41402 (a1ba@suya.place)'s status on Saturday, 18-May-2024 21:30:30 JST
-
@yakumo_izuru it's not inconvenient for most people. It just outright sucks.
Sure, it's OK for throwing patches at maintainer, especially if you don't want to make yet another centralized forge account. Sure, it enforces some development culture.
But it's all comes at the cost of headache of:
* Relying on mostly unreliable mail servers. I literally had my emails lost because of some insane anti-spam filters.
* Most mail clients are also fucking broken. You either have some bullshit that randomly enforces HTML and breaks patches, either some unix weenie vi-like bullshit that nobody ever used and had no real user feedback.
* Not having a real way to iterate over your changes. When you're sending a new revision of your patchset, all comments that was made for all previous revisions get lost (as they were replies to the older thread). I hate GitHub but look how it implements code review. It remembers the review comments were made even after rebasing.
* Most mailing lists are DEAD. Sometimes, the only way you can actually make somebody review your changes is to directly ping them in IRC or Matrix because nobody fucking reads ton of mails coming from centralized mailing list. As a generic contributor you can't really filter out everything you're not interested in. And, no, Cc'ing maintainers with a crappy perl script from Linux kernel doesn't always work.
* If maintainers didn't set up some kind of public archive (love mailing list archives that are walled behind cloudflare) or Patchwork you're screwed. Because the only people who have the patches you're interested in are those who were in To and Cc fields when OP sent their patchset.