Upgrade to release 0.35.0:
- Mark that Trio supports 3.15.
- Ensure that Trio consistently encodes all hostnames using IDNA
2008.
- Avoid erroring when running in a CPython built for an Android API
version less than 24.
- Ensure trio.lowlevel.notify_closing works even on events that
have been deleted, e.g. because they were registered as one shot
events.
- trio.Process no longer emits an EncodingWarning on Linux when
Python runs with -X warn_default_encoding
(or PYTHONWARNDEFAULTENCODING=1).
- The typing for async-file-objects is now consistent about what it
exposes based on what is exposed by the underlying object.
Signed-off-by: Leon Anavi <leon.anavi@konsulko.com>
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Upgrade to release 0.0.33:
This release lets you cap URL-encoded form fields as they stream in, and
fixes several upload and decoding edge cases. It also drops Python 3.10
and adds Python 3.15.
- Limit URL-encoded fields while streaming. The new opt-in
MAX_FIELDS and MAX_FIELD_SIZE options for FormParser raise
FormParserError before oversized data is buffered. Both default
to no limit, and multipart forms are unaffected.
- Nameless uploads no longer crash when spilled to disk.
Uploads without a filename used to raise AttributeError when
UPLOAD_KEEP_EXTENSIONS or UPLOAD_KEEP_FILENAME was set.
- UPLOAD_DIR accepts str with UPLOAD_KEEP_FILENAME. A str
upload directory used to fail on the first write to disk.
- Decoded data is complete before callbacks run (#336, #355).
Content-Transfer-Encoding decoders are finalized at the end of
each part, and quoted-printable escapes split across writes
decode correctly.
- parse_options_header lowercases content types with parameters
. Multipart/Form-Data; boundary=abc now returns
multipart/form-data, the same as headers without parameters.
- Faster opening boundary matching. Complete boundaries are
matched in one step.
- Faster parser state lookups. Parser states are
module-level enum members, so their repr is now multipart.END
instead of <MultipartState.END: 12>.
- Faster querystring callback dispatch.
- Python 3.10 is no longer supported. Python 3.11 is now the
minimum.
- Python 3.15 is supported.
Signed-off-by: Leon Anavi <leon.anavi@konsulko.com>
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Upgrade to release 5.0.4:
- Support pytest 9.2: pytest --codspeed no longer crashes at
session start when pytest-benchmark is installed
- build: pin -std=c11 for instrument-hooks core.c
- docs(readme): bump action version to v5
- feat: warn when running on an unsupported Python version
- ci: migrate to github default arm runners
- ci: use 26.04 image instead of latest
- ci: test against stable python 3.15 and 3.15t
- docs: add Python 3.15 to README version badge
Signed-off-by: Leon Anavi <leon.anavi@konsulko.com>
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Upgrade to release 1.4.1.post1:
Changed:
- Removed the upper version bound on the pytest dependency.
Signed-off-by: Leon Anavi <leon.anavi@konsulko.com>
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Upgrade to release 4.0.1:
- CVE-2026-49272 / GHSA-3jp2-3mrg-2vxc: The device code flow token
endpoint did not validate the device_code parameter. A request
with a known client_id and a bogus device_code could obtain a
bearer token. DeviceCodeGrant.validate_token_request now calls
DeviceCodeGrant.validate_device_code(), which rejects missing or
unknown device codes (invalid_request / invalid_grant) and
delegates the code's authorization status to the new
RequestValidator.validate_device_code() hook, per RFC 8628
Section 3.5.
Signed-off-by: Leon Anavi <leon.anavi@konsulko.com>
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Upgrade to release 1.12.0:
- feat(nodeenv): print shell completion scripts with shtab
- fix(nodeenv): point npm_config_prefix at the env in activate.bat
and Activate.ps1
- fix(nodeenv): install npm from the registry tarball on Windows
- fix(nodeenv): read quoted config values and mirror credentials
- fix: stop putting ^/dev/null on PATH in activate.fish
- fix(nodeenv): retry broken node downloads and stop hanging on
stalls
- fix: do not require python2 to build node from source
- fix(nodeenv): do not put the venv bin on PATH twice with -p
- fix(nodeenv): skip node.js pre-releases for latest and ranges
- fix: preserve fish activation environment
- fix(nodeenv): read a config value raw when interpolation fails
- fix(nodeenv): ignore source trees during prebuilt installs
- fix(nodeenv): escape glob characters in prebuilt paths
- fix(nodeenv): raise on bad config references again
- Make tox configuration more flexible
- chore(tests): port the Makefile test targets to pytest
- Align Python support metadata and runtime with Python 3.10+
- perf(nodeenv): unpack the node.js zip archive in parallel
- docs(site): new landing page for ekalinin.github.io/nodeenv
- docs(site): show GitHub stars and PyPI downloads badges
- docs: update isolated installation instructions
- docs(contributing): add contributing guide and templates
Signed-off-by: Leon Anavi <leon.anavi@konsulko.com>
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Upgrade to release 1.39:
- Test that the code examples in the docs run and match their
output
- Raise JsonPatchConflict when removing from immutable containers
- docs: explain copy and in-place behavior when a patch fails
Signed-off-by: Leon Anavi <leon.anavi@konsulko.com>
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Upgrade to release 9.18.1:
- The main change in this release is that themes are now JSON data:
you can install a theme by putting a file in your IPython
directory, or from a package, without writing any code.
- The release also fixes a number of completion regressions, fixes
WebP handling in ~IPython.display.Image, stops get_tail from
wiping the history database when ipykernel asks for n=None, and
adds ~IPython.core.formatters.DisplayFormatter.register_formatter.
- theme_table can no longer be mutated, and the individual theme
objects are no longer importable from IPython.utils.PyColorize.
Signed-off-by: Leon Anavi <leon.anavi@konsulko.com>
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Upgrade to release 3.5.4:
- Fix possible hang if worker holding the stop flag was killed
while holding it.
- Lots of refactoring/cleanup.
Signed-off-by: Leon Anavi <leon.anavi@konsulko.com>
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Upgrade to release 1.85.1:
- Fixes Denial of service in Health Watch RPC
- Fixes Memory leak in AsyncIO client-streaming calls with
interceptors
Signed-off-by: Leon Anavi <leon.anavi@konsulko.com>
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Upgrade to release 1.85.1:
- Fixes Denial of service in Health Watch RPC
- Fixes Memory leak in AsyncIO client-streaming calls with
interceptors
Signed-off-by: Leon Anavi <leon.anavi@konsulko.com>
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Upgrade to release 5.7.1:
- Update version of python-socketio dependency
- Add Python 3.15 and PyPy 3.12 to CI
Signed-off-by: Leon Anavi <leon.anavi@konsulko.com>
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Upgrade to release 2.22.3:
- Fixed property names with { or } being evaluated as format string
in error messages
- Fixed numeric keywords (minimum, maximum, multipleOf,
exclusiveMinimum, exclusiveMaximum) being applied to booleans
- Fixed propertyNames: false crashing on non-object values
- Fixed the date-time format accepting months, days and hours out
of range (such as 2020-19-39T29:00:00Z) and rejecting a leap
second (23:59:60Z)
- Fixed the date and date-time formats accepting days that do not
exist in their month, such as 2021-02-29 and 2021-04-31
- Fixed the time format accepting hours, minutes, seconds and
offsets out of range (such as 29:99:99)
- Fixed minimum and maximum being exclusive in draft 06 and later
when exclusiveMinimum or exclusiveMaximum is set to a number
other than zero
- Fixed bools in uniqueItems
- Added type annotations for public API and py.typed marker
(PEP 561)
- Added message to JsonSchemaValuesException joining messages of
all collected errors
Signed-off-by: Leon Anavi <leon.anavi@konsulko.com>
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Upgrade to release 4.4.0:
- Fix history ranges starting past the end
- Fix History.truncate() leaving session start index stale
- Flush window title updates without waiting for a prompt
- Fix two races behind the flaky pipeline job-control test
- Add security policy naming the Security Maintainer
- Preserve history session boundaries across list mutations
- Track history session membership by entry instead of index
- Preserve live history while persisting a bounded copy
- Address review feedback on persisted history copy
Signed-off-by: Leon Anavi <leon.anavi@konsulko.com>
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Upgrade to release 6.2.0.post1:
- Build Pyodide (pyemscripten) wheels for Python 3.14
From 6.2.0:
- Added the strict_hashable_types parameter to CBORDecoder, load()
and loads() which, when enabled, restricts map keys and set
elements to text strings and byte strings in order to prevent
denial-of-service attacks via hash collisions
(GHSA-p6xp-9h6j-r278)
- Documented decoding concatenated CBOR objects with EOF checks
between objects
- Fixed CBOREncoder.encode_to_bytes() sharing the string reference
namespace of the enclosing encode operation when called from a
default hook
- Fixed the encoder writing the repr() of a bytes regular
expression pattern (tag 35) instead of the pattern itself, so the
regular expression decoded back as the text pattern b'...' and no
longer matched the same input
- Fixed shared references (tag 29) allowing the decoder to
construct object graphs nested deeper than max_depth, which could
crash the interpreter with a segmentation fault when such an
object was hashed (as a map key or set member)
(GHSA-55vx-2phh-7grr)
- Fixed mission conversion of date values (tags 100 and 1004) to
JSON as ISO formatted strings
- Fixed a PanicException on Python 3.10 when decoding a string-form
datetime (tag 0) whose fractional seconds part contains a
multi-byte Unicode numeric character, as the microsecond padding
treated those as digits and then truncated the part on a byte
index inside one of them
- Fixed the decoder accepting a set (tag 258) whose payload is not
an array, folding a text string, byte string or map into a set of
its characters, bytes or keys instead of rejecting it
- Fixed the decoder accepting a non-integer exponent or mantissa in
a decimal fraction (tag 4) or a bigfloat (tag 5)
Signed-off-by: Leon Anavi <leon.anavi@konsulko.com>
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Upgrade to release 7.2.8:
- Fix type stubs for TTLCache and TLRUCache with default timer.
- Minor documentation and unit test improvements.
Signed-off-by: Leon Anavi <leon.anavi@konsulko.com>
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Upgrade to release 4.16.0:
- Added BlockingPortal.call_async() for asynchronously calling into
an event loop with optional abandonment on cancellation
- Added SOCK_SEQPACKET support to connect_unix() and
create_unix_listener()
- Changed UDP sockets on the asyncio backend to make send() wait
until the datagram has been passed to the operating system
- Fixed documentation implying that use_uvloop=True falls back to
the standard event loop when uvloop (winloop on Windows) is not
installed, and showed how to skip the corresponding pytest
backend variant when the dependency is missing
- Fixed get_current_task() on the asyncio backend raising an opaque
TypeError: cannot create weak reference to 'NoneType' object
instead of a clear error when called from a callback scheduled
via loop.call_soon_threadsafe() (e.g. from within
from_thread.run_sync()), where asyncio.current_task() is
legitimately None
- Fixed CancelScope on the asyncio backend swallowing a native
Task.cancel() that was delivered in the same event loop iteration
as the scope's own cancellation (such as a move_on_after()
deadline expiring), which left the host task with a pending
cancellation request that never raised CancelledError (\#1214)
- Fixed UDP socket closing hanging if a datagram send was still in
flight
- Fixed concurrent aclose_forcefully() calls on asyncio socket
streams returning before the underlying socket was closed
- Fixed aclose() on sockets not checkpointing, and thus failing to
raise a cancellation exception when called inside a cancelled
scope (affected UNIX socket streams and UNIX datagram sockets on
asyncio, and all sockets on Trio)
- Fixed aclose() on asyncio TCP socket listeners not checkpointing,
and leaking the listening socket if cancelled while another task
was waiting in accept()
- Fixed accept() on asyncio UNIX socket listeners hanging forever,
and leaving the listening socket registered with the event loop,
when the listener was closed by another task
- Fixed send() on asyncio TCP and UDP sockets returning normally,
despite the data being discarded, when the socket was closed by
another task (or, for TCP, the connection was lost) while send()
was waiting for the operating system to accept the data
- Fixed SocketStream.send() on the asyncio backend handing its data
to a paused transport after a previous send() was cancelled
- Fixed sleep() treating a negative delay (including -inf)
inconsistently across backends (asyncio returned immediately;
Trio raised ValueError) by raising ValueError on all backends
- Fixed create_tcp_listener() occasionally failing with a message
like Could not create 2 listeners with a consistent port when an
ephemeral port is requested and IPv6 is enabled and the
dual-stack path is not available or a specific local host name
was given
- Fixed AsyncFile not shielding against cancellation while closing
- Fixed TemporaryDirectory.cleanup() reacting to cancellation
prematurely when the host task was already cancelled, causing the
actual cleanup operation to be skipped
- Fixed a NaN deadline being accepted on asyncio while Trio raised
ValueError, which silently disabled the timeout
(fail_after(math.nan) returned instead of raising TimeoutError);
CancelScope now rejects it on both backends
- Fixed FileReadStream and FileWriteStream leaving files open when
cancelled during context manager exit, including unflushed writes
- Fixed Condition.notify() and Condition.notify_all() failing when
the underlying lock was acquired directly rather than through the
condition
- Fixed importing AnyIO with loaders that do not provide __file__,
such as PyOxidizer, raising TypeError instead of falling back to
eager imports
- Fixed text receive streams silently discarding incomplete
characters at end-of-file instead of applying the configured
decoding error handler
- Fixed Semaphore(fast_acquire=True) losing its fast-acquire
behavior when instantiated outside a running event loop
- Fixed BlockingPortalProvider permanently failing with an
AttributeError on every subsequent use if the first attempt to
start its blocking portal failed
- Fixed create_udp_socket() silently ignoring the local_port
argument when local_host was omitted, making the socket bind to a
random ephemeral port instead of the requested one
- Fixed worker threads on the asyncio backend never returning to
the idle pool (and thus never being reused or pruned) when a
to_thread.run_sync() call was cancelled after it had been queued
for the worker but before the worker thread had picked it up
- Fixed pytest plugin running Hypothesis async tests only on the
first parametrized backend instead of each backend
- Fixed TypeError when subclassing an existing TypedAttributeSet
subclass, and this subclass contains type annotations
- Fixed Lock.locked() (and Condition.locked()) raising
NoEventLoopError when the lock was created outside a running
event loop
- Fixed connect_tcp() leaking a connected socket when an enclosing
cancel scope was cancelled before the connection attempt task
group exited
- Fixed create_connected_udp_socket() silently ignoring local_port
when local_host was not given: the local address was left unset,
so the OS picked an ephemeral port instead of the requested one
- Fixed BlockingPortalProvider not passing the exception raised in
its with block on to the start_blocking_portal() context manager
it wraps
- Fixed anyio.itertools.islice() not consuming the first start
items when stop <= start
- Fixed BufferedByteReceiveStream.receive_exactly() and
receive_until() returning buffered data after explicit closure
instead of raising ClosedResourceError
- Fixed a TLS stream sending records the peer could no longer
decrypt after a cancelled send() lost an already-encrypted
record; the stream now raises BrokenResourceError on further use
instead
- Fixed close() on memory object streams not waking up tasks
blocked in send() or receive() on the same stream
- Fixed anyio.itertools.product() acquiring and consuming unused
input iterables when repeat=0
- Fixed SpooledTemporaryFile rolling over to disk on writes when
max_size is zero (the default)
- Fixed Trio UDP sockets remaining open when initialization failed
or was cancelled
- Fixed SpooledTemporaryFile.rollover() moving the current file
position to the end of the file
- Fixed cancellation during temporary directory creation leaving
the directory behind while its context manager is retained
- Fixed cancellation during SpooledTemporaryFile.rollover() losing
buffered contents
- Fixed TextSendStream.send() re-emitting the byte order mark on
every item with BOM-emitting encodings (utf-16, utf-32,
utf-8-sig), which made the peer decode a stray \ufeff in front of
every message after the first
- Fixed buffered byte streams returning empty bytes when a wrapped
object stream received empty messages instead of waiting for data
or raising EndOfStream
- Fixed BlockingPortal crashing with InvalidStateError when a
future it returned was cancelled just as the callable was
completing
- Fixed UNIXSocketStream.send_fds() silently truncating messages on
short socket writes
- Fixed cancelled TCP receives on the asyncio backend leaving
transport reads enabled, allowing incoming data to accumulate
while the application was not receiving.
- Fixed CapacityLimiter.acquire_on_behalf_of() on asyncio releasing
the wrong borrower when the acquiring task was cancelled during
the cancel-shielded checkpoint that runs after a non-blocking
acquisition
Signed-off-by: Leon Anavi <leon.anavi@konsulko.com>
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
For native recipes, the prefix variable is not '/usr', resulting
in iperf3-native searching for wrong location for ssl.h.
Fix to use the correct location to find ssl.h.
With this fix, iperf3, iperf3-native, nativesdk-iperf3 all succeed
to build.
Signed-off-by: Chen Qi <chenqi_hycx@cetc.com.cn>
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
python3-tomlkit was added to oe-core (46d3b26a63 "python3-tomlkit: add
new recipe (from meta-oe)") and is at 0.15.2 there, so the meta-python
0.15.1 copy is not used. Drop it, and drop it from the meta-python ptest
list as oe-core's ptest-packagelists.inc already lists it.
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Cc: Ross Burton <ross.burton@arm.com>
libfaketime was imported into oe-core (f3a4faaf4c "libfaketime: Import
recipe from meta-oe") and has had fixes there since: a 32-bit build fix,
the clang 23 fix, -O1 for the clobbered-variable error and the man page
install. Both copies are 0.9.13 at the same layer priority, so bitbake
picks one by parse order and currently ends up with the outdated meta-oe
copy.
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Cc: Ross Burton <ross.burton@arm.com>
Add leancrypto-module to build and package leancrypto.ko
independently of the userspace library.
Move common source and license metadata into leancrypto.inc and include the
submitted upstream patch to generate headers using the target kernel
architecture instead of the build host.
Signed-off-by: Ayoub Zaki <ayoub.zaki@embetrix.com>
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Upgrade to release 1.0.0:
- Configurable layout spaces through BoxStyle.gfx["SPACE"], with an
ordinary-space fallback for existing palettes.
- Declared the existing public API stable, without breaking changes
from 0.4.0.
Fixes:
WARNING: python3-asciitree-1.0.0-r0 do_check_backend: QA Issue:
inherits setuptools3 but has pyproject.toml specifying backend
setuptools.build_meta, use the correct class [pep517-backend]
License-Update: Update copyright years to 2015-2026
Signed-off-by: Leon Anavi <leon.anavi@konsulko.com>
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
"make install" depends on a doc target which runs "make -C docs man html".
With -j both goals start their own sphinx-build sharing _build/doctrees,
which sphinx does not write atomically, so one can read a doctree the
other is rewriting. do_install then fails at random:
AttributeError: 'dict' object has no attribute 'append'
make: *** [Makefile:138: doc] Error 2
Add .NOTPARALLEL to docs/Makefile so that the goals run one after the
other. Running "make -j64 man html" on an oversubscribed host failed 71
out of 576 times without this change and never with it.
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
AI-Generated: Uses Claude Code
lcov 2.6 makes "make install" depend on a new xs_lib target which runs
"perl Makefile.PL" with the host perl. That loads lcovutil.pm, so
do_install fails on hosts without Capture::Tiny:
Can't locate Capture/Tiny.pm in @INC
make: *** [Makefile:169: xs_lib] Error 2
The extension can't be cross-built this way, since it would use the host
perl configuration. It is an optional speedup and lcovutil.pm falls back
to pure Perl without it, so set LCOV_NO_XS=1.
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Changes from v2.5 release:
- new html2lcov tool
* Capture source diff and 'baseline' coverage from lcov HTML report - suitable
as input to generate differential coverage report versus 'now'. See the html2lcov documentation for details.
- new jacoco2lcov tool
* Translate JaCoCo java coverage data to lcov format - e.g., for multi-language
coverage reporting or to take advantage of lcov features. See the jacoco2lcov documentation for details.
- new XS back end (on plaforms which support it).
* C++ implementation of some key types and operations - to improve execution
time and memory footprint compared to pure Perl.
- various bug fixes and test coverage improvements.
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Brings following changes
3c04036 Fix: make static ECDH selectable with the br_ssl_client_set_single_ec() helper function.
3d9be2f Added dependencies to better support parallel or shuffled compilation.
3479195 Fixed chunked decoding in case of errors (if decoding failed at some point, subsequent chunks should be ignored, trying to reenter the decoder after a failure is a recipe for Bad Thing). Impacted functions were not used over malicious on-the-wire data for "normal" SSL/TLS usage.
8f795e5 Fixed Makefile dependencies.
7bea48e Fixed bug in handling incoming records with invalid length (impacted CBC encryption with 3DES or with the aes_small or aes_big AES implementations; only 3DES was selectable by default).
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
.circleci and scripts/ci are upstream's tooling for its AWS Ubuntu test
runners. They hardcode /home/ubuntu, which trips the buildpaths QA check
on build hosts whose HOME has the same path:
File .../influxdb/.circleci/config.yml in package influxdb-dev contains
a reference to the build host HOME directory.
Remove them before do_install instead of setting OEQA_BUILDPATHS_SKIP,
as they are of no use in influxdb-dev and a skip would also hide real
leaks. run_perftest.sh goes with them, so drop the sed fixing its shebang.
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
libaccountsservice uses sd-login for its session tracking and links
against libsystemd or libelogind unconditionally. There is no elogind
recipe to satisfy the alternative, so on sysvinit distros configure fails:
meson.build:206:2: ERROR: Assert failed: systemd required but not found
Declare the requirement so that the recipe is skipped there.
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
The systemdsystemunitdir option defaults to auto, which makes meson look
up systemd.pc. That only worked because systemd ends up in the sysroot
through other dependencies, and it fails on sysvinit distros:
meson.build:36:18: ERROR: Dependency "systemd" not found
Pass the unit directory explicitly. The unit is installed unconditionally
and systemd.bbclass removes it again on distros without systemd.
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
malcontent and malcontent-ui depend on accountsservice, which needs
systemd now, so they can not be built on sysvinit distros. Declare it so
the skip message names the actual requirement.
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
fwupd-efi finds the linker with find_program(cc.get_linker_id()),
which looks for a program named ld.bfd. meson.cross only sets ld, so
meson falls back to PATH and picks up ld.bfd from HOSTTOOLS. That breaks
cross builds on an x86-64 host:
hosttools/ld.bfd: cannot represent machine `aarch64'
and on x86-64 targets it silently links with the host linker.
Set ld.bfd in an additional cross file
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Universe builds every recipe, so the unconditional PROVIDES puts this
recipe and polkit in the same build, and bitbake -c fetch universe
reports that multiple .bb files provide polkit.
Claim the name only when PREFERRED_PROVIDER_polkit selects smolkit. With
the overlap gone, smolkit is no longer dropped from the world targets.
AI-Generated: Uses Claude Code
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Universe builds every recipe and ignores EXCLUDE_FROM_WORLD, so the
unconditional PROVIDES puts this recipe and GNU coreutils in the same
build, and bitbake -c fetch universe reports that multiple .bb files
provide coreutils, coreutils-native and nativesdk-coreutils.
Claim the name only when this recipe is the preferred provider, checked
separately for the target, native and nativesdk variants so they can
still be selected independently.
AI-Generated: Uses Claude Code
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
The GPL-2.0 entry dates back to the first recipe (AOSP android-5.1.1_r37,
2016) and was carried over unchanged to the Debian android-platform-tools
based 35.0.2 recipe.
In the android-5.1.1_r37 repos the recipe fetched, the only GPLv2 mentions
are headers which are dual licensed "Apache 2.0 OR GPLv2"
(system/core include/cutils/ashmem.h, include/utils/ashmem.h,
include/log/logger.h), kernel headers in libion which are not built, and
documentation templates. Apache-2.0 can be chosen for all of them.
The 35.0.2 sources come from Debian's android-platform-tools, whose
debian/copyright lists no GPL license: the files are Apache-2.0, BSD,
Expat, Khronos, PSF and public-domain, and a scan of the unpacked tarball
finds no GPL licensed source files.
Remove GPL-2.0-only from LICENSE and LIC_FILES_CHKSUM so that the recipe is
not wrongly treated as GPL by license based tooling (e.g. copyleft source
archiving).
Signed-off-by: Etienne Cordonnier <ecordonnier@snap.com>
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Release notes: https://github.com/fastapi/fastapi/releases
* opentelemetry-api is now a runtime dependency
* Add opentelemetry-sdk and git to ptest dependencies, used by the new
telemetry and translation script tests
* Ignore telemetry tests that need logfire, the opentelemetry OTLP
exporter and the opentelemetry fastapi instrumentation, which are not
packaged in meta-python
* Fixes the ptest hang in
test_frontend_dependency_restores_existing_dependency_stacks with
starlette 1.10, fixed upstream in 0.142.0
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
opentelemetry-sdk requires opentelemetry-api, opentelemetry-semantic-conventions
and typing-extensions, without them importing opentelemetry.sdk fails with:
ModuleNotFoundError: No module named 'opentelemetry.semconv'
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
test_runtime_dir_unreachable_xdg_runtime_dir_warns hardcodes the glibc
strerror(ELOOP) text "Too many levels of symbolic links", whereas musl
reports "Symbolic link loop". Backport the upstream fix which uses
os.strerror() instead.
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
pyroute2/ndb/task_manager.py imports the resource module and it is
pulled in when importing pyroute2, so every ptest module failed to
import with:
ModuleNotFoundError: No module named 'resource'
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
setup.py imports tuna.version from the source tree. The setuptools
build_meta backend does not put the source directory on sys.path, so
do_compile fails:
File "<string>", line 9, in <module>
ModuleNotFoundError: No module named 'tuna'
ERROR Backend subprocess exited when trying to invoke get_requires_for_build_wheel
It only worked so far because python3targetconfig added an empty entry,
i.e. the current directory, to PYTHONPATH. The same happens with a plain
"python3 -m build" outside of OE.
pyproject.toml already gets the version from tuna.version via
[tool.setuptools.dynamic], so drop the import from setup.py.
AI-Generated: Uses Claude Code
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
pyproject.toml specifies setuptools.build_meta as the build backend.
Inheriting setuptools3 with INSANE_SKIP for pep517-backend was a
workaround for the top level build.py shadowing pyproject-build's own
"build" module, which is fixed in oe-core's python3targetconfig class
("python3targetconfig: Do not add the current directory to PYTHONPATH").
pyproject.toml pins sip == 6.12.0, which is only used to regenerate the
wrapper sources that the sdist already ships, so skip the build
dependency check.
AI-Generated: Uses Claude Code
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Upgrade to release 0.6.1:
- Update CI dependencies
- Drop support of Python 3.10, support Python 3.15
- Replace deprecated license classifier
- Use HTTP server for Doconfly
Fixes:
WARNING: python3-webencodings-0.6.1-r0 do_check_backend: QA Issue:
inherits setuptools3 but has pyproject.toml specifying backend
flit_core.buildapi, use the correct class [pep517-backend]
License-Update: Use LICENSE file
Signed-off-by: Leon Anavi <leon.anavi@konsulko.com>
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Upgrade to release 1.1.0:
- Cancel queued requests on close
- Remove the unused build attribute and the Python 2.7 NullHandler
fallback
- Close owned sessions after their background requests finish,
including when using a supplied executor, and reject requests
after close
- Pickle check covers request arguments, not just the callable
- background_callback now emits a DeprecationWarning instead of a
log line
- Fix pool sizing for supplied sessions
- Preserve falsy callback results
Fixes:
WARNING: python3-requests-futures-1.1.0-r0 do_check_backend:
QA Issue: inherits setuptools3 but has pyproject.toml specifying
backend setuptools.build_meta, use the correct class [pep517-backend]
Signed-off-by: Leon Anavi <leon.anavi@konsulko.com>
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Upgrade to release 18.0.0:
- Upgrade to Unicode 18.0.0.
- Drop support for Python 3.8 (requires Python 3.9+).
- Refactor architecture and optimize table loading for better
performance.
- Modernize packaging (pyproject.toml) and adopt Nox for testing.
Fixes:
WARNING: python3-pyunormalize-18.0.0-r0 do_check_backend: QA Issue:
inherits setuptools3 but has pyproject.toml specifying backend
setuptools.build_meta, use the correct class [pep517-backend]
License-Update: Copyright years updated
Signed-off-by: Leon Anavi <leon.anavi@konsulko.com>
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Leon Anavi
Chen Qi
Khem Raj
Kris Gavvala
Ayoub Zaki
Jason Schonberg
Martin Siegumfeldt
Etienne Cordonnier
Khem Raj