KDE/kdelibs/kdecore

[Benjamin Meyer]

On Friday 02 December 2005 11:46 am, Ian Reinhart Geiser wrote:
> On Friday 02 December 2005 09:29 am, Oswald Buddenhagen wrote:
> > On Fri, Dec 02, 2005 at 01:20:22PM +0000, David Faure wrote:
> > > SVN commit 484988 by dfaure:
> > >
> > > Ingo's wish is my command: adding writeEntry(key,QList<QByteArray>)
> > > and readByteArrayListEntry.
> >
> > i think inflating the KConfigBase api with increasingly specialized and
> > weird read*Entry and writeEntry methods goes into the wrong direction.
> >
> > the imo right approach is having to/fromString methods in the classes
> > themselves. the remaining classes can have marshallers in KConfigBase
> > or (preferably?) some other class - that would particularly apply to
> > QList<>.
> > in any case it is essential to have the un-/marshallers available
> > separately, so one can nest them freely. i missed such a possibility
> > repeatedly.
> >
> > the {read,write}Entry core should be template-based, with explicit
> > instanciations for the common cases (PODs, string lists and various
> > other Q types - well, the sane part of the current api).
>
> Really I think the QSettings direction is way more useful to me as a
> developer.  With the current KConfig api I have to remember a million
> methods, and if they even exist.  With the QSettings api, i just have to
> remember that the value is a valid QVariant.
>
> Just my 2c though.
>
> Cheers
>  -ian reinhart geiser

I would have to agree, Qt4's QSettings simplification to only having a value() 
and setValue() make much more sense then attempting to add every possible 
case.


-Benjamin Meyer

-- 
aka icefox
Public Key: http://www.icefox.net/public_key.asc
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
URL: <http://mail.kde.org/pipermail/kde-core-devel/attachments/20051205/7b812910/attachment.sig>