SDL2

@Raziel

Thanks for testing.

@All

RC5 is available: https://github.com/AmigaPorts/SDL-2.0/ ... /tag/v2.0.20-rc5-amigaos4

Changes:

Quote:

Please relink your apps and test.

@Capehill
Thank you so much for your work. I will test it tonight and give you feedback

@Capehill

Thank you very much!

@Capehill
I just tested both issues and seem to be fine and working as they should.

Thanl you so much for the fixes.

@walkero

Thanks, hopefully this means we are close to the "final" 2.0.20 release :)

@Capehill
Sorry if it will be not related to SDL2, but at least i have an issue found when tried to check one of SDL2's examples.

The example I talked about is that one :
https://github.com/AmigaPorts/SDL-2.0/blob/main/test/testshader.c

That a test case that uses ARB shaders. I do not know if they work with MiniGL, but probably not?

So i tested with GL4ES, and find out the strange issue, which i remember to have in some other, bigger, SDL-related projects). And it wasn't with ARB exactly, it was something different.

Now problem is, if i compile that test case, it didn't work as it. Function "static SDL_bool InitShaders()" fails with saying "unable to compile shaders!" (while i tested those shaders, they of course compiles/executes fine by ogles2/warp3dnova). Error come exactly from the InitShaders() function, from that part:



Now, what is interesting, is that i enough to add printf("asdfasdf") or IDOS->Delay(1) to any place before that code block to make everything works as expected.

And I do not understand what it that. I tried to compile it without optimization at all as BSzili suggests: the same problem. All that i understand, for now, is that when we build with and without _any_ stub call, we have: 1) code block a bit differs by size in .text segment 2) we add some ms of "time wait" difference when with have/didn't have call to any function, and adding one, add some "delay", which by some reason help us.

ptitSeb suggests that maybe the way the function pointer is gathered is the issue?

Maybe it's something about SDL_GL_GetProcAddress() being not too fast / not checking something?

@kas1e

No ideas at the moment, sorry. Could you provide a debug version of this?

Or does serial debug change the behaviour already?

EDIT:

Unless I'm mistaken, code is branching to gl4es here:

https://github.com/kas1e/SDL2_GL4ES/bl ... igaos4/SDL_os4gl4es.c#L53

(By the way, it might make sense to change those agl* functions to gl4es* or similar, to avoid confusion with similarly named OGLES2-functions which have different parameters.)

Then, code is executed about here (?): https://github.com/ptitSeb/gl4es/blob/ ... ed83/src/agl/lookup.c#L25

and finally

https://github.com/ptitSeb/gl4es/blob/ ... da/src/gl/gl_lookup.c#L37

?

Quote:

Not supported by MiniGL, AFAIK this is only supported by gl4es over OGLES2.

@Capehill

Are there something else to do before releasing the public version of SDL 1.2.16 ?

@samo79

Off-topic ;) However, since there won't be official 1.2.16 from SDL team ( see README at https://github.com/libsdl-org/SDL-1.2 ), maybe we can have our "unofficial" 1.2.16 release anyhow.

@Capehill
+1 for unnoficial 1.2.16 :)

Quote:

As far as i read, the shaders langauge were introduced in 2002 with those extensions like ARB_vertex_program and ARB_fragment_program for OpenGL 1.4, and then in 2006 was transfered to OpenGL directly.

Our MiniGL is ogl 1.3, so we out of it. I was hope that maybe still, as we have some functions in our MiniGL which is not part of 1.3, but more, but then i checked the extensions we had for MiniGL, and yeah, there not a lot, only about arays, textures,etc mostly, but nothing about ARB shaders.


Quote:

Yeah, that right, but check other topic about "inserting printf blabbla", we probabaly almost find the roots.

@Capehill

Sure, just maybe merge it according to the latest changes (29 Jan 2022) and then if no problem occur release it

@Capehill

I can see that 2.20.0 went live, is that correct?

@kas1e

Could you please compile a new release of the gl4es sdl stuff?
I am in need of SDL_SetWindowMouseRect which is missing from the available SDL2_gl4es.

Thank you both

@Raziel

2.0.20 was released recently, yes.

http://os4depot.net/?function=showfile&file=library/misc/sdl2.lha

@Raziel
Quote:

Yeah of course hope to do so in the next few days. Just overflowed with all other amiga stuff.

@kas1e

Thank you

@Capehill

I don't want to be annoying, but did you see https://github.com/AmigaPorts/SDL/issues/195
?

Any ideas?

Thanks a lot

@Raziel

Yep, saw an e-mail but didn't have time to react yet.

After a very quick check, we don't have a mapping for numlock: https://github.com/AmigaPorts/SDL/blob ... /events/scancodes_amiga.h

I don't find any info about numlock:

https://wiki.amigaos.net/wiki/Keymap_Library

Need more study.

@Capehill

/me hugs @capehill

@Capehill

Quick question...c(w)ould SDL apps benefit from VA (Hardware Video Acceleration)?

@Capehill
@Raziel

It looks like the numlock key does not generate a keycode on the Amiga. At least, when I run the Input Event Sniffer (http://os4depot.net/?function=showfil ... isc/inputeventsniffer.lha) and watch for RAWKEY and EXTENDEDRAWKEY events nothing happens when I press numlock.