[libcxx-commits] [PATCH] D148432: [libc++] Simplify the tuple constructor overload set

[Louis Dionne via Phabricator via libcxx-commits]

ldionne added a comment.

Summarizing the discussion I had yesterday with @Mordante:

I think some of the problems here come from the fact that communication in libc++ is pretty siloed. I meet with most of the core contributors on a weekly basis (and sometimes more), but those are one-on-one meetings. As a result, I know what's going on and what discussion happened with who, but not everyone has the same context. So for example, Nikolas and I actually discussed this change in depth and we considered potential user implications carefully, but that isn't necessarily clear from this review alone.

To try to alleviate this, we agreed to:

1. Create a recurring monthly meeting for libc++ developers, similar to the regular meetings that happen for Clang. We can discuss general libc++ topics "in person" during those meetings and let everyone sync up with what's going on in the project.
2. We will document on our Contributing page that significant changes or changes that could meaningfully affect users should usually go through RFCs. I want to be careful about not mandating a RFC for changes too systematically, since almost all libc++ changes can affect users in one way or another (I added a `const` somewhere and broke some code inside Google last week!). That would just make the process heavier than it needs to be. However, going through a RFC on Discourse for significant changes (and this is a significant change in the sense that it sets the precedent for using extensions in the codebase) is reasonable, so we'll document that.

In terms of this change specifically: Let's wait for the above things to be implemented, then take this proposal through a RFC on Discourse and come back to it with the information we'll have gained.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D148432/new/

https://reviews.llvm.org/D148432