[PATCH] Re: KMix

[Frans Englich]

On Wednesday 24 December 2003 19:43, Aaron J. Seigo wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> On Wednesday 24 December 2003 08:14, Christian Esken wrote:
> > - MainWindow hidden / MainWindow shown / MainWindow minmized (iconified)
> > - User allows KMix to use Systray / User does not allow to use Systray
> > - Quit via CTRL-Q / Quit via Close-Button / Quit via Logout
>
> if you already know which ones work and which ones don't, that'd be great
> (as in: save me some time). otherwise, i'll do it myself between Christmas
> dinners and drinking ;-)

I can't reproduce this bug. This is the steps I do:
1) (Having no kmix running)  I start kmix.
2) Change the levels in the mainwindow.
3) Quit(and I don't forget quitting the applet too).
4) Start kmix again and checks if the levels are as I left them. And they are.

But.
KMixWindow::quit() is AFAICT never called. A KStdAction is connected to it but 
that one is only activated when the user uses the menu or the toolbar ie. 
AFAIK.
The attached patch simply renames the function to close() which QWidget 
has(quit() is no member of QWidget, but QApplication, perhaps they were mixed 
up somehow), and close() is called on shutdown. This is perhaps the reason to 
the latent save bug. Ofcourse, it doesn't necessarily means all aspects of 
the bug(or any..) is fixed.
But considering my shallow C++ knowledge I could be totally wrong.

Further, my initial patch could *really* not have broken saving. Seriously. 
All it did was removing the possibility to manually save. But it could have 
made the save-bug more visible since the user( Christian :) ) couldn't 
explicit/manually save. And that we should be thankful for! :)

> this is fixable (and will be fixed) and has nothing to do with the actual
> issue of whether or not it's a good idea to do it this way or not...

Yes, but Christian otoh didn't say he would revert or disliked the patch. He 
mentions that it's bad that "restore to default" is not available, that's a 
slightly different matter but the argument is anyway wrong(explained in the 
lengthy KMix threads).
Further, there haven't been one single valid argument to why my patches is 
wrong. All our discussion have been noisy and repetative. I've found one 
argument(never mentioned) for why my proposal is bad: The documentation 
mentions the kcm_kmix and kmix functionality I remove. So, in case the user 
reads the manual from beginning to end and then goes looking for the 
functionality there will be confusion. It's a priority issue.


			Frans



-------------- next part --------------
A non-text attachment was scrubbed...
Name: kmix_quit_bug.diff
Type: text/x-diff
Size: 1306 bytes
Desc: not available
URL: <http://mail.kde.org/pipermail/kde-core-devel/attachments/20031226/1822e049/attachment.diff>