)]}'
{"/COMMIT_MSG":[{"author":{"_account_id":13252,"name":"Dr. Jens Harbott","display_name":"Jens Harbott (frickler)","email":"frickler@offenerstapel.de","username":"jrosenboom"},"change_message_id":"ffee6e66fa4be026068300952a670f96501c9664","unresolved":true,"context_lines":[{"line_number":8,"context_line":""},{"line_number":9,"context_line":"Cinder stable/2026.1 raised its py-pure-client minimum to 1.89.0 when it"},{"line_number":10,"context_line":"fixed QoS handling for the SDK\u0027s new VolumePatch/VolumeGroupPatch models"},{"line_number":11,"context_line":"(cinder commit c3d5787c2c621b20d6ffad7b205b66f041cd8664, bug 2158313)."},{"line_number":12,"context_line":"The branch\u0027s upper constraint was never moved and still sits at 1.83.1,"},{"line_number":13,"context_line":"below Cinder\u0027s declared floor."},{"line_number":14,"context_line":""}],"source_content_type":"text/x-gerrit-commit-message","patch_set":1,"id":"f7ae5dfb_ec8449c2","line":11,"updated":"2026-11-15 09:51:26.000000000","message":"well this change was reverted, since it doesn\u0027t match stable policy","commit_id":"c40f05762483ae946d7100e9a23dc2bba1a54ae0"}],"/PATCHSET_LEVEL":[{"author":{"_account_id":5314,"name":"Brian Rosmaita","email":"rosmaita.fossdev@gmail.com","username":"brian-rosmaita"},"change_message_id":"933f95ac930b200baffd1a63243696831d9e3c73","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":1,"id":"cea3897c_e8de2e0e","updated":"2026-11-12 21:24:32.000000000","message":"I think it makes sense to allow this change.\n\n1. As Simon points out, the py-pure-client dependencies themselves respect the Gazpacho upper-constraints, so there won\u0027t be co-installability problems.\n\n2. The py-pure-client is an optional cinder requirement listed in cinder\u0027s pyproject.toml, so a \"normal\" install doesn\u0027t use it anyway.\n\n3. We had a discussion two years ago on a requirements patch [0] about removing the cinder driver requirements from global-requirements.txt and upper-constraints.txt; we wound up not removing them because (a) the global requirements review is worthwhile, and (b) since some of these drivers are used by cinder, glance, and manila, so it\u0027s worth having the upper-constraints co-installability guarantee.  I still think this is the case.  But given the optional nature of these driver requirements, I think it would be OK to allow more leeway than with a \"normal\" requirement in global-requirements.txt.  The leeway shouldn\u0027t allow us to make version changes whenever we want to with these driver requirements, but I think this case could be a legitimate exception, given that version 1.89.0 of the py-pure-client respects the other upper-constraints.\n\n[0] https://review.opendev.org/c/openstack/requirements/+/915165","commit_id":"c40f05762483ae946d7100e9a23dc2bba1a54ae0"},{"author":{"_account_id":13252,"name":"Dr. Jens Harbott","display_name":"Jens Harbott (frickler)","email":"frickler@offenerstapel.de","username":"jrosenboom"},"change_message_id":"ffee6e66fa4be026068300952a670f96501c9664","unresolved":true,"context_lines":[],"source_content_type":"","patch_set":1,"id":"c75d903a_6b8b638d","in_reply_to":"cea3897c_e8de2e0e","updated":"2026-11-15 09:51:26.000000000","message":"we could bump this as an exception if CI testing requires it, but I don\u0027t see where that is the case. the change cited in the commit message was reverted for good reasons, it isn\u0027t a good idea to bump minimum reqs on a stable branch. so what I think you could do, is rework that bug fix to work with both older and newer py-pure-client, then we could do the bump here to have the CI test what I assume would be the \"better\" code path","commit_id":"c40f05762483ae946d7100e9a23dc2bba1a54ae0"}]}