Pierre Lamot authored 3 years ago and Jean-Baptiste Kempf committed 3 years ago

Co-authored-by: Benjamin Arnaud <benjamin.arnaud@videolabs.io>

Pierre Lamot authored 3 years ago and Jean-Baptiste Kempf committed 3 years ago

Co-authored-by: Benjamin Arnaud <benjamin.arnaud@videolabs.io>

Alaric Senat authored 3 years ago and Jean-Baptiste Kempf committed 3 years ago

`ops` being NULL is a possibility when encoder loading failed.

Fixes #26624

Thomas Guillem authored 3 years ago and Rémi Denis-Courmont committed 3 years ago

Fixes GeoVision video playback via the avi demuxer.

Lyndon Brown authored 3 years ago and Hugo Beauzée-Luyssen committed 3 years ago

The update/clear icon-based button here gives no indication that it has
the focus if you tab to it via keyboard navigation. Switching the button
type from `QPushButton` to `QToolButton` fixes this.

Note that the similar button in the sync tab of the effects dialog is of
type `QToolButton` and did not exhibit the problem.

(Non-functional change; no other change in appearance).

Lyndon Brown authored 3 years ago and Hugo Beauzée-Luyssen committed 3 years ago

(No functional change).

Lyndon Brown authored 3 years ago and Hugo Beauzée-Luyssen committed 3 years ago

The effects dialog sync tab widget should belong to the main tab widget,
not the secondary tab widget within the video tab, which makes no sense.

(No functional change).

Lyndon Brown authored 3 years ago and Hugo Beauzée-Luyssen committed 3 years ago

If you used the `Tab` key to navigate between controls within the
synchronization tab of the effects dialog, once you landed upon the refresh
button (the circular icon, top-right) navigation got stuck, refusing to
jump to a new control.

I tracked down the cause to be related to the fact that the next control in
the tab list will be the 'apply' button, but in a tab change event handler
we hide the apply button for the sync tab. This seems to confuse Qt, though
navigating in reverse (`Shift`+`Tab`) works fine. Flipping the focus policy
of the button depending upon whether or not it is hidden works as a fix.

Fixes #25923.

Hugo Beauzée-Luyssen authored 3 years ago and Jean-Baptiste Kempf committed 3 years ago

Fix #26491

Steve Lhomme authored 3 years ago and Jean-Baptiste Kempf committed 3 years ago

Patch merged upstream

Martin Storsjö authored 3 years ago and Jean-Baptiste Kempf committed 3 years ago

This fixes build breakage with latest nightly libc++. Since
https://github.com/llvm/llvm-project/commit/2e2f3158c604adb8401a2a44a03f58d4b6f1c7f9,
fewer libc++ headers transitively include <algorithm>.

This protobuf header uses std::fill unconditionally, which is a
function provided by <algorithm>. (The surrounding code comment was
wrong/outdated - <algorithm> was needed on all platforms.)

The same seems to have been fixed in upstream protobuf as part
of a larger code sync:
https://github.com/protocolbuffers/protobuf/commit/ab4585a6956675ce14a1cba5d321fde980bbf12b#diff-5cc9e8e347380b7bbc8fac8c06dc98100f3b5c076fcffa177e8ce11e753d4f7c

Zhao Zhili authored 3 years ago and Jean-Baptiste Kempf committed 3 years ago

st3d/sv3d can appear in any VisualSampleEntry. It's easy to miss
entry for new codecs.

Tristan Matthews authored 3 years ago and Hugo Beauzée-Luyssen committed 3 years ago

2022-01-28 v3.3.0
  This release includes compression efficiency and perceptual quality
  improvements, speedup and memory optimizations, some new features, and
  several bug fixes.
  - New Features
    * AV1 RT: Introducing CDEF search level 5
    * Changed real time speed 4 to behave the same as real time speed 5
    * Add --deltaq-strength
    * rtc: Allow scene-change and overshoot detection for svc
    * rtc: Intra-only frame for svc
    * AV1 RT: Option 2 for codec control AV1E_SET_ENABLE_CDEF to disable
      CDEF on non-ref frames
    * New codec controls AV1E_SET_LOOPFILTER_CONTROL and
      AOME_GET_LOOPFILTER_LEVEL
    * Improvements to three pass encoding
  - Compression Efficiency Improvements
    * Overall compression gains: 0.6%
  - Perceptual Quality Improvements
    * Improves the perceptual quality of high QP encoding for delta-q mode 4
    * Auto select noise synthesis level for all intra
  - Speedup and Memory Optimizations
    * Added many SSE2 optimizations.
    * Good quality 2-pass encoder speedups:
      o Speed 2: 9%
      o Speed 3: 12.5%
      o Speed 4: 8%
      o Speed 5: 3%
      o Speed 6: 4%
    * Real time mode encoder speedups:
      o Speed 5: 2.6% BDRate gain, 4% speedup
      o Speed 6: 3.5% BDRate gain, 4% speedup
      o Speed 9: 1% BDRate gain, 3% speedup
      o Speed 10: 3% BDRate gain, neutral speedup
    * All intra encoding speedups (AVIF):
      o Single thread - speed 6: 8%
      o Single thread - speed 9: 15%
      o Multi thread(8) - speed 6: 14%
      o Multi thread(8) - speed 9: 34%
  - Bug Fixes
    * Issue 3163: Segmentation fault when using --enable-keyframe-filtering=2
    * Issue 2436: Integer overflow in av1_warp_affine_c()
    * Issue 3226: armv7 build failure due to gcc-11
    * Issue 3195: Bug report on libaom (AddressSanitizer: heap-buffer-overflow)
    * Issue 3191: Bug report on libaom (AddressSanitizer: SEGV on unknown
      address)
    * Issue 3176: Some SSE2/SADx4AvgTest.* tests fail on Windows
    * Issue 3175: Some SSE2/SADSkipTest.* tests fail on Windows

Source:
https://aomedia.googlesource.com/aom/+/refs/tags/v3.3.0/CHANGELOG

Remove vlc_gl_HasExtension helper from the public API and implement it
as a private helper in modules. The reason for such helper was to help
ensuring the core stays compatible with Core OpenGL profile, which is
broken as soon as glGetString(GL_EXTENSIONS) is called instead of the
glGetStringi(GL_EXTENSION, i) variant.

Having the helper in the core led to issues against the MacOSX
implementation.

The extension store in gl_util.h prevents reloading the functions
across multiple calls from vlc_gl_HasExtension, and the helper can
easily be used in any OpenGL client which needs access to the extension
list.

Fix #26606 regression as a side effect

Co-Authored-by: Zhao Zhili <quinkblack@foxmail.com>

The API is not used anymore, and with the removal of the reference
counting pattern in objects, not required.

Indeed, the OpenGL providers are neither multi-thread nor reentrant
so it needs a locked stated when multiple clients needs to use it,
which will also needs to be reference counted across the clients, so
the whole reference counting can be done there.

In addition, if the context was given from module to other module, like
in the filter chain, the video contexts forwarded in those chains would
already be doing the reference counting. As a side note, there would
probably be few reasons to implement such forwarding currently anyway
since it would means that the beginning of the chain would lose usage
of the context, which also encompass a pool of output picture too
currently.

refs #26602

upstreamed at: https://github.com/taglib/taglib/pull/1040
Fix #26602

This is already done implicitly

Fixes CVE-2021-31566, CVE-2021-36976