-
Notifications
You must be signed in to change notification settings - Fork 0
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
pypy alongside cpython + base/interpreter for noarch:python packages #1
Comments
That is a decision for conda to make. As packaged in the PyPy-provided tarballs, PyPy's stdlib and site-packages paths are different, but not for any good reason. Typically packages like debian, arch modify this as they wish. If conda wishes to put the two iplementaitons into the same environment they should ensure the layouts are different, and that the typical symlink of |
Would this design be valid for building feedstocks that use c-extension modules like NumPy and lxml? They cannot be shared across implementations. |
yes. Those packages would use the python-devel package shown in the recipe here. That would add a runtime dependency on either pypy or cpython to prevent ambiguity and problems. It might need tweaking beyond the example here, but the rough idea here should work. |
Good to know, thanks!
👍 TBH, that was what I anticipated to hear. IMHO, there is not too much value in having CPython and PyPy in the same environment. |
In reference to
@mattip in conda-forge/conda-forge.github.io#867 (comment):
Will Conda packages for CPython and PyPy themselves have path conflicts, e.g., both having
bin/python
orlib/pythonX.Y/some-file
? IIRC, PyPy uses a differentlib/some-thing
path, right?If there are no conflicts, we could potentially let them be installable side by side.
How would
conda
handle itsnoarch: python
packages for PyPy? Currently, it usespython.exe
orbin/pythonX.Y
. Either thepypy
package, or thepython=*=pypy
package should provide those binaries. If we put them into thepython
package, we could do the same forcpython
and ensure exclusivity. (I guess this may put off some people sincecpython
wouldn't be considered "complete" then. Well, I'm just putting down thoughts as them come right now and still lack a view of the whole picture...)The text was updated successfully, but these errors were encountered: