From: Grzegorz A. H. <gr...@te...> - 2002-01-03 17:53:00
|
On Thu, Jan 03, 2002 at 01:03:26AM -0500, Alberto Griggio wrote: > The same happens to me too. I think you know much better than me the > details of Allegro internals, so if you say this is a modules issue, I'll > look in that direction: anyway, I switched from Allegro 3.9.36 (without > modules) to 4.0.0 only a couple of weeks ago, so I didn't notice this > until last week. I am not that expert about Allegro internals, just happen to have dived into the waters a few times for such things like the fbcon driver. As for determining what happens with pyallegro that it doesn't load it's modules, you only need to have one module, it doesn't matter which one, to know if it get's loaded or not. I have for example modified allegro/src/linux/fbcon.c, around line 956 in function _module_init to make the following first call: printf("Entering fbcon module initialization\n"); It would be much "cleaner" to use AL_TRACE, but that one only get's included in debug versions and I haven't bothered yet to play with pyallegro and debug allegro versions. Once you have an installed module with that, you can go to X and run the C version of exhello. While the window pops up you should see at least once that printf appearing in the xterm. This is because each module get's registered before calling it's module dependant set_gfx_mode funcion. Now, if you repeat that with exhello.py, you won't see any printf message. That's as far as I have reached yet, because I still don't know how the dynamic loader works with paths and such things. I will tell you if I find out, of course. -- Grzegorz Adam Hankiewicz gr...@te... http://gradha.infierno.org/ |