Jörg Sommer <[EMAIL PROTECTED]> wrote: ... > I know this is a misusage of cat, but cat shouldn't die with SIGSEG > > $ cat audio.wav > /dev/dsp > Segmentation fault > $ ltrace cat audio.wav > /dev/dsp ... > malloc(4096) = 0x10015080 > read(3, ",\003", 671089218) = 4096 > write(1, "RIFF\364\036l\002WAVEfmt \020", 4096 <unfinished ...> > +++ killed by SIGSEGV +++
Thanks for the report. It looks like it malloc'd a 4096-byte buffer, filled it via read(), and then tried to write() it to standard output. I don't see how redirecting to /dev/dsp could cause *cat* to misbehave. Maybe there's something wrong with your audio device driver? Can you investigate further? If using strace shows the same symptom (write evokes segv) and `valgrind cat audio.wav > /dev/dsp' reports nothing suspicious, then I'd try using a different kernel sound set-up.
