[Evolution-hackers] Re: dlopen / API hooks for evolution ...
- From: Frank Schönheit <Frank Schoenheit Sun COM>
- To: michael meeks novell com
- Cc: Jayant Balraj Madavi <MJayant novell com>, JP Rosevear <jpr novell com>, evolution <evolution-hackers ximian com>
- Subject: [Evolution-hackers] Re: dlopen / API hooks for evolution ...
- Date: Mon, 31 Jan 2005 15:45:08 +0200
Hello Michael,
> Frank - I'm assuming here that it's ok to directly link to the glib API
> - splitting that out would be an even bigger pain, and I believe that's
> a platform pre-requisite for OO.o now.
According to Heiner Rechtien: No way. His points are
- we're developing C++, not C, so what do you need from this lib what
isn't (semantically equivalent) available elsewhere?
- glib has just recently been removed - the only thing still depending
on it is the Gnome integrator. RE would be quite unhappy with
re-introducing it.
- an architectural change such as this (one of the libs depending on
glib) needs approval by at least Matthias Huetsch (and is unlikely to
get it)
- This would create yet another "version hell", if we'd do this.
So I fear we need to find another solution :(
What kind of code do you need from glib?
> d) call EApiInit() as early as we can at a well defined
> point before we use any e_ API
> e) throw an exception if that init fails
I suggest the driver, which is the initial entry point, anyway. It would
do an EApiInit in it's first |acceptURL| or |connect| call. In the first
case, it could silently return sal_False if the init fails, in the
latter, it could throw an exception.
> JP - since you're ultimately responsible for the gratitous ABI breakage
Thanks & Ciao
Frank
--
- Frank Schönheit, Software Engineer, OpenOffice.org Database Access -
- frank schoenheit sun com http://dba.openoffice.org -
- -
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
[
Date Prev][
Date Next] [
Thread Prev][
Thread Next]
[
Thread Index]
[
Date Index]
[
Author Index]