[packman] regression: handbrake-unstable stops working (x264)

[Stefan Seyfried]

Hi Dominik,

Am 16.11.2014 um 13:39 schrieb Dominik Kopp:
> Hi,
> 
> handbrake-unstable-gtk 0.9.8 was working fine (I'm using as video encoder 
> x264) but after update to 0.9.9 some days/weeks ago it stopped working.
> It suffers now the same problem as the "stable" version handbrake-gtk has 
> for more than one year. (reported 24.06.2013)
> The result is now: both (handbrake-unstable-gtk and handbrake-gtk) versions 
> are failing(segfault?) when the video encoder x264 is used.
> 
> Can you please replace handbrake-unstable-gtk-0.9.9.9 with the older 0.9.8 
> version?

No (IMO). Handbrake is reaching the 0.10 release and it is certainly a
good idea to give it some testing before, and being "unstable" is what
"handbrake-unstable" in the name of the package says :-)

If you want an "handbrake-oldstable" version, go ahead and fork the
package. But I think it would be better to fix the issues you are seeing.

> regards
> Dominik
> 
> PS: explame of crash:
> HandBrakeCLI -i /home/dom/video/test2/1min.mpg  -o output.m4v -e x264
> ...
> [13:37:22] encx264: encoding at average bitrate 1000
> x264 [info]: using cpu capabilities: MMX2 SSE2Fast SSSE3 SSE4.2 AVX AVX2 
> FMA3 LZCNT BMI2
> [13:37:22] mpeg2video: "Chapter 1" (1) at frame 0 time 4666
> x264 [info]: profile High, level 3.0
> [13:37:22] sync: first pts is 4666
> Encoding: task 1 of 1, 52.07 %HandBrakeCLI: libavcodec/mpegvideo.c:353: 
> mpeg_er_decode_mb: Assertion `ref == 0' failed.
> Abgebrochen

This is not a segfault, this is an failed assertion. Have you reported
this to upstream handbrake? Especially if it is also happening in
released version 0.9.9?
Is the test video available somewhere (for reproducing the issue)?
-- 
Stefan Seyfried

"For a successful technology, reality must take precedence over
 public relations, for nature cannot be fooled." -- Richard Feynman