[fritsch]

Hi Fritsch,
so sorry to hear that you left the team. We will really miss you: you were the example to any other dev out there, always willing to help, investigate, develop, and had the willingness to help even novice users. Not to mention the extraordinary job you did with Kodi making picture quality so much better. I'm very sorry that you decided to leave before your VAAPI-EGL work hit master in v17.

Thanks for all your contributions!

Of course, this goes to other devs too, and I don't want to mention names here.

I do have some very bad feeling that this (and many more) project(s) are abandoned now, which will be such a shame. I hope @FernetMenta reads this...

Just a short note here, as your last sentence seems to go into a wrong direction: FernetMenta is 100% (!) not the reason I left - he was one of the single reasons I stayed even as we had such discussions (*), concerning the future of kodi with others in the team 6 months before. Everything I learned over the last 4 years was guided and in cooperation with Fernet. Btw. VAAPI-EGL has already landed in v17 master.

This discussion now got into two directions:
(*) It was proposed, that platforms (Windows, Android, whatever) should get special branches and features should be allowed to be commited without honoring the cross-platform aspect, Those branches are then released as final versions. This does not fit into my philosopy of cross-platform development at all. I am not against features (as you see in this thread), but features that will degrade, break our code platform over time is a feature not worth it. Btw. the current discussion about this Android stuff - this is shield only (!) code - it won't work anywhere else (no other android box that you can buy without hacking) as of v16 release ... so risking a complete go down of complete kodi, just cause one out of thousand boxes out there would benefit from it ... is a very high price. A price not worth it at all, as fernet and I have implemented the way passthrough needs to work for shield's android version, a version that most likely be supported with the next android version (Android 6), not even in later beta yet (**). This code is already in v17. Yes, fernet and I have rewritten the entire audio engine for v17 - platform agnoistic, so that the new Android platform can use it without any major hickups, which is a proof that one can hold up high "platform agnostic" _and_ features.

You all in here wait since pre version 15.x for a stable version with VAAPI-EGL. It won't be in v16, cause other platforms were not ready for the jump (VideoPlayer adjustments for specific platforms, Renderer and so on), so in order to keep our cross-platform spirit wide open - fernet and I decided: Hey, let's do it right and correct and do it with v17 - just that others can catch up. If you see some of our github histories, you see that ~ 10 people constantly helped to work on this, so that VideoPlayer merge was possible. That was the spirit I see in kodi and which made everything worth it. We will loose all this for "short time honor".

(**) Also then it is totally not clear if it works at all. Current implementation works (it is to my knowledge a shield specific backport of future android audio stack), cause the sink is fed with one packet always the same size (matching AC3, DTS, whatever packet), the very moment only a single byte of such a packet is not fully consumed - the approach won't work as the engine starts fragmentation ... and yeah even more funny is, that whenever a background process will run - it should cause stutter and dropouts, cause the packages are really damn small and underrun is easily possible, which is hacked arround by pausing the sink. So whenever you see your AVR blinking with DTS / AC3 and so on - you know why.