On Tue, Jun 7, 2016 at 11:20 AM, William Stein <wst...@gmail.com> wrote: > On Tue, Jun 7, 2016 at 11:18 AM, Nathan Dunfield <nat...@dunfield.info> wrote: >> Erik, >> >> Interesting, I didn't realize that build-time (as opposed to run-time) >> dependencies were even *possible* with setuptools, even if it does require a >> hack. With SnapPy, we just ship the Cython-generated C/C++-files in the >> tarball that we post on PyPI. > > I was going to post that my understanding is that shipping > Cython-generated files is the current standard way of shipping Cython > code (for better or worse).
No, In fact if that were the case, cythonize would also be run as a sub-command of `setup.py sdist` (ensuring that the C sources are always included in source distributions without having to remember to manually write `setup.py build_ext` first--I mistake I've seen a few times. One also has to cobble together a way to switch between compiling from the C sources (if they're already available) and running and not require Cython in that case. There's no out-of-the box solution for this in either Cython or setuptools. Might be worth agitating for. >> It adds a bit to the tarball's size, but is >> more robust than requiring Cython. For example, there are sometimes >> incompatibilities between different versions of Cython, and it's a bit rude >> to force the end-user to upgrade or downgrade their version of Cython just >> so they can install some other package. >> >> In our setup.py we use code like: >> >> # If have Cython, check that .c files are up to date: >> >> try: >> >> from Cython.Build import cythonize >> >> if 'clean' not in sys.argv: >> >> targets = ['planarity_src/planarity.pyx', pmap_dir + >> 'planarmap.pyx'] >> >> cythonize(targets) >> >> except ImportError: >> >> pass >> >> >> Even if the user has the "wrong" Cython, the PyPI tarball will still build >> since the Cythonized files are up-to-date. >> >> >> Best, >> >> Nathan >> >> -- >> You received this message because you are subscribed to the Google Groups >> "sage-devel" group. >> To unsubscribe from this group and stop receiving emails from it, send an >> email to sage-devel+unsubscr...@googlegroups.com. >> To post to this group, send email to sage-devel@googlegroups.com. >> Visit this group at https://groups.google.com/group/sage-devel. >> For more options, visit https://groups.google.com/d/optout. > > > > -- > William (http://wstein.org) > > -- > You received this message because you are subscribed to the Google Groups > "sage-devel" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to sage-devel+unsubscr...@googlegroups.com. > To post to this group, send email to sage-devel@googlegroups.com. > Visit this group at https://groups.google.com/group/sage-devel. > For more options, visit https://groups.google.com/d/optout. -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at https://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
