Crash

[Stephen J. Turnbull]

>>>>> "Valdis" == Valdis Kletnieks <Valdis.Kletnieks at vt.edu> writes:

    Valdis> Here's a (totally untested) of what I mean..  Yes, the X
    Valdis> code innards are ugly - but if we're going to assert(), we
    Valdis> should at least give the poor user a better hint about the
    Valdis> problem.....

Uh, if _we_ have a clue, we shouldn't be asserting at all.  But I
don't think we have one: isn't the _first_ thing we should do to tell
the user which display is having problems?  Uh, right ... we've
apparently successfully gotten past that first call to XOpenDisplay,
so Xlib should know but won't tell us---we can't trust it.

I guess it would be reasonable to just say what we do know and abort:

if (!d)
  {
    stderr_out("\n%s: Can't handle X error:\n",
    	(STRINGP (Vinvocation_name) ?
               /* #### this file needs definite Mule review ... */
               (char *) XSTRING_DATA (Vinvocation_name) : "xemacs"));
    stderr_out("Xlib reports NULL instead of a connection to an X server.  Aborting!\n",
    abort ();
  }

but this isn't much more help to the user.  Maybe "Xlib is a gibbering
psychotic mess; XEmacs is bailing out, grab your 'chute!"  ;-)

-- 
Institute of Policy and Planning Sciences     http://turnbull.sk.tsukuba.ac.jp
University of Tsukuba                    Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
               Ask not how you can "do" free software business;
              ask what your business can "do for" free software.