regro-cf-autotick-bot not pulling pypi sha

[michaelbilow]

I'm not sure whether this is intended behavior or a bug, but it would be nice (at least for me) if the autotick-bot took the sha of the new version, for example:

conda-forge/issho-feedstock@fb6ef96 was pushed by the autotick-bot, but did not update the sha available at https://pypi.io/packages/source/i/issho/issho-0.5.1.tar.gz, the source url defined by the rendered meta.yaml.

[scopatz]

Normally it does update the sha256sum. I wonder what went wrong.

[michaelbilow]

Thanks for the response; let me know if there's something I can do to pitch in.

[scopatz]

If you can find the job that this failed on in the regro/circle_worker repo, that'd be great! It might be kind of a slog, though...

[chrisjsewell]

FYI, this just happened to me as well: conda-forge/pytest-notebook-feedstock@41ea768

[scopatz]

hmm maybe this is because the sha value is in quotes.

[scopatz]

In the meta.yaml, that is

[scopatz]

Specifcally,

cf-scripts/conda_forge_tick/migrators.xsh

Line 265 in 46008e5

p = '{}:\s*[0-9A-Fa-f]+'.format(hash_type)

and surrounding logic does not seem to capture quote characters. PRs welcome!

[ericpre]

Another instance, where the sha256 was not updated by the bot: conda-forge/matplotlib-scalebar-feedstock@31815f6.

hmm maybe this is because the sha value is in quotes.

In this case (conda-forge/matplotlib-scalebar-feedstock@31815f6), the sha256 is in quotes in the meta.yml file.

[CJ-Wright]

Yea it seems that quotes are a problem with our current regex.

[phil-blain]

I'm not sure if this is the same issue, but in the the recent PR for Git 2.24 the bot failed to update the hashes (in this case there are no quotes) :
conda-forge/git-feedstock@9c03fb8#diff-e178b687b10a71a3348107ae3154e44cL1-L25

conda-forge/git-feedstock#68

I'll open a new issue if people feel this is a separate issue.