Bug 181453 - krunner does nothing when running command the first time
Summary: krunner does nothing when running command the first time
Status: RESOLVED FIXED
Alias: None
Product: krunner
Classification: Plasma
Component: general (show other bugs)
Version: unspecified
Platform: Compiled Sources Linux
: NOR normal
Target Milestone: ---
Assignee: Plasma Bugs List
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2009-01-21 02:16 UTC by Gábor Lehel
Modified: 2009-05-08 22:17 UTC (History)
6 users (show)

See Also:
Latest Commit:
Version Fixed In:


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Gábor Lehel 2009-01-21 02:16:14 UTC
Version:           r912664 (using Devel)
Compiler:          g++ (Ubuntu 4.3.2-1ubuntu11) 4.3.2 
OS:                Linux
Installed from:    Compiled sources

When I run a program with krunner, many times, nothing happens. This is not the same as bug 175176. In this case, the command is selected, krunner closes and everything -- and -then- nothing happens. If I then bring krunner up a second time, press up to get the last command from the history, and try the same thing again, it then works.
Comment 1 Thomas Capricelli 2009-02-03 00:32:41 UTC
i have exactly the same behaviour here. It has happened for the last few month and is still present in 4.2.
Comment 2 Aurelien Gateau 2009-02-10 23:12:05 UTC
Same thing as orzel here, it's really annoying.
Comment 3 Mads Bondo Dydensborg 2009-02-20 12:48:24 UTC
Same thing here.

It makes me crazy. Please fix it.

Thanks

Mads
Comment 4 Thomas Capricelli 2009-02-20 17:51:25 UTC
alt-f2 is absolutely not usable
Most users probably do not notice, but i'm sure all developers do. 
i can't believe none of the KDE  developers is hurt by that..
Comment 5 Aurelien Gateau 2009-02-24 09:58:57 UTC
It seems to work again for me in recent (few days ago) svn.
Comment 6 David Nolden 2009-03-04 04:16:06 UTC
Really, this renders ALT+F2 nearly useless. If there's a fix in recent svn, it would be great if someone could backport it.

Sometimes I wonder if any of the plasma developers actually use krunner...
Comment 7 David Nolden 2009-03-04 04:21:53 UTC
Sorry if this sounded a bit harsh, it's just a very frustrating issue.

Greetings
Comment 8 Thomas Capricelli 2009-03-04 04:25:43 UTC
well, i agree. At first i didn't even think about reporting it. I really thought devs would make as much use of alt-f2 as i do and would fix that quickly. It's so strange... i think they probably do not have the bug, for some reasons.

Aurelien says it seems to fixed in trunk, i would really like to know if this was back-ported to 4.2.1  (or 4.2 branch..)
Comment 9 David Nolden 2009-03-05 15:36:22 UTC
I've updated to most recent opensuse KDE 4.2 packages, as well as Qt 4.5, and I haven't encountered this problem again yet.
Comment 10 David Nolden 2009-03-05 23:15:12 UTC
Ok it just happened again, so it definitely is not fixed. It just seems to happen less often(Note: Everything runs faster and smoother now, maybe it has something to do with that)
Comment 11 Thomas Capricelli 2009-03-06 00:44:10 UTC
i haven't used 4.2.1 for long yet, but it seems indeed to be at least far better. Maybe even fixed. Let me test for some more days.
Comment 12 David Nolden 2009-03-26 21:54:51 UTC
This really drives me crazy, isn't this app supposed to be useful? Why cannot it run "konsole" when I type "konsole"?
Comment 13 Thomas Capricelli 2009-03-27 16:19:15 UTC
it makes me crazy too... i'm almost used now to always ask twice to get something done..... ;-(
Comment 14 Jacopo De Simoi 2009-03-28 00:51:36 UTC
We've been working on krunner quite hard last month, and many issues are solved with 4.2.2. 
Also, lots of issues (crashes) are apparently gone with qt 4.5.
 
I personally backported everything I think is relevant from trunk. Still, occasionally I have this bug with qt 4.4.3 in branch. Also I believe it to be related to another bug which causes a crash in KRun (not KRunner).

I would be interested to know if you can reproduce the issue with all or just some runners (the shell runner especially has some weird behaviour sometimes here, whereas the location runner had never had a single problem), so if you could please try to experiment with that (with 4.2.2 at least) and report here I'd be very grateful. 

I've to find time for a serious attack on this bug so, before then, any information that could narrow the testcases will be extremely appreciated
Thanks

  --J
Comment 15 Thomas Capricelli 2009-03-28 01:00:47 UTC
hello.
I'm using qt 4.5 here, and it's still a lot buggy. I'm not sure what runners are, but i'm using alt-f2 stuff mainly for web (through direct urls, and lot of gg, ggl, wp, and others shortened names), and to launche application. Nothing else (i dont even know what can be done related to 'console'..)

I'll try 4.2.2 next week, after it has been packaged, and report here!
Comment 16 Jacopo De Simoi 2009-03-28 03:02:03 UTC
I might have found something... please check if this is a reliable way to reproduce this bug: choose two programs, e.g. konsole and konqueror. 

try to run them in krunner in this order, *each time typing the full command*:
konsole
konqueror
konsole

the first two times it should work, the third you should either get a crash or nothing happens (i.e. konsole is NOT launched)
Comment 17 David Nolden 2009-03-28 09:06:15 UTC
I'm using Qt 4.5, and KDE trunk(opensuse packages, so a week back probably).

It happens with the url runner as well as with shell runner, I don't use others too much.

You're right, I can reproduce the problem with the sequence you told.

I have started krunner from the terminal now, and it actually seems that krunner _always_ crashes when it doesn't run a command, the terminal always says "KCrash: Application 'krunner' crashing...", even if the crash-handler window doesn't always show up.
Comment 18 Mads Bondo Dydensborg 2009-03-28 12:16:18 UTC
> try to run them in krunner in this order, *each time typing the full command*:
> konsole
> konqueror
> konsole
> 
> the first two times it should work, the third you should either get a crash or
> nothing happens (i.e. konsole is NOT launched)

I can confirm this on ubuntu 8.10 with kde 4.2.1, and qt. 4.4.3.

Man, I would love to get this fixed. I use krunner to start konsoles, and gg:'s about 100 times a day, and after upgrading to kde 4, it fails me about 30-40% of the time. (Yes, I should probably look into the way I work, but its just so damn convenient).

Getting this fixed would be so great. Please let us know how we can help.
Comment 19 Thomas Capricelli 2009-03-28 13:43:40 UTC
#16:

(using qt-4.5 and kde-4.2.1)
I did what you say.
The first time, it worked as expected (all three applications start as it should, no crash).
The second time, the first konsole started, konqueror did not start, and second konsole started.
The third test : same as second one..
Comment 20 Jacopo De Simoi 2009-03-28 15:28:02 UTC
@David:
 
 Yeah the problem is present with all runners indeed
The crashes you're getting are prolly #188186 and #186036 which are triggered by this very same bug. It depends on the fact that you use the shell runner to run konsole.

@Orzel:

 The bug depends on the position of what you type in the history of krunner, if the history is full, etc etc... I can see that you are basically getting the bug with konsole and konqueror switched, so maybe there already was either one recently in the history 

@all:
The bug I found depends on a variety of conditions regarding your history, so it's difficult to be able to reproduce reliably.
I have a one-line patch which is currently under review, as I'm unsure if we have to patch krunner, the libs, or both :)
Comment 21 Mads Bondo Dydensborg 2009-03-28 15:57:33 UTC
(In reply to comment #20)

> @all:
> The bug I found depends on a variety of conditions regarding your history, so
> it's difficult to be able to reproduce reliably.
> I have a one-line patch which is currently under review, as I'm unsure if we
> have to patch krunner, the libs, or both :)

Looking eagerly forward to a fix! Your effort is hugely appreciated.

Thanks,

Mads
Comment 22 Jacopo De Simoi 2009-03-29 21:52:35 UTC
SVN commit 946618 by jacopods:

Avoid the context being reset()-ed before running the item
(not yet closing #181453 for now, even if it should be fixed - awaiting feedback)

BUG: 186036
BUG: 188186
CCBUG: 181453


 M  +3 -1      interface.cpp  


WebSVN link: http://websvn.kde.org/?view=rev&revision=946618
Comment 23 Jacopo De Simoi 2009-03-29 21:53:43 UTC
SVN commit 946619 by jacopods:

Avoid the context being reset()-ed before running the item
(not yet closing #181453 for now, even if it should be fixed - awaiting feedback)

CCBUG:181453
CCBUG:186036
CCBUG:188186


 M  +3 -1      interface.cpp  


WebSVN link: http://websvn.kde.org/?view=rev&revision=946619
Comment 24 Jacopo De Simoi 2009-03-29 21:59:46 UTC
Here we go! 
The commit *should* fix this bug, please update (svn), cross your fingers and let me know.
This shouldn't make it for 4.2.2, maybe some distros will upgrade it at some point, but let me know as soon as you try it out so we can hopefully mark this nasty bug as closed =)

--J
Comment 25 David Nolden 2009-03-29 22:01:52 UTC
@Jacopo: Great! Thanks a lot.

I have to wait for opensuse packages to catch up before I can test, but if the konsole-konqueror-konsole sequence works properly now, I think you can close this bug. :)
Comment 26 Thomas Capricelli 2009-04-06 14:30:57 UTC
The bug is still there using 4.2.2 for me. I have not (yet?) been able to test trunk though.
Comment 27 Mads Bondo Dydensborg 2009-04-06 14:36:02 UTC
Samme situation here as for poster of #26, although it is my understanding from #24, that we will have to wait for 4.2.3 to get the fix?

I am sorry that I am not currently able to compile/run trunk. If someone would post a short tutorial on getting/configuring/building/installing *krunner* with the fix I would be happy to try it out, but otherwise, I'll have to wait for 4.2.3 and kubuntu packages.

Thanks

Mads
Comment 28 Thomas Capricelli 2009-05-07 18:24:39 UTC
seems fixed with 4.2.3 indeed, thanks!
Comment 29 Jacopo De Simoi 2009-05-08 22:17:15 UTC
Cool, thanks everybody for the feedback.
I'm closing then. Please reopen if it is the case.

 --J