* encode: flush headers immediately for server-sent events responses
The encode middleware withholds the response header until the first body
write so it can sniff content-type and apply the minimum_length threshold.
For a text/event-stream response the upstream typically writes headers and
flushes to establish the event stream before any event body is available,
so the client never received the headers and the stream stalled; the same
buffering also delayed individual events.
When WriteHeader sees a text/event-stream content type, initialize encoding
and write the header through immediately. Forcing the header out also marks
the response as started, so subsequent event writes bypass the minimum_length
buffering and stream to the client as they arrive.
Fixes#6293
* encode: add WriteHeader benchmark covering SSE fast path
* encode: replace mime.ParseMediaType with bound-checked SSE check
WriteHeader runs an SSE Content-Type check on every call once headers
haven't been written yet. mime.ParseMediaType parses the full media
type, including parameters, even when nothing matches, which shows up
on the hot header-write path.
Replace it with a bound-checked manual prefix/boundary check (isSSE),
skipping parameter parsing for the common non-SSE case.
* encode: reject content types with junk after text/event-stream
isSSE accepted any suffix after a space, so a value like
"text/event-stream nonsense" was treated as an SSE response. After the
media type, skip optional whitespace and require either the end of the
value or a parameter separator. The check remains allocation-free, so
the hot-path motivation for the manual matcher is preserved.
---------
Co-authored-by: SillyZir <269283839+SillyZir@users.noreply.github.com>
Co-authored-by: Kévin Dunglas <kevin@les-tilleuls.coop>
AcceptedEncodings runs for every request when response compression is enabled,
so its allocations contribute directly to per-request garbage and GC pressure.
While the gains are modest, this cuts allocations by two-thirds and bytes
nearly in half on a hot path that executes on every compressed request. Fewer
allocations means less work for the garbage collector, which at high request
rates translates into lower GC frequency and steadier tail latency, not just
faster execution of this one function in isolation.
For each token in the Accept-Encoding header it called strings.Split(accepted,
";"), allocating a throwaway slice per encoding, and grew the prefs slice from
empty. This commit replaces the per-token Split with strings.Cut (zero
allocation) and presize prefs from the comma count.
As requested per the policy, this commit adds a benchmark to exercise the
change. With header "gzip, deflate, br;q=0.9, zstd;q=0.8":
AcceptedEncodings
- sec/op 1158.5n -> 771.8n (-33.38%)
- B/op 408 -> 216 (-47.06%)
- allocs/op 9 -> 3 (-66.67%)
Co-authored-by: @dunglas
Header.Get/Set re-canonicalize and allocate whenever the passed key
isn't already in canonical MIME header form. Sec-WebSocket-Key,
WWW-Authenticate, and content-type all miss the fast path; switch to
their canonical forms (Sec-Websocket-Key, Www-Authenticate,
Content-Type).
Enable the canonicalheader linter to catch future regressions,
excluded in test files since those already use non-canonical casing
in several places without a perf-sensitive path behind them.
* fix(encode): prioritize zstd and br over gzip in content negotiation
* test(encode): update unit tests to reflect new default priority ties
* fix(encode): move default preferences to dynamic encode handler and restore generic negotiation helper
* test(encode): call real Provision function in served-response test
* test(encode): rename served-response test to TestServeHTTPDefaultEncodingPreference
* refactor(encode): use slices.SortStableFunc and httptest.NewRecorder as recommended
* refactor(encode): simplify sorting with cmp.Compare and check request error in test
* test(encode): fix variable redeclaration in TestServeHTTPDefaultEncodingPreference
Fix 'no new variables on left side of :=' error by changing 'err :=' to 'err ='
on line 347, since err was already declared on line 332.
This fixes the build failure in the encode module tests.
* feat: add bcrypt cost parameter to hash-password
* revert: typos
* refactor: take the cost out of interface
* fix: default bcrypt cost to 14
* fix: follow bcrypt library for min and max cost
* doc: mention defaultBcryptCost in cost parameter description
* chore: gci format
* fix: more specific bcrypt cost algorithm flag
* feat: bcrypt cost provisioning
* Revert "feat: bcrypt cost provisioning"
This reverts commit e09d4bd036.
* chore: gci format
* chore: gci format
* chore: gci format
* chore: golangcilint fmted
---------
Co-authored-by: Mohammed Al Sahaf <msaa1990@gmail.com>
* encode: write status immediate for success response for CONNECT requests
* fix compile
* fix test
* fix lint
* treat first write and flush for encode response writer to CONNECT request as success if status is not set explicitly
* Add zstd compression level support
* Refactored zstd levels to string arguments
fastest, default, better, best
* Add comment with list of all available levels
* Corrected data types for config
---------
Co-authored-by: Evgeny Blinov <e.a.blinov@gmail.com>
Co-authored-by: Matt Holt <mholt@users.noreply.github.com>
* use gofmput to format code
* use gci to format imports
* reconfigure gci
* linter autofixes
* rearrange imports a little
* export GOOS=windows golangci-lint run ./... --fix
* caddyhttp: Make use of http.ResponseController
Also syncs the reverseproxy implementation with stdlib's which now uses ResponseController as well https://github.com/golang/go/commit/2449bbb5e614954ce9e99c8a481ea2ee73d72d61
* Enable full-duplex for HTTP/1.1
* Appease linter
* Add warning for builds with Go 1.20, so it's less surprising to users
* Improved godoc for EnableFullDuplex, copied text from stdlib
* Only wrap in encode if not already wrapped
* reverseproxy: Mask the WS close message when we're the client
* weakrand
* Bump golangci-lint version so path ignores work on Windows
* gofmt
* ugh, gofmt everything, I guess
* fix encode handler header manipulation
also avoid implementing ReadFrom because it breaks when io.Copied to directly
* strconv.Itoa should be tried as a last resort
WriteHeader during Close
* encode: ignore flushing until after first write (fix#4314)
The first write will determine if encoding has to be done and will add an Content-Encoding. Until then Flushing has to be delayed so the Content-Encoding header can be added before headers and status code is written. (A passthrough flush would write header and status code)
* Update modules/caddyhttp/encode/encode.go
Co-authored-by: Matt Holt <mholt@users.noreply.github.com>
* Tweak compression settings
zstd: Limit window sizes to 128K to keep memory in control both server and client size.
zstd: Write 0 length frames. This may be needed for compatibility.
zstd: Create fewer encoders. Small memory improvement.
gzip: Allow -2 (Huffman only) and -3 (stateless) compression modes.
* Update modules/caddyhttp/encode/zstd/zstd.go
Update docs.
Co-authored-by: Francis Lavoie <lavofr@gmail.com>
Co-authored-by: Francis Lavoie <lavofr@gmail.com>
Followup to #4150, #4151 /cc @ueffel @polarathene
After a bit of discussion with @mholt, we decided to remove `prefer` as a subdirective and just go with using the order implicitly always. Simpler config, simpler docs, etc.
Effectively changes 7776471 and reverts a small part of f35a7fa.
