-
Notifications
You must be signed in to change notification settings - Fork 12.8k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Treat projections with infer as placeholder during fast reject in new solver #108830
Treat projections with infer as placeholder during fast reject in new solver #108830
Conversation
Some changes occurred in src/tools/clippy cc @rust-lang/clippy Some changes occurred to the core trait solver cc @rust-lang/initiative-trait-system-refactor |
ab25527
to
540a9a0
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
As we share TreatParams
with the DeepRejectCtxt
, should we instead add TreatProjections
separately as another argument to simplify_type
and keep the current TreatParams
?
that would work and is probably a bit cleaner 🤔 |
☔ The latest upstream changes (presumably #109019) made this pull request unmergeable. Please resolve the merge conflicts. |
7661cf3
to
dd430d7
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
r=me after nits
compiler/rustc_trait_selection/src/traits/specialize/specialization_graph.rs
Outdated
Show resolved
Hide resolved
@@ -0,0 +1,43 @@ | |||
// check-pass | |||
// compile-flags: -Ztrait-solver=next | |||
// Issue 96230 |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
we should add a fixed-by-trait-solver-next
label for issues like this, so that we don't forget to close them when we stabilize the new solver
dd430d7
to
84d254e
Compare
@bors r=lcnr |
…ject-faster, r=lcnr Treat projections with infer as placeholder during fast reject in new solver r? `@lcnr` Kind of a shame that we need to change all of the call sites for `for_each_relevant_impl`, etc. to pass an extra parameter. I guess I could have the "default" fn which calls a configurable fn?
…iaskrgr Rollup of 9 pull requests Successful merges: - rust-lang#108419 (Stabilize `atomic_as_ptr`) - rust-lang#108507 (use `as_ptr` to determine the address of atomics) - rust-lang#108607 (Don't use fd-lock on Solaris in bootstrap) - rust-lang#108830 (Treat projections with infer as placeholder during fast reject in new solver) - rust-lang#109055 (create `config::tests::detect_src_and_out` test for bootstrap) - rust-lang#109058 (Document BinOp::is_checkable) - rust-lang#109081 (simd-wide-sum test: adapt for LLVM 17 codegen change) - rust-lang#109083 (Update books) - rust-lang#109088 (Gracefully handle `#[target_feature]` on statics) Failed merges: r? `@ghost` `@rustbot` modify labels: rollup
This caused a perf regression (see results). @lcnr / @compiler-errors do we have a sense of whether this was expected or not? Any thoughts on possible fixes? |
ahh, that code is really hot so this small change has such a big impact 🤔 we should maybe change |
I can take a look at doing that unless lcnr beats me to it |
I am surprised by this regression. I thought the fast reject code was very important for |
…ct-faster-2, r=lcnr Don't pass `TreatProjections` separately to `fast_reject` Don't pass `TreatProjections` separately to `fast_reject`, and instead use the original approach of switching on two variants of `TreatParams` (undoes this: rust-lang#108830 (review)). Fixes the regression introduced in rust-lang#108830 (comment)
…-2, r=lcnr Don't pass `TreatProjections` separately to `fast_reject` Don't pass `TreatProjections` separately to `fast_reject`, and instead use the original approach of switching on two variants of `TreatParams` (undoes this: rust-lang/rust#108830 (review)). Fixes the regression introduced in rust-lang/rust#108830 (comment)
r? @lcnr
Kind of a shame that we need to change all of the call sites for
for_each_relevant_impl
, etc. to pass an extra parameter. I guess I could have the "default" fn which calls a configurable fn?