Bug#289481: acknowledged by developer (Re: Bug#289481: kdebase-bin: doesn't open file or html)
Alejandro Exojo <suy@badopi.org> writes:
> El Miércoles, 12 de Enero de 2005 20:01, Michael Below escribió:
>> As I said, I am sure you know what you are doing. But I had the
>> impression it might make sense to rebuild that ksyscoca database
>> during the upgrade 3.2 -> 3.3 in general. You said that someone else
>> had encountered a similar problem too, and solved it by deleting this
>> database?
>
> How did you performed the upgrade? Did you restarted KDE?
I don't think I was using KDE parts during the upgrade. Usually, I
have a fvwm desktop, and I am using aptitude (in rxvt) for
upgrades. The X apps I am using regularly are emacs, xmms and firefox,
I think they don't use KDE libs. For the last months I have been
tracking testing, doing upgrades every few days.
As far as I can tell, it was the next day when I tried to use a KDE
app (kmid or konqueror, I think) and it wouldn't work. Then I
remembered that there had been a large update in testing recently,
involving the KDE update.
> I'm no ksycoca expert, but AFAIK, it will be regenerated properly if you log
> out and log in again. If don't, simply run "kbuildsycoca" in a shell.
I don't think ksyscoca is being regenerated on each log in. There is a
status message in my xterm like "reusing ksyscoca" when I start a KDE
app for the first time. I guess you know this message -- I just saw it
when starting konqueror, but then it flooded the xterm with tons of
warnings like:
kbuildsycoca: WARNING: '/usr/share/applications/dia.desktop'
specifies undefined mimetype/servicetype 'application/x-dia-diagram'
(I didn't see them before deleting that ksyscoca file, but I didn't
notice any missing functionality, so I think I can ignore this?)
My xterm backlog buffer isn't big enough to keep that first message now.
If you want me to, I can log out and start a KDE app with stderr
redirection, then I can send you the exact words.
> Maybe i'm saying something stupid, but perhaps this regeneration can be
> triggered in postinst.
Thats what I thought (especially the first part ;)
Michael Below
Reply to: