Steve Mayhew
Steve Mayhew
@icbaker The decoder is producing 2 for 1 output frames (each input buffer produces **two** decoded frames. So the timestamps don't exists for half of the frames (as there was...
@tonihei I reviewed the pull request #10468 and tested the code, while it is sort of on a possible correct path... It has issues so it does not fix this...
> I think that a decoder is not allowed to create additional output buffers. The [`MediaCodec` javadoc](https://developer.android.com/reference/android/media/MediaCodec) seems to make it pretty clear there's a 1:1 correspondence between input and...
@icbaker I promised to keep you in the loop, our reseller worked with SEI/AMLogic to get a version of their firmware that "fixes the issue". There is no change required...
Also, so it's clear, the pull request https://github.com/google/ExoPlayer/pull/10468 is not needed to fix the problem, only updated firmware from SEI/AMLogic.
I cannot build from source on my workstation. Here's the versions I have: ```` SteveMayhewsMacBook:haxe smayhew$ git status HEAD detached at 3.4.7 nothing to commit, working tree clean SteveMayhewsMacBook:haxe smayhew$...
> @stevemayhew can you try compiling from outside the "haxe" directory? Interesting... that worked for the Neko target just fine: stevemayhewsmacbook:opensource smayhew$ export HAXE_STD_PATH=~/opensource/haxe/std/ stevemayhewsmacbook:opensource smayhew$ ~/opensource/haxe/haxe -main Main --interp...
Ah.. So the gotcha is if your class path is a parent of HAXE_STD_LIB? Given the way we construct our build we will have to be careful of that since...
Feel free to simply take this, change it around a bit or suggest a completely different approach. I've done developer testing on the change with the sample URL I sent...
> The reason this is happening as described above is that the declared segment durations of **some** segments in the HLS playlist don't really match the actual duration of the...