From Github #17
I was trying to use this as part of a general video pipeline, but the script doesn't exit with error codes if there are any problems (other than the one check to verify that ffmpeg exists). This makes it hard to know if it succeeded or not.
With the forked processes for multiple bitrates this may be a little tricky to implement. But it might also be that we can switch away from forking the processes (if I can find my notes from previous tinkering to see why I chose not to use it).
In any case, we should at least attempt to return an appropriate exit status on failure.
Activity
2019-07-28 12:05:02
Due to an NFS fault, mux jobs were running for a while and then FFMPEG segfaulted, leaving an incomplete encode, with no way for the wrapper script to identify it.
2019-07-28 12:12:42
Because of the difficulty of getting data from a forked process, in the event of a non-zero exit,
I don't really like using files for this, it's messy, but is probably the simplest way to implement for now - at least without a substantial refactor.
The main question mark over the PoC really, is whether or not I'm checking for the correct codes, it seems - https://lists.ffmpeg.org/pipermail/ffmpeg-user/2013-July/016245.html -
2019-07-28 12:15:52
Webhook User-Agent
View Commit
2019-07-28 12:55:52
Webhook User-Agent
View Commit
2019-11-23 12:09:16
2019-11-23 12:09:16