GitLab

Reverted 4082c803 as AudioDeviceIOProcID and its related methods require Darwin 9

This improves readability a lot and prevent us from reinventing the wheel.

auhal: Fix kAudioDevicePropertyDeviceIsAlive check (vlc_bool_t size has changed and doesn't match anymore the needed param size for kAudioDevicePropertyDeviceIsAlive).

Missing some cases that I could not test.
Also ffmpeg/chroma.c is locked, so I can but leave it broken.
Fix is for the remaining modules is obvious and the issue is easy to detect.

* Headers in include must contain the structures and prototypes needed
  by the plugins and should be named vlc_*.
* Headers in include/vlc must contain the structures and prototypes
  needed by external libvlc clients

* Moved and renamed some things in headers. 
   - Removed vlc_cpu.h, vlc_error.h (merged in vlc_common)
   - Removed snapshot.h (merged in vlc_vout.h)
   - Removed vlc_spu.h (merged in vlc_osd.h)
   - Removed intf_eject.h and vlc_interaction.h (merged in vlc_interface)
   - Moved all internal headers to src
   - Merged vlc_video.h and video_output.h and move private things to src
   - Removed vlc/intf.h, vlc/aout.h, vlc/vout.h, vlc/decoder.h and vlc/input.h
     (meta headers for use in modules, and often implying too large 
     dependencies)
* Removed some useless dependencies
* Unexported a bunch of functions and structures used only in src/
  (--> Still some work here)
* Finally made input_thread_t and input_source_t (mostly) private. 
  Added input_GetItem to fetch the input_item of an input_thread
* Cleaned up deprecated entries in vlc_symbols.h and bumped up symbol 
  prefix

This commit has a 99% probability of breaking the build, 0.1% of killing
your cat and 0% of getting you hot chicks.

Also, make it clearer what is shared and instance-specific

So,
libvlc_t ==> libvlc_global_t
vlc_t ==> libvlc_int_t (internal instance)

p_object->p_libvlc ==> p_object->p_libvlc_globale
p_object->p_vlc ==> p_object->p_libvlc

VLC_OBJECT_VLC ==> VLC_OBJECT_LIBVLC

And by the way, there is some cleanup required :) (ie, some things are created in the global object instead of the instance-specific one)

* browsed through all code files starting with A to F and added non-blocking intf_UserFatal calls next to msg_err where I think that it is necessary.
I tried to select the most important msg_errs. If I missed your favourite, feel free to add it.
Feedback on the wording and the placement of the interaction calls is strongly welcomed.

For the records: I'm using some general topics for most of the dialogues to produce less work for the translators. The current ones go like this:
"Audio output failed", "Network interaction failed", "Streaming / Transcoding failed", "Audio filtering failed", "File reading failed", "Playback failure"

More to come.

  Grmbl @ Digidesign coders...

  - fix last bugs with SPDIF stream selection
  - code and debug cleanup
  - added more comments to clearify the code, so it is useful to others :D

SetProperty in CoreAudio is not only Asynchronious, but also not atomic in it's bahaviour. Check the actual set format multiple times to make sure the SetProperty completly finished.

Use the original samplerate of the a52 or dts stream, if the device is capable of it, current samplerate if it is not, and highest other available samplerate as a last resort.

* Don't upmix the channels if we only have Stereo or Mono audio. Let the device setup take care of that.

Mac OS X audio module is now unified for analog and digital output. todo: test on a shitload of devices, test for stability (removal of USB audio device during playback), rename the plugin

First steps towards integrating coreaudio and auhal module into one. It does not actually render digital audio yet.

(da big courmischage)

* Add some more comments on speaker layouts etc.

* Update NEWS and README with info on the new module.

* Tell the auhal which channel ordering VLC uses. 7 and 8 channel configs are not supported because VLC uses the channel ordering L R Lm Rm Lr Rr C LFE which isn't a normal ITU, DVD, MPEG, AAC or WAV standard and therefore not supported by default by OS X.

* Fix several problems with the auhal module. only the problem with the changing inNumberFrames should be left. We need to add a buffer system to take care of this issue...