Re: [evolution-patches] Bug 21974: Make editor component configurable

[Not Zed <notzed ximian com>]

On Fri, 2004-01-09 at 10:29 -0600, Jason Hildebrand wrote:
> On Thu, 2004-01-08 at 20:35, Not Zed wrote:
> > looks good.
> 
> > the only problem i have with it is what happens when the editor
> > version
> > changes?  you would be left with the configuration value that was
> > previously saved pointing to the wrong thing.  or is the assumption
> > that
> > since its hidden, if you change it its your own fault?
> 
> That scenario had crossed my mind as well.  My original thinking was
> that the user just has to deal with it themself, as they were the one
> who messed with the key in the first place.  Since it's the power users who
> will be changing editors, they can probably figure it out.
> 
> However, here are two ideas for how we could reduce the impact of the problem.
> 
> 1. Embed a version number in the gconf key.  I.e. instead of looking for the 
> key "editor_oafiid", we look for "editor_oafiid_1".  This number can then
> be bumped whenever the default OAFIID is bumped.  
>     CONS: - gconf pollution
>           - power users get their editor set back to default whether or 
>             not they want

I don't think this is a practical solution.

> 2.  Having the key set to the wrong version will likely manifest itself in not
> being able to instantiate the editor component.  If the instantiation fails
> we could check if the gconf key differs from the default.  If it does we could
> pop up a dialog and offer to reset to the default editor (i.e. delete the
> key).
>     CONS: - would not catch the problem if the previous gtkhtml version is 
>     still installed

How about a variation of 2:

 if we can't activate it
 and the comnponent iid is different
 fallback silently (or with a g_warning) to the source default

I'm a lover of 'graceful degredation', i.e. offer something that works
rather than nothing or a crash.

(and i'm not a fan of popups).

> I don't know if either of these ideas would help all that much -- I think
> I'm still inclined just to leave the problem up to the user.  
> 
> Thoughts? 
> 
> peace,
> Jason
> 
> --
> Jason D. Hildebrand
> jason peaceworks ca
> 
> _______________________________________________
> Evolution-patches mailing list
> Evolution-patches lists ximian com
> http://lists.ximian.com/mailman/listinfo/evolution-patches