Hi Robin, On 2026-03-18T17:47:56+0100, Robin Sommer wrote: > > > > On Wed, Mar 18, 2026 at 11:21 +0800, you wrote: > > > Before I drop the proposal, though, are there any Mutt devs/contributors who > > would like to play devil's advocate? > > It's been a long time since I have contributed to mutt, but seeing the > recent activity I do actually have some patches pending still that I'm > planning to (re-)submit. So I'll bite: I believe that it would be > super helpful to have a standard clang-format configuration for mutt > that's consistently applied across the code base. > > The key to formatters like clang-format is *not* expecting them to get > every little piece of the layout precisely like you want it. They > won't, and that's ok. Humans also differ in where they break their > lines.
But the whole point of formatting is to make the code more readable to humans, which can help work more efficiently with code. Using clang-format(1) means accepting worse formatting, just because it's less direct work. But indirectly, it would result in less efficient reading of code, because its formatting has been degraded. I'd love to be able to automate some formatting rules, but the design of clang-format(1), which is take it all or leave it, is not usable for this. I hope someone eventually develops a similar tool that allows partial enforcement. Have a lovely day! Alex > > The value of formatters is somewhere else: once a project starts > consistently using them, suddenly nobody needs to spend any brain > power anymore on getting all those indentations and white space and > line break positions right. You can just ignore that whole minefield > because you know that clang-format will fix it before anything goes > upstream. Especially for external contributors, who don't have the > project's particular style down in their muscle memory, this > substantially lowers the barrier for getting a patch in shape. And it > helps the reviewers, too: no more "please move this brace over there". > > The projects I'm primarily involved with have all switched to > automatic formatting at this point, and never looked back. > > There are different ways to implement this of course. I'm also not a > fan of CI automatically reformatting code. But CI can reject code > that's not passing clang-format, with the developer then re-running it > locally, which is usually quick and easy. I know that there can be > some technical pitfalls here (like different clang-format versions > etc.), but honestely, the cases where I've run into this over many > years and compiler versions and groups of developers, remain very rare > and certainly don't outweigh the value of such tools. > > So, bottom line: I realize that my word won't count much here, but > nevertheless I would suggest you all re-consider dropping this so > easily. > > Best, > > Robin > > > -- > Robin Sommer * ICSI * [email protected] * www.icir.org/robin -- <https://www.alejandro-colomar.es>
signature.asc
Description: PGP signature
