[jira] [Commented] (ARROW-430) Python: Better version handling

["Rok Mihevc (Jira)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/ARROW-430?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17657464#comment-17657464 ] 

Rok Mihevc commented on ARROW-430:
----------------------------------

This issue has been migrated to [issue #15338|https://github.com/apache/arrow/issues/15338] on GitHub. Please see the [migration documentation|https://github.com/apache/arrow/issues/14542] for further details.

> Python: Better version handling
> -------------------------------
>
>                 Key: ARROW-430
>                 URL: https://issues.apache.org/jira/browse/ARROW-430
>             Project: Apache Arrow
>          Issue Type: Improvement
>          Components: Packaging, Python
>    Affects Versions: 0.1.0
>            Reporter: Uwe Korn
>            Assignee: Uwe Korn
>            Priority: Major
>             Fix For: 0.2.0
>
>
> We need better version handling for Python releases. {{setuptools_scm}} sadly conflicts a bit a with the workflow of conda-forge+git. In general, it seems that generating a version from a non-tagged git archive is not possible.
> For tagged archives, we can use git's {{export-subst}} functionality but that will result in non-deterministic tarballs from Github.
> My currently preferred (when not really clean but at least integrated into the release process) solution would be to read {{java/pom.xml}} and use this version number. For PEP-440 compability we would need to replace {{SNAPSHOT}} with {{pre1}}.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)