FS#40949 - [xf86-video-intel] upgrade to 2.99.912-1 breaks font rendering on dual screen setup
Attached to Project:
Arch Linux
Opened by W Len (wlen) - Monday, 23 June 2014, 12:37 GMT
Last edited by Laurent Carlier (lordheavy) - Tuesday, 05 August 2014, 10:51 GMT
Opened by W Len (wlen) - Monday, 23 June 2014, 12:37 GMT
Last edited by Laurent Carlier (lordheavy) - Tuesday, 05 August 2014, 10:51 GMT
|
Details
Description:
After the upgrade of xf86-video-intel from 2.99.911-2 -> 2.99.912-1, I'm getting font rendering glitches on my dual screen setup. See attachments for the glitches; pieces of text disappear in certain applications (Claws-mail, LibreOffice, Gvim), sometimes re-appearing when scrolling. The problem only manifests itself after switching to a dual-screen setup using xrandr. I use: xrandr --output VGA1 --auto --output LVDS1 --right-of VGA1 The problem appears on both screens. Switching to a single-screen setup using xrandr fixes the problem. Downgrading to 2.99.911-2 also solves the problem. See also: https://bbs.archlinux.org/viewtopic.php?id=183152 |
This task depends upon
Closed by Laurent Carlier (lordheavy)
Tuesday, 05 August 2014, 10:51 GMT
Reason for closing: Fixed
Additional comments about closing: xf86-video-intel-2.99.914-3
Tuesday, 05 August 2014, 10:51 GMT
Reason for closing: Fixed
Additional comments about closing: xf86-video-intel-2.99.914-3
the display controller is : Intel Corporation Mobile 945GM/GMS/GME, 943/940GML Express Integrated Graphics Controller (rev 03)
Btw, the problem does NOT exist with my samsung laptop (NP900x4c) with :
00:02.0 VGA compatible controller: Intel Corporation 3rd Gen Core processor Graphics Controller (rev 09)
Font rendering works fine on either display LVDS1 or VGA1 with the other off, or if they both display the same screen area. As soon as I put one display --output VGA1 --right-of LVDS1 the fonts become invisible.
Some fonts are affected, others not. It appears to be all scalable fonts or possibly only all ttf fonts that become invisible. Indeed libreoffice becomes unusable exactly like the first attachment of original report. Same problem in xterm with xterm*faceName: resource.