I'm a bit confused about certain behavior, and not sure if I have a mistake or I found a bug. The following all refers to this branch: https://github.com/Nashenas88/coi/tree/features . I have a workspace, and the main crate, coi, has multiple features, one of which cause a feature of a dependency to turn on and off (async feature of coi-derive). I have several crates in my workspace that are used solely to test that my crate's functionality works with all combination of features. Going into each directory and running
cargo test works as expected. https://gist.github.com/Nashenas88/50d7e99e78805ad1a50c239cf6fb2930 Running
cargo test at the root of the workspace does not: https://travis-ci.org/Nashenas88/coi/jobs/632791155 . Does running
cargo test at the workspace root only compile each crate once and then reuse them, ignore the feature requirements of the other crates? Or is there some detail I'm missing?
Maybe this is still an unsolved issue? This comment from a while back seems to imply so: https://github.com/rust-lang/cargo/issues/6195#issuecomment-431613030
does anyone know where the libs team has their meetings?
@bstrie You can read this topic, in short they don't have meetings or place for discussion. https://rust-lang.zulipchat.com/#narrow/stream/122651-general/topic/T-libs.20discussion.20forum
@bstrie if there is an issue that you need to bring to their attention, open a thread on internals or in rust-lang/rust
if you are in a rust team, you can ping them in rust-lang/rust, but otherwise someone else can ping them for you
I pinged them in https://github.com/rust-lang/rust/pull/67686 and https://github.com/rust-lang/rust/pull/67725 without getting much of a reaction, unfortunately (the holidays would explain this, though this is consistent with other pings over the last few months). I feel I have to pick someone from the libs team and Cc or r? them if I want to get a reaction.
@RalfJ I'd ping @Gankra for those, but I don't know if they are available for reviewing those PRs, otherwise @Amanieu or @SimonSapin are probably good fits as well
yeah I already explicitly pinged Gankra, I didnt know if they are in rust-lang/libs
@Lukas Kalbertodt maybe ?
Amanieu and Simon sound like good options (though last time I pinged them SImon indicated they are also pretty busy). not sure how long I should wait before trying to summon them. for now, the PRs are assigned to dtolnay
Amanieu is familiar with the hashmap implementation, but BTreeSet is quite different. The only person I recall that was deeply familiar with it is Gankra :/
Yea I didn't respond because I don't know anything about the btree internals.
for the 2nd PR I almost convinced myself I could r+ it. But I dont feel quite confident enough...
these are pretty high-risk changes -- if something goes wrong, we have some very hard to detect UB at our hands