As MONOTONIC is defined by POSIX to be the one that ticks during sleep while
mach absolute time does not.

For now (and for 3.0) vlc_tick_now() has the same source than the mach one.

It seems that the client_id API magic now gets invalidated much faster,
rendering the current model, where it is hard-coded into the script,
not viable anymore. Instead, this fetches and parses javascript assets
linked from the web page to search and extract the up-to-date magic from
them.

This is significantly slower as for now it fetches all of several
indiscriminate javascript assets, but at least it works, and should not
require that kind of maintenance anymore.

Which fixes non-optimized win64 builds

VLC needs Mingw-w64 v5.0 or higher , the define mentioned in the patch
was added in v3.0.0

Pierre Lamot authored 5 years ago and Alexandre Janniaux committed 5 years ago

The table is only used by the C file but still duplicated in every file
including vlc_text_style.h. If needed in other file, a public function
might be better than managing the table itself.

Signed-off-by: Alexandre Janniaux <ajanni@videolabs.io>

Found by lgtm.com

It may not be TRIPLET-pkg-config in some cases.

Romain Vimont authored 5 years ago and Steve Lhomme committed 5 years ago

The helper fragment_shaders.c was responsible to create the whole
fragment shaders. For now on, make it responsible to generate only the
piece of GLSL code necessary to access the input picture (i.e. to
implement vlc_texture()).

Make the renderer create its own fragment shader, using the
vlc_texture() function generated by fragment_shaders.c.

Signed-off-by: Alexandre Janniaux <ajanni@videolabs.io>

Romain Vimont authored 5 years ago and Steve Lhomme committed 5 years ago

Extract a fragment shader function to provide the pixel color of a
picture from pictures coordinates.

Concretely, it replaces:

    void main(void) {
     vec3 pic_hcoords = vec3(PicCoords, 1.0);
     ...
     gl_FragColor = result * FillColor;
    }

by:

    vec4 vlc_texture(vec2 pic_coords) {
     vec3 pic_hcoords = vec3(pic_coords, 1.0);
     ...
     return result * FillColor;
    }
    void main() {
     gl_FragColor = vlc_texture(PicCoords);
    }

Signed-off-by: Alexandre Janniaux <ajanni@videolabs.io>

And remove the patch that was merged upstream.

In the unlikely event that multiple threads want to check the CPU
flags concurrently, simply compute them as many times.

This removes the useless barriers, and the only usage of vlc_once() in
the core.

user does no longer override min values

min buffer <= max buffer <= live offset

This is not needed anymore since the PTS is the input one, and not the clock
fixed one.

This fixes the window size that was ignored when the size was updated from the
open callback of the window plugin.

Fixes #22674

It should be used before the initialisation of the vd plugin. During the
lifetime of the vd plugin, only cfg->display size should be used.

Refs #22674

Unused since 74da7a42, see conversation
on the ML [1] for more details on the intent.

[1] https://mailman.videolan.org/pipermail/vlc-devel/2007-September/034922.html

The profile is not used if the define doesn't exist.

SLresult is defined as an SLuint32 and might not be represented as an
unsigned long int. It is raising warnings when compiling for android
aarch64, so cast it to ensure it's an uint32_t and display it with
PRIu32 instead.

block_t->i_buffer is size_t too and must be displayed with %zu.

Instead of checking new requests every seconds, wait for actual new
requests to be enqueued before trying to do any processing. This avoid
the harmful sleep warning on vlc_tick_sleep as well as providing better
latency for new fingerprinting requests on a cold fingerprinter.

The vlc_tick_sleep() was a vlc_cond_timedwait before and became a msleep
since 832d5a6d, so there was no clear
reason why it is there.

The processing queue is using the player lock with the processing
condition variable, so this has become a unused variable since player
refactor at least.

Using memset to zero a struct with pointers isn't guaranteed to set
those pointers to NULL and is less elegant than using partial
initialization.

If breakpad is not used, the crash_handling variable is not used too,
but as it's used otherwise this should be silenced.

This is an update of 30 releases...
It might break some setups, hopefully not.

Patch submitted upstream