math/octave-forge-instrument-control: New port.
Octave low level I/O functions for serial, i2c, parallel, tcp, gpib,
udp and usbtmc interfaces. Not vxi11 as there doesn't seem to be
FreeBSD support.
workflows/release-lit: Update workflow and enable trusted publishing with pypi (#174907)
This makes some small improvements to the workflow including using some
more modern python packaging modules and also enables the trusted
publishing for pypi. This will allow us to publish lit packages to pypi
without needing to use an access token.
This action also now uses the pypi environment which will only publish
files when triggered by an llvm-* tag.
[NFCI][AMDGPU] Use X-macro to reduce boilerplate in `GCNSubtarget.h`
`GCNSubtarget.h` contained a large amount of repetitive code following the pattern `bool HasXXX = false;` for member declarations and `bool hasXXX() const { return HasXXX; }` for getters. This boilerplate made the file unnecessarily long and harder to maintain.
This patch introduces an X-macro pattern `GCN_SUBTARGET_HAS_FEATURE` that consolidates 129 simple subtarget features into a single list. The macro is expanded twice: once in the protected section to generate member variable declarations, and once in the public section to generate the corresponding getter methods. This reduces the file by approximately 265 lines while preserving the exact same API and functionality. Features with complex getter logic or inconsistent naming conventions are left as manual implementations for future improvement.
Ideally, these could be generated by TableGen using `GET_SUBTARGETINFO_MACRO`, similar to the X86 backend. However, `AMDGPU.td` has several issues that prevent direct adoption: duplicate field names (e.g., `DumpCode` is set by both `FeatureDumpCode` and `FeatureDumpCodeLower`), and inconsistent naming conventions where many features don't have the `Has` prefix (e.g., `FlatAddressSpace`, `GFX10Insts`, `FP64`). Fixing these issues would require renaming fields in `AMDGPU.td` and updating all references, which is left for future work.
editors/openoffice-devel: Fix with python >= 3.12
Fix build with python >= 3.12.
Since this is only a build fix, and builds with older python versions
are not changed other than an additional BUILD_DEPENDS, no PORTREVISION
bump is necessary.
PR: 292190
Reported by: George Mitchell <george at m5p.com>
[mlir][scf] Skip ops having results with warning in forall-to-for pass (#175926)
Avoid converting scf.forall ops that have results in forall-to-for pass.
Emit a warning instead of failing the pass, so mlir-opt can still
produce output on mixed IR.
Fixes https://github.com/llvm/llvm-project/issues/174319
[LifetimeSafety] Detect dangling references to field members (#176805)
Add support for detecting dangling references to struct fields in the CFG-based lifetime analysis.
The tests now include cases where accessing fields of temporary objects leads to dangling references, which are properly detected.
- Added `VisitMemberExpr` method to the `FactsGenerator` class to handle field member expressions
- Implemented the method to create origin flow facts for field member expressions
- Updated tests to include CFG-based warnings for dangling references to struct fields
- Fixed previously disabled tests that can now detect these issues
Fixes https://github.com/llvm/llvm-project/issues/176144
NAS-139371 / 26.04 / Add support for clustered local SMB accounts (#18057)
This commit adds support for managing clustered accounts database for
SMB users and groups. This is accomplished by wrapping around existing
TDBHandle utils and expanding their functionality so that they also work
with clustered databases. The passdb-related functions now take a
clustered keyword argument, and group-mapping now has alternative
CLUSTERED path. On config change for `stateful_failover` we re-run the
`smb.configure` endpoint to ensure
that accounts are properly inserted into running configuration.
GlobalISel: Use LibcallLowering to get libcall calling conventions (#176837)
0e304e6d9f306ead81fc5177b8a497af0d416a73 converted the name queries,
but missed some of the calling conventions.
[Clang][AMDGPU] Handle `wavefrontsize32` and `wavefrontsize64` features more robustly (#176599)
We should not allow `-wavefrontsize32` and `-wavefrontsize64` to be
specified at the same time. We should also not allow `-wavefrontsize32`
on a target that only supports `wavefrontsize32`, and the vice versa.
[Clang][AMDGPU] Add a Sema check for the imm argument of ` __builtin_amdgcn_s_setreg`
Our backend cannot select the corresponding intrinsic if the imm argument is not a `int16_t` or `uint16_t`, which is not really helpful.
GlobalISel: Use LibcallLowering to get libcall calling conventions
0e304e6d9f306ead81fc5177b8a497af0d416a73 converted the name queries,
but missed some of the calling conventions.
