Re: Insight 6.4 GUI does not start on Mac OS X

[Hermann-Simon Lichte <hermann dot lichte at gmail dot com>]

Hello Keith,

It's not the startup code on OS X that is giving me trouble. I reckon  
that a memcpy() in xstrdup() fails. I did some single stepping and  
this is how it looks like:

[snip] (we're in gdb's captured_main() now)

241	  interpreter_p = xstrdup (context->interpreter_p);
---xstrdup (s=0x315af0 "insight") at ../../insight-6.4/libiberty/ 
xstrdup.c:33
33	  register size_t len = strlen (s) + 1;
34	  register char *ret = XNEWVEC (char, len);
---xmalloc (size=8) at ../../insight-6.4/gdb/utils.c:973
973	  if (size == 0)
976	  val = malloc (size);		/* OK: malloc */
977	  if (val == NULL)
981	}
---xstrdup (s=0x315af0 "insight") at ../../insight-6.4/libiberty/ 
xstrdup.c:36
36	}
---xstrdup (s=0x315af0 "insight") at ../../insight-6.4/libiberty/ 
xstrdup.c:35
35	  return (char *) memcpy (ret, s, len);
36	}
35	  return (char *) memcpy (ret, s, len);
---exceptions_state_mc_action_iter () at ../../insight-6.4/gdb/ 
exceptions.c:203
203	  return exceptions_state_mc (CATCH_ITER);
---exceptions_state_mc (action=CATCH_ITER) at ../../insight-6.4/gdb/ 
exceptions.c:126
126	  switch (current_catcher->state)
173	      switch (action)
177		    struct gdb_exception exception = *current_catcher->exception;
178		    if (current_catcher->mask & RETURN_MASK (exception.reason))
177		    struct gdb_exception exception = *current_catcher->exception;
178		    if (current_catcher->mask & RETURN_MASK (exception.reason))
183			catcher_pop ();
---catcher_pop () at ../../insight-6.4/gdb/exceptions.c:107
107	  struct catcher *old_catcher = current_catcher;
108	  current_catcher = old_catcher->prev;
113	  restore_cleanups (old_catcher->saved_cleanup_chain);
---restore_cleanups (chain=0x0) at ../../insight-6.4/gdb/utils.c:416
416	  restore_my_cleanups (&cleanup_chain, chain);
---restore_my_cleanups (pmy_chain=0x3edbd8, chain=0x0) at ../../ 
insight-6.4/gdb/utils.c:428
428	  *pmy_chain = chain;
429	}
---catcher_pop () at ../../insight-6.4/gdb/exceptions.c:115
115	  uiout = old_catcher->saved_uiout;
118	}
---catcher_pop () at ../../insight-6.4/gdb/exceptions.c:115
115	  uiout = old_catcher->saved_uiout;
117	  xfree (old_catcher);
115	  uiout = old_catcher->saved_uiout;
118	}
117	  xfree (old_catcher);
---xfree (ptr=0x2804600) at ../../insight-6.4/gdb/utils.c:1032
1032	  if (ptr != NULL)
1033	    free (ptr);		/* OK: free */

Program exited with code 01

The program exits within the first 400 instructions after  
captured_main() has been called. This stuff might be more related to  
the gdb folks. Do you have any quick pointer at the moment?

Cheers,
Hermann

> > Ouch. You're not even in insight's code yet. I hope you have the  
> > source available. ;-)
>
> God bless www.opensource.apple.com!
>
> > Yeah, you just hit a wall with me: I'm afraid I probably won't be  
> > much help from here on out, but I'm willing to help any way I can  
> > if you keep the list updated with your efforts.
>
> Line 272 of crt.c just equals exit(main(argc, argv, envp, apple));
>
> But main() does not appear on the stack frame.
>
> I quote: "_start() is called from the machine dependent assembly  
> entry point "start:" . It takes care of setting up the stack so 'C'  
> routines can be called and
> passes argc, argv and envp to here."
>
> This is no particularly sophisticated code. So I am surprised that  
> it fails. What I am going to do now is to single step this baby and  
> see what exactly is going wrong there.
>
> Cheers,
> Hermann