Jeroen van Rijn
6cfdf2bb3a
Merge pull request #7797 from corleypc/align-regression-tests
...
Adds alignment related regression tests
2026-10-06 15:55:21 -07:00
corley
8b573fcf5e
adds alignment regression tests
2026-10-07 01:38:31 +03:00
Jeroen van Rijn
ed3dc46dfb
Merge pull request #7795 from corleypc/temp-cpy-fix
...
Fixes regressions caused by temp copy of underaligned values
2026-10-06 11:57:13 -07:00
Jeroen van Rijn
57528bfa70
Merge pull request #7796 from karl-zylinski/parapoly-group-enumerated-array-fix
...
Parapoly: Fix regression with proc group containing parapoly enumerated array specialization
2026-10-06 11:54:21 -07:00
Karl Zylinski
05d4c79dc5
Parapoly: Fix regression where proc group containing parapoly enumerated array specialization was marked as ambigious.
2026-10-06 20:42:29 +02:00
gingerBill
072c6bc227
Merge pull request #7759 from tibret/box3d-sethull-bug
...
Changing #by_ptr to a pointer to make sure box3d gets long-lived poin…
2026-10-06 18:24:39 +01:00
Mihail Moskov
55c6cf3dea
fixes regressions at sites needing true address caused by temp copy
2026-10-06 20:21:05 +03:00
gingerBill
0de10624fe
Name single-module output after the main package when -out is a directory
2026-10-06 18:14:29 +01:00
gingerBill
3956328809
Merge branch 'master' of https://github.com/odin-lang/Odin
2026-10-06 17:42:52 +01:00
gingerBill
0a076fa8fe
Fix nested procedure types in results losing names
2026-10-06 17:42:40 +01:00
gingerBill
25946eb2f1
Merge pull request #7789 from Barinzaya/bounds_check_with_context
...
Context-aware bounds check handling
2026-10-06 15:55:45 +01:00
Jeroen van Rijn
84bc3fc210
Merge pull request #7788 from A1029384756/windows-cross-compile
...
[windows-cross-compile] add missing `help` message
2026-10-06 07:08:36 -07:00
A1029384756
1c8600caf2
[windows-cross-compile] add missing help message
2026-10-06 09:59:07 -04:00
Barinzaya
8075b343c5
Split bounds check handlers into contextless and with_context.
...
When a `context` is available, the `with_context` variant will be called
and the `context.assertion_failure_proc` will be called, instead of
forcibly trapping. This allows bounds check errors able to be handled by
custom assertion failure procs.
2026-10-06 09:45:36 -04:00
gingerBill
ef46cfe326
Add vendor:directx/d3d_compiler to examples/all
2026-10-06 14:22:38 +01:00
gingerBill
b924d12fd9
Fix bug in the docs writer
2026-10-06 14:18:20 +01:00
gingerBill
e28e60a80c
Point "Expected a comma/semicolon, got a newline" errors just after the previous token
2026-10-05 21:49:35 +01:00
Rob Luppi
f8749e08e0
Changing #by_ptr to a pointer to avoid potential bugs when copying the HullData struct #7612
...
In the example given in #7612 the issue was caused by the lines
box_hull2 := b3d.MakeBoxHull(1, 1, 1)
b3d.Shape_SetHull(box_shape, box_hull2.base)
There is a note about potential errors during copy in the box3d code
that seems to be causing the issue
/// A convex hull.
/// @note This data structure has data hanging off the end and cannot be
directly copied.
typedef struct b3HullData
update to include Shape_SetMesh as well
2026-10-05 15:48:47 -04:00
gingerBill
3300b2c35a
Fix typo with linker section name problem
2026-10-05 18:48:25 +01:00
gingerBill
3b700c4656
Fix merge conflict with the same code
2026-10-05 18:47:24 +01:00
gingerBill
e3eaad0e98
Merge pull request #7747 from odin-lang/bill/llvm-backend-improvements
...
Compiler Performance Improvements to Reduce Compile Times Part 3
2026-10-05 18:40:53 +01:00
gingerBill
579bc065ea
Merge branch 'master' into bill/llvm-backend-improvements
2026-10-05 18:40:31 +01:00
gingerBill
20b3fdcd58
Accept an explicit -linker:radlink on Windows again, write no BOM for the default radlink's response file, and only require link.exe for -linker:msvc
2026-10-05 18:27:58 +01:00
gingerBill
881ad21336
Walk the global graph's declarations in small batches taken by a task for each thread, and fill its edges from the walk in parallel rather than through a serial list of edges
2026-10-05 17:26:41 +01:00
gingerBill
8c4d65289e
Write each package's doc entries in the natural order of their names to make the output deterministic
2026-10-05 16:59:58 +01:00
gingerBill
811658f915
Collect a global foreign block with its file when its attribute values are literals.
2026-10-05 16:55:03 +01:00
gingerBill
0e26e26a0d
List each file's possible type aliases as they are collected, gather the global entities whilst they are corrected, and only destroy the checker once parsing reported no errors
2026-10-05 16:36:17 +01:00
gingerBill
6aac694c18
Export the entities and scan for global whens and foreign blocks while the import graph and package order are being made
2026-10-05 16:12:20 +01:00
gingerBill
1669cf6192
Collect each package's entities as soon as it has parsed all of its files; hold any non-syntactic errors until all of the parsing is done and only report if there are no parsing errors
2026-10-05 15:40:33 +01:00
gingerBill
b2b03b68aa
Create the file scopes, imports, and canning of whens in parallel
2026-10-05 15:08:35 +01:00
gingerBill
4ad95d54b8
Gather the global entities in order from a list for each file instead of sorting them all by only sorting the files with some added out of order
2026-10-05 14:43:36 +01:00
gingerBill
90c44058bd
Merge pull request #7654 from A1029384756/windows-cross-compile
...
windows cross-linking - initial effort
2026-10-05 14:33:37 +01:00
gingerBill
b818d45e63
Keep each global group node's index on the entity itself
2026-10-05 14:29:54 +01:00
gingerBill
b6435bf7b7
Remove the declarations' child list; take entity ids in blocks per thread; collect the files with the most declarations first
2026-10-05 14:19:07 +01:00
Jeroen van Rijn
0191c579cb
Merge pull request #7756 from tibret/box3d-typo
...
Fixing typo from userMateriald to userMaterialId
2026-10-05 05:59:34 -07:00
Rob Luppi
4fdb889c81
Fixing typo from userMateriald to userMaterialId
...
I want to look at other issues, this is mostly a test of getting the
process down with something easy
2026-10-05 08:44:04 -04:00
A1029384756
b7e607737a
[windows-cross-compile] merge conflict
2026-10-05 08:34:06 -04:00
gingerBill
1082e0df9f
Wait for the type and inline cycle checks only after the min dep set
2026-10-05 13:32:39 +01:00
A1029384756
ba04d81d40
Merge remote-tracking branch 'upstream/master' into windows-cross-compile
2026-10-05 08:31:04 -04:00
gingerBill
ba13344c1f
Pass a local declaration's dependencies to those it is in once checked, remove the dependency tree update, find the global initialization order in the background, and stop linking every polymorphic call's scope into its declaration's
2026-10-05 13:20:26 +01:00
gingerBill
c500e68445
Pass a nested procedure's dependencies on to every declaration it is nested in so they do not depend on the order the bodies are checked in
2026-10-05 12:35:35 +01:00
gingerBill
bcf0ac7b21
Match polymorphic procedure specializations by parameter names at every depth so they no longer depend on which call site is checked first
2026-10-05 12:28:46 +01:00
gingerBill
c5cfff4e3c
With -use-separate-modules, make the type info in modules of its own of about ~1000 entries each rather than all in the default module
2026-10-05 12:25:58 +01:00
Jeroen van Rijn
8ca9cc4486
Merge pull request #7755 from pJotoro/patch-1
...
Add GetCursor and GetDpiForSystem procedures
2026-10-05 04:15:50 -07:00
Jeroen van Rijn
f61a61648f
Merge pull request #7754 from Kelimion/llvm-22
...
Upgrade Ubuntu CI to LLVM 22
2026-10-05 03:59:10 -07:00
gingerBill
8c55db676a
Revert running each module's backend stages as one task, as it did not help in terms of performance
2026-10-05 11:57:07 +01:00
gingerBill
8cadb56bbb
Run each module's stages after procedure generation as one task, from verification through its object, ordered by the module's cost
2026-10-05 11:45:46 +01:00
Jeroen van Rijn
bd37cf8bdd
Upgrade Ubuntu CI to LLVM 22
2026-10-05 12:42:16 +02:00
gingerBill
b57da82909
Find the global initialization order from a sorted array of the ready variables and a heap for those ready later and build the predecessors in its task rather than on the main thread
2026-10-05 10:55:13 +01:00
Jeroen van Rijn
4781f71276
Merge pull request #7751 from alexthed1rk/fix-simd-x86-sse2-insert
...
Impl simd x86 sse2 _mm_insert_epi16 fix
2026-10-05 02:41:21 -07:00
Jeroen van Rijn