Emmanuele Bassi wrote:
> On Fri, 2009-02-27 at 11:49 +0100, Michal Kolodziejczyk wrote:
>  
>>> it's probably a wrong implementation of mipmapping, if only the text is
>>> wrong.
>>>
>>> you can disable text mipmapping by using:
>>>
>>>   clutter_set_use_mipmapped_text (FALSE);
>> I will try this.

I confirm: this resolved the issue.

>> I have also problems with test-texture-quality (its get killed while
>> switching to high quality), but as you said, it can be driver related. I
>> will report back.
> 
> ...this is the strongest indication that it's a bug in the mipmapping
> implementation of the openchrome driver.

I am attaching a gdb backtrace if it would help:

(gdb) bt
#0  0x00000000 in ?? ()
#1  0xb7612171 in _mesa_get_teximage ()
   from /usr/lib/xorg/modules/dri/libdricore.so
#2  0xb7605218 in _mesa_GetTexImage ()
   from /usr/lib/xorg/modules/dri/libdricore.so
#3  0xb7d02f33 in cogl_texture_get_data ()
   from /usr/lib/libclutter-glx-0.8.so.0
#4  0xb7cf5e76 in clutter_texture_unrealize ()
   from /usr/lib/libclutter-glx-0.8.so.0

I am not sure if its a driver or mesa opengl library. I would appreciate
any clues.

> writing a simple GL application that enables mipmapping should
> definitely prove the theory, and allow you to open a bug against the
> openchrome drivers.

Well, the program compiled from source code available here:
http://www.videotutorialsrock.com/opengl_tutorial/mipmapping/text.php

works OK.


Regards,
        miko
-- 
To unsubscribe send a mail to [email protected]

Reply via email to