Oh and one more thing, i had to manually add geos and geos_c.dll someplace 
thats also in the system path of course, i just put them to c:\python33\dlls. 
Since they are not included in the git repo, i downloaded the appropriate egg 
file for shapely 1.2.18 for py 2.7 for windows 64bit and got them from there. 
Otherwise shapely would complain about not being able to load these dlls.

On 27. avg. 2013, at 15:18, Mike Toews <mwto...@gmail.com> wrote:

On 27 August 2013 06:35, Sean Gillies <sean.gill...@gmail.com> wrote:
> I intend that Mike's (mwtoews) branch will become the new Shapely 1.3.
> Details at:
> 
> https://github.com/Toblerity/Shapely/issues/56

It looks like there are folks looking to have Python 3 support (e.g.
https://github.com/SciTools/cartopy/pull/308). How should we proceed?
I have two branches on the go, so let me describe them.

py3: https://github.com/mwtoews/shapely/commits/py3

I initially had a `py3` branch, which had the minimal changes to
support both Python 2 and 3 from a single code base. The biggest
change here, besides the obvious, is the switch in setup.py from
setuptools to distutils. (Somehow, another branch of mine, `v13`,
slipped into this branch for a few weeks up until now.. sorry if you
pulled this and it is different. Admittedly, I'm still learning the
dark art of git, which I also used to restore the branch to an
intended previous state, plus merging the latest commits from upstream
master.) This branch should be the prime place for testing to make
sure most of the Python 2/3 features work. For instance, do speedups
work for everyone?

v13: https://github.com/mwtoews/shapely/commits/v13

The `v13` branch is a somewhat parallel branch to implement some more
aggressive changes for version 1.3, including reader/writer classes
for WKT/WKB formats, use of GEOS version numbers (rather than C API
version numbers) to determine capabilities, and a few other things.
(At present, it is a few commits behind upstream master, but I can
merge this). The wkt output is different from this branch, and I'd
like to do some more testing to see if the performance is
good/reliable.

Furthermore, with both branches I'm using Travis CI to do an automated
QA/QC, which are mostly looking good. Except for some GEOS
functionality mismatches, which can be skipped with some features from
a newer unittest package
(https://github.com/Toblerity/Shapely/issues/70), which are yet to be
committed pending on any comments on the direction of unit tests.

Could I suggest testing/merging py3 this week? Then some parts of v13
after that? As always, please test and communicate anything good/bad
to https://github.com/Toblerity/Shapely/issues/56 otherwise I don't
know if I'm breaking things or not.

-Mike
_______________________________________________
Community mailing list
Community@lists.gispython.org
http://lists.gispython.org/mailman/listinfo/community
_______________________________________________
Community mailing list
Community@lists.gispython.org
http://lists.gispython.org/mailman/listinfo/community

Reply via email to