Re: [Evolution-hackers] Roadmap Feedback (LAST CALL)

[Not Zed <notzed ximian com>]

I thought the main reason we didn't do it like this is that, well, gconf sucked and we were told this might not work very well.  I thought you could attach a schema at every level - i mean its not like we're using the schema stuff anyway for anything very useful.

The solution is of course, to replace gconf, again, but well.  Perhaps not.

The current xml blob hack is a nasty hack.  It turns a pretty mediocre solution into a clearly bad one.


On Mon, 2004-08-02 at 13:44 -0400, JP Rosevear wrote:
> On Mon, 2004-08-02 at 13:07 -0400, David Malcolm wrote:
> > On Mon, 2004-08-02 at 12:31 -0400, Dan Winship wrote:
> > > On Mon, 2004-08-02 at 11:52 -0400, David Malcolm wrote:
> > > > On Wed, 2004-07-28 at 16:56 -0400, JP Rosevear wrote:
> > > > > Last call for Evolution road map feedback.
> > > > > 
> > > > > -JP
> > > > 
> > > > Somewhat late here, but something I'd like to hack on for 2.2 is a
> > > > rewrite of the way Evolution stores its configuration information - last
> > > > time I looked, it stores hierarchical configuration for the various
> > > > objects as single strings in GConf (containing XML streams).  I'd prefer
> > > > it if the hierarchy was stored in expanded form within the GConf
> > > > database as separate keys and values.
> > > 
> > > The problem with this is that there's (currently) no way to express this
> > > in a GConf schema. (You could specify properties for a single account,
> > > but the only way to have an arbitrary number of something is to have a
> > > single list-valued property like we do now.)
> > 
> > Or have the things as subdirectories/subfolders within GConf, and set up
> > the schema on-the-fly based on the type of the object?
>
> Possibly, but then you can't get atomic notifications for the object as
> a whole afaik.
>
> -JP

-- 







Michael Zucchi <notzed ximian com>

"born to die, live to work, it's all downhill from here"

Novell's Evolution and Free Software Developer