ok.
On today's build,
[root@Peppy lame]# date
Wed Feb 16 06:04:56 EST 2000
[root@Peppy lame]#
if one compiles with
CC_OPTS = -UNDEBUG -O -Wall -g -DABORTFP
lame_current behaves.
If you do
# these options were suggested with egcs-990524
# CC = egcs #egcc for Debian systems
# CC_OPTS = -O9 -fomit-frame-pointer -march=pentium \
# -ffast-math -funroll-loops
then the current lame seg faults. I have not figured out how to gather
more info because it is in a pipe, like this.
mrec 239.1.2.3/24431|./lame -b 96 -m s -x -a -h - -|msend
239.1.110.1/24434
Thoughts?
_J
[root@Peppy lame]# mrec 239.1.2.3/24431|./lame -b 96 -m s -x -a -h -
-|mgroupsock write failed:
send 239.1.110.1/24434
WARNING packet -1001948823 out of order. old value 1 (diff -1001948824)
Assuming raw pcm input file : Forcing byte-swapping
LAME version 3.63 (www.sulaco.org/mp3)
GPSYCHO: GPL psycho-acoustic and noise shaping model version 0.76.
Autoconverting from stereo to mono. Setting encoding to mono mode.
Encoding stdin to stdout
Encoding as 44.1 kHz 96 kbps single-ch MPEG1 LayerIII ( 7.3x) qual=2
ERROR read():: Success
bufsize is 0Broken pipe
[root@Peppy lame]# echo $?
3
[root@Peppy lame]# ls core
ls: core: No such file or directory
[root@Peppy lame]#
In the new year, Mark Taylor wrote:
>
> >
> > The RTP code still does not seem to be generating any packets, so I will
> > still make lc generate the RTP versions, but if I can figure out what is
> > broken with the rtp thingie, I'll update that and then cross the bridge on
> > how to get my changes back into the repository. Is there a way to build a
> > debug version of lame so that assertions are in the code? if I use the
> > debug options in the Makefile?
> >
> > _J
> >
> Yes, that should work. All you need is -UNDEBUG to enable the
> assertions. (i.e. do not define NDEBUG).
>
>
>
>
> --
> MP3 ENCODER mailing list ( http://geek.rcc.se/mp3encoder/ )
>
--
MP3 ENCODER mailing list ( http://geek.rcc.se/mp3encoder/ )
