Bugzilla – Bug 30157
[GM45 wine] Strong lines in some parts of Shareaza and SQLYog
Last modified: 2011-03-15 17:15:36 UTC
Forwarding this bug from Ubuntu launchpad (https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/622068)
I have just upgrade to Ubuntu 10.10 and found that some lines in SQLYog and Shareaza while using Wine looks more intense than they have to be. I have made a regression test in Wine and found that this commit trigger the problem:
c29cf0591976f96c3adb30c3c3b6db59f4983251 is the first bad commit
Author: Roderick Colenbrander <firstname.lastname@example.org>
Date: Tue Apr 6 19:07:25 2010 +0200
winex11: Add support for 32-bit DDBs.
:040000 040000 73b8486e86ccd609ea2f0c8cf0c920b6ec3783dd
04bcf74196ee257fa4b8391041c627b89e6cbffe M dlls
I have reported the bug to Wine ( http://bugs.winehq.org/show_bug.cgi?id=23098 ) but after testing on a Radeon HD3450 running FGLRX and with a Radeon HD 4670 and the open source driver (xorg-x11-drv-ati-6.13.0-1.fc13.x86_64 and Mesa 7.8.1) they were unable to reproduce the problem, when I was using Ubuntu 10.04 I didn't have this issue (Only when I installed the xorg-edgers PPA). So my conclusion is that this must be an issue in the Intel driver.
There are screenshots in the Wine bug and some more information.
An easy way to test this bug is type 'winecfg' in the terminal, and the go to the 'Graphics' tab. The slider in the Screen Resolution section appear incorrectly.
This bug is not just creating wider lines, is also creating black tabs and in some areas of other programs like Ares is just showing black areas.
This is the report about black tabs in Office 2007
This is regression in xf86-video-driver. I've managed to bisect part of problem:
Author: Chris Wilson <email@example.com>
Date: Wed May 26 10:12:57 2010 +0100
uxa: Use Composite rather than solid blitter for PolyRect
Due to the relocation overhead, using a single composite with many
rectangles outperforms many solid blits.
Signed-off-by: Chris Wilson <firstname.lastname@example.org>
with this commit reverted winecfg slider is still black, but line it slides on is of normal text. Also reverting this commit fixes my inability to see tabs (they were black) during visio 2007 installation (which is the major regression for me). Actually if using driver at ec2437f958ec4f5ac5222b37cba4cd403b5c8855 wine works fine.
Maybe bug http://bugs.winehq.org/show_bug.cgi?id=23424 is related to this issue too, however it isn't (any longer) possible to revert commit 3055d40 on current TRUNK of the git tree to test whether that commit breaks something in wine or not.
Until the moment I read the comment #2, I thought that this bug is related to Xrender extension, but now I'm not sure.
When I set up wine registry by importing this registry file
then everything works fine.
Peter, will you check that this registry workaround eliminates your problem?
The bug is in the ddx's core rendering routines. The workaround that you've found is for wine to do it's rendering locally and then push the image as a whole to the xserver (hopefully only updating the damage regions!). This may in fact be quicker than hitting the X core rendering paths...
I can confirm that the workaround had solved the issue for me and I haven't noticed any slowdown or artifacts in redering. Thank You very much, I quite happy with my desktop roght now. :)
The wine - people are convinced that this issue is an issue of at least a X.org software piece. Now I'm wondering how I can help to solve this issue - b/c setting the registry key seems to be only a work around - and I rely on a working Lotus Notes at work.
Toralf Förster <email@example.com> writes:
> Alexandre Julliard wrote at 23:01:42
>> No, setting that registry value is a bad idea and should be discouraged.
>> It will break all sorts of things, particularly anything to do with
>> alpha blending.
> Oh - thx.
> Then I'm wondering what I can do to solve this issue :
Not a whole lot I'm afraid, except continue to push the X.org guys to
fix the bug, or revert to an old driver, or get a different graphics
card with working drivers...
The workaround had solved most of the issue but for MS Office 2007 the problem is even worst when I set the registry key so is not useful in all the cases.
(In reply to comment #8)
> The workaround had solved most of the issue but for MS Office 2007 the problem
> is even worst when I set the registry key so is not useful in all the cases.
I can confirm worse results with Office:
this bug affects Linux vector graphics program Inkscape.
the rulers are overwritten by black mouse pointers.
For Wine "ClientSideWithRender"="N" regkey tells Wine to not use XRender extension for font rendering, which partially solves wine issues but not Linux issues.
(In reply to comment #9)
> (In reply to comment #8)
> > The workaround had solved most of the issue but for MS Office 2007 the problem
> > is even worst when I set the registry key so is not useful in all the cases.
> I can confirm worse results with Office:
I confirm worse results with Office 2007 too.
It seems to happen with any Intel driver newer than 2.11.
Tested on a Core i3's integrated graphics.
I'm unable to reproduce the problem in both a VirtualBox VM running Arch Linux and my PC with an ATI HD5770.
I'ts affecting also Metatrader 4. You can see the issue when using a color scheme Yellow over black or when inserting a Parabolic sar indicator that should hava lime color but appear black.
Actually this bug is a duplicate of bug 28871 which was filed two months earlier.
(In reply to comment #13)
> Actually this bug is a duplicate of bug 28871 which was filed two months
Yes, as I pointed in a comment in bug 28871 .
It's the same bug.
Important to notice it affects other Linux apps other than Wine.
*** This bug has been marked as a duplicate of bug 28798 ***