--- Comment #4 from Michael Dale <> 2011-02-24 23:46:06 UTC ---
Error handling is important. If the encoder caches or the server dies
unexpectedly the file will never be considered "encoded" since it only gets
moved after a successful encode. Its certainly possible the encoder thinks the
file was successfully encoded while human viewers disagree. But re-running the
encode with the same settings every time someone requests ?action=purge will
likely just result in the same broken file being regenerated over and over. 

It would be nice to let users register a given video has bad encode or has
"playback issues", but that is a separate feature than the bug being discussed
here. ( probably a phase 2 feature )

Purges should be managed via maintenance scripts so we it can be run in
conjunction with new encoder deployments or deployments of changes in encoding

Users are not completely dependent on these maintenance scripts they can
"upload another version of the file" which will of result in purge of

If in practice we run into the situation where derivatives being purged and
re-encoded with the same settings will 'make the files work', then yes I am
open to bump the priority on letting some users purge transcodes somehow. 

*But the original bug is about differentiating from image purges* which I think
we should do regardless.

Configure bugmail:
------- You are receiving this mail because: -------
You are the assignee for the bug.
You are on the CC list for the bug.

Wikibugs-l mailing list

Reply via email to