Shrink `Nonterminal`
These commits shrink `Nonterminal` from 240 bytes to 40 bytes. When building `serde_derive` they reduce the number of `memcpy` calls from 9.6M to 7.4M, and it's a tiny win on a few other benchmarks.
r? @petrochenkov
This commit reduces the size of `Nonterminal` from a whopping 240 bytes
to 72 bytes (on x86-64), which gets it below the `memcpy` threshold.
It also removes some impedance mismatches with `Annotatable`, which
already uses `P` for these variants.
```
error[E0412]: cannot find type `T` in this scope
--> file.rs:3:12
|
3 | impl Trait<T> for Struct {}
| - ^ not found in this scope
| |
| help: you might be missing a type parameter: `<T>`
```
Fix#64298.
Implement `?const` opt-out for trait bounds
For now, such bounds are treated exactly the same as unprefixed ones in all contexts. [RFC 2632](https://github.com/rust-lang/rfcs/pull/2632) does not specify whether such bounds are forbidden outside of `const` contexts, so they are allowed at the moment.
Prior to this PR, the constness of a trait bound/impl was stored in `TraitRef`. Now, the constness of an `impl` is stored in `ast::ItemKind::Impl` and the constness of a bound in `ast::TraitBoundModifer`. Additionally, constness of trait bounds is now stored in an additional field of `ty::Predicate::Trait`, and the combination of the constness of the item along with any `TraitBoundModifier` determines the constness of the bound in accordance with the RFC. Encoding the constness of impls at the `ty` level is left for a later PR.
After a discussion in \#wg-grammar on Discord, it was decided that the grammar should not encode the mutual exclusivity of trait bound modifiers. The grammar for trait bound modifiers remains `[?const] [?]`. To encode this, I add a dummy variant to `ast::TraitBoundModifier` that is used when the syntax `?const ?` appears. This variant causes an error in AST validation and disappears during HIR lowering.
cc #67794
r? @oli-obk
Remove `rustc_error_codes` deps except in `rustc_driver`
Remove dependencies on `rustc_error_codes` in all crates except for `rustc_driver`.
This has some benefits:
1. Adding a new error code when hacking on the compiler only requires rebuilding at most `rustc_error_codes`, `rustc_driver`, and the reflexive & transitive closure of the crate where the new error code is being added and its reverse dependencies. This improves time-to-UI-tests (TTUT).
2. Adding an error description to an error code only requires rebuilding `rustc_error_codes` and `rustc_driver`. This should substantially improve TTUT.
r? @petrochenkov
cc @rust-lang/wg-diagnostics
Don't require `allow_internal_unstable` unless `staged_api` is enabled.
#63770 changed `qualify_min_const_fn` to require `allow_internal_unstable` for *all* crates that used an unstable feature, regardless of whether `staged_api` was enabled or the `fn` that used that feature was stably const. In practice, this meant that every crate in the ecosystem that wanted to use nightly features added `#![feature(const_fn)]`, which skips `qualify_min_const_fn` entirely.
After this PR, crates that do not have `#![feature(staged_api)]` will only need to enable the feature they are interested in. For example, `#![feature(const_if_match)]` will be enough to enable `if` and `match` in constants. Crates with `staged_api` (e.g., `libstd`) require `#[allow_internal_unstable]` to be added to a function if it uses nightly features unless that function is also marked `#[rustc_const_unstable]`. This prevents proliferation of `#[allow_internal_unstable]` into functions that are not callable in a `const` context on stable.
r? @oli-obk (author of #63770)
cc @Centril
This flag opts out of the min-const-fn checks entirely, which is usually
not what we want. The few cases where the flag is still necessary have
been annotated.
This is used for both the `?const` syntax in bounds as well as the `impl
const Trait` syntax. I also considered handling these separately by
adding a variant of `TraitBoundModifier` and a field to
`ItemKind::Impl`, but this approach was less intrusive.
Add a check for swapped words when we can't find an identifier
Fixes#66968
Couple things here:
1. The matches take the precedence of case insensitive match, then levenshtein match, then swapped words match. Doing this allows us to not even check for swapped words unless the other checks return `None`.
2. I've assumed that the swapped words check is not held to the limits of the max levenshtein distance threshold (ie. we want to try and find a match even if the levenshtein distance is very high). This means that we cannot perform this check in the `fold` that occurs after the `filter_map` call, because the candidate will be filtered out. So, I've split this into two separate `fold` calls, and had to collect the original iterator into a vec so it can be copied (I don't think we want to change the function signature to take a vec or require the `Copy` trait). An alternative implemenation may be to remove the `filter_map`, `fold` over the entire iterator, and do a check against `max_dist` inside the relevant cases there.
r? @estebank
