Re: Getting libgnome* into shape

[Martin Baulig <martin home-of-linux org>]

[Hmm, my mail setup is still broken, so quoting the archives]

George> Anders, Jonathan and I have been discussing what to do to bring
George> libgnome* into shape quickly.  Our current thoughts are that this
George> could be finished within a fortnight and then frozen.

Michael replied> A fortnight ! this seems like an incredibly long time to me; and
Michael replied> for what gain ? what precicely is going to take a fortnight ?

Ok, just to make it clear for the non-english people on this list: a fortnight means
14 days and not "a long night" like I assumed before I looked into my dictionary.

So, as I releazed 1.102.0 "Roswell" - things were already FROZEN except for a few
minor API issues - and now George wants to break the freeze for two more weeks - this
means that we'd slip for more than 6 (!) weeks - and to make things even worse, he
realized this two weeks after I did a FROZEN libgnome and libgnomeui release.

To make things even worse, he actually committed his changes without asking - for
changes like this, I do not consider posting to the list and committing less than 3 days
later as asking - people had hardly any chance at all to comment on it, and we just had
a weekend.

M> It seems to me that forcing people to link against a load of
M> deprecated code - when we already have a clean separation into
M> libgnome1-compat is just a gobsmacking idea ? I mean ... I can see no
M> possible benefit from it whatsoever - apart from re-integrating a load of
M> crufty code that is best left outside.

I totally agree.

M> There are indeed some places where this is most useful, I agree
M> entirely with the following:

M> > *  bring back entries, remove selectors non-deprecated (george, anders)
M> >     (done: gnome-entry, gnome-file-entry)
M> >    (not-done: gnome-pixmap-entry, gnome-icon-entry)
M> >
M> > *  finish gnome-about (anders)

Yes, agreed as well - and they were already on my list of open API issues.

M>
M>
M> > *  s/bonobo_config/gconf/
M>
M>         Why !? rational ? this has been hammered out extensively - I
M> really object to this change - very strongly - are there any substantial
M> reasons for this ( large ) change ? this will consume considerable time
M> for precicely 0 gain - why delay freezing everything, and introduce the
M> manifestly broken gconf C client headers into the project again ? sigh,
M> this is really a very tedious suggestion.

Ok, I'm not alone on this planet :-)

M> > *  bring back gnome-triggers, gnome-sound (non-deprecated)
M> >    (done)
M> 
M>         This should be in libgnome1-compat - we don't want this API in
M> libgnome IHMO - it's not something we should be supporting ad nauseum.

And George added an API back to libgnome which is only used in deprecated (!)
code - Eeeeek.

M> > *  bring back gnome-score (non-deprecated)
M> >
M>        One could make arguments for this I'm sure - but privatizing it
M> would be good.

This was only used in gnome-games, so why not move it into gnome-games ?

-- 
Martin Baulig
martin gnome org (private)
baulig suse de (work)