Version: (using Devel) Installed from: Compiled sources [ I initially filed this at https://bugzilla.redhat.com/show_bug.cgi?id=468451 and Rex Dieter suggested that I file it here upstream instead. My original bug was with a Fedora 10 pre-release, but I'm still seeing this with Fedora 10 + KDE 4.2 beta 2 from kde-redhat ] I'm running KDE 4 on an Acer One mini-laptop with a DPI of 133x135 (according to xdpyinfo). It seems that the fonts in window title bars, the start menu, and the task bar are using a different DPI than the rest of the system: they show up as significantly smaller than what is ostensibly the same font inside windows. I'll attach a screenshot of the system settings dialogue -- note the difference between the 9-point text in the "Window title" field and the actual window title bar. A workaround is to force the fonts DPI to 120 -- then all of the fonts on the screen agree with each other. I've also seen similar symptoms on a laptop with 122x122 DPI.
Created attachment 30019 [details] System settings window Note the difference between the selected "Window titlebar" font and the actual window titlebar
Sorry for the bad formatting (lack of paragraphs) above -- probably a cut-and-paste issue
This has also been an issue on a new laptop with a similarly high DPI: 15.4" screen, 1920x1200 resolution => 147x148 (or so) DPI. Some fonts showed up really big, others showed up really tiny, and forcing font DPI to 120 gets things mostly working okay but that's not satisfying.
I cannot reproduce this on Kubuntu with KDE 4.2.4. Which version of KDE do you run now?
I can confirm that this issue exists on Fedora 11, KDE 4.2.4 on a Thinkpad x301. It also seems to involve a race, since every now and then when I log into KDE the fonts are displayed correctly. Most of the times though KWin's fonts are smaller, as Mary Ellen has pointed out.
This might be a duplicate of bug 171683. If you use ATI, try open source driver (radeon/radeonhd) instead of fglrx.
I'm using the Intel driver on both laptops where I see this, for what it's worth ...
Same here, the i915 driver.
Someone with nvidia driver is able to reproduce this bug too? Anyway I agree with Christoph, this could be a duplicate of bug #171683
assuming dupe. if anybody can cause this with "kwin --replace&" during the session, please stand up :-) *** This bug has been marked as a duplicate of bug 260943 ***