| 17 May 2024 |
 jade_ | it also /forces/ good commit hygiene | 00:51:27 |
| jade_ | idk if you have noticed the commit history of lix but it is probably one of the cleanest and clearest commit histories i have ever worked with in any project. | 00:51:48 |
 strawberry🍓 (it/pup/she/they) 🏳️⚧️ 💜 🦴 (puppygock.gay) | In reply to @jade_:matrix.org
it also /forces/ good commit hygiene most underrated thing about projects ever | 00:52:05 |
| strawberry🍓 (it/pup/she/they) 🏳️⚧️ 💜 🦴 (puppygock.gay) | idk if it's just the places i hang out in but apparently it's such an unpopular opinion to want clean commit history | 00:52:30 |
 Qyriad | see, the thing about clean commit history | 00:53:13 |
| jade_ | the thing about gerrit is that it /forces/ it because your commit message is your change description. you have to write something in there for someone to be able to review it and understand where you're coming from | 00:53:35 |
| Qyriad | is it sounds like potential dogma right? it sounds like "yeah this is clearly best practices but shrugs" for a lot of people. but clean commit history doesn't just look nice, and it's not even just nice for rebases, it is necessary for code review | 00:54:46 |
| jade_ | yeah. github having the change description in two places is wild though | 00:55:23 |
| Qyriad | clean commit history is nice for looking back on. but it is necessary to be able to review changes | 00:55:25 |
| jade_ | oh right i think i remember that the forgejo AGit flow doesn't update PR descriptions into line with commit messages | 00:56:00 |
| strawberry🍓 (it/pup/she/they) 🏳️⚧️ 💜 🦴 (puppygock.gay) | i dont even have high standards, i just dont want PRs to be full of "Update <filename>", "address reviews", "Update <filename>, "address more reviews", etc | 00:56:07 |
| strawberry🍓 (it/pup/she/they) 🏳️⚧️ 💜 🦴 (puppygock.gay) | and merged like that | 00:56:16 |
| Qyriad | gods ikr | 00:56:20 |
| strawberry🍓 (it/pup/she/they) 🏳️⚧️ 💜 🦴 (puppygock.gay) | does no one amend or.. | 00:56:24 |
| jade_ | oh, yeah, gerrit actively prevents "address reviews" ever becoming a commit message | 00:56:29 |
| jade_ | that's pretty cool that it does that | 00:56:35 |
| Qyriad | In reply to@strawberry:puppygock.gay
does no one amend or.. well no because amending makes understanding the history of a PR harder! | 00:56:46 |
| Qyriad | and this is why people these days just fucking squash everything | 00:56:57 |
| jade_ | yeah like, the fundamental way that gh pr reviews work is broken | 00:57:09 |
| strawberry🍓 (it/pup/she/they) 🏳️⚧️ 💜 🦴 (puppygock.gay) | yeah but like, i dont get why amending a 2 line diff from a review is bad | 00:57:11 |
| strawberry🍓 (it/pup/she/they) 🏳️⚧️ 💜 🦴 (puppygock.gay) | if it's big yes make it a separate commit | 00:57:21 |
| Qyriad | it's not, really, but no one cares about commit history because of this whole cycle | 00:57:30 |
| jade_ | In reply to @strawberry:puppygock.gay
yeah but like, i dont get why amending a 2 line diff from a review is bad oh, another thing that gerrit Fixes is that it has a way better online editor to Just Fix someone's change as a reviewer if you want to submit it with changes. github doesn't let you stage multiple things together into one commit amend, or even let you make a commit amend from web. | 00:58:20 |
| strawberry🍓 (it/pup/she/they) 🏳️⚧️ 💜 🦴 (puppygock.gay) | i tried once fixing someone's PR from github UI, and i never ever want to do it again | 00:58:44 |
| strawberry🍓 (it/pup/she/they) 🏳️⚧️ 💜 🦴 (puppygock.gay) | * i tried once fixing someone's PR from github UI (literally something so small), and i never ever want to do it again | 00:58:57 |
| jade_ | overall this reduces review latency massively because reviewers can just fix their minor complaints like typos or whatever without taking another cycle | 00:58:59 |
| strawberry🍓 (it/pup/she/they) 🏳️⚧️ 💜 🦴 (puppygock.gay) | sometimes i'd rather just merge it and do the 2 line fix myself | 00:59:51 |
| strawberry🍓 (it/pup/she/they) 🏳️⚧️ 💜 🦴 (puppygock.gay) | * sometimes i'd rather just merge it and do the 2 line fix myself (on github) | 01:00:00 |
| Qyriad | yeah people don't do that nearly enough | 01:01:24 |
 julia | In reply to @jade_:matrix.org
overall this reduces review latency massively because reviewers can just fix their minor complaints like typos or whatever without taking another cycle meanwhile rust ppl, complaining about how github's pr review workflow thingy sucks lol | 01:08:42 |
