Login
Username:

Password:

Remember me



Lost Password?

Register now!

Sections

Who's Online
109 user(s) are online (93 user(s) are browsing Forums)

Members: 1
Guests: 108

cobale, more...

Support us!

Headlines

Recent files - OS4Depot - Your one stop for AmigaOS4 files





« 1 (2) 3 4 5 ... 9 »


Re: The MiniGL thread
Posted on: 2019/3/17 18:53   #21
Home away from home
Joined:
2006/12/2 3:55
From Italy, Perugia
Posts: 3339
@amig_os

Oh my.. this is the most nicer off topic i've seen
I've bought this game years ago but i was never
Able to play it under os4!
   Report Go to top

Re: The MiniGL thread
Posted on: 2019/3/17 18:58   #22
Just popping in
Joined:
2016/3/10 14:15
From Poland
Posts: 53
@samo79

Customers who*** already bought the BOXED VERSION *** will be able to download the AmigaOS 4.1 Final Edition version *** FREE *** of charge after registration of their copy.

PS.
I have a BOX version but I bought it again.
   Report Go to top

Re: The MiniGL thread
Posted on: 2019/3/17 19:02   #23
Quite a regular
Joined:
2013/10/17 16:21
From Hungary
Posts: 635
@kas1e
If libtxc_dxtn was a false alarm, then I definitely recommend it over S2TC, since it produces much better looking textures. It should be well tested, since it's part of Mesa. (Technically it was an external library because of licensing issues, but you get the idea).
_________________
I see the jimmies have been rustled.
   Report Go to top

Re: The MiniGL thread
Posted on: 2019/3/17 19:03   #24
Not too shy to talk
Joined:
2015/6/11 9:51
From Cologne
Posts: 406
And.... here we are off topic again
Give Shogo what it deserves -> at least its own dedicated thread

@BSzili
Yes, looks like it was false alarm and the true culprit was that extenstion string. So yes, let's go for the new dxt lib!
_________________
[Facebook] [YouTube Channel] [Atomic Bomberman Discord]
   Report Go to top

Re: The MiniGL thread
Posted on: 2019/3/17 19:19   #25
Home away from home
Joined:
2007/9/11 12:31
From Russia
Posts: 6378
@All
I made thread about Shogo, so we can continue there about minigl only :)

@BSzili
At least from tests i do, libtxc_dxtn version for sure works now.

Probably before you/daniel think about unstability is just via testing quake3 (with enabled extensions) and it failed (as issues with extensions come up) and as result false alarm
_________________
Join us to improve dopus5!
zerohero's mirror of os4/os3 crosscompiler suites
   Report Go to top

Re: The MiniGL thread
Posted on: 2019/3/17 19:39   #26
Not too shy to talk
Joined:
2015/6/11 9:51
From Cologne
Posts: 406
@kas1e
I updated the mgl.zip one my homepage on last time (crossing fingers ). The new dtx-lib is inside.
I also re-enabled the replacements for the deprecated CreateIO functions etc. Yesterday I disabled those when searching for the crash-culprit.
Please try that one again! Would certainly be best if we'd use the modern functions.
http://www.goldencode.de/tmp/mgl.zip
_________________
[Facebook] [YouTube Channel] [Atomic Bomberman Discord]
   Report Go to top

Re: The MiniGL thread
Posted on: 2019/3/17 19:54   #27
Just can't stay away
Joined:
2007/7/14 21:30
From Lothric
Posts: 1102
@Daytona675x

I tried Q3 three times and some other MiniGL apps, worked fine, on X5000/SI.
   Report Go to top

Re: The MiniGL thread
Posted on: 2019/3/17 19:56   #28
Home away from home
Joined:
2007/9/11 12:31
From Russia
Posts: 6378
@Daniel
That version also fine. I doing that way: reboot with that new libraries, and:

1). run 5 instances of "testgl" , and exit from
2). few run/exit from "cube" and on last one played a bit
3). run q1-darkplace, play a bit, exit
4). run sdl2 version of q3, timedemo on it, exit
5). run sdl1 version of q3, timedemo on it, exit
6). for sake of messing the memory a bit run gl4es version of q3, timedemo on it, exit
7). run Blender, move a bit in all directions, exit
8). run RCTW, played for about 10 minutes, exit
9). run Smoking Guns (damn how horrible slow is it even on x5k, need to made gl4es port for sure)
10). run LodePaint, paint a little for minute, exit
11). fire up odyssey to download some stuff from os4depot (and to mess memory more), download from there some minigl crap, run it as well one by one, all works.

At this point exit, and write that post. So, probably all still fine and good. At least i can't see any problems. Expectually if i run those mosnters like RCTW, q3 all different versions, odyssey on top of it. If it still didn't crash, i'd say go for that one :)
_________________
Join us to improve dopus5!
zerohero's mirror of os4/os3 crosscompiler suites
   Report Go to top

Re: The MiniGL thread
Posted on: 2019/3/17 21:10   #29
Home away from home
Joined:
2007/9/11 12:31
From Russia
Posts: 6378
@Daniel
Played also new Shogo via last minigl in 1920x1080 : all fine too.
_________________
Join us to improve dopus5!
zerohero's mirror of os4/os3 crosscompiler suites
   Report Go to top

Re: The MiniGL thread
Posted on: 2019/3/17 21:19   #30
Home away from home
Joined:
2006/12/2 3:55
From Italy, Perugia
Posts: 3339
@kas1e

Here it worked (Shogo) but then hanged at exit with some repetitive sound in background... will retest to see if was for casuality or if reproducible
   Report Go to top

Re: The MiniGL thread
Posted on: 2019/3/18 10:49   #31
Just can't stay away
Joined:
2007/7/14 21:30
From Lothric
Posts: 1102
Regarding compiled vertex array, it seems to me that there is an issue if "first" parameter is > 0. My interpretation of https://www.khronos.org/registry/OpenG ... compiled_vertex_array.txt is that range is [first, first + count[.

Quote:

void cgl_GLLockArrays(struct GLContextIFace *Self, GLint first, GLsizei count)
{
...

if (MGLEnsureVertexBufferSize(context, count-first)) {


Also GL error checking seems to missing.
   Report Go to top

Re: The MiniGL thread
Posted on: 2019/3/18 14:27   #32
Not too shy to talk
Joined:
2015/6/11 9:51
From Cologne
Posts: 406
@kas1e
@samo79
I now asume the lib is the most stable MGL version ever and that samo79's issue is nothing new or unrelated

@Capehill
Yes, right, you just found another package of MGL bugs (almost exactly 14 years old bugs)
And I just fixed it. Not yet comitted, but the lib on my homepage has been updated accordingly, please check it out.

However, please, report bugs in the usual way. This whole story here is about making sure that the upcoming 2.23 with its current feature list can safely replace any former lib.
Let's bring this to an end for now. I at least don't want to spent more time with this right now, I got my own projects to care for


Edited by Daytona675x on 2019/3/18 16:01:41
_________________
[Facebook] [YouTube Channel] [Atomic Bomberman Discord]
   Report Go to top

Re: The MiniGL thread
Posted on: 2019/3/18 17:03   #33
Just can't stay away
Joined:
2007/7/14 21:30
From Lothric
Posts: 1102
@Daytona675x

Works for me. Tried Q3, testgl and some other MiniGL apps.
   Report Go to top

Re: The MiniGL thread
Posted on: 2019/3/18 17:21   #34
Home away from home
Joined:
2006/12/2 3:55
From Italy, Perugia
Posts: 3339
@Daytona675x

Haha )

Retested Shogo and found that problem in reality was related to AHI, Units need to be set the same as in AHI's prefs and in game prefs, in this case Unit 0 for both, otherwise game seems crashing at exit


   Report Go to top

Re: The MiniGL thread
Posted on: 2019/3/18 17:43   #35
Home away from home
Joined:
2007/9/11 12:31
From Russia
Posts: 6378
@Daniel
Tried last version too, yeah, all fine. May i also suggest a bit change readme too ? On top we have now:

Quote:

MiniGL is a 3D API which is compatible to OpenGL(tm). It implements a subset
of that API and was designed for speed, therefore it directly interfaces with
Warp3D. This also means that there is currently no software mode.

This release is the second release of MiniGL as a set of shared libraries: minigl.library, which is the GL part, and mglut.library, which implements (partially) the GLUT API.

The new shared library interface drops a lot of the original functions (mgl*)
in favor of GLUT. Unfortunately, no documentation exists yet. Refer to the
glut source code and src/context.c to get an idea on how it works.


Imho today it will be better to just be like:

Quote:

MiniGL is a 3D API which is compatible to OpenGL 1.x. It implements a subset of that API and directly interfaces with Warp3D. There is also no software mode.

MiniGL reliased as a set of shared libraries. There are two such libraries: minigl.library, which is the GL part, and mglut.library, which implements some of the GLUT API.

The new shared library interface drops a lot of the original functions (mgl*) in favor of GLUT. Unfortunately, no documentation exists yet. Refer to the glut source code and src/context.c to get an idea on how it works.


Maybe even worht to delete 3st block at all about dropping of *mgl fucntions in favor of glut and no documentation.
_________________
Join us to improve dopus5!
zerohero's mirror of os4/os3 crosscompiler suites
   Report Go to top

Re: The MiniGL thread
Posted on: 2019/3/18 17:49   #36
Not too shy to talk
Joined:
2015/6/11 9:51
From Cologne
Posts: 406
@Capehill
Great!

@samo79
Very, hm, good

@all
I just bumped minigl and mglut to version 2.23, committed everything to the repo, incl. an updated 7z of precompiled binaries.

@kas1e
The very latest temporary mgl.zip from 5 minutes ago on my homepage has the same structure as Hans' old 2015 minigl-archive on os4depot.
Please check it out!
_________________
[Facebook] [YouTube Channel] [Atomic Bomberman Discord]
   Report Go to top

Re: The MiniGL thread
Posted on: 2019/3/18 17:57   #37
Quite a regular
Joined:
2013/10/17 16:21
From Hungary
Posts: 635
@kas1e
I think the OpenGL 1.x is less correct, since I added a few OpenGL 3 functions when I implemented fake VBOs

_________________
I see the jimmies have been rustled.
   Report Go to top

Re: The MiniGL thread
Posted on: 2019/3/18 18:16   #38
Home away from home
Joined:
2007/9/11 12:31
From Russia
Posts: 6378
@BSzili
Haha :) Not big deal to touch readme anyway , maybe later someday :)

@Daniel
Yeah, all looks fine, and like you use that mkrelease script , which have only one problem : icons didn't sorted visually correctly (expectually those ones in the Demos). If you doesn't mind, i can sort it all (so visually when any user will unpack it, all will fits nicely in wb windowses, etc).

But if not want to bother anymore, then yeah, that mgl.zip exactly like one we have on os4depot, just much better :)
_________________
Join us to improve dopus5!
zerohero's mirror of os4/os3 crosscompiler suites
   Report Go to top

Re: The MiniGL thread
Posted on: 2019/3/18 18:26   #39
Home away from home
Joined:
2007/9/11 12:31
From Russia
Posts: 6378
@Daniel

The only issue i can notice now, is that seems "data" of demos directory somehow changes a bit. At least, and by size of some files, and , by visuall look for example of "warp" demo.

It uses stars.ppm and flare.ppm.

In 2.20 archive, sizes are :

stars.ppm: 49.190
flare.ppm: 49.167

In 2.23 archive, sizes are :

stars.ppm: 49.452
flare.ppm: 50.390

As result, with new data files "warp" demo looks a bit strange. Instead of some circle in the middle its now has rectangle, which didn't fits in.

Probabaly data files somehow was regenerated or something ?

And that for sure datafiles, not minigl itself, as i tried demo and from 2.20 and from 2.23 archives on old datafiles they both fine, on new datafiles, they both looks wrong.

See:

old datafiles:

Resized Image

new datafiles:

Resized Image

See there some rectangle now.

In worse case, we can just recopy data files from 2.20 to new archive so all will be fine, but through it probably interesting to know why and when they changes.
_________________
Join us to improve dopus5!
zerohero's mirror of os4/os3 crosscompiler suites
   Report Go to top

Re: The MiniGL thread
Posted on: 2019/3/18 18:41   #40
Home away from home
Joined:
2006/12/2 3:55
From Italy, Perugia
Posts: 3339
@kas1e

Oh I just writed the same report but you anticipate me ... !
The warp demo inside the precompiled archive seems broken also on my system

Grab: https://ibb.co/sVnMNyD
   Report Go to top


« 1 (2) 3 4 5 ... 9 »



[Advanced Search]



Powered by XOOPS 2.0 © 2001-2016 The XOOPS Project