This lengthens dependency chains unnecessarily;
also -native versions are already included in DEPENDS via binutils.inc:
DEPENDS = "flex-native bison-native zlib-native gnu-config-native autoconf-native"
I have also checked that nothing in binutils needs anything from target
sysroots of flex/bison:
https://lists.openembedded.org/g/openembedded-core/message/155131
(From OE-Core rev: 8e9d3b5c8c994d307b2cca6a053b29403af9d5ef)
Signed-off-by: Alexander Kanavin <alex.kanavin@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The merge of libdl into libc in glibc 2.34 causes problems for pseudo. Add a fix
that works around this issue.
(From OE-Core rev: dd3e46a043c81cd4d81731a0f691868d3c059742)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Drop big-inodes-for-small-fs.patch: upstream made the same fix.
Drop 0001-lib-ext2fs-unix_io.c-do-unlock-on-error.patch
0001-lib-ext2fs-unix_io.c-revert-parts-of-libext2fs-fix-p.patch
(upstream has fixed the issue).
Add 0001-lib-ext2fs-unix_io.c-revert-parts-of-libext2fs-fix-p.patch
to correct a ptest failure due to incorrectly expected inode size
(recent change that wasn't run against the tests upstream?).
(From OE-Core rev: a4fc0af1050e5e0cc3d241279b92ea8c75aeeb8e)
Signed-off-by: Alexander Kanavin <alex.kanavin@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
6.15.0 - 2021-08-22
This release emits a more useful error message when @given() is applied to
a coroutine function, i.e. one defined using async def (issue #3054).
This was previously only handled by the generic return_value health check,
which doesn’t direct you to use either a custom executor or a library such
as pytest-trio or pytest-asyncio to handle it for you.
6.14.9 - 2021-08-20
This patch fixes a regression in Hypothesis 6.14.8, where from_type() failed
to resolve types which inherit from multiple parametrised generic types,
affecting the returns package (issue #3060).
References:
https://github.com/HypothesisWorks/hypothesis/issues/3054https://github.com/HypothesisWorks/hypothesis/issues/3060
(From OE-Core rev: ca0039a6e5524ab1e74df285ce16460dc980da65)
Signed-off-by: Tim Orling <timothy.t.orling@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Alongside GLIBC and UCLIBC, also relocate the musl loader.
(From OE-Core rev: 24c4e01892df1510e5ba64b89a0060f62f2384e6)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
In newer gcc versions the headers changed locations and our multiconfig
and nativesdk tweaks to loader paths wasn't working. The broke
buildtools-extended-tarball, particularly on arm. Update to fix this.
Ultimately we should dump the gcc specs and check for hardcoded paths.
This isn't possible quite yet as this patch doesn't fix the musl ones
as we don't use that in mutlilib or nativesdk scenarios at present.
(From OE-Core rev: 4ed6505a18ac76c0aa148deabb143214ac0c4552)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The elfutils ptests require debug symbols for the libc to be available, else
we see failures such as those on the autobuilder for the fast ptest image
on arm (the dbg symbols are pulled in by other recipes in other images).
Also fix various test skips/error messages due to missing gcc/ld and
development headers.
(From OE-Core rev: 41ecc76c7fab8f9805d3271255bcd027d87298bb)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Since upgrade of binutils to 2.37 builds of qtwebengine failed to link even
with ulimits -n 1000000 (!!).
Fix that by applying a patch from stable 'binutils-2_37-branch'.
(From OE-Core rev: 9f4660e1c6b251c55f9e7e8072b602edf843b952)
Signed-off-by: Andreas Müller <schnitzeltony@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Add PACKAGECONFIG for selinux rather than disable it directly. This is
useful for selinux distro feature.
(From OE-Core rev: d25b644741f5b941aa3051545e86e16a365c51c3)
Signed-off-by: Yi Zhao <yi.zhao@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ensure builds as deterministic by covering all compression configuration
options.
(From OE-Core rev: c78224b8546aa8d6bd238c2516c445b80de4c205)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The do_create_rdepends_inc function is used to recreate
the perl-rdepends.txt file. Update RDEPENDS_ to RDEPENDS:
for the new override syntax.
(From OE-Core rev: efc6ff1053aad393bbbd031f2504feb02891ae37)
Signed-off-by: Tim Orling <timothy.t.orling@intel.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Update patch as setup.cfg changes are now in upstream.
License-Update: Update Copyright years
NOTE: The 4.2.0 Release of SCons will deprecate Python 3.5 Support.
Python 3.5 support will be dropped in the next major release.
RELEASE 4.2.0 - Sat, 31 Jul 2021 18:12:46 -0700
From Byron Platt:
- Fix Install() issue when copytree recursion gives bad arguments
that can lead to install side-effects including keeping dangling
symlinks and silently failing to copy directories (and their
subdirectories) when the directory already exists in the target.
From Joseph Brill:
- Internal MSVS update: Remove unnecessary calls to find all
installed versions of msvc when constructing the installed visual
studios list.
From William Deegan:
- Improve Subst()'s logic to check for proper callable function or
class's argument list. It will now allow callables with expected
args, and any extra args as long as they have default arguments.
Additionally functions with no defaults for extra arguments as
long as they are set using functools.partial to create a new
callable which set them.
- Fix Issue #3035 - mingw with SHLIBVERSION set fails with either
not a dll error or "Multiple ways to build the same target were
specified for:". Now mingw will disable creating the symlinks
(and adding version string to ) dlls. It sets
SHLIBNOVERSIONSYMLINKS, IMPLIBNOVERSIONSYMLINKS and
LDMODULENOVERSIONSYMLINKS to True.
- Added --experimental flag, to enable various experimental
features/tools. You can specify 'all', 'none', or any combination
of available experimental features.
- Fix Issue #3933 - Remove unguarded print of debug information in
SharedLibrary logic when SHLIBVERSION is specified.
- Fix versioned shared library naming for MacOS platform.
(Previously was libxyz.dylib.1.2.3, has been fixed to
libxyz.1.2.3.dylib. Additionally the sonamed symlink had the same
issue, that is now resolved as well)
- Add experimental ninja builder. (Contributed by MongoDB, Daniel
Moody and many others).
- Fix#3955 - _LIBDIRFLAGS leaving $( and $) in *COMSTR output.
Added affect_signature flag to _concat function. If set to False,
it will prepend and append $( and $). That way the various
Environment variables can use that rather than
"$( _concat(...)$)".
- Fix issue with exparimental ninja tool which would fail on windows
or when ninja package wasn't installed but --experimental=ninja was
specified.
- As part of experimental ninja tool, allow SetOption() to set both
disable_execute_ninja and disable_ninja.
From David H:
- Fix Issue #3906 - `IMPLICIT_COMMAND_DEPENDENCIES` was not properly
disabled when set to any string value (For example ['none','false',
'no','off']) Also previously 'All' wouldn't have the desired affect.
From Ivan Kravets:
- Provide a custom argument escape function for `TempFileMunge`
using a new `TEMPFILEARGESCFUNC` variable. Useful if you need to
apply extra operations on a command argument before writing to a
temporary file (fix Windows slashes, normalize paths, etc.)
From Henrik Maier:
- DocbookXslt tool: The XSLT stylesheet file is now initialized to an
env.File() Node, such that dependencies work correctly in hierarchical
builds (eg when using DocbookXslt in SConscript('subdir/SConscript')
context.
From Daniel Moody:
- Update CacheDir to use uuid for tmpfile uniqueness instead of pid.
This fixes cases for shared cache where two systems write to the
same cache tmpfile at the same time because the happened to get the
same pid.
- Added support for passing custom CacheDir derived classes to
SCons. Moved copy_from_cache attribute from the Environment class to
CacheDir class. Code contributed by MongoDB.
- Update BuildTask to pass all targets to the progress object fixing
an issue where multi-target build nodes only got the first target
passed to the progress object.
- Fix a potential race condition in shared cache environments where
the permissions are not writeable for a moment after the file has
been renamed and other builds (users) will copy it out of the cache.
Small reorganization of logic to copy files from cachedir. Moved
CacheDir writeable permission code for copy to cache behind the atomic
rename operation.
- Added marking of intermediate and and multi target nodes generated
from SConf tests so that is_conftest() is more accurate.
- Added test for configure check failing to ensure it didn't break
generating and running ninja.
From Mats Wichmann:
- Initial support in tests for Python 3.10 - expected bytecode and
one changed expected exception message. Change some more regexes
to be specified as rawstrings in response to DeprecationWarnings.
- Add an example of adding an emitter to User Guide (concept from
Jeremy Elson)
- Add timing information for sconsign database dump when --debug=time
is selected. Also switch to generally using time.perf_counter,
which is the Python recommended way for timing short durations.
- Drop remaining definitions of dict-like has_key methods, since
Python 3 doesn't have a dictionary has_key (maintenance)
- Do not treat --site-dir=DIR and --no-site-dir as distinct options.
Allows a later instance to override an earlier one.
- Ignore empty cmdline arguments when computing targets (issue 2986)
- Remove long-deprecated construction variables PDFCOM, WIN32_INSERT_DEF,
WIN32DEFPREFIX, WIN32DEFSUFFIX, WIN32EXPPREFIX, WIN32EXPSUFFIX.
All have been replaced by other names since at least 1.0.
- Add a __iadd__ method to the CLVar class so that inplace adds
(+=) also work as expected (issue 2399)
- Remove local copy of CLVar in EnvironmentTests unittest file -
should be testing against the production version, and they
didn't really differ.
- Don't strip spaces in INSTALLSTR by using raw subst (issue 2018)
- Deprecate Python 3.5 as a supported version.
- CPPDEFINES now expands construction variable references (issue
2363)
- Restore behavior that Install()'d files are writable (issue 3927)
- Simplified Mkdir(), the internal mkdir_func no longer needs to
handle existing directories, it can now pass exist_ok=True to
os.makedirs().
- Avoid WhereIs exception if user set a tool name to empty (from
issue 1742)
- Maintenance: remove obsolete __getslice__ definitions (Py3 never
calls); add Node.fs.scandir to call new (Py3.5) os.scandir;
Node.fs.makedirs now passes the exist_ok flag; Cachedir creation
now uses this flag.
- Maintenance: remove unneeded imports and reorganize some. Fix
uses of warnings in some tools which instantiated the class but did
nothing with them, need to instead call SCons.Warnings.warn with the
warn class.
- Drop overridden changed_since_last_build method in Value class.
- Resync the SetOption implementation and the manpage, making sure
new options are available and adding a notes column for misc
information. SetOption equivalents to --hash-chunksize,
--implicit-deps-unchanged and --implicit-deps-changed are enabled.
- Add tests for SetOption failing on disallowed options and value
types.
- Maintenance: eliminate lots of checker complaints about Util.py.
- Maintenance: fix checker-spotted issues in Environment (apply_tools)
and EnvironmentTests (asserts comparing with self). For consistency,
env.Tool() now returns a tool object the same way Tool() has done.
- Change SConscript() missing SConscript behavior - if must_exist=False,
the warning is suppressed.
- Make sure TEMPFILEPREFIX can be set to an empty string (issue 3964)
From Dillan Mills:
- Add support for the
(TARGET,SOURCE,TARGETS,SOURCES,CHANGED_TARGETS,CHANGED_SOURCES}.relpath
property. This will provide a path relative to the top of the build
tree (where the SConstruct is located) Fixes#396
From Andrew Morrow:
- Fix issue #3790: Generators in CPPDEFINES now have access to
populated source and target lists
(From OE-Core rev: 37a0a8821838a15cc24e1fc84b632bebacb44fb0)
Signed-off-by: Tim Orling <timothy.t.orling@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
e2fsprogs calls filesystems larger than 3MB but smaller than 512MB
"small", which has some implications:
- blocksize 1024 instead of 4096
- inode_ratio 4096 instead of 16384
- inode_size 128 instead of 256
The outcome of the inode size dropping to 128 bytes is that they cannot
store 64-bit timestamps, so are not Y2038-safe.
A previous attempt to solve this problem[1] changed some of the canned
wic files to pass -T default to mkfs.ext4, but this only covered wic
images and not traditional images. Also, actually small filesystems,
for example a core-image-minimal, will happily be tens of megabytes and
with the "default" options will result in an image which runs out of
blocks before it runs out of space:
mkfs.ext4: Could not allocate block in ext2 filesystem while populating file system
Considering that many OpenEmbedded images are in fact "small", being
2038-safe is worth the marginal increase is disk usage. This patch
alters the small configuration in native builds so that it also has
256-byte inodes. Target is unchanged so that standard behaviour is
maintained outside of the build.
This is actually the same underlying patch that Mathieu Dubois-Briand
sent in April, but the wic change in [1] was accepted instead. I believe
that is the wrong approach and this approach covers more cases.
[ YOCTO #14478 ]
[1] openembedded-core eecbe62
[2] https://lists.openembedded.org/g/openembedded-core/message/150298
(From OE-Core rev: 9ab0ae83a24ee99e69f8ac54256b253a122aef8a)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
If lto is enabled, we need the prefix-map variables to be passed to the linker.
Add these to the list of options libtool passes through.
(From OE-Core rev: 2c26d2c00b47df856fb2d9c35486b135094d46ac)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
If we add DEBUG_PREFIX_MAP into LDFLAGS, ruby and ruby-dbg are no
longer reproducible. Fix this.
(From OE-Core rev: c34d5e409b291e69d6ff28ee71acba7055ef3f08)
Signed-off-by: Tony Battersby <tonyb@cybernetics.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Backport ustream gcc patch that enables -fdebug-prefix-map to cover
additional cases with LTO enabled to make LTO builds more reproducible.
[YOCTO #14481]
(From OE-Core rev: 8923253b9bc91b14b474abc4923ca916fb8a12ec)
Signed-off-by: Tony Battersby <tonyb@cybernetics.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Add a patch being discussed upstream to fix a ptest issue with glibc 2.34.
(From OE-Core rev: 8921f2acfd566d2c03cea7bdb9f0b1883994148b)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This includes changes to work with glibc 2.34.
(From OE-Core rev: 58cdb3ff8689c5e87769eb53d154a395492cdeb5)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
6.14.5 - 2021-07-27
This patch fixes hypothesis.strategies._internal.types.is_a_new_type. It
was failing on Python 3.10.0b4, where NewType is a function.
6.14.4 - 2021-07-26
This patch fixes from_type() and register_type_strategy() for
typing.NewType on Python 3.10, which changed the underlying
implementation (see bpo-44353 for details).
References:
https://bugs.python.org/issue44353
(From OE-Core rev: 19bc156c8f9d43c963b2db3535d11d05c6916600)
Signed-off-by: Tim Orling <timothy.t.orling@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The timeout for threading.Lock, threading.Condition, etc, is not using
a monotonic clock, it is affected if the system time (realtime clock)
is set.
This patch will make condvar use monotonic clock.
Refence: https://bugs.python.org/issue41710
(From OE-Core rev: 5a268f95a5bf5ee8c244a8af685d6c84aad9a4ac)
Signed-off-by: Zqiang <qiang.zhang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
In sdk, call createrepo-c failed with:
...
$ createrepo_c --update ./test_repo/rpm
Directory walk started Critical: Failed to detect compression for file
./test_repo/rpm/cortexa72/hello-2.10-r0.cortexa72.rpm: magic_load() failed: could not find any valid magic files!
...
Since commit [ea666fbc74 createrepo-c: set path to magic database for
native and nativesdk] applied, the MAGIC is incorrectly assigned.
The variable datadir will be expanded automatically for nativesdk,
do not need to add prefix ${SDKPATHNATIVE} to MAGIC
(From OE-Core rev: 54368f1b02e1ac4aa068515730a8c8bcd3683eb3)
Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Those were missed in previous rounds of automated and manual conversion.
(From OE-Core rev: 22f9c7268b542baf6cd8aa0e34c8fb7aa1579e08)
Signed-off-by: Denys Dmytriyenko <denis@denix.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Process
The source distribution re-installation feature removal has been delayed to 21.3.
(From OE-Core rev: bc744bad1e338445d5fa02471cdd6389ce44fabf)
Signed-off-by: Zang Ruochen <zangrc.fnst@fujitsu.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
It does not show correct version info of python3-pytest:
$ pip3 list | grep pytest
pytest 0.0.0
pytest requires setuptools-scm in setup.cfg. It could be met by adding
python3-setuptools-scm-native to dependency and then it will not
download egg file of setuptools-scm during do_compile any more. So
remove 0001-setup.py-remove-the-setup_requires-for-setuptools-scm.patch
and depend on python3-setuptools-scm-native to make it show the correct
version info.
(From OE-Core rev: 9159db64d3b3c7633deb135e683317fd3e98852c)
Signed-off-by: Kai Kang <kai.kang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
