Re: Bonobo Unique Application v5
Michael Meeks <[email protected]> Fri, 14 Nov 2003 12:52:45 +0000
| Newsgroups | gmane.comp.gnome.components |
|---|---|
| Organization | Ximian. |
| Message-ID | <[email protected]> |
Hi Gustavo, On Thu, 2003-10-30 at 22:27, Gustavo J. A. M. Carneiro wrote: > Here's a fresh new patch to implement the "unique application" feature, > already discussed in this list. OK, unfortunately I havn't had time to review this as thoroughly as I'd like. However - I'd like the functionality, if you can commit time to fixing a few things up before 2.6 then please commit it with the following changes: * bin redundant g_warnings on misc. methods * remove over commenting - just a few too many here and there * the new-instance thing should be a special case separate method on the IDL I think (discuss this one) * would it be nice to have a 'getAppName' thing on the interface for a GUI tool to allow poking at running apps ? > Minor Changes: > > 1. No longer uses BonoboArg<->GValue conversion functions, since their > API basically sucks; instead I created new functions for the same > purpose, albeit based on the old ones. If you have a better API for this (doing allocation) please expose it in bonobo-arg.h (I'd like to see the patch for that first). Also, we should really add an API to allow register/unregister external conversion functions: bonobo_arg_add_conversion (GType, to_arg_fn, from_arg_fn); or somesuch; so we can do more powerful stuff nicely (perhaps in libbonoboui) eg. GdkPixbuf <-> CORBA_any (etc.). > 2. A GValue containing a CORBA::any is now transported through CORBA > as one CORBA::any inside another CORBA::any. Sounds fine. My other concern is that we need to beef up the bonobo-activation-server stuff. I'd quite like you to write a tweaking tool that allows people to list the running servers, and poke at their messages [ if that's feasible ]. I'm particularly interested in the multi-display, multi-lang / that sort of angle on all this. Currently b-a-s is not good at dealing with these sort of uniqueness factors in a sensible way I think. Anyway, thanks for this - it looks nice :-) Regards, Michael. -- [email protected] <><, Pseudo Engineer, itinerant idiot