Il 21/10/2010 00:46, Martin Dobias ha scritto: > I completely agree that having an infrastructure for easy development > and bug tracking for plugins would help the authors with deployment > and cooperation.
It seems that we all agree on this: if there are different opinions, please stand out. > - what users should be granted write access to this SVN? My suggestion is to take a liberal approach, allowing all those who ask and accept the licence and guidelines (maybe the guidelines should be simplified in this case? I have to check). We can always remove malicious users in case of trouble, and remove their code. > - who would manage the permissions to the individual plugins? I agree with Frank Warmerdam that having fine-grained permissions is not worth the fuss. Reverting unwanted changes is simple enough, and the shame of making public mistakes should prevent most problems. > - should we require all plugins to be within this SVN repository or > allow 'external' ones? My position is: this repo should be the standard, embedded in QGIS, but the user should be allowed to add any additional repo by hand. I may be wrong, of course. > - how to deal with official/contributed plugins? To me this distinction does not make much sense: I see core plugins (in trunk) and additional plugins (in the separate svn). Among the additional there will be all sort of nuances, from the very strong but of limited use to the proof-of-concept. > We should maybe have a look around how other projects deal with the > infrastructure and security model for plugins - Firefox, KDE etc. Exactly: this is a big win IMHO. I would be happy if we could set up the svn+trac before the hackfest, and build up the webapp during it. All the best. -- Paolo Cavallini: http://www.faunalia.it/pc _______________________________________________ Qgis-developer mailing list Qgis-developer@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/qgis-developer
