Re: [dvd::rip] Re: Debugging the transcode freeze problem

[Davide Capodaglio <davidecapod@xxxxxxxxx>]

No change even with PERLIO=stdio...

This is what happens while transcode is working properly

ghost     6819  6409 10 22:44 pts/2    00:00:09       /usr/bin/perl 
/usr/bin/dvdrip
ghost     6945  6819  0 22:46 pts/2    00:00:00         sh -c mkdir -m 
0775 /tmp/dvdrip6819.ppm; cd /tmp/dvdrip6819.ppm; execflow transcode  -H
ghost     6947  6945  1 22:46 pts/2    00:00:00           transcode -H 
10 --print_status 25 -o snapshot -y ppm,null -T 1,-1,1 -x dvd,null -i /m
ghost     6958  6947  0 22:46 pts/2    00:00:00             sh -c tccat 
-T 1,-1,1 -i "/media/data/dvd/THE_SHINING" -t dvd -d 0 | tcdemux -s 0x8
ghost     6959  6958  4 22:46 pts/2    00:00:00               tccat -T 
1,-1,1 -i /media/data/dvd/THE_SHINING -t dvd -d 0
ghost     6960  6958  1 22:46 pts/2    00:00:00               tcdemux -s 
0x80 -x mpeg2 -S 0 -M 1 -d 0
ghost     6961  6958  0 22:46 pts/2    00:00:00               tcextract 
-t vob -a 0 -x mpeg2 -d 0
ghost     6962  6958 12 22:46 pts/2    00:00:00               tcdecode 
-x mpeg2 -d 0 -y yv12

and when transcode ends but tccat goes on streaming:

ghost     6819  6409  9 22:44 pts/2    00:00:09       /usr/bin/perl 
/usr/bin/dvdrip
ghost     6945  6819  0 22:46 pts/2    00:00:00         sh -c mkdir -m 
0775 /tmp/dvdrip6819.ppm; cd /tmp/dvdrip6819.ppm; execflow transcode  -H
ghost     6947  6945  2 22:46 pts/2    00:00:00           transcode -H 
10 --print_status 25 -o snapshot -y ppm,null -T 1,-1,1 -x dvd,null -i /m
ghost     6958  6947  0 22:46 pts/2    00:00:00             sh -c tccat 
-T 1,-1,1 -i "/media/data/dvd/THE_SHINING" -t dvd -d 0 | tcdemux -s 0x8
ghost     6959  6958 10 22:46 pts/2    00:00:01               tccat -T 
1,-1,1 -i /media/data/dvd/THE_SHINING -t dvd -d 0

As you can see tcdemux is no longer running (and so the pipe is broken) 
but tccat does not end and goes on reading the DVD.
Bye
Davide


Jörn Reder ha scritto:
> Davide Capodaglio wrote:
>
>   
> > I did the two tests, as usual 0.52 works while 0.98 stops.
> > I attached the two projects and the two logs, although they don't tell 
> > me so much, hope they are more useful for you...
> >     
>
> I'm not sure ;) The executed commands are nearly identical, only that 
> 0.98.x adds "--print_status 25" to all transcode commands to reduce 
> progress reporting overhead a bit. But since the transcode command 
> itself works (when executed on a shell), I think we don't need to pay 
> much attention to this.
>
> Another, more interesting, difference is that dvd::rip 0.52 sets a magic
> environment variable to workaround an old/strange Perl bug:
>
>   PERLIO=stdio
>
> Although I can't believe that this helps, please set this variable 
> before starting dvd::rip 0.98 and try your luck:
>
>   export PERLIO=stdio
>   dvdrip
>
> If it still doesn't work, I have to think about what we can do next. The
> good thing is, we can reproduce the bug reliably, which is the most 
> important part of the bug fixing process ;)
>
> Regards,
>
> Jörn
>
>