)]}'
{"/PATCHSET_LEVEL":[{"author":{"_account_id":13252,"name":"Dr. Jens Harbott","display_name":"Jens Harbott (frickler)","email":"frickler@offenerstapel.de","username":"jrosenboom"},"change_message_id":"86e27d39543c5216d52dd35acc26270245a29370","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":2,"id":"e3258bbe_935a471b","updated":"2025-09-12 19:13:49.000000000","message":"looks like the CI failure is real, the ordering of rc releases was changed\n\n```\nValidating uniform formatting on /home/zuul/src/opendev.org/openstack/requirements/global-requirements.txt\n-pysaml2!\u003d4.0.3,!\u003d4.0.4,!\u003d4.0.5,!\u003d4.0.5rc1,!\u003d4.1.0,!\u003d4.2.0,!\u003d4.3.0,!\u003d4.4.0,!\u003d4.6.0  # Apache-2.0\n+pysaml2!\u003d4.0.3,!\u003d4.0.4,!\u003d4.0.5rc1,!\u003d4.0.5,!\u003d4.1.0,!\u003d4.2.0,!\u003d4.3.0,!\u003d4.4.0,!\u003d4.6.0  # Apache-2.0\n```\n\nI think we might proceed with this early next cycle anyway, but not now while we\u0027re in freeze","commit_id":"c9608bf65764ddbf58d48ca3637a5cdbce7c8371"},{"author":{"_account_id":15334,"name":"Stephen Finucane","display_name":"stephenfin","email":"stephenfin@redhat.com","username":"sfinucan"},"change_message_id":"c353b9213e6e56bb08daf0d9a6a2d94cd2f02cfe","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":3,"id":"fb6f7b60_7778b40e","updated":"2025-09-15 16:16:01.000000000","message":"\u003e To reiterate my earlier comment, you could just install SetupTools and then the script needs no changes at all since import distutils.version will continue to work. Maybe adding Setuptools to the requirements.txt file for this package would be sufficient?\n\nYes, and I suspect that is why this job continues to pass. However, I can\u0027t imagine they plan to keep that around forever, and it seems sensible to me to use packaging instead (like we do everywhere else)?","commit_id":"10f7679f7db2ed4f75a15e94b07ca41e22f360c8"},{"author":{"_account_id":11904,"name":"Sean McGinnis","email":"sean.mcginnis@gmail.com","username":"SeanM"},"change_message_id":"4da8b998ca0bf248fe71d7b82fbe4dce19083f6d","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":3,"id":"13a796ff_f59427d4","updated":"2025-10-13 23:31:44.000000000","message":"Based on last comments, this sounds like the right way to go.","commit_id":"10f7679f7db2ed4f75a15e94b07ca41e22f360c8"},{"author":{"_account_id":5263,"name":"Jeremy Stanley","display_name":"fungi","email":"fungi@yuggoth.org","username":"fungi","status":"missing, presumed fed"},"change_message_id":"6115ae6601b4f8da7231ac60a928c9d372604d1b","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":3,"id":"b9ffaeb8_784a6826","in_reply_to":"044cb57f_0d9c7922","updated":"2025-09-19 14:14:48.000000000","message":"Wow, I suppose I shouldn\u0027t be surprised, but that doesn\u0027t bode well for all the projects and distros that added new SetupTools dependencies just to be able to keep using distutils. My recollection of when it was excised from the stdlib is that nobody created a stand-alone fork on PyPI precisely because the SetupTools maintainers had agreed to take it over and continue providing the interface. I expect a lot of other distutils users got the same impression.","commit_id":"10f7679f7db2ed4f75a15e94b07ca41e22f360c8"},{"author":{"_account_id":15334,"name":"Stephen Finucane","display_name":"stephenfin","email":"stephenfin@redhat.com","username":"sfinucan"},"change_message_id":"95bb23207827444ba6a4f11dee2f3aac1fb1e91a","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":3,"id":"044cb57f_0d9c7922","in_reply_to":"7be1e5a4_5f3ba44f","updated":"2025-09-19 09:32:24.000000000","message":"\u003e Replacing it with a different implementation in our scripts is fine for consistency\u0027s sake, but I wouldn\u0027t go ripping out distutils calls on the unfounded assumption that SetupTools might remove it at some date in the distant future.\n\nOh no, I know they will:\n\n\u003e We are working to remove a lot of distutils stuff, and trying to integrate everything so that \u0027distutils\u0027 is an internal implementation detail and not part of any supported interface, but it\u0027s taking time, as there\u0027s a lot of Hyrum\u0027s law manifesting.\n\nFrom https://bugs.launchpad.net/pbr/+bug/2111459/comments/5","commit_id":"10f7679f7db2ed4f75a15e94b07ca41e22f360c8"},{"author":{"_account_id":5263,"name":"Jeremy Stanley","display_name":"fungi","email":"fungi@yuggoth.org","username":"fungi","status":"missing, presumed fed"},"change_message_id":"3cb7627cb4145523f1ad098933cbabe46cd89377","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":3,"id":"7be1e5a4_5f3ba44f","in_reply_to":"fb6f7b60_7778b40e","updated":"2025-09-15 16:36:48.000000000","message":"The only reason distutils was in the stdlib originally was that the build process for CPython itself relied on it until recently. Since distutils is integral to SetupTools, maintainership was moved from the stdlib to SetupTools which continues to supply it now. I\u0027ve seen nobody mention dropping it from SetupTools in the future; I expect that would require some substantial work on the part of the maintainers. As an example, Debian\u0027s packages of Python libraries which require distutils were given SetupTools as either a run-time or build-time dependency, rather than carrying a separate distutils fork, under the expectation that it will remain a part of SetupTools more or less indefinitely.\n\nReplacing it with a different implementation in our scripts is fine for consistency\u0027s sake, but I wouldn\u0027t go ripping out distutils calls on the unfounded assumption that SetupTools *might* remove it at some date in the distant future.","commit_id":"10f7679f7db2ed4f75a15e94b07ca41e22f360c8"}]}