FreeBSD/ports d0654a3x11-wm/obconf Makefile, x11-wm/obconf/files patch-src_tree.c

x11-wm/obconf: fix crash on theme change

Fix crashing on theme change caused by incorrect
error handling.

PR:             295921
Reported by:    fernape
DeltaFile
+20-0x11-wm/obconf/files/patch-src_tree.c
+1-1x11-wm/obconf/Makefile
+21-12 files

LLVM/project e12a557clang/lib/CodeGen CGCoroutine.cpp, clang/test/CodeGenCoroutines coro-cwg2935.cpp

[Clang][CodeGen][Coroutines] Make coroutine startup exception-safe (C… (#202279)

…WG2935)

This patch attempts to implement the solution I proposed for [CWG2935
(Github)](https://github.com/cplusplus/CWG/issues/575), aligning Clang's
behavior with GCC and MSVC instead of leaving it undefined. When
`initial_suspend` (as well as `ready` and `suspend`) throws an
exception, Clang fails to destroy the task even though the task has
already been initialized (see https://godbolt.org/z/E4Y4bEn54).

This patch updates CGCoroutine.cpp to clean up the coroutine return
value after an exception is thrown when it is constructed in place,
addressing CWG2935.

I would like to hear more opinions on the solution and seek help to fix
Clang.
DeltaFile
+46-0clang/lib/CodeGen/CGCoroutine.cpp
+35-0clang/test/CodeGenCoroutines/coro-cwg2935.cpp
+81-02 files

FreeBSD/ports 075aae1math/py-openTSNE Makefile distinfo

math/py-openTSNE: update 1.0.2 → 1.0.4
DeltaFile
+4-3math/py-openTSNE/Makefile
+3-3math/py-openTSNE/distinfo
+7-62 files

FreeBSD/ports fcfb972net/nekobox distinfo Makefile

net/nekobox: update 5.11.16 → 5.11.17
DeltaFile
+3-3net/nekobox/distinfo
+1-1net/nekobox/Makefile
+4-42 files

FreeBSD/ports 026da16devel/py-rchitect distinfo Makefile

devel/py-rchitect: update 0.4.9 → 0.4.10
DeltaFile
+3-3devel/py-rchitect/distinfo
+1-1devel/py-rchitect/Makefile
+4-42 files

FreeBSD/ports 154a8acdevel/py-pycompilation distinfo Makefile

devel/py-pycompilation: update 0.4.12 → 0.4.13
DeltaFile
+3-3devel/py-pycompilation/distinfo
+1-2devel/py-pycompilation/Makefile
+4-52 files

FreeBSD/ports a231cbamath/py-gau2grid distinfo Makefile, math/py-gau2grid/files patch-CMakeLists.txt

math/py-gau2grid: update 2.0.7 → 2.0.9
DeltaFile
+23-9math/py-gau2grid/files/patch-CMakeLists.txt
+3-3math/py-gau2grid/distinfo
+2-2math/py-gau2grid/Makefile
+28-143 files

FreeBSD/ports f114b58math/py-animatplot-ng distinfo Makefile

math/py-animatplot-ng: update 0.4.4 → 0.4.6
DeltaFile
+3-3math/py-animatplot-ng/distinfo
+4-2math/py-animatplot-ng/Makefile
+7-52 files

FreeBSD/ports 42af614devel/py-distlib distinfo Makefile

devel/py-distlib: update 0.4.1 → 0.4.2
DeltaFile
+3-3devel/py-distlib/distinfo
+2-2devel/py-distlib/Makefile
+5-52 files

FreeBSD/ports 6514ae0math/py-sparsediffpy distinfo Makefile

math/py-sparsediffpy: update 0.3.0 → 0.5.0
DeltaFile
+3-3math/py-sparsediffpy/distinfo
+1-1math/py-sparsediffpy/Makefile
+4-42 files

FreeBSD/ports e53dd1afinance/py-python-binance distinfo Makefile

finance/py-python-binance: update 1.0.36 → 1.0.37
DeltaFile
+3-3finance/py-python-binance/distinfo
+1-1finance/py-python-binance/Makefile
+4-42 files

FreeBSD/ports ba7cfccgraphics/xournalpp distinfo Makefile

graphics/xournalpp: update 1.3.3 → 1.3.5
DeltaFile
+3-3graphics/xournalpp/distinfo
+1-2graphics/xournalpp/Makefile
+1-0graphics/xournalpp/pkg-plist
+5-53 files

LLVM/project 19250ebllvm/lib/Analysis ValueTracking.cpp LazyValueInfo.cpp, llvm/lib/IR Verifier.cpp

[InstCombine] Drop zero size dereferenceable assumptions (#202411)
DeltaFile
+35-0llvm/test/Transforms/InstCombine/assume.ll
+4-8llvm/lib/Analysis/ValueTracking.cpp
+4-5llvm/lib/Analysis/LazyValueInfo.cpp
+8-1llvm/lib/Transforms/InstCombine/InstCombineCalls.cpp
+6-2llvm/test/Verifier/assume-bundles.ll
+5-2llvm/lib/IR/Verifier.cpp
+62-183 files not shown
+69-219 files

LLVM/project c04c945llvm/lib/Analysis ValueTracking.cpp, llvm/test/Transforms/LICM hoist-deref-load.ll

[ValueTracking] Handle chain of single-pred blocks in willNotFreeBetween() (#202308)

willNotFreeBetween() currently handles the case where both instructions
are in the same block, or one is in the single predecessor of the other.
This patch extends this to handle a chain of single predecessor blocks.

The budget now applies to all checked instructions, rather than per
block. Also increase the budget by a factor of two (which means that new
budget interpretation should never regress relative to the previous).
DeltaFile
+23-17llvm/lib/Analysis/ValueTracking.cpp
+1-0llvm/test/Transforms/LICM/hoist-deref-load.ll
+1-0llvm/test/Transforms/SLPVectorizer/AArch64/masked-loads-side-effects-after-vec.ll
+25-173 files

OPNSense/core 5364cacsrc/etc rc.bootup, src/etc/inc/plugins.inc.d core.inc

rc: move system_powerd_configure() to bootup plugin hook
DeltaFile
+2-1src/etc/inc/plugins.inc.d/core.inc
+0-2src/etc/rc.bootup
+2-32 files

LLVM/project 69215c5clang/docs SafeStack.rst, clang/lib/Sema SemaChecking.cpp

[SafeStack] Introduce public interface header and deprecate builtins (#198292)

Introduce `sanitizer/safestack_interface.h` to expose SafeStack utility
functions as a public interface, similar to other sanitizers.

Exposed functions:
* `__safestack_get_unsafe_stack_ptr`
* `__safestack_get_unsafe_stack_bottom`
* `__safestack_get_unsafe_stack_top`

Deprecate the existing `__builtin___get_unsafe_stack_*` builtins and
emit warnings suggesting the new `__safestack_` equivalents.

Runtime compatibility aliases `__get_unsafe_stack_*` are maintained.
DeltaFile
+35-8clang/docs/SafeStack.rst
+35-0compiler-rt/include/sanitizer/safestack_interface.h
+23-10compiler-rt/lib/safestack/safestack.cpp
+13-9compiler-rt/test/safestack/sigaltstack.c
+17-0clang/test/Sema/safestack-deprecated-builtins.c
+16-0clang/lib/Sema/SemaChecking.cpp
+139-273 files not shown
+146-279 files

LLVM/project 6b90b6bclang/lib/AST/ByteCode Interp.h, clang/test/AST/ByteCode cxx20.cpp

[clang][bytecode] Check Ptr primtype in Store op (#202314)
DeltaFile
+9-0clang/test/AST/ByteCode/cxx20.cpp
+4-0clang/lib/AST/ByteCode/Interp.h
+13-02 files

LLVM/project ae85794clang/lib/CIR/CodeGen CIRGenBuiltinAArch64.cpp, clang/test/CodeGen/AArch64 neon-intrinsics.c

[CIR][AArch64] Lower NEON vpmax intrinsics (#201495)

### summary

part of : https://github.com/llvm/llvm-project/issues/185382

Adds ClangIR support for all AArch64 NEON pairwise-maximum intrinsics
(vpmax*) in
https://arm-software.github.io/acle/neon_intrinsics/advsimd.html#pairwise-maximum

Also enables the scalar pairwise reductions vpmaxs_f32, vpmaxqd_f64,
vpmaxnms_f32, and vpmaxnmqd_f64 by routing them through the common NEON
SISD path (fmaxv / fmaxnmv).
DeltaFile
+264-0clang/test/CodeGen/AArch64/neon/intrinsics.c
+0-240clang/test/CodeGen/AArch64/neon-intrinsics.c
+8-4clang/lib/CIR/CodeGen/CIRGenBuiltinAArch64.cpp
+272-2443 files

LLVM/project 994b47bflang/lib/Semantics mod-file.cpp symbol.cpp

code-format
DeltaFile
+1-1flang/lib/Semantics/mod-file.cpp
+1-1flang/lib/Semantics/symbol.cpp
+2-22 files

LLVM/project 588130aflang/include/flang/Semantics symbol.h semantics.h, flang/lib/Semantics mod-file.cpp symbol.cpp

[flang][OpenMP] Handle groupprivate device_type like declare_target
DeltaFile
+8-12flang/lib/Semantics/mod-file.cpp
+12-3flang/lib/Semantics/symbol.cpp
+8-6flang/lib/Semantics/resolve-directives.cpp
+10-2flang/include/flang/Semantics/symbol.h
+0-1flang/include/flang/Semantics/semantics.h
+38-245 files

LLVM/project 9f31e57flang/lib/Lower/OpenMP OpenMP.cpp, flang/lib/Semantics resolve-directives.cpp

code-foramt fix
DeltaFile
+3-5flang/lib/Lower/OpenMP/OpenMP.cpp
+3-3flang/lib/Semantics/resolve-directives.cpp
+6-82 files

LLVM/project 6b765c7flang/include/flang/Semantics semantics.h symbol.h, flang/lib/Lower/OpenMP OpenMP.cpp

[flang][OpenMP] Store groupprivate device_type on symbol details
DeltaFile
+26-46flang/lib/Semantics/mod-file.cpp
+21-26flang/lib/Semantics/resolve-directives.cpp
+19-17flang/lib/Lower/OpenMP/OpenMP.cpp
+0-20flang/include/flang/Semantics/semantics.h
+0-14flang/lib/Semantics/semantics.cpp
+11-0flang/include/flang/Semantics/symbol.h
+77-1236 files

LLVM/project c031f3dflang/lib/Semantics resolve-directives.cpp

code-format fix
DeltaFile
+2-2flang/lib/Semantics/resolve-directives.cpp
+2-21 files

LLVM/project 507e371flang/include/flang/Semantics semantics.h, flang/lib/Lower/OpenMP OpenMP.cpp

[flang][OpenMP] Propagate groupprivate device_type via .mod files
DeltaFile
+105-37flang/test/Lower/OpenMP/groupprivate.f90
+31-22flang/lib/Lower/OpenMP/OpenMP.cpp
+46-2flang/lib/Semantics/mod-file.cpp
+36-0flang/test/Lower/OpenMP/groupprivate-modfile.f90
+33-0flang/lib/Semantics/resolve-directives.cpp
+21-0flang/include/flang/Semantics/semantics.h
+272-614 files not shown
+286-8810 files

LLVM/project f7a5524flang/include/flang/Lower OpenMP.h, flang/lib/Lower/OpenMP OpenMP.cpp

NFCcode changes
DeltaFile
+29-29flang/lib/Lower/OpenMP/OpenMP.cpp
+1-1flang/include/flang/Lower/OpenMP.h
+30-302 files

LLVM/project b05320bflang/include/flang/Lower OpenMP.h AbstractConverter.h, flang/lib/Lower/OpenMP OpenMP.cpp

[flang][OpenMP] Replace OMPGroupprivateDeviceTypeInfo struct with a using alias
DeltaFile
+3-5flang/include/flang/Lower/OpenMP.h
+2-6flang/include/flang/Lower/AbstractConverter.h
+2-3flang/lib/Lower/OpenMP/OpenMP.cpp
+7-143 files

LLVM/project f7455fdflang/include/flang/Lower OpenMP.h AbstractConverter.h, flang/lib/Lower Bridge.cpp

support device_type groupprivate lowering
DeltaFile
+22-3flang/lib/Lower/OpenMP/OpenMP.cpp
+22-0flang/test/Lower/OpenMP/groupprivate.f90
+11-0flang/lib/Lower/Bridge.cpp
+11-0flang/include/flang/Lower/OpenMP.h
+11-0flang/include/flang/Lower/AbstractConverter.h
+77-35 files

LLVM/project 613a810flang/include/flang/Lower OpenMP.h, flang/lib/Lower/OpenMP OpenMP.cpp

[flang][mlir] Add flang to mlir lowering for groupprivate
DeltaFile
+186-0flang/test/Lower/OpenMP/groupprivate.f90
+124-2flang/lib/Lower/OpenMP/OpenMP.cpp
+0-9flang/test/Lower/OpenMP/Todo/groupprivate.f90
+1-0flang/include/flang/Lower/OpenMP.h
+311-114 files

LLVM/project 1ad448fllvm/lib/Transforms/InstCombine InstCombineSimplifyDemanded.cpp, llvm/test/Transforms/InstCombine vec_demanded_elts.ll

[InstCombine] Don't reuse a sibling binop with extra flags via demanded elts (#199782) (#201545)

SimplifyDemandedVectorElts can replace a vector binop with a sibling
binop. But if the sibling carries a flag the original lacked, it could
produce a result the original never would. E.g.

  Reusing `sub nuw`   could make the result poison on wrap.
  Reusing `fmul nsz`  could flip the sign of a zero.
  Reusing `fdiv arcp` could make the result less accurate.

Fix by only reusing the sibling when its flags are a subset of the
replaced binop's.

https://alive2.llvm.org/ce/z/9yEaG7
Fixes #199782.
DeltaFile
+53-5llvm/test/Transforms/InstCombine/vec_demanded_elts.ll
+8-3llvm/lib/Transforms/InstCombine/InstCombineSimplifyDemanded.cpp
+61-82 files

LLVM/project 188c048llvm/lib/Analysis LoopAccessAnalysis.cpp, llvm/test/Analysis/LoopAccessAnalysis early-exit-runtime-checks.ll

[LAA] Don't check for free in evaluatePtrAddRecAtMaxBTCWillNotWrap() (#202341)

The fact that an object of a certain size existed at the the location at
some point is sufficient to prove the desired nowrap fact. Whether the
object still exists there doesn't matter in this context.
DeltaFile
+3-4llvm/lib/Analysis/LoopAccessAnalysis.cpp
+3-2llvm/test/Analysis/LoopAccessAnalysis/early-exit-runtime-checks.ll
+1-0llvm/test/Transforms/LoopVectorize/early-exit-calls.ll
+7-63 files