Merge python_version and python_full_version markers#388
Merge python_version and python_full_version markers#388neersighted merged 2 commits intopython-poetry:mainfrom
Conversation
radoering
force-pushed
the
merge-python-version-markers
branch
from
69a8b76 to
ed6418b
Compare
neersighted
left a comment
neersighted
left a comment
There was a problem hiding this comment.
Given how brittle this has been I'm a little nervous about your change as I'm not quite familiar enough with the subtleties here -- however the set of test cases look quite robust and I can easily reason about that part of the change.
@dimbleby could you PTAL?
|
I think this looks correct, I'll leave a couple of style quibbles. I continue to have a slight feeling of unease that this code has become so elaborate: I like a marker simplification as much as the next person, but there's so much of it! |
I don't necessarily disagree, but I think ultimately we get higher quality results if we try to make sure the markers are as consistent and clean as possible. I think our best tool for maintainability is robust and through testing, and incremental refactoring as better patterns make themselves apparent. That being said, yes there's getting to be a lot of it and if I don't pay attention for a few weeks I forget how it works again 😆 |
radoering
force-pushed
the
merge-python-version-markers
branch
from
ed6418b to
38a03a1
Compare
|
Kudos, SonarCloud Quality Gate passed!
|
It's just necessary for the solver to not try solving some bogus overrides. Solving not just for one environment is complex, but that's one of poetry's USPs. |
3 participants
Relates-to: python-poetry/poetry#5717
Reintroduces merging of
python_versionandpython_full_version, which was removed in #382 because of python-poetry/poetry#5717The
constraintof thepython_markeris converted viaget_python_constraint_from_marker()and thus should be consistent with the latest findings described in #385.