Details
Diff Detail
Event Timeline
Since you are going to the effort of changing the CodeOwners file format from what is in LLVM, can you go ahead and add Discourse and Discord handles? It is often very helpful to tag someone on Discourse/Discord. I'll be proposing this change for LLVM soon.
lldb/CodeOwners.rst | ||
---|---|---|
8–9 | This could be taken to mean every review must have approval from a code owner, on top of whatever other review has been done. Is that the intent? Someone coming from a project with strong maintainer rules (e.g. GDB, so I gather) may take it that way. | |
215 | Is the extra space an RST thing, seems placed randomly. |
lldb/CodeOwners.rst | ||
---|---|---|
8–9 | I copied this from the Clang CodeOwners.rst with the aim of being consistent, but I'm happy to tweak it. We could qualify the last sentence with something like "when consensus cannot be reached" or if we think "gatekeeper" is too strong of a work maybe we can use "tie-breaker", though I like that the former implies a sense of duty. Happy to take suggestions! | |
215 | Nope that's just me: I'll clear that up 👍 |
lldb/CodeOwners.rst | ||
---|---|---|
8–9 | My understanding was that llvm in general didn't have this hard requirement for an owner to acknowledge every review. So yeah: Sounds good to me. |
- Update wording
- Remove double newline
lldb/CodeOwners.rst | ||
---|---|---|
8–9 |
Yup, that's my understanding as well! |
lldb/CodeOwners.rst | ||
---|---|---|
8–9 | I would not put policy regarding code owners in this document. The policy is already in the DeveloperPolicy for the project. You could reference back to that document if you want. |
lldb/CodeOwners.rst | ||
---|---|---|
8–9 | Good suggestion! I'm going to make similar changes to the Clang code owners file. |
lldb/CodeOwners.rst | ||
---|---|---|
133 | I accept this. |
Alright, seems like we have consensus. The only person that hasn't chimed in yet is Greg, but based on a comment in another thread he might be OOO. We can alway address concerns post-commit.
This could be taken to mean every review must have approval from a code owner, on top of whatever other review has been done. Is that the intent? Someone coming from a project with strong maintainer rules (e.g. GDB, so I gather) may take it that way.