On Thu, 22 Jul 2004 06:02, Christos Zoulas wrote:
> >>  % matlab
> >>  ??? MATLAB was unable to open a pseudo-tty: Invalid argument [22,1]
> >
> >Hmm, I did see that with matlab 6.5 R12.  In R13 they apparently stopped
> >using /dev/ptmx (I think), which allowed NetBSD to do the pty
> >allocation.  Now we have ptmx usage back, eh?
>
> Yes, but our ptmx is not compatible with the linux one [I am guessing].

I suspect so.  Here is the relevant bit of the trace:

  1300 MATLAB   CALL  open(0x4827d96c,2,0x48297184)
  1300 MATLAB   NAMI  "/emul/linux/dev/ptmx"
  1300 MATLAB   NAMI  "/dev/ptmx"
  1300 MATLAB   NAMI  "/dev/ptyp8"
  1300 MATLAB   RET   open 3
  1300 MATLAB   CALL  ioctl(3,_IO('T',0x1,0),0xbfff9bdc)
  1300 MATLAB   RET   ioctl 0
  1300 MATLAB   CALL  ioctl(3,_IOW('T',0x30,0x4),0xbfff9c9c)
  1300 MATLAB   RET   ioctl -1 errno -22 Invalid argument
  1300 MATLAB   CALL  fstat64(3,0xbfff9cc0)
  1300 MATLAB   RET   fstat64 0
  1300 MATLAB   CALL  fcntl64(3,1,0x16)
  1300 MATLAB   RET   fcntl64 0
  1300 MATLAB   CALL  rt_sigaction(0x11,0xbfffaa90,0xbfffaa00,8)
  1300 MATLAB   RET   rt_sigaction 0
  1300 MATLAB   CALL  rt_sigprocmask(1,0xbfffac00,0,8)
  1300 MATLAB   RET   rt_sigprocmask 0
  1300 MATLAB   CALL  fstat64(1,0xbfff46e4)
  1300 MATLAB   RET   fstat64 0
  1300 MATLAB   CALL  ioctl(1,_IO('T',0x1,0),0xbfff4638)
  1300 MATLAB   RET   ioctl 0
  1300 MATLAB   CALL  old_mmap(0xbfff46cc)
  1300 MATLAB   RET   old_mmap 1255079936/0x4acf0000
  1300 MATLAB   CALL  write(1,0x4acf0000,0x44)
  1300 MATLAB   GIO   fd 1 wrote 68 bytes
       "??? MATLAB was unable to open a pseudo-tty: Invalid argument [22,1]

cheers
mark