Richard Purdie [Mon, 6 Jul 2026 13:53:37 +0000 (14:53 +0100)]
staging: Fix vardepsexclude
Set staging_populate_sysroot_dir vardepsexclude correctly (which does reference the variables)
and remove the incorrect pieces from extend_recipe_sysroot.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
selftest/locales: opt out of ptests in DISTRO_FEATURES
selftests/locales tests configure glibc-locales to produce a restricted set,
and build core-image-minimal. This is problematic if something in
the dependency chain of that image relies on locales not in that set.
Until now, by coincidence, nothing did, but recent version updates
(particularly, json-c which relies on xxd in its ptest) pull in glib-2.0,
which for its ptest requires various locales not in the set. This causes
package_qa errors.
To address this particular error, and to minimize chances of it happening
again, let's build images for this selftest without ptest enabled; typically
only ptests want locales.
Signed-off-by: Alexander Kanavin <alex@linutronix.de> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ross Burton [Fri, 3 Jul 2026 10:07:10 +0000 (11:07 +0100)]
classes/insane: increase shebang size limit
Since Linux 5.1 the shebang buffer has been 256 bytes[1], so update the
check to match.
Also rewrite the test, create a variable for the magic number (that has
the same name as the kernel #define), read double that so we either have
a full line, or definitely are over the buffer, don't pointlessly try to
decode the bytes as UTF-8, and consolidate the test logic to entirely
inside the try block.
Also update the tests: sysroot-shebang-test needs updating to write a
longer shebang, and generalise the test case so that the string being
searched for isn't so specific.
[1] linux 6eb3c3d0a52dc ("exec: increase BINPRM_BUF_SIZE to 256")
Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ross Burton [Fri, 3 Jul 2026 10:07:09 +0000 (11:07 +0100)]
classes/insane: clean up do_qa_sysroot
do_populate_sysroot copies SYSROOT_DIRS into SYSROOT_DESTDIR, so there's
no need to loop over SYSROOT_DIRS when checking files: qa_check_staged
can just recurse down the entire tree.
Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Alex Kiernan [Wed, 1 Jul 2026 11:58:12 +0000 (12:58 +0100)]
tune-cortexa32: enable Thumb-2 and fix AArch32 crypto FPU selection
Cortex-A32 is an AArch32-only ARMv8-A core, but its tune was configured
as if it were a plain 64-bit-style tune: TUNE_FEATURES lacked both 'arm'
and 'thumb', so userspace and the kernel were built as fixed-width ARM
rather than Thumb-2, and the package architecture names did not match a
Thumb-2 build.
Add 'thumb' so the core is built as Thumb-2, and 'arm' so that
feature-arm-thumb.inc honours a recipe's explicit ARM_INSTRUCTION_SET =
"arm" instead of silently forcing it to thumb. Without 'arm', ARM_M_OPT
is pinned to "thumb" and recipes that request ARM warn, e.g.:
Recipe 'libmad' selects ARM_INSTRUCTION_SET to be 'arm', but tune
configuration overrides it to 'thumb'
On AArch32 the ARMv8 crypto extensions (AES/SHA) are part of the FPU,
not just the -march ISA. The 'crypto' feature only appends '+crypto' to
-march and leaves -mfpu at plain 'neon', so building the AES intrinsics
(e.g. mbedtls aesce.c) fails:
error: inlining failed in call to 'always_inline' 'vaesdq_u8':
target specific option mismatch
Select the crypto FPU by appending crypto-neon-fp-armv8 as the last word
of TUNE_CCARGS_MFPU when both 'crypto' and 'neon' are active, mirroring
the vfpv4+neon -> neon-vfpv4 idiom in feature-arm-neon.inc.
Update PACKAGE_EXTRA_ARCHS to match the resulting package arches:
cortexa32hf-neon becomes cortexa32t2hf-neon, and the crypto variant
becomes cortexa32t2hf-neon plus cortexa32t2hf-crypto-neon-fp-armv8.
AI-Generated: Claude Code (Claude Opus 4.8) Signed-off-by: Alex Kiernan <alex.kiernan@gmail.com> Signed-off-by: Alex Kiernan <alexk@a-squared-projects.uk> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Alex Kiernan [Wed, 1 Jul 2026 11:58:11 +0000 (12:58 +0100)]
rust-target-config: armv8a AArch32 Thumb build fixes
llvm_features_from_tune() calls target_is_armv7() to gate both the +v7
LLVM feature and the +thumb2 feature. target_is_armv7() (in
rust-common.bbclass) only recognises armv7a/armv7r/armv7m/armv7ve -- it
returns False for armv8a. As a result an ARMv8-A AArch32 Thumb build
gets "+neon,+thumb-mode" in the target JSON but neither "+v8" nor
"+thumb2".
LLVM's ParseARMTriple() returns "" for "arm-..." triples
(ARM::parseArch("arm") = INVALID), leaving HasV4TOps unset in the
MCSubtargetInfo used for module-level inline asm parsing. The ARM ASM
parser's hasThumb() then returns false and ".thumb" directives in
compiler_builtins naked-function trampolines (__aeabi_uldivmod etc.)
are rejected.
This produces two failures when building libstd-rs / compiler_builtins:
1. core: "LLVM ERROR: Cannot select: br" in core::fmt::num -- LLVM ARM
ISel patterns for Thumb-2 unconditional branches (t2B) are gated on
HasV8Ops as well as IsThumb2; without +v8 the pattern never fires.
2. compiler_builtins: "target does not support Thumb mode" on .thumb
directives in naked-function trampolines (__aeabi_uldivmod etc.) --
two separate paths both require +thumb2: HasV4TOps in the
MCSubtargetInfo (set transitively via ParseARMTriple) and
FeatureThumb2 in the code-gen ARMSubtarget.
The natural fix would be to change RUST_TARGET_SYS to "armv8a-...", but
"armv8a" is not a recognised Rust target architecture so the triple is
not valid there. Instead rust_sys_to_llvm_target() rewrites only the
JSON llvm-target field, leaving RUST_TARGET_SYS and rust-common.bbclass
unchanged.
- Add +v8 to llvm_features_from_tune for armv8a AArch32.
- Add +thumb2 to llvm_features_from_tune for armv8a AArch32 Thumb.
- Extend rust_sys_to_llvm_target() to accept the datastore and rewrite
the "arm-..." triple to "armv8a-..." for armv8a AArch32 targets, so
the JSON llvm-target field uses "armv8a-..." while RUST_TARGET_SYS
keeps "arm-...". ParseARMTriple then adds "+armv8-a" to the
MCSubtargetInfo, transitively setting HasV4TOps and making hasThumb()
return true.
The TARGET_ARCH == 'arm' guard is required because AArch64 tunes also
carry armv8a in TUNE_FEATURES (via arch-armv8a.inc) but must not be
affected.
AI-Generated: Claude Code (Claude Opus 4.8) Signed-off-by: Alex Kiernan <alex.kiernan@gmail.com> Signed-off-by: Alex Kiernan <alexk@a-squared-projects.uk> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ross Burton [Fri, 19 Jun 2026 14:12:22 +0000 (15:12 +0100)]
classes/archiver: remove SRPM mode
The archiver supports saving the source code in either a tarball or a
Source RPM (SRPM). Tarballs are a logical choice as they're a standard
format for source code, but SRPMs are primarily a way to distribute
source plus packaging scripts for RPM-based distributions, and the SRPMs
generated by the archiver do not include any packaging scripts.
This can result in confusion when people may think that you can take the
SRPM generated by the archiver and rebuild it with rpmbuild, whereas this
will not work.
I think we should remove the SRPM mode from the archiver on the grounds
that tarballs are idiomatic for sharing software, and SRPMs are actively
confusing.
Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
- Added default feature to disable ArgumentParser coloring
- Trying to fix RTD refusing to build
- Moved color-disabling into parse_parser to fix subparsers
- Unpinned ruff for CI + Cleaned up workflow files a little
- Updated license metadata
- Changed filename field to be explicit
- Fixed MyPy
- Added try-catch for autodoc mock-import
License file was renamed but its content unchanged.
webkitgtk: allow usign clang to compile for arm target
This commit reverts: be459bf17d: "webkitgtk: Use gcc to compile for arm target"
The issue #132322 [1] was resolved, and the corresponding fix [2] is
available in LLVM's v22.1.8 (currently used in oe-core master). It is
actually available since v22.1.0.
I tested it with MACHINE="qemuarm", and it builds correctly.
Signed-off-by: João Marcos Costa <joaomarcos.costa@bootlin.com> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
opensbi: don't override ELFFLAGS, silence ldflags QA for bare-metal ELFs
opensbi_1.8.1.bb passes ELFFLAGS="${LDFLAGS}" via EXTRA_OEMAKE. Being a
make command-line assignment, it overrides OpenSBI's in-Makefile
"ELFFLAGS +=" and drops the mandatory firmware link flags:
Losing --gc-sections retains unused fdt_*/atomic_* code and grows .text;
losing -pie/--no-dynamic-linker drops the M-mode firmware hardening.
The distro LDFLAGS were not reaching the link regardless: OpenSBI
rebuilds CC from CROSS_COMPILE (Makefile: CC = $(CROSS_COMPILE)gcc), so
oe's CC and its TARGET_CC_ARCH += "${LDFLAGS}" do not apply, and passing
them via ELFFLAGS was the only path in. They are also dynamic-linking
flags with no role in bare-metal firmware, so drop the override;
REPRODUCIBLE=y and CROSS_COMPILE stay.
Removing it re-exposes an ldflags QA warning that the injected
--hash-style=gnu had been satisfying. These firmware ELFs are bare-metal
M-mode binaries (--build-id=none, no dynamic linker) that legitimately
carry no GNU_HASH, so add INSANE_SKIP += "ldflags", as linux-firmware
does.
Tested on oe-core master (DISTRO=nodistro, MACHINE=qemuriscv64) via the
oe-nodistro-master bitbake-setup config:
Before, the fw_jump.elf link carries only the LDFLAGS content (-Wl,-O1,
--hash-style=gnu, --as-needed, -z relro/now) with OpenSBI's flags absent,
and do_package_qa reports ldflags errors on all three firmware ELFs.
After, the link carries OpenSBI's flags (-Wl,--gc-sections,
--exclude-libs,ALL, --build-id=none, --no-dynamic-linker, -pie) and QA
passes. On a single-SoC defconfig, restoring --gc-sections roughly
halves fw_jump.bin.
Markus Volk [Wed, 1 Jul 2026 16:55:35 +0000 (18:55 +0200)]
webkitgtk: add PACKAGECONFIG for librice
This avoids:
| -- Checking for module 'rice-proto'
| -- Package 'rice-proto' not found
| -- Checking for module 'rice-io'
| -- Package 'rice-io' not found
| -- Could NOT find Rice (missing: Io Proto) (Required is at least version "0.1.1")
| CMake Error at Source/cmake/GStreamerChecks.cmake:67 (message):
| librice-{io,proto} is needed for USE_LIBRICE.
Note: librice is availabe in meta-oe
Signed-off-by: Markus Volk <f_l_k@t-online.de> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ross Burton [Wed, 1 Jul 2026 16:12:49 +0000 (17:12 +0100)]
meson: remove obsolete environment unsets
We no longer need to remove environment variables because their behaviour
was fixed in meson 0.54.0[1]. The corresponding cleanup was done in
meson.bbclass some time ago[2] but was not done here.
[1] https://mesonbuild.com/Release-notes-for-0-54-0.html#environment-variables-with-cross-builds
[2] oe-core 20a5af2583d ("meson: use native-file instead of environment variables")
Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Currently, patchtest run with --directory responds to a CTRL+C press by
aborting only the current patch being tested, and moves onto the next.
Instead, this key press should stop the test entirely. Remove usage of
the unittest.installHandler() function (which intercepts the signal) and
handle it ourselves.
Also make sure that the branch is properly reset with git am --abort
afterwards, and that the return code is properly set in the sigint
handler function. Finally, update patchtest_parser.py so that the helper
info reflects this change.
AI-Generated: Uses Claude Code
Signed-off-by: Trevor Gamblin <tgamblin@baylibre.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
patchtest: tests: test_python_pylint.py: use more Pythonic comparison, fix indent
- Use "!=" instead of "is not" when checking for a renamed file
- Fix the pylint readlines() block so that indentation is consistent
with the rest of the module (four spaces instead of eight)
Signed-off-by: Trevor Gamblin <tgamblin@baylibre.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
patchtest: tests: test_patch.py: simplify source patch and upstream status checking
- Only check for new patches once in setUp(), rather than in each
function
- Fix misleading "CVE" messages when we're really looking for any new
source patches which are added
- Move patterns for checking patch upstream status into the relevant
test
Signed-off-by: Trevor Gamblin <tgamblin@baylibre.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The pretest_src_uri_left_files() and test_src_uri_left_files() functions
both use the same logic to check for changes, so encapsulate that into a
function for easier re-use. Also remove a pyparsing import which isn't
needed, since it already gets pulled in through other patchtest modules.
Signed-off-by: Trevor Gamblin <tgamblin@baylibre.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
patchtest: tests: test_mbox.py: improve variable and message clarity, fix whitespace
- Rename "l" to "length" so it's clearer
- Make the message printed for an invalid author (AUH, in particular)
more explicit
- Remove an extra space in an if block inside test_target_mailing_list()
Signed-off-by: Trevor Gamblin <tgamblin@baylibre.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The libarchive fuzz tests (test_fuzz_ar, test_fuzz_cab, test_fuzz_cpio,
test_fuzz_iso9660, ...) run 1000 randomised iterations per archive format
and are intended for dedicated fuzzing infrastructure rather than routine
ptest execution. On the time and memory constrained autobuilder QEMU
targets they regularly time out, e.g.:
158: test_fuzz_iso9660 Timeout! System state:
Upstream provides the SKIP_TEST_FUZZ environment variable to skip these
tests. Export it from run-ptest so the fuzz tests are skipped rather than
run, while the rest of the test suite continues to execute normally.
[YOCTO #16340]
Signed-off-by: Siva Balasubramanian <sivakumar.bs@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ross Burton [Thu, 2 Jul 2026 12:56:25 +0000 (13:56 +0100)]
at-spi2-core: inherit python3-dir not python3targetconfig
Now that meson.bbclass tells meson where to install Python files this
recipe just needs to know the value of PYTHON_SITEPACKAGES_DIR for
FILES, so just inherit python3-dir to remove a dependency on the target
python recipe.
Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ross Burton [Thu, 2 Jul 2026 12:56:24 +0000 (13:56 +0100)]
python3-pycairo: inherit python3-dir not python3targetconfig
Now that meson.bbclass tells meson where to install Python files this
recipe just needs to know the value of PYTHON_SITEPACKAGES_DIR for
FILES, so just inherit python3-dir to remove a dependency on the target
python recipe.
Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ross Burton [Thu, 2 Jul 2026 12:56:23 +0000 (13:56 +0100)]
blueprint-compiler: inherit python3-dir not python3targetconfig
Now that meson.bbclass tells meson where to install Python files this
recipe just needs to know the value of PYTHON_SITEPACKAGES_DIR for
FILES, so just inherit python3-dir to remove a dependency on the target
python recipe.
Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ross Burton [Thu, 2 Jul 2026 12:56:22 +0000 (13:56 +0100)]
classes/meson: set python install locations in the cross file
Instead of needing recipes to inherit python3targetconfig so that Python
when asked will report the correct target directory to install into, we
can pass the right directory via meson's cross file.
This means recipes that currently inherit python3targetconfig to get
this path no longer need to.
Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Joshua Watt [Tue, 30 Jun 2026 21:01:44 +0000 (15:01 -0600)]
wic-tools: Change to MACHINE_ARCH
Changes wic-tools to be MACHINE_ARCH. This isn't exactly an ideal, but
wic-tools is attempting to provide a target sysroot that (potentially)
contains the EFI binaries for wic (and, this is the only way wic is
willing to find those binaries). Since the EFI binaries are
MACHINE_ARCH, wic-tools must also be MACHINE_ARCH otherwise it causes
problems.
Signed-off-by: Joshua Watt <JPEWhacker@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Joshua Watt [Tue, 30 Jun 2026 21:01:42 +0000 (15:01 -0600)]
systemd-boot: Change to MACHINE_ARCH
Machines can change EFI_PROVIDER which changes the way systemd-boot is
built, meaning two machine may not build it the same way. As such, the
arch should be MACHINE_ARCH.
Signed-off-by: Joshua Watt <JPEWhacker@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Joshua Watt [Tue, 30 Jun 2026 21:01:41 +0000 (15:01 -0600)]
grub-efi: Change to MACHINE_ARCH
Machines can change EFI_PROVIDER which changes the way grub-efi is
built, meaning two machine may not build it the same way. As such, the
arch should be MACHINE_ARCH.
Signed-off-by: Joshua Watt <JPEWhacker@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Sun, 5 Jul 2026 09:43:12 +0000 (10:43 +0100)]
libjpeg-turbo: Add PIC flags
Our cross compiler generates PIE by default but the host may not. Add
oppropriate cmake flags to avoid build failures in the native variant
on hosts like OpenSUSE 16.0.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ross Burton [Thu, 2 Jul 2026 13:24:16 +0000 (14:24 +0100)]
libevent: upgrade 2.1.12 -> 2.1.13
Security Fixes (evtag, evrpc):
Fix an out-of-bounds read in decode_tag_internal.
(Found by @Brubbish. GHSA-fj29-64w6-73h6)
Fix an integer overflow in evtag_unmarshal_header.
(Found by @Brubbish. GHSA-45c6-qx49-89m8)
Security Fixes (evhttp):
Discard HTTP trailers, to prevent header smuggling attacks.
(Found by @sebastianosrt. GHSA-2gmv-p5m7-98p6)
Restrict HTTP header parsing to prevent request smuggling.
(Originally reported by @xclow3n; and then by @kodareef5,
@nstaller0490, @AsafMeizneer, and @yaotushaozhu.
GHSA-q39v-w2g7-gr8j.)
Treat CRLF and %00 more strictly in HTTP headers, to prevent
parser mismatch attacks.
(Reported by @xclow3n and @AsafMeizner. See GHSA-q39v-w2g7-gr8j,
GHSA-jcwh-pvf2-73p2.)
Fix a heap out-of-bound write that could occur when using
AF_UNIX sockets and compiling libevent with -DNDEBUG.
(Found by @mat-mo. GHSA-cvq5-vrvr-j338)
Security fixes (evbuffer, bufferevent):
Fixed a dangling pointer in evbuffer_add_reference.
(Found by @DarkaMaul. GHSA-c2pj-cg4r-88c8)
Security fixes (evdns):
Fix an out-of-bounds write in dnsname_to_labels
when building a DNS response of 2^16 bytes.
(Found by @sectroyer. GHSA-58rx-7448-jw47)
Security fixes (example code):
Avoid using strcpy() in sample/http-server.c.
(Reported by @sectroyer. GHSA-5rgj-2c58-7jrc.)
Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Roland Kovacs [Wed, 1 Jul 2026 08:26:39 +0000 (10:26 +0200)]
gnupg: fix CVE-2026-57062
CMS (Cryptographic Message Syntax) parsing in gpgsm in GnuPG through 2.5.20
mishandles the CMS format for AES-GCM because aes-ICVlen is supposed to be
12 bytes but 4 bytes is accepted.
Signed-off-by: Roland Kovacs <roland.kovacs@est.tech> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Roland Kovacs [Wed, 1 Jul 2026 08:26:38 +0000 (10:26 +0200)]
gnupg: Upgrade 2.5.17 -> 2.5.20
Bug fixes included in this release:
- gpg: Fix wrong assertion failure which could very rarely occur
during key signature checking. [rG693f5642f6]
- gpg: Consider certify-only keys for revocation signature check.
[T8196]
- gpgsm: Fix possible double free in the CMS parser. [T8240]
- gpgsm: Fix possible too early removal of ephemeral keys. [T8236]
- gpgsm: Avoid emitting a final FAILURE status line if --status-fd
is not used. [rG69c27fe377]
- gpgsm: Fix a regression in 2.5.19 for password encrypted GCM
data. [rG60a823c97b]
- agent: Fix not using cache for pinentry loopback. [rGd4b608a31f]
- agent: Fix command PUT_SECRET by saving input line. [rG1875bc185e]
- keyboxd: Mark keys searched but not imported via LDAP correctly
as ephemeral. [T8048]
- scdaemon: Avoid buffer overflow with SC-HSM cards providing RSA
keys > 2k. [T8244]
- dirmngr: Fix uninitialized use of the dns_any union in
dns_rr_cmp. [T8251]
Release-info: https://dev.gnupg.org/T7997
Signed-off-by: Roland Kovacs <roland.kovacs@est.tech> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Tasks should never be dependning on SSTATETASKS since that can change for things
which code should be independent from (e.g. package backend), breaking sstate reuse.
Add the appropriate vardepexclude entres.
This currently isn't an issue due to extend_recipe_sysroot being globally
excluded from hashes.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
sanity.bbclass: warn on cargo config files outside the build tree
Cargo walks from CWD up to the filesystem root merging every
.cargo/config[.toml] it finds. Any such file above BASE_WORKDIR is
silently picked up and can override Yocto's linker, registry or
compiler settings, leading to build failures.
Until cargo provides a proper fix upstream, add a warning so users
get a clear diagnostic instead of a build error.
Yoann Congal [Tue, 30 Jun 2026 13:48:54 +0000 (15:48 +0200)]
dropbear: Add missing WTFPL & Unlicense in LICENSE and LIC_FILES_CHKSUM
Vendored libtomcrypt and libtommath are respectively WTFPL and
Unlicense: update LICENSE to reflect that and include vendored LICENSE
files in LIC_FILES_CHKSUM to detect changes.
Note: libtomcrypt/LICENSE also specify "Public domain" but that is not a
real licence, so choose the alternative: WTFPL.
Ross Burton [Mon, 29 Jun 2026 14:15:57 +0000 (15:15 +0100)]
classes/meson: use ninja explicitly when compiling
"meson compile" is essentially a glorified wrapper around calling ninja
that doesn't support all of the options that ninja does, so calling it
directly means builds are fractionally faster and we get direct control
over the flags that ninja is passed.
A longer rationale can be found in the Gentoo change by Eli Schwartz
that caused this patch, who is both a Gentoo and Meson developer:
This patch was badly rebased onto upstream changes and ended up setting
install_dir twice. Refresh the patch to change the existing value
instead of adding another.
Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
CMAKE_VERBOSE is already enabled in cmake.bbclass (inherited in this
recipe). I would have considering uncommenting the line if it originally
meant to disable the verbose mode.
Signed-off-by: João Marcos Costa <joaomarcos.costa@bootlin.com> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
projects / thirdparty / openembedded / openembedded-core.git / log
