Add `*_value` methods to proc_macro lib
This is the (re-)implementation of https://github.com/rust-lang/libs-team/issues/459.
It allows to get the actual value (unescaped) of the different string literals.
It was originally done in https://github.com/rust-lang/rust/pull/136355 but it broke the artifacts build so we decided to move the crate to crates.io to go around this limitation.
Part of https://github.com/rust-lang/rust/issues/136652.
Considering this is a copy-paste of the originally approved PR, no need to go through the whole process again. \o/
r? `@Urgau`
Autodiff batching
Enzyme supports batching, which is especially known from the ML side when training neural networks.
There we would normally have a training loop, where in each iteration we would pass in some data (e.g. an image), and a target vector. Based on how close we are with our prediction we compute our loss, and then use backpropagation to compute the gradients and update our weights.
That's quite inefficient, so what you normally do is passing in a batch of 8/16/.. images and targets, and compute the gradients for those all at once, allowing better optimizations.
Enzyme supports batching in two ways, the first one (which I implemented here) just accepts a Batch size,
and then each Dual/Duplicated argument has not one, but N shadow arguments. So instead of
```rs
for i in 0..100 {
df(x[i], y[i], 1234);
}
```
You can now do
```rs
for i in 0..100.step_by(4) {
df(x[i+0],x[i+1],x[i+2],x[i+3], y[i+0], y[i+1], y[i+2], y[i+3], 1234);
}
```
which will give the same results, but allows better compiler optimizations. See the testcase for details.
There is a second variant, where we can mark certain arguments and instead of having to pass in N shadow arguments, Enzyme assumes that the argument is N times longer. I.e. instead of accepting 4 slices with 12 floats each, we would accept one slice with 48 floats. I'll implement this over the next days.
I will also add more tests for both modes.
For any one preferring some more interactive explanation, here's a video of Tim's llvm dev talk, where he presents his work. https://www.youtube.com/watch?v=edvaLAL5RqU
I'll also add some other docs to the dev guide and user docs in another PR.
r? ghost
Tracking:
- https://github.com/rust-lang/rust/issues/124509
- https://github.com/rust-lang/rust/issues/135283
In the AST, currently we use `BinOpKind` within `ExprKind::AssignOp` and
`AssocOp::AssignOp`, even though this allows some nonsensical
combinations. E.g. there is no `&&=` operator. Likewise for HIR and
THIR.
This commit introduces `AssignOpKind` which only includes the ten
assignable operators, and uses it in `ExprKind::AssignOp` and
`AssocOp::AssignOp`. (And does similar things for `hir::ExprKind` and
`thir::ExprKind`.) This avoids the possibility of nonsensical
combinations, as seen by the removal of the `bug!` case in
`lang_item_for_binop`.
The commit is mostly plumbing, including:
- Adds an `impl From<AssignOpKind> for BinOpKind` (AST) and `impl
From<AssignOp> for BinOp` (MIR/THIR).
- `BinOpCategory` can now be created from both `BinOpKind` and
`AssignOpKind`.
- Replaces the `IsAssign` type with `Op`, which has more information and
a few methods.
- `suggest_swapping_lhs_and_rhs`: moves the condition to the call site,
it's easier that way.
- `check_expr_inner`: had to factor out some code into a separate
method.
I'm on the fence about whether avoiding the nonsensical combinations is
worth the extra code.
They are no longer needed.
This does slightly worsen the error message for a single test, but that
test contains code that is so badly broken that I'm not worried about
it.
Notes about tests:
- tests/ui/rfcs/rfc-2294-if-let-guard/feature-gate.rs: some messages are
now duplicated due to repeated parsing.
- tests/ui/rfcs/rfc-2497-if-let-chains/disallowed-positions.rs: ditto.
- `tests/ui/proc-macro/macro-rules-derive-cfg.rs`: the diff looks large
but the only difference is the insertion of a single
invisible-delimited group around a metavar.
- `tests/ui/attributes/nonterminal-expansion.rs`: a slight span
degradation, somehow related to the recent massive attr parsing
rewrite (#135726). I couldn't work out exactly what is going wrong,
but I don't think it's worth holding things up for a single slightly
suboptimal error message.
`ast::Item` has an `ident` field.
- It's always non-empty for these item kinds: `ExternCrate`, `Static`,
`Const`, `Fn`, `Mod`, `TyAlias`, `Enum`, `Struct`, `Union`,
`Trait`, `TraitAlias`, `MacroDef`, `Delegation`.
- It's always empty for these item kinds: `Use`, `ForeignMod`,
`GlobalAsm`, `Impl`, `MacCall`, `DelegationMac`.
There is a similar story for `AssocItemKind` and `ForeignItemKind`.
Some sites that handle items check for an empty ident, some don't. This
is a very C-like way of doing things, but this is Rust, we have sum
types, we can do this properly and never forget to check for the
exceptional case and never YOLO possibly empty identifiers (or possibly
dummy spans) around and hope that things will work out.
The commit is large but it's mostly obvious plumbing work. Some notable
things.
- `ast::Item` got 8 bytes bigger. This could be avoided by boxing the
fields within some of the `ast::ItemKind` variants (specifically:
`Struct`, `Union`, `Enum`). I might do that in a follow-up; this
commit is big enough already.
- For the visitors: `FnKind` no longer needs an `ident` field because
the `Fn` within how has one.
- In the parser, the `ItemInfo` typedef is no longer needed. It was used
in various places to return an `Ident` alongside an `ItemKind`, but
now the `Ident` (if present) is within the `ItemKind`.
- In a few places I renamed identifier variables called `name` (or
`foo_name`) as `ident` (or `foo_ident`), to better match the type, and
because `name` is normally used for `Symbol`s. It's confusing to see
something like `foo_name.name`.
"Missing" patterns are possible in bare fn types (`fn f(u32)`) and
similar places. Currently these are represented in the AST with
`ast::PatKind::Ident` with no `by_ref`, no `mut`, an empty ident, and no
sub-pattern. This flows through to `{hir,thir}::PatKind::Binding` for
HIR and THIR.
This is a bit nasty. It's very non-obvious, and easy to forget to check
for the exceptional empty identifier case.
This commit adds a new variant, `PatKind::Missing`, to do it properly.
The process I followed:
- Add a `Missing` variant to `{ast,hir,thir}::PatKind`.
- Chang `parse_param_general` to produce `ast::PatKind::Missing`
instead of `ast::PatKind::Missing`.
- Look through `kw::Empty` occurrences to find functions where an
existing empty ident check needs replacing with a `PatKind::Missing`
check: `print_param`, `check_trait_item`, `is_named_param`.
- Add a `PatKind::Missing => unreachable!(),` arm to every exhaustive
match identified by the compiler.
- Find which arms are actually reachable by running the test suite,
changing them to something appropriate, usually by looking at what
would happen to a `PatKind::Ident`/`PatKind::Binding` with no ref, no
`mut`, an empty ident, and no subpattern.
Quite a few of the `unreachable!()` arms were never reached. This makes
sense because `PatKind::Missing` can't happen in every pattern, only
in places like bare fn tys and trait fn decls.
I also tried an alternative approach: modifying `ast::Param::pat` to
hold an `Option<P<Pat>>` instead of a `P<Pat>`, but that quickly turned
into a very large and painful change. Adding `PatKind::Missing` is much
easier.
Mostly parser: Eliminate code that's been dead / semi-dead since the removal of type ascription syntax
**Disclaimer**: This PR is intended to mostly clean up code as opposed to bringing about behavioral changes. Therefore it doesn't aim to address any of the 'FIXME: remove after a month [dated: 2023-05-02]: "type ascription syntax has been removed, see issue [#]101728"'.
---
By commit:
1. Removes truly dead code:
* Since 1.71 (#109128) `let _ = { f: x };` is a syntax error as opposed to a semantic error which allows the parse-time diagnostic (suggestion) "*struct literal body without path // you might have forgotten […]*" to kick in.
* The analysis-time diagnostic (suggestion) from <=1.70 "*cannot find value \`f\` in this scope // you might have forgotten […]*" is therefore no longer reachable.
2. Updates `is_certainly_not_a_block` to be in line with the current grammar:
* The seq. `{ ident:` is definitely not the start of a block. Before the removal of ty ascr, `{ ident: ty_start` would begin a block expr.
* This shouldn't make more code compile IINM, it should *ultimately* only affect diagnostics.
* For example, `if T { f: () } {}` will now be interpreted as an `if` with struct lit `T { f: () }` as its *condition* (which is banned in the parser anyway) as opposed to just `T` (with the *consequent* being `f : ()` which is also invalid (since 1.71)). The diagnostics are almost the same because we have two separate parse recovery procedures + diagnostics: `StructLiteralNeedingParens` (*invalid struct lit*) before and `StructLiteralNotAllowedHere` (*struct lits aren't allowed here*) now, as you can see from the diff.
* (As an aside, even before this PR, fn `maybe_suggest_struct_literal` should've just used the much older & clearer `StructLiteralNotAllowedHere`)
* NB: This does sadly regress the compiler output for `tests/ui/parser/type-ascription-in-pattern.rs` but that can be fixed in follow-up PRs. It's not super important IMO and a natural consequence.
3. Removes code that's become dead due to the prior commit.
* Basically reverts #106620 + #112475 (without regressing rustc's output!).
* Now the older & more robust parse recovery procedure (cc `StructLiteralNotAllowedHere`) takes care of the cases the removed code used to handle.
* This automatically fixes the suggestions for \[[playground](https://play.rust-lang.org/?version=stable&mode=debug&edition=2024&gist=7e2030163b11ee96d17adc3325b01780)\]:
* `if Ty::<i32> { f: K }.m() {}`: `if Ty::<i32> { SomeStruct { f: K } }.m() {}` (broken) → ` if (Ty::<i32> { f: K }).m() {}`
* `if <T as Trait>::Out { f: K::<> }.m() {}`: `if <T as Trait>(::Out { f: K::<> }).m() {}` (broken) → `if (<T as Trait>::Out { f: K::<> }).m() {}`
4. Merge and simplify UI tests pertaining to this issue, so it's easier to add more regression tests like for the two cases mentioned above.
5. Merge UI tests and add the two regression tests.
Best reviewed commit by commit (on request I'll partially squash after approval).
Visitors track whether an assoc item is in a trait impl or an inherent impl
`AssocCtxt::Impl` now contains an `of_trait` field. This allows ast lowering and nameres to not have to track whether we're in a trait impl or an inherent impl.
They're dodgy, covering all the keywords, including weak ones, and
edition-specific ones without considering the edition. They have a
single use in rustfmt. This commit changes that use to
`is_reserved_ident`, which is a much more widely used alternative and is
good enough, judging by the lack of effect on the test suite.
Handle spans of `~const`, `const` and `async` trait bounds in macro expansion
r? `@compiler-errors`
`visit_span` is actually only used in one place (the `transcribe::Marker`), and all of this syntax is unstable, so while it would still be nice to write a test for it, I wager there's lots more interesting things in `transcribe::Marker` to write tests for. And the worst is some diagnostics being weird or incremental being not as incremental as it could be
The idea is to identify cases of symbols/identifiers that are not
expected to be used. There isn't a perfectly sharp line between "dummy"
and "not dummy", but I think it's useful nonetheless.
Add `#[define_opaques]` attribute and require it for all type-alias-impl-trait sites that register a hidden type
Instead of relying on the signature of items to decide whether they are constraining an opaque type, the opaque types that the item constrains must be explicitly listed.
A previous version of this PR used an actual attribute, but had to keep the resolved `DefId`s in a side table.
Now we just lower to fields in the AST that have no surface syntax, instead a builtin attribute macro fills in those fields where applicable.
Note that for convenience referencing opaque types in associated types from associated methods on the same impl will not require an attribute. If that causes problems `#[defines()]` can be used to overwrite the default of searching for opaques in the signature.
One wart of this design is that closures and static items do not have generics. So since I stored the opaques in the generics of functions, consts and methods, I would need to add a custom field to closures and statics to track this information. During a T-types discussion we decided to just not do this for now.
fixes#131298
Reduce formatting `width` and `precision` to 16 bits
This is part of https://github.com/rust-lang/rust/issues/99012
This is reduces the `width` and `precision` fields in format strings to 16 bits. They are currently full `usize`s, but it's a bit nonsensical that we need to support the case where someone wants to pad their value to eighteen quintillion spaces and/or have eighteen quintillion digits of precision.
By reducing these fields to 16 bit, we can reduce `FormattingOptions` to 64 bits (see https://github.com/rust-lang/rust/pull/136974) and improve the in memory representation of `format_args!()`. (See additional context below.)
This also fixes a bug where the width or precision is silently truncated when cross-compiling to a target with a smaller `usize`. By reducing the width and precision fields to the minimum guaranteed size of `usize`, 16 bits, this bug is eliminated.
This is a breaking change, but affects almost no existing code.
---
Details of this change:
There are three ways to set a width or precision today:
1. Directly a formatting string, e.g. `println!("{a:1234}")`
2. Indirectly in a formatting string, e.g. `println!("{a:width$}", width=1234)`
3. Through the unstable `FormattingOptions::width` method.
This PR:
- Adds a compiler error for 1. (`println!("{a:9999999}")` no longer compiles and gives a clear error.)
- Adds a runtime check for 2. (`println!("{a:width$}, width=9999999)` will panic.)
- Changes the signatures of the (unstable) `FormattingOptions::[get_]width` methods to use a `u16` instead.
---
Additional context for improving `FormattingOptions` and `fmt::Arguments`:
All the formatting flags and options are currently:
- The `+` flag (1 bit)
- The `-` flag (1 bit)
- The `#` flag (1 bit)
- The `0` flag (1 bit)
- The `x?` flag (1 bit)
- The `X?` flag (1 bit)
- The alignment (2 bits)
- The fill character (21 bits)
- Whether a width is specified (1 bit)
- Whether a precision is specified (1 bit)
- If used, the width (a full usize)
- If used, the precision (a full usize)
Everything except the last two can simply fit in a `u32` (those add up to 31 bits in total).
If we can accept a max width and precision of u16::MAX, we can make a `FormattingOptions` that is exactly 64 bits in size; the same size as a thin reference on most platforms.
If, additionally, we also limit the number of formatting arguments, we can also reduce the size of `fmt::Arguments` (that is, of a `format_args!()` expression).
Revert <https://github.com/rust-lang/rust/pull/138084> to buy time to
consider options that avoids breaking downstream usages of cargo on
distributed `rustc-src` artifacts, where such cargo invocations fail due
to inability to inherit `lints` from workspace root manifest's
`workspace.lints` (this is only valid for the source rust-lang/rust
workspace, but not really the distributed `rustc-src` artifacts).
This breakage was reported in
<https://github.com/rust-lang/rust/issues/138304>.
This reverts commit 48caf81484b50dca5a5cebb614899a3df81ca898, reversing
changes made to c6662879b27f5161e95f39395e3c9513a7b97028.
By naming them in `[workspace.lints.rust]` in the top-level
`Cargo.toml`, and then making all `compiler/` crates inherit them with
`[lints] workspace = true`. (I omitted `rustc_codegen_{cranelift,gcc}`,
because they're a bit different.)
The advantages of this over the current approach:
- It uses a standard Cargo feature, rather than special handling in
bootstrap. So, easier to understand, and less likely to get
accidentally broken in the future.
- It works for proc macro crates.
It's a shame it doesn't work for rustc-specific lints, as the comments
explain.
This involves replacing `nt_pretty_printing_compatibility_hack` with
`stream_pretty_printing_compatibility_hack`.
The handling of statements in `transcribe` is slightly different to
other nonterminal kinds, due to the lack of `from_ast` implementation
for empty statements.
Notable test changes:
- `tests/ui/proc-macro/expand-to-derive.rs`: the diff looks large but
the only difference is the insertion of a single invisible-delimited
group around a metavar.
