)]}'
{"/PATCHSET_LEVEL":[{"author":{"_account_id":7118,"name":"Ian Wienand","email":"iwienand@redhat.com","username":"iwienand"},"change_message_id":"c46dc977cf28a131ff141ddb997843e19b8c1c90","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":1,"id":"1beb0062_7f699585","updated":"2022-06-09 06:50:39.000000000","message":"I feel like we\u0027re stuck in a bit of a loop here.  I was opposed to this previously because we\u0027re shipping something clearly relied upon (Python 3.6 support) but not testing it.  This doesn\u0027t seem helpful.\n\nBut, the openstack-tox python 3.6 job is failing, because the master requirements are specifying a networkx that doesn\u0027t have 3.6 support any more.\n\nI\u0027m wondering, in the history of all this, why we are using the openstack tox job templates.  If we\u0027ve discussed it, I can\u0027t see it?  I guess to be co-installable; but is that actually something we care about?  Does anyone actually install dib alongside everything else?  And it seems we can\u0027t be both using master requirements AND supporting 3.6 -- so something has to give.  We either start branching or find another way...\n\nWhy don\u0027t we just use the tox-py* jobs, that don\u0027t use requirements?  That ensures we\u0027re remaining syntax compatible with all the versions we want to support.","commit_id":"9a6da8f8fb415767ebf92fdaead834073beae8fe"}]}
