SFML2 Error / Crash message on Vista with ATI drivers

[Figs]

Hello. I am running Windows Vista and built SFML2 using MSYS/MinGW. The examples run but display a "[name of program] has stopped working." error message every time I quit. The problem appears to be the result of a segfault in an ATI graphics driver (atioglxx.dll), in particular, gdb indicates that the segfault occurs at 0x693839fa in atioglxx!DrvGetProcAddress (). To the best of my knowledge, I am using the most up to date drivers. My naive guess is that this may be a result of some sort of multithreading/concurrency problem in the interaction between window cleanup and OpenGL, although I'm not familiar enough with OpenGL or the SFML2 codebase to know where to look next. Any ideas?

[Laurent]

It's a known bug, don't worry.

[Figs]

This is a rather significant bug for my purposes. Is there a plan to fix it soon? If not, can you point me in the correct direction towards working on a fix myself?

[Laurent]

Is there a plan to fix it soon?

To fix it yes, definitely, but I don't know if it's going to happen soon. It's very hard to fix.

If not, can you point me in the correct direction towards working on a fix myself?

Use static link to avoid the crash. As for fixing the crash itself, don't even try to do that, it's a pure nightmare
And there's probably no clean solution anyway, so...

[Mr. X]

Has this bug been reported to AMD (As far as I know, its a bug in the driver and not in SFML, right?)? If they do not know of it, it will never be fixed.

[Laurent]

Well, I don't know if it's a bug. It happens at global exit of a DLL, so I don't know if the expected behaviour is well defined.

But I can report it and see what they say.

[Mr. X]

Any updates on this? (Have you informed AMD about it? Whats their response (if you already have one)?)

[Xurix]

This isn't a vista Specific bug. I had it on xp to.
But I expect Laurent will get it fixed.. somehow... lol Maybe ATI will fix it.. who knows.