This update contains the optimization for rendering cues
with large line values. It also reverts the change to clamp
parsed line values to between -1000 and 1000.
* Update vtt.js to the newest version which has the capability
of reporting back ParserErrors.
* Update the update-vtt.js script a bit as well since
the symbols that it exports have changed a bit since the last import.
Some of the general information it outputs has since changed as well
like where the repo exists.
* Update WebVTTListener to set the track readiness state
to 'failed to load' if the WebVTT file is not loaded correctly.
This completes the initial implementation of the processing algorithm.
The algorithm doesn't support vertical text yet, but that's not a
big issue since Firefox doesn't support it.
- Current source code for vtt.js can be found at:
https://github.com/andreasgal/vtt.js
- vtt.js code taken from commit 6bca57d265fc58cff091148b426ec69f576057d7.
- Added vtt.js, a JS WebVTT parser, to Gecko.
- Added a wrapper to interface between it and Gecko.
- Updated TextTrackCue and WebVTTListener to work with it.
--HG--
rename : content/media/WebVTTLoadListener.cpp => content/media/WebVTTListener.cpp
rename : content/media/WebVTTLoadListener.h => content/media/WebVTTListener.h
