OPNSense/core 0e7e4a5src/etc/inc system.inc, src/etc/inc/plugins.inc.d webgui.inc

backend: shift the ca/cert/crl code to the newer config_read/write_array
DeltaFile
+6-9src/www/interfaces.php
+1-4src/etc/inc/plugins.inc.d/webgui.inc
+2-2src/etc/inc/system.inc
+1-1src/www/system_advanced_admin.php
+0-2src/www/system_authservers.php
+10-185 files

LLVM/project 7897d92lldb/source/Commands CommandObjectType.cpp, lldb/test/Shell/Commands command-type-synthetic-add.test

[lldb][CommandObjectType] Print name of Python class when emitting warning about invalid synthetic provider (#181829)

Before:
```
(lldb) type synthetic add -l blah foo
warning: The provided class does not exist - please define it before attempting to use this synthetic provider
```

After:
```
(lldb) type synthetic add -l blah foo
warning: The provided class 'blah' does not exist - please define it before attempting to use this synthetic provider
```

Useful when many of these registration commands happen as part of
`~/.lldbinit`. Previously it wasn't immediately obvious which of those
commands failed.
DeltaFile
+6-4lldb/source/Commands/CommandObjectType.cpp
+6-0lldb/test/Shell/Commands/command-type-synthetic-add.test
+12-42 files

LLVM/project aa5d820clang/lib/CodeGen/TargetBuiltins ARM.cpp, clang/lib/Sema SemaARM.cpp

[clang][ARM] Refactor argument handling in `EmitAArch64BuiltinExpr` (2/2) (NFC)

Refactor `EmitAArch64BuiltinExpr` so that all AArch64/NEON builtins
handled by this hook _and marked as overloaded_ share a common path
for generating LLVM IR arguments (collected into the `Ops`
`SmallVector<Value*>`) (*). This is a follow-up for #181794 - please
refer to that PR for more context.

As in the previous PR, the key change is implemented in
`HasExtraNeonArgument` , i.e. in the hook that identifies Builtins with
the extra argument. In this PR, I am replacing the ad-hoc switch
statement with a more principled approach borrowed from SemaARM.cpp,
namely:
```cpp
uint64_t mask = 0;
  switch (BuiltinID) {
  #define GET_NEON_OVERLOAD_CHECK
  #include "clang/Basic/arm_fp16.inc"
  #include "clang/Basic/arm_neon.inc"

    [28 lines not shown]
DeltaFile
+46-253clang/lib/CodeGen/TargetBuiltins/ARM.cpp
+3-1clang/lib/Sema/SemaARM.cpp
+49-2542 files

FreeBSD/ports c2535f9finance/fava distinfo Makefile

finance/fava: upgrade to 1.30.12.
DeltaFile
+3-3finance/fava/distinfo
+1-1finance/fava/Makefile
+4-42 files

OPNSense/core 8e8a707src/etc/inc config.inc

config: add config_write_array() magic

With this we can insert new entries without the need to grab the entry
beforehand using config_read_array().

    config_write_array('foo', 'bar', 'hello');

becomes

    <foo>
        <bar>hello</bar>
    </foo>
DeltaFile
+9-0src/etc/inc/config.inc
+9-01 files

LLVM/project ebaa9bfclang/lib/Driver/ToolChains FreeBSD.cpp FreeBSD.h, clang/test/Driver freebsd.cpp

[Toolchains][FreeBSD] Honor system libgcc

On Linux, the system libgcc may act as the compiler runtime, and Clang
provides -rtlib=compiler-rt to switch implementations.

FreeBSD ships libcompiler_rt.a (LLVM’s builtins) in the base system,
with libgcc.a as a symlink to it. These libraries are linked via -L and
-lgcc.

An interesting detail is that even if the Clang resource directory
(llvm-xx/lib/clang/xx) appears before the system path in the search
order, it is still not used. This is because the linker looks for
libgcc.a rather than libclang_rt.builtins.a.

Since FreeBSD does not currently support -rtlib, honor the system libgcc
by hardcoding its path. Detect and handle cases where a custom
compiler-rt is injected via -L as an override workaround.
DeltaFile
+15-0clang/lib/Driver/ToolChains/FreeBSD.cpp
+4-0clang/test/Driver/freebsd.cpp
+2-0clang/lib/Driver/ToolChains/FreeBSD.h
+21-03 files

LLVM/project 5dc9f1aclang/lib/CodeGen/TargetBuiltins ARM.cpp, clang/test/CodeGen arm64-microsoft-intrinsics.c

[clang][ARM] Refactor argument handling in `EmitAArch64BuiltinExpr` (1/2) (NFC)

Refactor `EmitAArch64BuiltinExpr` so that all AArch64/NEON builtins
handled by this hook _and marked as non-overloaded_ share a common path
for generating LLVM IR arguments (collected into the `Ops`
`SmallVector<Value*>`) (*)

Previously, the argument emission loop unconditionally skipped the
trailing argument:

```cpp
  for (unsigned i = 0, e = E->getNumArgs() - 1; i != e; ++i)
```

This was originally intended to ignore the extra Sema-only argument
used by overloaded NEON builtins (e.g. the type discriminator passed
by `__builtin_neon_*` intrinsics). However, this logic was applied
unconditionally.


    [37 lines not shown]
DeltaFile
+225-79clang/lib/CodeGen/TargetBuiltins/ARM.cpp
+16-16clang/test/CodeGen/arm64-microsoft-intrinsics.c
+241-952 files

FreeBSD/ports 5c5be37x11/hyprviz distinfo Makefile.crates

x11/hyprviz: Update to 0.8.0

Changelog: https://github.com/timasoft/hyprviz/releases/tag/v0.8.0

Reported by:    GitHub (watch releases)
DeltaFile
+19-9x11/hyprviz/distinfo
+9-4x11/hyprviz/Makefile.crates
+2-3x11/hyprviz/Makefile
+30-163 files

LLVM/project 449f99allvm/lib/CodeGen AtomicExpandPass.cpp, llvm/test/CodeGen/X86 atomic-nocx16-win.ll

[X86] Correctly call 16 byte atomic helpers on x86_64 Windows (#181356)

On x86_64 Windows, i128 values are returned in xmm0, and as arguments to
functions, they are passed indirectly (rather than in a GPR pair) - see
WinX86_64ABIInfo::classify in Clang [1].

When AtomicExpandPass generates a libcall, it normally generates one
with i128 arguments and return types. This type isn't lowered correctly
(it ends up passed or returned in a pair of GPRs).

To fix this issue, make AtomicExpandPass use the v2i64 type instead of
i128 here, just like Clang would do for an explicit function call; this
type ends up lowered correctly.

This fixes https://github.com/mstorsjo/llvm-mingw/issues/538 and
https://github.com/llvm/llvm-project/issues/136428.

[1]
https://github.com/llvm/llvm-project/blob/llvmorg-21.1.8/clang/lib/CodeGen/Targets/X86.cpp#L3390-L3408

    [2 lines not shown]
DeltaFile
+38-0llvm/test/CodeGen/X86/atomic-nocx16-win.ll
+10-0llvm/lib/CodeGen/AtomicExpandPass.cpp
+48-02 files

Illumos/gate dadc95eusr/src/lib/libc/port/gen getmntent.c, usr/src/man/man3c getmntent.3c

17728 hasmntopt should take const pointers
Reviewed by: Dan McDonald <danmcd at edgecast.io>
Reviewed by: Andy Fiddaman <illumos at fiddaman.net>
Approved by: Gordon Ross <gordon.w.ross at gmail.com>
DeltaFile
+234-175usr/src/man/man3c/getmntent.3c
+6-5usr/src/lib/libc/port/gen/getmntent.c
+4-4usr/src/uts/common/sys/mnttab.h
+244-1843 files

LLVM/project 83833eelibcxx/src atomic.cpp

[libc++] Fix FreeBSD atomic timed wait system call (#180400)

This PR fixes atomic timed wait on FreeBSD platforms.

When we added timed wait support, on FreeBSD platform, we were looking
only the `UMTX_OP_WAIT` section of the FreeBSD doc
https://man.freebsd.org/cgi/man.cgi?query=_umtx_op&apropos=0&sektion=2&manpath=FreeBSD+16.0-CURRENT&format=html
. Somehow we missed the generic section that describes the timeout
parameter that applies to all operations. As a consequence, we missed
the part that the FreeBSD is expecting the `size` to be casted to
`uintptr_t`, then passed in to the `void*` parameter.

This PR fixes the issue by casting the type to what the system requires.

As drive by, this PR also
- uses simple `timespec` instead of the extended one
- as documentation suggests, "Interval counting is always performed by
the monotonic wall clock", and "If the flag is absent, the timeout value
is relative", which means that using the simple version is enough and

    [9 lines not shown]
DeltaFile
+12-11libcxx/src/atomic.cpp
+12-111 files

Illumos/gate b3cd65eusr/src Makefile.master, usr/src/tools/cw cw.c cw.1onbld

17743 cw: want time the compiler runs
Reviewed by: Bill Sommerfeld <sommerfeld at hamachi.org>
Approved by: Dan McDonald <danmcd at edgecast.io>
DeltaFile
+19-1usr/src/tools/cw/cw.c
+15-1usr/src/tools/cw/cw.1onbld
+8-4usr/src/Makefile.master
+42-63 files

LLVM/project b1dbdf3llvm/test/CodeGen/AArch64 atomic-ops-lse.ll trampoline.ll

[AArch64] Consider MOVaddr* as cheap if fuse-adrp-add

These pseudo-instructions usually translate into a pair of adrp+add and
have a single cycle latency on some micro-architectures.
DeltaFile
+88-88llvm/test/CodeGen/AArch64/atomic-ops-lse.ll
+43-34llvm/test/CodeGen/AArch64/trampoline.ll
+55-0llvm/test/CodeGen/AArch64/cheap-as-a-move-MOVaddr.ll
+34-19llvm/test/CodeGen/AArch64/machine-outliner-loh.ll
+14-12llvm/test/CodeGen/AArch64/cfguard-checks.ll
+12-10llvm/test/CodeGen/AArch64/atomic-ops.ll
+246-16310 files not shown
+291-19416 files

LLVM/project 5739f43cross-project-tests lit.site.cfg.py.in, cross-project-tests/dtlto path.test

[DTLTO][Windows] Expand short 8.3 form paths in ThinLTO module IDs (#178303)

Windows supports short 8.3 form filenames (e.g. `compile_commands.json`
-> `COMPIL~1.JSO`) for legacy reasons. See:
https://learn.microsoft.com/en-us/windows/win32/fileio/naming-a-file#short-vs-long-names.

Short 8.3 form paths are undesirable in distributed compilation
scenarios because they are local mappings tied to a specific directory
layout on a specific machine. As a result, they can break or defeat
sandboxing and path-based isolation mechanisms used by distributed build
systems.

We have observed such failures with DTLTO even in simple scenarios. For
example, on Windows, running:

```
clang.exe hello.c -flto=thin -fuse-ld=lld -fthinlto-distributor=fastbuild.exe -###
```


    [32 lines not shown]
DeltaFile
+92-0cross-project-tests/dtlto/path.test
+59-9llvm/lib/DTLTO/DTLTO.cpp
+14-1cross-project-tests/lit.site.cfg.py.in
+6-1llvm/include/llvm/DTLTO/DTLTO.h
+171-114 files

LLVM/project 147dcd4cross-project-tests/dtlto path.test

[TEST][FIX] Fix typo in tool name: 'llvm-ar'

(cherry picked from commit 5c114d80f88ef418b5f3ba3c919d300e21c4e677)
DeltaFile
+2-2cross-project-tests/dtlto/path.test
+2-21 files

LLVM/project 68a700cllvm/include/llvm/DTLTO DTLTO.h, llvm/include/llvm/LTO LTO.h

[DTLTO][NFC] Minor improvements to the input file preparation class (#180824)

This change performs a small set of NFC refactors to improve clarity. In
particular, we make it clear that the responsibilities of this class now
extend beyond its original archive-handling role.

(cherry picked from commit 8b57b9730222cdc098d6a8708099e5b4986b1e2c)
DeltaFile
+71-88llvm/lib/DTLTO/DTLTO.cpp
+26-16llvm/include/llvm/DTLTO/DTLTO.h
+1-1llvm/include/llvm/LTO/LTO.h
+1-1llvm/lib/LTO/LTO.cpp
+99-1064 files

LLVM/project 86407b6llvm/include/llvm/Support/Windows WindowsSupport.h, llvm/lib/Support/Windows Path.inc

[Windows][Support] Add helper to expand short 8.3 form paths (#178480)

Windows supports short 8.3 form filenames (for example,
`compile_commands.json` -> `COMPIL~1.JSO`) for legacy reasons. See:
https://learn.microsoft.com/en-us/windows/win32/fileio/naming-a-file#short-vs-long-names

Such paths are not unusual because, on Windows, the system temporary
directory is commonly derived from the `TMP`/`TEMP` environment
variables. For historical compatibility reasons, these variables are
often set to short 8.3 form paths on systems where user names exceed
eight characters.

Introduce `windows::makeLongFormPath()` to convert paths to their long
form by expanding any 8.3 components via `GetLongPathNameW`.

As part of this change:
- Extended-length path prefix handling is centralized by adding
`stripExtendedPrefix()` and reusing it in `realPathFromHandle()`.
- `widenPath()` is cleaned up to use shared prefix constants.

    [5 lines not shown]
DeltaFile
+159-0llvm/unittests/Support/Path.cpp
+77-15llvm/lib/Support/Windows/Path.inc
+5-0llvm/include/llvm/Support/Windows/WindowsSupport.h
+241-153 files

LLVM/project affc059llvm/lib/Target/Hexagon/MCTargetDesc HexagonAsmBackend.cpp, llvm/test/MC/Hexagon newvalue_jump_ext_imm.s

[Hexagon] Fix encoding of packets with fixups followed by alignment (#179168)

When a packet containing extended immediates and new-value compare-jump
instructions with fixups was followed by a .p2align directive, we would
incorrectly add nops to the packet. After reshuffling, the fixup offsets
would become invalid, causing corrupted encodings.

Fixes round-trip assembly for patterns like:

    {
      r18 = ##65536
      if (!cmp.gtu(r1,r18.new)) jump:t .L1
    }
    .p2align 4

(cherry picked from commit 87d73f74713039cacf9d174a9d74de0db8eccb8b)
DeltaFile
+49-0llvm/test/MC/Hexagon/newvalue_jump_ext_imm.s
+7-0llvm/lib/Target/Hexagon/MCTargetDesc/HexagonAsmBackend.cpp
+56-02 files

FreeBSD/src efcfba9tools/build/mk OptionalObsoleteFiles.inc

OptionalObsoleteFiles: Add missing figpar MLINKS

MFC after:      3 days
Fixes:          bc6c827078b7 ("OptionalObsoleteFiles: Add figpar to dialog section")
DeltaFile
+7-0tools/build/mk/OptionalObsoleteFiles.inc
+7-01 files

LLVM/project 9b5ea37clang/cmake/caches hexagon-unknown-linux-musl-clang-dylib.cmake hexagon-unknown-linux-musl-clang-cross.cmake

[hexagon] Add dylib cmake + toolchain_only (#177247)

The toolchains take up much less space when we enable dylib, so let's
create an option to build them that way.

Also: TOOLCHAIN_ONLY was ineffective in
hexagon-unknown-linux-musl-clang-cross.cmake because cmake takes the
first setting from hexagon-unknown-linux-musl-clang.cmake with
precedence. FORCE it to fix that issue.

Signed-off-by: Brian Cain <brian.cain at oss.qualcomm.com>
(cherry picked from commit cf4ce383b3c7776c06ee744acebfec1e0c1c946d)
DeltaFile
+7-0clang/cmake/caches/hexagon-unknown-linux-musl-clang-dylib.cmake
+3-1clang/cmake/caches/hexagon-unknown-linux-musl-clang-cross.cmake
+10-12 files

OPNSense/core 6fe3d78src/etc/inc system.inc, src/etc/inc/plugins.inc.d core.inc

backend: two unused global $config references

(cherry picked from commit 83f8b6b70c6857fb461773f1acbfa00ea27a083c)
DeltaFile
+0-2src/etc/inc/plugins.inc.d/core.inc
+0-2src/etc/inc/system.inc
+0-42 files

NetBSD/pkgsrc wWSJTOumisc/xjdic Makefile

   xjdic: comment out dead link
VersionDeltaFile
1.23+2-2misc/xjdic/Makefile
+2-21 files

OPNSense/core 808e46dsrc/etc/inc interfaces.inc

interfaces: safe to drop ip_in_interface_alias_subnet()

Only used by isc-dhcp plugin which has its own copy now.
DeltaFile
+0-18src/etc/inc/interfaces.inc
+0-181 files

NetBSD/pkgsrc AkWCOICgames/xfrisk Makefile

   xfrisk: comment out dead link
VersionDeltaFile
1.35+3-3games/xfrisk/Makefile
+3-31 files

OPNSense/plugins 1bfb448net/isc-dhcp Makefile, net/isc-dhcp/src/etc/inc/plugins.inc.d dhcpd.inc

net/isc-dhcp: move ip_in_interface_alias_subnet() here

Only called by this plugin.  So we can ditch it from core.
DeltaFile
+18-0net/isc-dhcp/src/etc/inc/plugins.inc.d/dhcpd.inc
+1-1net/isc-dhcp/src/www/services_dhcp_edit.php
+1-1net/isc-dhcp/Makefile
+1-1net/isc-dhcp/src/www/services_dhcp.php
+21-34 files

NetBSD/pkgsrc QDB0NA7sysutils/xfburn Makefile

   xfburn: update HOMEPAGE
VersionDeltaFile
1.37+2-2sysutils/xfburn/Makefile
+2-21 files

NetBSD/pkgsrc DAXUZgpsysutils/xe-guest-utilities Makefile

   xe-guest-utilities: switch to https
VersionDeltaFile
1.19+2-2sysutils/xe-guest-utilities/Makefile
+2-21 files

NetBSD/pkgsrc e38x5m8x11/xdm3d Makefile

   xdm3d: comment out dead link
VersionDeltaFile
1.32+2-2x11/xdm3d/Makefile
+2-21 files

NetBSD/pkgsrc PgMDZHndevel/xdelta3 Makefile

   xdelta3: switch to github
VersionDeltaFile
1.14+2-2devel/xdelta3/Makefile
+2-21 files

OpenBSD/ports MH5pWuVsysutils/gemini-cli distinfo Makefile, sysutils/gemini-cli/pkg PLIST

   Update to gemini-cli-0.29.0.
VersionDeltaFile
1.3+425-1,057sysutils/gemini-cli/pkg/PLIST
1.3+2-2sysutils/gemini-cli/distinfo
1.3+1-1sysutils/gemini-cli/Makefile
+428-1,0603 files