[Patch] fix for the ultra nasty #68954, and this one works

[Otto Bruggeman]

On Tuesday 04 May 2004 20:29, Leo Savernik wrote:
> Am Dienstag 04 Mai 2004 15:28 schrieb Petter E. Stokke:
> [...]
>
> > > Huh? When this code was enabled, almost every webpage I would click on,
> > > would crash.
> > > gcc-3.3, but I don't think this is related.
> > > Maybe the Qt version? But I guess we all use qt-copy.
> >
> > FWIW, I just commented out the "return false;" as above, and now
> > selection works beautifully. That's with GCC 3.3.3, Qt 3.3.2 and
> > yesterday's KDE HEAD.
>
> Ok, let's summarize: For two people, the code crashes; for two people, it
> works beautifully. So what does this tell us?
>
> I suggest a poll on kdedevelopers.org:
>
> Does KHtmlView::moveCaretTo crash for you on every page?
> ( ) yes
> ( ) no

I was the one who complained to dirk that it crashed continuously about the 
day after that you committed the code. Dirk suggested adding some box && to 
the if statements in that method and odd as it may seem that seemed to fix 
it. I wanted to send the patch[*] to the list but i forgot about it and then 
dirk totally disabled it so i did not bother anymore.

And at that time i was using SuSE's gcc 3.3.2-26 rpms, qt-copy 3.3.1 iirc and 
the rest basically a SuSE 9.0 system.

But initial testing with the code from after david's remove return commit 
seems to indicate that it works just fine now even without my "patch". But 
i'll do some more testing tonight.

Otto

[*] I'm hesitant to use the word patch for it since it does not do anything 
special since box is checked for in the for loop and i cant find anything that 
might set box to zero in there.