)]}'
{"/PATCHSET_LEVEL":[{"author":{"_account_id":35201,"name":"Michael Kelly","display_name":"Michael Kelly","email":"mkelly@arista.com","username":"michael_kelly"},"change_message_id":"2a2468c901270a585fd82e9dde226cec6437d079","unresolved":true,"context_lines":[],"source_content_type":"","patch_set":2,"id":"201a0a1b_87cf5ba4","updated":"2022-08-18 20:09:19.000000000","message":"\u003e Patch Set 2: Code-Review-1\n\u003e \n\u003e Can we just lift the version cap?  Normally we try to use uncapped versions so that we get updates automatically and then put in caps temporarily if something breaks.  (The current \u003c cap is an indication that we dropped the ball on that.)\n\nI\u0027m fine to do that if that\u0027s what\u0027s preferred from a project standpoint. \n\nFor context, typically I avoid having ambiguous version numbers for imports because it results in less predictable build artifacts even for the same commit hash.  eg: with the version locked here I know that any image built from this will always have the expected version, which makes debugging easier later.","commit_id":"5f97bbea93f612664e9885808e3fc9d1cd8f2821"},{"author":{"_account_id":1,"name":"James E. Blair","email":"jim@acmegating.com","username":"corvus"},"change_message_id":"92cd65f8786c66d1cd28b3621d5c5244a98a04f7","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":2,"id":"dcc78703_02795d5f","in_reply_to":"201a0a1b_87cf5ba4","updated":"2022-08-18 20:13:10.000000000","message":"Yeah, I think this would match most closely what we\u0027re doing elsewhere, so I\u0027d like to try that.  I totally get the locking versions argument, and I think that\u0027s a perfectly legit thing for a downstream consumer to want to do.  From an upstream project standpoint, we try to roll with the latest of everything, continuously, except when it breaks (which it does, quite often, and we take joy in how quickly we become aware of such problems).","commit_id":"5f97bbea93f612664e9885808e3fc9d1cd8f2821"},{"author":{"_account_id":35201,"name":"Michael Kelly","display_name":"Michael Kelly","email":"mkelly@arista.com","username":"michael_kelly"},"change_message_id":"dfff75d6ce0f88150fece50d8452c8242ea0f66c","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":2,"id":"371f1bf1_b9972724","in_reply_to":"dcc78703_02795d5f","updated":"2022-08-18 20:44:45.000000000","message":"sure. sounds good to me","commit_id":"5f97bbea93f612664e9885808e3fc9d1cd8f2821"},{"author":{"_account_id":9311,"name":"Tristan Cacqueray","email":"tdecacqu@redhat.com","username":"tristanC"},"change_message_id":"64bf102e8f7d5ea633aa392180f295147b6fef8b","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":4,"id":"914446f7_7a2b238d","updated":"2022-09-06 19:28:56.000000000","message":"I don\u0027t mind keeping the upper bound, but it sounded like you meant to remove it.","commit_id":"db7f5349e43470d54c21986ea56ff68e138ddaab"},{"author":{"_account_id":1,"name":"James E. Blair","email":"jim@acmegating.com","username":"corvus"},"change_message_id":"c2bbdd300c2c2159e8e813d5754efafcf6a76b2e","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":4,"id":"b20503c9_5ed79a8b","in_reply_to":"914446f7_7a2b238d","updated":"2022-10-14 13:14:25.000000000","message":"There should be no cap at all.  The only time we will add an upper version is when we know it won\u0027t work, and that is temporary.  There is no \"qualification\" process other than the test failure that will occur the next time we try to propose a change.  That is how we handle *all* python dependencies in the zuul project.","commit_id":"db7f5349e43470d54c21986ea56ff68e138ddaab"},{"author":{"_account_id":35201,"name":"Michael Kelly","display_name":"Michael Kelly","email":"mkelly@arista.com","username":"michael_kelly"},"change_message_id":"34c5ae6bf61b386c00c8588e2a85b0ffe8a9dcba","unresolved":true,"context_lines":[],"source_content_type":"","patch_set":4,"id":"f1e3f28f_8afb2925","in_reply_to":"914446f7_7a2b238d","updated":"2022-09-07 21:30:45.000000000","message":"To clarify in plain language, are we wanting to have the upper bound mean \u0027we support any 1.35 or less version\u0027 or is it ok to say \u0027we support 1.35.x up to 1.35.6, but have not qualified \u003e.7\u0027? I\u0027m ok with either option.","commit_id":"db7f5349e43470d54c21986ea56ff68e138ddaab"},{"author":{"_account_id":35201,"name":"Michael Kelly","display_name":"Michael Kelly","email":"mkelly@arista.com","username":"michael_kelly"},"change_message_id":"59946a0dd413a720082fccc71f1cd07748a016be","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":4,"id":"abc1a6af_6a7f3a82","in_reply_to":"b20503c9_5ed79a8b","updated":"2022-10-14 15:40:25.000000000","message":"Sure.  As I think about it, this probably does need a minimum bound though as it\u0027ll be broken with earlier version","commit_id":"db7f5349e43470d54c21986ea56ff68e138ddaab"}]}