OpenBSD/src yEDBZycusr.bin/tmux format.c

   Fix session_*_flag format variables which loop over the windows (they should
   only be false if all windows do not have the flag, not the first one). GitHub
   issue 5599.
VersionDeltaFile
1.418+6-6usr.bin/tmux/format.c
+6-61 files

LLVM/project 3822ee5llvm/include/llvm/CodeGen CodeGenTargetMachineImpl.h, llvm/lib/CodeGen CodeGenTargetMachineImpl.cpp

CodeGen: Move DataLayout computation to CodeGenTargetMachineImpl's ctor

Every target's TargetMachine constructor passed TT.computeDataLayout() as
the DL string argument to the base constructor, duplicating the same call
across all backends. Some backends just didn't bother passing in the ABI
name.

Co-Authored-By: Claude <noreply at anthropic.com> (Claude Opus 4.8)
DeltaFile
+5-4llvm/lib/CodeGen/CodeGenTargetMachineImpl.cpp
+4-4llvm/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp
+4-4llvm/include/llvm/CodeGen/CodeGenTargetMachineImpl.h
+3-4llvm/lib/Target/RISCV/RISCVTargetMachine.cpp
+3-4llvm/lib/Target/Mips/MipsTargetMachine.cpp
+3-4llvm/lib/Target/ARM/ARMTargetMachine.cpp
+22-2421 files not shown
+46-5727 files

OpenBSD/src ZVgBDF2usr.bin/tmux utf8.c

   Do not loop forever if someone tries to give WCHAR_MAX a width, GitHub issue
   5602.
VersionDeltaFile
1.73+7-2usr.bin/tmux/utf8.c
+7-21 files

NetBSD/pkgsrc-wip aaa3a01quickshell TODO distinfo, quickshell/patches patch-src_wayland_buffer_dmabuf.cpp patch-src_services_pam_subprocess.cpp

quickshell: fix pkglint
DeltaFile
+9-9quickshell/distinfo
+6-11quickshell/TODO
+14-0quickshell/patches/patch-src_wayland_buffer_dmabuf.hpp
+3-0quickshell/patches/patch-src_core_logging.cpp
+2-0quickshell/patches/patch-src_wayland_buffer_dmabuf.cpp
+2-0quickshell/patches/patch-src_services_pam_subprocess.cpp
+36-205 files not shown
+46-2011 files

FreeBSD/ports d96b343devel Makefile, devel/py-pytest-rich pkg-descr distinfo

devel/py-pytest-rich: New port: Leverage rich for richer test session output

required by misc/py-transformers

PR:             298597
Approved by:    Baptiste Daroussin <bapt at FreeBSD.org> (on behalf of portmgr@)
DeltaFile
+27-0devel/py-pytest-rich/Makefile
+3-0devel/py-pytest-rich/distinfo
+2-0devel/py-pytest-rich/pkg-descr
+1-0devel/Makefile
+33-04 files

FreeBSD/ports d9379d2science/py-trx-python distinfo Makefile

science/py-trx-python: update 0.4.0 → 0.6
DeltaFile
+3-4science/py-trx-python/Makefile
+3-3science/py-trx-python/distinfo
+6-72 files

FreeBSD/ports 248e2b5audio/cardinal Makefile

audio/cardinal: remove unnecessary USE_CXXSTD
DeltaFile
+0-1audio/cardinal/Makefile
+0-11 files

FreeBSD/ports 56969cedevel/py-propcache distinfo Makefile, devel/py-propcache/files patch-cython

devel/py-propcache: update 0.4.1 → 0.5.4
DeltaFile
+11-6devel/py-propcache/Makefile
+0-11devel/py-propcache/files/patch-cython
+3-3devel/py-propcache/distinfo
+14-203 files

FreeBSD/ports a506f60textproc Makefile, textproc/py-sacrebleu pkg-descr distinfo

textproc/py-sacrebleu: New port: Compute shareable BLEU, chrF, and TER scores for machine translation

required by misc/py-transformers

PR:             298607
Approved by:    Baptiste Daroussin <bapt at FreeBSD.org> (on behalf of portmgr@)
DeltaFile
+34-0textproc/py-sacrebleu/Makefile
+3-0textproc/py-sacrebleu/pkg-descr
+3-0textproc/py-sacrebleu/distinfo
+1-0textproc/Makefile
+41-04 files

OpenBSD/src 63EtMpCusr.bin/tmux cmd-run-shell.c

   Expand -c for run-shell like the other -c flags, reported by Saúl Nogueras.
VersionDeltaFile
1.95+2-2usr.bin/tmux/cmd-run-shell.c
+2-21 files

LLVM/project b7b9ba8llvm/include/llvm/Transforms/Vectorize/SandboxVectorizer/Passes LoadStoreVec.h, llvm/lib/Transforms/Vectorize/SandboxVectorizer/Passes LoadStoreVec.cpp

[SandboxVec][LoadStoreVec] Support constant vectors of mixed types

createConstantVector() previously packed the constant store operands
as-is, which only worked when every store had the same element type.
Take the lane type from VecUtils::getCombinedVectorTypeFor() instead and
reinterpret each constant's bits as that type, going through an integer
of matching width via ptrtoint/inttoptr/bitcast. Constants wider than a
lane (e.g. an i64 in an <N x i32>) are split across several lanes in
memory order. Bail out when a constant cannot be reinterpreted, such as
a non-integral pointer or a relocatable address that needs splitting.

Also flatten vector-typed ConstantPointerNull into per-lane nulls, and
bail out on the remaining vector constants such as poison rather than
packing them into the result.

Co-Authored-By: Claude Opus 5 <noreply at anthropic.com>
DeltaFile
+487-0llvm/test/Transforms/SandboxVectorizer/Passes/LoadStoreVec/load_store_vec_mixed_types.ll
+99-3llvm/lib/Transforms/Vectorize/SandboxVectorizer/Passes/LoadStoreVec.cpp
+8-3llvm/include/llvm/Transforms/Vectorize/SandboxVectorizer/Passes/LoadStoreVec.h
+1-1llvm/test/Transforms/SandboxVectorizer/Passes/LoadStoreVec/load_store_vec.ll
+595-74 files

FreeBSD/ports f90704dfilesystems/amazon-efs-utils Makefile

filesystems/amazon-efs-utils: Fix build on aarch64

Sponsored by:   Netflix
DeltaFile
+11-1filesystems/amazon-efs-utils/Makefile
+11-11 files

OpenBSD/src s0dU75Qusr.bin/tmux window.c

   Reset layout manually instead of calling window_unzoom which can go down the
   notification path and end up double freeing the pane (this was previously
   removed in 2015 but added back to fix a problem with late destroy - this is a
   better fix). GitHub issue 5591 from Romain Francoise.
VersionDeltaFile
1.375+10-2usr.bin/tmux/window.c
+10-21 files

FreeBSD/ports c562b94net/gerbera pkg-plist Makefile, net/gerbera/files patch-git-01-3e73aeb54f5f4be70bc4c74f6c7362287bdbb13f

net/gerbera: Update to 3.3.0

Changelog: https://github.com/gerbera/gerbera/releases/tag/v3.3.0
DeltaFile
+0-56net/gerbera/files/patch-git-01-3e73aeb54f5f4be70bc4c74f6c7362287bdbb13f
+3-3net/gerbera/distinfo
+1-2net/gerbera/Makefile
+1-0net/gerbera/pkg-plist
+5-614 files

FreeBSD/ports 0ce7f93graphics/blender4 Makefile

graphics/blender4: Mark BROKEN

Fails to find Python 3.11 while port depends on default version which is 3.12

Reported by:    pkg-fallout
DeltaFile
+1-0graphics/blender4/Makefile
+1-01 files

FreeBSD/ports f2fc742security/lego Makefile distinfo

security/lego: Update to 5.5.1

Changelog: https://github.com/go-acme/lego/releases/tag/v5.5.1
DeltaFile
+5-5security/lego/distinfo
+1-1security/lego/Makefile
+6-62 files

FreeBSD/ports 1396390devel/pkgconf Makefile distinfo

devel/pkgconf: Update to 3.0.6

Changelog: https://github.com/pkgconf/pkgconf/blob/pkgconf-3.0.6/NEWS

PR:             297828
Approved by:    maintainer timeout, 3+ weeks
Exp-run by:     antoine
DeltaFile
+3-3devel/pkgconf/distinfo
+1-1devel/pkgconf/Makefile
+4-42 files

FreeBSD/ports 1fdc842net/libpcap Makefile distinfo

net/libpcap: Update to 1.11.0

Changelog:
https://github.com/the-tcpdump-group/libpcap/blob/libpcap-1.11.0/CHANGES

PR:             298501
Reviewed by:    garga (maintainer)
DeltaFile
+2-4net/libpcap/Makefile
+3-3net/libpcap/distinfo
+5-72 files

FreeBSD/ports 7ef4151audio/taglib Makefile distinfo

audio/taglib: Update to 2.3.2

Changelog: https://github.com/taglib/taglib/releases/tag/v2.3.2

PR:             298215
Approved by:    maintainer timeout, 2+ weeks
DeltaFile
+3-3audio/taglib/distinfo
+2-2audio/taglib/Makefile
+5-52 files

FreeBSD/src a0bb961sys/conf kern.mk

kern.mk: make clang 23 -Wunused-but-set-global non-fatal

This warning triggers in a few places in contributed code, and would
therefore be annoying to fix. Use -Wno-error= to at least show the
warnings so there is some incentive to submit them upstream.

MFC after:      3 days

(cherry picked from commit bcd0d76d4882dd8d19a49c050411b5c58a0deb74)
DeltaFile
+3-0sys/conf/kern.mk
+3-01 files

FreeBSD/src 6fa0670share/mk bsd.sys.mk

bsd.sys.mk: make clang 23 -Wunused-but-set-global non-fatal

This warning triggers in a few places in contributed code, and would
therefore be annoying to fix. Use -Wno-error= to at least show the
warnings so there is some incentive to submit them upstream.

MFC after:      3 days

(cherry picked from commit d577b4ed588c0f259d7298188233b17dc0435dd4)
DeltaFile
+3-0share/mk/bsd.sys.mk
+3-01 files

FreeBSD/src 1aea303contrib/llvm-project/libunwind/src CompactUnwinder.hpp

Merge commit 1f332ae4f1b3 from llvm-project (by Alexander Kornienko):

  Fix -Wformat diagnostic after #190965 (#193704)

  Fixes libunwind compiler diagnostic when building with clang after
  034d4dcad6396d1241e8262e69871b8d61da7e4f:
  ```
  In file included from libunwind/src/libunwind.cpp:31:
  In file included from libunwind/src/UnwindCursor.hpp:52:
  libunwind/src/CompactUnwinder.hpp:339:46: error: format specifies type 'unsigned long long' but the argument has type 'uint64_t' (aka 'unsigned long') [-Werror,-Wformat]
    338 |                            "function starting at 0x%llX",
        |                                                    ~~~~
        |                                                    %lX
    339 |                             compactEncoding, functionStart);
        |                                              ^~~~~~~~~~~~~
  libunwind/src/config.h:215:63: note: expanded from macro '_LIBUNWIND_DEBUG_LOG'
    215 |   #define _LIBUNWIND_DEBUG_LOG(msg, ...)  _LIBUNWIND_LOG(msg, __VA_ARGS__)
        |                                                          ~~~  ^~~~~~~~~~~
  libunwind/src/config.h:181:45: note: expanded from macro '_LIBUNWIND_LOG'

    [23 lines not shown]
DeltaFile
+5-4contrib/llvm-project/libunwind/src/CompactUnwinder.hpp
+5-41 files

FreeBSD/src ecf5686sys/conf kern.mk

kern.mk: make clang 23 -Wunused-but-set-global non-fatal

This warning triggers in a few places in contributed code, and would
therefore be annoying to fix. Use -Wno-error= to at least show the
warnings so there is some incentive to submit them upstream.

MFC after:      3 days

(cherry picked from commit bcd0d76d4882dd8d19a49c050411b5c58a0deb74)
DeltaFile
+3-0sys/conf/kern.mk
+3-01 files

FreeBSD/src 4cff1abshare/mk bsd.sys.mk

bsd.sys.mk: make clang 23 -Wunused-but-set-global non-fatal

This warning triggers in a few places in contributed code, and would
therefore be annoying to fix. Use -Wno-error= to at least show the
warnings so there is some incentive to submit them upstream.

MFC after:      3 days

(cherry picked from commit d577b4ed588c0f259d7298188233b17dc0435dd4)
DeltaFile
+3-0share/mk/bsd.sys.mk
+3-01 files

FreeBSD/src 672f2c5contrib/llvm-project/libunwind/src CompactUnwinder.hpp

Merge commit 1f332ae4f1b3 from llvm-project (by Alexander Kornienko):

  Fix -Wformat diagnostic after #190965 (#193704)

  Fixes libunwind compiler diagnostic when building with clang after
  034d4dcad6396d1241e8262e69871b8d61da7e4f:
  ```
  In file included from libunwind/src/libunwind.cpp:31:
  In file included from libunwind/src/UnwindCursor.hpp:52:
  libunwind/src/CompactUnwinder.hpp:339:46: error: format specifies type 'unsigned long long' but the argument has type 'uint64_t' (aka 'unsigned long') [-Werror,-Wformat]
    338 |                            "function starting at 0x%llX",
        |                                                    ~~~~
        |                                                    %lX
    339 |                             compactEncoding, functionStart);
        |                                              ^~~~~~~~~~~~~
  libunwind/src/config.h:215:63: note: expanded from macro '_LIBUNWIND_DEBUG_LOG'
    215 |   #define _LIBUNWIND_DEBUG_LOG(msg, ...)  _LIBUNWIND_LOG(msg, __VA_ARGS__)
        |                                                          ~~~  ^~~~~~~~~~~
  libunwind/src/config.h:181:45: note: expanded from macro '_LIBUNWIND_LOG'

    [23 lines not shown]
DeltaFile
+5-4contrib/llvm-project/libunwind/src/CompactUnwinder.hpp
+5-41 files

FreeBSD/src aca53fcsys/compat/linuxkpi/common/include/linux xarray.h

Fix statement with no effect in linuxkpi's xarray.h

When compiling the kernel with gcc 14, errors similar to the following
are emitted:

  sys/dev/cxgbe/iw_cxgbe/ev.c: In function 'c4iw_ev_handler':
  sys/compat/linuxkpi/common/include/linux/xarray.h:132:23: error: statement with no effect [-Werror=unused-value]
    132 |                 flags == 0; \
  sys/dev/cxgbe/iw_cxgbe/ev.c:274:17: note: in expansion of macro 'xa_unlock_irqrestore'
    274 |                 xa_unlock_irqrestore(&dev->cqs, flag);
        |                 ^~~~~~~~~~~~~~~~~~~~
  sys/compat/linuxkpi/common/include/linux/xarray.h:132:23: error: statement with no effect [-Werror=unused-value]
    132 |                 flags == 0; \
  sys/dev/cxgbe/iw_cxgbe/ev.c:283:17: note: in expansion of macro 'xa_unlock_irqrestore'
    283 |                 xa_unlock_irqrestore(&dev->cqs, flag);
        |                 ^~~~~~~~~~~~~~~~~~~~

It looks like the intent of the "flags == 0" statement was to make the
'flags' macro argument not unused, but it still results in a warning.

    [5 lines not shown]
DeltaFile
+1-1sys/compat/linuxkpi/common/include/linux/xarray.h
+1-11 files

LLVM/project 383f4cfllvm/include/llvm/ExecutionEngine/Orc DlfcnSPS.h Dlfcn.h, llvm/include/llvm/ExecutionEngine/Orc/Shared/SPSCI DlfcnSPSCI.h

[ORC] Drive LLJIT dlopen/dlupdate/dlclose via proxies (#224881)

Add Dlfcn proxies (Dlfcn.h), their SPS specs (DlfcnSPS.h) and CI
descriptors (DlfcnSPSCI.h), and route ORCPlatformSupport's three
callSPSWrapper sites through them via recordProxy.
DeltaFile
+38-45llvm/lib/ExecutionEngine/Orc/LLJIT.cpp
+51-0llvm/include/llvm/ExecutionEngine/Orc/Shared/SPSCI/DlfcnSPSCI.h
+38-0llvm/include/llvm/ExecutionEngine/Orc/Dlfcn.h
+36-0llvm/include/llvm/ExecutionEngine/Orc/DlfcnSPS.h
+163-454 files

LLVM/project 08daaa8libcxx/test/std/containers/container.adaptors/flat.map types.verify.cpp, libcxx/test/std/containers/container.adaptors/flat.multimap types.verify.cpp

[libc++] Update container adaptors' type verification tests (#224810)

1. Improves the error message for LWG2566.
2. Converts related compile.fail.cpp tests to verify.cpp tests.
3. Updates all related tests.
4. Removes some redundant tests (e.g. `[[nodiscard]]` is tested
elsewhere).

References:
- https://wg21.link/lwg2566
- https://wg21.link/container.adaptors
- https://wg21.link/flat.map.overview#7
- https://wg21.link/flat.map.overview#8
- https://wg21.link/flat.multimap.overview#7
- https://wg21.link/flat.multimap.overview#8
- https://wg21.link/flat.multiset.overview#7
- https://wg21.link/flat.multiset.overview#8
- https://wg21.link/flat.set.overview#7
- https://wg21.link/flat.set.overview#8

    [4 lines not shown]
DeltaFile
+0-35libcxx/test/std/containers/container.adaptors/stack/stack.defn/types.compile.fail.cpp
+33-0libcxx/test/std/containers/container.adaptors/flat.multimap/types.verify.cpp
+33-0libcxx/test/std/containers/container.adaptors/flat.map/types.verify.cpp
+26-0libcxx/test/std/containers/container.adaptors/flat.set/types.verify.cpp
+26-0libcxx/test/std/containers/container.adaptors/flat.multiset/types.verify.cpp
+0-22libcxx/test/std/containers/container.adaptors/stack/stack.defn/empty.verify.cpp
+118-5713 files not shown
+194-15219 files

FreeNAS/freenas e2cd46esrc/middlewared/middlewared/plugins/zfs resource_processes.py update_rules.py, src/middlewared/middlewared/pytest/unit/plugins/zfs test_update_rules.py

Round out zfs.resource with list, update and the dataset parity methods

`zfs.resource` is replacing `pool.dataset`, but it had no way to change a property, its listing method was called `query` while taking a single object rather than the `(filters, options)` that name implies everywhere else, and the choices methods, the zvol blocksize recommendation and the processes family still lived only under `pool.dataset`, so retiring that namespace would have taken those capabilities with it.

**`query` becomes `list`.** The behaviour is unchanged - a single object describing what to walk, a flat result unless `nest_results` is set - but the name no longer promises a filterable query it never was. This is a hard rename with no compatibility shim, so 25.10.x and 26.0 clients lose `zfs.resource.query`; that needs a release note.

**`update` is new.** It takes `{path, properties, user_properties, inherit}` and returns the updated entry, rather than splitting into the `set` and `inherit` that `zfs(8)` uses. A dataset edit is one save, the underlying handle already performs the property set, the user-property set and the per-property inherit in a single pass, and splitting the API would only export sequencing and partial rollback to every caller. Its property vocabulary subclasses the create model so the public/private split is inherited rather than restated, minus what ZFS cannot change after creation.

**The write primitive moves.** `update_impl` now lives on `zfs.resource` as an untyped keyword-argument primitive, since internal callers legitimately write property names outside the public vocabulary. `pool.dataset.update_impl` becomes a shim that unpacks its existing argument dict, which keeps that dict intact as the HA wire format it is and leaves its callers untouched. `zfs.tier` calls the primitive directly through a typed call and moves onto the `ZFS_RESOURCE_*` roles that match its namespace - a custom privilege built on `DATASET_*` alone therefore loses `zfs.tier`, which also needs a release note. `pool.dataset.update` keeps its own policy layer and is not shimmed onto the new method; dedup licensing, the acltype-versus-SMB-share block and the LUN resync side effects belong above the raw ZFS layer, not on it.

**`rename` and `promote` become public**, typed and single-object, held by `ZFS_RESOURCE_WRITE`, with the private entry points renamed `rename_impl` and `promote_impl`. `rename` requires `force` as the same acknowledgement `pool.dataset.rename` requires and drops `recursive`, which a dataset rename can never be; the message for that moved to the `pool.dataset.rename` shim so the caller-visible error is unchanged. `pool.snapshot.rename` is re-pointed at the snapshot implementation, which is where snapshot renames actually live - it has been dead since the rename guard landed.

**The read-parity methods move** to `zfs.resource` behind one-line `pool.dataset` shims that keep their existing models and roles, with the pure parts split into modules that need no service or filesystem access.

**Fixes carried along.** Normalising sources no longer dereferences a null `properties`; the overlapping-path check now runs whenever the walk descends rather than only when `get_children` is set, and names the option that actually triggered it; `max_depth` is bounded below at zero; and nesting no longer raises when a parent's `children` is absent because the walk never descended into it.
DeltaFile
+324-93src/middlewared/middlewared/plugins/zfs/resource.py
+379-0tests/api2/test_zfs_resource_list.py
+0-353tests/api2/test_zfs_resource_query.py
+244-0src/middlewared/middlewared/plugins/zfs/update_rules.py
+241-0src/middlewared/middlewared/plugins/zfs/resource_processes.py
+215-0src/middlewared/middlewared/pytest/unit/plugins/zfs/test_update_rules.py
+1,403-44663 files not shown
+2,931-1,19569 files

OpenBSD/ports q3A7ttWdevel/proj/patches patch-src_lib_proj_cmake

   devel/proj: unbreak configure in graphics/mapnik, reported by sthen@

   https://github.com/OSGeo/PROJ/pull/4847 adds INTERFACE_COMPILE_FEATURES
   "c_std_99;cxx_std_17" to proj-targets.cmake, and that trips cmake in
   mapnik. to be further analyzed/discussed with upstream.
VersionDeltaFile
1.3+14-11devel/proj/patches/patch-src_lib_proj_cmake
+14-111 files