| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2023-09-22 | |||
| 12:17:30 | bauzas | those need to be reasonabily needing a reason, because it makes any other backports a pain | |
| 12:17:54 | bauzas | we suffered a lot with functional tests refactoring, but this was kinda understansable | |
| 12:18:11 | bauzas | since it was improving tests reusability | |
| 12:18:16 | stephenfin | bauzas: out of curiosity, could you point to a recent case where a backport was impeded because of one of these linting change? | |
| 12:18:20 | sean-k-mooney | honetsly not automating this makes it harder for use to maintain nova | |
| 12:19:02 | sean-k-mooney | and it makes it harder for new peopel to continbue. with the reduced upstream comunity i think it would be in our intereste to embrance this automation where we can | |
| 12:19:08 | bauzas | stephenfin: I don't have any examples of the linting changes in mind, but be sure we faced some problems for other bits | |
| 12:19:25 | bauzas | sean-k-mooney: I understand that point for refactoring things | |
| 12:19:37 | bauzas | sean-k-mooney: but for linting, I seriously argue | |
| 12:20:06 | sean-k-mooney | yep and ill happy argue back | |
| 12:20:17 | bauzas | code style modifications being even less convincing to me | |
| 12:20:59 | sean-k-mooney | honestly i have spend enough time this cycle on sytle issues in code review | |
| 12:21:17 | sean-k-mooney | that if this is the only thing we got done in caracal i woudl consider it a success | |
| 12:21:45 | bauzas | I know it's Friday, but is this a reason to argue now ? | |
| 12:22:38 | sean-k-mooney | i mean its friday what else would we do | |
| 12:22:45 | stephenfin | :) | |
| 12:22:48 | sean-k-mooney | go home early and realx | |
| 12:24:07 | bauzas | I already sweated and now I have endorphins | |
| 12:24:24 | bauzas | not sure my stress level should jump, not good for my heart | |
| 12:24:27 | sean-k-mooney | if the wether stays ok i might go for a bike ride later | |
| 12:24:49 | sean-k-mooney | althogh its a bit humid out so we will see | |
| 12:24:53 | bauzas | seriously, I'm half-kidding, but feel free to put code styling and autoformatters on the PTG agenda | |
| 12:25:05 | opendevreview | Merged openstack/nova master: db: Replace use of backref https://review.opendev.org/c/openstack/nova/+/860829 | |
| 12:25:12 | bauzas | not sure my opinion will evolve but I'm still open to discuss about it | |
| 12:25:36 | bauzas | I just don't think the benefits overcome the pain | |
| 12:26:01 | sean-k-mooney | honestly the pain of not having them makes me somethimes condier if i wnat to keep manintianing nova | |
| 12:26:11 | bauzas | and about exceprts of code changes that pulled extra complexity into backports, I can do some homework and find some series | |
| 12:26:32 | sean-k-mooney | bauzas: i would happly backport the autofromating if that woudl make you feel beeter | |
| 12:26:36 | bauzas | not out of my head at first sight unfortunately, but there are experiences for sure | |
| 12:26:49 | bauzas | sean-k-mooney: down to Train ? :) | |
| 12:26:53 | sean-k-mooney | sure | |
| 12:27:03 | sean-k-mooney | the patches would be generated by the tool | |
| 12:27:15 | sean-k-mooney | we woudl jsut have to review it once and its done for ever | |
| 12:27:40 | sean-k-mooney | if that is your main objection its sovleabel. but im also oke with fixing this when backporting | |
| 12:28:27 | bauzas | that's indeed my main objection, beyond just a feeling | |
| 12:28:52 | sean-k-mooney | as you know i have been working in both go and python lately | |
| 12:29:07 | bauzas | I'm personnally against such things because I think it negates the stance of writing, but this is a personal opinion and absolutely not arguable | |
| 12:29:22 | sean-k-mooney | and ansible... but it has renifoced that feelting that auto liniting is good | |
| 12:29:58 | sean-k-mooney | and auto formating is both good for the reviewer and author | |
| 12:30:31 | bauzas | I can't wait for a ChatGPT bot on Visual Code | |
| 12:30:43 | bauzas | VS Code* | |
| 12:30:52 | stephenfin | https://marketplace.visualstudio.com/items?itemName=GitHub.copilotvs | |
| 12:30:55 | sean-k-mooney | hehe i already use code poilte ot help writhe my comments | |
| 12:31:04 | sean-k-mooney | its why they have had less typos lately | |
| 12:31:27 | bauzas | and yeah, I heard about GH CoPilot | |
| 12:32:00 | sean-k-mooney | it daydreams a bit so unless you know the code base i would not recomend it to new commers | |
| 12:33:24 | sean-k-mooney | i have mostly been using it of go/ansible but i find i have to check its work and rewite it but its often close to what i would write | |
| 12:33:26 | bauzas | that's the problem with those AI tools | |
| 12:33:36 | bauzas | 95% of the job is perfectly done | |
| 12:33:48 | bauzas | and then the rest is errorprone | |
| 12:33:59 | sean-k-mooney | yep and that last 5% is where being a software engineer comes in | |
| 12:34:25 | bauzas | I'm not sure my productivity will raise because of a bot writing my code | |
| 12:34:26 | sean-k-mooney | using your expirnce ot understand its probmpts and correct them | |
| 12:35:21 | bauzas | most of the time it takes for me to write a feature is : 1/ identify the gaps, 2/ come with a proposal, 3/ let the community agree on the proposal, 4/ write that | |
| 12:35:33 | bauzas | the 4/ isn't exactly hard to do | |
| 12:36:41 | sean-k-mooney | sure and most of the process of adding a feature into nova is code review | |
| 12:36:53 | sean-k-mooney | not wirting the code in the first place | |
| 12:37:41 | bauzas | honestly, I feel the entry barrier on Nova isn't because we're not having black style and autoformatters | |
| 12:38:19 | bauzas | but rather because it's a very large distributed software contributed by many in different times | |
| 12:38:34 | bauzas | and I don't think autoformatters will lower this barrier | |
| 12:38:51 | sean-k-mooney | sure but that misses the point | |
| 12:39:45 | bauzas | which is that people can autotab their brackets ? | |
| 12:40:04 | sean-k-mooney | no | |
| 12:40:05 | stephenfin | it's a quality of life improvement | |
| 12:40:09 | sean-k-mooney | yep | |
| 12:40:42 | stephenfin | being a surgeon is hard: working in a clean operating theatre makes it slightly easier | |
| 12:41:01 | bauzas | I see myself more like a butcher :D | |
| 12:41:15 | bauzas | anyway, I stink and I need to get a shower | |
| 12:41:17 | stephenfin | then black is a sharp knife | |
| 12:41:20 | stephenfin | ;) | |
| 12:41:38 | stephenfin | or rather an automatic knife sharpener | |
| 12:41:55 | sean-k-mooney | i was going to play the frech card and say its a Charcuterie | |
| 12:42:01 | bauzas | stephenfin: if we continue the analogy, people can get hurt by a sharp knife if they don't know how to use it :) | |
| 12:42:29 | sean-k-mooney | which is another reason my analogy is better :) | |
| 12:43:08 | sean-k-mooney | be a charcutier not a butcher of our lovely codebase :) | |
| 12:55:40 | sean-k-mooney | stephenfin: since your about can you look at https://review.opendev.org/c/openstack/nova/+/895655 | |
| 12:56:17 | sean-k-mooney | once thats merged i can propsoe the revert adn ping it to rodolfo and co so we can renable it once neutron says the test is fixed for ovn | |
| 12:56:32 | stephenfin | ack, done | |
| 12:56:56 | sean-k-mooney | that test is one of the thing that cased the hackign patch to fail | |
| 12:57:12 | sean-k-mooney | the other failure was volume detach related | |
| 13:05:01 | bauzas | fwiw, starting to look at the rest of sql14 series | |
| 13:09:04 | opendevreview | Felix Huettner proposed openstack/nova master: ignore errors when deleting volume attachments https://review.opendev.org/c/openstack/nova/+/896235 | |
| 13:18:05 | bauzas | okay, this is Friday, I'm not a better and decent position to start a debate https://luminousmen.com/post/my-unpopular-opinion-about-black-code-formatter | |
| 13:18:30 | bauzas | now* | |
| 13:18:57 | sean-k-mooney | for what its worht a number of opentack proejct have arelady adopted black | |
| 13:19:09 | bauzas | that doesn't make the point valid :) | |
| 13:19:12 | sean-k-mooney | we already have autopep8 which was a compomise | |
| 13:19:58 | bauzas | yup, and which doesn't force your code to change, right? | |
| 13:20:14 | bauzas | just because a ending bracket has to be on a separate line | |
| 13:20:17 | sean-k-mooney | it only fixes code if it not pep8 complient | |
| 13:20:33 | sean-k-mooney | and it undersands the things we disable in tox.ini | |
| 13:20:57 | bauzas | black is *very* noisy in terms of code change since it can't adapt to the existing code style | |
| 13:21:11 | sean-k-mooney | i.e. it reads https://github.com/openstack/nova/blob/master/tox.ini#L307 | |
| 13:21:21 | bauzas | and it forces us to do the way blacks thinks being the best | |
| 13:22:16 | sean-k-mooney | i think that is better then use disagreeing on what is the best style | |
| 13:22:30 | bauzas | btw. this twitter post which is replied by Guido is absolutly gold | |
| 13:23:16 | sean-k-mooney | bauzas: by the way i actully dont really care what the tool is provide we have one that does it | |
| 13:23:23 | bauzas | it shows the mindset of many developers : in order to follow *best* practices of coding, it says "install X and Y" | |
| 13:23:37 | bauzas | my head explodes when I read this | |
| 13:23:54 | bauzas | I thought for 2 decades that coding best practices were about code patterns and reusability | |