Using StyleSheets is bad WAS Re: Konqueror lineedit Bug

Aaron J. Seigo aseigo at
Tue Nov 20 20:30:32 GMT 2007

On Tuesday 20 November 2007, you wrote:
> A Dimarts 20 Novembre 2007, Aaron J. Seigo va escriure:

first the work related part:

> > and/or fixing things with 10 line patches, we're
> > going to end up writing custom proxy styles for everything we want to do?
> It *IS* a 10 line patch, if you don't count the proxy that actually is dumb
> code.
> So, can we have this patch in for KDE 4.0 so things actually work or should
> we just pray for uncle TT to be good and fix it fast?

ignoring the sarcasm in your email: sure, i'm fine with; can you please add 
a //FIXME in there to make it clear why this is like it is and what needs to 
happen to have it replaced?

my concern is that we end up with these hacks all over the place in our code 
which makes it uglier and not as easy to maintain.

the social / attitude part:

> I'm beginning to be fed up of this "Trolltech employee trend" telling us to

my viewpoint isn't limited to Qt, has nothing to do with my affiliation with 
Trolltech and i do the same thing to people in Trolltech when it comes to 
things like phonon. ;) you may have noticed i also take this same sort 
of "fix it in the right place" position in many other areas, e.g. with 
okular: we shouldn't have a stand alone ps app shipped with kde if we can put 
that into okular properly; that's "fixing it in the right place".

> not do anything because Qt is going to do it for us, but not now, just in
> an uncertain future.

erm. the last time a CSS related bug got reported, we got a patch for qt-copy 
within the week. no uncertain futures; though the future is absolutely 
uncertain if we don't try and communicate.

there's a reason we use libraries, and with open source one of the nice things 
is that we can fix them.

> Girish is probably *very* responsive to you, probably you even have his
> phone number and can phone him directly,

no, i use his publicly available email address like any other mortal.

> Even in the case they accept it to be a thing that should not happen it'll
> be a wish which will probably make it be "fixed" on Qt 4.4, 4.5 or never.
> That means KDE 4.x will have that bug until TrollTech decides to fix this,
> and given the severity of the bug, i think this is totally unacceptable.

hm. see, i don't have that experience. this is *why* Trolltech put forward 
individuals as KDE liaisons, to get these issues triaged properly. *use 
them*, don't complain.

if we wish to talk about things we're all fed up with, it's the attitude of 
KDE developers who view Trolltech (and other upstream providers of 
technology) as hostile-by-default. when we take a moment to learn how to work 
with these other groups, great things can and do happen.

like when i noted that CSS was stomping on margin hints in line edits and 
girish had a fix for it wthin the week, rather than applying Casper's patch 
to kdelibs to hack around the bug in Qt. it's not that Casper's patch didn't 
work, it's that it was a fix to the wrong thing.

i dunno about you, but if i can get bugs fixed upstream rather than hack 
around them in my own code, that sounds like a win. especially since if this 
is a problem in CSS support, everyone who uses it will end up with these 
problems. fix it somewhere that makes sense if at all possible and we get a 
lot more out of it.

Trolltech is trying to be more open and accessible; i (along with others) am 
trying to poke them along even further; i'm also trying to encourage the 
necessary responding behaviour from this community.

there's interesting, an unexpected, amounts of pushback from both sides. we 
all need to take a slightly bigger picture view of how this relationship 
should be working.

> > i understand that probably feels much more hard core and therefore makes
> > one feel good .. but, it seems to me that addressing these issues
> > upstream or patching things properly seems a lot more ... sane.
> That sentence reads like a personal attack, 

it isn't. it's a statement about how some solutions are more correct (fix 
something upstream) while others are much more "hack it in for now" which 
while gratifying in the short term leads to crap in our libraries over time.
sorry if you took it personally, it wasn't about you.

Aaron J. Seigo
humru othro a kohnu se
GPG Fingerprint: 8B8B 2209 0C6F 7C47 B1EA  EE75 D6B7 2EB1 A7F1 DB43

KDE core developer sponsored by Trolltech
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
URL: <>

More information about the kde-core-devel mailing list