| opendevreview | Akihiro Motoki proposed openstack/horizon master: Allow both Django 2.2 and 3.2 for smooth transition https://review.opendev.org/c/openstack/horizon/+/815393 | 00:27 |
|---|---|---|
| opendevreview | Akihiro Motoki proposed openstack/horizon master: Allow both Django 2.2 and 3.2 for smooth transition https://review.opendev.org/c/openstack/horizon/+/815393 | 00:46 |
| amotoki | vishalmanchanda: hi | 05:07 |
| vishalmanchanda | amotoki: hello. | 05:08 |
| amotoki | vishalmanchanda: is the failure in the integration tests really a random failure? | 05:08 |
| amotoki | vishalmanchanda: I am not sure as "no such method error" does not sound like a random failure.... | 05:09 |
| vishalmanchanda | amotoki: maybe because I have seen the same tests failing in past and then passed after recheck. | 05:10 |
| vishalmanchanda | amotoki: Let me try these in my local env. but it will going to take time. | 05:11 |
| vishalmanchanda | amotoki: just checked the status for that test https://zuul.openstack.org/stream/5d273832842f45b6af440b0532c7c86f?logfile=console.log | 05:13 |
| vishalmanchanda | amotoki: it is passing right now. | 05:14 |
| vishalmanchanda | amotoki: both are passing now | 05:14 |
| amotoki | vishalmanchanda: looking at https://opendev.org/openstack/horizon/src/commit/ab2021eb1da17a9350f0bcb9784f088a83859ce0/horizon/test/webdriver.py, | 05:17 |
| amotoki | vishalmanchanda: when an error at https://zuul.opendev.org/t/openstack/build/ddf2805c71ce4498b1996465fa1407ee/log/job-output.txt#26002 happens, reload_request does not work. | 05:17 |
| amotoki | vishalmanchanda: it looks like a horizon code error in horizon/horizon/test/webdriver.py | 05:18 |
| amotoki | vishalmanchanda: the reason 'recheck' works seems the original exception did not happen. | 05:18 |
| opendevreview | Akihiro Motoki proposed openstack/horizon master: Allow both Django 2.2 and 3.2 for smooth transition https://review.opendev.org/c/openstack/horizon/+/815393 | 05:24 |
| opendevreview | Merged openstack/horizon master: Allow both Django 2.2 and 3.2 for smooth transition https://review.opendev.org/c/openstack/horizon/+/815393 | 11:07 |
| *** tosky_ is now known as tosky | 11:25 | |
| vishalmanchanda | amotoki: hi, If I am not wrong, now I have to cut a new release for horizon with latest patch merged? | 11:36 |
| vishalmanchanda | amotoki: So that openstack/requirements or plugins can use horizon with these changes? | 11:38 |
| amotoki | vishalmanchanda: yes, that's my understanding. Otherwise I see no way to allow horizon plugins to use Django 3.2 with a released version of horizon. | 13:11 |
| amotoki | vishalmanchanda: yes, that's my understanding. Otherwise I see no way to allow horizon plugins to use Django 3.2 with a released version of horizon. | 13:14 |
| amotoki | vishalmanchanda: I am not a fan to force to land something by skipping tests. In most cases it is a sign that we miss something. | 13:16 |
| amotoki | vishalmanchanda: out side of horizon plugins, this kind of confusion happened around Django req change in horizon. https://review.opendev.org/c/openstack/requirements/+/815464 | 13:34 |
| vishalmanchanda | amotoki: hmm, I was thinking if we should add document about such migration to avoid any confusion. | 13:40 |
| amotoki | vishalmanchanda: yes. it is in my plan. when I switched the horizon release model to cycle-with-intermediary to allow plugins to consume horizon as a lib, perhaps I was too familiar with the requirements process, so I know what we should do, | 13:42 |
| amotoki | vishalmanchanda: but I was lazy enough and did not document it so far :p | 13:42 |
| vishalmanchanda | amotoki: I took a brutal way to get them merged skipping CI job:(. | 13:42 |
| amotoki | vishalmanchanda: when switching the release model, I discussed the detail with the requirements team and what was the problem. | 13:43 |
| vishalmanchanda | amotoki: I was also lazy to send a mail about it on open-discuss because I thought I will take care of it asap. | 13:44 |
| amotoki | vishalmanchanda: np | 13:44 |
| amotoki | vishalmanchanda: another solution is that horizon catches up the latest Django release promptly | 13:44 |
| vishalmanchanda | amotoki: anyways thanks for the help:) | 13:44 |
| amotoki | it allows us to drop the cap in global-requirements.txt | 13:44 |
| vishalmanchanda | amotoki: But I can see you are the one who push a patch to raise cap for horizon at first place https://review.opendev.org/c/openstack/requirements/+/701213 | 13:50 |
| vishalmanchanda | amotoki: Sorry I didn't understand the context completely . | 13:52 |
| amotoki | vishalmanchanda: the main reason is that the lack of dev resources in horizon and it is not easy to catch up with the latest Django releases. I discussed it with the release team and the conclusion at that time resulted in my patch you mentioned. | 13:58 |
| vishalmanchanda | amotoki: ok gotcha! | 13:59 |
| amotoki | vishalmanchanda: on the other hand, ideally horizon tracks Django changes (possibly testing beta releases allows horizon to work when a new Django is released) | 13:59 |
| amotoki | vishalmanchanda: The general consensus in the requirements management is to avoid capping a upper-bound version as much as possible. Capping the upper-bound needs to be well-considered. | 14:01 |
| opendevreview | Tatiana Ovchinnikova proposed openstack/horizon master: Make existing empty metadata properties optional https://review.opendev.org/c/openstack/horizon/+/812009 | 19:44 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!