The move to libgda 4.0

[Hans de Goede]

Denis Leroy wrote:
> I've been investigating the move to libgda 3.99.8 (4.0 API) for Fedora 
> 11. The 4.0 API is apparently stable now, and based on the F-11 release 
> timeframe, upstream does recommend it. There's also a growing number of 
> projects working with the 4.0 API already and I've received a request 
> for this for an Anjuta plugin 
> (https://bugzilla.redhat.com/show_bug.cgi?id=479298).
> 
> There are currently 4 packages that depend on the 3.0 API: glom, 
> gnome-python2-extras, qof and gnumeric-plugins-extras. The port to 4.0 
> appears to be non-trivial 
> (http://library.gnome.org/devel/libgda-4.0/3.99/migration-2.html). We 
> can probably work out a patch for something like the gnumeric plugin, 
> but I don't know how much work would be required to port 
> gnome-python2-extras for example. Glom still uses the 3.0 API, but the 
> 4.0 port is almost finished in SVN.
> 
> 2 options here:
> 
> 1. Wait for glom to release its 4.0 API port, ping upstream for other 
> dependencies or work on patches. Not yet clear how much work is required 
> for this.
> 
> 2. Revive the compat-libgda package (currently a dead.package, we used 
> to have it for the 2.0 API when 2.99 was packaged) for the 3.0 API and 
> move libgda to 3.99. Would this require a new package review ?
> 

Depends on if there is more then just a name change. Are libgda3 and libgda4 
parallel installable (including their -devel) without requiring any hacks to 
libgda3 ? (We do not want to hack libgda4, as then we would need to carry those 
hacks for a potential long time).

If its just a rename a review is not needed IMHO.

> I'm strongly leaning toward 2, mostly because of the uncertainty of 
> option 1 plus the fact it will hold up other updates waiting for the 4.0 
> API and delay 4.0 API testing.
> 

+1 for option 2

Regards,

Hans