MNT: migrate static metadata out of setup.cfg #145
Conversation
|
I missed that |
neutrinoceros
force-pushed
the
mnt/pyproject.toml
branch
from
dd5caa4 to
e38246a
Compare
neutrinoceros
force-pushed
the
mnt/pyproject.toml
branch
from
e38246a to
e4ad476
Compare
pyproject.toml
Outdated
|
|
||
| [tool.pytest.ini_options] | ||
| minversion = "4.6" | ||
| testpaths = ['"erfa"', '"docs"', '"README.rst"'] |
There was a problem hiding this comment.
I don't think this needs the single+double quotes -- at least, astropy just has the double ones.
erfa/_dev/scm_version.py
Outdated
| @@ -6,43 +6,55 @@ | |||
| import os.path as pth | |||
| from warnings import warn | |||
|
|
|||
| try: | |||
| from setuptools_scm import git, Configuration, get_version as _get_version | |||
There was a problem hiding this comment.
I'm a bit confused why this is needed, given that astropy has the simpler direct try/except that we had here. What other attribute is being gotten that the __getattr__ solves? Or is it just that timing now is different?
There was a problem hiding this comment.
I didn't have any recollection of it, luckilly the answer to your question is in the commit message:
RFC: refactor scm_version.py to survive pytest collection if setuptools_scm isn't installed
There was a problem hiding this comment.
So, isn't the answer then to ensure that pytest doesn't look at this directory? Is that what astropy does?
There was a problem hiding this comment.
I don't know. I couldn't reproduce the problem I was trying to solve here so I just reverted the change. The PR is more self-contained as a result.
neutrinoceros
force-pushed
the
mnt/pyproject.toml
branch
2 times, most recently
from
41cfa7a to
65619a0
Compare
Follow up to #144
migration from
setup.cfgtopyproject.tomlwas mostly automated usingini2toml