Re: "Invalid or undefined bitmap" when launched with -Q

[Peter Dyballa]

Am 14.08.2006 um 01:55 schrieb Nick Roberts:

> > I think today is not my day to have success:
> >
> > When entering static int load_pixmap (f, name, w_ptr, h_ptr):
> >
> >      Breakpoint 1, load_pixmap (f=0x9a8420, name=60552283,
> > w_ptr=0x9f4d54, h_ptr=0x9f4d58) at xfaces.c:1163
> >      (gdb) pr name
> >      No symbol "print_output_debug_flag" in current context.
> >      (gdb) xtype
> >      The history is empty.
>
> According to an earlier e-mail you compiled with -O2.  Maybe it's been
> optimised away as it only seems to be used on Windows.

Yes, there were indeed some details optimised away. So I re- 
configured and re-compiled with -O0.

> But I don't quite follow what Eli suggestd anyway.  I would have  
> thought you
> need to type
>
> (gdb) p name
> (gdb) pr
>
> or
>
> (gdb) p name
> (gdb) xpr
>
> To do the former I think you either need to recompile without -O2 or
> redefine pr:
>
> define pr
>   set debug_print ($)
> end
>
> but I think you can do the latter without changing anything.

This setting is already active in .gdbinit:

   60   # Set up something to print out s-expressions.
   61	# We save and restore print_output_debug_flag to prevent the  
w32 port
   62	# from calling OutputDebugString, which causes GDB to display  
each
   63   # character twice (yuk!).
   64   define pr
   65     set $output_debug = print_output_debug_flag
   66     set print_output_debug_flag = 0
   67     set debug_print ($)
   68     set print_output_debug_flag = $output_debug
   69   end
   70   document pr
   71   Print the emacs s-expression which is $.
   72   Works only when an inferior emacs is executing.
   73   end

Although this "inferior emacs" seems to be running – I see a frame,  
although without mode-line — the debug_print seems to come too early ...

	Breakpoint 1, load_pixmap (f=0x9a8420, name=60552283,  
w_ptr=0x9f4d54, h_ptr=0x9f4d58) at xfaces.c:1163
        (gdb) p name
        $1 = 60552283
        (gdb) xpr
        No enum type named Lisp_Type.
        (gdb) n
        
	Breakpoint 3, x_create_bitmap_from_file (f=0x9a8420, file=60552283)  
at image.c:486
        (gdb) p file
        $2 = 60552283
        (gdb) pr
        No symbol "print_output_debug_flag" in current context.

Or do I need to set it again and again in *gud-emacs* buffer?


When GNU Emacs starts in Gdb I also get informed:

        No struct type named Lisp_Symbol.


Gdb comes with Apple's Mac OS X:

	GNU gdb 6.3.50-20050815 (Apple version gdb-477) (Sun Apr 30 20:06:22  
GMT 2006)
        Copyright 2004 Free Software Foundation, Inc.
	GDB is free software, covered by the GNU General Public License, and  
you are
	welcome to change it and/or distribute copies of it under certain  
conditions.
        Type "show copying" to see the conditions.
	There is absolutely no warranty for GDB.  Type "show warranty" for  
details.
        This GDB was configured as "powerpc-apple-darwin".


I'm going to re-configure and re-compile without -dead_strip in  
LDFLAGS. Do I need a more up-to-date Gdb, too?

> Incidentally if I start "gdb emacs" with "r -Q", it doesn't hit  
> load_pixmap.

The default in my copy of GNU Emacs is:

        gdb --annotate=3 emacs

Then I get a report about the breakpoints and type run and cont and  
the first breakpoint is reached. My addition to .gdbinit is:

1066    set args -geometry 80x40+30+40
1067    set args -Q
1068    break load_pixmap
1069    break add_to_log
1070    break x_create_bitmap_from_file
1071    #break openp
1072    #set args -Q -nw --debug-init
1073    #break Fsqrt
1074    #break extract_float
1075    
1076    define weiter
1077      continue
1078      until 1209
1079      print fn
1080    end


--
Greetings

  Pete

To most people solutions mean finding the answers. But to chemists  
solutions
are things that are still all mixed up.