Simplify generation of test packages used in test_check #1208
+104
−183
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
dnicolodi
force-pushed
the
rm-setuptools
branch
from
1d76b1c to
05207ec
Compare
|
ping? |
This will allow to reuse it in later commits. Change the function to return a pathlib.Path instead of a string because that is what is used internally and because it is easy to turn that into a string when needed, but it is a bit more cumbersome to do the opposite.
These tests used `build` to invoke the `setuptools.build_meta` build backend on a package directory generated on the fly. The tests are only interested in the content of the package metadata and use variations of `setup.cfg` to generate the desired metadata. This can be simplified to the direct generation of sdist archives with synthetic `PKG-INFO` metadata files. This makes the actual metadata the tests are checking obvious and avoid two test dependencies. The tests simplification highlighted that two tests are actually testing the exact same thing. Remove one.
woodruffw
reviewed
Jan 9, 2025
| assert b"Metadata-Version: 1.1" in metadata | ||
| assert b"Name: test" in metadata | ||
| assert b"Version: 1.2.3" in metadata | ||
|
|
||
|
|
||
| def test_missing_pkg_info(archive_format, tmp_path): | ||
| """Raise an exception when sdist does not contain PKG-INFO.""" | ||
| filename = build_archive( | ||
| filepath = build_archive( |
There was a problem hiding this comment.
As an observation: we end up calling str(filepath) repeatedly, so maybe it's fine to have build_archive return a str instead of a pathlib.Path?
(No strong preference, however.)
There was a problem hiding this comment.
The reason for the change is explained in the commit message. In subsequent patches I'll use the Path object instead of the string. I've a few more cleanups ready to submit but I split them in smaller PRs for easy of review.
dnicolodi
force-pushed
the
rm-setuptools
branch
from
05207ec to
2ca55db
Compare
woodruffw
reviewed
Jan 9, 2025
There was a problem hiding this comment.
LGTM overall, thanks @dnicolodi! One small comment about a return type, which you can shoot down based on your taste.
(Thanks for doing this -- this should make test standups/runs decently faster.)
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
These tests used
buildto invoke thesetuptools.build_metabuild backend on a package directory generated on the fly. The tests are only interested in the content of the package metadata and use variations ofsetup.cfgto generate the desired metadata.This can be simplified to the direct generation of sdist archives with synthetic
PKG-INFOmetadata files. This makes the actual metadata the tests are checking obvious and avoid two test dependencies.The tests simplification highlighted that two tests are actually testing the exact same thing. Remove one.