Details
- Reviewers
Mordante philnik - Group Reviewers
Restricted Project - Commits
- rG87acf6ddc73d: [libc++][docs] Add note about RFCs for significant changes
Diff Detail
- Repository
- rG LLVM Github Monorepo
Event Timeline
LGTM % nit.
libcxx/docs/Contributing.rst | ||
---|---|---|
31 | I think just saying "that might have a significant impact on the user" is enough. I don't see the point of making an RFC for changes that are purely project-internal. We should just discuss this kind of stuff on the monthly meetup once we have that set up. If there are concerns about user interactions we can still write an RFC. (e.g. I don't think there is a point in asking whether we can rename _LIBCPP_CONSTEXPR_AFTER_CXXab to _LIBCPP_CONSTEXPR_SINCE_CXXcd) |
libcxx/docs/Contributing.rst | ||
---|---|---|
31 | @Mordante Are you fine with that? I am, but IIUC one of your concerns was that we often didn't do a RFC on Discourse for project-internal discussions as well, which could sometimes lead to a consensus that isn't clear to everyone. I'd be fine with keeping the wording as-is or to changing it to only include decisions that potentially have significant impact on users. |
LGTM!
libcxx/docs/Contributing.rst | ||
---|---|---|
31 | I'm fine with it. I think it would be good to do Discourse too for large internal things. I still Discord is not great medium for these discussions. Good to start them but not good if it becomes a larger discussion. One of my main issues is that proposals often don't have a conclusion whether we reached some sort of consensus. (But that seems to happen in more RFCs of LLVM.) Here I think having meetings will help to alleviate that. However I don't feel we need to write all that down. |
I think just saying "that might have a significant impact on the user" is enough. I don't see the point of making an RFC for changes that are purely project-internal. We should just discuss this kind of stuff on the monthly meetup once we have that set up. If there are concerns about user interactions we can still write an RFC. (e.g. I don't think there is a point in asking whether we can rename _LIBCPP_CONSTEXPR_AFTER_CXXab to _LIBCPP_CONSTEXPR_SINCE_CXXcd)