On Wed, Sep 30, 2009 at 07:39:41AM +0300, Xan Lopez wrote: > 2009/9/30 Josselin Mouette <joss debian org>: > > Le lundi 28 septembre 2009 à 16:11 +0300, Xan Lopez a écrit : > >> Yeah, Seed support is pretty much in the "experimental" camp for the > >> moment. It would be great if someone started a wiki page with the > >> biggest blockers, or started opening bugs and marked them with some > >> relevant keywords. I opened a few (mostly seed bugs) a few months ago, > >> but then got too busy with other ephy/webkit issues :/ > >> > >> Any volunteer? > > > > Can I volunteer for re-introducing Python support instead? > > :) > > It's not really possible to have python and seed support at the same > time, since they would conflict in their memory management of the > objects they see. And I really don't feel like spending time in python > makes any sense when the long term plan is to use javascript for > extensions. > > That's pretty much what I said in the email about the topic last time, > and I still believe the same thing :) Even though I have myself felt the pain of trying to translate my Python extension to Seed, I do feel that in the long run it's the right thing to drop Python. Personally I don't feel Javascript is a very good language technically, but it's a strong contender mostly because it's used in other browsers (especially FF). What I'd like to see someone do is work out the bindings necessary to write extensions in Vala. I actually gave up on using Seed for my bookmark sharing extension and have now written the most basic functionality in Vala. It's a very pleasant language to work with, and since it compiles via C it ought to be possible to write Epiphany extensions completely in it. That requires a .vapi file for the extensions API though. The way I've currently done it is to write the minimal amount of C possible and then call into Vala, where all the heavy lifting happens. Then the result is passed back to C for some minor post-processing. The nature of my extension allows this sort of division, but that's of course not true for all extensions. /M -- Magnus Therning (OpenPGP: 0xAB4DFBA4) magnus＠therning．org Jabber: magnus＠therning．org http://therning.org/magnus identi.ca|twitter: magthe
Attachment:
pgpAZHShKZYmC.pgp
Description: PGP signature