Re: [Evolution-hackers] OO.o / evolution integration ...

[Fridrich Strba <fridrich strba bluewin ch>]

Michael, one possible way to fight the "voracious demands of 
bureaucratic process, and the commensurate stifling of the will-to-live" 
from Hamburg would be writing a wrapper library that would be integrated 
in the e-d-s and maintained there. If carefully designed, the API of 
this library could be very stable. I assume that the OOo uses just basic 
addressbook operations. The only thing would be to assure (inside the 
e-d-s) that this library is calling the right functions of the 
libedata-server.

If the proposal is not making sense, just forget about me.

Cheers

Fridrich

Michael Meeks wrote:
> 	Ultimately - it's going to be _extremely_ hard to convince Sun to use
> this / ship with it, if we are going to have to write an extra
> compatibility layer per-different-version of e-d-s; and (of course) it
> will involve yet more development effort we don't have time for.
> 	Why was this change done ? is it going to happen again ? how
> regularly ? and what is the recommended course for OO.o[1] ? how can
> this be made future compatible - dlopening libedata-server-%d.%d
> starting with 1.8 and moving down ? [ and then grabbing symbols out etc.
> as is done for many other unstable platform libraries ].
> [1] - a product where a single binary has to be engineered to install on
> multiple platforms - ( where the merits of long term ABI stability are
> deeply appreciated ).