[sane-devel] mustek_pp trouble (was: saned doesn't release the parallel port when finished)

[edg1 at tiscali.be]

Op 26-12-03 16:36:05 schreef Henning Meier-Geinitz:
> So either the mustek_pp backend is in an endless loop here or it
> exited for some reason. If it were in an endless loop, saned wouldn't
> exit (but it does, doesn't it?). You don't get a segmantation fault
> after saned has exited?
> 
> Ok, let's look at the code. I'm a bit surprised about the name of  
> that function (cis_config_ccd) :-)

Yes, I know, it's a poor name choice :-)

> So the exit/freeze/whatever happens in cis_config_ccd or
> cis_return_home.

Or in sanei_pa4s2 or libieee1284.
 
> Maybe Jochen Eisinger or Eddy De Greef (the authors of that backend)
> have any clues on what's going on?

I have no idea. I could add additional debug print statements
to the driver around the point where things seem to get stuck,
but IIRC, Jim cannot compile the backend.

There is one other option: running saned in a debugger.
Jim, can you install gdb on your system? If so, you
can try the following:

  * As root, enter
     SANE_DEBUG_MUSTEK_PP=255 gdb saned
  * At the gdb prompt, enter
     r -d128
  * In another terminal window, enter
     scanimage -d net:localhost:mustek_pp:Mustek-1200CP > /dev/null
  * When the debug output in the gdb window stops and saned hangs,
    hit "CTRL-C" in the gdb window.
  * Now enter "bt" at the gdb prompt. This will print a stack trace.
  * You can quit gdb with "q".

I'd like to see the stack trace and the last 50 lines or so of the
debug output.

Eddy