jellyMar 19
Kees Cook

RE: https://fedi.lwn.net/@lwn/116239303710146388

And when you make changes to 20,000 kmalloc calls, you start getting CCed on a LOT of patches since you touched "neighboring code" recently. 😭

Mar 19 at 9:08pmWeb
Show threadTed GouldMar 19
@kees teach you to try and make things better 😜
Show threadJonathan CorbetMar 19
@kees Gee, sounds like being the docs maintainer :)
Show threadermo | Rune MorlingMar 19
@kees Oof.
Show threadJani NikulaMar 19
@kees I'm eagerly waiting for your upcoming patches to get_maintainer.pl next. 😉
Show threadljsMar 20

@jani @kees it should default to showing maintainers and reviewers only. The git stuff is really silly and unhelpful.

Maybe if you run it on a patch with tags add those people too...

Show threadJani NikulaMar 20

@ljs @kees

Yeah, returning stuff from git history should only be a fallback for files without maintainers. I get my fair share of these Cc's and I'm always like WTF why am I included.

IMO it should also only include the most specific maintainers by default. If you have the maintainer for file foo/bar/baz, don't include the maintainer for foo/bar or foo.

Show threadJani NikulaMar 20

@ljs @kees

Oh, and the primary reason I only whine and moan here instead of sending patches is that the damn thing is written in Perl. Plus I don't want to get Cc'd for patches touching get_maintainer.pl in the future. No wait. 😉

Show threadljsMar 20
@jani @kees yeah I'm not really keen on fiddling with perl either... I normally put my money where my mouth is on whines re: kernel but in this case, no :))
Show threadljsMar 20

@jani @kees the whole inheritance thing is kinda broken actually see [0] for an example of this being broken :)

Stuff lower in hierarchy that potentially impacts stuff higher up is not cc'd to those subject to it, stuff higher up does include lower-down etc.

I guess we need to distinguish between 'anything in this glob that's not already handled elsewhere' vs. 'I want to be cc'd on everything here'.

cc: @vbabka who loves editing get_maintainer.pl :P

https://lore.kernel.org/all/2026030820[email protected]/T/#u

[PATCH] MAINTAINERS: add Josh Law as reviewer for library code

Show threadPetr TesaříkMar 20
@ljs @jani @kees @vbabka Ugh, why is this still written in Perl?
Why didn't someone rewrite it in—I don't know—Brainfuck?
Show threadljsMar 20
@ptesarik @jani @kees @vbabka cc: @lkundrak
Show threadNEPŘÁTELSKÉ EMOCE 🇺🇦🇨🇿Mar 20
@ptesarik @ljs @jani @kees @vbabka brother repent
Show threadJani NikulaMar 20

@ljs @kees @vbabka

Right, MAINTAINERS changes should also include all the other MAINTAINERS entries it impacts.

Show threadljsMar 20

@jani @kees @vbabka I realise there's some thorny issues as to how to simultaneously avoid noise but also make sure the right people are cc'd.

And I whine a lot about cc (maybe too much tbh) on list because my mail is a disaster and I stupidly somewhat rely on it for my workflow atm, but it's important to get it right.

But right now get_maintainer.pl is enough of a pain it's understandable that a lot of people ignore it + go by gut, big win is improve defaults I think

Show threadmarch38Mar 20
@kees LLVM recently switched to Discourse, away from mailing lists. Some people are not happy. Fans of this or that tool and everyone else are all missing the most important point: where is the high-level, _design document_ for "get_maintainer.pl" ? It's a really hard/impossible problem, so there's no such document and no design. The choice of Perl is only the icing on the cake.