ffmpeg output parser fix

[KurtisVargas]

The ffmpeg parser expects '\r' in the output to yield to the caller. For small or fast ffmpeg jobs, ffmpeg never outputs the '\r' so the job is detected as fails.

This patch handles the case where a job is finished but never yielded. Seems to not mess with legitimate errors.

[mdhiggins]

Thanks.

Also, wanted you to check this issue out and see if you think its involved

senko/python-video-converter#56

[KurtisVargas]

I don't think the other pull is accurate. I checked out output of the
ffmpeg daily build with a hex editor and it still uses '\r' like it did
before when refreshing the status line with fps/timeleft/etc. I've only
seen the '\n' when ffmpeg doesn't have to update the timeleft because it
processes it so quickly.

The pattern match (pat = re.compile(r'time=([0-9.:]+) ') is not that tight
so I could see it finding a bad match with the prodigious amount of ffmpeg
output.

Definitely hold onto the thought - it might be necessary if things change.

On Wed, May 6, 2015 at 9:57 PM, Michael Higgins [email protected]
wrote:

Thanks.

Also, wanted you to check this issue out and see if you think its involved

senko/python-video-converter#56
senko/python-video-converter#56

—
Reply to this email directly or view it on GitHub
#257 (comment)
.