Resolve dependencies of local wheels #1944
Conversation
|
@vphilippon @uranusjr @jtratner @ncoghlan would one of you mind taking a look at this (preferrably one of you + @vphilippon just in case this is a terrible idea) |
- Fixes #1937, #1683 - Modifies the resolved wheel info on the way into the lockfile to retain the relative path Signed-off-by: Dan Ryan <[email protected]>
techalchemy
force-pushed
the
bugfix/1937-resolve-wheels
branch
from
572578f to
a7e6cfa
Compare
| @@ -1128,6 +1128,12 @@ def do_lock( | |||
| # Add default dependencies to lockfile. | |||
| for dep in results: | |||
| # Add version information to lockfile. | |||
| pipfile_version = project.packages[dep['name']] if dep['name'] in project.packages else None | |||
There was a problem hiding this comment.
Does project.packages.get(dep['name']) work?
There was a problem hiding this comment.
Yes despite being a toml object this works fine
pipenv/patched/piptools/resolver.py
Outdated
| @@ -120,7 +120,7 @@ def resolve(self, max_rounds=12): | |||
| @staticmethod | |||
| def check_constraints(constraints): | |||
| for constraint in constraints: | |||
| if constraint.link is not None and not constraint.editable: | |||
| if constraint.link is not None and not constraint.editable and not (constraint.is_wheel or constraint.is_artifact): | |||
There was a problem hiding this comment.
What does is_artifact add to this check? (Asking because the title says this resolves wheels.)
There was a problem hiding this comment.
Ah yes. I set this up because I wasn’t sure if we wanted to resolve artifacts too. Thoughts?
There was a problem hiding this comment.
Doing that sounds plausible, but it would probably make sense to pursue it as a separate PR.
There was a problem hiding this comment.
changed and pushed, will let it build and then merge
…lve-wheels Signed-off-by: Dan Ryan <[email protected]>
Signed-off-by: Dan Ryan <[email protected]>
|
Hmm are we on a minor version bump now? |
|
How much harder would it be to make this work on zip and tarball urls like in #772? |
retain the relative path
Signed-off-by: Dan Ryan [email protected]