[FFmpeg-cvslog] r14956 - trunk/libavformat/matroskadec.c

[Mike Melanson]

Robert Swain wrote:
> 2008/8/26 Mike Melanson <mike at multimedia.cx>:
>> Robert Swain wrote:
>>>> One of the 3 revisions following 14755 broke WMA. According to an
>>>> earlier email, it was 14758, specifically.
>>> Is this only an issue on x86_64? I see no regressions on x86_32.
>> Check FATE. It also struck gcc 2.95.3 and icc on x86_32.
> 
> OK. Why is it 'one of the three revisions following r14755'? Does fate
> not build every revision?

If the commits are spaced far enough apart in time, sure, it builds each
revision. :) Click the history link next to a configuration to see which
revisions have actually been built. If 3 revisions are committed in as
many minutes, then FATE will pick up all 3 changes for the next
build/test cycle.

I brainstormed today about possibly noticing when a test broke and there
are multiple revisions when the break could have occurred, and then
scheduling to have the in-between builds tested when CPU time is
available. The mechanism might coincide with the oft-requested email
alert system I am scheming.

-- 
	-Mike Melanson