Struggling with CVS version of dia

[Martin Cracauer <cracauer cons org>]

Hi,

first of all, thanks for one of the few newer OpenSource projects
which really impressed me.  Nice work, congrats!

I have the following problems using the CVS version of dia.  Versions
tried are of today, and from a few weeks ago.

Please note that I don't know whom I am facing:
- maybe the CVS tree is just used by one person at this time
- maybe the autoconf stuff is brokenness on my side, maybe the
  developer(s) don't care auto the autoconf stuff, or maybe they could
  need help 

So, just let me know what you expect me to do, not the otehr way round
:-)

%%

The name of the perl interpreter in xml-i18n-merge doesn't get
substituted by configure (the file gets generated, though), I assume
this is an autoconf problem, but can't figure where to get the right
macro files

%%

In
  app/menus.c
I need 
  #include <gdk/gdkkeysyms.h>

%%

The top-level Makefile tries to call ./po/update.sh, which aborts
saying it is deactivated.

%%

The documentation Makefiles try to install *.html and *.css even though
those are not built by the default make target.  I figure they should
depend on a install-<something> target instead.

%%

After patching these things I run into a segfault.  Backtrace follows,
note the bad address for the accelerator.  Does this look familiar to
someone? 

(gdb) r 
Starting program: /usr/local/bin/dia 

Program received signal SIGSEGV, Segmentation fault.
0x4020e33b in gtk_accelerator_parse (
    accelerator=0xffbe <Address 0xffbe out of bounds>, 
    accelerator_key=0xbffff5e4, accelerator_mods=0xbffff5e8)
    at gtkaccelgroup.c:855
855     gtkaccelgroup.c: No such file or directory.
(gdb) bt
#0  0x4020e33b in gtk_accelerator_parse (
    accelerator=0xffbe <Address 0xffbe out of bounds>, 
    accelerator_key=0xbffff5e4, accelerator_mods=0xbffff5e8)
    at gtkaccelgroup.c:855
#1  0x4026a995 in gtk_item_factory_add_item (ifactory=0x81c6ab8, 
    path=0x81df800 "/Help/Manual", 
    accelerator=0xffbe <Address 0xffbe out of bounds>, 
    callback=0x806f3a8 <help_manual_callback>, callback_action=0, 
    callback_data=0x0, callback_type=1, item_type=0x402f1ac0 "<Item>",

    widget=0x81df2b4) at gtkitemfactory.c:537
#2  0x4026c161 in gtk_item_factory_create_item (ifactory=0x81c6ab8, 
    entry=0x80df518, callback_data=0x0, callback_type=1)
    at gtkitemfactory.c:1162
#3  0x4026b5ea in gtk_item_factory_create_items_ac (ifactory=0x81c6ab8, 
    n_entries=17, entries=0x80df400, callback_data=0x0, callback_type=1)
    at gtkitemfactory.c:838
#4  0x4026b4f9 in gtk_item_factory_create_items (ifactory=0x81c6ab8, 
    n_entries=17, entries=0x80df400, callback_data=0x0) at gtkitemfactory.c:816
#5  0x8084037 in menus_init ()
#6  0x8084248 in menus_get_toolbox_menubar ()
#7  0x8080fd1 in create_toolbox ()
#8  0x8070ffb in app_init ()
#9  0x809a2b2 in main ()

%%

Versions:

gtk-config: 1.2.10
glib-config: 1.2.10
autoconf: autoconf (GNU Autoconf) 2.52
automake: automake (GNU automake) 1.5
libtool: ltmain.sh (GNU libtool) 1.4.2 (1.922.2.53 2001/09/11 03:18:52)
gcc: 2.95.2
i686 Linux 2.4.6
libc-2.1.3.so

%%

I don't really need a newer version that 0.88.1, however I want to
build a plugin package and would prefer to test it on both release and
CVS version.

So, if you say the API will not change, that is fine for me also.

I also have bunch of questions regarding module making, but they're
not ripe yet and I figured I get going first.

Thanks!