RFC: KConfig XT (KDE 3.2)

[Maks Orlovich]

> Instead of relying on the defaults that are hard-coded in the application,
> rely on default configuration files being installed in $KDEDIR. The
> technical changes required for this are very minimal, it is mostly a change
> in policy.

One thing that is kind of worrisome is the potential for an increase in number 
of keys KConfig may need to store in its maps (from the profiling I did, I 
doubt comments would matter much, since the KConfig time is dominated by the 
map storage -- with the parser being blindingly fast), inducing a startup 
time penalty. I doubt it'd be anything to worry about, since we've been 
making enough across-the-board improvements to probably cover it, but I am 
just wondering whether anyone has any good ideas on how to assess the effect 
of something like this w/o spending a ton of time implementing things, so we 
have the numbers in front of us... 

Other than that, I think this is a great plan and goal.