Romain Vimont authored 3 years ago and Hugo Beauzée-Luyssen committed 3 years ago

This filter crops the input pictures by adjusting the format (offsets
and size), without any copy (contrary to croppadd).

This is especially useful to receive pictures with arbitrary visible
area (smaller than the full size), to debug vouts or filters.

User agents are apparently now expected to do this; failure to do so
results in the video file data transfer getting throttled down to rates
such as 80 kB/s, 60 kB/s or 40 kB/s, below playback rate, and usually
resulting in a video that hangs upon loading or every few seconds, and
is impossible to play. This behavior seems to have first appeared in
June, but been fully rolled out only last week.

Just like with URL signatures, we interoperate with YouTube by
fulfilling what's apparently expected from us, using the same approach
as so far: we parse the descrambling rules from the javascript code, and
apply them.

Fixes #26174

This should help against transient errors, and parsing of the javascript
URL isn't the part that's most likely to break.

We'll be descrambling the "n" parameter in addition to the URL signature
using this same javascript web asset, so we want to be able to share and
reuse it.

Use a more specific name as this isn't the only parameter anymore that
we'll be descrambling by parsing and emulating javascript.

Javascript variables can contain other, special characters, also %a
depends on the locale.

After tightening access restrictions to it, the get_video_info YouTube
API was completely retired around July 2021, with an HTTP 410 Gone code.
All this fallback achieves anymore is poor UX.

Alexandre Janniaux authored 3 years ago and Jean-Baptiste Kempf committed 3 years ago

The support for detached threads has been removed completely. There is
no need to document then in vlc_join now.

See the following commits:

android: thread: remove unused detached thread support
043d7ebf

os2: thread: remove unused detached thread support
f09937dc

win32: thread: remove unused detached thread support
8675f083

threads: remove vlc_clone_detach()
a10ac09d

Deprecate vlc_clone_detach()
c5f960ff

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

shortcut did exists in VLC 3.0

fix: #25892

Rémi Denis-Courmont authored 3 years ago and Jean-Baptiste Kempf committed 3 years ago

The input format is fixed for the lifetime of the decoder thread.
The output format is not always available as the packetiser or the
decoder may delay the determination of their output format until they
start outputting packets.

The downside is the assumption that input and output category are the
same. This assumption is already enforced by an assertion, and relied
heavily upon by the ES output for track selection, so it is hardly a
problem.

Fixes #21975

Romain Vimont authored 3 years ago and Jean-Baptiste Kempf committed 3 years ago

A m×n matrix contains m rows and n columns.

My usage (m columns and n rows) did not match the common convention, so
what I called 3x2 matrices were in fact 2x3 matrices.

Felix Paul Kühne authored 3 years ago and Jean-Baptiste Kempf committed 3 years ago

This adds support for the spatial audio API introduced in iOS 15.

Romain Vimont authored 3 years ago and Romain Vimont committed 3 years ago

To multiply two 2x3 matrices, the right matrix must be expanded to 3x3,
by adding a row [0 0 1], so that the dimensions match (2x3 x 3x3 = 2x3).

The matrix multiplication was erroneous: it behaved as if the matrix was
expanded by an implicit row [1 1 1] instead.

The error has been introduced by
31c20009.

According to the specification, EGL displays are not reference-counted
by default. As such, `eglTerminate()` will terminate a display even if
there are other users for it in the same process.

This becomes a fatal problem if platform display is shared by multiple
components, and the EGL display parameters are identical or deemed
compatible by the EGL driver. Typically, this will occur with the
window provider and the display.

To fix this, the Khronos group defined an EGL extension to negotiate
reference counting explicitly.

* On Android, this patch makes no differences as the EGL enables
  reference counting by default, regardless of the baseline
  specifications.
* On X11, this patch makes no differences because the module will
  create its own private Xlib `Display` as the platform display,
  such that the EGL display is not shareable.
* This patch matters on Wayland where the `wl_display` pointer must be
  shared between the window provider and display. Without this patch,
  Either the window provider or the display cannot use EGL.
* On Windows, the situation ought to be similar to Wayland, though this
  is much less of an issue, as EGL is not typically used.

This patch does **not** require reference counting if not available as
this would needlessly break on Android and X11, as well as with the
non-embedded XDG-shell window provider on Wayland.

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

Since 6b7a6d86 the shaders are stored in a
ComPtr. Copying the value also adds a reference automatically.

Romain Vimont authored 3 years ago and Jean-Baptiste Kempf committed 3 years ago

The OpenGL states must not be set on Open, but on each Draw call if
necessary (they may impact other filters).

In practice, it is not necessary to set them at all:
 - GL_CULL_FACE is useless for the renderer;
 - the other states are assumed to be the default values in all other
   filters.

Suggested-by: Niklas Haas <git@haasn.dev>

Rémi Denis-Courmont authored 3 years ago and Jean-Baptiste Kempf committed 3 years ago

For historical reasons, the TLS cleanup was serialised on a critical
section. Now that there is a read/write lock, there are no reasons to
serialise exit. The thread exit process only reads the list head and
elements, does not modify the list per se; only vlc_threadvar_create()
and vlc_threadvar_delete() add and remove elements respectively.

This allows threads to exit in parallel.

Rémi Denis-Courmont authored 3 years ago and Jean-Baptiste Kempf committed 3 years ago

The VLC API "wastes" time preserving the thread's system error code.
There are no ways to consume those errors in the thread exit code, so
there is really no point in going out of our way here.

Rémi Denis-Courmont authored 3 years ago and Jean-Baptiste Kempf committed 3 years ago

Old-style TLS keys only need to be in the list if they have a
per-thread destructor. If they don't, then don't put them in that list,
so iterating the list is faster.

Rémi Denis-Courmont authored 3 years ago and Jean-Baptiste Kempf committed 3 years ago

Old-style TLS keys only need to be in the list if they have a
per-thread destructor. If they don't, then don't put them in that list,
so iterating the list is faster.

Romain Vimont authored 3 years ago and Rémi Denis-Courmont committed 3 years ago

If there is no crop, the right and bottom position are the width and the
height, respectively.

Fix regression introduced by 13282687.

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

If not we are leaking a TlsAlloc ID and it make indicate other leaks in
more threads.

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

If not we are leaking a TlsAlloc ID and it make indicate other leaks in
more threads.

The API is available since Vista and is also available in winstore app. We
don't have to support 2 different versions anymore in 4.0.

The "ifdef" doesn't seem to be taken in account. It's as if the define was
always set. Now that the define is properly used we need to pass it to moc.

If http-proxy exists, it is already checked by the common core code.