Patch for bug #200795

[Germain Garand]

Le vendredi 11 décembre 2009, Andrea Iacovitti a écrit :
> Hi,
> I revisited the patch sent previously because probably I had not fully
> understood what was suggested by the KHTML developers... i hope now it's
> the case :-)

Hi Andrea, sorry for this very late review, I didn't notice your reply ;-/

> Function shouldDisableNativeBorders() now really handles the case where
> native borders have not to be painted (either because we have to paint css
> borders or we have not to paint borders at all because border:none)
> I added a partial fix (or workaround?) for bug #216795: force setting of
> qlineedit native frame (just in case previously disabled); may be the
> real/better fix here is to handle qlineedit widget borders through proxy
> style (if yes, i have not been able to achieve this :-( )
>
> Is that waht's expected?

yes, that looks fine, as well as the rest of your patch, except for

-    if (shouldPaintBorder() && style()->hasBorder())
+    if (shouldPaintCSSBorders())
     {
         paintBorder(paintInfo.p, _tx, _ty, width(), height(), style());
     }

as I believe the check for hasBorder() is valid here, if only for 
optimization's sake.

I'll apply it today, thanks.

>
> (I'm sorry if I'm only causing noise but I am new to khtml and qt4)
>

no by all means, if you ever feel the pain of Warnock's dilemma, just send a 
reminder, there's no reason for a patch to ever remain unreviewed.
(you may also try reviewboard.kde.org/groups/kdelibs, for maximum nagging 
effect :-)

Greetings,
Germain