On Wed, Apr 4, 2012 at 14:27, Thomas Petazzoni <thomas.petazz...@enix.org> wrote: > Le Wed, 4 Apr 2012 14:23:53 +0200, > David MENTRE <dmen...@linux-france.org> a écrit : > >> Regarding dependency in README, I would prefer to depend on an >> official version (2.0, 2.1, ...) than on a random commit. > > A commit hash is not a random commit, it's a precise commit. Just like > the SVN revision numbers we use for mapnik-osm, for example. > > As of today, since Mapnik 2.0 cannot support our Printable stylesheet, > and Mapnik 2.1 isn't released, I think the best solution is to just > update the documentation to use the Mapnik Git, and give the commit > hash we've tested, as Jeroen proposed. > > Thomas
Sounds good. For the record, I've just tested python-mapnik as supplied by oneiric, and it both works with 'import mapnik' and passes my assert. And Thomas is right, the git SHA hashes are to git what r12342 is to SVN, so telling someone to 'git clone ....; git checkout [hash]' gives us a tested version, whereas 'git clone ....' automatically checks out HEAD which is the newest commit, therefore random. I'll send a new version of the patch momentarily. Thanks Jeroen -- ↑↑↓↓←→←→BA[Start]
