Smaller and more correct generator codegen
This removes unnecessary panicking branches in the resume function when the generator can not return or unwind, respectively.
Closes https://github.com/rust-lang/rust/issues/66100
It also addresses the correctness concerns wrt poisoning on unwind. These are not currently a soundness issue because any operation *inside* a generator that could possibly unwind will result in a cleanup path for dropping it, ultimately reaching a `Resume` terminator, which we already handled correctly. Future MIR optimizations might optimize that out, though.
r? @Zoxc
tidy: Better license checks.
This implements some improvements to the license checks in tidy:
* Use `cargo_metadata` instead of parsing vendored crates. This allows license checks to run without vendoring enabled, and allows the checks to run on PR builds.
* Check for stale entries.
* Check that the licenses for exceptions are what we think they are.
* Verify exceptions do not leak into the runtime.
Closes#62618Closes#62619Closes#63238 (I think)
There are some substantive changes here. The follow licenses have changed from the original comments:
* openssl BSD+advertising clause to Apache-2.0
* pest MPL2 to MIT/Apache-2.0
* smallvec MPL2 to MIT/Apache-2.0
* clippy lints MPL2 to MIT OR Apache-2.0
rustc: don't resolve Instances which would produce malformed shims.
There are some `InstanceDef` variants (shims and drop "glue") which contain a `Ty`, and that `Ty` is used in generating the shim MIR. But if that `Ty` mentions any generic parameters, the generated shim would refer to them (but they won't match the `Substs` of the `Instance`), or worse, generating the shim would fail because not enough of the type is known.
Ideally we would always produce a "skeleton" of the type, e.g. `(_, _)` for dropping any tuples with two elements, or `Vec<_>` for dropping any `Vec` value, but that's a lot of work, and they would still not match the `Substs` of the `Instance` as it exists today, so `Instance` would probably need to change.
By making `Instance::resolve` return `None` in the still-generic cases, we get behavior similar to specialization, where a default can only be used if there are no more generic parameters which would allow a more specialized `impl` to match.
<hr/>
This was found while testing the MIR inliner with #68965, because it was trying to inline shims.
cc @rust-lang/wg-mir-opt
Properly handle Spans that reference imported SourceFiles
Previously, metadata encoding used DUMMY_SP to represent any spans that
referenced an 'imported' SourceFile - e.g. a SourceFile from an upstream
dependency. This currently has no visible consequences, since these
kinds of spans don't currently seem to be emitted anywhere. However,
there's no reason that we couldn't start using such spans in
diagnostics.
This PR changes how we encode and decode spans in crate metadata. We
encode spans in one of two ways:
* 'Local' spans, which reference non-imported SourceFiles, are encoded
exactly as before.
* 'Foreign' spans, which reference imported SourceFiles, are encoded
with the CrateNum of their 'originating' crate. Additionally, their
'lo' and 'high' values are rebased on top of the 'originating' crate,
which allows them to be used with the SourceMap data encoded for that
crate.
To support this change, I've also made the following modifications:
* `DefId` and related structs are now moved to `rustc_span`. This allows
us to use a `CrateNum` inside `SourceFile`. `CrateNum` has special
handling during deserialization (it gets remapped to be the proper
`CrateNum` from the point of view of the current compilation session),
so using a `CrateNum` instead of a plain integer 'workaround type' helps
to simplify deserialization.
* The `ExternalSource` enum is renamed to `ExternalSourceKind`. There is
now a struct called `ExternalSource`, which holds an
`ExternalSourceKind` along with the original line number information for
the file. This is used during `Span` serialization to rebase spans onto
their 'owning' crate.
Previously, metadata encoding used DUMMY_SP to represent any spans that
referenced an 'imported' SourceFile - e.g. a SourceFile from an upstream
dependency. These leads to sub-optimal error messages in certain cases
(see the included test).
This PR changes how we encode and decode spans in crate metadata. We
encode spans in one of two ways:
* 'Local' spans, which reference non-imported SourceFiles, are encoded
exactly as before.
* 'Foreign' spans, which reference imported SourceFiles, are encoded
with the CrateNum of their 'originating' crate. Additionally, their
'lo' and 'high' values are rebased on top of the 'originating' crate,
which allows them to be used with the SourceMap data encoded for that
crate.
The `ExternalSource` enum is renamed to `ExternalSourceKind`. There is
now a struct called `ExternalSource`, which holds an
`ExternalSourceKind` along with the original line number information for
the file. This is used during `Span` serialization to rebase spans onto
their 'owning' crate.
ci: use python from the correct path
Apparently the old path we were using for Python 2 on Windows was not documented, and eventually got removed. This switches our CI to use the correct path.
See https://github.com/rust-lang/rust/pull/70112#issuecomment-600760786 for the actual failure.
Apparently the old path we were using for Python 2 on Windows was not
documented, and eventually got removed. This switches our CI to use the
correct path.
Tidy: fix running rustfmt twice
`./x.py test tidy` runs rustfmt twice. This is because `Build::build` runs `execute_cli` twice (once dry, once not). This can be quite slow (and prints a bunch of things twice).
I'm not sure if this is really the best place to check the dry_run status.
Remove some imports to the rustc crate
- When we have `NestedVisitorMap::None`, we use `type Map = dyn intravisit::Map<'v>;` instead of the actual map. This doesn't actually result in dynamic dispatch (in the future we may want to use an associated type default to simplify the code).
- Use `rustc_session::` imports instead of `rustc::{session, lint}`.
r? @Zoxc
Make methods declared by `newtype_index` macro `const`
Crates that use the macro to define an `Idx` type need to enable `#![feature(const_if_match, const_panic)]`.
Expansion-driven outline module parsing
After this PR, the parser will not do any conditional compilation or loading of external module files when `mod foo;` is encountered. Instead, the parser only leaves `mod foo;` in place in the AST, with no items filled in. Expansion later kicks in and will load the actual files and do the parsing. This entails that the following is now valid:
```rust
#[cfg(FALSE)]
mod foo {
mod bar {
mod baz; // `foo/bar/baz.rs` doesn't exist, but no error!
}
}
```
Fixes https://github.com/rust-lang/rust/issues/64197.
r? @petrochenkov
Use smaller discriminants for generators
Closes https://github.com/rust-lang/rust/issues/69815
I'm not yet sure about the runtime performance impact of this, so I'll try running this on some benchmarks (if I can find any). (Update: No impact on the benchmarks I've measured on)
* [x] Add test with a generator that has exactly 256 total states
* [x] Add test with a generator that has more than 256 states so that it needs to use a u16 discriminant
* [x] Add tests for the size of `Option<[generator]>`
* [x] Add tests for the `discriminant_value` intrinsic in all cases
Erase regions in writeback
Regions in `TypeckTables` (except canonicalized user annotations) are now erased. Further, we no longer do lexical region solving on item bodies with `-Zborrowck=mir`.
cc #68261
r? @nikomatsakis
Fix bugs in Peekable and Flatten when using non-fused iterators
I fixed a couple of bugs with regard to the `Peekable` and `Flatten`/`FlatMap` iterators when the underlying iterator isn't fused. For testing, I also added a `NonFused` iterator wrapper that panics when `next` or `next_back` is called on an iterator that has returned `None` before, which will hopefully make it easier to spot these mistakes in the future.
### Peekable
`Peekable::next_back` was implemented as
```rust
self.iter.next_back().or_else(|| self.peeked.take().and_then(|x| x))
```
which is incorrect because when the `peeked` field is `Some(None)`, then `None` has already been returned from the inner iterator and what it returns from `next_back` can no longer be relied upon. `test_peekable_non_fused` tests this.
### Flatten
When a `FlattenCompat` instance only has a `backiter` remaining (i.e. `self.frontiter` is `None` and `self.iter` is empty), then `next` will call `self.iter.next()` every time, so the `iter` field needs to be fused. I fixed it by giving it the type `Fuse<I>` instead of `I`, I think this is the only way to fix it. `test_flatten_non_fused_outer` tests this.
Furthermore, previously `FlattenCompat::next` did not set `self.frontiter` to `None` after it returned `None`, which is incorrect when the inner iterator type isn't fused. I just delegated it to `try_fold` because that already handles it correctly. `test_flatten_non_fused_inner` tests this.
r? @scottmcm
