[prev in list] [next in list] [prev in thread] [next in thread] 

List:       dri-patches
Subject:    [Dri-patches] [Bug #121897] G400 glReadPixel() affected by overlaping windows
From:       noreply () sourceforge ! net
Date:       2000-11-07 18:54:23
[Download RAW message or body]

Bug #121897, was updated on 2000-Nov-07 10:54
Here is a current snapshot of the bug.

Project: Direct Rendering Infrastructure
Category: MGA X Server
Status: Open
Resolution: None
Bug Group: Rendering Error
Priority: 5
Summary: G400 glReadPixel() affected by overlaping windows

Details: Setup:
	G400, PIII BX
	linux 2.4.0-test10
	DRI from CVS 7th Nov 2000
	xserver run in 32bpp.
	dga enabled.
	Backing store enabled (disabling makes no difference)

I have an application that makes extensive use of gltexsub and glreadpixels. It opens \
an double (or single) buffered  Xwindow and maps it. 

Textures are rendered into this window and read back using glReadPixcels. 

If the glx window is obsured by another, then the contents of the array returned are \
corrupted.

This happens if all of the following cases:-
  Double buffered:
	render to back buffer and read from back
	render to back, swap and read from front
 Single buffer:	
	render and read to front buffer.

This used to work ok, Infact I used only used to map the window if I was debuging. I \
readpixcels now returns black if I dont map the window.

The textures are RGB8 and the visual is RGBA.

I'll try and work on a simple testcase to demonstate it.


Possibly related:-
If I hit Ctrl-C, next time I run the application the textures are corrupted with blue \
and yellow tint on the display. Some times blobs are splattered arround the screen on \
start up. This can also causes the xserver to hang, but this can normaly be freed by \
sshing in and killing the glx application.


This used to work many months ago, just after 32 bit was working with mga. Sorry I \
dont have the date.

Mike


For detailed info, follow this link:
http://sourceforge.net/bugs/?func=detailbug&bug_id=121897&group_id=387


[prev in list] [next in list] [prev in thread] [next in thread]