From kde-core-devel  Thu Jun 26 04:56:03 2003
From: Benjamin Meyer <ben () meyerhome ! net>
Date: Thu, 26 Jun 2003 04:56:03 +0000
To: kde-core-devel
Subject: Re: KConfig Bug?
X-MARC-Message: https://marc.info/?l=kde-core-devel&m=105660341211614

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Wednesday 25 June 2003 6:05 pm, Benjamin Meyer wrote:
> When I call KConfigBase::deleteEntry("foo") and entry foo exists in my
> global config a "foo[$d]" is created in the local config and from then on
> when I call readEntry("foo") it doesn't see the global config.  Now if I
> change my deleteEntry("foo") to deleteEntry("foo", false, true) the local
> config foo line is removed entirely and the global config shows up by
> default on the next readEntry.  Now here is the confusing part: the
> deleteEntry API doc states that the second bool statement:
>
> void KConfigBase::deleteEntry (
> const QString &   pKey,
> bool bNLS = false,
> bool bGlobal = false
> )
>  If bGlobal is true, the pair is not removed from the application specific
> config file, but to the global KDE config file.
>
> But it seems to be doing the opposite.  When false (by default) it will add
> the [$d] thus disabling the global entry and when true it removes the line
> entirely thus keeping the global entry.  Is this a bug?  After wading
> through KConfig the last few hours this is the best I can tell.  If so who
> do I report it too?

I have gone through the code in kconfig/kconfiginibackend/kconfigbase and I am 
pretty sure that to do what the api says it should do, line 1176 of 
kconfigbase.cpp should be changed to 
"aEntryData.bGlobal = !bGlobal;"  But I am nor the author nor any sort of an 
expert on kconfig and simply offer this up as a suggestion.

- -Benjamin Meyer

P.S.  After changing KAutoConfig to match what is stated in 
kdelibs/kdecore/KCONFIG_DESIGN I was "flamed" and told to revert.  If 
KCONFIG_DESIGN is incorrect and I should not be using kconfigbase class can 
someone please fix the design file because as a developer to spend time 
updating code to match the design doc only to be yelled at is really 
irritating.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iD8DBQE++nzr1rZ3LTw38vIRAlCrAJ0aYs8js4BZtIkGTGUvg4qjF70wswCgkDJC
NebHEw3KDwG7ycGUcI8rkk0=
=svcK
-----END PGP SIGNATURE-----