=== release 1.28.4 ===

2026-06-12 13:19:43 +0100  Tim-Philipp Müller <tim@centricular.com>

	* gstreamer.doap:
	* meson.build:
	  Release 1.28.4

2026-05-12 16:08:30 +0200  Carlos Falgueras García <cfalgueras@fluendo.com>

	* tests/check/gst/gstmemory.c:
	  tests: fix support with glib <= 2.67.2
	  These tests use "g_memdup2()" but they don't include
	  "gst/glib-compat-private.h", which makes the compilation fail when linking
	  against Glib <= 2.67.2 as "g_memdup2()" was included at 2.67.3
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/11618>

2026-06-03 12:00:58 +0200  Stéphane Cerveau <scerveau@igalia.com>

	* gst/gstvalue.c:
	  gstvalue: free existing value in list append
	  In gst_value_unique_list_append_and_take_value, the append_value
	  is transfer-full, so if this value is already existing, it should
	  be unset inside the method.
	  Tensor elements append one caps per model output into a unique list
	  (transfer-full). When two outputs share the same caps signature (e.g.
	  YOLOv8-seg), the list deduplicates the second — but the drop path leaked
	  that caps instead of releasing it.
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/11756>

2026-06-02 18:55:35 +0200  Stéphane Cerveau <scerveau@igalia.com>

	* gst/gstcaps.c:
	* tests/check/gst/gstcaps.c:
	  caps: flag nested caps in static caps as may-be-leaked
	  As static caps are usually leaked, then the nested caps
	  are also leaked and should be set GST_MINI_OBJECT_FLAG_MAY_BE_LEAKED
	  as well.
	  The leak has been spotted with leak tracer and the onnx elements which
	  use for example:
	  static GstStaticCaps scaps = GST_STATIC_CAPS (
	  "video/x-raw,"
	  "tensors=(structure)["
	  "tensorgroups,"
	  "out=(/uniquelist){"
	  "(GstCaps)["
	  "tensor/strided,"
	  "tensor-id=(string)out,"
	  "type=(string)float32"
	  "]"
	  "}"
	  "]");
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/11756>

2026-05-28 11:07:24 +0200  Stéphane Cerveau <scerveau@igalia.com>

	* gst/gstvalue.c:
	* tests/check/gst/gstvalue.c:
	  gstvalue: fix crash when converting NULL G_TYPE_VALUE_ARRAY to G_TYPE_STRING
	  In the case of Rust binding, converting to G_TYPE_STRING
	  a G_TYPE_VALUE_ARRAY might lead to a crash if the GValue contains NULL.
	  Consider it as an empty string.
	  Reproduce with vp8enc:temporal-scalability-target-bitrate and
	  GstPipelineStudio.
	  Add defensive check in gst_value_transform_any_list_string
	  Add additional tests for array, list and set.
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/11737>

2026-05-20 11:21:45 +0800  Qian Hu (胡骞) <qian.hu@mediatek.com>

	* gst/gstbufferpool.c:
	  bufferpool: avoid leaking partially preallocated buffers
	  Preallocate buffers into a temporary array in default_start() before handing
	  them to release_buffer(). If allocation fails midway, free the buffers that
	  are still owned by the base class instead of leaving already released buffers
	  behind.
	  This avoids leaking buffers and corrupting cur_buffers when preallocation
	  fails, while still allowing subclasses to take ownership in release_buffer()
	  only after all allocations succeeded.
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/11664>

2026-05-15 01:46:30 +0530  Nirbheek Chauhan <nirbheek@centricular.com>

	* gst/gsttask.c:
	  gsttask: Fix racy tests by making unref deterministic
	  gst_check_teardown_element() asserts that the refcount of the element
	  is 1 after its state has been set to NULL which stops all pad tasks.
	  However, after 1179387b9f, the pad task also holds a reference to the
	  element and the finalize hits after the task is stopped and releases
	  the task lock (sync) and before the task thread exits (async).
	  This made the assert flaky, because it would see 1 or 2 depending on
	  how slowly things are scheduled.
	  Co-authored-by: Sebastian Dröge <sebastian@centricular.com>
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/11638>

2026-05-11 16:33:44 +0530  Nirbheek Chauhan <nirbheek@centricular.com>

	* gst/gstinfo.c:
	  gstinfo: Don't use fwrite() on Windows for debug logging
	  When setting GST_DEBUG_FILE, the file was opened with fopen() and
	  written to with fwrite(). This required an expensive fflush() on
	  Windows because fwrite() doesn't support line-buffered I/O.
	  Bypass the CRT userspace buffering by submitting whole lines in
	  a single write to the kernel. This is equivalent to line-buffered FILE
	  writes, and is about 12% faster.
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/11632>

2026-05-11 12:49:55 +0530  Nirbheek Chauhan <nirbheek@centricular.com>

	* gst/gstinfo.c:
	  gstinfo: Use stack allocated memory when printing on Win32
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/11624>

2026-05-10 03:45:43 +0530  Nirbheek Chauhan <nirbheek@centricular.com>

	* gst/printf/README:
	* gst/printf/gst-printf.h:
	* gst/printf/vasnprintf.c:
	  gstinfo: Don't ever use %n to know how much was written
	  The return value of snprintf will let us know that already. %n is only
	  needed for ancient snprintf implementation that didn't do that. We do
	  not support any of those.
	  This allows us to use snprintf instead of sprintf which is marginally
	  faster.
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/11624>

2026-05-10 01:58:42 +0530  Nirbheek Chauhan <nirbheek@centricular.com>

	* gst/gstinfo.c:
	* gst/printf/printf.c:
	* gst/printf/printf.h:
	  gstinfo: Use stack allocation for <=1KB messages
	  It should be super low-risk to add an extra KB to the stack at all
	  call-sites.
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/11624>

2026-05-12 22:46:00 +0300  Sebastian Dröge <sebastian@centricular.com>

	* gst/gstdatetime.c:
	  datetime: Improve correctness of ISO-8601 string parsing
	  - Use unsigned integers of the correct sizes everywhere for parsing
	  to properly detect errors.
	  - Validate values first before using them.
	  - Actually validate timezone information and don't assert on invalid
	  timezone information.
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/11616>

2026-05-09 14:26:55 +0000  L. E. Segovia <amy@centricular.com>

	* gst/gstregistry.c:
	  registry: detect libgstreamer load from Android container and skip canonicalization
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/11612>

2026-05-11 18:33:25 +0100  Tim-Philipp Müller <tim@centricular.com>

	* meson.build:
	  Back to development after 1.28.3

=== release 1.28.3 ===

2026-05-11 18:28:12 +0100  Tim-Philipp Müller <tim@centricular.com>

	* gstreamer.doap:
	* meson.build:
	  Release 1.28.3

2026-04-15 16:31:02 +0300  Sebastian Dröge <sebastian@centricular.com>

	* gst/gstvalue.c:
	  value: On buffer deserialization errors first unmap the buffer and then unref it
	  Also unref the buffer if mapping fails (which is technically impossible).
	  Fixes https://gitlab.freedesktop.org/gstreamer/gstreamer/-/work_items/5042
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/11590>

2026-05-08 10:48:38 +0300  Sebastian Dröge <sebastian@centricular.com>

	* gst/gstallocator.c:
	  allocator: Use g_try_malloc() instead of g_malloc() for sysmem
	  The latter will just abort the process if allocation fails but all the
	  surrounding code is already expecting it to return NULL on failure and this
	  change now allows to recover from impossible allocations.
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/11569>

2026-05-07 16:47:47 +0100  Arthur Chan <arthur.chan@adalogics.com>

	* libs/gst/base/gstbaseparse.c:
	  baseparse: Fix memory leak when subclass returns error
	  Signed-off-by: Arthur Chan <arthur.chan@adalogics.com>
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/11567>

2026-05-07 08:38:01 +0300  Sebastian Dröge <sebastian@centricular.com>

	* gst/gststreams.c:
	  streams: Add METADATA to the valid stream flags for serialization
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/11566>

2026-05-07 08:35:47 +0300  Sebastian Dröge <sebastian@centricular.com>

	* gst/gststreams.c:
	  streams: Don't assert on empty stream flags during serialization
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/11566>

2026-04-07 13:03:19 +0100  Nirbheek Chauhan <nirbheek@centricular.com>

	* gst/gstmacros.h:
	  gstmacros: Remove restrict definition
	  We now require recent GCC, Clang, and MSVC for building, all of which
	  provide the `restrict` keyword. Our define is now breaking external
	  code. F.ex., Qt 6.11 on MSVC uses `__declspec(restrict)` which gets
	  converted to `__declspec(__restrict)` which is not a valid keyword.
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/11563>

2026-04-07 13:09:24 +0100  Nirbheek Chauhan <nirbheek@centricular.com>

	* meson.build:
	  meson: Require C std gnu11 or c11
	  When using gcc or clang, gnu11 will be used, and when using MSVC c11
	  will be used which will pass /std:c11. This provides us with the
	  `restrict` keyword on all supported platforms.
	  We can do this now because we require Visual Studio 2019.
	  This feature was added in Meson 1.3
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/11563>

2026-05-05 10:18:23 -0400  Xavier Claessens <xclaessens@netflix.com>

	* gst/gstdevicemonitor.c:
	* tests/check/gst/gstdevice.c:
	  devicemonitor: Wait for start thread to finish when listing devices
	  This fixes a regression introduced by:
	  https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/9698
	  Since 1.28.0, gst_device_monitor_start() is not blocking any more, which
	  causes the list of devices returned by gst_device_monitor_get_devices()
	  to be incomplete. For example v4l2 provider would not fill udev
	  properties yet.
	  Solve this by blocking in gst_device_monitor_get_devices() for backward
	  compatibility. That way, applications that wish to not block can wait
	  for %GST_MESSAGE_DEVICE_MONITOR_STARTED before calling get_devices().
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/11554>

2026-05-03 11:48:44 +0300  Sebastian Dröge <sebastian@centricular.com>

	* libs/gst/base/gstaggregator.c:
	  aggregator: Fix documentation
	  Parameter docs for various functions were missing.
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/11488>

2026-04-27 20:12:37 +0300  Sebastian Dröge <sebastian@centricular.com>

	* libs/gst/base/gstbitwriter.c:
	  bitwriter: Zero-initialize uninitialized data
	  Otherwise we will read uninitialized data when combining it with newly
	  written bits.
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/11471>

2026-04-27 18:18:22 +0300  Sebastian Dröge <sebastian@centricular.com>

	* libs/gst/base/gstbitwriter.h:
	* tests/check/libs/bitwriter.c:
	  bitwriter: Allow unsetting set bits when overwriting them
	  Previously the original bit pattern would always be preserved and new bits
	  only OR-ed into the array. This made it impossible to actually unset bits.
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/11471>

2026-04-28 14:52:16 +0100  Tim-Philipp Müller <tim@centricular.com>

	* tools/gst-inspect.c:
	  gst-inspect-1.0: type for string caps fields should be 'string' not 'gchararray'
	  See e.g. gst-inspect-1.0 vorbisenc
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/11460>

2026-04-26 18:18:35 +0530  Nirbheek Chauhan <nirbheek@centricular.com>

	* docs/meson.build:
	  meson: Docs build is not supported when gstcheck is disabled
	  subprojects/gstreamer/docs/meson.build:147:13: ERROR: Unknown variable "gst_check_gir".
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/11444>

2026-04-24 17:29:42 +0530  Nirbheek Chauhan <nirbheek@centricular.com>

	* libs/gst/check/gstcheck.c:
	  gstcheck: Update 'Since' marker to 1.28.3
	  We backported this new symbol for CI use, but it's a good idea to
	  update the Since marker
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/11387>

2026-03-31 11:58:28 +0200  Piotr Brzeziński <piotr@centricular.com>

	* libs/gst/check/gstcheck.c:
	* libs/gst/check/gstcheck.h:
	  check: Add API to run a test suite with forking disabled
	  We already needed to set CK_FORK=no in some tests (e.g. for NVIDIA elements). Turns out Vulkan tests on macOS also need
	  that, so this adds a helper API to use instead of having to override the envvar.
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/11387>

2026-02-06 21:42:25 +0000  Thibault Saunier <tsaunier@igalia.com>

	* gst/gstpad.c:
	  gstpad: fix potential buffer leak in get_range_failed error handler
	  In gst_pad_get_range_unchecked(), the get_range_failed error handler
	  did not clean up res_buf when the getrange function returned an error.
	  If a getrange implementation allocates a buffer but then returns a
	  non-OK flow, the buffer would be leaked since it was never written
	  back to the caller's buffer pointer.
	  Add defensive cleanup matching the pattern already used in the
	  probe_stopped_unref handler in the same function.
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/11284>

2026-02-06 17:35:26 +0000  Thibault Saunier <tsaunier@igalia.com>

	* plugins/elements/gsttypefindelement.c:
	  typefind: fix buffer leak on pull_range failure during shutdown
	  When gst_pad_pull_range() fails during a racy pipeline teardown, the
	  already-allocated buffer may still be set in the output parameter even
	  though the return value is not GST_FLOW_OK. The code jumps to the
	  pause label without clearing the buffer, leaking it.
	  Use gst_clear_buffer() before the goto to properly release the buffer
	  on all error paths.
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/11284>

2026-04-07 20:07:19 +0100  Tim-Philipp Müller <tim@centricular.com>

	* meson.build:
	  Back to development after 1.28.2

=== release 1.28.2 ===

2026-04-07 20:02:23 +0100  Tim-Philipp Müller <tim@centricular.com>

	* gstreamer.doap:
	* meson.build:
	  Release 1.28.2

2026-03-31 14:39:52 +0900  Seungha Yang <seungha@centricular.com>

	* plugins/elements/gstqueue.c:
	  queue: Fix potential use-after-free in log function
	  Avoid accessing already pushed event object
	  Regression introduced by
	  https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/10068
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/11192>

2026-03-18 23:42:43 +0100  Tim-Philipp Müller <tim@centricular.com>

	* po/LINGUAS:
	* po/kk.po:
	* po/pt_BR.po:
	* po/sv.po:
	* po/zh_CN.po:
	  gstreamer: update translations
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/11077>

2026-03-10 23:05:17 +0530  Nirbheek Chauhan <nirbheek@centricular.com>

	* libs/gst/base/gstbaseparse.c:
	  baseparse: Fix out_buffer leak in frame_free and missing ref in frame_copy
	  gst_base_parse_frame_free() only unrefs frame->buffer but not
	  frame->out_buffer. If finish_frame() exits early (e.g. due to a
	  g_return_val_if_fail check on adapter size, or via the scanning path),
	  out_buffer is left non-NULL and leaked when frame_free() zeroes the
	  frame struct via memset without unreffing it.
	  This can be triggered via h264parse (and other video parsers) in the
	  packetized path: gst_h264_parse_parse_frame() takes a buffer from the
	  frame_out adapter and sets frame->out_buffer. If finish_frame() then
	  fails its precondition checks, that buffer is never freed.
	  Also fix gst_base_parse_frame_copy() which uses g_memdup2() to copy
	  the frame struct, giving the copy the same out_buffer pointer as the
	  original, but only adds a ref for buffer and not for out_buffer.
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/11056>

2026-03-07 11:47:29 -0500  Xavier Claessens <xclaessens@netflix.com>

	* gst/gstbin.c:
	  bin: iterator is not nullable
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/11036>

2026-01-29 17:02:49 +0900  Seungha Yang <seungha@centricular.com>

	* libs/gst/base/gstbaseparse.c:
	  baseparse: Preserve upstream buffer duration if possible
	  When upstream buffer timestamps are reliable and subclass has
	  timestamp interpolation disabled, preserve the upstream buffer
	  duration instead of replacing it with GST_CLOCK_TIME_NONE.
	  Subclasses typically derive buffer duration from stream header
	  information, but this can be inaccurate for variable framerate
	  streams or when the signaled framerate itself is unreliable. By
	  forwarding the original buffer duration, this change allows
	  subclasses to handle timing more accurately in such cases.
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/11016>

2026-03-09 00:31:25 +0300  Vitaly Vlasov <vnigtha@gmail.com>

	* plugins/elements/gstfilesink.c:
	  filesink: Fix wrong open() in overwrite mode
	  Fix gst_fopen() when File-sink-file-mode=overwrite is selected on non Windows host
	  Fixes  #4946
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/10990>

2026-03-03 18:24:41 -0300  L. E. Segovia <amy@centricular.com>

	* gst/gstregistry.c:
	  registry: Skip recursion into and from .dSYM bundles
	  The filtering of dSYM bundles was done *after* an early return that approved
	  non-dot directories.
	  For good measure, let's also add a filtering entry to ban directories when their
	  parent is a dSYM bundle.
	  See https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/10850#note_3357038
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/10970>

2026-01-29 12:34:46 +0100  Fabian Orccon <forccon@fluendo.com>

	* gst/gstsystemclock.c:
	* tests/validate/gst-tester.c:
	  all: GThreadFunc return type fixes
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/10955>

2026-02-26 01:53:07 +0000  Tim-Philipp Müller <tim@centricular.com>

	* meson.build:
	  Back to development after 1.28.1

=== release 1.28.1 ===

2026-02-26 01:44:06 +0000  Tim-Philipp Müller <tim@centricular.com>

	* gstreamer.doap:
	* meson.build:
	  Release 1.28.1

2026-02-23 19:42:18 +0900  Seungha Yang <seungha@centricular.com>

	* gst/gstutils.c:
	  gst: Add explanatory comment to call_async implementation
	  Clarify why notify is conditionally called for deprecated
	  gst_element_call_async()
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/10872>

2026-02-20 09:28:15 +0200  Sebastian Dröge <sebastian@centricular.com>

	* tools/gst-stats.c:
	  gst-stats: Also allow ANSI colored logs without 0x in front of the thread id
	  The non-ANSI colored regex already handled this correctly and this was required
	  for Windows before already anyway.
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/10857>

2026-02-19 13:08:39 -0300  L. E. Segovia <amy@centricular.com>

	* gst/gstregistry.c:
	  registry: Skip .dSYM bundles when loading plugins, try 2
	  If paths within the .dSYM folder are being passed, the current filter won't cover them.
	  See https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/10466#note_3330939
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/10853>

2026-02-12 02:51:07 +0530  Nirbheek Chauhan <nirbheek@centricular.com>

	* gst/gstregistry.c:
	* meson.build:
	  gstreamer: Stop using deprecated G_MODULE_SUFFIX
	  It's deprecated on macOS because shared libs can be either .so or
	  .dylib. We already account for that, we might as well stop using
	  G_MODULE_SUFFIX and get rid of deprecation warnings.
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/10807>

2026-01-28 11:54:11 +0530  Nirbheek Chauhan <nirbheek@centricular.com>

	* meson.build:
	  gstcheck: Disable fork on tvOS and watchOS
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/10807>

2026-01-28 11:45:20 +0530  Nirbheek Chauhan <nirbheek@centricular.com>

	* gst/meson.build:
	* libs/gst/helpers/ptp/meson.build:
	* meson.build:
	  meson: Deprecate `system = 'ios'` in cross files, use subsystem
	  Since we require Meson 1.4.0, we can now use subsystem names (added in
	  1.2.0).
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/10807>

2023-04-02 16:53:45 +0100  Tim-Philipp Müller <tim@centricular.com>

	* scripts/dist-common-files.py:
	* scripts/meson.build:
	  modules: dist common files from monorepo root
	  Less noise when making releases, and just need to maintain one copy.
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/10822>

2026-02-15 17:45:19 +0000  Tim-Philipp Müller <tim@centricular.com>

	* README.md:
	  modules: remove subproject README.md from git
	  Will be added to the tarballs based on the monorepo README on dist.
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/10822>

2026-02-15 17:20:59 +0000  Tim-Philipp Müller <tim@centricular.com>

	* RELEASE:
	  modules: remove RELEASE from git, will be generated from template on dist
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/10822>

2026-02-15 15:02:07 +0000  Tim-Philipp Müller <tim@centricular.com>

	* NEWS:
	  modules: Remove NEWS from git which is generated from full release notes
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/10822>

2026-02-13 19:50:56 +0900  Seungha Yang <seungha@centricular.com>

	* tests/check/libs/bitwriter.c:
	  tests: bitwriter: Fix leak in test
	  Fix leak in test revealed after bugfix. The data returned from
	  gst_bit_writer_reset_and_get_data() must be freed with g_free()
	  as documented.
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/10805>

2026-02-13 15:57:13 +0900  Seungha Yang <seungha@centricular.com>

	* libs/gst/base/gstbitwriter.c:
	  bitwriter: Steal owned data in reset_and_get_data()
	  Avoid unnecessary data copy/free by returning the owned
	  data directly to the caller
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/10805>

2026-02-11 16:49:02 +0900  Seungha Yang <seungha@centricular.com>

	* libs/gst/base/gstaggregator.c:
	  aggregator: Handle gap event before segment
	  If gap events are received before segment event, log the situation and
	  gracefully return FALSE instead of relying on the critical warning path
	  in gst_segment_clip()
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/10789>

2026-02-05 16:31:30 +0100  François Laignel <francois@centricular.com>

	* plugins/elements/gstmultiqueue.c:
	  multiqueue: reverse playback: use segment stop position as start time
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/10770>

2026-02-05 16:24:33 +0900  Seungha Yang <seungha@centricular.com>

	* plugins/elements/gstelements_private.c:
	  filesink: Report write error correctly on Windows
	  Propagate write() return value from the writev emulation path
	  as intended
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/10758>

2026-02-05 15:48:30 +0900  Seungha Yang <seungha@centricular.com>

	* plugins/elements/gstelements_private.c:
	  filesink: Add debug category to file write helper
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/10758>

2026-02-02 09:56:59 +0200  Sebastian Dröge <sebastian@centricular.com>

	* libs/gst/base/gstaggregator.h:
	  aggregator: Various introspection annotations / docs fixes for vfuncs
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/10723>

2026-01-26 23:21:06 +0800  He Junyan <junyan.he@intel.com>

	* gst/gstcaps.c:
	  caps: gst_caps_append_structure should not allow NULL structure
	  This check can also help to fix the possible memory leak of features
	  in gst_caps_append_structure_full() when the structure is NULL.
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/10689>

2026-01-28 13:53:28 +0200  Sebastian Dröge <sebastian@centricular.com>

	* libs/gst/net/gstptpclock.c:
	  ptpclock: Initialize PTP message to avoid invalid uninitialized variable warning
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/10656>

2026-01-28 11:57:56 +0200  Sebastian Dröge <sebastian@centricular.com>

	* libs/gst/helpers/ptp/parse.rs:
	  ptpclock: Fix copy-paste mistake in error messages
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/10656>

2026-01-28 11:52:13 +0200  Sebastian Dröge <sebastian@centricular.com>

	* libs/gst/helpers/ptp/parse.rs:
	* libs/gst/net/gstptpclock.c:
	  ptpclock: Fix potential out-of-bounds read when parsing ANNOUNCE messages
	  The required length of the message payloader after the common header is 30
	  bytes, not 20 bytes.
	  Also fix the same bug in the Rust code, but here it wouldn't have caused any
	  other problems than a less optimal error message because there is actual bounds
	  checking.
	  Fixes https://gitlab.freedesktop.org/gstreamer/gstreamer/-/issues/4871
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/10656>

2026-01-28 11:49:41 +0200  Sebastian Dröge <sebastian@centricular.com>

	* gst/gstdatetime.c:
	* gst/gsturi.c:
	  gst: Fix a couple of new const-ness warnings around strstr() usage
	  Fixes https://gitlab.freedesktop.org/gstreamer/gstreamer/-/issues/4871
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/10656>

2026-01-14 23:47:04 +0100  Ruben Gonzalez <rgonzalez@fluendo.com>

	* libs/gst/base/gsttypefindhelper.c:
	  typefindhelper: EOS case explained in more details
	  Related to https://gitlab.freedesktop.org/gstreamer/gstreamer/-/issues/4848
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/10632>

2026-01-09 20:17:52 +0100  François Laignel <francois@centricular.com>

	* plugins/elements/gstinputselector.c:
	  input-selector: don't wait for active pad when eos (active-segment mode)
	  When `sync-mode` is set to `active-segment` and current pad is not active,
	  don't wait depending on the active segment runtime if the active pad is EOS.
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/10621>

2026-01-09 18:20:51 +0100  François Laignel <francois@centricular.com>

	* plugins/elements/gstinputselector.c:
	  input-selector: fix regression receiving eos event
	  When a pad receives an eos event, it must first check for if the active pad
	  should be changed. Failure to do that could lead to the event not being
	  forwarded if current pad is the last pad remaining and it was not active
	  before.
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/10621>

2026-01-08 11:54:52 +0100  François Laignel <francois@centricular.com>

	* plugins/elements/gstinputselector.c:
	  input-selector: don't consider pad being released checking for eos
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/10621>

2026-01-06 15:06:57 +0100  François Laignel <francois@centricular.com>

	* plugins/elements/gstinputselector.c:
	  input-selector: gracefully handle active pad disappearance
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/10621>

2026-01-06 11:43:59 +0100  François Laignel <francois@centricular.com>

	* plugins/elements/gstinputselector.c:
	  input-selector: fix race condition selecting next active pad
	  When a pad is being released, the pad is marked as flushing and the selector
	  lock is released before the pad is removed from the element pad list. This is
	  necessary so any in-progress event or buffer handling for this pad terminates
	  before the pad is deactivated.
	  If another thread was trying to select the next active pad at this particular
	  moment, it could end up picking the pad being released before it was removed
	  from the element pad list.
	  This commit adds a flag which is set when the pad is being released, so it won't
	  get picked as the next active pad.
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/10621>

2026-01-27 17:10:54 +0000  Tim-Philipp Müller <tim@centricular.com>

	* meson.build:
	  Back to development after 1.28.0

=== release 1.28.0 ===

2026-01-27 17:02:33 +0000  Tim-Philipp Müller <tim@centricular.com>

	* NEWS:
	* README.md:
	* RELEASE:
	* gstreamer.doap:
	* meson.build:
	  Release 1.28.0

2026-01-26 18:51:01 +0530  Sanchayan Maity <sanchayan@centricular.com>

	* libs/gst/base/gstaggregator.h:
	  aggregator: Annotate some GstAggregatorClass methods
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/10596>

2026-01-26 14:54:59 +0530  Sanchayan Maity <sanchayan@centricular.com>

	* gst/gstelement.h:
	  element: Annotate request_new_pad
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/10596>

2026-01-25 17:17:25 +0000  Tim-Philipp Müller <tim@centricular.com>

	* po/LINGUAS:
	* po/ar.po:
	* po/bg.po:
	* po/de.po:
	* po/es.po:
	* po/hr.po:
	* po/nl.po:
	* po/pl.po:
	* po/ro.po:
	* po/sl.po:
	* po/sr.po:
	* po/uk.po:
	  gstreamer: update translations
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/10598>

2026-01-08 22:16:19 +0100  Ruben Gonzalez <rgonzalez@fluendo.com>

	* plugins/elements/gstidentity.c:
	  identity: cosmetic change to mimic fakesink
	  Remove extra space at chain log:
	  ```
	  $ gst-launch-1.0 -v videotestsrc num-buffers=1
	  ! identity silent=false name=i1 ! fakesink name=i2 silent=false
	  ...pts: 0:00:00.000000000, duration: 0:00:00.033333333, offset: 0, offset_end:  1, flags: 00000040..
	  ...pts: 0:00:00.000000000, duration: 0:00:00.033333333, offset: 0, offset_end: 1, flags: 00000040...
	  ```
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/10518>

2026-01-07 10:51:59 +0100  Samuel Thibault <samuel.thibault@ens-lyon.org>

	* gst/gstinfo.c:
	  gstinfo: Fix build on systems where pthread_t is int
	  Otherwise we get
	  error: returning ‘pthread_t’ {aka ‘int’} from a function with return type
	  ‘GstTid’ {aka ‘void *’} makes pointer from integer without a cast [-Wint-conversion]
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/10508>

2026-01-05 20:20:51 +0000  Tim-Philipp Müller <tim@centricular.com>

	* meson.build:
	  Back to development after 1.27.90

=== release 1.27.90 ===

2026-01-05 20:15:10 +0000  Tim-Philipp Müller <tim@centricular.com>

	* NEWS:
	* RELEASE:
	* gstreamer.doap:
	* meson.build:
	  Release 1.27.90

2026-01-05 18:57:01 +0000  Tim-Philipp Müller <tim@centricular.com>

	* po/af.po:
	* po/ast.po:
	* po/az.po:
	* po/be.po:
	* po/bg.po:
	* po/ca.po:
	* po/cs.po:
	* po/da.po:
	* po/de.po:
	* po/el.po:
	* po/en_GB.po:
	* po/eo.po:
	* po/es.po:
	* po/eu.po:
	* po/fi.po:
	* po/fr.po:
	* po/fur.po:
	* po/gl.po:
	* po/hr.po:
	* po/hu.po:
	* po/id.po:
	* po/it.po:
	* po/ja.po:
	* po/ka.po:
	* po/ko.po:
	* po/lt.po:
	* po/lv.po:
	* po/nb.po:
	* po/nl.po:
	* po/pl.po:
	* po/pt_BR.po:
	* po/ro.po:
	* po/ru.po:
	* po/rw.po:
	* po/sk.po:
	* po/sl.po:
	* po/sq.po:
	* po/sr.po:
	* po/sv.po:
	* po/tr.po:
	* po/uk.po:
	* po/vi.po:
	* po/zh_CN.po:
	* po/zh_TW.po:
	  gstreamer: update translations
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/10497>

2025-12-30 17:53:22 -0500  Doug Nazar <nazard@nazar.ca>

	* tests/check/elements/queue.c:
	  tests: Fix several memory leaks
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/10476>

2025-12-28 17:28:22 +0000  L. E. Segovia <amy@centricular.com>

	* gst/gstregistry.c:
	  registry: Skip .dSYM bundles when loading plugins
	  The .dylibs here aren't actual binaries, they're mach-O files with debug
	  sections only. These will be naturally skipped/blacklisted, but after a
	  long stream of dlopen errors.
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/10466>

2025-12-28 17:27:10 +0000  L. E. Segovia <amy@centricular.com>

	* libs/gst/helpers/ptp/meson.build:
	  ptp: Fix helper test naming convention and gate it
	  This ensures it doesn't get built unless -Dtests=enabled, and that its symbols
	  don't get confused for gst-ptp-helper's.
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/10466>

2025-12-26 18:03:19 -0500  Doug Nazar <nazard@nazar.ca>

	* tests/check/gst/gstinfo.c:
	  info: Increase test interval when running under valgrind
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/10456>

2025-12-19 15:40:26 +0530  Nirbheek Chauhan <nirbheek@centricular.com>

	* meson.build:
	  meson: Solve some cyclic dependencies caused by test-only deps
	  gstreamer => gobject-introspection => cairo => fontconfig => freetype2 => harfbuzz => cairo
	  gst-plugins-base => libdrm => cairo => fontconfig => freetype2 => harfbuzz => cairo
	  gst-plugins-good => cairo => librsvg => cairo
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/10422>

2025-12-19 15:19:42 +0530  Nirbheek Chauhan <nirbheek@centricular.com>

	* meson.build:
	  meson: Disable introspection option on glib
	  Allows us to update glib and gobject-introspection. We don't use the
	  new introspection infrastructure merged in glib anyway.
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/10422>

2025-12-19 18:49:46 +0100  Alicia Boya García <aboya@igalia.com>

	* gst/gstutils.c:
	  utils: Fix race in gst_element_link_pads_filtered()
	  gst_element_link_pads_filtered() can occur at the same time the bin
	  is in the middle of a PAUSED->PLAYING transition.
	  Since before this patch the capsfilter state lock bit wasn't being set,
	  in the case of a failure to link it was possible for the state change to
	  NULL in gst_element_link_pads_filtered() to race against a state change
	  to PLAYING inside gst_bin_change_state_func() and lose.
	  This caused errors when the unreffed capsfilter was disposed moments
	  later, as the element dispose function found the element to be
	  accidentally in PLAYING state.
	  As of writing, this particular race reproduces consistently when running
	  check.gst-plugins-good.elements_qtdemux.test_qtdemux_sample_interleaving
	  under valgrind with the the default fair scheduler (valgrind
	  --fair-sched=try), both locally and in the CI. This patch fixes it.
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/10383>

2025-12-18 09:19:52 +0100  Carlos Falgueras García <cfalgueras@fluendo.com>

	* plugins/elements/gstqueue.c:
	  queue: Log the status of the queue at debug level
	  Log the queue status at "GST_LEVEL_DEBUG" when it's empty or full by
	  reusing the current macro "STATUS()".
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/10403>

2025-12-15 15:08:02 -0500  Daniel Morin <daniel.morin@collabora.com>

	* gst/gstvalue.c:
	* gst/gstvalue.h:
	* tests/check/gst/gstcaps.c:
	* tests/check/gst/gststructure.c:
	* tests/check/gst/gstvalue.c:
	* tools/gst-inspect.c:
	  gst: Rename GstValueSet to GstValueUniqueList
	  GObject-Introspection has an issue with GstSet because anything that starts with
	  'gst_value_set' becomes something that belongs to 'GstSet' but we have
	  gst_value_set_bitmask and gst_value_set_SOMETHING (), which all would become
	  methods of GstSet.
	  To avoid this, rename GstSet (aka GstValueSet) to GstUniqueList (aka
	  GstValueUniqueList).
	  Fixes https://gitlab.freedesktop.org/gstreamer/gstreamer/-/issues/4813
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/10384>

2025-12-13 22:17:41 +0100  François Laignel <francois@centricular.com>

	* plugins/elements/gstinputselector.c:
	* plugins/elements/gstinputselector.h:
	  input-selector: implement a two-phase sinkpad switch
	  Switching the sinkpad was racy: when a new sinkpad was selected, previous
	  active sinkpad could continue pushing events or buffers, while the new sinkpad
	  already started pushing its sticky events, leading to inconsistencies
	  downstream.
	  This commit implements a a two-phase switch approach:
	  1. When the user selects the active pad, it is first set as pending.
	  2. Functions handling buffers or events first check whether a new pad is pending
	  activation, in which case the new pad activation is commited.
	  3. After completing step 2, the functions handles the buffer or event.
	  The active pad can not change before the buffer or event is pushed.
	  Mutability of the pending pad is protected by SELECTOR_LOCK.
	  Mutability of the active pad is protected by a dedicated RW lock which can be
	  locked independently from SELECTOR_LOCK and can prevent active pad changes when
	  that would interfere with current processing.
	  The RW lock allows preventing the active_pad from being changed while we
	  are processing a buffer or event.
	  Only `gst_input_selector_maybe_commit_active_pad ()` can get this lock in
	  writer mode for a short period of time (committing the `pending_active_pad`
	  to `active_pad`). Other concurrent code paths lock it in reader mode.
	  When a `pending_active_pad` is available (and only in that case),
	  code paths calling `gst_input_selector_maybe_commit_active_pad ()` would
	  hang until the lock is available for writer mode, which is expected since
	  their processing should be performed with respect to the expected `active_pad`.
	  Note that a `GRecMutex` was also considered to prevent possible deadlocks
	  in callbacks, but only a `GRWLock` can deal with the concurrency which occurs
	  during preroll.
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/10204>

2025-12-09 19:13:20 +0000  Tim-Philipp Müller <tim@centricular.com>

	* meson.build:
	  Back to development after 1.27.50

=== release 1.27.50 ===

2025-12-09 19:08:48 +0000  Tim-Philipp Müller <tim@centricular.com>

	* NEWS:
	* RELEASE:
	* gstreamer.doap:
	* meson.build:
	  Release 1.27.50

2025-12-05 11:19:25 +0200  Sebastian Dröge <sebastian@centricular.com>

	* libs/gst/base/gstbasesrc.c:
	  basesrc: Fix decide allocation fallback
	  Fallback to generic pool, even if the change to the config was validated. Pools
	  may have other restrictions that may render them unusable.
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/10298>

2025-12-04 16:48:37 -0500  Nicolas Dufresne <nicolas.dufresne@collabora.com>

	* libs/gst/base/gstbasetransform.c:
	  basetransform: Fix decide allocation fallback
	  Fallback to generic pool, even if the change to the config was validated. Pools
	  may have other restrictions that may render them unusable.
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/10298>

2025-11-28 14:12:53 -0600  Olivier Crête <olivier.crete@collabora.com>

	* tools/gst-inspect.c:
	  gst-inspect: Pretty print tensor caps
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/9109>

2025-08-03 21:39:36 -0400  Daniel Morin <daniel.morin@collabora.com>

	* libs/gst/base/gstbasetransform.c:
	  basetransform: Explicitly set sinkpad flag to GST_PAD_FLAG_ACCEPT_INTERSECT
	  - Note this is NOT a breaking change.
	  - Base transform accept caps evaluation is based on intersection between caps
	  and template while the default behaviour should be to verify that caps is a
	  subset of the template. There's a specific pad flag to indicate accept caps
	  should be based on intersect (GST_PAD_FLAG_ACCEPT_INTERSECT). Since alot of
	  elements base based on bastransform and the expected behaviour from the base
	  class to evaluate accept caps is to use intersect we set the flat
	  GST_PAD_FLAG_ACCEPT_INTERSECT inside gst_base_transform_init() on the sinkpad
	  and inside gst_base_transform_acceptcaps_default() we use interect if the flag
	  is set otherwise we use gst_caps_is_subset (caps, template). This way
	  basetransform will honor the flag and allow element that want accept caps to
	  be base on subset to clear this flag.
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/9109>

2025-10-30 17:03:35 -0400  Daniel Morin <daniel.morin@collabora.com>

	* docs/random/caps_grammar:
	  doc: adding caps grammar doc
	  - Include new grammar for GST_TYPE_SET
	  - Converted doc to markdown
	  - Updated grammar description based on what exist.
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/10262>

2025-07-18 15:14:52 -0400  Daniel Morin <daniel.morin@collabora.com>

	* tests/check/gst/gstcaps.c:
	* tests/check/gst/gststructure.c:
	* tests/check/gst/gstvalue.c:
	  test: Tests for GstSet
	  - tests for GstSet
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/10262>

2025-07-29 09:12:46 -0400  Daniel Morin <daniel.morin@collabora.com>

	* gst/gststructure.c:
	  gststructure: fix issue when structure contain GST_TYPE_SET
	  - 'set' already use cast annotation for divisor, we don't add another one, to
	  avoid generating '(set)(/set){}'.
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/10262>

2025-11-28 14:24:53 -0500  Daniel Morin <daniel.morin@collabora.com>

	* gst/gstvalue.c:
	  gstvalue: remove code duplication from GstList
	  - Improve code maintainability by removing duplicated code
	  Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/10262>

2025-07-18 15:06:03 -0400  Daniel Morin <daniel.morin@collabora.com>

	* gst/gst_private.h:
	* gst/gststructure.c:
	* gst/gstvalue.c:
	* gst/gstvalue.h:
	  gstvalue: adding GstSet
