Re: Towards better OAF/Bonobo integration
- From: Maciej Stachowiak <mjs noisehavoc org>
- To: Havoc Pennington <hp redhat com>
- Cc: Maciej Stachowiak <mjs noisehavoc org>, michael ximian com, gnome-components-list gnome org, gnome-hackers gnome org
- Subject: Re: Towards better OAF/Bonobo integration
- Date: Sun, 29 Jul 2001 01:59:28 -0700
On 28Jul2001 11:36PM (-0400), Havoc Pennington wrote:
> What are the implications of this for backward compat? (i.e. what
> happens if I have both oaf1 and bonobo-base client applications
I think I posted earlier on gnome-components-list what my plan for
coexistence of OAF1 and OAF2 is. I think that plan is not much
affected by whether or not this other change is made.
The short version of my coexistence plan is:
* Make the two oaf libraries parallel installable cleanly.
* Rename oafd for GNOME2.
* Put a special API in oaf2 for activating oaf1-based components by
talking to the GNOME1 oafd (the GNOME2 gconf client library could in
theory use this and gconfd could remain an oaf1 app; or perhaps there
is a race-free way to make a single gconfd that can be activated by
either GNOME1 or GNOME2 (my proposal would be to install an activation
record for both, always register with both, and use some sort of
filesystem-based locking to avoid races when GNOME1 and GNOME2 clients
are activating at the same time.
Now, I've heard fancier proposals for doing things like having a
special OAF attribute to specify GNOME platform version, having two
separate databases of activated components in one oafd, etc etc, but I
am really wary of doing something complicated because I think it would
be more work and more risk for probably not much benefit.
But I'm happy to hear other proposals.
gnome-hackers mailing list
gnome-hackers gnome org
] [Thread Prev