From kde-core-devel  Fri Mar 07 18:32:17 2003
From: Waldo Bastian <bastian () kde ! org>
Date: Fri, 07 Mar 2003 18:32:17 +0000
To: kde-core-devel
Subject: Re: glib dependancy in KDE3.x
X-MARC-Message: https://marc.info/?l=kde-core-devel&m=104706206502983

On Friday 07 March 2003 18:50, Neil Stevens wrote:
> On Friday March 07, 2003 09:43, Alexander Kellett wrote:
> > yoyo,
> >
> > On Fri, Mar 07, 2003 at 06:28:35PM +0100, Waldo Bastian wrote:
> > > But this is all pretty much limited to applications that link directly
> > > to arts, which aren't that many.
> >
> > thats what i'd thought, but according to neil's
> > interpretation stefan says otherwise. most confusing.
>
> Stefan just says apps that link to arts will link to glib.  I say that
> knotify and konqueror link to arts, therefore every user ends up linking
> to glib.

I don't think konqueror does actually. Maybe some plugin does. Anyway, that's 
not really important.

Yes, every user ends up linking to glib, but if arts where to add some 
arts-specific C++ code to arts it would just as well end up with every user. 
Or e.g. if arts would decide to use C++ exceptions it could easily have a 
larger impact on memory usage.

The issue would have been much more important if it would affect the link 
dependency of every application because then you suddenly need to multiply 
effects of e.g. link-time and non-shared memory usage with 20 or so to get 
the numbers for a base desktop and consider the negative impact on 
scalability in general. (Think thin clients, where you run 50 or 100 clients 
from a single server.)

But this is just a one-time hit that IMO isn't any worse than e.g. an 
additional konqueror plugin.

Cheers,
Waldo
-- 
bastian@kde.org -=|[ SuSE, The Linux Desktop Experts ]|=- bastian@suse.com