Re: [AD] Some new Annie/Unicode stuff |
[ Thread Index |
Date Index
| More lists.liballeg.org/allegro-developers Archives
]
On 26 Oct 2001, Annie wrote:
>
> *** From Allegro todo:
>
> > Wishlist:
> > - make allegro not use uconvert static buffer; reserve for user
>
> It's done, it seems to work...
Sorry to tell you this, but I'd already done that in CVS.
I'll (try to) see if you found things I missed.
> *** Another patch, for fbcon: convert the strings when needed...
Committed.
> *** Other problems:
> * Under XWindow, this code hangs:
> #include <allegro.h>
> int main(void)
> {
> allegro_init();
> return 0;
> }
> END_OF_MAIN()
> Seems to hang inside _xwin_destroy_window and seems to be related to pthread
> The ps command shows that there are 2 processes.
I can't reproduce it. It could be a race condition.
> * Under the console, this code initialize correctly fbcon:
...
> But this one does not:
...
> And quits with:
> Fatal error: unable to set GFX_SAFE <--- printed by set_gfx_mode
> Not a VGA mode-X resolution <--- allegro_error
Fixed in CVS.
diff -u -r1.3 umodules.c
--- src/unix/umodules.c 2001/10/04 05:52:39 1.3
+++ src/unix/umodules.c 2001/10/26 22:53:30
@@ -99,7 +99,7 @@
filename = buf3;
}
- if (!exists(filename))
+ if (!exists(uconvert_ascii(filename, buf)))
continue;
handle = dlopen(filename, RTLD_NOW);
> * With this program:
> #include <allegro.h>
> int main(void)
> {
> allegro_init();
> install_keyboard();
> readkey();
> return 0;
> }
> END_OF_MAIN()
> Using Ctrl-Alt-End stops only one of the 3 processes/threads.
A long standing known bug (it's now in the todo.txt file). In case
someone wants to fix it, I'm fairly sure it's due to the interaction
between threads and signals. A quick fix would be to change:
raise(SIGTERM);
in lkeybd.c to:
exit(EXIT_FAILURE);
(or EXIT_SUCCESS, as the X port is doing).