Uh oh!
There was an error while loading. Please reload this page.
Fix two const_trait_impl issues - #100336
Conversation
This comment has been minimized.
This comment has been minimized.
fb617e9 to
7de7c9cCompareThere was a problem hiding this comment.
Shouldn't you also be applying this to arg types too?
There was a problem hiding this comment.
Also, I guess a deeper question is whether this should be fixed in rustc_trait_selection::traits::wf instead..?
Or I guess whether we should be doing wfcheck with a Constness::Const param-env at all? e.g. should we just be doing param_env.without_const when constructing the wfcx
Like, is it ever required for correctness to register a wf obligation with Constness::Const? What would break/be unsound/be incorrect if we just made the param-env not const during wfcheck?
There was a problem hiding this comment.
whether we should be doing wfcheck with a
Constness::Constparam-env at all?
We should. For an impl to be well-formed, all super traits must be fulfilled. In the case of const super traits, we need to ensure that the parent trait has a constimpl.
There was a problem hiding this comment.
I guess I mean for the nominal obligations of the projection types in the signature, for example. Not necessarily when we verify that a trait's supertraits are satisfied.
Put in a different way, any time we instantiate a WellFormed predicate, is it ever wrong to pass a param env with no constness there instead?
There was a problem hiding this comment.
That seems desirable. This is related to #100543. I'm going to look into this further.
There was a problem hiding this comment.
I tend to agree with @compiler-errors' reasoning here (it was my first thought when looking at the code).
There was a problem hiding this comment.
This has been done, but only from one producer of wf obligations, to fix this issue. The second fix is unfortunate but is the only solution I can find without breaking the tests.
estebank
commented
Aug 15, 2022
@oli-obk is currently out. CC @rust-lang/types r? rust-lang/types |
7de7c9c to
8429cd0Compare
This comment has been minimized.
This comment has been minimized.
85c785a to
758b550Compare#[const_trait] return types758b550 to
7f02890Comparebors
commented
Aug 22, 2022
☔ The latest upstream changes (presumably #100654) made this pull request unmergeable. Please resolve the merge conflicts. |
oli-obk
commented
Aug 22, 2022
r? @oli-obk |
There was a problem hiding this comment.
So... the problem was that this is not Self: ~const Index, but it was tested as such? Do we have tests for various combinations of #[const_trait] being on the super trait or not, and being on the trait itself or not?
7f02890 to
d744f36Compareoli-obk
commented
Aug 22, 2022
@bors r+ |
bors
commented
Aug 22, 2022
oli-obk
commented
Aug 22, 2022
r=me with CI passing ^^ |
fee1-dead
commented
Aug 22, 2022
@bors r=oli-obk |
bors
commented
Aug 22, 2022
…it, r=oli-obk Fix two const_trait_impl issues r? `@oli-obk` Fixesrust-lang#100222. Fixesrust-lang#100543.
Rollup of 8 pull requests Successful merges: - rust-lang#98200 (Expand potential inner `Or` pattern for THIR) - rust-lang#99770 (Make some const prop mir-opt tests `unit-test`s) - rust-lang#99957 (Rework Ipv6Addr::is_global to check for global reachability rather than global scope - rebase) - rust-lang#100331 (Guarantee `try_reserve` preserves the contents on error) - rust-lang#100336 (Fix two const_trait_impl issues) - rust-lang#100713 (Convert diagnostics in parser/expr to SessionDiagnostic) - rust-lang#100820 (Use pointer `is_aligned*` methods) - rust-lang#100872 (Add guarantee that Vec::default() does not alloc) Failed merges: r? `@ghost` `@rustbot` modify labels: rollup
r? @oli-obk
Fixes#100222.
Fixes#100543.