Re: [Texmacs-dev] Re: [E-devel] Segfault with Imlib2

[The Rasterman]

On Wed, 13 Jul 2005 12:42:20 +0200 Joris van der Hoeven <address@hidden>
babbled:

> Hi,
> 
> I am continuing trying to use Imlib2 in combination with
> a static TeXmacs binary.
> 
> I got a bit further when using dynamic linking at run time when
> using the -fPIC -rdynamic options (I did not need these in the case
> of freetype; I don't know why). When doing this, the software does
> no longer crash immediately, but no images show up.
> 
> I also tried static linking of the effectively used loader libraries
> together with libImlib2 (so everything should be statically linked
> in that case, right?). However, I still experience problems with
> the memory allocation in this situation. Notice that these problems
> occur more or less at random times: sometimes I have to load
> 50 pictures before an actual crash occurs.
> 
> For my static TeXmacs distribution, I really need at least dynamic
> linking at run time or static linking at compile time to work.
> Do you have any other ideas of what I might try? It is very well
> possible that I, or you, or we both just don't use the right
> compiler/linker options at a certain point. But I am running
> out of ideas...

i'm a bit baffled myself too. i'd need a chunk of time to dig into it and i'm
definitely short of that! :(

> Best wishes, Joris
> 
> -----------------------------------------------------------
> > > > #0  0x083c39ed in chunk_free (ar_ptr=0x84f0f68, p=0x8be3000) at
> > > > malloc.c:3231 #1  0x083c6311 in __libc_free (mem=0x8becfd8) at
> > > > malloc.c:3154 #2  0x0835c233 in __imlib_ConsumeImage (im=0x8bc68c0) at
> > > > image.c:176 #3  0x0835c505 in __imlib_CleanupImageCache () at
> > > > image.c:381 #4  0x0835d28a in __imlib_FreeImage (im=0x8bc68c0) at
> > > > image.c:1172 #5  0x0834824a in imlib_free_image () at api.c:1327
> > > > #6  0x08190818 in imlib2_image_size (u={rep = 0x8a58458}, 
> > > > address@hidden,
> > > > address@hidden) at ./Plugins/Imlib2/imlib2.cpp:123 #7  0x081901a8 in
> > > > image_size (image={rep = 0x8a58458}, address@hidden, address@hidden) at
> > > > ./System/Files/image_files.cpp:145 #8  0x0813d509 in
> > > > image_widget_rep::handle_repaint (this=0x8798678, ev=
> > > >         {<event_ptr_base> = {rep = 0x8858688}, <No data fields>}) at
> > > >         ./Window/Widget/Misc/file_chooser_widget.cpp:323
> > > > ...
> > > > 
> > > > The routine imlib2_image_size is mine. Here it is:
> > > > 
> > > > -----------------------------------------------------
> > > > void
> > > > imlib2_image_size (url u, int& w, int& h) {
> > > >   if (imlib2_size_cache->contains (u->t)) {
> > > >     w= as_int (imlib2_size_cache[u->t][0]);
> > > >     h= as_int (imlib2_size_cache[u->t][1]);
> > > >   }
> > > >   else {
> > > >     Imlib_Image image= imlib2_load_image (u);
> > > >     if (image) {
> > > >       IMLIB2_context_set_image (image);
> > > >       w= IMLIB2_image_get_width ();
> > > >       h= IMLIB2_image_get_height ();
> > > >       IMLIB2_free_image ();
> > > >       imlib2_size_cache (u->t)= tuple (as_string (w), as_string (h));
> > > >     }
> > > >   }
> > > > }
> > > > -----------------------------------------------------
> > > > 
> > > > The routines IMLIB2_x are just copies of the corresponding
> > > > routines imlib_x (so as to allow for dynamic linking).
> > > > For your knowledge, here is also the routine imlib2_load_image:
> > > > 
> > > > -----------------------------------------------------
> > > > static Imlib_Image
> > > > imlib2_load_image (url u) {
> > > >   url name= resolve (u);
> > > >   if (is_none (name))
> > > >     name= resolve ("$TEXMACS_PIXMAP_PATH/TeXmacs-gnu.xpm");
> > > >   char *_name= as_charp (as_string (name));
> > > >   Imlib_Image image;
> > > >   image = IMLIB2_load_image (_name);
> > > >   delete[] _name;
> > > >   return image;
> > > > }
> > > > -----------------------------------------------------
> > > > 
> > > > Am I doing something wrong with the memory management?
> > > > 
> > > > Thanks, Joris
> > 
> > 
> > _______________________________________________
> > Texmacs-dev mailing list
> > address@hidden
> > http://lists.gnu.org/mailman/listinfo/texmacs-dev
> 
> 
> -------------------------------------------------------
> This SF.Net email is sponsored by the 'Do More With Dual!' webinar happening
> July 14 at 8am PDT/11am EDT. We invite you to explore the latest in dual
> core and dual graphics technology at this free one hour event hosted by HP,
> AMD, and NVIDIA.  To register visit http://www.hp.com/go/dualwebinar
> _______________________________________________
> enlightenment-devel mailing list
> address@hidden
> https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
> 


-- 
------------- Codito, ergo sum - "I code, therefore I am" --------------
The Rasterman (Carsten Haitzler)    address@hidden
裸好多                              address@hidden
Tokyo, Japan (東京 日本)