David Neary <[EMAIL PROTECTED]> writes:

> One question... do you think it would be worthwhile to make the
> API change, document what needs to be done to change from the old
> API to the new, and then have a number of people work on porting
> the plug-ins? Each plug-in in itself would probably be a small
> job, but the lot of them might take you a while.
> Would this be worthwhile? I would certainly do a few plug-ins,
> are there others who would be able to help out here?


Thanks for offering help on this huge hack.

While there is certainly help needed (my fingers start to break when
i try to type 'GTK_RESPONSE_FOO' :-), I don't really know how this
could be done by several people without temporarily breaking
half of the tree.

gimp_dialog_new()'s API has changed in an incompatible way so
unported plug-ins break. I thought about adding a temporary new
API first, port them step by step, and then do a perl hack on
the tree to rename all new functions back to gimp_dialog_new().
However this will hardly result in a nicely formated source,
they would al have to be edited manually again.
(guess why i hesitated so long before starting this ;)

After all, I think this API change is best committed atomically,
so I'll just continue hacking the plug-ins and commit everything
today or tomorrow.

Gimp-developer mailing list

Reply via email to