Bug 231571 - Analog Clock plasmoid is painted incorrectly after KDE startup
Summary: Analog Clock plasmoid is painted incorrectly after KDE startup
Status: RESOLVED DUPLICATE of bug 198770
Alias: None
Product: plasma4
Classification: Plasma
Component: widget-clock (show other bugs)
Version: unspecified
Platform: Compiled Sources Linux
: NOR normal
Target Milestone: ---
Assignee: Plasma Bugs List
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2010-03-21 19:31 UTC by Alexey Chernov
Modified: 2010-05-13 21:30 UTC (History)
3 users (show)

See Also:
Latest Commit:
Version Fixed In:


Attachments
Picture with wrong painted clock (4.91 KB, image/png)
2010-03-21 19:35 UTC, Alexey Chernov
Details
Correctly painted clock (30.42 KB, image/png)
2010-03-21 19:37 UTC, Alexey Chernov
Details

Note You need to log in before you can comment on or make changes to this bug.
Description Alexey Chernov 2010-03-21 19:31:16 UTC
Version:            (using KDE 4.4.1)
Compiler:          GCC 4.4.3 Target: x86_64-unknown-linux-gnu Configured with: ../gcc-4.4.3/configure --prefix=/usr --libexecdir=/usr/lib --enable-shared --enable-threads=posix --enable-__cxa_atexit --enable-clocale=gnu --enable-languages=c,c++,fortran --disable-multilib --disable-bootstrap Thread model: posix
OS:                Linux
Installed from:    Compiled From Sources

Clock plasmoid isn't painted correctly after KDE startup. Actually, just some artifacts are painted (they are moving, i.e. if second arrow moves, the wrong texture of it moves somewhere, too). See the screenshot for additional information.
Comment 1 Alexey Chernov 2010-03-21 19:35:57 UTC
Created attachment 41808 [details]
Picture with wrong painted clock
Comment 2 Alexey Chernov 2010-03-21 19:37:51 UTC
Created attachment 41809 [details]
Correctly painted clock

Wrong remaining texture on the left of shot disappears on next desktop repaint.
Comment 3 Anne-Marie Mahfouf 2010-03-21 20:28:36 UTC
You mention "compiled from source": what revision is kdebase please?

You wrote "Wrong remaining texture on the left of shot disappears on next desktop repaint." what do you mean with "next desktop repaint"?
Comment 4 Alexey Chernov 2010-03-21 20:34:46 UTC
It's 4.4.1 release package.

Yes, it disappears on opening and closing a window and switching to a desktop with opened windows for me.

I forgot to say, the clock become painted OK if you resize it or move it etc.
Comment 5 Anne-Marie Mahfouf 2010-03-21 20:49:40 UTC
I'd like to know what 4.4.1 release package you used please as I thought we fixed this. A revision number for that package would help or the distro release package.

Thanks in advance.
Comment 6 Alexey Chernov 2010-03-21 20:54:57 UTC
Is there any way to find it out from unpacked tarball?
Comment 7 Anne-Marie Mahfouf 2010-03-21 21:03:03 UTC
As you don't indicate your distro (I guess you can ask them the revision number) please see if you have this commit in your sources
http://websvn.kde.org/?view=revision&revision=1092156
as it's the one that I thought fixed the painting problem.
Comment 8 Alexey Chernov 2010-03-21 21:24:36 UTC
Yes, I use LFS so it's only up to me to know :)

As to commit, it seems to partially applied but something is definitely changed. It's hard to check all the lines, but I've checked bigger ones - most of them are applied, but some seems to be old.
Comment 9 Anne-Marie Mahfouf 2010-03-21 21:29:09 UTC
Either you have the files as after the commit or not. If not or if in doubt, please download the updated files and replace yours with them and rebuild kdebase and see whether the analog clock painting problem is gone or not at the next start after the installation of this updated build. Thanks in advance.
Comment 10 Alexey Chernov 2010-03-21 21:34:39 UTC
Ok, I'll try to make it the next several days and post the result here. Thank you for references.
Comment 11 Alexey Chernov 2010-04-25 14:44:06 UTC
Sorry for delay, I've checked 4.4.1 with the commit which you pointed but there was no effect, bug still existed.
Now I switched to 4.4.2 and this bug is still here. What's more, previously it affected only clock on the 2nd monitor, but now it hurts clocks of both monitors.
Comment 12 Anne-Marie Mahfouf 2010-04-25 15:18:30 UTC
So let's define the bug again as I lost track of what it really is. Can you check first with one monitor and explain the bug. You can use 
kquitapp plasma-desktop to quit plasma and then plasma-desktop in a terminal to restart plasma and thus test the painting of the analog clock.
For example put an analog clock on your desktop and restart plasma. Do the same with the second hand enabled. Write down the exact steps to reproduce.
Once you explain this, I can see if I reproduce and fix it!
Thanks in advance.
Comment 13 Alexey Chernov 2010-04-25 17:03:13 UTC
Ok, I've deeply tested the case and here're the results:
1. With one monitor turned on it's 100% not reproducible. I've checked with 3 logouts and starts and it didn't appear.
2. When switching to 2 monitors mode (I mean, not clone but 'shared' desktop) in a desktop _with_no_windows_are_to_be_opened_on_the_2nd_screen it's 100% reproducible. The screenshot is the same as I posted previously.
3. When there were some windows on the 2nd screen then you minimize them and the bug isn't reproducible, everything is ok.
4. When restarting plasma with kquitapp in console both with one and two screens everything is ok, too. The bug isn't reproducible.

So here's steps to reproduce the bug:
1. Start KDE in one monitor mode (or in clone mode). Check that clock is ok.
2. Switch to some desktop with no windows opened on it. So, on second screen won't appear any window, too.
3. Change the settings to 2 monitor mode in 'Screen' section of systemsettings.
4. Look at the 2nd monitor and check the bug on the clock appeared. If somehow the window on the 1st monitor disappeared (for example, you quit the settings window and apply changes at once) then the bug would appear on the 1st monitor, too.
Comment 14 Marco Martin 2010-05-13 21:30:07 UTC
slightly different corruption but i think is the very same problem

*** This bug has been marked as a duplicate of bug 198770 ***