From noreply at github.com  Mon Aug 17 00:17:12 2026
From: noreply at github.com (Lucas Holt)
Date: Sun, 16 Aug 2026 21:17:12 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports] be04f2: magus: fix
 multi-file YARA scans
Message-ID: <MidnightBSD/mports/push/refs/heads/master/1e3a37-be04f2@github.com>

  Branch: refs/heads/master
  Home:   https://github.com/MidnightBSD/mports
  Commit: be04f29d9776b58eaf8f5b8b8df358cf0b4b366a
      https://github.com/MidnightBSD/mports/commit/be04f29d9776b58eaf8f5b8b8df358cf0b4b366a
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-17 (Mon, 17 Aug 2026)

  Changed paths:
    M Tools/lib/Magus/App/Slave/Worker.pm

  Log Message:
  -----------
  magus: fix multi-file YARA scans

Use YARA scan-list mode so candidate port files are treated as scan targets instead of additional rule files. Classify scanner execution failures as internal errors rather than suspicious-content warnings.

AI-Assisted-by: OpenAI Codex <noreply at openai.com>
Signed-off-by: Lucas Holt <luke at foolishgames.com>



To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications

From noreply at github.com  Mon Aug 17 10:31:38 2026
From: noreply at github.com (Lucas Holt)
Date: Mon, 17 Aug 2026 07:31:38 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/src] 33c18e: netipsec: validate
 PF_KEY socket address lengths
Message-ID: <MidnightBSD/src/push/refs/heads/stable/4.0/b283db-3d0065@github.com>

  Branch: refs/heads/stable/4.0
  Home:   https://github.com/MidnightBSD/src
  Commit: 33c18eee95512aa9c86f46221bc8c45326749aa5
      https://github.com/MidnightBSD/src/commit/33c18eee95512aa9c86f46221bc8c45326749aa5
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-17 (Mon, 17 Aug 2026)

  Changed paths:
    M sys/netipsec/key.c

  Log Message:
  -----------
  netipsec: validate PF_KEY socket address lengths

Validate new PF_KEY security association addresses before copying them into fixed-size index fields.  Convert the index setup macros to checked helpers so future callers retain the same bound invariant.

This ports FreeBSD src commits cba481a7bff2 and fffaf471280b.

AI-Assisted-by: OpenAI Codex (GPT-5)
Signed-off-by: Lucas Holt <luke at foolishgames.com>


  Commit: 3d006548093a5588ec5499db3e5da592aac9c8be
      https://github.com/MidnightBSD/src/commit/3d006548093a5588ec5499db3e5da592aac9c8be
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-17 (Mon, 17 Aug 2026)

  Changed paths:
    M UPDATING

  Log Message:
  -----------
  UPDATING: note PF_KEY socket address fix

AI-Assisted-by: OpenAI Codex (GPT-5)
Signed-off-by: Lucas Holt <luke at foolishgames.com>


Compare: https://github.com/MidnightBSD/src/compare/b283db294471...3d006548093a

To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/src/settings/notifications

From noreply at github.com  Mon Aug 17 10:31:50 2026
From: noreply at github.com (Lucas Holt)
Date: Mon, 17 Aug 2026 14:31:50 +0000 (UTC)
Subject: [Midnightbsd-cvs] [MidnightBSD/src] b66e4c: netipsec: validate
 PF_KEY socket address lengths
Message-ID: <MidnightBSD/src/push/refs/heads/stable/4.1/6f1bdd-5ee85c@github.com>

  Branch: refs/heads/stable/4.1
  Home:   https://github.com/MidnightBSD/src
  Commit: b66e4cd1377d58a692d6b9ada2e81ea41ea40b67
      https://github.com/MidnightBSD/src/commit/b66e4cd1377d58a692d6b9ada2e81ea41ea40b67
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-17 (Mon, 17 Aug 2026)

  Changed paths:
    M sys/netipsec/key.c

  Log Message:
  -----------
  netipsec: validate PF_KEY socket address lengths

Validate new PF_KEY security association addresses before copying them into fixed-size index fields.  Convert the index setup macros to checked helpers so future callers retain the same bound invariant.

This ports FreeBSD src commits cba481a7bff2 and fffaf471280b.

AI-Assisted-by: OpenAI Codex (GPT-5)
Signed-off-by: Lucas Holt <luke at foolishgames.com>


  Commit: 5ee85c7b0bbb03df3ed4eb1e14499c4067f2aa03
      https://github.com/MidnightBSD/src/commit/5ee85c7b0bbb03df3ed4eb1e14499c4067f2aa03
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-17 (Mon, 17 Aug 2026)

  Changed paths:
    M UPDATING

  Log Message:
  -----------
  UPDATING: note PF_KEY socket address fix

AI-Assisted-by: OpenAI Codex (GPT-5)
Signed-off-by: Lucas Holt <luke at foolishgames.com>


Compare: https://github.com/MidnightBSD/src/compare/6f1bdd4dcb21...5ee85c7b0bbb

To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/src/settings/notifications

From noreply at github.com  Mon Aug 17 10:31:50 2026
From: noreply at github.com (Lucas Holt)
Date: Mon, 17 Aug 2026 14:31:50 +0000 (UTC)
Subject: [Midnightbsd-cvs] [MidnightBSD/src] 1132bf: netipsec: validate
 PF_KEY socket address lengths
Message-ID: <MidnightBSD/src/push/refs/heads/master/6a27a4-950fb3@github.com>

  Branch: refs/heads/master
  Home:   https://github.com/MidnightBSD/src
  Commit: 1132bf4a3d6ef46f0fc42f0cbdc9c48522f5a6e4
      https://github.com/MidnightBSD/src/commit/1132bf4a3d6ef46f0fc42f0cbdc9c48522f5a6e4
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-17 (Mon, 17 Aug 2026)

  Changed paths:
    M sys/netipsec/key.c

  Log Message:
  -----------
  netipsec: validate PF_KEY socket address lengths

Validate new PF_KEY security association addresses before copying them into fixed-size index fields.  Convert the index setup macros to checked helpers so future callers retain the same bound invariant.

This ports FreeBSD src commits cba481a7bff2 and fffaf471280b.

AI-Assisted-by: OpenAI Codex (GPT-5)
Signed-off-by: Lucas Holt <luke at foolishgames.com>


  Commit: 950fb32dcefbc94fee97253c77818106a0ac3f3e
      https://github.com/MidnightBSD/src/commit/950fb32dcefbc94fee97253c77818106a0ac3f3e
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-17 (Mon, 17 Aug 2026)

  Changed paths:
    M UPDATING

  Log Message:
  -----------
  UPDATING: note PF_KEY socket address fix

AI-Assisted-by: OpenAI Codex (GPT-5)
Signed-off-by: Lucas Holt <luke at foolishgames.com>


Compare: https://github.com/MidnightBSD/src/compare/6a27a477ec74...950fb32dcefb

To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/src/settings/notifications

From noreply at github.com  Mon Aug 17 17:03:15 2026
From: noreply at github.com (Lucas Holt)
Date: Mon, 17 Aug 2026 14:03:15 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports] f77687: www/node*: update
 releases and fix c-ares build
Message-ID: <MidnightBSD/mports/push/refs/heads/master/be04f2-f77687@github.com>

  Branch: refs/heads/master
  Home:   https://github.com/MidnightBSD/mports
  Commit: f77687f152566a1b209143a9e8895e2526007d1a
      https://github.com/MidnightBSD/mports/commit/f77687f152566a1b209143a9e8895e2526007d1a
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-17 (Mon, 17 Aug 2026)

  Changed paths:
    M www/node20/Makefile
    M www/node20/Makefile.version
    M www/node20/distinfo
    M www/node20/files/patch-src_cares__wrap.h
    M www/node22/Makefile
    M www/node22/Makefile.version
    M www/node22/distinfo
    R www/node22/files/patch-common.gypi
    M www/node22/files/patch-src-cares_wrap.h
    M www/node23/Makefile
    M www/node23/files/patch-src-cares_wrap.h
    M www/node24/Makefile
    M www/node24/Makefile.version
    M www/node24/distinfo
    M www/node24/files/patch-src-cares_wrap.h
    M www/node25/Makefile
    M www/node25/Makefile.version
    M www/node25/distinfo
    M www/node25/files/patch-src-cares_wrap.h

  Log Message:
  -----------
  www/node*: update releases and fix c-ares build

Update maintained Node.js releases and make the reverse DNS callback const-correct for current c-ares. Keep node23 at its final release with a revision bump and drop an obsolete PowerPC-only node22 patch.

AI-Assisted-by: OpenAI Codex <noreply at openai.com>
Signed-off-by: Lucas Holt <luke at foolishgames.com>



To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications

From noreply at github.com  Mon Aug 17 18:26:55 2026
From: noreply at github.com (Lucas Holt)
Date: Mon, 17 Aug 2026 15:26:55 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports] 09b318:
 devel/py-poetry-core: update to 2.4.1
Message-ID: <MidnightBSD/mports/push/refs/heads/master/f77687-3e8bc4@github.com>

  Branch: refs/heads/master
  Home:   https://github.com/MidnightBSD/mports
  Commit: 09b3188d32e86c281cf7c8af62cd35aba4173db7
      https://github.com/MidnightBSD/mports/commit/09b3188d32e86c281cf7c8af62cd35aba4173db7
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-16 (Sun, 16 Aug 2026)

  Changed paths:
    M devel/py-poetry-core/Makefile
    M devel/py-poetry-core/distinfo

  Log Message:
  -----------
  devel/py-poetry-core: update to 2.4.1

1.9.1 only understands the legacy [tool.poetry] metadata, so projects that
declare PEP 621 [project] metadata fail to build against it:

  RuntimeError: The Poetry configuration is invalid:
    - The fields ['authors', 'description', 'name', 'version'] are required
      in package mode.        (audio/py-audioread)
  ERROR Missing dependencies: poetry-core>=2.0.0   (math/py-pandas-stubs)

2.x still supports the legacy layout, so existing consumers keep working.
Both ports above build against 2.4.1 here.

AI-Assisted-by: Claude Opus 5 (1M context) <noreply at anthropic.com>
Signed-off-by: Lucas Holt <luke at foolishgames.com>


  Commit: e396fc034bc1a944c78428ff1d7bc81034e6b45f
      https://github.com/MidnightBSD/mports/commit/e396fc034bc1a944c78428ff1d7bc81034e6b45f
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-16 (Sun, 16 Aug 2026)

  Changed paths:
    M devel/py-boto/Makefile

  Log Message:
  -----------
  devel/py-boto: depend on py-six at build time

setup.py imports boto, which imports boto.compat, which imports six, so the
module is needed before the package is built and not only at run time:

  File "boto/compat.py", line 49, in <module>
    import six
  ModuleNotFoundError: No module named 'six'

AI-Assisted-by: Claude Opus 5 (1M context) <noreply at anthropic.com>
Signed-off-by: Lucas Holt <luke at foolishgames.com>


  Commit: 924748e58993711ac1423fdea082884d987738c6
      https://github.com/MidnightBSD/mports/commit/924748e58993711ac1423fdea082884d987738c6
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-16 (Sun, 16 Aug 2026)

  Changed paths:
    M x11-fonts/py-compreffor/Makefile

  Log Message:
  -----------
  x11-fonts/py-compreffor: add the missing py-wheel build dependency

The pep517 build stops with "ERROR Missing dependencies: wheel".

AI-Assisted-by: Claude Opus 5 (1M context) <noreply at anthropic.com>
Signed-off-by: Lucas Holt <luke at foolishgames.com>


  Commit: 194e19281339f480cbff5b1bf3f45a499d6ee678
      https://github.com/MidnightBSD/mports/commit/194e19281339f480cbff5b1bf3f45a499d6ee678
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-16 (Sun, 16 Aug 2026)

  Changed paths:
    M devel/thrift-cpp/Makefile

  Log Message:
  -----------
  devel/thrift-cpp: fix the staging path in post-install

post-install used ${STAGEDIR}${PREFIX}.  During fake, PREFIX already contains
FAKE_DESTDIR, so the path was doubled and strip could not find the libraries:

  strip: open .../fake-inst-amd64/magus/work/.../fake-inst-amd64/usr/local/lib/lib*.so
  failed: No such file or directory

The three compatibility symlinks were created in the same bogus directory.

AI-Assisted-by: Claude Opus 5 (1M context) <noreply at anthropic.com>
Signed-off-by: Lucas Holt <luke at foolishgames.com>


  Commit: 6ac23e80144b4f8d02cc7b3bb00995a8ee627936
      https://github.com/MidnightBSD/mports/commit/6ac23e80144b4f8d02cc7b3bb00995a8ee627936
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-16 (Sun, 16 Aug 2026)

  Changed paths:
    M filesystems/httpdirfs/Makefile
    A filesystems/httpdirfs/files/patch-meson.build

  Log Message:
  -----------
  filesystems/httpdirfs: do not build the test suite

tests/meson.build pulls in the "unity" subproject through a wrap that meson
resolves by cloning it with git.  Ports must not fetch anything at build time,
and git is not a build dependency, so configure fails:

  ERROR: Subproject unity is buildable: NO
  tests/meson.build:5:12: ERROR: Git program not found, cannot download
  unity.wrap via git.

There is no meson option to turn the tests off, so patch out the subdir call.

AI-Assisted-by: Claude Opus 5 (1M context) <noreply at anthropic.com>
Signed-off-by: Lucas Holt <luke at foolishgames.com>


  Commit: e6d5fec3de444e03203d203081f113823d5ca422
      https://github.com/MidnightBSD/mports/commit/e6d5fec3de444e03203d203081f113823d5ca422
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-16 (Sun, 16 Aug 2026)

  Changed paths:
    M ai/github-copilot-cli/Makefile

  Log Message:
  -----------
  ai/github-copilot-cli: run the phdr relocation after the patch phase

The program header relocation needs python, but python:build registers a
BUILD_DEPENDS, and those are only installed at the start of the configure
phase -- well after post-extract runs:

  make: exec(/usr/local/bin/python3.12) failed (No such file or directory)

Use python:patch and do the work in post-patch, which runs after
patch-depends.  The relocation now reports "moved 14 program headers to
offset 64" and the port packages.

AI-Assisted-by: Claude Opus 5 (1M context) <noreply at anthropic.com>
Signed-off-by: Lucas Holt <luke at foolishgames.com>


  Commit: dc0f9d3025da542aee6a760f63c8438944d14b0e
      https://github.com/MidnightBSD/mports/commit/dc0f9d3025da542aee6a760f63c8438944d14b0e
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-16 (Sun, 16 Aug 2026)

  Changed paths:
    M databases/py-pg8000/Makefile
    M databases/py-pg8000/distinfo

  Log Message:
  -----------
  databases/py-pg8000: update to 1.31.5

1.17.0 builds through an old versioneer.py that calls
configparser.SafeConfigParser, removed in Python 3.12:

  AttributeError: module 'configparser' has no attribute 'SafeConfigParser'

1.31.5 uses hatchling and PEP 517.  Unlike FreeBSD this depends on
devel/py-hatchling rather than the py-hatch CLI, which is what pyproject.toml
actually lists in build-system.requires and is the port mports has.

AI-Assisted-by: Claude Opus 5 (1M context) <noreply at anthropic.com>
Signed-off-by: Lucas Holt <luke at foolishgames.com>


  Commit: 2a78def1fe8be4a920b4e9015628b122bc0a365f
      https://github.com/MidnightBSD/mports/commit/2a78def1fe8be4a920b4e9015628b122bc0a365f
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-17 (Mon, 17 Aug 2026)

  Changed paths:
    M audio/libsamplerate/Makefile
    M audio/libsamplerate/distinfo
    A audio/libsamplerate/files/patch-cmake_FindSndFile.cmake
    R audio/libsamplerate/files/patch-examples_audio__out.c
    R audio/libsamplerate/files/patch-src_samplerate.h
    M audio/libsamplerate/pkg-plist

  Log Message:
  -----------
  audio/libsamplerate: update to 0.2.2

0.1.9 predates src_clone(), so audio/py-samplerate does not build:

  samplerate.cpp:129:14: error: use of undeclared identifier 'src_clone'

Upstream moved to github.com/libsndfile/libsamplerate, switched to cmake and
relicensed from GPLv2 to BSD 2-clause in 0.2.0; the LICENSE change is
confirmed against COPYING in the 0.2.2 tarball.  The two local patches were
0.1.9-era and no longer apply, so take FreeBSD's FindSndFile patch instead.

The shared library version stays libsamplerate.so.0, so the 17 dependent
ports need no revision bump.

do-install-EXAMPLES-on uses ${FAKE_DESTDIR}${PREFIX}: option helper targets
run in the main make, where PREFIX is not stage-prefixed and STAGEDIR is
empty.

AI-Assisted-by: Claude Opus 5 (1M context) <noreply at anthropic.com>
Signed-off-by: Lucas Holt <luke at foolishgames.com>


  Commit: 2f810f7e4c7c0f43dd187136f48b4b46aa555b11
      https://github.com/MidnightBSD/mports/commit/2f810f7e4c7c0f43dd187136f48b4b46aa555b11
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-17 (Mon, 17 Aug 2026)

  Changed paths:
    M devel/py-pyproject-metadata/Makefile
    M devel/py-pyproject-metadata/distinfo

  Log Message:
  -----------
  devel/py-pyproject-metadata: update to 0.12.1

Needed by devel/meson-python 0.18.0, which requires >= 0.9.0.  Upstream moved
to github.com/pypa/pyproject-metadata, renamed the sdist to
pyproject_metadata-*, and switched the build backend to flit-core.

AI-Assisted-by: Claude Opus 5 (1M context) <noreply at anthropic.com>
Signed-off-by: Lucas Holt <luke at foolishgames.com>


  Commit: fec85a962f6d7a16afb3ea49834e3fa236098c27
      https://github.com/MidnightBSD/mports/commit/fec85a962f6d7a16afb3ea49834e3fa236098c27
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-17 (Mon, 17 Aug 2026)

  Changed paths:
    M devel/meson-python/Makefile
    M devel/meson-python/distinfo

  Log Message:
  -----------
  devel/meson-python: update to 0.18.0

math/py-numpy 2.x requires meson-python >= 0.18.0.  Bump the
pyproject-metadata requirement to >= 0.9.0 and packaging to >= 23.2 to match,
and add the sysutils/patchelf run dependency the newer backend uses.

The MidnightBSD-local meson-${PYTHON_VER} naming and the MESON make
environment override are kept.

AI-Assisted-by: Claude Opus 5 (1M context) <noreply at anthropic.com>
Signed-off-by: Lucas Holt <luke at foolishgames.com>


  Commit: 20c0595129d03e66722976f38fcb4c7d4e6f87b8
      https://github.com/MidnightBSD/mports/commit/20c0595129d03e66722976f38fcb4c7d4e6f87b8
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-17 (Mon, 17 Aug 2026)

  Changed paths:
    M math/py-numpy/Makefile
    M math/py-numpy/distinfo
    R math/py-numpy/files/patch-clang
    R math/py-numpy/files/patch-numpy-distutils-exec_command.py
    R math/py-numpy/files/patch-numpy-distutils-fcompiler-__init__.py
    R math/py-numpy/files/patch-numpy-distutils-fcompiler-gnu.py
    R math/py-numpy/files/patch-numpy-distutils-system_info.py
    R math/py-numpy/files/patch-numpy_core_setup.py
    R math/py-numpy/files/patch-numpy_core_src_common_npy__cpu__features.c
    R math/py-numpy/files/patch-numpy_distutils_mingw32ccompiler.py
    R math/py-numpy/files/patch-numpy_f2py_f2py2e.py
    R math/py-numpy/files/patch-numpy_random_setup.py
    R math/py-numpy/files/patch-pyproject.toml
    R math/py-numpy/files/patch-swig44
    R math/py-numpy/files/site.cfg

  Log Message:
  -----------
  math/py-numpy: update to 2.4.6

mports has carried math/py-numpy1 at 1.26.4 for a while, so the 1.x line is
already provided; math/py-numpy itself was simply never moved to 2.x.  This
blocks consumers that require the new ABI, e.g. math/py-ml-dtypes:

  py312-ml-dtypes-0.5.1_2 depends on package: py312-numpy>=2.0,1<3,1 - not found

None of the 56 ports depending on math/py-numpy pin an upper bound below 2, so
nothing needed relaxing.

numpy 2 removed numpy.distutils, so all twelve local patches and site.cfg are
gone along with the distutils-era configure plumbing; the build is meson via
PEP 517.  FreeBSD carries no patches at this version either.

The BLAS choice keeps ATLAS, NETLIB and OPENBLAS: mports' blaslapack extension
knows atlas, gotoblas, netlib and openblas only, so FreeBSD's BLIS and
FlexiBLAS options have nothing to select here.  CONFLICTS_INSTALL against
py*-numpy1 is added, and post-install-DOCS-on keeps ${FAKE_DESTDIR} because
option helper targets run in the main make.

AI-Assisted-by: Claude Opus 5 (1M context) <noreply at anthropic.com>
Signed-off-by: Lucas Holt <luke at foolishgames.com>


  Commit: 39de5aa1ca13d665b14d0532e38a8dc3caeaa59b
      https://github.com/MidnightBSD/mports/commit/39de5aa1ca13d665b14d0532e38a8dc3caeaa59b
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-17 (Mon, 17 Aug 2026)

  Changed paths:
    M math/or-tools/Makefile

  Log Message:
  -----------
  math/or-tools: take libbzip2 from misc/vxl

archivers/bzip2 installs libbz2.so, not libbzip2.so, so the dependency could
never be satisfied:

  or-tools-9.14 depends on shared library: libbzip2.so - not found

misc/vxl bundles and installs lib/libbzip2.so, which is where FreeBSD points
this dependency too.

AI-Assisted-by: Claude Opus 5 (1M context) <noreply at anthropic.com>
Signed-off-by: Lucas Holt <luke at foolishgames.com>


  Commit: a4f8d3237916330fbbffb3640546fda613c9a242
      https://github.com/MidnightBSD/mports/commit/a4f8d3237916330fbbffb3640546fda613c9a242
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-17 (Mon, 17 Aug 2026)

  Changed paths:
    M ai/py-sagemaker-schema-inference-artifacts/Makefile

  Log Message:
  -----------
  ai/py-sagemaker-schema-inference-artifacts: fix the staging path of the install

PEP517_INSTALL_CMD passed --destdir ${STAGEDIR} --prefix ${PREFIX}, but
do-install runs with PREFIX already stage-prefixed, so the wheel unpacked into
a doubled path and the RECORD lookup that follows failed:

  FileNotFoundError: .../fake-inst-amd64/usr/local/lib/python3.12/site-packages/
  sagemaker_schema_inference_artifacts-0.0.5*.dist-info/RECORD

Use --destdir ${FAKE_DESTDIR} --prefix ${TRUE_PREFIX}, matching the default
PEP517_INSTALL_CMD in Mk/extensions/python.mk.

AI-Assisted-by: Claude Opus 5 (1M context) <noreply at anthropic.com>
Signed-off-by: Lucas Holt <luke at foolishgames.com>


  Commit: 60fba952dc9154b61880cf6a9b4fd05b3d447b38
      https://github.com/MidnightBSD/mports/commit/60fba952dc9154b61880cf6a9b4fd05b3d447b38
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-17 (Mon, 17 Aug 2026)

  Changed paths:
    R www/gitlab-workhorse/files/patch-internal_gitaly_blob.go

  Log Message:
  -----------
  www/gitlab-workhorse: drop the obsolete blob.go patch

patch-internal_gitaly_blob.go widened a len() comparison against
math.MaxUint32 in BlobClient.SendListBlobs.  18.8.2 no longer has that
function or the comparison, so the patch fails to apply and the port stops in
the patch phase:

  No such line 69 in input file, ignoring
  1 out of 1 hunks failed--saving rejects to internal/gitaly/blob.go.rej

The only remaining math.MaxUint32 uses in workhorse are in vendored third
party code.  The patch phase completes cleanly without it.

AI-Assisted-by: Claude Opus 5 (1M context) <noreply at anthropic.com>
Signed-off-by: Lucas Holt <luke at foolishgames.com>


  Commit: 5aaf83a4bdc4d59af1f9ab50367111ce4480caee
      https://github.com/MidnightBSD/mports/commit/5aaf83a4bdc4d59af1f9ab50367111ce4480caee
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-17 (Mon, 17 Aug 2026)

  Changed paths:
    M japanese/nethack34/Makefile

  Log Message:
  -----------
  japanese/nethack34: keep the stage directory out of the launcher script

sys/unix/Makefile.top builds bin/jnethack by substituting GAMEDIR, which is
${PREFIX}/share/${HACKNAME}.  PREFIX is stage-prefixed during install, so the
launcher shipped with the stage path baked in and fake rejected it:

  bin/jnethack contains the fake destdir

  HACKDIR=.../work/fake-inst-amd64/usr/local/share/jnethack

Strip FAKE_DESTDIR back out in post-install; the launcher now reads
HACKDIR=/usr/local/share/jnethack.

AI-Assisted-by: Claude Opus 5 (1M context) <noreply at anthropic.com>
Signed-off-by: Lucas Holt <luke at foolishgames.com>


  Commit: 0b784101c001d9647e7f6d8366f4a6494655dceb
      https://github.com/MidnightBSD/mports/commit/0b784101c001d9647e7f6d8366f4a6494655dceb
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-17 (Mon, 17 Aug 2026)

  Changed paths:
    M lang/cython0/Makefile

  Log Message:
  -----------
  lang/cython0: drop the py27 flavor

Cython 0.29.37 fails to cythonize its own sources under python 2.7:

  Cython.Compiler.Errors.CompileError: .../Cython/Compiler/FlowControl.py

FreeBSD no longer has python 2.7 at all, so its python:2.7-3.12 range never
generates a py27 flavor there; mports still ships lang/python27, so it does.
Nothing in the tree depends on lang/cython0 and the port is already marked
DEPRECATED with an expiration date that has passed, so narrow the range rather
than chase the 2.7 build.

AI-Assisted-by: Claude Opus 5 (1M context) <noreply at anthropic.com>
Signed-off-by: Lucas Holt <luke at foolishgames.com>


  Commit: db588224cdf6b478e4a30b68377d352e53877d26
      https://github.com/MidnightBSD/mports/commit/db588224cdf6b478e4a30b68377d352e53877d26
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-17 (Mon, 17 Aug 2026)

  Changed paths:
    M lang/gnat12/Makefile

  Log Message:
  -----------
  lang/gnat12: build as FreeBSD and fix the PLIST_SUB values

gcc's configure has no midnightbsd case, so _PLATFORM=${OPSYS:tl}${OSREL} gave
a triplet it rejects outright:

  *** Configuration x86_64-portbld-midnightbsd4 not supported
  gmake[1]: *** [Makefile:4549: configure-gcc] Error 1

Use freebsd13.0, which is what lang/gcc12 does and what this port's own
bootstrap assets (gnat-*-freebsd.13-*) already are.

With the build fixed, packaging then failed because the architecture and
multilib PLIST_SUB values were a single space rather than empty, so every
entry using them gained a leading space:

  Could not stat .../usr/local/ gnat12/lib/gcc/.../adainclude/g-sse.ads

Set AMD64, AARCH64, X86, 64BIT and LIB32 to "" for the enabled cases.  Builds
and packages here with MULTILIB on.

AI-Assisted-by: Claude Opus 5 (1M context) <noreply at anthropic.com>
Signed-off-by: Lucas Holt <luke at foolishgames.com>


  Commit: 9d352b56b9c63299ccf87fa66828e715028ceaad
      https://github.com/MidnightBSD/mports/commit/9d352b56b9c63299ccf87fa66828e715028ceaad
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-17 (Mon, 17 Aug 2026)

  Changed paths:
    M databases/firebird40-server/Makefile
    M databases/firebird40-server/distinfo

  Log Message:
  -----------
  databases/firebird40-server: build msg.fdb for the client too, update to 4.0.5

databases/firebird40-client failed in the message file step:

  I/O error during "open" operation for file "msg.fdb"
  (E) src/msgs/build_file.epp:47: Couldn't access database DB = 'msg.fdb'

gen/msg.fdb is produced by the "databases" target, and gen/Makefile only makes
that a prerequisite of gpre when IsServer=Y:

  gen/Makefile:434  ifeq ($(IsServer), Y)
                        gpre: $(GPRE) databases
                    else
                        gpre: $(GPRE)

but "rest" depends on message_file either way, and the client package installs
the compiled .msg files.  IsServer=Y was set only for the server, which is why
firebird40-server built while firebird40-client did not.  Move it out of the
CLIENT_ONLY guard; USERS and GROUPS stay server-only.

The update to 4.0.5 is unrelated to that failure -- 4.0.4 and 4.0.5 fail
identically without the IsServer change -- and is included only to track the
release FreeBSD ships.  Both ports build and package.

AI-Assisted-by: Claude Opus 5 (1M context) <noreply at anthropic.com>
Signed-off-by: Lucas Holt <luke at foolishgames.com>


  Commit: c7b22106ec4f1283fd9e83ac5aee79a48f5b9d51
      https://github.com/MidnightBSD/mports/commit/c7b22106ec4f1283fd9e83ac5aee79a48f5b9d51
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-17 (Mon, 17 Aug 2026)

  Changed paths:
    A lang/rust-nightly/files/patch-src_bootstrap_bootstrap.py

  Log Message:
  -----------
  lang/rust-nightly: teach bootstrap.py about MidnightBSD

The build stopped immediately:

  unknown OS type: MidnightBSD
  Build completed unsuccessfully in 0:00:00

bootstrap.py maps uname(1) output to an LLVM triple and aborts on anything
missing from its table.  lang/rust already carries this hunk as
patch-src_bootstrap_bootstrap.py; rust-nightly never got it.

This is unrelated to the SIGSEGV failures in lang/rust and lang/rust188.

AI-Assisted-by: Claude Opus 5 (1M context) <noreply at anthropic.com>
Signed-off-by: Lucas Holt <luke at foolishgames.com>


  Commit: 3e8bc46a608755e23da1e40e3bd6ff681ba4ea48
      https://github.com/MidnightBSD/mports/commit/3e8bc46a608755e23da1e40e3bd6ff681ba4ea48
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-17 (Mon, 17 Aug 2026)

  Changed paths:
    M Tools/lib/Magus/App/Slave/Worker.pm
    M www/firefox/Makefile
    M www/firefox/distinfo
    M www/firefox/files/patch-python_sites_build.txt
    M www/firefox/files/patch-third__party_sqlite3_ext_sqlite-vec_sqlite-vec.c
    A www/firefox/files/patch-toolkit_components_protobuf_src_google_protobuf_io_zero__copy__stream__impl.cc
    R www/firefox/files/patch-toolkit_components_protobuf_src_google_protobuf_stubs_port.h
    M www/node20/Makefile
    M www/node20/Makefile.version
    M www/node20/distinfo
    M www/node20/files/patch-src_cares__wrap.h
    M www/node22/Makefile
    M www/node22/Makefile.version
    M www/node22/distinfo
    R www/node22/files/patch-common.gypi
    M www/node22/files/patch-src-cares_wrap.h
    M www/node23/Makefile
    M www/node23/files/patch-src-cares_wrap.h
    M www/node24/Makefile
    M www/node24/Makefile.version
    M www/node24/distinfo
    M www/node24/files/patch-src-cares_wrap.h
    M www/node25/Makefile
    M www/node25/Makefile.version
    M www/node25/distinfo
    M www/node25/files/patch-src-cares_wrap.h

  Log Message:
  -----------
  Merge branch 'master' of ssh://github.com/MidnightBSD/mports


Compare: https://github.com/MidnightBSD/mports/compare/f77687f15256...3e8bc46a6087

To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications

From noreply at github.com  Tue Aug 18 13:11:28 2026
From: noreply at github.com (Lucas Holt)
Date: Tue, 18 Aug 2026 10:11:28 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports] 73ba9d: lang/rust: update to
 1.95.0
Message-ID: <MidnightBSD/mports/push/refs/heads/lang/rust/000000-73ba9d@github.com>

  Branch: refs/heads/lang/rust
  Home:   https://github.com/MidnightBSD/mports
  Commit: 73ba9ddb7fe1e101a6c140c31b1256b20fcf3085
      https://github.com/MidnightBSD/mports/commit/73ba9ddb7fe1e101a6c140c31b1256b20fcf3085
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-18 (Tue, 18 Aug 2026)

  Changed paths:
    M lang/rust/Makefile
    M lang/rust/distinfo
    M lang/rust/files/patch-AddLLVM.cmake
    M lang/rust/files/patch-vendor_cc.rs

  Log Message:
  -----------
  lang/rust: update to 1.95.0

Bump PORTVERSION to 1.95.0, reset PORTREVISION, and move the bootstrap
compiler to 1.94.0 / 2026-03-05 to match src/stage0, mirroring FreeBSD
ports commit 3c971390fa39.

Two patches needed retargeting for the new sources:

- patch-AddLLVM.cmake: upstream LLVM now quotes "${CMAKE_SYSTEM_NAME}"
  in the rpath origin check and the line moved from 2552 to 2568.
- patch-vendor_cc.rs: the vendored cc crate moved from 1.2.50 to 1.2.55.
  All ten vendored cc-* directories remain covered.

Announce:	https://blog.rust-lang.org/2026/04/16/Rust-1.95.0/
ChangeLog:	https://github.com/rust-lang/rust/releases/tag/1.95.0

Validated on amd64: checksum, patch, build, fake, and package all
succeed; the staged toolchain reports rustc/cargo 1.95.0 and the port
produces a 130M rust-1.95.0.mport.  bmake test is skipped (NO_TEST=yes).
The checksums for the other seven architectures come from the FreeBSD
distinfo and were not verified locally.

The build required two manual restarts after transient SIGSEGVs in the
1.94.0 bootstrap cargo (at rust-analyzer-proc-macro-srv and again at
generate-copyright).  There was no memory pressure, the exact failing
invocations succeeded on retry, and repeated manual runs of the same
cargo command never crashed, so this looks like host flakiness rather
than a port defect - but a cluster build may still report a failure at
one of those steps.

lang/rust-bootstrap and the cargo.mk minimum version pin are left alone
on purpose; they are separate concerns and raising the Mk pin would
force unrelated rebuilds.

AI-Assisted-by: Claude Opus 5 <noreply at anthropic.com>
Signed-off-by: Lucas Holt <luke at foolishgames.com>



To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications

From noreply at github.com  Tue Aug 18 14:11:46 2026
From: noreply at github.com (Lucas Holt)
Date: Tue, 18 Aug 2026 11:11:46 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports] f8dc27: lang/rust: update to
 1.95.0 (#754)
Message-ID: <MidnightBSD/mports/push/refs/heads/master/3e8bc4-f8dc27@github.com>

  Branch: refs/heads/master
  Home:   https://github.com/MidnightBSD/mports
  Commit: f8dc2742ccf4a26aa7c62f55414f96d42fc7cd70
      https://github.com/MidnightBSD/mports/commit/f8dc2742ccf4a26aa7c62f55414f96d42fc7cd70
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-18 (Tue, 18 Aug 2026)

  Changed paths:
    M lang/rust/Makefile
    M lang/rust/distinfo
    M lang/rust/files/patch-AddLLVM.cmake
    M lang/rust/files/patch-vendor_cc.rs

  Log Message:
  -----------
  lang/rust: update to 1.95.0 (#754)

Updates `lang/rust` from 1.94.0 to 1.95.0, mirroring FreeBSD ports
commit `3c971390fa39`.

## Changes

- `PORTVERSION` 1.94.0 -> 1.95.0, `PORTREVISION` 2 -> 0
- `BOOTSTRAPS_DATE` 2026-01-22 -> 2026-03-05, `RUST_BOOTSTRAP_VERSION`
1.93.0 -> 1.94.0 (matches `src/stage0`)
- `distinfo` taken from FreeBSD's 1.95 distinfo; their 1.94 distinfo was
byte-identical to ours apart from the timestamp, so all eight
architectures transfer safely
- `files/patch-AddLLVM.cmake` retargeted: upstream LLVM now quotes
`"${CMAKE_SYSTEM_NAME}"` in the rpath origin check and the line moved
from 2552 to 2568
- `files/patch-vendor_cc.rs` retargeted: the vendored cc crate moved
from 1.2.50 to 1.2.55; verified all ten vendored `cc-*` directories are
still covered

Announce: https://blog.rust-lang.org/2026/04/16/Rust-1.95.0/
ChangeLog: https://github.com/rust-lang/rust/releases/tag/1.95.0

## Validation (amd64)

| Step | Result |
|---|---|
| `bmake checksum` | OK (amd64 distfiles + bootstraps) |
| `bmake patch` | clean from a clean tree |
| `bmake` | completed successfully |
| `bmake fake` | clean; staged `rustc 1.95.0 (59807616e)` / `cargo
1.95.0 (f2d3ce0bd)` |
| `bmake package` | `rust-1.95.0.mport`, 130M |
| `bmake test` | skipped (`NO_TEST=yes`) |
| `portlint -A` | 22 warnings + 1 fatal, all pre-existing on untouched
files; the fatal is only the `work` directory being present |

Checksums for the other seven architectures come from the FreeBSD
distinfo and were not verified locally.

## Note for the build cluster

The build needed two manual restarts after transient SIGSEGVs in the
1.94.0 bootstrap cargo, first at `rust-analyzer-proc-macro-srv` and
later at `generate-copyright`. There was no memory pressure (32G RAM and
34G swap both idle), the exact failing invocations succeeded when rerun
by hand, and eight repeated manual runs of the same `cargo metadata`
command never crashed. This looks like host flakiness in the
FreeBSD-built bootstrap cargo rather than a port defect, but Magus does
not self-resume, so a cluster run may still report a failure at one of
those steps.

## Deliberately out of scope

`lang/rust-bootstrap` (still 1.94.0) and the `Mk/extensions/cargo.mk`
minimum version pin (`>=1.87.0`). FreeBSD bumps both in the same commit,
but they are separate ports/concerns here and raising the Mk pin would
force unrelated rebuilds. Neither `lang/rust-nightly` nor `lang/rust188`
inherits these variables, so both are unaffected.

? Generated with [Claude Code](https://claude.com/claude-code)

## Summary by Sourcery

Update the lang/rust port to Rust 1.95.0 and align its bootstrap data,
checksums, and platform patches with the new release.

Enhancements:
- Update the Rust compiler port to version 1.95.0 with the corresponding
bootstrap toolchain and source metadata.
- Refresh platform-specific patches for the updated LLVM and vendored cc
sources while retaining MidnightBSD support.

Build:
- Refresh Rust 1.95.0 distfiles and checksums.

Signed-off-by: Lucas Holt <luke at foolishgames.com>



To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications


From noreply at github.com  Tue Aug 18 14:19:09 2026
From: noreply at github.com (Lucas Holt)
Date: Tue, 18 Aug 2026 18:19:09 +0000 (UTC)
Subject: [Midnightbsd-cvs] [MidnightBSD/mports] a0bc59: lang/rust-bootstrap:
 update to 1.95.0
Message-ID: <MidnightBSD/mports/push/refs/heads/lang/rust-bootstrap/000000-a0bc59@github.com>

  Branch: refs/heads/lang/rust-bootstrap
  Home:   https://github.com/MidnightBSD/mports
  Commit: a0bc59139776ef6c5abcc175124a665994b74d65
      https://github.com/MidnightBSD/mports/commit/a0bc59139776ef6c5abcc175124a665994b74d65
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-18 (Tue, 18 Aug 2026)

  Changed paths:
    M lang/rust-bootstrap/Makefile
    M lang/rust-bootstrap/distinfo
    M lang/rust-bootstrap/files/patch-AddLLVM.cmake
    M lang/rust-bootstrap/files/patch-vendor_cc.rs

  Log Message:
  -----------
  lang/rust-bootstrap: update to 1.95.0

Follow lang/rust to 1.95.0 so the port keeps producing bootstrap
tarballs that match the compiler it bootstraps, and drop the now stale
PORTREVISION.

Sync the two patches that needed retargeting for the 1.95 sources, both
identical to the copies committed in lang/rust:

- patch-AddLLVM.cmake: upstream LLVM now quotes "${CMAKE_SYSTEM_NAME}"
  in the rpath origin check and the line moved.
- patch-vendor_cc.rs: the vendored cc crate moved from 1.2.50 to 1.2.55.

FreeBSD's rust-bootstrap distinfo for 1.95.0 records a different
rustc-1.95.0-src.tar.xz than their own lang/rust distinfo, because they
build bootstraps from the pre-release tarball.  Both ports here share
DIST_SUBDIR=rust, so use the released tarball that lang/rust already
uses to keep one checksum per distfile.

Validated on amd64: checksum OK, extract and patch apply cleanly with no
rejects, and configure produces the expected config.toml.  portlint
reports 15 warnings, all pre-existing on files this commit does not
touch.  A full build of the amd64 flavor was still running when this was
committed and the i386 flavor was not built.

AI-Assisted-by: Claude Opus 5 <noreply at anthropic.com>
Signed-off-by: Lucas Holt <luke at foolishgames.com>



To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications

From noreply at github.com  Tue Aug 18 14:40:57 2026
From: noreply at github.com (Lucas Holt)
Date: Tue, 18 Aug 2026 11:40:57 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports] eefc4e: lang/rust-bootstrap:
 update to 1.95.0 (#755)
Message-ID: <MidnightBSD/mports/push/refs/heads/master/f8dc27-eefc4e@github.com>

  Branch: refs/heads/master
  Home:   https://github.com/MidnightBSD/mports
  Commit: eefc4ef445937c21adafa1a6649e27795562431e
      https://github.com/MidnightBSD/mports/commit/eefc4ef445937c21adafa1a6649e27795562431e
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-18 (Tue, 18 Aug 2026)

  Changed paths:
    M lang/rust-bootstrap/Makefile
    M lang/rust-bootstrap/distinfo
    M lang/rust-bootstrap/files/patch-AddLLVM.cmake
    M lang/rust-bootstrap/files/patch-vendor_cc.rs

  Log Message:
  -----------
  lang/rust-bootstrap: update to 1.95.0 (#755)

Follows #754 by updating `lang/rust-bootstrap` from 1.94.0 to 1.95.0, so
the port keeps producing bootstrap tarballs matching the compiler it
bootstraps.

## Changes

- `PORTVERSION` 1.94.0 -> 1.95.0, dropped the now stale `PORTREVISION=1`
- `distinfo` updated for `rustc-1.95.0-src.tar.xz`
- `files/patch-AddLLVM.cmake` and `files/patch-vendor_cc.rs` synced from
`lang/rust`; both needed retargeting for the 1.95 sources (LLVM now
quotes `"${CMAKE_SYSTEM_NAME}"` and the line moved; the vendored cc
crate moved 1.2.50 -> 1.2.55)

## Note on the distfile

FreeBSD's rust-bootstrap distinfo for 1.95.0 records a *different*
`rustc-1.95.0-src.tar.xz` (`deb399b9?`, 238942468 bytes) than their own
lang/rust distinfo (`62b67230?`, 238952004 bytes) ? upstream re-rolled
the tarball, and they build bootstraps from the pre-release copy. Both
ports here share `DIST_SUBDIR=rust`, so this uses the released tarball
that `lang/rust` already carries, keeping one checksum per distfile.

## Validation (amd64)

| Step | Result |
|---|---|
| `bmake checksum` | OK |
| `bmake extract` / `bmake patch` | clean, no rejects |
| `bmake configure` | OK, expected `config.toml` generated |
| `portlint` | 15 warnings, all pre-existing on files this PR does not
touch |

A full build of the amd64 flavor was still running when this was opened;
I will comment with the result. The i386 flavor was not built. Note that
the identical sources and patch set built end to end in #754.

? Generated with [Claude Code](https://claude.com/claude-code)

## Summary by Sourcery

Update the Rust bootstrap port to 1.95.0 while keeping its
FreeBSD-specific patches compatible with the new sources.

Enhancements:
- Update the Rust bootstrap port to version 1.95.0 and align its source
patches with the newer Rust and LLVM sources.

Chores:
- Refresh the bootstrap source checksum metadata and remove the obsolete
port revision.

Signed-off-by: Lucas Holt <luke at foolishgames.com>



To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications


From noreply at github.com  Tue Aug 18 14:41:36 2026
From: noreply at github.com (Lucas Holt)
Date: Tue, 18 Aug 2026 11:41:36 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports]
Message-ID: <MidnightBSD/mports/push/refs/heads/lang/rust-bootstrap/a0bc59-000000@github.com>

  Branch: refs/heads/lang/rust-bootstrap
  Home:   https://github.com/MidnightBSD/mports

To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications

From noreply at github.com  Tue Aug 18 14:11:58 2026
From: noreply at github.com (Lucas Holt)
Date: Tue, 18 Aug 2026 18:11:58 +0000 (UTC)
Subject: [Midnightbsd-cvs] [MidnightBSD/mports]
Message-ID: <MidnightBSD/mports/push/refs/heads/lang/rust/73ba9d-000000@github.com>

  Branch: refs/heads/lang/rust
  Home:   https://github.com/MidnightBSD/mports

To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications

From noreply at github.com  Tue Aug 18 17:01:48 2026
From: noreply at github.com (Lucas Holt)
Date: Tue, 18 Aug 2026 14:01:48 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports] e5b07e: lang/rust-bootstrap:
 build the cargo bootstrap tar...
Message-ID: <MidnightBSD/mports/push/refs/heads/lang/rust-bootstrap-cargo/000000-e5b07e@github.com>

  Branch: refs/heads/lang/rust-bootstrap-cargo
  Home:   https://github.com/MidnightBSD/mports
  Commit: e5b07e2e3a5f7db3a26234786932936ad8f4d664
      https://github.com/MidnightBSD/mports/commit/e5b07e2e3a5f7db3a26234786932936ad8f4d664
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-18 (Tue, 18 Aug 2026)

  Changed paths:
    M lang/rust-bootstrap/Makefile

  Log Message:
  -----------
  lang/rust-bootstrap: build the cargo bootstrap tarball

lang/rust bootstraps from three distfiles - rustc, rust-std and cargo -
but this port ran x.py dist with extended=false and no tools list, so it
never produced a cargo tarball.  The bootstraps it shipped were not
enough to bump lang/rust on their own and cargo still had to come from
the FreeBSD mirror.  Set tools=["cargo"] and extended=true so cargo is
built and packaged alongside the other two.

extended=true also emits the combined rust-${PORTVERSION} installer, so
drop it in do-install along with the rust-dev and rustc-dev tarballs,
which lang/rust never fetches.  Restrict [dist] to the xz format while
here; only the xz tarballs are installed and building the gzip copies of
everything was pure overhead.

The package now carries exactly cargo, rustc and rust-std and shrinks
from 469M to 75M.

Validated on the amd64 flavor: configure, build, fake and package all
succeed.  The resulting tarballs were then exercised the way
lang/rust's do-configure uses them - each install.sh run against a
scratch prefix - and the installed toolchain reports rustc and cargo
1.95.0 and builds and runs a small crate.  portlint reports 15
warnings, all pre-existing and unrelated.  The i386 flavor was not
built.

AI-Assisted-by: Claude Opus 5 <noreply at anthropic.com>
Signed-off-by: Lucas Holt <luke at foolishgames.com>



To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications

From noreply at github.com  Fri Aug 21 19:25:42 2026
From: noreply at github.com (Lucas Holt)
Date: Fri, 21 Aug 2026 16:25:42 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports] 3f4ac7: editors/emacs: fix
 CVE-2026-6861 (SVG stylesheet h...
Message-ID: <MidnightBSD/mports/push/refs/heads/editors/emacs/000000-3f4ac7@github.com>

  Branch: refs/heads/editors/emacs
  Home:   https://github.com/MidnightBSD/mports
  Commit: 3f4ac782e4d1a3c58ede59d71ad987837f5f1ae5
      https://github.com/MidnightBSD/mports/commit/3f4ac782e4d1a3c58ede59d71ad987837f5f1ae5
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-21 (Fri, 21 Aug 2026)

  Changed paths:
    M editors/emacs/Makefile
    A editors/emacs/files/patch-src_image.c

  Log Message:
  -----------
  editors/emacs: fix CVE-2026-6861 (SVG stylesheet heap overflow)

svg_load_image() in src/image.c wrote its NUL terminator one byte past
the end of the buffer allocated for an SVG's inline stylesheet data:

    css = xmalloc (SBYTES (lcss) + 1);
    strncpy (css, SSDATA (lcss), SBYTES (lcss));
    *(css + SBYTES (lcss) + 1) = 0;

Opening a crafted SVG file therefore corrupts the heap, which can crash
Emacs or disclose adjacent memory. Affects 28.1 through 30.2; there is
no fixed stable release upstream, so carry the fix as a port patch.

Only reachable when built with SVG support (librsvg), which is on by
default via OPTIONS_DEFAULT.

Verified the unpatched line is present in the 30.2 distfile, that the
patch applies cleanly, and that src/image.c compiles and links with
SVG support enabled.

Upstream bug:	https://debbugs.gnu.org/cgi/bugreport.cgi?bug=80851
Obtained from:	FreeBSD ports 915d199e510d
Security:	CVE-2026-6861

AI-Assisted-by: Claude Opus 5 <noreply at anthropic.com>
Signed-off-by: Lucas Holt <luke at foolishgames.com>



To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications

From noreply at github.com  Fri Aug 21 20:20:31 2026
From: noreply at github.com (Lucas Holt)
Date: Fri, 21 Aug 2026 17:20:31 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports] ff7a3b: lang/rust-bootstrap:
 build the cargo bootstrap tar...
Message-ID: <MidnightBSD/mports/push/refs/heads/master/eefc4e-ff7a3b@github.com>

  Branch: refs/heads/master
  Home:   https://github.com/MidnightBSD/mports
  Commit: ff7a3b1b7ef433b3676abaa6c540fab71874e5ea
      https://github.com/MidnightBSD/mports/commit/ff7a3b1b7ef433b3676abaa6c540fab71874e5ea
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-21 (Fri, 21 Aug 2026)

  Changed paths:
    M lang/rust-bootstrap/Makefile

  Log Message:
  -----------
  lang/rust-bootstrap: build the cargo bootstrap tarball (#756)

Follow-up to #755, fixing the gap noted there: `lang/rust` bootstraps
from **three** distfiles ? `rustc`, `rust-std` and `cargo` ? but this
port ran `x.py dist` with `extended=false` and no `tools` list, so it
never produced a cargo tarball. The bootstraps it shipped were not
enough to bump `lang/rust` on their own; cargo still had to come from
the FreeBSD mirror.

## Changes

- `tools=["cargo"]` and `extended=true` so cargo is built and packaged
with the other two
- `do-install` drops the combined `rust-${PORTVERSION}` installer (a
side effect of `extended=true`) plus the `rust-dev` / `rustc-dev`
tarballs, which `lang/rust` never fetches ? same set FreeBSD keeps
- `[dist] compression-formats=["xz"]`, since only the xz tarballs are
installed and building gzip copies of everything was pure overhead

The package now carries exactly `cargo`, `rustc` and `rust-std`, and
shrinks from **469M to 75M**.

## Validation (amd64 flavor)

| Step | Result |
|---|---|
| `bmake configure` | expected `config.toml` |
| `bmake` | Build completed successfully in 0:43:45 |
| `bmake fake` | stages exactly cargo + rustc + rust-std |
| `bmake package` | `amd64-rust-bootstrap-1.95.0.mport`, 75M |
| `portlint` | 15 warnings, all pre-existing and unrelated |

### Bootstrap functional test

The produced tarballs were exercised the way `lang/rust`'s
`do-configure` uses them ? each extracted and run through its own
`install.sh --prefix=<scratch>` ? and then used as a real toolchain:

```
$ cargo-1.95.0-x86_64-unknown-freebsd/install.sh --prefix=<scratch>   # cargo installed.
$ rust-std-1.95.0-x86_64-unknown-freebsd/install.sh --prefix=<scratch> # rust std installed.
$ rustc-1.95.0-x86_64-unknown-freebsd/install.sh --prefix=<scratch>    # rustc installed.

$ <scratch>/bin/rustc --version
rustc 1.95.0 (59807616e 2026-04-14) (built from a source tarball)
$ <scratch>/bin/cargo --version
cargo 1.95.0 (f2d3ce0bd 2026-03-21) (built from a source tarball)

$ <scratch>/bin/cargo init --name bstest && cargo build --offline && ./target/debug/bstest
Hello, world!
```

So the bootstrap set installs cleanly and cargo drives rustc to compile
and run a crate.

The i386 flavor was not built.

? Generated with [Claude Code](https://claude.com/claude-code)

## Summary by Sourcery

Build a complete, slimmer Rust bootstrap tarball set containing only the
cargo, rustc, and rust-std artifacts required by lang/rust.

New Features:
- Build and package Cargo alongside rustc and rust-std for Rust
bootstrap consumers.

Bug Fixes:
- Ensure the bootstrap package contains the complete set of distfiles
required to independently update lang/rust.

Enhancements:
- Limit generated and installed bootstrap artifacts to the required
xz-compressed cargo, rustc, and rust-std tarballs.

Signed-off-by: Lucas Holt <luke at foolishgames.com>



To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications


From noreply at github.com  Fri Aug 21 20:20:32 2026
From: noreply at github.com (Lucas Holt)
Date: Fri, 21 Aug 2026 17:20:32 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports]
Message-ID: <MidnightBSD/mports/push/refs/heads/lang/rust-bootstrap-cargo/e5b07e-000000@github.com>

  Branch: refs/heads/lang/rust-bootstrap-cargo
  Home:   https://github.com/MidnightBSD/mports

To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications

From noreply at github.com  Fri Aug 21 20:20:52 2026
From: noreply at github.com (Lucas Holt)
Date: Fri, 21 Aug 2026 17:20:52 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports] b0f1f5: editors/emacs: fix
 CVE-2026-6861 (SVG stylesheet h...
Message-ID: <MidnightBSD/mports/push/refs/heads/master/ff7a3b-b0f1f5@github.com>

  Branch: refs/heads/master
  Home:   https://github.com/MidnightBSD/mports
  Commit: b0f1f5fc8dc12b76c14168b0f7384ef4abf01def
      https://github.com/MidnightBSD/mports/commit/b0f1f5fc8dc12b76c14168b0f7384ef4abf01def
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-21 (Fri, 21 Aug 2026)

  Changed paths:
    M editors/emacs/Makefile
    A editors/emacs/files/patch-src_image.c

  Log Message:
  -----------
  editors/emacs: fix CVE-2026-6861 (SVG stylesheet heap overflow) (#757)

## Summary

`svg_load_image()` in `src/image.c` writes its NUL terminator one byte
past the end of the buffer allocated for an SVG's inline stylesheet
data:

```c
css = xmalloc (SBYTES (lcss) + 1);
strncpy (css, SSDATA (lcss), SBYTES (lcss));
*(css + SBYTES (lcss) + 1) = 0;   /* one past the allocation */
```

Opening a crafted SVG corrupts the heap, which can crash Emacs or
disclose adjacent memory.

- **CVE-2026-6861** ? affects 28.1 through 30.2 inclusive
- Only reachable when built with SVG support (librsvg), which is on by
default via `OPTIONS_DEFAULT`
- Upstream bug: https://debbugs.gnu.org/cgi/bugreport.cgi?bug=80851
- Obtained from: FreeBSD ports
[`915d199e510d`](https://cgit.freebsd.org/ports/commit/?id=915d199e510d)

There is no fixed stable release upstream ? `ftp.gnu.org/gnu/emacs/`
still tops out at 30.2, so the fix is carried as a port patch rather
than a version bump.

## Changes

- Add `files/patch-src_image.c`
- Bump `PORTREVISION` 5 ? 6

## Testing

- Confirmed the unpatched line is present verbatim in the 30.2 distfile
- `bmake patch` applies cleanly; result verified in `WRKSRC/src/image.c`
- `src/image.c` compiles without errors and `src/emacs` links, with
librsvg SVG support enabled

`portlint` reports 3 fatal + 2 warnings, all pre-existing and unrelated
to this change (the `COMMENT_nox`/`COMMENT_canna`/`COMMENT_wayland`
placement complaints are present in FreeBSD's port too).

## Not included

- The FreeBSD tree-sitter 0.26+ backport ? mports `devel/tree-sitter` is
0.25.5, so it is inapplicable
- The 31.0.91/31.1-rc1 pretest bump ? those are pretest/RC, not stable
releases

? Generated with [Claude Code](https://claude.com/claude-code)

## Summary by Sourcery

Patch Emacs 30.2 to prevent heap corruption when loading crafted SVG
images.

Bug Fixes:
- Fix the SVG inline stylesheet heap overflow in Emacs by correcting the
NUL terminator placement.

Chores:
- Bump the Emacs port revision to include the security fix.

Signed-off-by: Lucas Holt <luke at foolishgames.com>



To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications


From noreply at github.com  Fri Aug 21 20:20:53 2026
From: noreply at github.com (Lucas Holt)
Date: Fri, 21 Aug 2026 17:20:53 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports]
Message-ID: <MidnightBSD/mports/push/refs/heads/editors/emacs/3f4ac7-000000@github.com>

  Branch: refs/heads/editors/emacs
  Home:   https://github.com/MidnightBSD/mports

To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications

From noreply at github.com  Fri Aug 21 20:48:48 2026
From: noreply at github.com (Lucas Holt)
Date: Fri, 21 Aug 2026 17:48:48 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports] 1e92d7: ai/codex:
 de-duplicate distinfo
Message-ID: <MidnightBSD/mports/push/refs/heads/ai/codex/000000-1e92d7@github.com>

  Branch: refs/heads/ai/codex
  Home:   https://github.com/MidnightBSD/mports
  Commit: 1e92d7e83aa3cbc8d81da682eef07f68ce3fffff
      https://github.com/MidnightBSD/mports/commit/1e92d7e83aa3cbc8d81da682eef07f68ce3fffff
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-21 (Fri, 21 Aug 2026)

  Changed paths:
    M ai/codex/distinfo

  Log Message:
  -----------
  ai/codex: de-duplicate distinfo

distinfo contained the entire distfile list twice: 2424 SHA256 entries for
1212 distinct files, the second block byte-identical to the first. It was
introduced in 017ad5ddb3 (0.147); the file was clean at e292b391b8 (0.146.0),
so makesum appended rather than replaced, or ran twice.

Duplicated entries break fetching, because DISTINFO_DATA (Mk/bsd.mport.mk)
prints every matching line:

    CKSIZE=$(distinfo_data SIZE "${full_file}")   -> "17438 17438"

Mk/scripts/do-fetch.sh then interpolates that unquoted into "-S ${CKSIZE}",
so fetch(1) receives a stray extra argument. Because the crate paths contain
a '/', do-fetch.sh also passes "-o <path>", and fetch requires -o to name a
directory once it sees more than one URL:

    fetch: rust/crates/Inflector-0.11.4.crate is not a directory

Every mirror then fails identically, which is how this surfaced in Magus
run 647 as a fetch failure rather than a malformed distinfo.

Keeping only the first block leaves the file byte-identical to a correct
makesum run. Verified the 1212 remaining entries match DISTFILES exactly
(none missing, none extra) and that "bmake checksum" passes for all of them.

AI-Assisted-by: Claude Opus 5 <noreply at anthropic.com>
Signed-off-by: Lucas Holt <luke at foolishgames.com>



To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications

From noreply at github.com  Fri Aug 21 20:53:50 2026
From: noreply at github.com (Lucas Holt)
Date: Fri, 21 Aug 2026 17:53:50 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports] 4f0a2b: java/openjdk23: fix
 duplicated GH distfile in DIST...
Message-ID: <MidnightBSD/mports/push/refs/heads/java/openjdk23/000000-4f0a2b@github.com>

  Branch: refs/heads/java/openjdk23
  Home:   https://github.com/MidnightBSD/mports
  Commit: 4f0a2beed7a9663107513178b166ae7c3e1b0ae5
      https://github.com/MidnightBSD/mports/commit/4f0a2beed7a9663107513178b166ae7c3e1b0ae5
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-21 (Fri, 21 Aug 2026)

  Changed paths:
    M java/openjdk23/Makefile
    M java/openjdk23/distinfo

  Log Message:
  -----------
  java/openjdk23: fix duplicated GH distfile in DISTFILES and distinfo

DISTFILES hardcoded battleblow-jdk23u-jdk-23.0.2+7-1_GH0.tar.gz alongside
${BOOTSTRAPJDKFILE}, but USE_GITHUB with GH_ACCOUNT/GH_PROJECT/GH_TAGNAME
already appends exactly that filename. The distfile was therefore listed
twice, and makesum faithfully recorded it twice in distinfo.

Duplicated entries break fetching, because DISTINFO_DATA (Mk/bsd.mport.mk)
prints every matching line:

    CKSIZE=$(distinfo_data SIZE "${full_file}")   -> "117326728 117326728"

Mk/scripts/do-fetch.sh interpolates that unquoted into "-S ${CKSIZE}", so
fetch(1) sees a stray extra argument and reports

    fetch: 117326728: No such file or directory

on every mirror. The size check at do-fetch.sh:152 breaks the same way
("[ ... -eq ... ]" with too many arguments), so even a completed download is
rejected as a size mismatch. That is how this surfaced in Magus run 647 as a
fetch failure. The bootstrap JDK, which is not duplicated, fetched fine in the
same log.

Dropping the hardcoded line fixes the cause rather than the symptom: editing
distinfo alone would be undone by the next makesum. It also removes a
hand-maintained copy of the tag that would go stale on the next update.
FreeBSD's java/openjdk21 uses "DISTFILES= ${BOOTSTRAPJDKFILE}:boot" for the
same reason.

Verified DISTFILES now lists each file once, distinfo matches it exactly, and
"bmake fetch" and "bmake checksum" both succeed.

AI-Assisted-by: Claude Opus 5 <noreply at anthropic.com>
Signed-off-by: Lucas Holt <luke at foolishgames.com>



To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications

From noreply at github.com  Fri Aug 21 20:56:37 2026
From: noreply at github.com (Lucas Holt)
Date: Fri, 21 Aug 2026 17:56:37 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports] 4b6f0c:
 Mk/scripts/do-fetch.sh: diagnose duplicated distin...
Message-ID: <MidnightBSD/mports/push/refs/heads/mk/do-fetch-cksize/000000-4b6f0c@github.com>

  Branch: refs/heads/mk/do-fetch-cksize
  Home:   https://github.com/MidnightBSD/mports
  Commit: 4b6f0c769d3419282e25121bca13223d92911b22
      https://github.com/MidnightBSD/mports/commit/4b6f0c769d3419282e25121bca13223d92911b22
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-21 (Fri, 21 Aug 2026)

  Changed paths:
    M Mk/scripts/do-fetch.sh

  Log Message:
  -----------
  Mk/scripts/do-fetch.sh: diagnose duplicated distinfo entries

distinfo_data() prints one line per matching entry, so a distfile listed
twice in distinfo makes CKSIZE two values instead of one:

    CKSIZE=$(distinfo_data SIZE "${full_file}")   -> "117326728 117326728"

CKSIZE is interpolated unquoted into "-S ${CKSIZE}", so it word-splits into
fetch(1)'s argument list and the second value is taken as an extra URL. The
result is a failure on every mirror with an error that points nowhere near
the real problem:

    fetch: 117326728: No such file or directory
    fetch: rust/crates/Inflector-0.11.4.crate is not a directory
    => Couldn't fetch it - please try to retrieve this port manually

The size comparison further down breaks the same way, because
[ "${actual_size}" -eq "${CKSIZE}" ] gets too many arguments, so even a
completed download is discarded as a size mismatch.

Validate CKSIZE before it reaches the command line and report the actual
cause instead:

    => Invalid SIZE for zipmix_src.zip in .../distinfo.
    => Expected a single integer, got: 4759 4759
    => The entry is most likely present more than once.

The check sits inside the existing "-z dp_DISABLE_SIZE" guard, so makesum,
which runs with DISABLE_SIZE=yes and NO_CHECKSUM=yes, is unaffected and
remains usable to regenerate a duplicated distinfo.

Not changed: checksum.sh iterates "for chksum in $CKSUM" deliberately to
allow several recorded checksums, and the _sha256sum lookup here is only
tested for emptiness. Neither is affected by duplicates.

Tested with archivers/zipmix against a scratch DISTDIR: clean distinfo
fetches normally, a duplicated one now reports the message above instead of
"Couldn't fetch it", and makesum still rewrites a duplicated distinfo
correctly.

AI-Assisted-by: Claude Opus 5 <noreply at anthropic.com>
Signed-off-by: Lucas Holt <luke at foolishgames.com>



To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications

From noreply at github.com  Fri Aug 21 23:45:56 2026
From: noreply at github.com (Lucas Holt)
Date: Fri, 21 Aug 2026 20:45:56 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports] 6bb045: ai/codex:
 de-duplicate distinfo (#758)
Message-ID: <MidnightBSD/mports/push/refs/heads/master/b0f1f5-6bb045@github.com>

  Branch: refs/heads/master
  Home:   https://github.com/MidnightBSD/mports
  Commit: 6bb04539357549551523f4543aa1e3a1ab9da3cf
      https://github.com/MidnightBSD/mports/commit/6bb04539357549551523f4543aa1e3a1ab9da3cf
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-21 (Fri, 21 Aug 2026)

  Changed paths:
    M ai/codex/distinfo

  Log Message:
  -----------
  ai/codex: de-duplicate distinfo (#758)

## Problem

`ai/codex/distinfo` contained the entire distfile list **twice** ? 2424
`SHA256` entries for 1212 distinct files, with the second block
byte-identical to the first.

Introduced in `017ad5ddb3` (0.147). The file was clean at `e292b391b8`
(0.146.0, 1169 entries), so `makesum` appended rather than replaced, or
was run twice.

## Why it breaks fetching

`DISTINFO_DATA` in `Mk/bsd.mport.mk:2335` prints *every* matching line,
so a duplicated entry yields two values:

```sh
CKSIZE=$(distinfo_data SIZE "${full_file}")   # -> "17438 17438"
```

`Mk/scripts/do-fetch.sh:144` interpolates that unquoted:

```sh
_fetch_cmd="${_fetch_cmd} -S ${CKSIZE}"
```

so `fetch(1)` gets a stray extra argument. Because the crate paths
contain a `/`, `do-fetch.sh:130` also passes `-o <path>`, and `fetch`
requires `-o` to name a *directory* once it sees more than one URL:

```
fetch: rust/crates/Inflector-0.11.4.crate is not a directory
```

Every mirror then fails identically ? which is why this surfaced in
Magus run 647 as a *fetch* failure rather than a malformed distinfo.
`do-fetch.sh:152` is affected too: `[ "$actual_size" -eq "$CKSIZE" ]`
gets too many arguments, so even a successful download is rejected as a
size mismatch.

Reproduced standalone:

```console
$ fetch -S 5 5 -o /tmp/f https://example.com/
fetch: /tmp/f is not a directory
```

## Fix

Keep only the first block. The result is byte-identical to what a
correct `makesum` run produces.

## Testing

- Confirmed the two halves were byte-identical before removal, so the
de-dup is lossless
- The 1212 remaining entries match `bmake -V DISTFILES` exactly ? none
missing, none extra
- `bmake checksum` passes for all 1212 distfiles
- `distinfo_data SIZE` now returns a single value for both a crate path
and the GH tarball

A companion PR hardens `do-fetch.sh` so a duplicated distinfo fails
loudly instead of masquerading as a mirror outage.

? Generated with [Claude Code](https://claude.com/claude-code)

## Summary by Sourcery

Bug Fixes:
- Remove the duplicated distinfo entries so fetching and checksum
validation receive a single size and checksum value for each distfile.

Signed-off-by: Lucas Holt <luke at foolishgames.com>



To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications


From noreply at github.com  Fri Aug 21 23:45:56 2026
From: noreply at github.com (Lucas Holt)
Date: Fri, 21 Aug 2026 20:45:56 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports]
Message-ID: <MidnightBSD/mports/push/refs/heads/ai/codex/1e92d7-000000@github.com>

  Branch: refs/heads/ai/codex
  Home:   https://github.com/MidnightBSD/mports

To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications

From noreply at github.com  Fri Aug 21 23:46:44 2026
From: noreply at github.com (Lucas Holt)
Date: Fri, 21 Aug 2026 20:46:44 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports]
Message-ID: <MidnightBSD/mports/push/refs/heads/java/openjdk23/4f0a2b-000000@github.com>

  Branch: refs/heads/java/openjdk23
  Home:   https://github.com/MidnightBSD/mports

To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications

From noreply at github.com  Fri Aug 21 23:46:44 2026
From: noreply at github.com (Lucas Holt)
Date: Fri, 21 Aug 2026 20:46:44 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports] 4a4588: java/openjdk23: fix
 duplicated GH distfile in DIST...
Message-ID: <MidnightBSD/mports/push/refs/heads/master/6bb045-4a4588@github.com>

  Branch: refs/heads/master
  Home:   https://github.com/MidnightBSD/mports
  Commit: 4a4588c4675d454e30b2a5eb8d96545f8b2f7358
      https://github.com/MidnightBSD/mports/commit/4a4588c4675d454e30b2a5eb8d96545f8b2f7358
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-21 (Fri, 21 Aug 2026)

  Changed paths:
    M java/openjdk23/Makefile
    M java/openjdk23/distinfo

  Log Message:
  -----------
  java/openjdk23: fix duplicated GH distfile in DISTFILES and distinfo (#759)

## Problem

`DISTFILES` hardcoded `battleblow-jdk23u-jdk-23.0.2+7-1_GH0.tar.gz`
alongside `${BOOTSTRAPJDKFILE}`:

```make
DISTFILES=	${BOOTSTRAPJDKFILE}:boot \
		battleblow-jdk23u-jdk-23.0.2+7-1_GH0.tar.gz
```

But `USE_GITHUB=yes` with `GH_ACCOUNT`/`GH_PROJECT`/`GH_TAGNAME`
**already appends exactly that filename**. The distfile was listed
twice, and `makesum` faithfully recorded it twice in `distinfo`.

## Why it breaks fetching

`DISTINFO_DATA` in `Mk/bsd.mport.mk:2335` prints *every* matching line:

```sh
CKSIZE=$(distinfo_data SIZE "${full_file}")   # -> "117326728 117326728"
```

`Mk/scripts/do-fetch.sh:144` interpolates that unquoted into `-S
${CKSIZE}`, so `fetch(1)` receives a stray extra argument and reports,
on every mirror:

```
fetch: 117326728: No such file or directory
```

The size check at `do-fetch.sh:152` breaks the same way ? `[
"$actual_size" -eq "$CKSIZE" ]` with too many arguments ? so even a
*completed* 111 MB download is rejected as a size mismatch. That's how
this surfaced in Magus run 647 as a fetch failure rather than a
malformed distinfo.

The same log contains its own control: the bootstrap JDK, which is *not*
duplicated, fetched cleanly (171 MB) in the same invocation.

## Fix

Drop the hardcoded line. This fixes the cause rather than the symptom ?
editing `distinfo` alone would be undone by the next `makesum`. It also
removes a hand-maintained copy of the tag that would go stale on the
next update.

FreeBSD's `java/openjdk21` uses `DISTFILES= ${BOOTSTRAPJDKFILE}:boot`
for the same reason.

## Testing

- `bmake -V DISTFILES` now lists each file exactly once
- `distinfo` matches `DISTFILES` exactly ? none missing, none extra
- `bmake fetch` succeeds (exit 0); the `fetch: 117326728: ...` errors
are gone
- `bmake checksum` passes for both distfiles

## Unrelated observation

The bootstrap JDK is declared `MASTER_SITES=LOCAL/openjdk:boot`, but no
MidnightBSD mirror actually serves it ? all five `:boot` sites 404'd and
only `distcache.FreeBSD.org` had it. Fetching works, but the mirrors
look unseeded. Not addressed here.

? Generated with [Claude Code](https://claude.com/claude-code)

## Summary by Sourcery

Bug Fixes:
- Remove the duplicated OpenJDK 23 GitHub distfile entry so fetching and
checksum validation no longer fail due to repeated size metadata.

Signed-off-by: Lucas Holt <luke at foolishgames.com>



To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications


From noreply at github.com  Fri Aug 21 23:48:06 2026
From: noreply at github.com (Lucas Holt)
Date: Fri, 21 Aug 2026 20:48:06 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports] 11da37:
 Mk/scripts/do-fetch.sh: diagnose duplicated distin...
Message-ID: <MidnightBSD/mports/push/refs/heads/master/4a4588-11da37@github.com>

  Branch: refs/heads/master
  Home:   https://github.com/MidnightBSD/mports
  Commit: 11da37178655722145db4436f78cf5d30c812674
      https://github.com/MidnightBSD/mports/commit/11da37178655722145db4436f78cf5d30c812674
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-21 (Fri, 21 Aug 2026)

  Changed paths:
    M Mk/scripts/do-fetch.sh

  Log Message:
  -----------
  Mk/scripts/do-fetch.sh: diagnose duplicated distinfo entries (#760)

## Problem

`distinfo_data()` prints one line per matching entry, so a distfile
listed twice in `distinfo` makes `CKSIZE` **two** values instead of one:

```sh
CKSIZE=$(distinfo_data SIZE "${full_file}")   # -> "117326728 117326728"
```

`CKSIZE` is interpolated unquoted into `-S ${CKSIZE}`, so it word-splits
into `fetch(1)`'s argument list and the second value is taken as an
extra URL. Every mirror then fails with an error that points nowhere
near the real problem:

```
fetch: 117326728: No such file or directory
fetch: rust/crates/Inflector-0.11.4.crate is not a directory
=> Couldn't fetch it - please try to retrieve this port manually
```

The size comparison further down breaks the same way ? `[
"${actual_size}" -eq "${CKSIZE}" ]` gets too many arguments ? so even a
*completed* download is discarded as a size mismatch.

This cost real debugging time: two ports in Magus run 647 (`ai/codex`,
`java/openjdk23`) were reported as fetch failures that looked exactly
like mirror outages. Fixed separately in #758 and #759.

## Fix

Validate `CKSIZE` before it reaches the command line, and report the
actual cause:

```
=> Invalid SIZE for zipmix_src.zip in .../distinfo.
=> Expected a single integer, got: 4759 4759
=> The entry is most likely present more than once.
```

The check sits inside the existing `-z "${dp_DISABLE_SIZE}"` guard, so
**`makesum` is unaffected** ? it runs with `DISABLE_SIZE=yes
NO_CHECKSUM=yes` and remains usable to regenerate a duplicated distinfo.
That matters: a check placed outside the guard would break the very
command you'd use to recover.

## Deliberately not changed

- `checksum.sh` iterates `for chksum in $CKSUM` on purpose, to allow
several recorded checksums ? duplicates are harmless there.
- The `_sha256sum` lookup at line 60 is only tested for emptiness.

Neither is affected, so the `-S` path is the only genuinely broken call
site.

## Testing

With `archivers/zipmix` against a scratch `DISTDIR`:

| Case | Result |
|---|---|
| Clean distinfo | fetches normally, exit 0 (unchanged) |
| Duplicated distinfo, **without** this patch | `=> Couldn't fetch it -
please try to retrieve this port manually` |
| Duplicated distinfo, **with** this patch | the diagnostic above |
| `makesum` on a duplicated distinfo | still regenerates a clean
distinfo |

Also `sh -n` clean, and the `*[!0-9]*` pattern verified against
single-integer, empty, two-value, and non-numeric input.

? Generated with [Claude Code](https://claude.com/claude-code)

## Summary by Sourcery

Bug Fixes:
- Detect duplicated or otherwise invalid distinfo SIZE values before
fetching, replacing misleading fetch failures with a clear diagnostic.

Signed-off-by: Lucas Holt <luke at foolishgames.com>



To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications


From noreply at github.com  Fri Aug 21 23:48:08 2026
From: noreply at github.com (Lucas Holt)
Date: Fri, 21 Aug 2026 20:48:08 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports]
Message-ID: <MidnightBSD/mports/push/refs/heads/mk/do-fetch-cksize/4b6f0c-000000@github.com>

  Branch: refs/heads/mk/do-fetch-cksize
  Home:   https://github.com/MidnightBSD/mports

To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications

From noreply at github.com  Fri Aug 21 23:54:21 2026
From: noreply at github.com (Lucas Holt)
Date: Fri, 21 Aug 2026 20:54:21 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports] e90d5f:
 chinese/wordpress-zh_CN: regenerate distinfo for 7...
Message-ID: <MidnightBSD/mports/push/refs/heads/chinese/wordpress-zh_CN/000000-e90d5f@github.com>

  Branch: refs/heads/chinese/wordpress-zh_CN
  Home:   https://github.com/MidnightBSD/mports
  Commit: e90d5f29307bd9a7459fdd2f1b14a3c8583880b4
      https://github.com/MidnightBSD/mports/commit/e90d5f29307bd9a7459fdd2f1b14a3c8583880b4
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-21 (Fri, 21 Aug 2026)

  Changed paths:
    M chinese/wordpress-zh_CN/distinfo

  Log Message:
  -----------
  chinese/wordpress-zh_CN: regenerate distinfo for 7.0.3

Commit 6314d32ee7 updated the master port www/wordpress from 6.9 to 7.0.3 and
regenerated only www/wordpress/distinfo. This slave still carried the 6.9
checksums, so fetching failed:

    => wordpress-7.0.3-zh_CN.tar.gz is not in /usr/mports/chinese/wordpress-zh_CN/distinfo.
    => Either /usr/mports/chinese/wordpress-zh_CN/distinfo is out of date, or
    => wordpress-7.0.3-zh_CN.tar.gz is spelled incorrectly.

Reported by Magus run 647. All five www/wordpress slaves are affected; each is
fixed on its own branch.

No PORTREVISION bump: the version comes from the master port, which already
carries it.

Verified with bmake makesum followed by bmake checksum.

AI-Assisted-by: Claude Opus 5 <noreply at anthropic.com>
Signed-off-by: Lucas Holt <luke at foolishgames.com>



To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications

From noreply at github.com  Fri Aug 21 23:54:23 2026
From: noreply at github.com (Lucas Holt)
Date: Fri, 21 Aug 2026 20:54:23 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports] f6d534:
 chinese/wordpress-zh_TW: regenerate distinfo for 7...
Message-ID: <MidnightBSD/mports/push/refs/heads/chinese/wordpress-zh_TW/000000-f6d534@github.com>

  Branch: refs/heads/chinese/wordpress-zh_TW
  Home:   https://github.com/MidnightBSD/mports
  Commit: f6d5342d380dfc3f53d3b02672084b0ef34ab1c2
      https://github.com/MidnightBSD/mports/commit/f6d5342d380dfc3f53d3b02672084b0ef34ab1c2
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-21 (Fri, 21 Aug 2026)

  Changed paths:
    M chinese/wordpress-zh_TW/distinfo

  Log Message:
  -----------
  chinese/wordpress-zh_TW: regenerate distinfo for 7.0.3

Commit 6314d32ee7 updated the master port www/wordpress from 6.9 to 7.0.3 and
regenerated only www/wordpress/distinfo. This slave still carried the 6.9
checksums, so fetching failed:

    => wordpress-7.0.3-zh_TW.tar.gz is not in /usr/mports/chinese/wordpress-zh_TW/distinfo.
    => Either /usr/mports/chinese/wordpress-zh_TW/distinfo is out of date, or
    => wordpress-7.0.3-zh_TW.tar.gz is spelled incorrectly.

Reported by Magus run 647. All five www/wordpress slaves are affected; each is
fixed on its own branch.

No PORTREVISION bump: the version comes from the master port, which already
carries it.

Verified with bmake makesum followed by bmake checksum.

AI-Assisted-by: Claude Opus 5 <noreply at anthropic.com>
Signed-off-by: Lucas Holt <luke at foolishgames.com>



To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications

From noreply at github.com  Fri Aug 21 23:54:28 2026
From: noreply at github.com (Lucas Holt)
Date: Fri, 21 Aug 2026 20:54:28 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports] e49ff0: french/wordpress:
 regenerate distinfo for 7.0.3
Message-ID: <MidnightBSD/mports/push/refs/heads/french/wordpress/000000-e49ff0@github.com>

  Branch: refs/heads/french/wordpress
  Home:   https://github.com/MidnightBSD/mports
  Commit: e49ff0fa5f30da9ef1a2759beefdeb6f53fadd2e
      https://github.com/MidnightBSD/mports/commit/e49ff0fa5f30da9ef1a2759beefdeb6f53fadd2e
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-21 (Fri, 21 Aug 2026)

  Changed paths:
    M french/wordpress/distinfo

  Log Message:
  -----------
  french/wordpress: regenerate distinfo for 7.0.3

Commit 6314d32ee7 updated the master port www/wordpress from 6.9 to 7.0.3 and
regenerated only www/wordpress/distinfo. This slave still carried the 6.9
checksums, so fetching failed:

    => wordpress-7.0.3-fr_FR.tar.gz is not in /usr/mports/french/wordpress/distinfo.
    => Either /usr/mports/french/wordpress/distinfo is out of date, or
    => wordpress-7.0.3-fr_FR.tar.gz is spelled incorrectly.

Reported by Magus run 647. All five www/wordpress slaves are affected; each is
fixed on its own branch.

No PORTREVISION bump: the version comes from the master port, which already
carries it.

Verified with bmake makesum followed by bmake checksum.

AI-Assisted-by: Claude Opus 5 <noreply at anthropic.com>
Signed-off-by: Lucas Holt <luke at foolishgames.com>



To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications

From noreply at github.com  Fri Aug 21 23:54:38 2026
From: noreply at github.com (Lucas Holt)
Date: Fri, 21 Aug 2026 20:54:38 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports] 057014: german/wordpress:
 regenerate distinfo for 7.0.3
Message-ID: <MidnightBSD/mports/push/refs/heads/german/wordpress/000000-057014@github.com>

  Branch: refs/heads/german/wordpress
  Home:   https://github.com/MidnightBSD/mports
  Commit: 057014db1a3dbdca2420d6eb890dc7ae5e6bfe8e
      https://github.com/MidnightBSD/mports/commit/057014db1a3dbdca2420d6eb890dc7ae5e6bfe8e
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-21 (Fri, 21 Aug 2026)

  Changed paths:
    M german/wordpress/distinfo

  Log Message:
  -----------
  german/wordpress: regenerate distinfo for 7.0.3

Commit 6314d32ee7 updated the master port www/wordpress from 6.9 to 7.0.3 and
regenerated only www/wordpress/distinfo. This slave still carried the 6.9
checksums, so fetching failed:

    => wordpress-7.0.3-de_DE.tar.gz is not in /usr/mports/german/wordpress/distinfo.
    => Either /usr/mports/german/wordpress/distinfo is out of date, or
    => wordpress-7.0.3-de_DE.tar.gz is spelled incorrectly.

Reported by Magus run 647. All five www/wordpress slaves are affected; each is
fixed on its own branch.

No PORTREVISION bump: the version comes from the master port, which already
carries it.

Verified with bmake makesum followed by bmake checksum.

AI-Assisted-by: Claude Opus 5 <noreply at anthropic.com>
Signed-off-by: Lucas Holt <luke at foolishgames.com>



To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications

From noreply at github.com  Fri Aug 21 23:54:40 2026
From: noreply at github.com (Lucas Holt)
Date: Fri, 21 Aug 2026 20:54:40 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports] 428aa0: japanese/wordpress:
 regenerate distinfo for 7.0.3
Message-ID: <MidnightBSD/mports/push/refs/heads/japanese/wordpress/000000-428aa0@github.com>

  Branch: refs/heads/japanese/wordpress
  Home:   https://github.com/MidnightBSD/mports
  Commit: 428aa0a98ca75e7e9f026545fc84eca4ef9f3247
      https://github.com/MidnightBSD/mports/commit/428aa0a98ca75e7e9f026545fc84eca4ef9f3247
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-21 (Fri, 21 Aug 2026)

  Changed paths:
    M japanese/wordpress/distinfo

  Log Message:
  -----------
  japanese/wordpress: regenerate distinfo for 7.0.3

Commit 6314d32ee7 updated the master port www/wordpress from 6.9 to 7.0.3 and
regenerated only www/wordpress/distinfo. This slave still carried the 6.9
checksums, so fetching failed:

    => wordpress-7.0.3-ja.tar.gz is not in /usr/mports/japanese/wordpress/distinfo.
    => Either /usr/mports/japanese/wordpress/distinfo is out of date, or
    => wordpress-7.0.3-ja.tar.gz is spelled incorrectly.

Reported by Magus run 647. All five www/wordpress slaves are affected; each is
fixed on its own branch.

No PORTREVISION bump: the version comes from the master port, which already
carries it.

Verified with bmake makesum followed by bmake checksum.

AI-Assisted-by: Claude Opus 5 <noreply at anthropic.com>
Signed-off-by: Lucas Holt <luke at foolishgames.com>



To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications

From noreply at github.com  Fri Aug 21 23:55:09 2026
From: noreply at github.com (Lucas Holt)
Date: Fri, 21 Aug 2026 20:55:09 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports] f7a042: cad/cura: fix
 invalid LICENSE, set to lgpl3
Message-ID: <MidnightBSD/mports/push/refs/heads/cad/cura/000000-f7a042@github.com>

  Branch: refs/heads/cad/cura
  Home:   https://github.com/MidnightBSD/mports
  Commit: f7a042528d93716f47e541e6fb26f23880bd8f0e
      https://github.com/MidnightBSD/mports/commit/f7a042528d93716f47e541e6fb26f23880bd8f0e
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-21 (Fri, 21 Aug 2026)

  Changed paths:
    M cad/cura/Makefile

  Log Message:
  -----------
  cad/cura: fix invalid LICENSE, set to lgpl3

LICENSE was "apgl", which is not a recognized identifier. The framework
therefore treated it as a custom license and aborted before doing any work:

    ===>  License not correctly defined: for unknown licenses, defining
          LICENSE_PERMS is mandatory (otherwise use a known LICENSE)

Magus run 647 recorded this as a fetch failure, because the license check runs
inside the fetch target. Nothing was ever fetched.

"apgl" reads like a transposition of "agpl", and FreeBSD's cad/cura does use
AGPLv3, but that is not what this port builds. Cura relicensed at 4.0, and the
LICENSE file of Ultimaker/Cura 4.13.1 -- the very file LICENSE_FILE points at --
is the GNU Lesser General Public License version 3. So the correct identifier
is lgpl3, not agpl.

No PORTREVISION bump: the port could not build at all, so there is no package
whose metadata changes.

Verified with bmake check-license.

AI-Assisted-by: Claude Opus 5 <noreply at anthropic.com>
Signed-off-by: Lucas Holt <luke at foolishgames.com>



To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications

From noreply at github.com  Fri Aug 21 23:55:35 2026
From: noreply at github.com (Lucas Holt)
Date: Fri, 21 Aug 2026 20:55:35 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports] 70a054:
 japanese/p5-WWW-MobileCarrierJP: fix misspelled LI...
Message-ID: <MidnightBSD/mports/push/refs/heads/japanese/p5-WWW-MobileCarrierJP/000000-70a054@github.com>

  Branch: refs/heads/japanese/p5-WWW-MobileCarrierJP
  Home:   https://github.com/MidnightBSD/mports
  Commit: 70a0547c838bd1176d6d0f2b07381bbd4c8af416
      https://github.com/MidnightBSD/mports/commit/70a0547c838bd1176d6d0f2b07381bbd4c8af416
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-21 (Fri, 21 Aug 2026)

  Changed paths:
    M japanese/p5-WWW-MobileCarrierJP/Makefile

  Log Message:
  -----------
  japanese/p5-WWW-MobileCarrierJP: fix misspelled LICENSE identifier

LICENSE read "artitistic gpl+". "artitistic" is not a recognized identifier,
so the framework treated it as a custom license and aborted before doing any
work:

    ===>  License not correctly defined: for unknown licenses, defining
          LICENSE_PERMS_artitistic is mandatory (otherwise use a known LICENSE)

Magus run 647 recorded this as a fetch failure, because the license check runs
inside the fetch target. Nothing was ever fetched.

The intended value is "artistic". With LICENSE_COMB=dual this is the standard
Perl dual license, spelled "artistic gpl+" by every other p5-* port in the
tree.

No PORTREVISION bump: the port could not build at all, so there is no package
whose metadata changes.

Verified with bmake check-license.

AI-Assisted-by: Claude Opus 5 <noreply at anthropic.com>
Signed-off-by: Lucas Holt <luke at foolishgames.com>



To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications

From noreply at github.com  Fri Aug 21 23:55:49 2026
From: noreply at github.com (Lucas Holt)
Date: Fri, 21 Aug 2026 20:55:49 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports] 541122:
 textproc/gitlab-code-parser: fix LICENSE case
Message-ID: <MidnightBSD/mports/push/refs/heads/textproc/gitlab-code-parser/000000-541122@github.com>

  Branch: refs/heads/textproc/gitlab-code-parser
  Home:   https://github.com/MidnightBSD/mports
  Commit: 541122c949451c765b7f3042803677ca5fe12fc0
      https://github.com/MidnightBSD/mports/commit/541122c949451c765b7f3042803677ca5fe12fc0
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-21 (Fri, 21 Aug 2026)

  Changed paths:
    M textproc/gitlab-code-parser/Makefile

  Log Message:
  -----------
  textproc/gitlab-code-parser: fix LICENSE case

LICENSE was spelled "MIT". Identifiers are case-sensitive and the recognized
one is "mit", so the framework treated "MIT" as a custom license and aborted
before doing any work:

    ===>  License not correctly defined: for unknown licenses, defining
          LICENSE_PERMS is mandatory (otherwise use a known LICENSE)

Confirmed upstream: LICENSE.md, which LICENSE_FILE points at, is the MIT
license text.

Magus run 647 has not reached textproc yet, but this port will fail the same
way that cad/cura and japanese/p5-WWW-MobileCarrierJP already did -- as a
fetch failure, because the license check runs inside the fetch target.

No PORTREVISION bump: the port could not build at all, so there is no package
whose metadata changes.

Verified with bmake check-license.

AI-Assisted-by: Claude Opus 5 <noreply at anthropic.com>
Signed-off-by: Lucas Holt <luke at foolishgames.com>



To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications

From noreply at github.com  Fri Aug 21 23:55:52 2026
From: noreply at github.com (Lucas Holt)
Date: Fri, 21 Aug 2026 20:55:52 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports] be0c94:
 textproc/gitlab-elasticsearch-indexer: fix LICENSE...
Message-ID: <MidnightBSD/mports/push/refs/heads/textproc/gitlab-elasticsearch-indexer/000000-be0c94@github.com>

  Branch: refs/heads/textproc/gitlab-elasticsearch-indexer
  Home:   https://github.com/MidnightBSD/mports
  Commit: be0c94472ff780de00e3f4597000dbfbd15e7963
      https://github.com/MidnightBSD/mports/commit/be0c94472ff780de00e3f4597000dbfbd15e7963
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-21 (Fri, 21 Aug 2026)

  Changed paths:
    M textproc/gitlab-elasticsearch-indexer/Makefile

  Log Message:
  -----------
  textproc/gitlab-elasticsearch-indexer: fix LICENSE case

LICENSE was spelled "MIT". Identifiers are case-sensitive and the recognized
one is "mit", so the framework treated "MIT" as a custom license and aborted
before doing any work:

    ===>  License not correctly defined: for unknown licenses, defining
          LICENSE_PERMS is mandatory (otherwise use a known LICENSE)

Confirmed upstream: LICENSE.md, which LICENSE_FILE points at, is the MIT
license text.

Magus run 647 has not reached textproc yet, but this port will fail the same
way that cad/cura and japanese/p5-WWW-MobileCarrierJP already did -- as a
fetch failure, because the license check runs inside the fetch target.

No PORTREVISION bump: the port could not build at all, so there is no package
whose metadata changes.

Verified with bmake check-license.

AI-Assisted-by: Claude Opus 5 <noreply at anthropic.com>
Signed-off-by: Lucas Holt <luke at foolishgames.com>



To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications

From noreply at github.com  Fri Aug 21 23:57:33 2026
From: noreply at github.com (Lucas Holt)
Date: Fri, 21 Aug 2026 20:57:33 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports] 77192c:
 chinese/wordpress-zh_CN: regenerate distinfo for 7...
Message-ID: <MidnightBSD/mports/push/refs/heads/master/11da37-77192c@github.com>

  Branch: refs/heads/master
  Home:   https://github.com/MidnightBSD/mports
  Commit: 77192c1032df8fec186fdb05266e04d6a41f8414
      https://github.com/MidnightBSD/mports/commit/77192c1032df8fec186fdb05266e04d6a41f8414
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-21 (Fri, 21 Aug 2026)

  Changed paths:
    M chinese/wordpress-zh_CN/distinfo

  Log Message:
  -----------
  chinese/wordpress-zh_CN: regenerate distinfo for 7.0.3 (#761)

## Problem

Commit `6314d32ee7` updated the master port `www/wordpress` from 6.9 to
7.0.3 and regenerated only `www/wordpress/distinfo`. This Simplified
Chinese slave still carried the 6.9 checksums, so fetching failed:

```
=> wordpress-7.0.3-zh_CN.tar.gz is not in /usr/mports/chinese/wordpress-zh_CN/distinfo.
=> Either /usr/mports/chinese/wordpress-zh_CN/distinfo is out of date, or
=> wordpress-7.0.3-zh_CN.tar.gz is spelled incorrectly.
```

Reported by Magus run 647.

## Fix

`bmake makesum`.

No `PORTREVISION` bump ? the version comes from the master port, which
already carries it.

## Testing

`bmake makesum` followed by `bmake checksum`, which passes against the
freshly fetched 7.0.3 zh_CN tarball.

## Scope

All five `www/wordpress` slaves are affected by the same commit
(`chinese/wordpress-zh_CN`, `chinese/wordpress-zh_TW`,
`french/wordpress`, `german/wordpress`, `japanese/wordpress`). Each is
fixed on its own branch per the one-port-per-PR convention; this is one
of the five.

? Generated with [Claude Code](https://claude.com/claude-code)

## Summary by Sourcery

Bug Fixes:
- Update the Simplified Chinese WordPress distinfo checksums to support
fetching and verifying version 7.0.3.

Signed-off-by: Lucas Holt <luke at foolishgames.com>



To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications


From noreply at github.com  Fri Aug 21 23:57:34 2026
From: noreply at github.com (Lucas Holt)
Date: Fri, 21 Aug 2026 20:57:34 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports]
Message-ID: <MidnightBSD/mports/push/refs/heads/chinese/wordpress-zh_CN/e90d5f-000000@github.com>

  Branch: refs/heads/chinese/wordpress-zh_CN
  Home:   https://github.com/MidnightBSD/mports

To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications

From noreply at github.com  Fri Aug 21 23:57:58 2026
From: noreply at github.com (Lucas Holt)
Date: Fri, 21 Aug 2026 20:57:58 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports]
Message-ID: <MidnightBSD/mports/push/refs/heads/chinese/wordpress-zh_TW/f6d534-000000@github.com>

  Branch: refs/heads/chinese/wordpress-zh_TW
  Home:   https://github.com/MidnightBSD/mports

To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications

From noreply at github.com  Fri Aug 21 23:57:59 2026
From: noreply at github.com (Lucas Holt)
Date: Fri, 21 Aug 2026 20:57:59 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports] 3e7374:
 chinese/wordpress-zh_TW: regenerate distinfo for 7...
Message-ID: <MidnightBSD/mports/push/refs/heads/master/77192c-3e7374@github.com>

  Branch: refs/heads/master
  Home:   https://github.com/MidnightBSD/mports
  Commit: 3e73745d6b8cd35131ed81f141961ce460dfae65
      https://github.com/MidnightBSD/mports/commit/3e73745d6b8cd35131ed81f141961ce460dfae65
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-21 (Fri, 21 Aug 2026)

  Changed paths:
    M chinese/wordpress-zh_TW/distinfo

  Log Message:
  -----------
  chinese/wordpress-zh_TW: regenerate distinfo for 7.0.3 (#762)

## Problem

Commit `6314d32ee7` updated the master port `www/wordpress` from 6.9 to
7.0.3 and regenerated only `www/wordpress/distinfo`. This Traditional
Chinese slave still carried the 6.9 checksums, so fetching failed:

```
=> wordpress-7.0.3-zh_TW.tar.gz is not in /usr/mports/chinese/wordpress-zh_TW/distinfo.
=> Either /usr/mports/chinese/wordpress-zh_TW/distinfo is out of date, or
=> wordpress-7.0.3-zh_TW.tar.gz is spelled incorrectly.
```

Reported by Magus run 647.

## Fix

`bmake makesum`.

No `PORTREVISION` bump ? the version comes from the master port, which
already carries it.

## Testing

`bmake makesum` followed by `bmake checksum`, which passes against the
freshly fetched 7.0.3 zh_TW tarball.

## Scope

All five `www/wordpress` slaves are affected by the same commit
(`chinese/wordpress-zh_CN`, `chinese/wordpress-zh_TW`,
`french/wordpress`, `german/wordpress`, `japanese/wordpress`). Each is
fixed on its own branch per the one-port-per-PR convention; this is one
of the five.

? Generated with [Claude Code](https://claude.com/claude-code)

## Summary by Sourcery

Bug Fixes:
- Update the Traditional Chinese WordPress port checksums so version
7.0.3 can be fetched and verified successfully.

Signed-off-by: Lucas Holt <luke at foolishgames.com>



To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications


From noreply at github.com  Fri Aug 21 23:58:20 2026
From: noreply at github.com (Lucas Holt)
Date: Fri, 21 Aug 2026 20:58:20 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports] 78efea: french/wordpress:
 regenerate distinfo for 7.0.3 (#...
Message-ID: <MidnightBSD/mports/push/refs/heads/master/3e7374-78efea@github.com>

  Branch: refs/heads/master
  Home:   https://github.com/MidnightBSD/mports
  Commit: 78efea8e38c24cd4874f7e0cef2dc18631e3119a
      https://github.com/MidnightBSD/mports/commit/78efea8e38c24cd4874f7e0cef2dc18631e3119a
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-21 (Fri, 21 Aug 2026)

  Changed paths:
    M french/wordpress/distinfo

  Log Message:
  -----------
  french/wordpress: regenerate distinfo for 7.0.3 (#763)

## Problem

Commit `6314d32ee7` updated the master port `www/wordpress` from 6.9 to
7.0.3 and regenerated only `www/wordpress/distinfo`. This French slave
still carried the 6.9 checksums, so fetching failed:

```
=> wordpress-7.0.3-fr_FR.tar.gz is not in /usr/mports/french/wordpress/distinfo.
=> Either /usr/mports/french/wordpress/distinfo is out of date, or
=> wordpress-7.0.3-fr_FR.tar.gz is spelled incorrectly.
```

Reported by Magus run 647.

## Fix

`bmake makesum`.

No `PORTREVISION` bump ? the version comes from the master port, which
already carries it.

## Testing

`bmake makesum` followed by `bmake checksum`, which passes against the
freshly fetched 7.0.3 fr_FR tarball.

## Scope

All five `www/wordpress` slaves are affected by the same commit
(`chinese/wordpress-zh_CN`, `chinese/wordpress-zh_TW`,
`french/wordpress`, `german/wordpress`, `japanese/wordpress`). Each is
fixed on its own branch per the one-port-per-PR convention; this is one
of the five.

? Generated with [Claude Code](https://claude.com/claude-code)

## Summary by Sourcery

Bug Fixes:
- Regenerate the French WordPress 7.0.3 distribution checksums so the
localized slave can fetch and verify its source archive.

Signed-off-by: Lucas Holt <luke at foolishgames.com>



To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications


From noreply at github.com  Fri Aug 21 23:58:20 2026
From: noreply at github.com (Lucas Holt)
Date: Fri, 21 Aug 2026 20:58:20 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports]
Message-ID: <MidnightBSD/mports/push/refs/heads/french/wordpress/e49ff0-000000@github.com>

  Branch: refs/heads/french/wordpress
  Home:   https://github.com/MidnightBSD/mports

To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications

From noreply at github.com  Fri Aug 21 23:58:40 2026
From: noreply at github.com (Lucas Holt)
Date: Fri, 21 Aug 2026 20:58:40 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports] eaf943: german/wordpress:
 regenerate distinfo for 7.0.3 (#...
Message-ID: <MidnightBSD/mports/push/refs/heads/master/78efea-eaf943@github.com>

  Branch: refs/heads/master
  Home:   https://github.com/MidnightBSD/mports
  Commit: eaf943459e6bf1a2fa5c8db889d43d7592640a5d
      https://github.com/MidnightBSD/mports/commit/eaf943459e6bf1a2fa5c8db889d43d7592640a5d
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-21 (Fri, 21 Aug 2026)

  Changed paths:
    M german/wordpress/distinfo

  Log Message:
  -----------
  german/wordpress: regenerate distinfo for 7.0.3 (#764)

## Problem

Commit `6314d32ee7` updated the master port `www/wordpress` from 6.9 to
7.0.3 and regenerated only `www/wordpress/distinfo`. This German slave
still carried the 6.9 checksums, so fetching failed:

```
=> wordpress-7.0.3-de_DE.tar.gz is not in /usr/mports/german/wordpress/distinfo.
=> Either /usr/mports/german/wordpress/distinfo is out of date, or
=> wordpress-7.0.3-de_DE.tar.gz is spelled incorrectly.
```

Reported by Magus run 647.

## Fix

`bmake makesum`.

No `PORTREVISION` bump ? the version comes from the master port, which
already carries it.

## Testing

`bmake makesum` followed by `bmake checksum`, which passes against the
freshly fetched 7.0.3 de_DE tarball.

## Scope

All five `www/wordpress` slaves are affected by the same commit
(`chinese/wordpress-zh_CN`, `chinese/wordpress-zh_TW`,
`french/wordpress`, `german/wordpress`, `japanese/wordpress`). Each is
fixed on its own branch per the one-port-per-PR convention; this is one
of the five.

? Generated with [Claude Code](https://claude.com/claude-code)

## Summary by Sourcery

Bug Fixes:
- Update the German WordPress distribution checksums to support fetching
and verifying version 7.0.3.

Signed-off-by: Lucas Holt <luke at foolishgames.com>



To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications


From noreply at github.com  Fri Aug 21 23:58:41 2026
From: noreply at github.com (Lucas Holt)
Date: Fri, 21 Aug 2026 20:58:41 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports]
Message-ID: <MidnightBSD/mports/push/refs/heads/german/wordpress/057014-000000@github.com>

  Branch: refs/heads/german/wordpress
  Home:   https://github.com/MidnightBSD/mports

To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications

From noreply at github.com  Fri Aug 21 23:59:06 2026
From: noreply at github.com (Lucas Holt)
Date: Fri, 21 Aug 2026 20:59:06 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports]
Message-ID: <MidnightBSD/mports/push/refs/heads/japanese/wordpress/428aa0-000000@github.com>

  Branch: refs/heads/japanese/wordpress
  Home:   https://github.com/MidnightBSD/mports

To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications

From noreply at github.com  Fri Aug 21 23:59:04 2026
From: noreply at github.com (Lucas Holt)
Date: Fri, 21 Aug 2026 20:59:04 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports] 046bf5: japanese/wordpress:
 regenerate distinfo for 7.0.3 ...
Message-ID: <MidnightBSD/mports/push/refs/heads/master/eaf943-046bf5@github.com>

  Branch: refs/heads/master
  Home:   https://github.com/MidnightBSD/mports
  Commit: 046bf5b111c7821193368354484f4ca300a3fe51
      https://github.com/MidnightBSD/mports/commit/046bf5b111c7821193368354484f4ca300a3fe51
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-21 (Fri, 21 Aug 2026)

  Changed paths:
    M japanese/wordpress/distinfo

  Log Message:
  -----------
  japanese/wordpress: regenerate distinfo for 7.0.3 (#765)

## Problem

Commit `6314d32ee7` updated the master port `www/wordpress` from 6.9 to
7.0.3 and regenerated only `www/wordpress/distinfo`. This Japanese slave
still carried the 6.9 checksums, so fetching failed:

```
=> wordpress-7.0.3-ja.tar.gz is not in /usr/mports/japanese/wordpress/distinfo.
=> Either /usr/mports/japanese/wordpress/distinfo is out of date, or
=> wordpress-7.0.3-ja.tar.gz is spelled incorrectly.
```

Reported by Magus run 647.

## Fix

`bmake makesum`.

No `PORTREVISION` bump ? the version comes from the master port, which
already carries it.

## Testing

`bmake makesum` followed by `bmake checksum`, which passes against the
freshly fetched 7.0.3 ja tarball.

## Scope

All five `www/wordpress` slaves are affected by the same commit
(`chinese/wordpress-zh_CN`, `chinese/wordpress-zh_TW`,
`french/wordpress`, `german/wordpress`, `japanese/wordpress`). Each is
fixed on its own branch per the one-port-per-PR convention; this is one
of the five.

? Generated with [Claude Code](https://claude.com/claude-code)

## Summary by Sourcery

Bug Fixes:
- Update the Japanese WordPress port checksums for version 7.0.3 so the
distribution archive can be fetched and verified.

Signed-off-by: Lucas Holt <luke at foolishgames.com>



To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications


From noreply at github.com  Fri Aug 21 23:59:32 2026
From: noreply at github.com (Lucas Holt)
Date: Fri, 21 Aug 2026 20:59:32 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports]
Message-ID: <MidnightBSD/mports/push/refs/heads/cad/cura/f7a042-000000@github.com>

  Branch: refs/heads/cad/cura
  Home:   https://github.com/MidnightBSD/mports

To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications

From noreply at github.com  Fri Aug 21 23:59:32 2026
From: noreply at github.com (Lucas Holt)
Date: Fri, 21 Aug 2026 20:59:32 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports] a5cab7: cad/cura: fix
 invalid LICENSE, set to lgpl3 (#766)
Message-ID: <MidnightBSD/mports/push/refs/heads/master/046bf5-a5cab7@github.com>

  Branch: refs/heads/master
  Home:   https://github.com/MidnightBSD/mports
  Commit: a5cab7a1e854c3d119c696b510dbbfb240ba1182
      https://github.com/MidnightBSD/mports/commit/a5cab7a1e854c3d119c696b510dbbfb240ba1182
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-21 (Fri, 21 Aug 2026)

  Changed paths:
    M cad/cura/Makefile

  Log Message:
  -----------
  cad/cura: fix invalid LICENSE, set to lgpl3 (#766)

## Problem

`LICENSE= apgl` is not a recognized identifier. The framework treats an
unknown token as a *custom* license and aborts before doing any work:

```
===>  License not correctly defined: for unknown licenses, defining
      LICENSE_PERMS is mandatory (otherwise use a known LICENSE)
```

Magus run 647 recorded this as a **fetch failure**, because the license
check runs inside the fetch target ? but nothing was ever fetched. (The
trailing `make: exec(exit) failed (No such file or directory)` in that
log is framework noise masking the real error on the line above.)

## Why lgpl3 and not agpl

`apgl` reads like a transposition of `agpl`, and FreeBSD's `cad/cura`
does use `AGPLv3` ? but that is **not what this port builds**.

Cura relicensed at 4.0. This port is `DISTVERSION= 4.13.1`, `GH_ACCOUNT=
Ultimaker`, and the `LICENSE` file at that tag ? the very file
`LICENSE_FILE= ${WRKSRC}/LICENSE` points at ? is:

```console
$ fetch -qo - https://raw.githubusercontent.com/Ultimaker/Cura/4.13.1/LICENSE | head -2
    GNU LESSER GENERAL PUBLIC LICENSE
                       Version 3, 29 June 2007
```

So `lgpl3` is the value that matches the shipped source. Taking the
obvious typo fix would have recorded a license the source does not
carry, and left `LICENSE_FILE` contradicting `LICENSE`.

FreeBSD's port looks stale on this point.

## Fix

`LICENSE= apgl` ? `LICENSE= lgpl3`

No `PORTREVISION` bump ? the port could not build at all, so there is no
package whose metadata changes.

## Testing

`bmake check-license` ? `===>  License lgpl3 accepted by the user`

? Generated with [Claude Code](https://claude.com/claude-code)

## Summary by Sourcery

Bug Fixes:
- Correct the Cura port license identifier from the invalid `apgl` value
to `lgpl3`, allowing license checks and builds to proceed.

Signed-off-by: Lucas Holt <luke at foolishgames.com>



To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications


From noreply at github.com  Sat Aug 22 00:00:10 2026
From: noreply at github.com (Lucas Holt)
Date: Fri, 21 Aug 2026 21:00:10 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports]
Message-ID: <MidnightBSD/mports/push/refs/heads/japanese/p5-WWW-MobileCarrierJP/70a054-000000@github.com>

  Branch: refs/heads/japanese/p5-WWW-MobileCarrierJP
  Home:   https://github.com/MidnightBSD/mports

To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications

From noreply at github.com  Sat Aug 22 00:00:10 2026
From: noreply at github.com (Lucas Holt)
Date: Fri, 21 Aug 2026 21:00:10 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports] 591765:
 japanese/p5-WWW-MobileCarrierJP: fix misspelled LI...
Message-ID: <MidnightBSD/mports/push/refs/heads/master/a5cab7-591765@github.com>

  Branch: refs/heads/master
  Home:   https://github.com/MidnightBSD/mports
  Commit: 591765ae973af8314a8847f0432104e2f1f46e61
      https://github.com/MidnightBSD/mports/commit/591765ae973af8314a8847f0432104e2f1f46e61
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-21 (Fri, 21 Aug 2026)

  Changed paths:
    M japanese/p5-WWW-MobileCarrierJP/Makefile

  Log Message:
  -----------
  japanese/p5-WWW-MobileCarrierJP: fix misspelled LICENSE identifier (#767)

## Problem

`LICENSE= artitistic gpl+` ? `artitistic` is not a recognized
identifier, so the framework treated it as a *custom* license and
aborted before doing any work:

```
===>  License not correctly defined: for unknown licenses, defining
      LICENSE_PERMS_artitistic is mandatory (otherwise use a known LICENSE)
```

Magus run 647 recorded this as a **fetch failure**, because the license
check runs inside the fetch target ? but nothing was ever fetched. (The
trailing `make: exec(exit) failed (No such file or directory)` in that
log is framework noise masking the real error on the line above.)

## Fix

`artitistic` ? `artistic`.

With `LICENSE_COMB=dual` this is the standard Perl dual license, spelled
`artistic gpl+` by every other `p5-*` port in the tree
(`archivers/p5-Archive-Extract`, `archivers/p5-IO-Compress`, and ~20
more).

No `PORTREVISION` bump ? the port could not build at all, so there is no
package whose metadata changes.

## Testing

`bmake check-license` ? `===> License artistic gpl+ accepted by the
user`

? Generated with [Claude Code](https://claude.com/claude-code)

## Summary by Sourcery

Bug Fixes:
- Correct the misspelled Artistic license identifier so the port passes
license validation and can build.

Signed-off-by: Lucas Holt <luke at foolishgames.com>



To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications


From noreply at github.com  Sat Aug 22 00:00:45 2026
From: noreply at github.com (Lucas Holt)
Date: Fri, 21 Aug 2026 21:00:45 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports] ff7236:
 textproc/gitlab-code-parser: fix LICENSE case (#768)
Message-ID: <MidnightBSD/mports/push/refs/heads/master/591765-ff7236@github.com>

  Branch: refs/heads/master
  Home:   https://github.com/MidnightBSD/mports
  Commit: ff7236f2d03262651ba49a40d3a03e5e6b1a2f39
      https://github.com/MidnightBSD/mports/commit/ff7236f2d03262651ba49a40d3a03e5e6b1a2f39
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-22 (Sat, 22 Aug 2026)

  Changed paths:
    M textproc/gitlab-code-parser/Makefile

  Log Message:
  -----------
  textproc/gitlab-code-parser: fix LICENSE case (#768)

## Problem

`LICENSE= MIT`. Identifiers are case-sensitive and the recognized one is
`mit`, so the framework treats `MIT` as a *custom* license and aborts
before doing any work:

```
===>  License not correctly defined: for unknown licenses, defining
      LICENSE_PERMS is mandatory (otherwise use a known LICENSE)
```

Reproduced locally with `bmake check-license`.

## Fix

`MIT` ? `mit`.

The license itself is correct ? `LICENSE.md` upstream at
https://gitlab.com/gitlab-org/rust/gitlab-code-parser is the MIT license
text. Only the identifier spelling was wrong.

No `PORTREVISION` bump ? the port could not build at all, so there is no
package whose metadata changes.

## Context

Magus run 647 has not reached `textproc` yet, but this port will fail
the same way `cad/cura` (#766) and `japanese/p5-WWW-MobileCarrierJP`
(#767) already did ? reported as a *fetch* failure, because the license
check runs inside the fetch target.

Found by sweeping the tree for `LICENSE` tokens that are neither in
`bmake license-list` nor backed by a `LICENSE_PERMS` definition. That
sweep returned exactly four ports; these two and the two above are all
of them.

## Testing

`bmake check-license` ? `===>  License mit accepted by the user`

? Generated with [Claude Code](https://claude.com/claude-code)

## Summary by Sourcery

Bug Fixes:
- Correct the GitLab code parser port's license identifier so license
validation recognizes the existing MIT license.

Signed-off-by: Lucas Holt <luke at foolishgames.com>



To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications


From noreply at github.com  Sat Aug 22 00:00:47 2026
From: noreply at github.com (Lucas Holt)
Date: Fri, 21 Aug 2026 21:00:47 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports]
Message-ID: <MidnightBSD/mports/push/refs/heads/textproc/gitlab-code-parser/541122-000000@github.com>

  Branch: refs/heads/textproc/gitlab-code-parser
  Home:   https://github.com/MidnightBSD/mports

To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications

From noreply at github.com  Sat Aug 22 00:01:15 2026
From: noreply at github.com (Lucas Holt)
Date: Fri, 21 Aug 2026 21:01:15 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports]
Message-ID: <MidnightBSD/mports/push/refs/heads/textproc/gitlab-elasticsearch-indexer/be0c94-000000@github.com>

  Branch: refs/heads/textproc/gitlab-elasticsearch-indexer
  Home:   https://github.com/MidnightBSD/mports

To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications

From noreply at github.com  Sat Aug 22 00:01:16 2026
From: noreply at github.com (Lucas Holt)
Date: Fri, 21 Aug 2026 21:01:16 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports] a4d23f:
 textproc/gitlab-elasticsearch-indexer: fix LICENSE...
Message-ID: <MidnightBSD/mports/push/refs/heads/master/ff7236-a4d23f@github.com>

  Branch: refs/heads/master
  Home:   https://github.com/MidnightBSD/mports
  Commit: a4d23f996375ad90791eb996cd970092fe924aa8
      https://github.com/MidnightBSD/mports/commit/a4d23f996375ad90791eb996cd970092fe924aa8
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-22 (Sat, 22 Aug 2026)

  Changed paths:
    M textproc/gitlab-elasticsearch-indexer/Makefile

  Log Message:
  -----------
  textproc/gitlab-elasticsearch-indexer: fix LICENSE case (#769)

## Problem

`LICENSE= MIT`. Identifiers are case-sensitive and the recognized one is
`mit`, so the framework treats `MIT` as a *custom* license and aborts
before doing any work:

```
===>  License not correctly defined: for unknown licenses, defining
      LICENSE_PERMS is mandatory (otherwise use a known LICENSE)
```

Reproduced locally with `bmake check-license`.

## Fix

`MIT` ? `mit`.

The license itself is correct ? `LICENSE.md` upstream at
https://gitlab.com/gitlab-org/gitlab-elasticsearch-indexer is the MIT
license text. Only the identifier spelling was wrong.

No `PORTREVISION` bump ? the port could not build at all, so there is no
package whose metadata changes.

## Context

Magus run 647 has not reached `textproc` yet, but this port will fail
the same way `cad/cura` (#766) and `japanese/p5-WWW-MobileCarrierJP`
(#767) already did ? reported as a *fetch* failure, because the license
check runs inside the fetch target.

Found by sweeping the tree for `LICENSE` tokens that are neither in
`bmake license-list` nor backed by a `LICENSE_PERMS` definition. That
sweep returned exactly four ports; these two and the two above are all
of them.

## Testing

`bmake check-license` ? `===>  License mit accepted by the user`

? Generated with [Claude Code](https://claude.com/claude-code)

## Summary by Sourcery

Bug Fixes:
- Correct the GitLab Elasticsearch indexer port?s license identifier so
license validation recognizes the existing MIT license.

Signed-off-by: Lucas Holt <luke at foolishgames.com>



To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications


From noreply at github.com  Sat Aug 22 00:39:06 2026
From: noreply at github.com (Lucas Holt)
Date: Fri, 21 Aug 2026 21:39:06 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports] 2d1b47: archivers/zipmix:
 update to 20160430
Message-ID: <MidnightBSD/mports/push/refs/heads/archivers/zipmix/000000-2d1b47@github.com>

  Branch: refs/heads/archivers/zipmix
  Home:   https://github.com/MidnightBSD/mports
  Commit: 2d1b472fd0bd609f4ea401a7a04a67484d0db4f6
      https://github.com/MidnightBSD/mports/commit/2d1b472fd0bd609f4ea401a7a04a67484d0db4f6
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-22 (Sat, 22 Aug 2026)

  Changed paths:
    M archivers/zipmix/Makefile
    M archivers/zipmix/distinfo

  Log Message:
  -----------
  archivers/zipmix: update to 20160430

The port pinned the 2007-02-21 distfile, but upstream (advsys.net) replaced
zipmix_src.zip in place with a 2016-04-30 build years ago. Since the two
MASTER_SITES now serve different content under the same name, only the
BSDforge mirror still matched distinfo; the primary site failed its size
check on every fetch:

    fetch: http://www.advsys.net/ken/util/zipmix_src.zip: size mismatch:
           expected 4759, actual 5232

That left the port depending on a single third-party mirror for a file
upstream no longer publishes. Follow upstream instead.

Upstream changes between the two builds:
- guard the byte-swap helpers with __POWERPC__ rather than __APPLE__
- add getday() and isdatetimewithin1sec(), so files whose ZIP timestamps
  differ by a second still match

files/patch-zipmix.c still applies (hunk 2 shifts by 22 lines).

Verified: makesum, patch, build, fake, package, and portlint ("looks fine").

Note for whoever touches this next: the resulting binary hangs when asked to
mix two archives. That is NOT introduced here -- builds of the 2007 and 2016
sources hang identically, so it is pre-existing and orthogonal to this
update.

AI-Assisted-by: Claude Opus 5 <noreply at anthropic.com>
Signed-off-by: Lucas Holt <luke at foolishgames.com>



To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications

From noreply at github.com  Sat Aug 22 10:17:21 2026
From: noreply at github.com (Lucas Holt)
Date: Sat, 22 Aug 2026 07:17:21 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports] 1901cd: x11/xpr,
 x11/xload: update X.Org applications
Message-ID: <MidnightBSD/mports/push/refs/heads/master/a4d23f-01ce5d@github.com>

  Branch: refs/heads/master
  Home:   https://github.com/MidnightBSD/mports
  Commit: 1901cd1b23c67a5c22267e387ed1b4b7309b0948
      https://github.com/MidnightBSD/mports/commit/1901cd1b23c67a5c22267e387ed1b4b7309b0948
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-18 (Tue, 18 Aug 2026)

  Changed paths:
    M x11/xload/Makefile
    M x11/xload/distinfo
    M x11/xpr/Makefile
    M x11/xpr/distinfo

  Log Message:
  -----------
  x11/xpr, x11/xload: update X.Org applications

Update xpr to 1.2.1 and xload to 1.2.2, and use their Meson build support.

AI-Assisted-by: OpenAI Codex <noreply at openai.com>
Signed-off-by: Lucas Holt <luke at foolishgames.com>


  Commit: 26726e98b19fb8da951b90512041e8f85ffc05d3
      https://github.com/MidnightBSD/mports/commit/26726e98b19fb8da951b90512041e8f85ffc05d3
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-22 (Sat, 22 Aug 2026)

  Changed paths:
    M ai/claude-code/Makefile
    M ai/claude-code/distinfo

  Log Message:
  -----------
  ai/claude-code: 2.1.239


  Commit: 01ce5d01bc72e899d7d316b32b230da14bf574bc
      https://github.com/MidnightBSD/mports/commit/01ce5d01bc72e899d7d316b32b230da14bf574bc
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-22 (Sat, 22 Aug 2026)

  Changed paths:
    M Mk/scripts/do-fetch.sh
    M ai/codex/distinfo
    M cad/cura/Makefile
    M chinese/wordpress-zh_CN/distinfo
    M chinese/wordpress-zh_TW/distinfo
    M editors/emacs/Makefile
    A editors/emacs/files/patch-src_image.c
    M french/wordpress/distinfo
    M german/wordpress/distinfo
    M japanese/p5-WWW-MobileCarrierJP/Makefile
    M japanese/wordpress/distinfo
    M java/openjdk23/Makefile
    M java/openjdk23/distinfo
    M lang/rust-bootstrap/Makefile
    M textproc/gitlab-code-parser/Makefile
    M textproc/gitlab-elasticsearch-indexer/Makefile

  Log Message:
  -----------
  Merge branch 'master' of github.com:MidnightBSD/mports


Compare: https://github.com/MidnightBSD/mports/compare/a4d23f996375...01ce5d01bc72

To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications

From noreply at github.com  Sun Aug 23 01:21:50 2026
From: noreply at github.com (Lucas Holt)
Date: Sat, 22 Aug 2026 22:21:50 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports] 9726d9: ai/claude-code:
 2.1.241
Message-ID: <MidnightBSD/mports/push/refs/heads/master/01ce5d-9726d9@github.com>

  Branch: refs/heads/master
  Home:   https://github.com/MidnightBSD/mports
  Commit: 9726d9f6fd168b7833d14f9172614b1c640252c1
      https://github.com/MidnightBSD/mports/commit/9726d9f6fd168b7833d14f9172614b1c640252c1
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-23 (Sun, 23 Aug 2026)

  Changed paths:
    M ai/claude-code/Makefile
    M ai/claude-code/distinfo

  Log Message:
  -----------
  ai/claude-code: 2.1.241



To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications

From noreply at github.com  Sun Aug 23 01:44:16 2026
From: noreply at github.com (Lucas Holt)
Date: Sat, 22 Aug 2026 22:44:16 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports] c58f50: benchmarks: add
 p5-Benchmark-Forking and p5-Benchm...
Message-ID: <MidnightBSD/mports/push/refs/heads/benchmarks/p5-benchmark-modules/000000-c58f50@github.com>

  Branch: refs/heads/benchmarks/p5-benchmark-modules
  Home:   https://github.com/MidnightBSD/mports
  Commit: c58f50c87d9b6155fb0109f85c63eeb3c8d9501c
      https://github.com/MidnightBSD/mports/commit/c58f50c87d9b6155fb0109f85c63eeb3c8d9501c
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-23 (Sun, 23 Aug 2026)

  Changed paths:
    M benchmarks/Makefile
    A benchmarks/p5-Benchmark-Forking/Makefile
    A benchmarks/p5-Benchmark-Forking/distinfo
    A benchmarks/p5-Benchmark-Forking/pkg-descr
    A benchmarks/p5-Benchmark-Forking/pkg-plist
    A benchmarks/p5-Benchmark-Stopwatch/Makefile
    A benchmarks/p5-Benchmark-Stopwatch/distinfo
    A benchmarks/p5-Benchmark-Stopwatch/pkg-descr
    A benchmarks/p5-Benchmark-Stopwatch/pkg-plist

  Log Message:
  -----------
  benchmarks: add p5-Benchmark-Forking and p5-Benchmark-Stopwatch

Import two related Perl benchmarking modules from FreeBSD ports and
adapt them for mports:

- Benchmark::Forking runs each timed code block in a forked child so
  side-effects cannot leak between test cases.
- Benchmark::Stopwatch times the progression of a script via lap marks
  and prints a summary with elapsed and percentage-of-total times.

The imported Makefiles used FreeBSD license identifiers (ART10,
GPLv1+), which mports does not recognize; check-license aborted the
build before fetch. Use the mports spelling (artistic, gpl+) instead.
Point MAINTAINER at ports at MidnightBSD.org, drop the PORTREVISION values
carried over from FreeBSD, and register both directories in the
benchmarks category Makefile.

Both ports build, stage, and package cleanly, and their upstream test
suites pass. All dependencies (devel/p5-Clone, devel/p5-Test-LongString)
already exist in the tree.

AI-Assisted-by: Claude Opus 5 (1M context) <noreply at anthropic.com>
Signed-off-by: Lucas Holt <luke at foolishgames.com>



To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications

From noreply at github.com  Sun Aug 23 01:47:54 2026
From: noreply at github.com (Lucas Holt)
Date: Sat, 22 Aug 2026 22:47:54 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports] 0ea991: archivers/zipmix:
 update to 20160430 (#770)
Message-ID: <MidnightBSD/mports/push/refs/heads/master/9726d9-0ea991@github.com>

  Branch: refs/heads/master
  Home:   https://github.com/MidnightBSD/mports
  Commit: 0ea9911e92561cdb8c81e3914cb440bcf558e926
      https://github.com/MidnightBSD/mports/commit/0ea9911e92561cdb8c81e3914cb440bcf558e926
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-23 (Sun, 23 Aug 2026)

  Changed paths:
    M archivers/zipmix/Makefile
    M archivers/zipmix/distinfo

  Log Message:
  -----------
  archivers/zipmix: update to 20160430 (#770)

## Problem

The port pinned the **2007-02-21** distfile, but upstream (`advsys.net`)
replaced `zipmix_src.zip` **in place** with a **2016-04-30** build years
ago. The two `MASTER_SITES` now serve different content under the same
filename:

| Site | Size | Date | Matches distinfo |
|---|---|---|---|
| `advsys.net` (primary) | 5232 | 2016-04-30 | ? |
| `BSDforge.com` (mirror) | 4759 | 2007-02-21 | ? |

So every fetch failed its size check on the primary and fell through to
the mirror:

```
fetch: http://www.advsys.net/ken/util/zipmix_src.zip: size mismatch: expected 4759, actual 5232
```

That left the port depending on a single third-party mirror for a file
upstream no longer publishes ? one mirror outage away from being
unfetchable. Following upstream is the durable fix.

## Upstream changes between the two builds

- Guard the byte-swap helpers with `__POWERPC__` rather than `__APPLE__`
? the old guard misfired on non-PPC Apple hardware
- Add `getday()` and `isdatetimewithin1sec()`, so files whose ZIP
timestamps differ by a second still match

`files/patch-zipmix.c` still applies (hunk 2 shifts by 22 lines).

## Testing

`bmake makesum`, `patch`, `build`, `fake`, `package` all succeed;
`portlint` reports *"looks fine"*. The built binary runs and prints its
usage banner.

## Pre-existing bug found while testing ? not introduced here

The resulting binary **hangs** when actually asked to mix two archives
(`zipmix A.zip B.zip out.zip`), producing a 0-byte output and never
terminating. It is not an interactive prompt ? it hangs the same way
with input fed to stdin.

I built the patched **2007** and **2016** sources side by side and they
hang **identically**, so this is pre-existing and orthogonal to this
update. Flagging it rather than folding it in; it deserves its own
investigation.

? Generated with [Claude Code](https://claude.com/claude-code)

## Summary by Sourcery

Bug Fixes:
- Update the zipmix port to the upstream 2016-04-30 distfile so fetches
use the current primary source instead of relying on the outdated
mirror.

Signed-off-by: Lucas Holt <luke at foolishgames.com>



To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications


From noreply at github.com  Sun Aug 23 01:47:56 2026
From: noreply at github.com (Lucas Holt)
Date: Sat, 22 Aug 2026 22:47:56 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports]
Message-ID: <MidnightBSD/mports/push/refs/heads/archivers/zipmix/2d1b47-000000@github.com>

  Branch: refs/heads/archivers/zipmix
  Home:   https://github.com/MidnightBSD/mports

To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications

From noreply at github.com  Sun Aug 23 01:48:38 2026
From: noreply at github.com (Lucas Holt)
Date: Sat, 22 Aug 2026 22:48:38 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports]
Message-ID: <MidnightBSD/mports/push/refs/heads/benchmarks/p5-benchmark-modules/c58f50-000000@github.com>

  Branch: refs/heads/benchmarks/p5-benchmark-modules
  Home:   https://github.com/MidnightBSD/mports

To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications

From noreply at github.com  Sun Aug 23 01:48:38 2026
From: noreply at github.com (Lucas Holt)
Date: Sat, 22 Aug 2026 22:48:38 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports] e93ab7: benchmarks: add
 p5-Benchmark-Forking and p5-Benchm...
Message-ID: <MidnightBSD/mports/push/refs/heads/master/0ea991-e93ab7@github.com>

  Branch: refs/heads/master
  Home:   https://github.com/MidnightBSD/mports
  Commit: e93ab760a64499c05541a7a253aec9415617d35f
      https://github.com/MidnightBSD/mports/commit/e93ab760a64499c05541a7a253aec9415617d35f
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-23 (Sun, 23 Aug 2026)

  Changed paths:
    M benchmarks/Makefile
    A benchmarks/p5-Benchmark-Forking/Makefile
    A benchmarks/p5-Benchmark-Forking/distinfo
    A benchmarks/p5-Benchmark-Forking/pkg-descr
    A benchmarks/p5-Benchmark-Forking/pkg-plist
    A benchmarks/p5-Benchmark-Stopwatch/Makefile
    A benchmarks/p5-Benchmark-Stopwatch/distinfo
    A benchmarks/p5-Benchmark-Stopwatch/pkg-descr
    A benchmarks/p5-Benchmark-Stopwatch/pkg-plist

  Log Message:
  -----------
  benchmarks: add p5-Benchmark-Forking and p5-Benchmark-Stopwatch (#771)

Adds two related Perl benchmarking modules, imported from FreeBSD ports
and adapted for mports.

- **benchmarks/p5-Benchmark-Forking** 1.01 ? runs each timed code block
in a forked child so side-effects cannot leak between test cases.
- **benchmarks/p5-Benchmark-Stopwatch** 0.05 ? times the progression of
a script via lap marks, printing elapsed time and percentage of total
per event.

Kept as one PR since the two are companion benchmarking modules and
share the same `benchmarks/Makefile` SUBDIR change.

## mports adaptation

The imported Makefiles carried FreeBSD-only license identifiers (`ART10
GPLv1+`), which mports does not recognize ? `check-license` aborted
before fetch:

```
===>  License not correctly defined: for unknown licenses, defining LICENSE_PERMS_ART10 is mandatory
```

Changes applied to both ports:

- `LICENSE= ART10 GPLv1+` ? `artistic gpl+`
- `MAINTAINER` ? `ports at MidnightBSD.org`
- Dropped the `PORTREVISION` values carried over from FreeBSD
- `bmake makesum` added the missing `TIMESTAMP` lines (SHA256/SIZE were
already correct)
- Registered both directories in `benchmarks/Makefile`

## Dependencies

No new ports required. `devel/p5-Clone` and `devel/p5-Test-LongString`
(Stopwatch) already exist and are registered; Forking has no non-core
prerequisites.

## Validation

|  | Forking | Stopwatch |
|---|---|---|
| `bmake check-license` | pass | pass |
| `bmake build` / `fake` / `package` | pass | pass |
| `pkg-plist` vs staged tree | matches | matches |
| `bmake test` | 4/4 pass | 22/22 pass |
| `portlint` | looks fine | looks fine |

Built on amd64. The tree's `lang/perl5.40` is at 5.40.5 while the build
host had perl5-5.40.4 installed, so builds were run with `NO_DEPENDS=yes
PERL_VERSION=5.40.4` to avoid an unrelated perl upgrade. Both ports are
pure perl with no arch-specific components, but they have not been built
against 5.40.5.

? Generated with [Claude Code](https://claude.com/claude-code)

## Summary by Sourcery

Add two CPAN benchmarking modules for isolated forked benchmarks and
stopwatch-style timing to the benchmarks collection.

New Features:
- Add the Benchmark-Forking and Benchmark-Stopwatch Perl benchmarking
ports.
- Register both benchmarking modules in the benchmarks port collection.

Enhancements:
- Adapt both ports to mports licensing and packaging conventions.

Signed-off-by: Lucas Holt <luke at foolishgames.com>



To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications


From noreply at github.com  Sun Aug 23 01:58:09 2026
From: noreply at github.com (Lucas Holt)
Date: Sat, 22 Aug 2026 22:58:09 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports] 1feb63: magus: add optional
 ZFS chroot rollback
Message-ID: <MidnightBSD/mports/push/refs/heads/codex/magus-zfs-chroots/000000-1feb63@github.com>

  Branch: refs/heads/codex/magus-zfs-chroots
  Home:   https://github.com/MidnightBSD/mports
  Commit: 1feb63f71d2769fb3256bc2c02f84d7d306c98de
      https://github.com/MidnightBSD/mports/commit/1feb63f71d2769fb3256bc2c02f84d7d306c98de
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-23 (Sun, 23 Aug 2026)

  Changed paths:
    M Tools/lib/Magus/App/Slave/Worker.pm
    M Tools/lib/Magus/Chroot.pm
    M Tools/lib/Magus/Config.pm
    M Tools/lib/Magus/PortTest.pm
    M Tools/magus/config.yaml.example
    M Tools/magus/slave/magus.pl
    A Tools/t/magus-chroot-zfs.t

  Log Message:
  -----------
  magus: add optional ZFS chroot rollback

Add guarded per-worker ZFS datasets with fresh startup snapshots while preserving cpdup as the default backend. Mark workers dirty before dependency injection and cover the ZFS lifecycle with mocked tests.

AI-Assisted-by: OpenAI Codex <noreply at openai.com>
Signed-off-by: Lucas Holt <luke at foolishgames.com>



To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications

From noreply at github.com  Sun Aug 23 02:04:53 2026
From: noreply at github.com (Lucas Holt)
Date: Sat, 22 Aug 2026 23:04:53 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports] d05cfa: magus: include ctime
 in bootstrap cache key
Message-ID: <MidnightBSD/mports/push/refs/heads/codex/magus-zfs-chroots/1feb63-d05cfa@github.com>

  Branch: refs/heads/codex/magus-zfs-chroots
  Home:   https://github.com/MidnightBSD/mports
  Commit: d05cfa3f0e5cd1b16224006b6d6c32c9fb0bf0b6
      https://github.com/MidnightBSD/mports/commit/d05cfa3f0e5cd1b16224006b6d6c32c9fb0bf0b6
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-23 (Sun, 23 Aug 2026)

  Changed paths:
    M Tools/lib/Magus/Chroot.pm
    M Tools/t/magus-chroot-zfs.t

  Log Message:
  -----------
  magus: include ctime in bootstrap cache key

Invalidate cached bootstrap checksums for same-sized replacements that restore the original mtime, and cover the case with a regression test.

AI-Assisted-by: OpenAI Codex <noreply at openai.com>
Signed-off-by: Lucas Holt <luke at foolishgames.com>



To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications

From noreply at github.com  Sun Aug 23 09:57:42 2026
From: noreply at github.com (Lucas Holt)
Date: Sun, 23 Aug 2026 06:57:42 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports] 0ddfa0: magus: add optional
 ZFS chroot rollback (#772)
Message-ID: <MidnightBSD/mports/push/refs/heads/master/e93ab7-0ddfa0@github.com>

  Branch: refs/heads/master
  Home:   https://github.com/MidnightBSD/mports
  Commit: 0ddfa0aa839d1c05770d7dfb4c645f19cd92ab18
      https://github.com/MidnightBSD/mports/commit/0ddfa0aa839d1c05770d7dfb4c645f19cd92ab18
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-23 (Sun, 23 Aug 2026)

  Changed paths:
    M Tools/lib/Magus/App/Slave/Worker.pm
    M Tools/lib/Magus/Chroot.pm
    M Tools/lib/Magus/Config.pm
    M Tools/lib/Magus/PortTest.pm
    M Tools/magus/config.yaml.example
    M Tools/magus/slave/magus.pl
    A Tools/t/magus-chroot-zfs.t

  Log Message:
  -----------
  magus: add optional ZFS chroot rollback (#772)

## Summary
- add an opt-in per-worker ZFS chroot backend while retaining cpdup by
default
- rebuild each ZFS worker from the verified bootstrap and create a fresh
magus-clean snapshot at daemon startup
- roll dirty workers back between builds and mark them dirty before
dependency or distfile injection
- validate dataset names, mountpoints, local ownership, and unexpected
children before destructive ZFS operations

## Testing
- prove -I Tools/lib Tools/t/magus-chroot-zfs.t
- perl syntax checks for Magus::Chroot, Magus::PortTest, and the new
test
- parsed Tools/magus/config.yaml.example with YAML::LoadFile
- git diff --check

End-to-end ZFS, mount, chroot, and Magus build testing was not run
because the development host is not MidnightBSD.

## Summary by Sourcery

Add an optional ZFS chroot backend that provides verified per-worker
rollback baselines and safer lifecycle management while retaining
existing cpdup behavior by default.

New Features:
- Add an opt-in ZFS-backed chroot mode for Magus workers while
preserving cpdup as the default backend.
- Refresh per-worker ZFS chroot baselines from the verified bootstrap at
daemon startup and roll workers back to clean snapshots between builds.

Bug Fixes:
- Ensure chroots are marked dirty before dependency and distfile
injection so interrupted builds cannot leave modified state marked
clean.
- Detect replaced bootstrap tarballs reliably and refresh stale chroot
references when the bootstrap changes.

Enhancements:
- Validate ZFS dataset ownership, mountpoints, configuration, and child
contents before destructive operations.
- Standardize external command execution and strengthen mount cleanup
checks for chroot management.

Documentation:
- Document the new optional ZFS chroot configuration settings in the
example Magus configuration.

Tests:
- Add coverage for bootstrap checksum invalidation, ZFS configuration
validation, ownership and child safety checks, baseline refresh, and
rollback behavior.

---------

Signed-off-by: Lucas Holt <luke at foolishgames.com>



To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications

From noreply at github.com  Sun Aug 23 09:57:43 2026
From: noreply at github.com (Lucas Holt)
Date: Sun, 23 Aug 2026 06:57:43 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports]
Message-ID: <MidnightBSD/mports/push/refs/heads/codex/magus-zfs-chroots/d05cfa-000000@github.com>

  Branch: refs/heads/codex/magus-zfs-chroots
  Home:   https://github.com/MidnightBSD/mports

To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications

From noreply at github.com  Sun Aug 23 20:00:44 2026
From: noreply at github.com (Lucas Holt)
Date: Sun, 23 Aug 2026 17:00:44 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports] f0ed1a: databases/p5-DBD-Pg:
 update to 3.21.1
Message-ID: <MidnightBSD/mports/push/refs/heads/databases/p5-DBD-Pg/000000-f0ed1a@github.com>

  Branch: refs/heads/databases/p5-DBD-Pg
  Home:   https://github.com/MidnightBSD/mports
  Commit: f0ed1aa565249cdace89ee35ce5fc1006749c29e
      https://github.com/MidnightBSD/mports/commit/f0ed1aa565249cdace89ee35ce5fc1006749c29e
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-23 (Sun, 23 Aug 2026)

  Changed paths:
    M databases/p5-DBD-Pg/Makefile
    M databases/p5-DBD-Pg/distinfo
    M databases/p5-DBD-Pg/files/patch-Makefile.PL

  Log Message:
  -----------
  databases/p5-DBD-Pg: update to 3.21.1

Upstream fixes an incorrect allocation in quote_float() along with
related fixes in _dequote_bytea_hex() and pg_destringify_array().

The quote_float() bug is CVE-2026-78183: for the special literals NaN,
Inf, +Inf, -Inf, Infinity, +Infinity and -Infinity the function emitted
the literal wrapped in single quotes plus NUL (length + 3 bytes) while
allocating only length + 1, overflowing the heap buffer by two bytes.
It was reachable through $dbh->quote($val, DBI::SQL_NUMERIC).

The regression was introduced in 3.21.0 by the quote.c rewrite, so
mports was never vulnerable at 3.20.2.  This update is for currency and
to stay ahead of the fixed branch.

Regenerate patch-Makefile.PL: upstream reworked the INC line to use
qq{} quoting, so the first hunk no longer applied.  The patch keeps its
original purpose of appending $comp_opts to INC rather than letting
CCFLAGS clobber perl's own compiler flags.

Set NO_TEST: 3.21.1 calls plan() twice in t/03dbmethod.t,
t/10_pg_error_field.t and t/30unicode.t when no PostgreSQL server is
reachable, so those files die rather than skipping.

AI-Assisted-by: Claude Opus 5 (1M context) <noreply at anthropic.com>
Signed-off-by: Lucas Holt <luke at foolishgames.com>



To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications

From noreply at github.com  Sun Aug 23 20:10:45 2026
From: noreply at github.com (Lucas Holt)
Date: Sun, 23 Aug 2026 17:10:45 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports] e6ce42: databases/p5-DBD-Pg:
 update to 3.21.1 (#773)
Message-ID: <MidnightBSD/mports/push/refs/heads/master/0ddfa0-e6ce42@github.com>

  Branch: refs/heads/master
  Home:   https://github.com/MidnightBSD/mports
  Commit: e6ce42373a2bd51ec6095849b9c938ca36f2f687
      https://github.com/MidnightBSD/mports/commit/e6ce42373a2bd51ec6095849b9c938ca36f2f687
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-23 (Sun, 23 Aug 2026)

  Changed paths:
    M databases/p5-DBD-Pg/Makefile
    M databases/p5-DBD-Pg/distinfo
    M databases/p5-DBD-Pg/files/patch-Makefile.PL

  Log Message:
  -----------
  databases/p5-DBD-Pg: update to 3.21.1 (#773)

Updates `databases/p5-DBD-Pg` from 3.20.2 to 3.21.1.

## Security context

Upstream 3.21.1 fixes **CVE-2026-78183**, a heap out-of-bounds write in
`quote_float()`. For the special literals `NaN`, `Inf`, `+Inf`, `-Inf`,
`Infinity`, `+Infinity` and `-Infinity`, the function emitted the
literal wrapped in single quotes plus NUL (`length + 3` bytes) while
allocating only `length + 1` ? a two-byte overflow, reachable via
`$dbh->quote($val, DBI::SQL_NUMERIC)`.

**mports was never vulnerable.** The bug was introduced in 3.21.0 by the
quote.c rewrite, and the tree was on 3.20.2, which predates it. No other
port bundles a copy ? all consumers (`comms/atslog`, `databases/pgtop`,
`databases/postgresqltuner`, `mail/sympa`, `mail/spamassassin`,
`ports-mgmt/portscout`) depend on this port. This update is for
currency, not remediation.

Fix verified present in the new distfile ? the special-literal branch
now allocates `length + 3` and returns early (`quote.c:374`). 3.21.1
also fixes allocations in `_dequote_bytea_hex()` and
`pg_destringify_array()`.

## Port changes

- **`patch-Makefile.PL` regenerated.** Upstream reworked the `INC` line
to use `qq{}` quoting, so the first hunk no longer applied (`1 out of 2
hunks failed`). Regenerated with `bmake makepatch`; the patch keeps its
original purpose of appending `$comp_opts` to `INC` instead of letting
`CCFLAGS` clobber perl's own compiler flags. Note the new
`strbuf$(OBJ_EXT)` object in `OBJECT`.
- **`NO_TEST= yes` added.** 3.21.1 calls `plan tests => N` at the top of
`t/03dbmethod.t`, `t/10_pg_error_field.t` and `t/30unicode.t`, *before*
the `connect_database()` check that calls `plan skip_all`. With no
reachable PostgreSQL server that's two `plan` calls, so the files die
with `You tried to plan twice` instead of skipping. 3.20.2 had no early
`plan` and skipped cleanly. This is an upstream test-harness bug, not a
module defect ? the tests pass where a server is available.

`pkg-plist` needed no changes.

## Validation

| Check | Result |
|---|---|
| `bmake check-license` | pass |
| `bmake makesum` | SHA256 `f3fe2b2d?4db6`, 317973 bytes |
| `bmake patch` | pass (regenerated patch applies cleanly) |
| `bmake build` | pass ? builds `Pg.o dbdimp.o quote.o types.o strbuf.o`
|
| `bmake fake` | pass, staged tree matches `pkg-plist` |
| `bmake package` | pass |
| `portlint` | looks fine |

Built on amd64 against `postgresql14-client-14.20`. The build host has
perl5-5.40.4 installed while the tree's `lang/perl5.40` is at 5.40.5, so
builds ran with `NO_DEPENDS=yes PERL_VERSION=5.40.4` to avoid an
unrelated perl upgrade; this port does build an arch-specific `Pg.so`,
so a rebuild against 5.40.5 is worth confirming on the cluster.

? Generated with [Claude Code](https://claude.com/claude-code)

## Summary by Sourcery

Update databases/p5-DBD-Pg to version 3.21.1 and adjust port build and
test handling for the new release.

Enhancements:
- Update the DBD::Pg port to upstream version 3.21.1.

Build:
- Regenerate the Makefile patch to support the updated upstream build
configuration while preserving compiler flag handling.

Tests:
- Disable the port test target because the updated upstream suite
incorrectly plans twice when PostgreSQL is unavailable.

Signed-off-by: Lucas Holt <luke at foolishgames.com>



To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications


From noreply at github.com  Sun Aug 23 20:10:47 2026
From: noreply at github.com (Lucas Holt)
Date: Sun, 23 Aug 2026 17:10:47 -0700
Subject: [Midnightbsd-cvs] [MidnightBSD/mports]
Message-ID: <MidnightBSD/mports/push/refs/heads/databases/p5-DBD-Pg/f0ed1a-000000@github.com>

  Branch: refs/heads/databases/p5-DBD-Pg
  Home:   https://github.com/MidnightBSD/mports

To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications