#257: save list in memory to avoid reusing exhausted iterator#258
Conversation
|
Seems some Mypy assertion failed: At first sight it doesn't seem to be introduced by my change, I'll check later this evening to confirm or fix. |
|
@brettcannon I'll go ahead and make a bugfix release of packaging? |
Don't have packaging.tags.compatible_tags() reuse an iterable (pypa#258)
|
@pradyunsg SGTM, but should that wait until that issue in the pip repo involving what to do about PyPy wheels gets resolved (i.e. whether code will get put back in to add the old wheel names on top of the new ones)? |
|
@brettcannon @pradyunsg Is there a timeline for when this fix might get released? If it's still blocked on something, would you mind cross-referencing that issue from here, to make it easier to track progress? Thanks! |
|
The relevant pip issues are: pypa/pip#7626 and pypa/pip#7629 |
|
We could release a first patch version 20.0.1 now addressing pypa/pip#7626 and still release a 20.0.2 when a decision will be taken about PyPy wheels. |
|
I think I'm done for the day today - I'll do all the release packaging+pip work tomorrow. |
|
Following the precedence set by packaging 17.0/17.1, I've bumped the minor version number, instead of adding a patch number. |
This should resolve #257
The solution I proposed in the issue, I did not use, but I resorted to the same way this variable was handled in the other functions "generic_tags" and "cpython_tags"
The test added was also tested against the unmodified code to ensure failure.
Only tested with "tox -e py37", full environment list is still running.