Hi Dragomir,

On Wed, 17 Oct 2012 23:38:28 Dragomir Dragnev wrote:
> So i tested to change default GTK2 theme from QtCurve to Qt4 from system
> settings and without kde restart tested mysql-workbench again with full
> success no more memory leaks or anything - just looking ugly..
> 
> So for temporary sloution without global change ot GTK2 theme:
> GTK2_RC_FILES=/usr/share/themes/Qt4/gtk-2.0/gtkrc mysql-workbench
> 

You've found the problem, thank you. :)

Could you please also try with 

        --force-sw-render
and
        --force-opengl-render

to see if it have any effect on leak?


> Dmitry what GTK2 theme do you use? I wonder is it some general bug in
> QtCurve, some messed up settings of QtCurve or some misuse of it by
> mysql-workbench.

I'm with Oxygen (kde-style-oxygen, gtk2-engines-oxygen, gtk3-engines-oxygen, 
kde-config-gtk-style) and I don't have any *-qtcurve packages installed.


> 
> Also I tested with qtcurve experimental version 1.8.14-1 (current:
> 1.8.12-2) with the same result:
> sudo aptitude -t experimental install kde-style-qtcurve kwin-style-qtcurve
> qtcurve-i18n

Looks like problematic package is 

        gtk2-engines-qtcurve

See more

        https://bugs.freedesktop.org/show_bug.cgi?id=54657#c4


> And finally I downgraded mysql-workbench from 5.2.44+dfsg-1~exp1 to
> 5.2.40+dfsg-1+b1 and the above solution does work for the memory leak
> problem but there pops-up another bug that is missing in 5.2.44: with the
> same query: select * from mysql.no_such_table;
> it prints the error that the table does not exist, but on the status line
> stays "Executing Query..." and now the program can not be closed any more
> and acts stragnly...

Sorry about that. Version 5.2.40+dfsg-2 is fixed in that regards.
For a while I couldn't backport fix for "hangs on exit" because upstream have 
no public code repository (VCS). :(

Finally last week I uploaded fixed 5.2.40 to "unstable" and just today it 
migrated to "testing" probably hours after you tried buggy version.


> Apparently in 5.2.44 this bug is fixed so I am going back with experimental
> and staying there...

:)

Thank you for all the troubleshooting and feedback.

Cheers,
Dmitry.


-- 
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Reply via email to