Backport KPDF changes to 3.3?

[Frans Englich]

On Wednesday 25 August 2004 22:06, Matt Rogers wrote:
> On Wed, Aug 25, 2004 at 09:52:54PM +0000, Frans Englich wrote:
> > On Wednesday 25 August 2004 21:39, Albert Astals Cid wrote:
> > > Hi, i would like that you try the changes i have made into kpdf this
> > > afternoon, they fix lots of rendering problems (i've already closed 7
> > > bugs and think i can close some more)
> > >
> > > They work fine for me but i would like some more testing.
> >
> > Wouldn't it be good if there was a regression suit for this kind of
> > things(collection of PDFs in CVS)? Testing will always be needed, the
> > code base may radically change, and asking a mailing list is never a
> > bullet proof or consistent way of testing.
>
> sure, it'd be great if there was a regression test for this kind of thing,

I don't know of such a thing(but I bet Adobe adobe have it, they would be 
interested in open sourcing it..hmm..) I thought an ad hoc collection of 
attachements on bug reports etc. would grow as a useful resource(at least 
over time) -- better than nothing

> but a collection of pdfs in cvs is a bad idea. why? because in order to an
> adequate test suite for the possible problems an update might cause, the
> pdf collection would be huge.

Yes, what about FTP? If it's a question of an "ad-hoc" collection, it probably 
would probably not be updated, and that evens FTP/CVS out I think(without 
taking an overly technical discussion). But such a regression suite should be 
a SF project(as in independent and not KDE tied), at least in the long run..

>
> the testing doesn't necessarily have to be consistent, but mailing a list
> asking the people to test something, imho, is a great idea. The more people
> that test the various pdfs they might have, the more bugs found that can be
> worked out. it's part of that whole "more eyes (in the end) means less
> bugs" thing.

Yes, it's definitely better than nothing.


			Frans