Re: Clarifying confusion of our variable placement rules caused by cleanup.h
Dan Carpenter <[email protected]> Sun, 18 Jan 2026 22:33:42 +0300
| Newsgroups | dev.linux.lists.ksummit,org.kernel.vger.linux-kernel |
|---|---|
| Message-ID | <[email protected]> |
On Sun, Jan 18, 2026 at 02:17:30PM -0500, James Bottomley wrote: > > > Duplicate header are trivially caught by tooling. > > > > > > But such rules aren't useful either -- I've seen that Python IDEs > > > hide import list by default (and probably manage it) because it is > > > not "real" code. > > > > > > Rules for initializers can be harmful because ordering affects code > > > generation. > > > > I agree. I still prefer the upside-down x-mas tree approach for > > declaring variables, but obviously if they also get initialized, then > > that trumps aesthetic reasoning. > > How is any of this relevant to a style document? You're quibbling over > individual maintainer foibles which, while they may be deeply held to > you (and obviously are relevant to contributors to your subsystems > because they need to know your foibles), can't be part of our universal > advice because not all maintainers agree (not even on the direction of > the Christmas Tree). > The direction of the Christmas Tree is always upside down. That's a standard in networking and a bunch of other subsystems. Otherwise people don't care. I've seen people who write code in Right Side Up Christmas Tree style but they don't reject code which is in a different order. If you're working across the entire kernel like I do then it's safest to assume Upside Down Christmas Tree is the rule. regards, dan carpenter