staikos at kde.org
Tue Aug 1 23:36:54 CEST 2006
This is not about removing one approach for another. This is about finding
the right fix, and as far as I can tell, there is no supporting evidence for
this commit you made. Please don't do this. If you are going to commit code
into our live repository, make sure that it is well founded. If someone asks
for an explanation, please back up the change. The commit log should even
explain. SVN ping-pong is really not an acceptable development model.
Quoting Andrew Walker <arwalker at sumusltd.com>:
> As I said feel free.
> On August 1, 2006 2:25 pm, George Staikos wrote:
> > Quoting Andrew Walker <arwalker at sumusltd.com>:
> > > I have no particular attachment to my fix so feel free.
> > > However, there are clearly problems with normalize
> > > as normalizing a QRect with a negative width should
> > > never change the absolute value of that width. if you keep
> > > the width positive then obviously normalize works fine.
> > Normalize does not touch anything to do with width or height. Please
> > re-read the code again. All it does is swap coordinates if they are
> > inverted. The problem is very clearly elsewhere. Unless I can see some
> > sort of evidence of where the problem is beyond what I have of the
> > off-by-one in division, I'm going to back out all the changes to that file
> > and put in a correction for that division error.
More information about the Kst