Bug 374402 - startkde and qdbus-- logic in startkde make no sense
Summary: startkde and qdbus-- logic in startkde make no sense
Status: RESOLVED FIXED
Alias: None
Product: plasmashell
Classification: Plasma
Component: general (other bugs)
Version First Reported In: 5.8.4
Platform: Mageia RPMs Linux
: NOR normal
Target Milestone: 1.0
Assignee: David Edmundson
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2016-12-31 17:23 UTC by unruh
Modified: 2017-01-04 21:32 UTC (History)
2 users (show)

See Also:
Latest Commit:
Version Fixed/Implemented In:
Sentry Crash Report:


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description unruh 2016-12-31 17:23:15 UTC
If /usr/lib64/qt5.bin is not in pathe startx startkde produces
Cannot start D-Bus. Can you callqdbus
error message.
The problem is that starkde cannot find qdbus. It tries to run qtpaths sometimes, but since that is in exactly the same directory as qdbus, if it cannot find qdbus, it cannot find qtpaths either.

startkde looks to see if startkde was run with a full path or not. If it was run with a full path, it adds that directory to the internal PATH. It then tries to run qtpaths to add a full path to qdbus. But that makes no sense since if qdbus cannot be found neither can qdbus. And whether or not startkde is started with a full path is unrelated to whether qdbus can be fouund since startde is not in the same directory as qdbus./lib64/qt5/bin
a) at least a link to qtpaths should be in /usr/bin, so qtpaths can allways be found to be run.
b) qtpaths should always be run to create a call to qdbus which is independent of whether or not startkde is started with a full path or not, and whether or not /ust/lib64/qt5/bin is in the PATH or not.
c) Yhe error message should be changed to somethig usefull.The current errormessage reads like a joke (what is qdbus)

I could make a suggestion as to how startkde should be changed  if that would be helpful.
Comment 1 Christoph Feck 2016-12-31 17:38:10 UTC
> a) at least a link to qtpaths should be in /usr/bin, so qtpaths can allways be found to be run.

Yes, and that is the responsibility of the distribution packagers. Please report this issue to the bug tracker of your distribution.
Comment 2 David Edmundson 2017-01-02 18:22:19 UTC
Git commit bc46f163234f7a3882469c54bdd5d682937e5f78 by David Edmundson, on behalf of Kai Uwe Broulik.
Committed on 02/01/2017 at 18:22.
Pushed by davidedmundson into branch 'master'.

Sync startplasma with a startkde change

Summary:
bc1324ceebbd71d1571d3adb51371b17391ab425 in startkde had a change

    Don't try to set $qdbus to a path to qdbus... it wasn't used
    conherently, without issue don't really need it

this wasn't synced with startplasma.

Reviewers: #plasma

Subscribers: plasma-devel

Tags: #plasma

Differential Revision: https://phabricator.kde.org/D3915

M  +4    -17   startkde/startplasma.cmake

https://commits.kde.org/plasma-workspace/bc46f163234f7a3882469c54bdd5d682937e5f78
Comment 3 David Edmundson 2017-01-02 18:22:27 UTC
startkde has no reference to qtpaths (in Plasma 5.9 ) and so this bug is fixed.

startplasma does (the wayland version) but that's just because it's out of sync.
I'll fix that and close this.
Comment 4 unruh 2017-01-03 03:54:32 UTC
I disagree. startkde crashing makes the whole system inoperable. That situation should not happen. If qdbus cannot be found, then startkde crashes. This will happen if for exqmple the path to qdbus is not in the PATH.  On my system, it is in /usr/lib{,64}/qt5/bin and that is not a directory anyone would expect to have to be in the PATH. It never hs before and the lib directories should not be expected to be there. 
startkde needs to do a much better job of searching for qdbus than it does, and needs to throw a far more informative error message if it cannot find it. 
To blame the distributors for a crash which makes the system almost inoperative really should not be acceptable.

Note that there IS an issue. I and many others have had startkde crash on them because it turns out that /usr/lib64/qt5/bin was not in the PATH. To expect the user (or even the distrutor) to know this is asking too much of them, especially when the error message thrown by the crash is incomprehensible.

I cannot reopen this bug, but feel it should be reopened.
Comment 5 Christoph Feck 2017-01-04 21:32:43 UTC
Please add a comment if you still see this issue with Plasma 5.9.