OpenGL bug in 0.9.1
Hi, to Oliver Stiebel and anyone else interested: I reported about a bug in wine 20050930 in October. Then Oliver suggested he might write a little test application for this bug. Well, the bug is still present in wine 0.9.1. Searching the web turned up a simple app for windows which is equivalent to glxinfo under X11 (see [1], it's called wglinfo). As this program reproduces the bug, I thought it might provide a good starting point for a test app. regards, Daniel [1] http://www.cg.tuwien.ac.at/~wimmer/
On Sunday 20 November 2005 18:28, Daniel wrote:
Hi,
to Oliver Stiebel and anyone else interested: I reported about a bug in wine 20050930 in October. Then Oliver suggested he might write a little test application for this bug. Well, the bug is still present in wine 0.9.1. Searching the web turned up a simple app for windows which is equivalent to glxinfo under X11 (see [1], it's called wglinfo). As this program reproduces the bug, I thought it might provide a good starting point for a test app.
Hi, what is the bug ? (buzilla entry number?) Regards, Raphael
regards, Daniel
--- Raphael <fenix(a)club-internet.fr> wrote:
On Sunday 20 November 2005 18:28, Daniel wrote:
Hi,
to Oliver Stiebel and anyone else interested: I reported about a bug in wine 20050930 in October. Then Oliver suggested he might write a little test application for this bug. Well, the bug is still present in wine 0.9.1. Searching the web turned up a simple app for windows which is equivalent to glxinfo under X11 (see [1], it's called wglinfo). As this program reproduces the bug, I thought it might provide a good starting point for a test app.
Hi,
what is the bug ? (buzilla entry number?) wine3277..
I sent Daniel this small test app: void main(void) { Display *display; char *displayname = NULL; int nCfgs = 0; GLXFBConfig* cfgs = NULL; int att_list[64]; int att_pos = 0; display = XOpenDisplay (displayname); printf("%s %s \n",glXGetClientString(display,GLX_VENDOR), displayname); ADD2(GLX_RENDER_TYPE, GLX_RGBA_BIT); ADD2(GLX_BUFFER_SIZE, 0); att_list[att_pos] = None; cfgs = glXChooseFBConfig(display, DefaultScreen(display), att_list, &nCfgs); if (cfgs == NULL) { printf("cfgs is null\n"); } else { printf("Found configs\n"); } and it returns cfgs is null which I wouldn't have expected, but it looks like there's a basic problem with glXChooseFBConfig Oliver
Regards, Raphael
regards, Daniel
___________________________________________________________ Yahoo! Messenger - NEW crystal clear PC to PC calling worldwide with voicemail http://uk.messenger.yahoo.com
On Monday 21 November 2005 12:30, Oliver Stieber wrote:
--- Raphael <fenix(a)club-internet.fr> wrote:
On Sunday 20 November 2005 18:28, Daniel wrote:
Hi,
to Oliver Stiebel and anyone else interested: I reported about a bug in wine 20050930 in October. Then Oliver suggested he might write a little test application for this bug. Well, the bug is still present in wine 0.9.1. Searching the web turned up a simple app for windows which is equivalent to glxinfo under X11 (see [1], it's called wglinfo). As this program reproduces the bug, I thought it might provide a good starting point for a test app.
Hi,
what is the bug ? (buzilla entry number?)
wine3277..
I sent Daniel this small test app:
void main(void) { Display *display; char *displayname = NULL; int nCfgs = 0; GLXFBConfig* cfgs = NULL; int att_list[64]; int att_pos = 0;
display = XOpenDisplay (displayname); printf("%s %s \n",glXGetClientString(display,GLX_VENDOR), displayname);
ADD2(GLX_RENDER_TYPE, GLX_RGBA_BIT); ADD2(GLX_BUFFER_SIZE, 0); att_list[att_pos] = None; cfgs = glXChooseFBConfig(display, DefaultScreen(display), att_list, &nCfgs);
if (cfgs == NULL) { printf("cfgs is null\n"); } else { printf("Found configs\n"); }
and it returns cfgs is null which I wouldn't have expected, but it looks like there's a basic problem with glXChooseFBConfig
Yak, can you test what is the error X provide ? and without BUFFER_SIZE at 0 ? Regards, Raphael
--- Raphael <fenix(a)club-internet.fr> wrote:
On Monday 21 November 2005 12:30, Oliver Stieber wrote:
--- Raphael <fenix(a)club-internet.fr> wrote:
On Sunday 20 November 2005 18:28, Daniel wrote:
Hi,
to Oliver Stiebel and anyone else interested: I reported about a bug in wine 20050930 in October. Then Oliver suggested he might write a little test application for this bug. Well, the bug is still present in wine 0.9.1. Searching the web turned up a simple app for windows which is equivalent to glxinfo under X11 (see [1], it's called wglinfo). As this program reproduces the bug, I thought it might provide a good starting point for a test app.
Hi,
what is the bug ? (buzilla entry number?)
wine3277..
I sent Daniel this small test app:
void main(void) { Display *display; char *displayname = NULL; int nCfgs = 0; GLXFBConfig* cfgs = NULL; int att_list[64]; int att_pos = 0;
display = XOpenDisplay (displayname); printf("%s %s \n",glXGetClientString(display,GLX_VENDOR), displayname);
ADD2(GLX_RENDER_TYPE, GLX_RGBA_BIT); ADD2(GLX_BUFFER_SIZE, 0); att_list[att_pos] = None; cfgs = glXChooseFBConfig(display, DefaultScreen(display), att_list, &nCfgs);
if (cfgs == NULL) { printf("cfgs is null\n"); } else { printf("Found configs\n"); }
and it returns cfgs is null which I wouldn't have expected, but it looks like there's a basic problem with glXChooseFBConfig
Yak, can you test what is the error X provide ? and without BUFFER_SIZE at 0 ?
Daniel is putting a patch together, his card only supported glx 1.2 and glXChooseFBConfig required 1.3. There was some old code that works with 1.2 that he's using. Oliver.
Regards, Raphael
___________________________________________________________ Yahoo! Messenger - NEW crystal clear PC to PC calling worldwide with voicemail http://uk.messenger.yahoo.com
On Tuesday 22 November 2005 02:09, you wrote: <snip>
Yak, can you test what is the error X provide ? and without BUFFER_SIZE at 0 ?
Daniel is putting a patch together, his card only supported glx 1.2 and glXChooseFBConfig required 1.3. There was some old code that works with 1.2 that he's using.
No, all cards "supports" glx 1.3. GLX is implemented on X server and is independant from graphic card capabilities (GLX 1.3 is only a API evolution form 1.2). Graphic cards driver shouldn't influe this APIs. He must find what happened (and maybe open a bug report on X)
Oliver.
Regards, Raphael
On Tue, Nov 22, 2005 at 08:55:05AM +0100, Raphael wrote:
Daniel is putting a patch together, his card only supported glx 1.2 and glXChooseFBConfig required 1.3. There was some old code that works with 1.2 that he's using.
No, all cards "supports" glx 1.3. GLX is implemented on X server and is independant from graphic card capabilities (GLX 1.3 is only a API evolution form 1.2). Graphic cards driver shouldn't influe this APIs. He must find what happened (and maybe open a bug report on X)
Actually, I think Raphael is right. glXChooseFBConfig shouldn't be hardware related (it's nothing that needs acceleration like e.g. matrix transforms). However, as glxinfo shows, my system just doesn't implement GLX 1.3, only version 1.2. So, while it is strange that glXChooseFBConfig exists at all on my system it is not required to work. Thus, no bug in X11 (I think). Daniel
--- Daniel <daniel.skorka(a)stud.uni-karlsruhe.de> wrote:
On Tue, Nov 22, 2005 at 08:55:05AM +0100, Raphael wrote:
Daniel is putting a patch together, his card only supported glx 1.2 and glXChooseFBConfig required 1.3. There was some old code that works with 1.2 that he's using.
No, all cards "supports" glx 1.3. GLX is implemented on X server and is independant from graphic card capabilities (GLX 1.3 is only a API evolution form 1.2). Graphic cards driver shouldn't influe this APIs. He must find what happened (and maybe open a bug report on X)
Actually, I think Raphael is right. glXChooseFBConfig shouldn't be hardware related (it's nothing that needs acceleration like e.g. matrix transforms). However, as glxinfo shows, my system just doesn't implement GLX 1.3, only version 1.2. So, while it is strange that glXChooseFBConfig exists at all on my system it is not required to work. Thus, no bug in X11 (I think).
From ATI's documentation on pbuffers, but still about glx and X11
2. Description of the "pbuffer API" Pbuffer support was officially introduced in GLX 1.3 but Silicon Graphics, Inc. has publicly released only the source code of GLX 1.2 so far. This is included in XFree4.X. To provide pbuffer support to our customers, we decided to partially emulate the behavior of GLX 1.3 on top of GLX 1.2. As soon as the "real" GLX 1.3 API is available in XFree86, you can use the "genuine" pbuffers of GLX 1.3 with our drivers. This will not require any changes in your application code because the programmer has to use the same function calls. To use pbuffers, you need a custom version of "libGL.so.1.2" installed - including our GLX 1.3 emulation code. Our driver installation tool will automatically do that for you. NOTE: Your application will report GLX version 1.2 on glXQueryVersion. However, it is safe to call the GLX 1.3 functions as described below. and glxinfo on my pc shows client glx 1.3 and server glx 1.2, which seems to fit ATI's documentation. Oliver.
Daniel
___________________________________________________________ To help you stay safe and secure online, we've developed the all new Yahoo! Security Centre. http://uk.security.yahoo.com
On Tue, Nov 22, 2005 at 02:05:07PM +0000, Oliver Stieber wrote:
From ATI's documentation on pbuffers, but still about glx and X11 [...] To use pbuffers, you need a custom version of "libGL.so.1.2" installed - including our GLX 1.3 emulation code. Our driver installation tool will automatically do that for you.
Ahh, I think you're confused here. I am _not_ using the binary drivers provided by ATI. I _am_ using the stock drivers shipped with X11 and provided by the DRI project. Therefore, what you just quoted doesn't apply to my situation. Also, even if pbuffers are something hardware related, glXChooseFBConfig seems to be nothing more than a nicer interface to XGetVisualInfo and glXGetConfig than glXChooseVisual is.
and glxinfo on my pc shows client glx 1.3 and server glx 1.2, which seems to fit ATI's documentation.
It reports client glx 1.2 and server glx 1.2 for me. regards, Daniel
--- Daniel <daniel.skorka(a)stud.uni-karlsruhe.de> wrote:
On Tue, Nov 22, 2005 at 02:05:07PM +0000, Oliver Stieber wrote:
From ATI's documentation on pbuffers, but still about glx and X11 [...] To use pbuffers, you need a custom version of "libGL.so.1.2" installed - including our GLX 1.3 emulation code. Our driver installation tool will automatically do that for you.
Ahh, I think you're confused here. I am _not_ using the binary drivers provided by ATI. I _am_ using the stock drivers shipped with X11 and provided by the DRI project. Therefore, what you just quoted doesn't apply to my situation. Also, even if pbuffers are something hardware related, glXChooseFBConfig seems to be nothing more than a nicer interface to XGetVisualInfo and glXGetConfig than glXChooseVisual is.
I think it makes sense of why other peoples machines are working and yours isn't, many people have client glx version 1.3 that supports glXChooseFBConfig you only have glx 1.2 that doesn't support glXChooseFBConfig which is why I think wines opengl isn't working for you but other opengl application are.
and glxinfo on my pc shows client glx 1.3 and server glx 1.2, which seems to fit ATI's documentation.
It reports client glx 1.2 and server glx 1.2 for me.
My OpenGL book has some weird notes: glXChooseFBConfig is available only if the GXL version is 1.3 or greater. but then it says: If the GLX version is 1.1 or 1.0 the GL version must be 1.0, If the GLX version is 1.2 the GL version must be 1.1. If the GLX version is 1.3 the GL version must be 1.2. which doesn't seem to fit with it'self. It also says: Errors: NULL Is returned if ..... dpy doesn't support the GLX extension. I think the ChooseVisuial fallback if either, the client glx version is 1.3 or glXChooseFBConfig returns NULL is OK. Oliver.
regards, Daniel
___________________________________________________________ WIN ONE OF THREE YAHOO! VESPAS - Enter now! - http://uk.cars.yahoo.com/features/competitions/vespa.html
Oliver Stieber <oliver_stieber <at> yahoo.co.uk> writes:
I think the ChooseVisuial fallback if either, the client glx version is 1.3 or glXChooseFBConfig returns NULL is OK.
Falling back to glXChooseVisual() if glXChooseFBConfig() fails should be sufficient. We probably should not depend on glx 1.3 only since the binary-only ATI driver also only claims glx 1.2. However, it should be noted that while Xorg/XFree86 do not support glx 1.3, they _do_ support glXChooseFBConfig and friends... so there really should be no one failing to use glXChooseFBConfig, and if there is it might be a bug in either how it is being used, or in its implementation. Regards, Aric
participants (4)
-
Aric Cyr -
Daniel -
Oliver Stieber -
Raphael