Major release; see upstream release notes for details.
Recipe-side changes: patches / files moved to the new version directory.
AI-Generated: Uses Claude Code
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Major release; see upstream release notes for details.
Recipe-side changes: patches / files moved to the new version directory.
AI-Generated: Uses Claude Code
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Point release with bug fixes and minor improvements.
Recipe-side changes: patches / files moved to the new version directory.
AI-Generated: Uses Claude Code
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Minor release; see upstream release notes for details.
Recipe-side changes: patches / files moved to the new version directory.
AI-Generated: Uses Claude Code
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Minor release; see upstream release notes for details.
Recipe-side changes: patches / files moved to the new version directory.
AI-Generated: Uses Claude Code
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Point release with bug fixes and minor improvements.
Recipe-side changes: patches / files moved to the new version directory.
AI-Generated: Uses Claude Code
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Minor release; see upstream release notes for details.
Recipe-side changes: patches / files moved to the new version directory.
AI-Generated: Uses Claude Code
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Major release that tightens correctness and hardens the readers, per upstream
release notes. Full UTF-8 validation is now performed on every read (RFC 8259
8.1) including values reached through skip/raw_json and unmatched keys, with an
opt-out option; out-of-range integers are rejected rather than silently wrapped
(the range check previously ran after the operation that destroyed the value);
reader recursion is bounded at 256 levels so hostile buffers report
exceeded_max_recursive_depth instead of overflowing the stack; truncated
non-null-terminated input reports unexpected_end and blank input reports
no_read_input instead of spurious success; and BEVE variants move to the v2
wire format with tagging decided per variant rather than per alternative.
Buffers that used to parse may now correctly fail. Glaze remains a header-only
C++23 library (cxx_std_23 target feature unchanged) and no recipe in this layer
DEPENDS on it, so the ABI-less break has no in-layer consumers.
AI-Generated: Uses Claude Code
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
The recipe was pinned to the 4.2.4 tag but carried a +git PV suffix; 4.2.5 is
a real tagged release so the suffix is dropped. Upstream release notes list
three changes: support for uses-allocator construction, with std::uses_allocator
specialized for all function types (PR #62); std::aligned_storage_t is no longer
used, since it is deprecated in C++23 (PR #67); and an AppVeyor CI fix (PR #68).
Header-only library with no API break.
AI-Generated: Uses Claude Code
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Per upstream NEWS, 0.2.0 is a small additive release on the GtDBusQueue
mock-D-Bus harness: new functions to emit a signal from the mocked service
and to mock activatable bus names (!14), a switch to POSIX file functions
instead of GIO when writing activatable service files (!15), and
gt_dbus_queue_assert_pop_message() now accepts empty varargs (!16). The
remainder is build/CI work - gtkdoc-check now runs as part of the test suite
(!7) and the docs CI build and deployment were fixed (!13). The public API is
additive only, so no consumer changes were needed; the recipe just needed the
new tag SRCREV and PV.
AI-Generated: Uses Claude Code
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Adds Gen6 support across the CLI and library (new OSA sub-commands,
status query and capture-control variants) and reworks
switchtec_osa_capture_data() to fill in a caller supplied
switchtec_osa_capture_data structure rather than printing directly.
That rework replaced the run-time sized reply buffer, which 4.3 built
from a flexible array member plus alloca(), with a variable length array
declared as a struct member. That is a GCC-only extension which clang
rejects outright ('variable length array in structure' extension will
never be supported), so add a patch restoring the portable flexible
array member form and passing the computed reply size to switchtec_cmd().
AI-Generated: Uses Claude Code
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Picks up the 1.1.x line: the 1.1.0 release plus three subsequent bugfix
releases. valijson remains a header-only JSON Schema validation library
built with the same CMake options, so the recipe needs no changes beyond
PV/SRCREV; the BSD-2-Clause LICENSE is unchanged.
AI-Generated: Uses Claude Code
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Bugfix release in the 0.13 line (the newest snapper tag; the higher
numbered yast-snapper/2_2x_y tags in the same repository belong to the
separate YaST snapper module, not to snapper itself). No build system or
packaging changes; COPYING is unchanged.
AI-Generated: Uses Claude Code
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Three bugfix releases on the 3.15 line, per upstream release notes. 3.15.1
fixes a potential out-of-bounds access when locating the start of a broken
UTF-8 sequence during linebreaking (#3096) and TEMPLATE_LIST_TEST_CASE_METHOD
/ CATCH_TEMPLATE_PRODUCT_TEST_CASE potentially causing ODR violations (#3161).
3.15.2 stops --warn InfiniteGenerators from firing when the generator was
limited to a specific element by path filtering and silences
-Wunused-parameter in -fno-exceptions builds. 3.15.3 makes the JSON
reporter's number handling locale-independent (#3176), improves inf/NaN
handling in JSON output, and rewrites catch_discover_tests for a 4-5x speedup
with deterministic registration order. No API or ABI change; only the git tag
and SRCREV needed updating.
AI-Generated: Uses Claude Code
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Cap'n Proto 1.5.0 is primarily a KJ-HTTP security release. It hardens
the plain-HTTP serialization path against request smuggling / HTTP
desync: serializeRequest/serializeConnectRequest now reject request
targets containing whitespace or control characters and serializeResponse
rejects status text containing CR, LF or NUL (GHSL-2026-146, reported by
Kevin Stubbings), and the header parser no longer accepts whitespace
between a field name and the colon. It also picks up compatibility
fixes: builds against OpenSSL 4.0 while remaining compatible with older
OpenSSL, Bazel 9 support, boringssl Bzlmod repo_name, an MSVC build fix,
tolerating ERROR_INVALID_FUNCTION from GetFileInformationByHandleEx under
Wine, replacing the c++/LICENSE.txt symlink with a real file, and
scoping CTest to the project. LICENSE is unchanged and
0001-Export-binaries-only-for-native-build.patch still applies cleanly.
AI-Generated: Uses Claude Code
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Upstream released 2.36.4 on 2026-05-26 as a maintenance release on the
2.36 (atk-2.36/atkmm-2.36 ABI) series. It is documentation and
build-system work only: gnome.org/gtkmm.org doc links are refreshed, the
obsolete FSF postal address is dropped from the license texts,
codegen/extradefs is relabelled GPL-2.0 to match COPYING.tools, meson >=
0.60.0 and python3 >= 3.7 are now required, subprojects/atk.wrap is
repointed at at-spi2-core (GNOME/atk moved there), the Python
installation Meson itself uses is now reused, 'dot' is no longer
required to build the documentation, and the Visual Studio/NMake build
files are updated. No API or ABI change. LIC_FILES_CHKSUM is updated
for the FSF-address rewrite in COPYING and COPYING.tools; the license
terms (LGPL-2.1 / GPL-2.0) are unchanged.
AI-Generated: Uses Claude Code
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Upstream reworked the build plumbing in this release: most of the kconfig
machinery moved out of kconfig/Makefile into scripts/genconfig.sh and
scripts/kconfig.c, and the defconfig/allyesconfig/allnoconfig/randconfig
targets are now thin wrappers around genconfig.sh. As part of that
cleanup (commit 228cb983) the rules that build the host-side kconfig
frontends lost their $(HOSTCC) and were hardcoded to "cc", which breaks
cross builds where only an explicit HOSTCC is exported and no "cc" is on
PATH. Add a patch restoring $(HOSTCC) there; the rest of the recipe needs
no changes.
AI-Generated: Uses Claude Code
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Moves the Unicode Character Database from Unicode 14.0 to Unicode 16.0,
picking up the 15.0, 15.1 and 16.0 standard releases (roughly 5000 new
characters, including seven new scripts, plus the associated property
and normalization data file updates). 16.0.0 is the newest version
published under https://www.unicode.org/Public/zipped/. The license
file (license.txt) is byte-identical, so LIC_FILES_CHKSUM is unchanged.
AI-Generated: Uses Claude Code
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Two maintenance releases in the 4.1 stable line. No build system or
packaging changes were needed; the BSL-1.0 LICENSE_1_0.txt is unchanged
and all backend PACKAGECONFIG options are still present.
AI-Generated: Uses Claude Code
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Upstream added a configure-time check for libcrypt providing crypt_r()
and now unconditionally links against -lcrypt: auth.c uses crypt_r()
with a struct crypt_data for verifying yescrypt/scrypt/bcrypt/sha512crypt
and sha256crypt password hashes, so the build now hard-fails with
"No libcrypt with crypt_r()" when it is missing. Add virtual/crypt to
DEPENDS accordingly.
AI-Generated: Uses Claude Code
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Adobe's 2.042R-u/1.062R-i/1.026R-vf release re-compiles all fonts so
that the static instances correspond to the variable-font versions;
uprights are at 2.042 and italics at 1.062.
Upstream also reorganised the release assets: instead of a single
all-formats source-code-pro-<ver>.zip containing a top-level versioned
directory (plus EOT/OTF/TTF/WOFF trees and a bundled LICENSE.txt), each
format is now published as its own archive that unpacks a bare OTF/
directory with no license file. So SRC_URI now fetches
OTF-source-code-pro-2.042R-u_1.062R-i.zip with ;subdir=${BP}, S points
at that subdir, and LIC_FILES_CHKSUM references the common OFL-1.1 text
(the repository's LICENSE.md is the verbatim SIL OFL 1.1, only the
copyright year and Adobe notice differ, and it is no longer shipped in
the font archive). UPSTREAM_CHECK_REGEX is adjusted for the new
"R-u" tag suffix. All 14 OTF faces are packaged as before.
AI-Generated: Uses Claude Code
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Adobe released Source Han Sans 2.005 on 2025-06-18. Per the upstream
ReadMe version history it bumps the copyright range to 2014-2025,
refreshes the Latin component from Source Sans v3.052, and applies a
large batch of per-language glyph and CMap fixes across the
CN/TW/HK/JP/KR subsets (added U+31BC-U+31BF, U+2BB37, U+2D25D and
various Ext-B/C glyphs; corrected strokes, contour alignment and
interpolation bugs; removed and remapped a number of region-specific
glyphs to their corresponding shared forms; fixed IVS mappings in the
KR sequence file). Upstream also renumbered the release assets with an
ordinal prefix (SourceHanSansTW.zip is now 20_SourceHanSansTW.zip), so ZIPNAME is
updated to match.
AI-Generated: Uses Claude Code
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Adobe released Source Han Sans 2.005 on 2025-06-18. Per the upstream
ReadMe version history it bumps the copyright range to 2014-2025,
refreshes the Latin component from Source Sans v3.052, and applies a
large batch of per-language glyph and CMap fixes across the
CN/TW/HK/JP/KR subsets (added U+31BC-U+31BF, U+2BB37, U+2D25D and
various Ext-B/C glyphs; corrected strokes, contour alignment and
interpolation bugs; removed and remapped a number of region-specific
glyphs to their corresponding shared forms; fixed IVS mappings in the
KR sequence file). Upstream also renumbered the release assets with an
ordinal prefix (SourceHanSansKR.zip is now 18_SourceHanSansKR.zip), so ZIPNAME is
updated to match.
AI-Generated: Uses Claude Code
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Adobe released Source Han Sans 2.005 on 2025-06-18. Per the upstream
ReadMe version history it bumps the copyright range to 2014-2025,
refreshes the Latin component from Source Sans v3.052, and applies a
large batch of per-language glyph and CMap fixes across the
CN/TW/HK/JP/KR subsets (added U+31BC-U+31BF, U+2BB37, U+2D25D and
various Ext-B/C glyphs; corrected strokes, contour alignment and
interpolation bugs; removed and remapped a number of region-specific
glyphs to their corresponding shared forms; fixed IVS mappings in the
KR sequence file). Upstream also renumbered the release assets with an
ordinal prefix (SourceHanSansJP.zip is now 17_SourceHanSansJP.zip), so ZIPNAME is
updated to match.
AI-Generated: Uses Claude Code
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Adobe released Source Han Sans 2.005 on 2025-06-18. Per the upstream
ReadMe version history it bumps the copyright range to 2014-2025,
refreshes the Latin component from Source Sans v3.052, and applies a
large batch of per-language glyph and CMap fixes across the
CN/TW/HK/JP/KR subsets (added U+31BC-U+31BF, U+2BB37, U+2D25D and
various Ext-B/C glyphs; corrected strokes, contour alignment and
interpolation bugs; removed and remapped a number of region-specific
glyphs to their corresponding shared forms; fixed IVS mappings in the
KR sequence file). Upstream also renumbered the release assets with an
ordinal prefix (SourceHanSansCN.zip is now 19_SourceHanSansCN.zip), so ZIPNAME is
updated to match.
AI-Generated: Uses Claude Code
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Upstream 1.12.1 adds a meson build system alongside autotools (with CI
checks comparing the installed file list against xorgproto), aligns the
meson settables with the autotools ones, refreshes the gitlab CI
templates, and drops the "All rights reserved" phrasing from the Oracle
copyright notices. That copyright change alters COPYING, so
LIC_FILES_CHKSUM is updated accordingly (license terms themselves, MIT,
are unchanged). Upstream also switched the release tarball compression
from bzip2 to xz, so SRC_URI_EXT is set to "xz". The recipe continues
to build with autotools.
AI-Generated: Uses Claude Code
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Upstream 1.7.3 is a documentation-content release: it refreshes the
MAINTAINERS file (Xnest and Xft pick up new maintainers), drops the
"All rights reserved" phrasing from Oracle copyright notices, and
updates the gitlab CI templates and build-system boilerplate. Upstream
also switched the release tarball compression from bzip2 to xz for this
release, so SRC_URI_EXT is set to "xz". general/License.xml is
unchanged, so LIC_FILES_CHKSUM stays as-is.
AI-Generated: Uses Claude Code
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
The Python-language SWIG typemaps in OpenIPMI_lang.i use the Python 2
C-API names PyInt_Check/PyInt_AsLong/PyInt_FromLong/PyString_FromString(AndSize).
Python 3.14 no longer ships any fallback definitions for them, so the
generated OpenIPMI_wrap.c fails to compile:
OpenIPMI_wrap.c:13486:10: error: use of undeclared identifier 'PyInt_Check'
OpenIPMI_wrap.c:14950:21: error: use of undeclared identifier 'PyString_FromString'; did you mean 'PyLong_FromString'?
Add a patch mapping them to their Python 3 equivalents.
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
PyInt_AsLong is a Python 2 C-API name with no fallback definition in
Python 3.14, so any SWIG module that %includes this file (e.g. hmtrp)
fails to compile:
hmtrpPYTHON_wrap.cxx:7351:14: error: use of undeclared identifier 'PyInt_AsLong'; did you mean 'PyLong_AsLong'?
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
_rsa.i, _lib.i and _ec.i use the Python 2 C-API names PyInt_AsLong,
PyInt_FromLong and PyString_FromString. Python 3.14 no longer ships
any fallback definitions for them, so the generated _m2crypto_wrap.c
fails to compile:
src/SWIG/_m2crypto_wrap.c:4895:21: error: call to undeclared function 'PyInt_AsLong'
src/SWIG/_m2crypto_wrap.c:9842:26: error: call to undeclared function 'PyString_FromString'
Add a patch mapping them to their Python 3 equivalents in the shared
raw-C block emitted ahead of the %include'd modules that use them.
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Khem Raj