may want to either look into a tool designed to do this
+(e.g. <a href="http://freshmeat.net/projects/unison/">unison</a>), or you may
+wish to use an external wrapper for rsync that keeps extra data about what was
+in the last transfer so that it can figure out if a file is new or deleted
+(e.g. <a href="http://freshmeat.net/projects/drsync/">drsync</a>).
+
+</ul>
+
+<!--#include virtual="footer.html" -->

rsync-web/rsync-and-debian/.cvsignore

@@ -0,0 +1,2 @@
+rsync-and-debian.html
+rsync-and-debian.ps

rsync-web/rsync-and-debian/rsync-and-debian.sgml

@@ -0,0 +1,869 @@
+<!doctype linuxdoc system>
+<linuxdoc>
+  <article>
+    <titlepag>
+      <title>About integration of rsync and Debian</title>
+      <author>
+	<name>Martin Pool <tt>mbp@samba.org</tt></name>
+      </author>
+      <date>$Date: 2002/05/14 03:10:09 $</date>
+    </titlepag>
+    
+    <toc>
+    
+    <sect>
+      <heading>Introduction</heading>
+
+      <p>
+	It seems like there is an rsync thread on <tt/debian-devel/
+	every month or two.  Rather than going round in circles yet
+	again, I though I would summarise the issues in one place.
+      </p>
+
+      <p>
+	By way of background, I have been the maintainer of rsync for
+	about a year, and I wrote the <tt/librsync/ and <tt/rproxy/
+	packages.  Incidentally, I run Debian on most of the machines
+	I use, so I do care about the two projects working together
+	well.
+      </p>
+
+      <p>
+	I have tried to respond to all the issues raised on the list
+	recently, but possibly have missed some.  After putting many
+	hours into rsync I suppose I am quite fond of it and may be a
+	little biased, but I think I also know its strengths and
+	problems more than most.
+      
+      <p>
+	If there are issues ommitted by this document, or if you think
+	the answers are incomplete, unbalanced, or incorrect, then
+	please mail <tt/mbp@samba.org/, <tt/rsync@lists.samba.org/
+	and/or <tt/debian-devel@lists.debian.org/.
+
+    </sect>
+
+
+    <sect>
+      <heading>Background</heading>
+
+      
+      <sect1>
+	<heading>The rsync algorithm</heading>
+
+	<p>
+	  rsync is really two independent but related ideas: the
+	  <em/rsync algorithm/ for delta compression, and its
+	  implementation in a mirroring program.
+	</p>
+	
+	<p>
+	  The rsync algorithm is described in detail in the ANU
+	  Technical Report included with the distribution and on the
+	  <htmlurl url="http://rsync.samba.org/" name="rsync.samba.org"> web site.
+	  </p>
+
+	<p>
+	  Briefly, the rsync algorithm provides an efficient means to
+	  transfer a file <em/A/ from a source machine to a
+	  destination machine, when a similar file <em/A'/ is already
+	  present on the destination.  The destination machine sends a
+	  checksum of each consecutive block (of say 1kB) from <em/A'/
+	  to the source machine.  The source machine searches through
+	  <em/A/ for matching blocks.  Whatever does not match must be
+	  different, and a description of these differences is sent to
+	  the destination.
+	</p>
+
+	<p>
+	  The algorithm could be embodied in other forms and uses than
+	  the rsync program.  Indeed, the algorithm has been
+	  reimplemented in programs other than rsync itself: in other
+	  languages (Rexx, CAML, Java), in a C library (<tt/librsync/)
+	  and a derivative in xdelta.  xdelta is a specialization of
+	  the algorithm for the case where the two files are local and
+	  can be directly compared to compute a minimal binary delta.
+	</p>
+
+	<p>
+	  rsync deltas are much more efficient than diffs, for several
+	  reasons: most importantly, useful diff formats include
+	  context lines, which are wasted information.  rsync gets the
+	  same benefit of making sure that the change is applied to a
+	  compatible version of the file, but in a much more efficient
+	  way.  diffs suffer from being human-readable and therefore
+	  verbose, and cannot handle binary files.  Both can be
+	  transparently compressed.  diffs are an invaluable tool for
+	  software development, but not ideal for distribution of
+	  deltas.
+	</p>
+      </sect1>
+
+
+      
+      <sect1>
+	<heading/The rsync program/
+
+	<p>
+	  The rsync mirroring program is similar in functionality to
+	  other tools such as <tt/wget/, <tt/ftpmirror/, <tt/cvsup/
+	  and <tt/rdist/.  It has a number of functions that are
+	  useful in association with file mirroring, such as bandwidth
+	  limiting, access control, recursive mirroring, and selection
+	  of files to copy in various ways.
+	</p>
+
+	<p>
+	  In general, rsync is substantially faster than other
+	  tools, because it uses delta compression, and because the
+	  protocol is designed to be efficient both in traffic and
+	  round trips.  rsync can use <tt/zlib/ to compress traffic
+	  (and introduce double-<tt/free()/ security holes :-), at the
+	  option of the client.  Compression may also be disabled by
+	  the server administrator.
+	</p>
+
+	<p>
+	  rsync has the important property that it is generally
+	  idempotent: repeated runs of rsync, even if interrupted,
+	  make the contents of the destination machine converge
+	  towards those of the source machine.
+
+	<p>
+	  Unlike wget and ftpmirror, rsync must be installed on both
+	  the client and the server.  It uses its own protocol, rather
+	  than HTTP or FTP.  This is probably the biggest practical
+	  drawback at the moment: one cannot use rsync from arbitrary
+	  web sites, but only from systems  that have specifically
+	  chosen to support it.  On the other hand, many important
+	  sites for free software do now support rsync, and the
+	  developers see them as an important constituency to support.
+
+	<p>
+	  rsync can tunnel through a proxy server that supports
+	  <tt/HTTP CONNECT/, a SOCKS server (using <tt/socksify/), an
+	  ssh tunnel, or various other methods.
+
+	<p>
+	  rsync can run as a daemon, similar to an <tt/ftpd/, in which
+	  it is controlled by the <tt>/etc/rsyncd.conf</tt> file.  The
+	  archetypal use is to offer public, anonymous, read-only
+	  access to a software archive but as with <tt/ftpd/ other
+	  configurations are possible and common.
+	</p>
+
+	<p>
+	  rsync can also run across a tunnel program such as <tt/ssh/,
+	  in which case it is very similar to <tt/scp/.  This mode is
+	  commonly used for making network backups, or uploading
+	  information to a web server.
+	</p>
+	
+	<p>
+	  rsync may also be used locally, which is a degenerate case
+	  with client and server connected across a unix-domain or
+	  localhost socket.
+	</>
+
+	<p>
+	  rsync runs on many varieties of Unix under both gcc and
+	  native compilers, and also under Cygwin on Microsoft
+	  platforms.  There is somebody working on a VMS port.
+	</p>
+      </sect1>
+
+
+      
+
+      <sect1>
+	<heading/librsync/
+
+	<p>
+	  <tt/librsync/ is a library that reimplements the rsync
+	  algorithm in a very flexible way, with the goal of allowing
+	  any reasonable mode of operation, and integration with any
+	  existing program.  rsync does not currently link to
+	  librsync, but it might do so in the long term.
+
+	<p>
+	  <tt/librsync/ currently uses an encoding
+	  format different to that of rsync 2.5, mostly because I did
+	  not want to be constrained by historical code.
+
+	<p>
+	  <tt/librsync/ is at version 0.9.5 and roughly as stable as
+	  the version number suggests: it is used by Intermezzo and
+	  some other projects, but is not yet really mature.  The
+	  <tt/librsync/ distribution comes with a tool <tt/rdiff/ that
+	  exposes the functionality as a Unix tool to shell scripts,
+	  etc.  <tt/librsync/ is LGPL'd.
+
+	<p>
+	  For example, you can imagine the server calculating and
+	  caching the checksums, or the client calculating the delta
+	  once and then sending it over UDP multicast to several
+	  destinations, or uuencoded in email.  Neither of these would
+	  be straightforward to implement in the rsync codebase, but
+	  all can be done with rdiff.
+
+      </sect1>
+
+
+      <sect1>
+	<heading/rproxy/
+
+	<p>
+	  <tt/librsync/ is used in the <url
+	  url="http://rproxy.samba.org/" name="rproxy"> program, which
+	  is a prototype of transparent integration of delta
+	  compression into HTTP.
+	</p>
+
+	<p>
+	  rproxy implements delta compression of arbitrary HTTP
+	  content, whether dynamic or static.  For example, rproxy
+	  gives very good compression on repeated visits to portal
+	  sites such as Slashdot, since it can transmit only the
+	  sections of the page modified since last time it was
+	  viewed.  Regular HTTP caches, by contrast, must either hit
+	  on the whole page, or reload the whole page, and therefore
+	  do poorly on template-based dynamic content.
+
+	<p>
+	  rproxy adds compatible extensions to HTTP headers, and can
+	  pass through HTTP caches.  It requires upstream and
+	  downstream support, which at the moment means two
+	  installations of rproxy, but in the future could conceivably
+	  be integrated into Mozilla, Squid, Apache and similar
+	  programs.  
+
+	<p>
+	  rproxy is packaged in Debian and is moderately useful for
+	  people with slow links.
+
+	<p>
+	  rproxy is not HTML-specific, but HTML is by far the most
+	  common case of dynamic HTTP content that can be
+	  delta-compressed.  However, HTML documents tend to be fairly
+	  small, and as connectivity improves they're becoming of
+	  decreasing interest as a compression problem.
+
+	<p>
+	  (My personal interest in this project declined significantly
+	  when I went from a 56k6 modem to ADSL at home.  I realize
+	  many people in the world still have slow connections.)
+
+	<p>
+	  There are some Internet Drafts from Jeffrey Mogul and others
+	  that add similar delta compression based on the server
+	  storing all possible deltas.  Last time I looked, there did
+	  not seem much interest in wide adoption of these proposals.
+
+	<p>
+	  Documenting a protocol extension to the standard expected by
+	  an RFC can be a lot of work.  A beginning on this work has
+	  been made for rproxy, but more experiments with the
+	  implementation are needed before proposal as a standard.
+
+	<p>
+	  rproxy is not being actively developed at the moment.
+	  Obviously I cannot answer why every programmer in the world
+	  does not work on this.  Personally I think that developing
+	  rsync itself, and then librsync/rdiff, is likely to be more
+	  useful; I suspect other people interested in working in this
+	  area might have similar thoughts.  
+
+	<p>
+	  I don't think there are any problems in the code or project
+	  that would actively prevent anybody from working on it.  I'd
+	  be happy to hand over maintenance to somebody else.  Ben
+	  Elliston has expressed interest in looking after it in the
+	  last couple of months, and possibly it will be more active
+	  in the future.
+      </sect1>
+
+      
+
+      <sect1>
+	<heading>Introduction to Debian</heading>
+
+	<p>
+	  Debian is a free operating system developed by the
+	  cooperation of people all over the world.  The most
+	  important relation of rsync to Debian is copying of software
+	  packages from developers to distribution servers to end
+	  users.
+	</p>
+
+	<p>
+	  Many Debian users get their software by internet download
+	  from a public server, rather than pre-installed on a machine
+	  or on CD-ROM.
+	</p>
+
+	<p>
+	  Debian software archives include both software source,
+	  binary packages for various architectures, and index
+	  metadata, primarily the <tt/Packages/ files.
+	</p>
+
+	<p>
+	  Because of Debian's community development process, new
+	  packages are released very often.  On a typical day in
+	  Debian's <tt/unstable/ release, there might be fifty new
+	  uploads.
+	</p>
+
+	<p>
+	  Debian is quite different from other software distributions
+	  in shipping so many packages so frequently.  The BSDs (as I
+	  understand it) base most of their development out of a CVS
+	  or CVSup tree, which inherently distributes coarse-grained
+	  deltas.  Proprietary systems make binary releases, but much
+	  less frequently.  Possibly the development branches of other
+	  distributions, such as Redhat "Rawhide" and Mandrake
+	  "Cooker" are similar.
+	  </p>
+      </sect1>
+
+
+
+
+      <sect1>
+	<heading>apt-proxy</heading> 
+      <p><label id="apt-proxy">
+	<url url="http://apt-proxy.sourceforge.net/" name="apt-proxy">	is a caching proxy for Debian archives.  It appears as an HTTP
+	server to apt clients, and uses rsync, http, or ftp to connect
+	to an upstream server.  
+
+      <p>
+	Because rsync is less efficient than HTTP for transferring
+	compressed files, <tt/apt-proxy/ can selectively use rsync for
+	uncompressed Packages and files and HTTP or FTP for <tt/.deb/
+	files.
+
+      <p>
+	<tt/apt-proxy/, unlike Squid, has domain knowledge about
+	Debian archives and can therefore perform functions such as
+	purging old packages from the cache.
+
+      <p>
+	The original apt-proxy was written by Paul 'Rusty' Russel.
+	The current maintainer is Chris Halls.
+      </p>
+    </sect1>
+  </sect>      
+
+    
+
+    <sect>
+      <heading>Open Issues</heading>
+
+
+      <sect1>
+	<heading/Compressed files cannot be differenced/
+
+	<p>
+	  <tt/gzip/, like most compression algorithms has the property
+	  that a change in the source file at one point will cause
+	  cascading changes in the output file from that point
+	  onwards, and therefore make delta compression more or less
+	  useless.
+	</p>
+
+	<p>
+	  Although delta compression is not possible, rsync is still a
+	  very useful tool for mirroring compressed files.  The
+	  efficient network protocol mean that it will generally be
+	  slightly faster and use less traffic than HTTP or FTP.
+
+	<p>
+	  There is a patch called <tt/--rsyncable/ for gzip that fixes
+	  this behaviour: gzip files are basically broken up into
+	  blocks so that changes (including insertion or deletion) in
+	  the input file affect only the corresponding blocks in the
+	  output file.  (The blocks are not of fixed size, but rather
+	  delimited by marker patterns at which a checksum hits a
+	  particular value, so they move as data is inserted or
+	  removed.)
+
+	<p>
+	  I believe that this will merge into the upstream <tt/zlib/
+	  soon and be on by default, at which point <tt/.deb/ files
+	  will delta-compress well.  This patch does seem to be
+	  languishing, though, and it would be good to either get it
+	  into the upstream, or into Debian's own version.  Needless
+	  to say it must be extremely thoroughly tested.
+
+	<p>
+	  The scheme for containing changes is not specific to rsync,
+	  and might also be useful with <tt/xdelta/, <tt/rdiff/, or
+	  some other binary delta scheme invented in the future.  It
+	  also does not require a copy of the old file when
+	  compressing the new one.
+
+	<p>
+	  This scheme relies on the output file being determined only
+	  by the input file.  As far as I know compression schemes
+	  like gzip, lzw, and bzip2 are deterministic in this way.
+	  (GnuPG, for example, is not, since it uses a random session
+	  key.)
+
+	<p>
+	  Incidentally this would allow more efficient deltas between
+	  Debian ISO images.
+	</p>
+
+	<p>
+	  Alternatively, you could distribute an uncompressed package
+	  tree that would rsync efficiently, but since the gzip patch
+	  should merge soon this seems unnecessary.
+
+	<p>
+	  There is also the interesting possibility of using
+	  <tt/dpkg-repack/ to generate something similar to the
+	  previous <tt/.deb/ and then use it as the basis of an rsync
+	  download.
+
+	<p>
+	  It could be possible to make an equivalent patch to
+	  <tt/bzip2/, but possibly the large block size would cause
+	  trouble.
+
+	<p>
+	  A patch to gzip which implements this behaviour is available
+	  from rsync CVS.
+      </sect1>
+
+
+
+      <sect1>
+	<heading/rsync is too hard on servers/
+
+	<p>
+	  If it is, then I think we should fix the problems, rather
+	  than invent a new system from scratch.  I think the
+	  scalability problems are accidents of the current codebase,
+	  rather than anything inherent in the design.
+	</p>
+      </sect1>
+
+
+      
+      <sect1>
+	<heading/We should throw out the current code and start from
+	scratch/
+
+	<p>
+	  Some projects (Apache 2.0, Mozilla, ...) choose to do this;
+	  some concentrate on piecemeal improvement (Linux).
+
+	<p>
+	  As Joel Spolsky recently observed, throwing the code out to
+	  start from scratch always seems like a nice idea, but rarely
+	  works out.  Essentially you are opting for the devil you
+	  don't know, but he definitely exists.
+
+	<p>
+	  Having dealt with some fairly crufty old code I sometimes
+	  feel like doing this in rsync, but I can also see the
+	  arguments against it.  Starting from a new codebase would
+	  bring in plenty of new bugs (including security bugs),
+	  orphan existing users, and halt progress until it caught up
+	  to the same level.  It would require carefully documenting
+	  all the existing internal protocols and behaviours, which is
+	  a nice idea but not clearly a good use of the developers'
+	  time.
+
+	<p>
+	  If you throw out the code, you have the questions of whether
+	  or not to keep the current network protocol, and whether or
+	  not to keep the current command-line option.
+
+	<p>
+	  Not being able to talk to old remote versions would be a
+	  real pain, and would certainly slow adoption.  (Think of how
+	  unpopular SSH2 was when it wanted to be incompatible with
+	  SSH1, but use the same TCP port.)  On the other hand,
+	  keeping the same network protocol limits your ability to
+	  redesign things.
+
+	
+	<p>
+	  There is some cruft in the command line syntax, but throwing
+	  it out would also break everybody's scripts, mental maps,
+	  and documentation.
+
+
+	<p>
+	  Possibly rsync 3.0 will be largely rewritten.  <tt/librsync/
+	  is a start in that direction.  If you want to do this,
+	  <tt/librsync/ might help you, but please think about it
+	  before you begin hacking.
+
+      </sect1>
+
+      
+
+
+      <sect1>
+	<heading/rsync development roadmap/
+
+	<p>
+	  Debian's goals for rsync have to be balanced against the
+	  requirements of other users (including the mirror sites that
+	  distribute Debian) and the limited development resources.
+
+	<p>
+	  As of April 2002, the 2.5.5 release is current and 2.5.6 is
+	  pending.  These versions seem to work quite well and we
+	  encourage people to upgrade from stable 2.4.6 and previous
+	  versions.  It is very important to the developers to
+	  establish a stable base release before starting out on new
+	  development.
+	</p>
+
+	<p>
+	  A number of enhancements are planned for the 2.6 series.
+	  Some of them are discussed below, and more details can be
+	  found in the <tt/TODO/ file in the rsync distribution.
+
+	<p>
+	  If you want rsync to progress faster, the best thing you can
+	  do is find reproducible failure cases and report them well,
+	  or to help us write a regression test suite.
+
+	<p>
+	  Nobody is very actively working on rproxy or librsync as far
+	  as I know.  Personally at the moment I feel supporting rsync
+	  itself is more useful.
+      </sect1>
+
+
+
+      
+      <sect1>
+	<heading/Goswin Brederlow's proposal to use the reverse rsync
+	algorithm over HTTP Range requests/
+
+	<p>
+	  Goswin Brederlow has a nice proposal that allows rsync
+	  compression using a regular HTTP server.  It depends only
+	  upon rsyncable archives, and development of new client and
+	  support programs.
+
+	<p>
+	  The idea, as I understand it, is that the checksums for each
+	  package should be pre-calculated and stored on the HTTP
+	  server along with the files they describe.  (The checksum
+	  files should be on the order of a thousand times smaller
+	  than the archive file.)
+
+	<p>
+	  The client program will download the checksum file over HTTP
+	  if it exists.  The client can then search for common blocks
+	  in the local file, and therefore determine what new regions
+	  it must fetch from the server.  The client then sends an
+	  HTTP request using the optional <tt/Range/ header for the
+	  relevant sections of the new file, and patches them onto the
+	  old file.  The checksum file ought to also include a
+	  whole-file message digest which can be used to ensure that
+	  reassembly was successful.
+	
+	<p>
+	  This scheme has the great advantage that the server is
+	  entirely passive, and only needs to support standard HTTP.
+	  The checksum files can be generated once for each package
+	  either during upload, or by the administrator of a
+	  particular server.
+
+	<p>
+	  (This is not the same as rproxy, which uses a different
+	  encoding and requires upstream support, but can
+	  transparently compress any request without requiring
+	  signature files on the server.)
+
+	<p>
+	  This scheme could be fairly easily built on top of rdiff or
+	  librsync.
+
+	<p>
+	  I think this sounds like a very promising scheme.  I think
+	  rsync might still be better for people who want to copy
+	  large trees, but for users apt-getting single packages this
+	  would be a simple way to get delta-compression in, pending
+	  only --rsyncable files.
+
+
+	<p>
+	  I'm not sure if all HTTP servers currently handle <tt/Range/
+	  commands efficiently.  This proposal would probably stress
+	  them more than is common at the moment.
+
+	<p>
+	  Because it requires a special client, and special checksum
+	  files stored on the server it has a wider impact than just
+	  using rsync.  It ought to be done in a non-Debian-specific
+	  way.  Rather than adding knowledge about the deltas into
+	  apt-get itself, we ought to make a separate tool which
+	  downloads a delta over HTTP.
+
+      </sect1>
+
+
+
+
+       <sect1>
+	 <heading>rsync only compares files with the exact same
+	   name</heading>
+
+	 <p>
+	   Even for uncompressed files, rsync will not currently try to
+	   use <tt/linux-2.4.17.tar/ to do delta-compression against
+	   <tt/linux-2.4.18.tar/, because it cannot guess from the
+	   names that the two files are related.
+
+	 <p>
+	   Paul Russell has contributed a patch which adds a heuristic
+	   to detect this.  It will probably be merged in 2.6.
+	 </p>
+       </sect1>
+
+
+       <sect1>
+	 <heading/rsync uses too much memory/
+
+	 <p>
+	   rsync traverses the entire directory tree before it begins
+	   copying files.  On machines with little virtual memory and a
+	   lot of files to copy this can be a problem.
+
+	 <p>
+	   Some problems in this area have been fixed in the 2.5
+	   series.  Other solutions involving internal restructuring of
+	   the <tt/flist/ and <tt/hlink/ code will be attempted in 2.6,
+	   and they should yield a substantial improvement.
+
+	 <p>
+	   We are not likely to change the general approach of
+	   traversing the tree up-front in the near future, because it
+	   is tightly tied to the network protocol.  It might be
+	   attempted in the 3.0 timeframe, but it is not entirely clear
+	   that it is really a problem.
+       </sect1>
+
+
+
+       <sect1>
+	 <heading>rsync hangs?</heading>
+
+	 <p>
+	   There were a number of deadlock bugs in the past.  We
+	   believe they are all fixed in 2.5.5, but would welcome good
+	   bug reports to the contrary.  Some of them were in fact
+	   Linux and Solaris kernel bugs, but they've been fixed by the
+	   relevant parties quite some time ago.
+	 </p>
+       </sect1>
+
+
+
+       <sect1>
+	 <heading>Reduced network usage is not justified by increased CPU
+	   usage
+	 </heading>
+
+	 <p>
+	   The balance point will vary for each administrator.  It is
+	   hard to answer this universally.  The balance is very
+	   different in countries other than the US.
+	 </p>
+
+	 <p>
+	   I suspect CPU cycles are falling in price faster than
+	   network bandwidth, and for many people unused CPU cycles are
+	   wasted but network packets cost money.
+	 </p>
+       </sect1>
+
+
+
+       <sect1>
+	 <heading/rsync can't saturate a fast link/
+
+	 <p>
+	   rsync with <tt>--whole-file</tt> (to turn off the
+	   delta algorithm and be more comparable to ftp) floods a
+	   100Mbps link.  If you have something faster than that to
+	   your upstream Debian mirror, you're a lucky person.
+	 </p>
+
+	 <p>
+	   It is not inherently impossible to make rsync use
+	   <tt>sendfile()</tt> and <tt>mmap()</tt>; it's just that most
+	   people don't need them.  (Apache did not use them by default
+	   either last time I looked, but that doesn't mean it's not
+	   fast enough for most people.)
+	 </p>
+       </sect1>
+
+
+
+       <sect1>
+	 <heading>rsync is not actively developed</heading>
+
+	 <p>
+	   This was a pretty fair accusation a year ago, but I don't
+	   think it is true anymore.  We have made five releases this
+	   year, and the <tt/diff -u/ since 2.4.6, the last version
+	   from the previous maintainer, is 38305 lines.
+
+	 <p>
+	   Mail is answered promptly.  The web site has been revised
+	   and cleaned.  We're working on a regression test suite, code
+	   cleanup and documentation, and other quality issues.
+
+	 <p>
+	   I'm the most active developer at the moment, but there are a
+	   number of other people with experience in the code who
+	   regularly commit changes or send patches.
+
+	 <p>
+	   The stability expectations of our users require a somewhat
+	   disciplined approach to accepting patches, but I don't think
+	   that's a bad thing.  The goal is to iterate through
+	   <em/freeze/ and <em/flow/ phases similar to the kernel.
+
+	 <p>
+	   The Debian rsync package maintainer, Phil Hands, has been
+	   somewhat inactive, but apparently he's back now.  Colin
+	   Walters has been helping out.
+
+	 <p>
+	   If you have patches that were dropped, please send them
+	   through again.
+       </sect1>
+
+
+
+       <sect1>
+	 <heading>Servers should cache deltas</heading>
+
+	 <p>
+	   This also smells like premature optimization.  It might be
+	   useful, but it certainly seems less important than some
+	   other tasks.
+       </sect1>
+
+
+
+       <sect1>
+	 <heading>Possible patent on rsync?</heading>
+
+	 <p>
+	   It has been suggested that some uses of rsync may conflict
+	   with a US patent.  I am not sure of the current situation so
+	   I won't comment.  I don't know of any suggestion that the
+	   rsync program as it currently exists infringes any patent.
+	 </p>
+
+	 <p>
+	   I do know that rsync has been in use for many years by large
+	   and small organization with no complaints of patent
+	   infringement.
+	 </p>
+
+	 <p>
+	   I would point out that linked lists, mark-and-copy garbage
+	   collection, and the Tab key are all patented too.  Somebody
+	   who always carefully checked first for software patents
+	   would never write anything at all.
+	 </p>
+
+       </sect1>
+
+
+       <sect1>
+	 <heading>Debian should store more metadata outside of
+	 the package file</heading>
+
+	 <p>
+	   I can see some reasons why this might be a good idea, but
+	   it doesn't seem particularly relevant to rsync either way.
+	 </p>
+       </sect1>
+
+
+
+       <sect1>
+	 <heading>Debian should distribute diffs or xdeltas
+	   between different versions of their packages</heading>
+
+	 <p>
+	   A problem with this is that Debian releases packages very
+	   often, and so the number of delta files is large.  This will
+	   be a problem for mirror sites who cannot carry all of the
+	   delta files for history, although of course the client could
+	   fall back to directly fetching the new package if the
+	   relevant deltas are not present.
+	 </p>
+
+	 <p>
+	   Fetching a series of deltas which update the same area is
+	   less efficient than calculating the deltas directly.
+	 </p>
+
+	 <p>
+	   Although this idea does not seem very practical for people
+	   using <em/unstable/, it could work well for <em/stable/
+	   and other situations where there are relatively few
+	   releases, or for people without network access.  One could,
+	   for example, distribute a CD-ROM of xdeltas.  On the other
+	   hand, since CD-ROMs are relatively large compared to the
+	   distribution it might be simpler to just distribute the new
+	   packages directly as is currently done.
+      </sect1>
+      
+      <sect1>
+	<heading>rsync should be used by <tt/apt-get update/</heading>
+	
+	<p>
+	  Even if there are few gains from compressing <tt/.deb/s,
+	  rsync might help in providing a more efficient transfer of
+	  the <tt/Packages/ file, which has small but frequent
+	  changes.  This approach is used by <tt/apt-proxy/.
+      </sect1>
+    </sect>
+
+    
+    <appendix>
+    <sect>
+      <heading>Revision history</heading>
+
+      
+<p><tscreen><verb>
+$Log: rsync-and-debian.sgml,v $
+Revision 1.10  2002/05/14 03:10:09  mbp
+Add note about just transferring Package files.
+
+Revision 1.9  2002/05/14 02:35:43  mbp
+Update notes on apt-proxy based on mail from Chris Halls.
+
+Revision 1.8  2002/04/12 02:04:15  mbp
+More information about Goswin's idea, as I understand it.
+
+Revision 1.7  2002/04/12 01:39:52  mbp
+Lots more details about rproxy.  Thanks to Brian May for prompting me
+to do this.
+
+Revision 1.6  2002/04/12 00:43:06  mbp
+Fix revision history SGML stuff.
+
+Revision 1.5  2002/04/12 00:24:50  mbp
+Fix mailing list name.
+Add revision history.
+	
+	  </verb></tscreen></p>
+    </sect>
+  </article>
+</linuxdoc>

rsync-web/rsync-debug

@@ -0,0 +1,10 @@
+#!/bin/sh
+
+ulimit -c unlimited
+
+# Some systems have "truss" or "tusc" instead of "strace".
+# The -f option tells strace to follow children too.
+# The -t option asks for timestamps.
+# The -s 1024 option increases the string decoding limit per function call.
+# The -o option tells strace where to send its output.
+strace -f -t -s 1024 -o /tmp/rsync-$$.out rsync "${@}"

rsync-web/rsync_wrapper.pl

@@ -0,0 +1,471 @@
+#      __
+#     /\ \ From the mind of
+#    /  \ \
+#   / /\ \ \_____ Lee Eakin  <LEakin@Nostrum.COM>
+#  /  \ \ \______\       or  <Lee@Eakin.ORG>
+# / /\ \ \/____  /
+# \ \ \ \____\/ / Wrapper module for the rsync program
+#  \ \ \/____  /  rsync can be found at http://rsync.samba.org/rsync/
+#   \ \____\/ /
+#    \/______/
+
+package Rsync;
+require 5.004;
+
+use FileHandle;
+use IPC::Open3 qw(open3);
+use Carp;
+
+use strict;
+
+=head1 NAME
+
+Rsync - perl module interface to B<rsync> http://rsync.samba.org/rsync/
+
+=head1 SYNOPSIS
+
+use Rsync;
+
+$obj = Rsync->new(qw(-az C<-e> /usr/local/bin/ssh
+         --rsync-path /usr/local/bin/rsync));
+
+$obj->exec(qw(localdir rhost:remdir))
+         or warn "rsync failed\n";
+
+=head1 DESCRIPTION
+
+Perl Convenience wrapper for B<rsync> program.  Written for B<rsync> 2.3.1 but
+should perform properly with most versions.
+
+=cut
+
+# options from the rsync man pae
+###### Boolean flags ######
+# -h, --help                  show this help screen
+# -v, --verbose               increase verbosity
+# -q, --quiet                 decrease verbosity
+# -c, --checksum              always checksum
+# -a, --archive               archive mode
+# -r, --recursive             recurse into directories
+# -R, --relative              use relative path names
+# -b, --backup                make backups (default ~  suffix)
+# -u, --update                update only (don't overwrite newer files)
+# -l, --links                 preserve soft links
+# -L, --copy-links            treat soft links like regular files
+#     --copy-unsafe-links     copy links outside the source tree
+#     --safe-links            ignore links outside the destination tree
+# -H, --hard-links            preserve hard links
+# -p, --perms                 preserve permissions
+# -o, --owner                 preserve owner (root only)
+# -g, --group                 preserve group
+# -D, --devices               preserve devices (root only)
+# -t, --times                 preserve times
+# -S, --sparse                handle sparse files efficiently
+# -n, --dry-run               show what would have been transferred
+# -W, --whole-file            copy whole files, no incremental checks
+# -x, --one-file-system       don't cross filesystem boundaries
+# -C, --cvs-exclude           auto ignore files in the same way CVS does
+#                             RCS SCCS CVS CVS.adm RCSLOG cvslog.* tags TAGS
+#                             .make.state .nse_depinfo *~ #* .#* ,* *.old *.bak
+#                             *.BAK *.orig *.rej .del-* *.a *.o *.obj *.so *.Z
+#                             *.elc *.ln core
+#     --delete                delete files that don't exist on the sending side
+#     --delete-excluded       also delete excluded files on the receiving side
+#     --partial               keep partially transferred files
+#     --force                 force deletion of directories even if not empty
+#     --numeric-ids           don't map uid/gid values by user/group name
+# -I, --ignore-times          don't exclude files that match length and time
+#     --size-only             only use file size when determining if a file
+#                             should be transferred
+# -z, --compress              compress file data
+#     --version               print version number
+#     --daemon                run as a rsync daemon
+#     --stats                 give some file transfer stats
+#     --progress              show progress during transfer
+###### scalar values ######
+#     --csum-length=LENGTH    <=16 bit md4 checksum size
+# -B, --block-size=SIZE       checksum blocking size (default 700)
+#     --timeout=TIME          set IO timeout in seconds
+#     --port=PORT             specify alternate rsyncd port number
+# -e, --rsh=COMMAND           specify rsh replacement
+# -T, --temp-dir=DIR          create temporary files in directory DIR
+#     --compare-dest=DIR      also compare destination files relative to DIR
+#     --exclude-from=FILE     exclude patterns listed in FILE
+#     --include-from=FILE     don't exclude patterns listed in FILE
+#     --config=FILE           specify alternate rsyncd.conf file
+#     --password-file=FILE    get password from FILE
+#     --log-format=FORMAT     log file transfers using specified format
+#     --suffix=SUFFIX         override backup suffix
+#     --rsync-path=PATH       specify path to rsync on the remote machine
+###### array values ######
+#     --exclude=PATTERN       exclude files matching PATTERN
+#     --include=PATTERN       don't exclude files matching PATTERN
+#
+
+=over 4
+
+=item Rsync::new
+
+$obj = new Rsync;
+
+   or
+
+$obj = Rsync->new;
+
+   or
+
+$obj = Rsync->new(@options);
+
+=back
+
+Create an Rsync object.  Any options passed at creation are stored in
+the object as defaults for all future exec call on that object.  Options
+are the same as those in L<rsync> with the addition of
+--path-to-rsync which can be used to override the hardcoded default of
+/usr/local/bin/rsync, and --debug which causes the module functions to
+print some debugging information to STDERR.
+
+=cut
+
+sub new {
+   my $class=shift;
+
+   # seed the options hash, booleans, scalars, excludes, data,
+   # status, stderr/stdout storage for last exec
+   my $self={
+      # the full path name to the rsync binary
+      'path-to-rsync' => '/usr/local/bin/rsync',
+      # these are the boolean flags to rsync, all default off, including them
+      # in the args list turns them on
+      'flag' => {qw(
+         --archive         0   --backup          0   --checksum          0
+         --compress        0   --copy-links      0   --copy-unsafe-links 0
+         --cvs-exclude     0   --daemon          0   --delete            0
+         --delete-excluded 0   --devices         0   --dry-run           0
+         --force           0   --group           0   --hard-links        0
+         --help            0   --ignore-times    0   --links             0
+         --numeric-ids     0   --one-file-system 0   --owner             0
+         --partial         0   --perms           0   --progress          0
+         --quiet           0   --recursive       0   --relative          0
+         --safe-links      0   --size-only       0   --sparse            0
+         --stats           0   --times           0   --update            0
+         --verbose         0   --version         0   --whole-file        0
+      )},
+      # these have simple scalar args we cannot easily check
+      'scalar' => {qw(
+         --block-size      0   --compare-dest    0   --config            0
+         --csum-length     0   --exclude-from    0   --include-from      0
+         --log-format      0   --password-file   0   --port              0
+         --rsh             0   --rsync-path      0   --suffix            0
+         --temp-dir        0   --timeout         0
+      )},
+      # these can be specified multiple times and are additive, the doc also
+      # specifies that it is an ordered list so we must preserve that order
+      'exclude'     => [],
+      # source/destination path names and hostnames
+      'data'        => [],
+      # return status from last exec
+      'status'      => 0,
+      'realstatus'  => 0,
+      # whether or not to print debug statements
+      'debug'       => 0,
+      # stderr from last exec in array format (messages from remote rsync proc)
+      'err'         => [],
+      # stdout from last exec in array format (messages from local rsync proc)
+      'out'         => [],
+   };
+   if (@_) {
+      &defopts($self,@_) or return undef;
+   }
+   bless $self, $class;
+}
+
+=over 4
+
+=item Rsync::defopts
+
+defopts $obj @options;
+
+   or
+
+$obj->defopts(@options);
+
+=back
+
+Set default options for future exec calls for the object.  See L<rsync>
+for a complete list of valid options.  This is really the internal
+function that B<new> calls but you can use it too.  Presently there is no way
+to turn off the boolean options short of creating another object, but if it is
+needed and the B<rsync> guys don't use it, I may add hooks to let + and ++ or a
+leading no- toggle it back off similar to B<Getopt::Long> (the GNU way).
+
+=cut
+
+sub defopts {
+   my $self=shift;
+   my @opts=@_;
+
+   # need a conversion table in case someone uses the short options
+   my %short=(qw(
+      -B  --block-size   -C  --cvs-exclude     -D  --devices   -H  --hard-links
+      -I  --ignore-times -L  --copy-links      -R  --relative  -T  --temp-dir
+      -W  --whole-file   -a  --archive         -b  --backup    -c  --checksum
+      -e  --rsh          -g  --group           -h  --help      -l  --links
+      -n  --dry-run      -o  --owner           -p  --perms     -q  --quiet
+      -r  --recursive    -s  --sparse          -t  --times     -u  --update
+      -v  --verbose      -x  --one-file-system -z  --compress
+   ));
+   while (my $opt=shift @opts) {
+      my $arg;
+      print(STDERR "setting debug flag\n"),$self->{debug}=1,next
+         if $opt eq '--debug';
+      print STDERR "processing option: $opt\n" if $self->{debug};
+      if ($opt=~/^-/) {
+         # handle short opts first
+         if ($opt=~/^-(\w+)$/) {
+            foreach (split '',$1) {
+               print STDERR "short option: -$_\n" if $self->{debug};
+               if (exists $short{'-'.$_}) {
+                  $opt=$short{'-'.$_};
+                  # convert it to the long form
+                  $opt=$short{$opt} if exists $short{$opt};
+                  # handle the 3 short opts that require args
+                  $self->{scalar}{$opt}=shift(@opts),next if (/^[BeT]$/);
+                  # handle the rest
+                  $self->{flag}{$opt}=1,next if exists $self->{flag}{$opt};
+               }
+               carp "$opt - unknown option\n";
+               return undef;
+            }
+         }
+         # handle long opts with = args
+         if ($opt=~/^(--\w+[\w-]*)=(.*)$/) {
+            print STDERR "splitting longopt: $opt ($1 $2)\n" if $self->{debug};
+            ($opt,$arg)=($1,$2);
+         }
+         # handle boolean flags
+         $self->{flag}{$opt}=1,next if exists $self->{flag}{$opt};
+         # handle simple scalars
+         $self->{scalar}{$opt}=($arg || shift @opts),next
+            if exists $self->{scalar}{$opt};
+         # handle excludes
+         if ($opt eq '--exclude') {
+            $arg||=shift @opts;
+            # if they sent a reset, we will too
+            $self->{exclude}=[],next if $arg eq '!';
+            # otherwise add it to the list
+            push @{$self->{exclude}},$arg;
+            next;
+         }
+         # to preserve order we store both includes and excludes in the same
+         # array.  We use the leading '+ ' (plus space) trick from the man
+         # page to accomplish this.
+         if ($opt eq '--include') {
+            $arg||=shift @opts;
+            # first check to see if this is really an exclude
+            push(@{$self->{exclude}},$arg),next if $arg=~s/^- //;
+            # next see if they sent a reset, if they did, we will too
+            $self->{exclude}=[],next if $arg eq '!';
+            # if it really is an include, fix it first, since we use exclude
+            $arg='+ '.$arg unless $arg=~/^\+ /;
+            push @{$self->{exclude}},$arg;
+            next;
+         }
+         # handle our special case to override hard-coded path to rsync
+         $self->{'path-to-rsync'}=($arg || shift @opts),next
+            if $opt eq '--path-to-rsync';
+         # if we get this far nothing matched so it must be an error
+         carp "$opt - unknown option\n";
+         return undef;
+      } else { # must be data (source/destination info)
+         print STDERR "adding to data array: $opt\n" if $self->{debug};
+         push(@{$self->{data}},$opt);
+      }
+   }
+   1;
+}
+
+=over 4
+
+=item Rsunc::exec
+
+exec $obj @options or warn "rsync failed\n";
+
+   or
+
+$obj->exec(@options) or warn "rsync failed\n";
+
+=back
+
+This is the function that does the real work.  Any options passed to this
+routine are appended to any pre-set options and are not saved.  They effect
+the current execution of B<rsync> only.  It returns 1 if the return status was
+zero (or true), if the B<rsync> return status was non-zero it returns undef and
+stores the return status.  You can examine the return status from rsync and
+any output to stdout and stderr with the functions listed below.
+
+=cut
+
+sub exec {
+   my $self=shift;
+   my @cmd=($self->{'path-to-rsync'});
+
+   foreach (sort keys %{$self->{flag}}) {
+      push @cmd,$_ if $self->{flag}{$_};
+   }
+   foreach (sort keys %{$self->{scalar}}) {
+      push @cmd,$_.'='.$self->{scalar}{$_} if $self->{scalar}{$_};
+   }
+   foreach (@{$self->{exclude}}) {
+      push @cmd,'--exclude='.$_;
+   }
+   foreach (@{$self->{data}}) {
+      push @cmd,$_;
+   }
+   push @cmd,@_ if @_;
+   print STDERR "exec: @cmd\n" if $self->{debug};
+   my $in=FileHandle->new; my $out=FileHandle->new; my $err=FileHandle->new;
+   my $pid=eval{ open3 $in,$out,$err,@cmd };
+   if ($@) {
+      $self->{realstatus}=0;
+      $self->{status}=255;
+      $self->{err}=[$@,"Execution of rsync failed.\n"];
+      return undef;
+   }
+   $in->close; # we don't use it and neither should rsync (at least not yet)
+   $self->{err}=[ $err->getlines ];
+   $self->{out}=[ $out->getlines ];
+   $err->close;
+   $out->close;
+   waitpid $pid,0;
+   $self->{realstatus}=$?;
+   $self->{status}=$?>>8;
+   return undef if $self->{status};
+   return 1;
+}
+
+=over 4
+
+=item Rsync::status
+
+$rval = status $obj;
+
+   or
+
+$rval = $obj->status;
+
+=back
+
+Returns the status from last B<exec> call right shifted 8 bits.
+
+=cut
+
+sub status {
+   my $self=shift;
+   return $self->{status};
+}
+
+=over 4
+
+=item Rsync::realstatus
+
+$rval = realstatus $obj;
+
+   or
+
+$rval = $obj->realstatus;
+
+=back
+
+Returns the real status from last B<exec> call (not right shifted).
+
+=cut
+
+sub realstatus {
+   my $self=shift;
+   return $self->{realstatus};
+}
+
+=over 4
+
+=item Rsync::err
+
+$aref = err $obj;
+
+   or
+
+$aref = $obj->err;
+
+=back
+
+Returns an array or a reference to the array containing all output to stderr
+from the last B<exec> call.  B<rsync> sends all messages from the remote
+B<rsync> process to stderr.  This functions purpose is to make it easier for
+you to parse that output for appropriate information.
+
+=cut
+
+sub err {
+   my $self=shift;
+   return(wantarray ? @{$self->{err}} : $self->{err});
+}
+
+=over 4
+
+=item Rsync::out
+
+$aref = out $obj;
+
+   or
+
+$aref = $obj->out;
+
+=back
+
+Similar to the B<err> function, this returns an array or a reference to the
+array containing all output to stdout from the last B<exec> call.  B<rsync>
+sends all messages from the local B<rsync> process to stdout.
+
+=cut
+
+sub out {
+   my $self=shift;
+   return(wantarray ? @{$self->{out}} : $self->{out});
+}
+
+=head1 Author
+
+Lee Eakin E<lt>leakin@nostrum.comE<gt>
+
+=head1 Credits
+
+Gerard Hickey                             C<PGP::Pipe>
+
+Russ Allbery                              C<PGP::Sign>
+
+Graham Barr                               C<Net::*>
+
+Andrew Tridgell and Paul Mackerras        C<rsync(1)>
+
+John Steele   E<lt>steele@nostrum.comE<gt>
+
+Philip Kizer  E<lt>pckizer@nostrum.comE<gt>
+
+Larry Wall                                C<perl(1)>
+
+I borrowed many clues on wrapping an external program from the PGP modules,
+and I would not have had such a useful tool to wrap except for the great work
+of the B<rsync> authors.  Thanks also to Graham Barr, the author of the libnet
+modules and many others, for looking over this code.  Of course I must mention
+the other half of my brain, John Steele, and his good friend Philip Kizer for
+finding B<rsync> and bringing it to my attention.  And I would not have been
+able to enjoy writing useful tools if not for the creator of the B<perl>
+language.
+
+=head1 Copyrights
+
+      Copyleft (l) 1999, by Lee Eakin
+
+=cut
+
+1;

rsync-web/security.html

@@ -0,0 +1,284 @@
+<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
+<HTML>
+<HEAD>
+<TITLE>rsync</TITLE>
+<style>
+.security { color: red; }
+h3 { margin-bottom: 0px; }
+.date { color: #D25A0B; }
+</style>
+</HEAD>
+<!--#include virtual="header.html" -->
+
+<H2 align="center">Rsync Security Advisories</H2>
+
+<p>You should install a security fix for rsync when the rync you are running is:<ol>
+<li>older than 3.2.5 and pulling from an untrusted server
+<li>older than 3.2.5 and using the bundled zlib
+<li>older than 3.1.3 with --xattrs enabled
+<li>older than 3.1.3 with a writable rsync daemon
+<li>older than 2.6.6
+</ol>
+
+<p><a name="s3_2_5"></a><hr>
+<h3>Improved file-list validation in 3.2.5</h3>
+<i class=date>August 14th, 2022</i>
+
+<p>If you are running an rsync older than 3.2.5 and pulling files from an
+untrusted server, upgrade to 3.2.5 to get some added file-list validation rules
+that should prevent the sender from sneaking in extra top-level arguments
+and/or including files/dirs that should have been filtered out by the client's
+filter rules. Fixes CVE-2022-29154.
+
+<p><a name="s3_2_5-2"></a><hr>
+<h3>Zlib memory corruption bug in rsync 2.6.6 - 3.2.4</h3>
+<i class=date>August 14th, 2022</i>
+
+<p>If your rsync is configured to use the bundled zlib, you should upgrade to
+3.2.5 which contains the official zlib fix for a buffer overrun bug that was
+detailed in CVE-2022-37434. While you're at it, be sure to update your system's
+zlib.
+
+<p><a name="s3_2_4"></a><hr>
+<h3>Zlib memory corruption bug in rsync 2.6.6 - 3.2.3</h3>
+<i class=date>April 15th, 2022</i>
+
+<p>If your rsync is configured to use the bundled zlib, you should upgrade to
+3.2.4 which contains the official zlib fix for a memory corruption bug that was
+detailed in CVE-2018-25032. While you're at it, be sure to update your system's
+zlib.
+
+<p><a name="s3_1_3"></a><hr>
+<h3>File-list validation in 3.1.3</h3>
+<i class=date>January 28st, 2018</i>
+
+<p>If you are using a version of rsync older than 3.1.3 as a client and
+receiving xattrs from an rsync server that you might not fully trust, a
+malicious (modified) server could send a non-null-terminated xattr name to
+overflow the xattr read buffer.
+
+<p>If you are running a writable rsync daemon older than 3.1.3, you should add
+a rule "refuse options = protect-args" if you don't fully trust the users who
+are sending you files.
+
+<p><a name="s3_1_2"></a><hr>
+<h3>File-list validation in 3.1.2</h3>
+<i class=date>December 21st, 2015</i>
+
+<p>If you're using a version of rsync older than 3.1.2 as a client and
+receiving files from an rsync server that you might not fully trust, this
+version adds extra checking to the file list to prevent the sender from
+tweaking the paths and/or the transfer requests in a way that could cause
+a file to be received outside the transfer destination.
+
+<p><a name="s3_0_2"></a><hr>
+<h3>Xattr security fix in 3.0.2</h3>
+<i class=date>April 8th, 2008</i>
+
+<p>If you're using a version of rsync from 2.6.9 to 3.0.1 that has extended
+attribute (xattr) support enabled, you should upgrade to 3.0.2 to avoid a
+potential buffer overflow problem.  All 3.x versions have the potential to
+support xattrs (depending on OS availability and the configure options used),
+but version 2.6.9 had to be patched for this support.  You can run the command
+"rsync --version" and look for the string "xattrs" (as long as it is not
+"no&nbsp;xattrs") to see if your rsync is affected.
+
+<p>For those running affected versions, there is
+<a href="https://download.samba.org/pub/rsync/security/rsync-3.0.1-xattr-alloc.diff">a
+patch with the fix available</a>.
+
+<p>Those running a writable rsync daemon can opt to refuse the "xattrs" option
+as a way to avoid the problem without an upgrade:
+
+<blockquote><pre>refuse options = xattrs</pre></blockquote>
+
+<p>(If you already refuse options, be sure to append "xattrs" to your existing
+config parameter rather than adding another refuse directive.)
+
+<p><a name="s3_0_0"></a><hr>
+<h3>Daemon security fixes in 3.0.0 (with patches for 2.6.9)</h3>
+<i class=date>First published on November 28th, 2007;<br>
+Updated on December 16th, 2007;<br>
+Item 3 added on February 15th, 2008</i>
+
+<p>Three security advisories affect people who run a <b>writable</b> rsync
+daemon:  The first affects only those with "use chroot = no" (which is not a
+very safe combination in general), the second affects a daemon that has
+daemon-excluded files that are being hidden in a module's hierarchy, and
+the third affects only those with "use chroot = yes".
+Included are simple config-change suggestions that should help you to
+avoid the security issues and patches that make things safer.
+These advisories affect all rsync versions.
+
+<h4>1. Daemon advisory for "use chroot = no"</h4>
+
+<p>If you are running a writable rsync daemon with "use chroot = no", there
+is at least one way for someone to trick rsync into creating a symlink
+that points outside of the module's hierarchy.
+
+<p>This means that if you are allowing access from users who you don't
+trust, that you should either figure out a way to turn on "use chroot",
+or configure the daemon to refuse the "links" option (see "refuse
+options" in the rsyncd.conf manpage) which will disable the ability of
+the rsync module to receive symlinks.  After doing so, you should also
+check that any existing symlinks in the daemon hierarchy are safe.
+
+<p>Starting with the 3.0.0-pre6 release, there is a new daemon parameter
+available: "munge symlinks".  This allows an rsync daemon to accept
+symlinks and return them intact (with even a leading slash still there,
+which is new for a non-chroot daemon), but will not allow the symlinks
+to be used while they are in the daemon's hierarchy.
+
+<p>For those running
+2.6.9, there is
+<a href="https://download.samba.org/pub/rsync/security/rsync-2.6.9-munge-symlinks.diff">a
+patch for 2.6.9 to implement this parameter</a>.
+
+<p>Any admin applying that patch should read the "munge symlinks" section
+of the modified rsyncd.conf manpage for more information.  You can also
+read about this parameter in the
+<a href="https://download.samba.org/pub/rsync/rsyncd.conf.html">rsyncd.conf
+manpage from a 3.x version</a>.
+
+<h4>2. Daemon advisory for daemon excludes</h4>
+
+<p>If you are running a writable rsync daemon that is using one of the
+"exclude", "exclude from", or "filter" parameters in the rsyncd.conf file
+to hide data from your users, you should be aware that there are tricks
+that a user can play with symlinks and/or certain options that can allow
+a user that knows the name of a hidden file to access it or overwrite it
+(if file permissions allow that).
+
+<p>You can avoid the symlink problem using the suggestions in the advisory
+above.
+
+<p>When a daemon has "use chroot = no" set , there was some buggy
+exclude-checking for these options: <code>--compare-dest</code>,
+<code>--link-dest</code>, <code>--copy-dest</code>, <code>--partial-dir</code>,
+<code>--backup-dir</code>, <code>--temp-dir</code>, and
+<code>--files-from</code>.  These are fixed in the 3.0.0pre7 release.  For
+those running 2.6.9, there is <a
+href="https://download.samba.org/pub/rsync/security/rsync-2.6.9-daemon-exclude.diff">a patch for
+2.6.9</a> to fix these checks.
+
+<p>Some of the above options can still cause problems if an excluded
+sub-directory or filename is inside the option's directory hierarchy and the
+names of a transferred file clashes with it.  The affected options are the
+various <code>--*-dest</code> options (of which only <code>--link-dest</code>
+is particularly worrisome),
+<code>--backup-dir</code>, and <code>--partial-dir</code>.
+
+<p>You can avoid these sub-path problems by putting the following "refuse
+options" setting into your rsyncd.conf file:
+
+<blockquote><pre>refuse options = link-dest backup-dir partial-dir</pre></blockquote>
+
+<p>Those who aren't using an rsync with the latest exclude fixes may want to add
+some of the other affected options as well.
+
+<h4>3. Daemon advisory for "use chroot = yes"</h4>
+
+<p>If you are running a writable rsync daemon with "use chroot = yes", you
+should take care that users cannot upload their own library files and attempt
+to load them.
+
+<p>Beginning with rsync 3.0.0pre10, you can specify an inside-chroot path that
+makes the top of the transfer a subdirectory inside the chroot area, and that
+automatically makes library loading occur outside the transfer area (assuming
+you didn't pick an unwise subdirectory name for the transfer area and you
+don't have symlinks that point outside the transfer area).
+
+<p>If that solution is not good for you, the easiest way to protect your daemon
+is to create some appropriate directories in the top of your module's path
+hierarchy, such as "/etc", "/lib", and "/usr" (and any other top-level dirs
+that might be in the load path), chown those directories to some other user
+than the one that the module runs as (so that rsync will not be able to write
+files there, assuming that it is not run as root), and then hide the dirs using
+an exclude directive (either add a new one or extend your existing one):
+
+<blockquote><pre>exclude = /etc /lib /usr</pre></blockquote>
+
+<p>Doing all that will assure you that no user will be able to use rsync to
+upload a library that can be potentially loaded while rsync is attempting to
+perform an action, such as translating a username.  You can feel free to put
+trusted libraries that you want rsync to access in the protected hierarchies,
+as desired.
+
+<p>Also available in rsync 3.0.0pre10 is a new daemon parameter that allows you
+to avoid the accessing of user/group-name translation libraries by a chrooted
+rsync:  the "numeric ids" daemon parameter lets you turn on a forced
+numeric-only mode.  The default for a chroot module is to enable this
+parameter, while the default for a non-chroot module is to disable it.
+
+<p>For those running 2.6.9, there is <a
+href="https://download.samba.org/pub/rsync/security/rsync-2.6.9-daemon-ids.diff">a patch for
+2.6.9</a> to add the "numeric ids" daemon config parameter.  (The patch will
+only apply cleanly if you've already applied the munge-symlinks diff mentioned
+above.)
+
+<p><a name="s2_6_8"></a><hr>
+<h3>Xattr security fix in 2.6.8</h3>
+<i class=date>April 22th, 2006</i>
+
+<p>If you're using a version of rsync prior to 2.6.8 that was patched to
+include extended attribute (xattr) support, you should upgrade to 2.6.8 or
+later to avoid a potential buffer overflow problem.
+
+<p><a name="s2_6_6"></a><hr>
+<h3>Zlib security fix in 2.6.6</h3>
+<i class=date>July 28th, 2005</i>
+
+<p>If you're using a version of rsync prior to 2.6.6, there is a potential
+null-pointer security bug in the zlib code.  You can avoid its affect in an
+rsync daemon situation by configuring rsync to refuse the "compress" option.
+
+<p><a name="s2_6_3"></a><hr>
+<h3>Daemon path-sanitizing fix in 2.6.3</h3>
+<i class=date>August 12th, 2004</i>
+
+<p>There is a path-sanitizing bug that affects daemon-mode in
+rsync versions prior to 2.6.3, but only if "use chroot" is disabled.  It
+does <b>not</b> affect the normal send/receive filenames that specify what
+files should be transferred (this is because these names happen to get
+sanitized twice, and thus the second call removes any lingering leading
+slash(es) that the first call left behind).  It does affect certain
+option paths that cause auxiliary files to be read or written.
+
+<p>One potential fix that doesn't require recompiling rsync is to set
+"use chroot = true" for all the modules in the rsyncd.conf file.
+
+<p><a name="s2_6_1"></a><hr>
+<h3>Daemon security fix in 2.6.1</h3>
+<i class=date>April 26th, 2004</i>
+
+<p>There is a security problem in all versions prior to 2.6.1 that affects only
+people running a read/write daemon with "use chroot" disabled.  If the user privs
+of a module in the daemon config is anything above "nobody", you are at risk
+of someone crafting an attack that could write a file outside of the module's
+"path" setting (where all its files should be stored).  Please either enable
+chroot or upgrade to 2.6.1.  People not running a daemon, running a read-only
+daemon, or running a chrooted daemon are totally unaffected.
+
+<p><a name="s2_5_7"></a><hr>
+<h3>Memory overflow fix in 2.5.7</h3>
+<i class=date>December 4th, 2003</i>
+
+<p>Rsync versions prior to 2.5.7 contain a heap overflow vulnerability that
+could be used to remotely run arbitrary code, but this only affects the use of
+rsync as an "rsync daemon" (where rsync handles incoming socket connections,
+typically on port 873).
+
+<!--#include virtual="footer.html" -->
+<br>&nbsp;
+<br>&nbsp;
+<br>&nbsp;
+<br>&nbsp;
+<br>&nbsp;
+<br>&nbsp;
+<br>&nbsp;
+<br>&nbsp;
+<br>&nbsp;
+<br>&nbsp;
+<br>&nbsp;
+<br>&nbsp;
+<br>&nbsp;

rsync-web/susan.txt

@@ -0,0 +1,31 @@
+I use rsync to backup my wifes home directory across a modem link each
+night. The cron job looks like this
+
+    #!/bin/sh
+    cd ~susan
+    {
+    echo
+    date
+    dest=~/backup/`date +%A`
+    mkdir $dest.new
+    find . -xdev -type f \( -mtime 0 -or -mtime 1 \) -exec cp -aPv "{}"
+    $dest.new \;
+    cnt=`find $dest.new -type f | wc -l`
+    if [ $cnt -gt 0 ]; then
+      rm -rf $dest
+      mv $dest.new $dest
+    fi
+    rm -rf $dest.new
+    rsync -Cavze ssh . samba:backup
+    } >> ~/backup/backup.log 2>&1
+
+
+note that most of this script isn't anything to do with rsync, it just
+creates a daily backup of Susans work in a ~susan/backup/ directory so
+she can retrieve any version from the last week. The last line does
+the rsync of her directory across the modem link to the host
+samba. Note that I am using the -C option which allows me to add
+entries to .cvsignore for stuff that doesn't need to be backed up.
+
+
+

rsync-web/tech_report/footnode.html

@@ -0,0 +1,134 @@
+<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2 Final//EN">
+<!--Converted with LaTeX2HTML 98.1p1 release (March 2nd, 1998)
+originally by Nikos Drakos (nikos@cbl.leeds.ac.uk), CBLU, University of Leeds
+* revised and updated by:  Marcus Hennecke, Ross Moore, Herb Swan
+* with significant contributions from:
+  Jens Lippmann, Marek Rouchal, Martin Wilck and others -->
+<HTML>
+<HEAD>
+<TITLE>Footnotes</TITLE>
+<META NAME="description" CONTENT="Footnotes">
+<META NAME="keywords" CONTENT="tech_report">
+<META NAME="resource-type" CONTENT="document">
+<META NAME="distribution" CONTENT="global">
+<META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=iso-8859-1">
+<LINK REL="STYLESHEET" HREF="tech_report.css">
+<LINK REL="previous" HREF="node7.html">
+<LINK REL="up" HREF="tech_report.html">
+</HEAD>
+<BODY >
+
+<DL>
+<DT><A NAME="foot10">... bytes</A><A NAME="foot10"
+ HREF="node2.html#tex2html1"><SUP>1</SUP></A>
+<DD>We have found that
+  values of S between 500 and 1000 are quite good for most purposes
+<PRE>.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+</PRE>
+<DT><A NAME="foot24">... size.</A><A NAME="foot24"
+ HREF="node6.html#tex2html2"><SUP>2</SUP></A>
+<DD>All the tests in this section were
+  carried out using rsync version 0.5
+<PRE>.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+</PRE>
+<DT><A NAME="foot40">... files.</A><A NAME="foot40"
+ HREF="node6.html#tex2html3"><SUP>3</SUP></A>
+<DD>The wall
+  clock time was approximately 2 minutes per run on a 50 MHz SPARC 10
+  running SunOS, using rsh over loopback for communication.  GNU diff
+  took about 4 minutes.
+<PRE>.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+.
+</PRE>
+</DL><ADDRESS>
+<I>Andrew Tridgell</I>
+<BR><I>1998-11-09</I>
+</ADDRESS>
+</BODY>
+</HTML>

rsync-web/tech_report/images.aux

@@ -0,0 +1 @@
+\relax 

rsync-web/tech_report/images.log

@@ -0,0 +1,106 @@
+This is TeX, Version 3.14159 (C version 6.1) (format=latex 97.10.16)  9 NOV 1998 20:10
+**./images.tex
+(images.tex
+LaTeX2e <1996/12/01> patch level 1
+Babel <v3.6h> and hyphenation patterns for american, german, loaded.
+
+(/usr/lib/texmf/texmf/tex/latex/base/article.cls
+Document Class: article 1996/10/31 v1.3u Standard LaTeX document class
+(/usr/lib/texmf/texmf/tex/latex/base/size10.clo
+File: size10.clo 1996/10/31 v1.3u Standard LaTeX file (size option)
+)
+\c@part=\count79
+\c@section=\count80
+\c@subsection=\count81
+\c@subsubsection=\count82
+\c@paragraph=\count83
+\c@subparagraph=\count84
+\c@figure=\count85
+\c@table=\count86
+\abovecaptionskip=\skip41
+\belowcaptionskip=\skip42
+\bibindent=\dimen102
+) (/usr/lib/texmf/texmf/tex/latex/graphics/color.sty
+Package: color 1997/01/07 v1.0d Standard LaTeX Color (DPC)
+(/usr/lib/texmf/texmf/tex/latex/config/color.cfg)
+Package color Info: Driver file: dvips.def on input line 128.
+(/usr/lib/texmf/texmf/tex/latex/graphics/dvips.def
+File: dvips.def 1996/12/12 v3.0d Driver-dependant file (DPC,SPQR)
+) (/usr/lib/texmf/texmf/tex/latex/graphics/dvipsnam.def
+File: dvipsnam.def 1996/12/12 v3.0d Driver-dependant file (DPC,SPQR)
+))
+\sizebox=\box26
+\lthtmlwrite=\write3
+No file images.aux.
+LaTeX Font Info:    Checking defaults for OML/cmm/m/it on input line 74.
+LaTeX Font Info:    ... okay on input line 74.
+LaTeX Font Info:    Checking defaults for T1/cmr/m/n on input line 74.
+LaTeX Font Info:    ... okay on input line 74.
+LaTeX Font Info:    Checking defaults for OT1/cmr/m/n on input line 74.
+LaTeX Font Info:    ... okay on input line 74.
+LaTeX Font Info:    Checking defaults for OMS/cmsy/m/n on input line 74.
+LaTeX Font Info:    ... okay on input line 74.
+LaTeX Font Info:    Checking defaults for OMX/cmex/m/n on input line 74.
+LaTeX Font Info:    ... okay on input line 74.
+LaTeX Font Info:    Checking defaults for U/cmr/m/n on input line 74.
+LaTeX Font Info:    ... okay on input line 74.
+latex2htmlLength hsize=349.0pt
+latex2htmlLength vsize=682.0pt
+latex2htmlLength hoffset=0.0pt
+latex2htmlLength voffset=0.0pt
+latex2htmlLength topmargin=0.0pt
+latex2htmlLength topskip=0.00003pt
+latex2htmlLength headheight=0.0pt
+latex2htmlLength headsep=0.0pt
+latex2htmlLength parskip=0.0pt plus 1.0pt
+latex2htmlLength oddsidemargin=53.0pt
+latex2htmlLength evensidemargin=53.0pt
+LaTeX Font Info:    External font `cmex10' loaded for size
+(Font)              <7> on input line 94.
+LaTeX Font Info:    External font `cmex10' loaded for size
+(Font)              <5> on input line 94.
+l2hSize :tex2html_wrap_inline88:6.45831pt::0.0pt::6.43404pt.
+[1
+
+
+]
+l2hSize :tex2html_wrap_inline90:7.44444pt::7.44444pt::6.18402pt.
+[2
+
+
+]
+l2hSize :displaymath158:32.38896pt::0.0pt::349.0pt.
+[3
+
+
+]
+l2hSize :displaymath160:32.38896pt::0.0pt::349.0pt.
+[4
+
+
+]
+l2hSize :tex2html_wrap_inline166:7.33331pt::7.33331pt::39.55894pt.
+[5
+
+
+]
+l2hSize :displaymath170:14.5pt::0.0pt::349.0pt.
+[6
+
+
+]
+l2hSize :displaymath172:14.5pt::0.0pt::349.0pt.
+[7
+
+
+] (images.aux) ) 
+Here is how much of TeX's memory you used:
+ 518 strings out of 10906
+ 6196 string characters out of 72186
+ 46914 words of memory out of 262141
+ 3454 multiletter control sequences out of 9500
+ 3640 words of font info for 14 fonts, out of 150000 for 255
+ 14 hyphenation exceptions out of 607
+ 23i,5n,19p,182b,215s stack positions out of 300i,40n,60p,3000b,4000s
+
+Output written on images.dvi (7 pages, 2296 bytes).

rsync-web/tech_report/images.pl

@@ -0,0 +1,48 @@
+# LaTeX2HTML 98.1p1 release (March 2nd, 1998)
+# Associate images original text with physical files.
+
+
+$key = q/{inline}alpha{inline}MSF=1.6;AAT;/;
+$cached_env_img{$key} = q|<IMG
+ WIDTH="15" HEIGHT="13" ALIGN="BOTTOM" BORDER="0"
+ SRC="img1.gif"
+ ALT="$\alpha$">|; 
+
+$key = q/{displaymath}a(k+1,l+1)=(a(k,l)-X_k+X_l+1)bmodM{displaymath}MSF=1.6;AAT;/;
+$cached_env_img{$key} = q|<IMG
+ WIDTH="322" HEIGHT="28"
+ SRC="img6.gif"
+ ALT="\begin{displaymath}a(k+1,l+1) = (a(k,l) - X_k + X_{l+1}) \bmod M \end{displaymath}">|; 
+
+$key = q/{displaymath}a(k,l)=(sum_i=k^lX_i)bmodM{displaymath}MSF=1.6;AAT;/;
+$cached_env_img{$key} = q|<IMG
+ WIDTH="176" HEIGHT="56"
+ SRC="img3.gif"
+ ALT="\begin{displaymath}a(k,l) = (\sum_{i=k}^l X_i) \bmod M \end{displaymath}">|; 
+
+$key = q/{displaymath}b(k,l)=(sum_i=k^l(l-i+1)X_i)bmodM{displaymath}MSF=1.6;AAT;/;
+$cached_env_img{$key} = q|<IMG
+ WIDTH="240" HEIGHT="56"
+ SRC="img4.gif"
+ ALT="\begin{displaymath}b(k,l) = (\sum_{i=k}^l (l-i+1)X_i) \bmod M \end{displaymath}">|; 
+
+$key = q/{inline}X_kldotsX_l{inline}MSF=1.6;AAT;/;
+$cached_env_img{$key} = q|<IMG
+ WIDTH="67" HEIGHT="29" ALIGN="MIDDLE" BORDER="0"
+ SRC="img5.gif"
+ ALT="$X_k \ldots X_l$">|; 
+
+$key = q/{displaymath}b(k+1,l+1)=(b(k,l)-(l-k+1)X_k+a(k+1,l+1))bmodM{displaymath}MSF=1.6;AAT;/;
+$cached_env_img{$key} = q|<IMG
+ WIDTH="454" HEIGHT="28"
+ SRC="img7.gif"
+ ALT="\begin{displaymath}b(k+1,l+1) = (b(k,l) - (l-k+1) X_k + a(k+1,l+1)) \bmod M \end{displaymath}">|; 
+
+$key = q/{inline}beta{inline}MSF=1.6;AAT;/;
+$cached_env_img{$key} = q|<IMG
+ WIDTH="14" HEIGHT="29" ALIGN="MIDDLE" BORDER="0"
+ SRC="img2.gif"
+ ALT="$\beta$">|; 
+
+1;
+

rsync-web/tech_report/images.tex

@@ -0,0 +1,140 @@
+\batchmode
+\documentclass[a4paper]{article}
+\makeatletter
+\usepackage[dvips]{color}
+\pagecolor[gray]{.7}
+
+
+\makeatletter
+\count@=\the\catcode`\_ \catcode`\_=8 
+\newenvironment{tex2html_wrap}{}{} \catcode`\_=\count@
+\makeatother
+\let\mathon=$
+\let\mathoff=$
+\ifx\AtBeginDocument\undefined \newcommand{\AtBeginDocument}[1]{}\fi
+\newbox\sizebox
+\setlength{\hoffset}{0pt}\setlength{\voffset}{0pt}
+\addtolength{\textheight}{\footskip}\setlength{\footskip}{0pt}
+\addtolength{\textheight}{\topmargin}\setlength{\topmargin}{0pt}
+\addtolength{\textheight}{\headheight}\setlength{\headheight}{0pt}
+\addtolength{\textheight}{\headsep}\setlength{\headsep}{0pt}
+\setlength{\textwidth}{349pt}
+\newwrite\lthtmlwrite
+\makeatletter
+\let\realnormalsize=\normalsize
+\global\topskip=2sp
+\def\preveqno{}\let\real@float=\@float \let\realend@float=\end@float
+\def\@float{\let\@savefreelist\@freelist\real@float}
+\def\end@float{\realend@float\global\let\@freelist\@savefreelist}
+\let\real@dbflt=\@dbflt \let\end@dblfloat=\end@float
+\let\@largefloatcheck=\relax
+\def\@dbflt{\let\@savefreelist\@freelist\real@dbflt}
+\def\adjustnormalsize{\def\normalsize{\mathsurround=0pt \realnormalsize
+ \parindent=0pt\abovedisplayskip=0pt\belowdisplayskip=0pt}\normalsize}%
+\def\lthtmltypeout#1{{\let\protect\string\immediate\write\lthtmlwrite{#1}}}%
+\newcommand\lthtmlhboxmathA{\adjustnormalsize\setbox\sizebox=\hbox\bgroup}%
+\newcommand\lthtmlvboxmathA{\adjustnormalsize\setbox\sizebox=\vbox\bgroup%
+ \let\ifinner=\iffalse }%
+\newcommand\lthtmlboxmathZ{\@next\next\@currlist{}{\def\next{\voidb@x}}%
+ \expandafter\box\next\egroup}%
+\newcommand\lthtmlmathtype[1]{\def\lthtmlmathenv{#1}}%
+\newcommand\lthtmllogmath{\lthtmltypeout{l2hSize %
+:\lthtmlmathenv:\the\ht\sizebox::\the\dp\sizebox::\the\wd\sizebox.\preveqno}}%
+\newcommand\lthtmlfigureA[1]{\let\@savefreelist\@freelist
+       \lthtmlmathtype{#1}\lthtmlvboxmathA}%
+\newcommand\lthtmlfigureZ{\lthtmlboxmathZ\lthtmllogmath\copy\sizebox
+       \global\let\@freelist\@savefreelist}%
+\newcommand\lthtmldisplayA[1]{\lthtmlmathtype{#1}\lthtmlvboxmathA}%
+\newcommand\lthtmldisplayB[1]{\edef\preveqno{(\theequation)}%
+  \lthtmldisplayA{#1}\let\@eqnnum\relax}%
+\newcommand\lthtmldisplayZ{\lthtmlboxmathZ\lthtmllogmath\lthtmlsetmath}%
+\newcommand\lthtmlinlinemathA[1]{\lthtmlmathtype{#1}\lthtmlhboxmathA  \vrule height1.5ex width0pt }%
+\newcommand\lthtmlinlineA[1]{\lthtmlmathtype{#1}\lthtmlhboxmathA}%
+\newcommand\lthtmlinlineZ{\egroup\expandafter\ifdim\dp\sizebox>0pt %
+  \expandafter\centerinlinemath\fi\lthtmllogmath\lthtmlsetinline}
+\newcommand\lthtmlinlinemathZ{\egroup\expandafter\ifdim\dp\sizebox>0pt %
+  \expandafter\centerinlinemath\fi\lthtmllogmath\lthtmlsetmath}
+\def\lthtmlsetinline{\hbox{\vrule width.1em\vtop{\vbox{%
+  \kern.1em\copy\sizebox}\ifdim\dp\sizebox>0pt\kern.1em\else\kern.3pt\fi
+  \ifdim\hsize>\wd\sizebox \hrule depth1pt\fi}}}
+\def\lthtmlsetmath{\hbox{\vrule width.1em\vtop{\vbox{%
+  \kern.1em\kern0.8 pt\hbox{\hglue.17em\copy\sizebox\hglue0.8 pt}}\kern.3pt%
+  \ifdim\dp\sizebox>0pt\kern.1em\fi \kern0.8 pt%
+  \ifdim\hsize>\wd\sizebox \hrule depth1pt\fi}}}
+\def\centerinlinemath{%\dimen1=\ht\sizebox
+  \dimen1=\ifdim\ht\sizebox<\dp\sizebox \dp\sizebox\else\ht\sizebox\fi
+  \advance\dimen1by.5pt \vrule width0pt height\dimen1 depth\dimen1 
+ \dp\sizebox=\dimen1\ht\sizebox=\dimen1\relax}
+
+\def\lthtmlcheckvsize{\ifdim\ht\sizebox<\vsize\expandafter\vfill
+  \else\expandafter\vss\fi}%
+\makeatletter \tracingstats = 1 
+
+
+\begin{document}
+\pagestyle{empty}\thispagestyle{empty}%
+\lthtmltypeout{latex2htmlLength hsize=\the\hsize}%
+\lthtmltypeout{latex2htmlLength vsize=\the\vsize}%
+\lthtmltypeout{latex2htmlLength hoffset=\the\hoffset}%
+\lthtmltypeout{latex2htmlLength voffset=\the\voffset}%
+\lthtmltypeout{latex2htmlLength topmargin=\the\topmargin}%
+\lthtmltypeout{latex2htmlLength topskip=\the\topskip}%
+\lthtmltypeout{latex2htmlLength headheight=\the\headheight}%
+\lthtmltypeout{latex2htmlLength headsep=\the\headsep}%
+\lthtmltypeout{latex2htmlLength parskip=\the\parskip}%
+\lthtmltypeout{latex2htmlLength oddsidemargin=\the\oddsidemargin}%
+\makeatletter
+\if@twoside\lthtmltypeout{latex2htmlLength evensidemargin=\the\evensidemargin}%
+\else\lthtmltypeout{latex2htmlLength evensidemargin=\the\oddsidemargin}\fi%
+\makeatother
+\stepcounter{section}
+\stepcounter{section}
+{\newpage\clearpage
+\lthtmlinlinemathA{tex2html_wrap_inline88}%
+$\alpha$%
+\lthtmlinlinemathZ
+\hfill\lthtmlcheckvsize\clearpage}
+
+{\newpage\clearpage
+\lthtmlinlinemathA{tex2html_wrap_inline90}%
+$\beta$%
+\lthtmlinlinemathZ
+\hfill\lthtmlcheckvsize\clearpage}
+
+\stepcounter{section}
+{\newpage\clearpage
+\lthtmldisplayA{displaymath158}%
+\begin{displaymath}a(k,l) = (\sum_{i=k}^l X_i) \bmod M \end{displaymath}%
+\lthtmldisplayZ
+\hfill\lthtmlcheckvsize\clearpage}
+
+{\newpage\clearpage
+\lthtmldisplayA{displaymath160}%
+\begin{displaymath}b(k,l) = (\sum_{i=k}^l (l-i+1)X_i) \bmod M \end{displaymath}%
+\lthtmldisplayZ
+\hfill\lthtmlcheckvsize\clearpage}
+
+{\newpage\clearpage
+\lthtmlinlinemathA{tex2html_wrap_inline166}%
+$X_k \ldots X_l$%
+\lthtmlinlinemathZ
+\hfill\lthtmlcheckvsize\clearpage}
+
+{\newpage\clearpage
+\lthtmldisplayA{displaymath170}%
+\begin{displaymath}a(k+1,l+1) = (a(k,l) - X_k + X_{l+1}) \bmod M \end{displaymath}%
+\lthtmldisplayZ
+\hfill\lthtmlcheckvsize\clearpage}
+
+{\newpage\clearpage
+\lthtmldisplayA{displaymath172}%
+\begin{displaymath}b(k+1,l+1) = (b(k,l) - (l-k+1) X_k + a(k+1,l+1)) \bmod M \end{displaymath}%
+\lthtmldisplayZ
+\hfill\lthtmlcheckvsize\clearpage}
+
+\stepcounter{section}
+\stepcounter{section}
+\stepcounter{section}
+\stepcounter{section}
+
+\end{document}

rsync-web/tech_report/index.html

@@ -0,0 +1,94 @@
+<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2 Final//EN">
+<!--Converted with LaTeX2HTML 98.1p1 release (March 2nd, 1998)
+originally by Nikos Drakos (nikos@cbl.leeds.ac.uk), CBLU, University of Leeds
+* revised and updated by:  Marcus Hennecke, Ross Moore, Herb Swan
+* with significant contributions from:
+  Jens Lippmann, Marek Rouchal, Martin Wilck and others -->
+<HTML>
+<HEAD>
+<TITLE>The rsync algorithm</TITLE>
+<META NAME="description" CONTENT="The rsync algorithm">
+<META NAME="keywords" CONTENT="tech_report">
+<META NAME="resource-type" CONTENT="document">
+<META NAME="distribution" CONTENT="global">
+<META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=iso-8859-1">
+<LINK REL="STYLESHEET" HREF="tech_report.css">
+<LINK REL="next" HREF="node1.html">
+</HEAD>
+<BODY >
+<!--Navigation Panel-->
+<A NAME="tex2html4"
+ HREF="node1.html">
+<IMG WIDTH="37" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="next"
+ SRC="next.gif"></A> 
+<IMG WIDTH="26" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="up"
+ SRC="up.gif"> 
+<IMG WIDTH="63" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="previous"
+ SRC="previous.gif">   
+<BR>
+<B> Next:</B> <A NAME="tex2html5"
+ HREF="node1.html">The problem</A>
+<BR>
+<BR>
+<!--End of Navigation Panel-->
+
+<P>
+
+<P>
+
+<P>
+<H1 ALIGN="CENTER">The rsync algorithm</H1>
+<P ALIGN="CENTER"><STRONG>Andrew Tridgell &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;Paul Mackerras
+<BR>
+Department of Computer Science <BR>
+Australian National University <BR>
+Canberra, ACT 0200, Australia</STRONG></P>
+<P ALIGN="LEFT"></P>
+
+<P>
+
+<H3>Abstract:</H3>
+<DIV>
+This report presents an algorithm for updating a file on one machine
+  to be identical to a file on another machine.  We assume that the
+  two machines are connected by a low-bandwidth high-latency
+  bi-directional communications link.  The algorithm identifies parts
+  of the source file which are identical to some part of the
+  destination file, and only sends those parts which cannot be matched
+  in this way.  Effectively, the algorithm computes a set of
+  differences without having both files on the same machine.  The
+  algorithm works best when the files are similar, but will also
+  function correctly and reasonably efficiently when the files are
+  quite different.
+</DIV>
+<P>
+<P>
+<BR><HR>
+<!--Table of Child-Links-->
+<A NAME="CHILD_LINKS">&#160;</A>
+<UL>
+<LI><A NAME="tex2html6"
+ HREF="node1.html">The problem</A>
+<LI><A NAME="tex2html7"
+ HREF="node2.html">The rsync algorithm</A>
+<LI><A NAME="tex2html8"
+ HREF="node3.html">Rolling checksum</A>
+<LI><A NAME="tex2html9"
+ HREF="node4.html">Checksum searching</A>
+<LI><A NAME="tex2html10"
+ HREF="node5.html">Pipelining</A>
+<LI><A NAME="tex2html11"
+ HREF="node6.html">Results</A>
+<LI><A NAME="tex2html12"
+ HREF="node7.html">Availability</A>
+<LI><A NAME="tex2html13"
+ HREF="node8.html">About this document ... </A>
+</UL>
+<!--End of Table of Child-Links-->
+<BR><HR>
+<ADDRESS>
+<I>Andrew Tridgell</I>
+<BR><I>1998-11-09</I>
+</ADDRESS>
+</BODY>
+</HTML>

rsync-web/tech_report/labels.pl

@@ -0,0 +1,6 @@
+# LaTeX2HTML 98.1p1 release (March 2nd, 1998)
+# Associate labels original text with physical files.
+
+
+1;
+

rsync-web/tech_report/node1.html

@@ -0,0 +1,119 @@
+<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2 Final//EN">
+<!--Converted with LaTeX2HTML 98.1p1 release (March 2nd, 1998)
+originally by Nikos Drakos (nikos@cbl.leeds.ac.uk), CBLU, University of Leeds
+* revised and updated by:  Marcus Hennecke, Ross Moore, Herb Swan
+* with significant contributions from:
+  Jens Lippmann, Marek Rouchal, Martin Wilck and others -->
+<HTML>
+<HEAD>
+<TITLE>The problem</TITLE>
+<META NAME="description" CONTENT="The problem">
+<META NAME="keywords" CONTENT="tech_report">
+<META NAME="resource-type" CONTENT="document">
+<META NAME="distribution" CONTENT="global">
+<META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=iso-8859-1">
+<LINK REL="STYLESHEET" HREF="tech_report.css">
+<LINK REL="next" HREF="node2.html">
+<LINK REL="previous" HREF="tech_report.html">
+<LINK REL="up" HREF="tech_report.html">
+<LINK REL="next" HREF="node2.html">
+</HEAD>
+<BODY >
+<!--Navigation Panel-->
+<A NAME="tex2html22"
+ HREF="node2.html">
+<IMG WIDTH="37" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="next"
+ SRC="next.gif"></A> 
+<A NAME="tex2html20"
+ HREF="tech_report.html">
+<IMG WIDTH="26" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="up"
+ SRC="up.gif"></A> 
+<A NAME="tex2html14"
+ HREF="tech_report.html">
+<IMG WIDTH="63" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="previous"
+ SRC="previous.gif"></A>   
+<BR>
+<B> Next:</B> <A NAME="tex2html23"
+ HREF="node2.html">The rsync algorithm</A>
+<B> Up:</B> <A NAME="tex2html21"
+ HREF="tech_report.html">The rsync algorithm</A>
+<B> Previous:</B> <A NAME="tex2html15"
+ HREF="tech_report.html">The rsync algorithm</A>
+<BR>
+<BR>
+<!--End of Navigation Panel-->
+
+<H1><A NAME="SECTION00010000000000000000">
+The problem</A>
+</H1>
+
+<P>
+Imagine you have two files, <I>A</I> and <I>B</I>, and you wish to update <I>B</I> to be
+the same as <I>A</I>. The obvious method is to copy <I>A</I> onto <I>B</I>.
+
+<P>
+Now imagine that the two files are on machines connected by a slow
+communications link, for example a dial up IP link.  If <I>A</I> is large,
+copying <I>A</I> onto <I>B</I> will be slow.  To make it faster you could
+compress <I>A</I> before sending it, but that will usually only gain a
+factor of 2 to 4.
+
+<P>
+Now assume that <I>A</I> and <I>B</I> are quite similar, perhaps both derived
+from the same original file. To really speed things up you would need
+to take advantage of this similarity. A common method is to send just
+the differences between <I>A</I> and <I>B</I> down the link and then use this
+list of differences to reconstruct the file.
+
+<P>
+The problem is that the normal methods for creating a set of
+differences between two files rely on being able to read both files.
+Thus they require that both files are available beforehand at one end
+of the link.  If they are not both available on the same machine,
+these algorithms cannot be used (once you had copied the file over,
+you wouldn't need the differences).  This is the problem that rsync
+addresses.
+
+<P>
+The rsync algorithm efficiently computes which parts of a source file
+match some part of an existing destination file.  These parts need not
+be sent across the link; all that is needed is a reference to the part
+of the destination file.  Only parts of the source file which are not
+matched in this way need to be sent verbatim.  The receiver can then
+construct a copy of the source file using the references to parts of
+the existing destination file and the verbatim material.
+
+<P>
+Trivially, the data sent to the receiver can be compressed using any
+of a range of common compression algorithms, for further speed
+improvements.
+
+<P>
+<HR>
+<!--Navigation Panel-->
+<A NAME="tex2html22"
+ HREF="node2.html">
+<IMG WIDTH="37" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="next"
+ SRC="next.gif"></A> 
+<A NAME="tex2html20"
+ HREF="tech_report.html">
+<IMG WIDTH="26" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="up"
+ SRC="up.gif"></A> 
+<A NAME="tex2html14"
+ HREF="tech_report.html">
+<IMG WIDTH="63" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="previous"
+ SRC="previous.gif"></A>   
+<BR>
+<B> Next:</B> <A NAME="tex2html23"
+ HREF="node2.html">The rsync algorithm</A>
+<B> Up:</B> <A NAME="tex2html21"
+ HREF="tech_report.html">The rsync algorithm</A>
+<B> Previous:</B> <A NAME="tex2html15"
+ HREF="tech_report.html">The rsync algorithm</A>
+<!--End of Navigation Panel-->
+<ADDRESS>
+<I>Andrew Tridgell</I>
+<BR><I>1998-11-09</I>
+</ADDRESS>
+</BODY>
+</HTML>

rsync-web/tech_report/node2.html

@@ -0,0 +1,183 @@
+<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2 Final//EN">
+<!--Converted with LaTeX2HTML 98.1p1 release (March 2nd, 1998)
+originally by Nikos Drakos (nikos@cbl.leeds.ac.uk), CBLU, University of Leeds
+* revised and updated by:  Marcus Hennecke, Ross Moore, Herb Swan
+* with significant contributions from:
+  Jens Lippmann, Marek Rouchal, Martin Wilck and others -->
+<HTML>
+<HEAD>
+<TITLE>The rsync algorithm</TITLE>
+<META NAME="description" CONTENT="The rsync algorithm">
+<META NAME="keywords" CONTENT="tech_report">
+<META NAME="resource-type" CONTENT="document">
+<META NAME="distribution" CONTENT="global">
+<META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=iso-8859-1">
+<LINK REL="STYLESHEET" HREF="tech_report.css">
+<LINK REL="next" HREF="node3.html">
+<LINK REL="previous" HREF="node1.html">
+<LINK REL="up" HREF="tech_report.html">
+<LINK REL="next" HREF="node3.html">
+</HEAD>
+<BODY >
+<!--Navigation Panel-->
+<A NAME="tex2html32"
+ HREF="node3.html">
+<IMG WIDTH="37" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="next"
+ SRC="next.gif"></A> 
+<A NAME="tex2html30"
+ HREF="tech_report.html">
+<IMG WIDTH="26" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="up"
+ SRC="up.gif"></A> 
+<A NAME="tex2html24"
+ HREF="node1.html">
+<IMG WIDTH="63" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="previous"
+ SRC="previous.gif"></A>   
+<BR>
+<B> Next:</B> <A NAME="tex2html33"
+ HREF="node3.html">Rolling checksum</A>
+<B> Up:</B> <A NAME="tex2html31"
+ HREF="tech_report.html">The rsync algorithm</A>
+<B> Previous:</B> <A NAME="tex2html25"
+ HREF="node1.html">The problem</A>
+<BR>
+<BR>
+<!--End of Navigation Panel-->
+
+<H1><A NAME="SECTION00020000000000000000">
+The rsync algorithm</A>
+</H1>
+
+<P>
+Suppose we have two general purpose computers <IMG
+ WIDTH="15" HEIGHT="13" ALIGN="BOTTOM" BORDER="0"
+ SRC="img1.gif"
+ ALT="$\alpha$">
+and <IMG
+ WIDTH="14" HEIGHT="29" ALIGN="MIDDLE" BORDER="0"
+ SRC="img2.gif"
+ ALT="$\beta$">.
+Computer <IMG
+ WIDTH="15" HEIGHT="13" ALIGN="BOTTOM" BORDER="0"
+ SRC="img1.gif"
+ ALT="$\alpha$">
+has access to a file <I>A</I> and <IMG
+ WIDTH="14" HEIGHT="29" ALIGN="MIDDLE" BORDER="0"
+ SRC="img2.gif"
+ ALT="$\beta$">
+has access to
+file <I>B</I>, where <I>A</I> and <I>B</I> are ``similar''.  There is a slow
+communications link between <IMG
+ WIDTH="15" HEIGHT="13" ALIGN="BOTTOM" BORDER="0"
+ SRC="img1.gif"
+ ALT="$\alpha$">
+and <IMG
+ WIDTH="14" HEIGHT="29" ALIGN="MIDDLE" BORDER="0"
+ SRC="img2.gif"
+ ALT="$\beta$">.
+
+<P>
+The rsync algorithm consists of the following steps:
+
+<P>
+<DL COMPACT>
+<DT>1.
+<DD><IMG
+ WIDTH="14" HEIGHT="29" ALIGN="MIDDLE" BORDER="0"
+ SRC="img2.gif"
+ ALT="$\beta$">
+splits the file <I>B</I> into a series of non-overlapping
+fixed-sized blocks of size S bytes<A NAME="tex2html1"
+ HREF="footnode.html#foot10"><SUP>1</SUP></A>.
+  The last block may be shorter than <I>S</I> bytes.
+
+<P>
+<DT>2.
+<DD>For each of these blocks <IMG
+ WIDTH="14" HEIGHT="29" ALIGN="MIDDLE" BORDER="0"
+ SRC="img2.gif"
+ ALT="$\beta$">
+calculates two checksums:
+  a weak ``rolling'' 32-bit checksum (described below) and a strong
+  128-bit MD4 checksum.
+
+<P>
+<DT>3.
+<DD><IMG
+ WIDTH="14" HEIGHT="29" ALIGN="MIDDLE" BORDER="0"
+ SRC="img2.gif"
+ ALT="$\beta$">
+sends these checksums to <IMG
+ WIDTH="15" HEIGHT="13" ALIGN="BOTTOM" BORDER="0"
+ SRC="img1.gif"
+ ALT="$\alpha$">.
+  
+<DT>4.
+<DD><IMG
+ WIDTH="15" HEIGHT="13" ALIGN="BOTTOM" BORDER="0"
+ SRC="img1.gif"
+ ALT="$\alpha$">
+searches through <I>A</I> to find all blocks of length <I>S</I>  bytes (at any offset, not just multiples of <I>S</I>) that have the same
+  weak and strong checksum as one of the blocks of <I>B</I>. This can be
+  done in a single pass very quickly using a special property of the
+  rolling checksum described below.
+  
+<DT>5.
+<DD><IMG
+ WIDTH="15" HEIGHT="13" ALIGN="BOTTOM" BORDER="0"
+ SRC="img1.gif"
+ ALT="$\alpha$">
+sends <IMG
+ WIDTH="14" HEIGHT="29" ALIGN="MIDDLE" BORDER="0"
+ SRC="img2.gif"
+ ALT="$\beta$">
+a sequence of instructions for
+  constructing a copy of <I>A</I>.  Each instruction is either a reference
+  to a block of <I>B</I>, or literal data.  Literal data is sent only for
+  those sections of <I>A</I> which did not match any of the blocks of <I>B</I>.
+</DL>
+<P>
+The end result is that <IMG
+ WIDTH="14" HEIGHT="29" ALIGN="MIDDLE" BORDER="0"
+ SRC="img2.gif"
+ ALT="$\beta$">
+gets a copy of <I>A</I>, but only the pieces
+of <I>A</I> that are not found in <I>B</I> (plus a small amount of data for
+checksums and block indexes) are sent over the link. The algorithm
+also only requires one round trip, which minimises the impact of the
+link latency.
+
+<P>
+The most important details of the algorithm are the rolling checksum
+and the associated multi-alternate search mechanism which allows the
+all-offsets checksum search to proceed very quickly. These will be
+discussed in greater detail below.
+
+<P>
+<HR>
+<!--Navigation Panel-->
+<A NAME="tex2html32"
+ HREF="node3.html">
+<IMG WIDTH="37" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="next"
+ SRC="next.gif"></A> 
+<A NAME="tex2html30"
+ HREF="tech_report.html">
+<IMG WIDTH="26" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="up"
+ SRC="up.gif"></A> 
+<A NAME="tex2html24"
+ HREF="node1.html">
+<IMG WIDTH="63" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="previous"
+ SRC="previous.gif"></A>   
+<BR>
+<B> Next:</B> <A NAME="tex2html33"
+ HREF="node3.html">Rolling checksum</A>
+<B> Up:</B> <A NAME="tex2html31"
+ HREF="tech_report.html">The rsync algorithm</A>
+<B> Previous:</B> <A NAME="tex2html25"
+ HREF="node1.html">The problem</A>
+<!--End of Navigation Panel-->
+<ADDRESS>
+<I>Andrew Tridgell</I>
+<BR><I>1998-11-09</I>
+</ADDRESS>
+</BODY>
+</HTML>

rsync-web/tech_report/node3.html

@@ -0,0 +1,187 @@
+<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2 Final//EN">
+<!--Converted with LaTeX2HTML 98.1p1 release (March 2nd, 1998)
+originally by Nikos Drakos (nikos@cbl.leeds.ac.uk), CBLU, University of Leeds
+* revised and updated by:  Marcus Hennecke, Ross Moore, Herb Swan
+* with significant contributions from:
+  Jens Lippmann, Marek Rouchal, Martin Wilck and others -->
+<HTML>
+<HEAD>
+<TITLE>Rolling checksum</TITLE>
+<META NAME="description" CONTENT="Rolling checksum">
+<META NAME="keywords" CONTENT="tech_report">
+<META NAME="resource-type" CONTENT="document">
+<META NAME="distribution" CONTENT="global">
+<META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=iso-8859-1">
+<LINK REL="STYLESHEET" HREF="tech_report.css">
+<LINK REL="next" HREF="node4.html">
+<LINK REL="previous" HREF="node2.html">
+<LINK REL="up" HREF="tech_report.html">
+<LINK REL="next" HREF="node4.html">
+</HEAD>
+<BODY >
+<!--Navigation Panel-->
+<A NAME="tex2html42"
+ HREF="node4.html">
+<IMG WIDTH="37" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="next"
+ SRC="next.gif"></A> 
+<A NAME="tex2html40"
+ HREF="tech_report.html">
+<IMG WIDTH="26" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="up"
+ SRC="up.gif"></A> 
+<A NAME="tex2html34"
+ HREF="node2.html">
+<IMG WIDTH="63" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="previous"
+ SRC="previous.gif"></A>   
+<BR>
+<B> Next:</B> <A NAME="tex2html43"
+ HREF="node4.html">Checksum searching</A>
+<B> Up:</B> <A NAME="tex2html41"
+ HREF="tech_report.html">The rsync algorithm</A>
+<B> Previous:</B> <A NAME="tex2html35"
+ HREF="node2.html">The rsync algorithm</A>
+<BR>
+<BR>
+<!--End of Navigation Panel-->
+
+<H1><A NAME="SECTION00030000000000000000">
+Rolling checksum</A>
+</H1>
+
+<P>
+The weak rolling checksum used in the rsync algorithm needs to have
+the property that it is very cheap to calculate the checksum of a
+buffer 
+<!-- MATH: $X_2 .. X_{n+1}$ -->
+<I>X</I><SUB>2</SUB> .. <I>X</I><SUB><I>n</I>+1</SUB> given the checksum of buffer 
+<!-- MATH: $X_1 .. X_n$ -->
+<I>X</I><SUB>1</SUB> .. <I>X</I><SUB><I>n</I></SUB> and
+the values of the bytes <I>X</I><SUB>1</SUB> and <I>X</I><SUB><I>n</I>+1</SUB>.
+
+<P>
+The weak checksum algorithm we used in our implementation was inspired
+by Mark Adler's adler-32 checksum.  Our checksum is defined by
+<BR><P></P>
+<DIV ALIGN="CENTER">
+<!-- MATH: \begin{displaymath}
+a(k,l) = (\sum_{i=k}^l X_i) \bmod M
+\end{displaymath} -->
+
+
+<IMG
+ WIDTH="176" HEIGHT="56"
+ SRC="img3.gif"
+ ALT="\begin{displaymath}a(k,l) = (\sum_{i=k}^l X_i) \bmod M \end{displaymath}">
+</DIV>
+<BR CLEAR="ALL">
+<P></P>
+<BR><P></P>
+<DIV ALIGN="CENTER">
+<!-- MATH: \begin{displaymath}
+b(k,l) = (\sum_{i=k}^l (l-i+1)X_i) \bmod M
+\end{displaymath} -->
+
+
+<IMG
+ WIDTH="240" HEIGHT="56"
+ SRC="img4.gif"
+ ALT="\begin{displaymath}b(k,l) = (\sum_{i=k}^l (l-i+1)X_i) \bmod M \end{displaymath}">
+</DIV>
+<BR CLEAR="ALL">
+<P></P>
+<BR><P></P>
+<DIV ALIGN="CENTER">
+<!-- MATH: \begin{displaymath}
+s(k,l) = a(k,l) + 2^{16} b(k,l)
+\end{displaymath} -->
+
+
+<I>s</I>(<I>k</I>,<I>l</I>) = <I>a</I>(<I>k</I>,<I>l</I>) + 2<SUP>16</SUP> <I>b</I>(<I>k</I>,<I>l</I>)
+</DIV>
+<BR CLEAR="ALL">
+<P></P>
+<P>
+where <I>s</I>(<I>k</I>,<I>l</I>) is the rolling checksum of the bytes 
+<!-- MATH: $X_k \ldots X_l$ -->
+<IMG
+ WIDTH="67" HEIGHT="29" ALIGN="MIDDLE" BORDER="0"
+ SRC="img5.gif"
+ ALT="$X_k \ldots X_l$">.
+For simplicity and speed, we use 
+<!-- MATH: $M = 2^{16}$ -->
+<I>M</I> = 2<SUP>16</SUP>.  
+
+<P>
+The important property of this checksum is that successive values can
+be computed very efficiently using the recurrence relations
+
+<P>
+<BR><P></P>
+<DIV ALIGN="CENTER">
+<!-- MATH: \begin{displaymath}
+a(k+1,l+1) = (a(k,l) - X_k + X_{l+1}) \bmod M
+\end{displaymath} -->
+
+
+<IMG
+ WIDTH="322" HEIGHT="28"
+ SRC="img6.gif"
+ ALT="\begin{displaymath}a(k+1,l+1) = (a(k,l) - X_k + X_{l+1}) \bmod M \end{displaymath}">
+</DIV>
+<BR CLEAR="ALL">
+<P></P>
+<BR><P></P>
+<DIV ALIGN="CENTER">
+<!-- MATH: \begin{displaymath}
+b(k+1,l+1) = (b(k,l) - (l-k+1) X_k + a(k+1,l+1)) \bmod M
+\end{displaymath} -->
+
+
+<IMG
+ WIDTH="454" HEIGHT="28"
+ SRC="img7.gif"
+ ALT="\begin{displaymath}b(k+1,l+1) = (b(k,l) - (l-k+1) X_k + a(k+1,l+1)) \bmod M \end{displaymath}">
+</DIV>
+<BR CLEAR="ALL">
+<P></P>
+<P>
+Thus the checksum can be calculated for blocks of length S at all
+possible offsets within a file in a ``rolling'' fashion, with very
+little computation at each point.
+
+<P>
+Despite its simplicity, this checksum was found to be quite adequate as
+a first level check for a match of two file blocks.  We have found in
+practice that the probability of this checksum matching when the
+blocks are not equal is quite low.  This is important because the much
+more expensive strong checksum must be calculated for each block where
+the weak checksum matches.
+
+<P>
+<HR>
+<!--Navigation Panel-->
+<A NAME="tex2html42"
+ HREF="node4.html">
+<IMG WIDTH="37" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="next"
+ SRC="next.gif"></A> 
+<A NAME="tex2html40"
+ HREF="tech_report.html">
+<IMG WIDTH="26" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="up"
+ SRC="up.gif"></A> 
+<A NAME="tex2html34"
+ HREF="node2.html">
+<IMG WIDTH="63" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="previous"
+ SRC="previous.gif"></A>   
+<BR>
+<B> Next:</B> <A NAME="tex2html43"
+ HREF="node4.html">Checksum searching</A>
+<B> Up:</B> <A NAME="tex2html41"
+ HREF="tech_report.html">The rsync algorithm</A>
+<B> Previous:</B> <A NAME="tex2html35"
+ HREF="node2.html">The rsync algorithm</A>
+<!--End of Navigation Panel-->
+<ADDRESS>
+<I>Andrew Tridgell</I>
+<BR><I>1998-11-09</I>
+</ADDRESS>
+</BODY>
+</HTML>

rsync-web/tech_report/node4.html

@@ -0,0 +1,145 @@
+<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2 Final//EN">
+<!--Converted with LaTeX2HTML 98.1p1 release (March 2nd, 1998)
+originally by Nikos Drakos (nikos@cbl.leeds.ac.uk), CBLU, University of Leeds
+* revised and updated by:  Marcus Hennecke, Ross Moore, Herb Swan
+* with significant contributions from:
+  Jens Lippmann, Marek Rouchal, Martin Wilck and others -->
+<HTML>
+<HEAD>
+<TITLE>Checksum searching</TITLE>
+<META NAME="description" CONTENT="Checksum searching">
+<META NAME="keywords" CONTENT="tech_report">
+<META NAME="resource-type" CONTENT="document">
+<META NAME="distribution" CONTENT="global">
+<META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=iso-8859-1">
+<LINK REL="STYLESHEET" HREF="tech_report.css">
+<LINK REL="next" HREF="node5.html">
+<LINK REL="previous" HREF="node3.html">
+<LINK REL="up" HREF="tech_report.html">
+<LINK REL="next" HREF="node5.html">
+</HEAD>
+<BODY >
+<!--Navigation Panel-->
+<A NAME="tex2html52"
+ HREF="node5.html">
+<IMG WIDTH="37" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="next"
+ SRC="next.gif"></A> 
+<A NAME="tex2html50"
+ HREF="tech_report.html">
+<IMG WIDTH="26" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="up"
+ SRC="up.gif"></A> 
+<A NAME="tex2html44"
+ HREF="node3.html">
+<IMG WIDTH="63" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="previous"
+ SRC="previous.gif"></A>   
+<BR>
+<B> Next:</B> <A NAME="tex2html53"
+ HREF="node5.html">Pipelining</A>
+<B> Up:</B> <A NAME="tex2html51"
+ HREF="tech_report.html">The rsync algorithm</A>
+<B> Previous:</B> <A NAME="tex2html45"
+ HREF="node3.html">Rolling checksum</A>
+<BR>
+<BR>
+<!--End of Navigation Panel-->
+
+<H1><A NAME="SECTION00040000000000000000">
+Checksum searching</A>
+</H1>
+
+<P>
+Once <IMG
+ WIDTH="15" HEIGHT="13" ALIGN="BOTTOM" BORDER="0"
+ SRC="img1.gif"
+ ALT="$\alpha$">
+has received the list of checksums of the blocks of <I>B</I>,
+it must search <I>A</I> for any blocks at any offset that match the
+checksum of some block of <I>B</I>.  The basic strategy is to compute the
+32-bit rolling checksum for a block of length <I>S</I> starting at each
+byte of <I>A</I> in turn, and for each checksum, search the list for a
+match.  To do this our implementation uses a
+simple 3 level searching scheme.
+
+<P>
+The first level uses a 16-bit hash of the 32-bit rolling checksum and
+a 2<SUP>16</SUP> entry hash table.  The list of checksum values (i.e., the
+checksums from the blocks of <I>B</I>) is sorted according to the 16-bit
+hash of the 32-bit rolling checksum.  Each entry in the hash table
+points to the first element of the list for that hash value, or
+contains a null value if no element of the list has that hash value.
+
+<P>
+At each offset in the file the 32-bit rolling checksum and its 16-bit
+hash are calculated.  If the hash table entry for that hash value is
+not a null value, the second level check is invoked.
+
+<P>
+The second level check involves scanning the sorted checksum list
+starting with the entry pointed to by the hash table entry, looking
+for an entry whose 32-bit rolling checksum matches the current value.
+The scan terminates when it reaches an entry whose 16-bit hash
+differs.  If this search finds a match, the third level check is
+invoked.
+
+<P>
+The third level check involves calculating the strong checksum for the
+current offset in the file and comparing it with the strong checksum
+value in the current list entry.  If the two strong checksums match,
+we assume that we have found a block of <I>A</I> which matches a block of
+<I>B</I>.  In fact the blocks could be different, but the probability of
+this is microscopic, and in practice this is a reasonable assumption.
+
+<P>
+When a match is found, <IMG
+ WIDTH="15" HEIGHT="13" ALIGN="BOTTOM" BORDER="0"
+ SRC="img1.gif"
+ ALT="$\alpha$">
+sends <IMG
+ WIDTH="14" HEIGHT="29" ALIGN="MIDDLE" BORDER="0"
+ SRC="img2.gif"
+ ALT="$\beta$">
+the data in <I>A</I> between
+the current file offset and the end of the previous match, followed by
+the index of the block in <I>B</I> that matched.  This data is sent
+immediately a match is found, which allows us to overlap the
+communication with further computation.
+
+<P>
+If no match is found at a given offset in the file, the rolling
+checksum is updated to the next offset and the search proceeds.  If a
+match is found, the search is restarted at the end of the matched
+block.  This strategy saves a considerable amount of computation for
+the common case where the two files are nearly identical.  In
+addition, it would be a simple matter to encode the block indexes as
+runs, for the common case where a portion of <I>A</I> matches a series of
+blocks of <I>B</I> in order.
+
+<P>
+<HR>
+<!--Navigation Panel-->
+<A NAME="tex2html52"
+ HREF="node5.html">
+<IMG WIDTH="37" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="next"
+ SRC="next.gif"></A> 
+<A NAME="tex2html50"
+ HREF="tech_report.html">
+<IMG WIDTH="26" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="up"
+ SRC="up.gif"></A> 
+<A NAME="tex2html44"
+ HREF="node3.html">
+<IMG WIDTH="63" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="previous"
+ SRC="previous.gif"></A>   
+<BR>
+<B> Next:</B> <A NAME="tex2html53"
+ HREF="node5.html">Pipelining</A>
+<B> Up:</B> <A NAME="tex2html51"
+ HREF="tech_report.html">The rsync algorithm</A>
+<B> Previous:</B> <A NAME="tex2html45"
+ HREF="node3.html">Rolling checksum</A>
+<!--End of Navigation Panel-->
+<ADDRESS>
+<I>Andrew Tridgell</I>
+<BR><I>1998-11-09</I>
+</ADDRESS>
+</BODY>
+</HTML>

rsync-web/tech_report/node5.html

@@ -0,0 +1,85 @@
+<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2 Final//EN">
+<!--Converted with LaTeX2HTML 98.1p1 release (March 2nd, 1998)
+originally by Nikos Drakos (nikos@cbl.leeds.ac.uk), CBLU, University of Leeds
+* revised and updated by:  Marcus Hennecke, Ross Moore, Herb Swan
+* with significant contributions from:
+  Jens Lippmann, Marek Rouchal, Martin Wilck and others -->
+<HTML>
+<HEAD>
+<TITLE>Pipelining</TITLE>
+<META NAME="description" CONTENT="Pipelining">
+<META NAME="keywords" CONTENT="tech_report">
+<META NAME="resource-type" CONTENT="document">
+<META NAME="distribution" CONTENT="global">
+<META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=iso-8859-1">
+<LINK REL="STYLESHEET" HREF="tech_report.css">
+<LINK REL="next" HREF="node6.html">
+<LINK REL="previous" HREF="node4.html">
+<LINK REL="up" HREF="tech_report.html">
+<LINK REL="next" HREF="node6.html">
+</HEAD>
+<BODY >
+<!--Navigation Panel-->
+<A NAME="tex2html62"
+ HREF="node6.html">
+<IMG WIDTH="37" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="next"
+ SRC="next.gif"></A> 
+<A NAME="tex2html60"
+ HREF="tech_report.html">
+<IMG WIDTH="26" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="up"
+ SRC="up.gif"></A> 
+<A NAME="tex2html54"
+ HREF="node4.html">
+<IMG WIDTH="63" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="previous"
+ SRC="previous.gif"></A>   
+<BR>
+<B> Next:</B> <A NAME="tex2html63"
+ HREF="node6.html">Results</A>
+<B> Up:</B> <A NAME="tex2html61"
+ HREF="tech_report.html">The rsync algorithm</A>
+<B> Previous:</B> <A NAME="tex2html55"
+ HREF="node4.html">Checksum searching</A>
+<BR>
+<BR>
+<!--End of Navigation Panel-->
+
+<H1><A NAME="SECTION00050000000000000000">
+Pipelining</A>
+</H1>
+
+<P>
+The above sections describe the process for constructing a copy of one
+file on a remote system.  If we have a several files to copy, we can
+gain a considerable latency advantage by pipelining the process.
+
+<P>
+This involves <IMG
+ WIDTH="14" HEIGHT="29" ALIGN="MIDDLE" BORDER="0"
+ SRC="img2.gif"
+ ALT="$\beta$">
+initiating two independent processes. One of the
+processes generates and sends the checksums to <IMG
+ WIDTH="15" HEIGHT="13" ALIGN="BOTTOM" BORDER="0"
+ SRC="img1.gif"
+ ALT="$\alpha$">
+while the
+other receives the difference information from <IMG
+ WIDTH="15" HEIGHT="13" ALIGN="BOTTOM" BORDER="0"
+ SRC="img1.gif"
+ ALT="$\alpha$">
+and
+reconstructs the files. 
+
+<P>
+If the communications link is buffered then these two processes can
+proceed independently and the link should be kept fully utilised in
+both directions for most of the time. 
+
+<P>
+<BR><HR>
+<ADDRESS>
+<I>Andrew Tridgell</I>
+<BR><I>1998-11-09</I>
+</ADDRESS>
+</BODY>
+</HTML>

rsync-web/tech_report/node6.html

@@ -0,0 +1,299 @@
+<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2 Final//EN">
+<!--Converted with LaTeX2HTML 98.1p1 release (March 2nd, 1998)
+originally by Nikos Drakos (nikos@cbl.leeds.ac.uk), CBLU, University of Leeds
+* revised and updated by:  Marcus Hennecke, Ross Moore, Herb Swan
+* with significant contributions from:
+  Jens Lippmann, Marek Rouchal, Martin Wilck and others -->
+<HTML>
+<HEAD>
+<TITLE>Results</TITLE>
+<META NAME="description" CONTENT="Results">
+<META NAME="keywords" CONTENT="tech_report">
+<META NAME="resource-type" CONTENT="document">
+<META NAME="distribution" CONTENT="global">
+<META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=iso-8859-1">
+<LINK REL="STYLESHEET" HREF="tech_report.css">
+<LINK REL="next" HREF="node7.html">
+<LINK REL="previous" HREF="node5.html">
+<LINK REL="up" HREF="tech_report.html">
+<LINK REL="next" HREF="node7.html">
+</HEAD>
+<BODY >
+<!--Navigation Panel-->
+<A NAME="tex2html72"
+ HREF="node7.html">
+<IMG WIDTH="37" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="next"
+ SRC="next.gif"></A> 
+<A NAME="tex2html70"
+ HREF="tech_report.html">
+<IMG WIDTH="26" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="up"
+ SRC="up.gif"></A> 
+<A NAME="tex2html64"
+ HREF="node5.html">
+<IMG WIDTH="63" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="previous"
+ SRC="previous.gif"></A>   
+<BR>
+<B> Next:</B> <A NAME="tex2html73"
+ HREF="node7.html">Availability</A>
+<B> Up:</B> <A NAME="tex2html71"
+ HREF="tech_report.html">The rsync algorithm</A>
+<B> Previous:</B> <A NAME="tex2html65"
+ HREF="node5.html">Pipelining</A>
+<BR>
+<BR>
+<!--End of Navigation Panel-->
+
+<H1><A NAME="SECTION00060000000000000000">
+Results</A>
+</H1>
+
+<P>
+To test the algorithm, tar files were created of the Linux kernel
+sources for two versions of the kernel. The two kernel versions were
+1.99.10 and 2.0.0. These tar files are approximately 24MB in size and
+are separated by 5 released patch levels.
+
+<P>
+Out of the 2441 files in the 1.99.10 release, 291 files had changed in
+the 2.0.0 release, 19 files had been removed and 25 files had been
+added.
+
+<P>
+A ``diff'' of the two tar files using the standard GNU diff utility
+produced over 32 thousand lines of output totalling 2.1 MB.
+
+<P>
+The following table shows the results for rsync between the two files
+with a varying block size.<A NAME="tex2html2"
+ HREF="footnode.html#foot24"><SUP>2</SUP></A>
+
+<P>
+<BR>
+<BR>
+<TABLE CELLPADDING=3 BORDER="1">
+<TR><TD ALIGN="LEFT"><B> block</B></TD>
+<TD ALIGN="LEFT"><B> matches</B></TD>
+<TD ALIGN="LEFT"><B> tag</B></TD>
+<TD ALIGN="LEFT"><B> false</B></TD>
+<TD ALIGN="LEFT"><B> data</B></TD>
+<TD ALIGN="LEFT"><B> written</B></TD>
+<TD ALIGN="LEFT"><B> read</B></TD>
+</TR>
+<TR><TD ALIGN="LEFT"><B> size</B></TD>
+<TD ALIGN="LEFT">&nbsp;</TD>
+<TD ALIGN="LEFT"><B> hits</B></TD>
+<TD ALIGN="LEFT"><B> alarms</B></TD>
+<TD ALIGN="LEFT">&nbsp;</TD>
+<TD ALIGN="LEFT">&nbsp;</TD>
+<TD ALIGN="LEFT">&nbsp;</TD>
+</TR>
+<TR><TD ALIGN="LEFT"><P>
+300</TD>
+<TD ALIGN="LEFT">64247</TD>
+<TD ALIGN="LEFT">3817434</TD>
+<TD ALIGN="LEFT">948</TD>
+<TD ALIGN="LEFT">5312200</TD>
+<TD ALIGN="LEFT">5629158</TD>
+<TD ALIGN="LEFT">1632284</TD>
+</TR>
+<TR><TD ALIGN="LEFT">500</TD>
+<TD ALIGN="LEFT">46989</TD>
+<TD ALIGN="LEFT">620013</TD>
+<TD ALIGN="LEFT">64</TD>
+<TD ALIGN="LEFT">1091900</TD>
+<TD ALIGN="LEFT">1283906</TD>
+<TD ALIGN="LEFT">979384</TD>
+</TR>
+<TR><TD ALIGN="LEFT">700</TD>
+<TD ALIGN="LEFT">33255</TD>
+<TD ALIGN="LEFT">571970</TD>
+<TD ALIGN="LEFT">22</TD>
+<TD ALIGN="LEFT">1307800</TD>
+<TD ALIGN="LEFT">1444346</TD>
+<TD ALIGN="LEFT">699564</TD>
+</TR>
+<TR><TD ALIGN="LEFT">900</TD>
+<TD ALIGN="LEFT">25686</TD>
+<TD ALIGN="LEFT">525058</TD>
+<TD ALIGN="LEFT">24</TD>
+<TD ALIGN="LEFT">1469500</TD>
+<TD ALIGN="LEFT">1575438</TD>
+<TD ALIGN="LEFT">544124</TD>
+</TR>
+<TR><TD ALIGN="LEFT">1100</TD>
+<TD ALIGN="LEFT">20848</TD>
+<TD ALIGN="LEFT">496844</TD>
+<TD ALIGN="LEFT">21</TD>
+<TD ALIGN="LEFT">1654500</TD>
+<TD ALIGN="LEFT">1740838</TD>
+<TD ALIGN="LEFT">445204</TD>
+</TR>
+</TABLE>
+<BR>
+<BR>
+
+<P>
+In each case, the CPU time taken was less than the
+time it takes to run ``diff'' on the two files.<A NAME="tex2html3"
+ HREF="footnode.html#foot40"><SUP>3</SUP></A>
+
+<P>
+The columns in the table are as follows:
+
+<P>
+<DL>
+<DT><STRONG>block size</STRONG>
+<DD>The size in bytes of the checksummed blocks.
+<DT><STRONG>matches</STRONG>
+<DD>The number of times a block of <I>B</I> was found in <I>A</I>.
+<DT><STRONG>tag hits</STRONG>
+<DD>The number of times the 16 bit hash of the rolling
+checksum matched a hash of one of the checksums from <I>B</I>.
+<DT><STRONG>false alarms</STRONG>
+<DD>The number of times the 32 bit rolling checksum
+  matched but the strong checksum didn't.
+<DT><STRONG>data</STRONG>
+<DD>The amount of file data transferred verbatim, in bytes.
+<DT><STRONG>written</STRONG>
+<DD>The total number of bytes written by <IMG
+ WIDTH="15" HEIGHT="13" ALIGN="BOTTOM" BORDER="0"
+ SRC="img1.gif"
+ ALT="$\alpha$">
+including protocol overheads. This is almost all file data.
+<DT><STRONG>read</STRONG>
+<DD>The total number of bytes read by <IMG
+ WIDTH="15" HEIGHT="13" ALIGN="BOTTOM" BORDER="0"
+ SRC="img1.gif"
+ ALT="$\alpha$">
+including
+  protocol overheads. This is almost all checksum information.
+</DL>
+<P>
+The results demonstrate that for block sizes above 300 bytes, only a
+small fraction (around 5%) of the file was transferred. The amount
+transferred was also considerably less than the size of the diff file
+that would have been transferred if the diff/patch method of updating
+a remote file was used.
+
+<P>
+The checksums themselves took up a considerable amount of space,
+although much less than the size of the data transferred in each
+case. Each pair of checksums consumes 20 bytes: 4 bytes for the
+rolling checksum plus 16 bytes for the 128-bit MD4 checksum.
+
+<P>
+The number of false alarms was less than 1/1000 of the number of
+true matches, indicating that the 32 bit rolling checksum is quite
+good at screening out false matches. 
+
+<P>
+The number of tag hits indicates that the second level of the
+checksum search algorithm was invoked about once every 50
+characters.  This is quite high because the total number of blocks in
+the file is a large fraction of the size of the tag hash table. For
+smaller files we would expect the tag hit rate to be much closer to
+the number of matches.  For extremely large files, we should probably
+increase the size of the hash table.
+
+<P>
+The next table shows similar results for a much smaller set of files.
+In this case the files were not packed into a tar file first. Rather,
+rsync was invoked with an option to recursively descend the directory
+tree. The files used were from two source releases of another software
+package called Samba. The total source code size is 1.7 MB and the
+diff between the two releases is 4155 lines long totalling 120 kB.
+
+<P>
+<BR>
+<BR>
+<TABLE CELLPADDING=3 BORDER="1">
+<TR><TD ALIGN="LEFT"><B> block</B></TD>
+<TD ALIGN="LEFT"><B> matches</B></TD>
+<TD ALIGN="LEFT"><B> tag</B></TD>
+<TD ALIGN="LEFT"><B> false</B></TD>
+<TD ALIGN="LEFT"><B> data</B></TD>
+<TD ALIGN="LEFT"><B> written</B></TD>
+<TD ALIGN="LEFT"><B> read</B></TD>
+</TR>
+<TR><TD ALIGN="LEFT"><B> size</B></TD>
+<TD ALIGN="LEFT">&nbsp;</TD>
+<TD ALIGN="LEFT"><B> hits</B></TD>
+<TD ALIGN="LEFT"><B> alarms</B></TD>
+<TD ALIGN="LEFT">&nbsp;</TD>
+<TD ALIGN="LEFT">&nbsp;</TD>
+<TD ALIGN="LEFT">&nbsp;</TD>
+</TR>
+<TR><TD ALIGN="LEFT"><P>
+300</TD>
+<TD ALIGN="LEFT">3727</TD>
+<TD ALIGN="LEFT">3899</TD>
+<TD ALIGN="LEFT">0</TD>
+<TD ALIGN="LEFT">129775</TD>
+<TD ALIGN="LEFT">153999</TD>
+<TD ALIGN="LEFT">83948</TD>
+</TR>
+<TR><TD ALIGN="LEFT">500</TD>
+<TD ALIGN="LEFT">2158</TD>
+<TD ALIGN="LEFT">2325</TD>
+<TD ALIGN="LEFT">0</TD>
+<TD ALIGN="LEFT">171574</TD>
+<TD ALIGN="LEFT">189330</TD>
+<TD ALIGN="LEFT">50908</TD>
+</TR>
+<TR><TD ALIGN="LEFT">700</TD>
+<TD ALIGN="LEFT">1517</TD>
+<TD ALIGN="LEFT">1649</TD>
+<TD ALIGN="LEFT">0</TD>
+<TD ALIGN="LEFT">195024</TD>
+<TD ALIGN="LEFT">210144</TD>
+<TD ALIGN="LEFT">36828</TD>
+</TR>
+<TR><TD ALIGN="LEFT">900</TD>
+<TD ALIGN="LEFT">1156</TD>
+<TD ALIGN="LEFT">1281</TD>
+<TD ALIGN="LEFT">0</TD>
+<TD ALIGN="LEFT">222847</TD>
+<TD ALIGN="LEFT">236471</TD>
+<TD ALIGN="LEFT">29048</TD>
+</TR>
+<TR><TD ALIGN="LEFT">1100</TD>
+<TD ALIGN="LEFT">921</TD>
+<TD ALIGN="LEFT">1049</TD>
+<TD ALIGN="LEFT">0</TD>
+<TD ALIGN="LEFT">250073</TD>
+<TD ALIGN="LEFT">262725</TD>
+<TD ALIGN="LEFT">23988</TD>
+</TR>
+</TABLE>
+<BR>
+<BR>
+
+<P>
+<HR>
+<!--Navigation Panel-->
+<A NAME="tex2html72"
+ HREF="node7.html">
+<IMG WIDTH="37" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="next"
+ SRC="next.gif"></A> 
+<A NAME="tex2html70"
+ HREF="tech_report.html">
+<IMG WIDTH="26" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="up"
+ SRC="up.gif"></A> 
+<A NAME="tex2html64"
+ HREF="node5.html">
+<IMG WIDTH="63" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="previous"
+ SRC="previous.gif"></A>   
+<BR>
+<B> Next:</B> <A NAME="tex2html73"
+ HREF="node7.html">Availability</A>
+<B> Up:</B> <A NAME="tex2html71"
+ HREF="tech_report.html">The rsync algorithm</A>
+<B> Previous:</B> <A NAME="tex2html65"
+ HREF="node5.html">Pipelining</A>
+<!--End of Navigation Panel-->
+<ADDRESS>
+<I>Andrew Tridgell</I>
+<BR><I>1998-11-09</I>
+</ADDRESS>
+</BODY>
+</HTML>