| 6 May 2020 |
|  Ash joined the room. | 18:53:15 |
| 11 May 2020 |
 jorgepiloto | Hi all, I just saw that the Azure Pipelines badge on the main page for the repo shows a failing message | 16:10:57 |
| jorgepiloto | However, last merged changes some hours ago passel all the tests without problem | 16:12:22 |
 astrojuanlu | hi jorgepiloto ! thanks for the shout out, let me take a look | 16:18:38 |
| astrojuanlu | oh, I see. numba 0.39.1 was released | 16:19:30 |
| astrojuanlu | https://github.com/poliastro/poliastro/pull/958 | 16:21:19 |
| jorgepiloto | Cool! 🚀 | 17:34:55 |
| jorgepiloto | Hmmm... This is weird, the issue is still present. Might the badge be pointing towards the last stable version? | 17:37:02 |
| astrojuanlu | notice that I could not merge the PR | 18:02:08 |
| astrojuanlu | another error appeared | 18:02:11 |
| jorgepiloto | Is this again the same problem from some weeks ago? That one passing in Linux but failing in other OS? | 18:23:14 |
| astrojuanlu | not really:
astropy.utils.iers.iers.IERSRangeError: (some) times are outside of range covered by IERS table.
| 18:24:48 |
| astrojuanlu | it's weird because these tests were passing before. but I've heard bad things about the IERS service being down, and also there seems to be a new Astropy release... I still didn't have time to investigate the problem | 18:25:20 |
| 13 May 2020 |
| astrojuanlu | * oh, I see. numba 0.49.1 was released | 07:58:39 |
| astrojuanlu | jorgepiloto: the problem with the latest numba release was fixed, but we still get random failures on Windows. hopefully the next pipeline on master will pass | 07:59:03 |
| jorgepiloto | So for the next merged pr the problem should be fixed by itself? | 08:52:28 |
| astrojuanlu | well... https://github.com/poliastro/poliastro/issues/959 | 09:07:44 |
| 17 May 2020 |
| Ash changed their display name from ashvoid to Ash. | 05:49:10 |
| 3 Jul 2020 |
| jorgepiloto | astrojuanlu Did you find the reason behind those random failures? I just saw the pr | 14:14:09 |
| astrojuanlu | there's one test that is having floating point inaccuracies - I'm only addressing that one: https://github.com/poliastro/poliastro/pull/990 | 14:20:05 |
| astrojuanlu | the other one, the Tcl or Tk on Windows... no idea | 14:20:12 |
| 4 Jul 2020 |
| jorgepiloto | Cool! By the way, a new version of isort has been released today under the name Penny and the flag --recursive has been removed. As said in official changelog:
--recursive option has been removed. Directories passed in are now automatically sorted recursive.
https://timothycrosley.github.io/isort/CHANGELOG/
| 08:37:00 |
| jorgepiloto | I thought it would be as simple as removing the flag but noticed the following:
isort now formats all nested "as" imports in the "from" form. import x.y as a becomes from x import y as a.
| 08:44:16 |
 juanluastro | hahaha omg... | 08:44:35 |
| juanluastro | well, our policy is to obey whatever black and isort say, so... 😅 | 08:45:21 |
| jorgepiloto | Do you want me to open a quick fix on this? | 08:45:45 |
| juanluastro | is your documentation PR failing because of that? | 08:46:19 |
| juanluastro | if so, you can push to the same branch | 08:46:34 |
| jorgepiloto | Yes, it is installing by default the last isort version | 08:46:59 |
| juanluastro | I guess all the PRs will start failing then, so let's fix it ASAP | 08:53:26 |
