Dennis Schridde schreef: > Am Dienstag, 20. Februar 2007 schrieb Giel van Schijndel: > >> Giel van Schijndel schreef: >> >>> [EMAIL PROTECTED] schreef: >>> >>>> Using latest svn, I notice that on menu screen, where you pick >>>> map/player for multiplayer/skirmish games, on right side where >>>> show version, it get corrupted if you hover mouse over one of the >>>> options like power, and it displays low/medium/high, the version is >>>> then replaced with corruption. >>>> >>>> Same also happen when play mission, and first see mission text, it >>>> is also corrupted. >>>> >>>> Anyone else see this? >>>> >>>> I can post pictures if need. >>>> >>> Filed a bugreport: https://gna.org/bugs/index.php?8514 >>> >>> I'm guessing that ^ is what you're describing with "Same also happen >>> when play mission, and first see mission text, it is also corrupted."? >>> >>> As for the version text cluttering in the Skirmish set-up screen. I >>> haven't been able to reproduce that with r717. With HEAD I am however >>> able to reproduce that, so I'm guessing it are two separately introduced >>> bugs. Although they seem to be related because of their similarities >>> >> Ok, scratch that last part. >> I am able to reproduce the text cluttering in the Skirmish set-up screen >> with r717. >> >> I probably just didn't look hard enough for it. >> > I am very certain that the ability to reproduce this directly related to the > result of "rand()%2 == 0". ;) > > For those who don't get it immediately: If my theory is correct that the RGBA > nature of the PNGs (JPEGs are just RGB) and a broken drawing order in the > menu is responsible for this, then the above chunk of code decides whether a > PNG or a JPEG is loaded and thus whether the bug is exposed or not. > Nope, I actually found the culprit.
The problem seems to be an excess call to `glDisable(GL_TEXTURE_2D);` by pie_SetTexturePage in lib/ivis_opengl/piestate.c due to a removed call to pie_SetTexturePage(-1); The first call (to pie_SetTexturePage) with -1 as param actually calls glDisable and makes sure the second call with -1 as param *won't* call glDisable. This second call issued somewhere else in the source actually caused glDisable(GL_TEXTURE_2D) to be called (because without the first call, which has been removed in r717 it becomes the first). Whatever glDisable(GL_TEXTURE_2D) does I don't know. I'm no OpenGL expert. I do know however that if it is called the text cluttering problem appears and that it disappears if it isn't called. Anyway a fix is provided in r794. Whether it is a good/real fix or just a hack that works is something I leave up to those more experienced with OpenGL. -- Giel
signature.asc
Description: OpenPGP digital signature
_______________________________________________ Warzone-dev mailing list Warzone-dev@gna.org https://mail.gna.org/listinfo/warzone-dev
