[prev in list] [next in list] [prev in thread] [next in thread] 

List:       kde-core-devel
Subject:    Re: Question: KIO Scheduler
From:       Dawit Alemayehu <adawit () kde ! org>
Date:       2000-11-05 7:27:10
[Download RAW message or body]

On Saturday 04 November 2000 08:40 am, David Faure wrote:
> On Monday 30 October 2000 05:44, Waldo Bastian wrote:

[snipped my long attempt at an explanation]

> > > Thus, the question is where is the correct place to put this stuff such
> > > that when this static object is destroyed or about to be destroyed the
> > > cached entries, if any, can be cleaned as well...
> >
> > The scheduler gets deleted after the KApplication gets deleted, there
> > isn't a whole lot of stuff you are allowed to do at that point.
> >
> > It's like a house that falls apart and you are running for the door....

Hey!  That is what I would exactly be doing.  How else would I get out ??? :)))

> Yup. Does the new method for cleaning up when kapplication exits
> (the one with "Simon" in the debug message :-)
> help for that ? Maybe the house is a bit less falling apart at that point ?

I figured out the problem here.  The actual code to delete the cached keys still
works under the circumstances described by Waldo since the calls are going
to a still running daemon and the function calls are member functions of the
object being destroyed.  The problem was the very calls I put in to make sure
that this methods were executed on the other hand does not work: calls to kdDebug
in the dtor and the function(s) called from from the dtor that is.  Replacing these
with fprintf statements seems to show that the code actually works.  Now back to
re-modifying the caching code yet again to accomodate this deletions.  Some of
the previous assumptions did not quite work out :(  Thanks for the help..

Regards,
Dawit A.

[prev in list] [next in list] [prev in thread] [next in thread]