From: Dennis Taylor <corbeau@execpc.com>
  To  : ggi-develop@eskimo.com
  Date: Sun, 14 Mar 1999 23:02:27 -0600

giiClose() coredump

        So I wrote a really simple test GGI2D program the other night
to see how GGI2D handles some weird cases in the even-odd
algorithm. It compiles and runs fine in the X target... but when it
exits, it coredumps on giiClose() with what looks like a problem in
X. I figure, given that I know pretty much zero about X programming,
that I should post it here and see if this will be transparent to
someone with a higher clue factor. Here's the source code of the
program (yeah, it's a bit messy, but it's just a test):


#include <stdio.h>
#include <ggi/ggi.h>
#include <ggi/ggi2d.h>

#define DEF_X      640
#define DEF_Y      480
#define DEF_DEPTH  GT_16BIT

ggi2d_coord diamond[ 4 ] = { { 100, 0 },
							 { 200, 100 },
							 { 100, 200 },
							 { 0, 100 }, };

int main( void ) {
  int err;
  ggi_mode mode;
  ggi_visual_t vis;
  ggi_color color;

  /* Set up the visual. */
  err = ggiInit();
  if (err)
	ggiPanic( "*** startup panic, can't initialize GGI (err %d)\n", err );
  vis = ggiOpen( NULL );
  if (vis == NULL)
	ggiPanic( "*** startup panic, can't open display\n" );
  if (ggiCheckGraphMode( vis, DEF_X, DEF_Y, DEF_X, DEF_Y, DEF_DEPTH, &mode )) {
	err = ggiSetMode( vis, &mode );
	fputs( "*** using system-suggested graphics mode.\n", stderr );
	ggiFPrintMode( stderr, &mode );
	fputc( '\n', stderr );
  } else {
	err = ggiSetGraphMode( vis, DEF_X, DEF_Y, DEF_X, DEF_Y, DEF_DEPTH );
  }
  if (err)
	ggiPanic( "*** startup panic, can't set graphics mode (err %d)\n", err );
  err = ggi2dInit();
  if (err)
	ggiPanic( "*** startup panic, can't initialize GGI2D (err %d)\n", err );
  err = ggi2dOpen( vis );
  if (err)
	ggiPanic( "*** startup panic, can't open GGI2D visual (err %d)\n", err );

  /* Fill the visual with white. */
  color.r = color.g = color.b = 0xFFFF;
  ggiSetGCBackground( vis, ggiMapColor( vis, &color ) );
  ggiSetGCForeground( vis, ggiMapColor( vis, &color ) );
  ggiFillscreen( vis );

  /* Draw the polygon in black. */
  ggiSetGCForeground( vis, (ggi_pixel) 0 );
  ggi2dFillPoly( vis, diamond, 4 );
  ggiFlush( vis );

  /* Wait for a keypress, then die. */
  ggiGetc( vis );
  ggi2dClose( vis );
  ggiClose( vis );
  ggi2dExit();
  ggiExit();

  return( 0 );
}


        It compiles fine with "gcc -Wall -pedantic -g -O -o 2dtest
2dtest.c -lggi -lggi2d". Here's the output with GGI_DEBUG=255 on....


	(It starts up, draws the screen fine, and then spews a whole lot
	 of do_blits while it idles for a couple seconds...)

LibGGI: GGI_X_flush(0x804c480, 0) called
LibGGI: _ggi_x_do_blit(0x804dce8, 0x0, 640x480) called
LibGGI: GGI_X_flush(0x804c480, 0) called
LibGGI: _ggi_x_do_blit(0x804dce8, 0x0, 640x480) called

	(Here's where I press any key to kill it.)

LibGII: _giiPollAll(0x804e138, 0xa0, 0xbffffa04) called
LibGII: GII_xwin_eventpoll(0x804e138) called
LibGII: GII_xwin_eventpoll: KeyPress
LibGII: _giiEvQueueAdd(0x804e138, 0xbffff750) called
LibGII: _giiEvQueueSetup() called
LibGII: _giiEvQueueSetup alloced 0x806a848
LibGII: Adding event type 5, size 36 at pos 0
LibGII: _giiEvQueueRelease(0x804e138, 0xbffffabc, 0xa0) called
LibGII: queue = 0x804e2b8, queue->queues = 0x804e2bc, queue->queues[queue] = (nil)
LibGII: queue = 0x804e2b8, queue->queues = 0x804e2bc, queue->queues[queue] = 0x804e6d8
LibGII: queue = 0x804e2b8, queue->queues = 0x804e2bc, queue->queues[queue] = (nil)
LibGII: queue = 0x804e2b8, queue->queues = 0x804e2bc, queue->queues[queue] = (nil)
LibGII: queue = 0x804e2b8, queue->queues = 0x804e2bc, queue->queues[queue] = (nil)
LibGII: queue = 0x804e2b8, queue->queues = 0x804e2bc, queue->queues[queue] = 0x806a848
LibGII: _giiEvQueueRelease: Plausible found.
LibGII: queue = 0x804e2b8, queue->queues = 0x804e2bc, queue->queues[queue] = 0x8066828
LibGII: queue = 0x804e2b8, queue->queues = 0x804e2bc, queue->queues[queue] = (nil)
LibGII: queue = 0x804e2b8, queue->queues = 0x804e2bc, queue->queues[queue] = (nil)
LibGII: queue = 0x804e2b8, queue->queues = 0x804e2bc, queue->queues[queue] = 0x8068838
LibGII: queue = 0x804e2b8, queue->queues = 0x804e2bc, queue->queues[queue] = (nil)
LibGII: queue = 0x804e2b8, queue->queues = 0x804e2bc, queue->queues[queue] = (nil)
LibGII: queue = 0x804e2b8, queue->queues = 0x804e2bc, queue->queues[queue] = (nil)
LibGII: queue = 0x804e2b8, queue->queues = 0x804e2bc, queue->queues[queue] = (nil)
LibGII: Retrieved event type 5, size 36.
LibGGI: ggiExtensionDetach(0x804c480, 0) called
LibGGI: ggiClose("0x804c480") called
LibGGI: ggiClose: closing
LibGGI: _ggiZapDL(0x804c480, 0x804bf24) called
LibGGI: Disposing "display-x"
LibGGI: _GGI_mansync_stop() (MANSYNC_CHILD) called.
LibGGI: _GGI_X_freedbs: XShmDetach(0x804cd60, 58720263)
LibGGI: X-target closed
LibGGI: Closing handle: 0x804c848
LibGGI: _ggiZapDL(0x804c480, 0x804c67c) called
LibGGI: Disposing "generic-color"
LibGGI: Closing handle: 0x8050b90
LibGGI: _ggiZapDL(0x804c480, 0x804c6e4) called
LibGGI: Disposing "generic-linear-16"
LibGGI: Closing handle: 0x8050e20
LibGGI: Disposing "generic-stubs"
LibGGI: Closing handle: 0x8050930
LibGGI: _ggiZapDL(0x804c480, 0x804c538) called
LibGGI: Disposing "generic-stubs-2d"
LibGGI: Closing handle: 0x8065ce0
LibGGI: Disposing "helper-mansync"
LibGGI: Closing handle: 0x804de88
LibGII: giiClose(0x804e138) called
LibGII: _giiEvQueueDestroy(0x804e138) called
LibGII: Desttroying 0x804e2b8, 0x804e2bc
LibGII: _giiEvQueueDestroy going 0, (nil)
LibGII: _giiEvQueueDestroy going 1, 0x804e6d8
LibGII: _giiEvQueueDestroy going 2, (nil)
LibGII: _giiEvQueueDestroy going 3, (nil)
LibGII: _giiEvQueueDestroy going 4, (nil)
LibGII: _giiEvQueueDestroy going 5, 0x806a848
LibGII: _giiEvQueueDestroy going 6, 0x8066828
LibGII: _giiEvQueueDestroy going 7, (nil)
LibGII: _giiEvQueueDestroy going 8, (nil)
LibGII: _giiEvQueueDestroy going 9, 0x8068838
LibGII: _giiEvQueueDestroy going 10, (nil)
LibGII: _giiEvQueueDestroy going 11, (nil)
LibGII: _giiEvQueueDestroy going 12, (nil)
LibGII: _giiEvQueueDestroy going 13, (nil)
LibGII: _giiEvQueueDestroy done
Segmentation fault (core dumped)


        Ouch. So I look at it in gdb, naturally.


Program received signal SIGSEGV, Segmentation fault.
0x40004122 in _dl_lookup_symbol (undef_name=0x4010937b "_XimLocalIMFree",
    ref=0xbffffb00, symbol_scope=0x40009fc8, reference_name=0x804c9a0
    "/usr/X11R6/lib/libX11.so.6", flags=2)
    at dl-lookup.c:70
dl-lookup.c:70: No such file or directory.
(gdb) bt
#0  0x40004122 in _dl_lookup_symbol (undef_name=0x4010937b "_XimLocalIMFree",
    ref=0xbffffb00, symbol_scope=0x40009fc8, reference_name=0x804c9a0
    "/usr/X11R6/lib/libX11.so.6", flags=2)
    at dl-lookup.c:70
#1  0x40005711 in fixup (l=0x804c9d0, reloc_offset=134531536) at
    ../sysdeps/i386/dl-machine.h:252
#2  0x4000584f in _dl_runtime_resolve () at dl-runtime.c:138
#3  0x4014191d in XCloseIM ()
#4  0x401a34c3 in GII_xwin_close (inp=0x804e138) at input.c:386
#5  0x400c451b in giiClose (inp=0x804e138) at gii.c:723
#6  0x40017263 in _ggiDestroyVisual (vis=0x804c480) at visual.c:280
#7  0x40013b91 in ggiClose (visual=0x804c480) at init.c:387
#8  0x8048b17 in main () at 2dtest.c:60
(gdb)


        This is XF86 3.3.3 on a 2.0.36 Linux kernel, running the March
14th snapshot of GGI, which I just compiled tonight. I'm a little
boggled. Is there anything blatantly wrong about this situation, or is
this a GII bug, or an X bug, or... what? I'll poke around it more
tomorrow, but I thought I'd throw it out here in case anyone who knows
this stuff better could spot the problem.

        Hope this helps...


__________________________________________________________________________
Dennis Taylor           "Anyone whose days are all the same and free from
corbeau@execpc.com       want inhabits eternity of a sort."  - Peter Hoeg
__________________________________________________________________________
   PGP Fingerprint: E8D6 9670 4FBD EEC3 6C6B  810B 2B30 E529 51BD 7B90