KConfig borkage

[Oswald Buddenhagen]

On Tue, Oct 30, 2007 at 01:37:25PM -0500, Thomas Braxton wrote:
> On 10/29/07, Oswald Buddenhagen <ossi at kde.org> wrote:
> > procedure:
>
> this should now be fixed, update & let me know if anything doesn't
> work right
>
you did a good job on the the pre-approved parts. :-)
the tests pass. but i trust no test i didn't fake myself. :)

i'd still add the test that checks for the [$d] entries in the file
itself.

other stuff:
- KConfigGroup::revertToDefault - should use entryMap.revertEntry() (via
  kconfig indirection, presumably), no?
- i think some marginal cases like deleting localized entries, changing
  their globality and reverting them are not handled properly. somebody
  coming up with clear semantics on that in the first place would be
  good, though. :}
- the handling of opening kdeglobals explicitly seems weird, at least.
  couldn't it simply mask out IncludeGlobals and thus handle it as any
  other config without globals instead of spreading kdeglobals special
  cases all over the place?
- expandString - there are still some severere inconsistencies between
  data types. i started on that one, but suspended work when i noticed
  that i'm rewriting half of the backend api. need to revisit that ...

ugh, i should stop reading lib sources, otherwise the urge to rewrite
major parts overcomes me again - and we don't want aaron doing something
nasty to me, right? :)

-- 
Hi! I'm a .signature virus! Copy me into your ~/.signature, please!
--
Chaos, panic, and disorder - my work here is done.