implement Ord for OutlivesPredicate and other types
It became necessary while implementing https://github.com/rust-lang/rust/pull/50070 to have `Ord` implemented for `OutlivesPredicate`.
This PR implements `Ord` for `OutlivesPredicate` as well as other types needed for the implementation.
Fix span for type-only arguments
Currently it points to the comma or parenthesis before the type, which is broken
cc @mark-i-m this is what broke #48309
r? @estebank
rustc: Correctly pretty-print macro delimiters
This commit updates the `Mac_` AST structure to keep track of the delimiters
that it originally had for its invocation. This allows us to faithfully
pretty-print macro invocations not using parentheses (e.g. `vec![...]`). This in
turn helps procedural macros due to #43081.
Closes#50840
This commit updates the `Mac_` AST structure to keep track of the delimiters
that it originally had for its invocation. This allows us to faithfully
pretty-print macro invocations not using parentheses (e.g. `vec![...]`). This in
turn helps procedural macros due to #43081.
Closes#50840
rustc: Fix procedural macros generating lifetime tokens
This commit fixes an accidental regression from #50473 where lifetime tokens
produced by procedural macros ended up getting lost in translation in the
compiler and not actually producing parseable code. The issue lies in the fact
that a lifetime's `Ident` is prefixed with `'`. The `glue` implementation for
gluing joint tokens together forgot to take this into account so the lifetime
inside of `Ident` was missing the leading tick!
The `glue` implementation here is updated to create a new `Symbol` in these
situations to manufacture a new `Ident` with a leading tick to ensure it parses
correctly.
Closes#50942
rustc: Fix joint-ness of stringified token-streams
This commit fixes `StringReader`'s parsing of tokens which have been stringified
through procedural macros. Whether or not a token tree is joint is defined by
span information, but when working with procedural macros these spans are often
dummy and/or overridden which means that they end up considering all operators
joint if they can!
The fix here is to track the raw source span as opposed to the overridden span.
With this information we can more accurately classify `Punct` structs as either
joint or not.
Closes#50700
This commit fixes an accidental regression from #50473 where lifetime tokens
produced by procedural macros ended up getting lost in translation in the
compiler and not actually producing parseable code. The issue lies in the fact
that a lifetime's `Ident` is prefixed with `'`. The `glue` implementation for
gluing joint tokens together forgot to take this into account so the lifetime
inside of `Ident` was missing the leading tick!
The `glue` implementation here is updated to create a new `Symbol` in these
situations to manufacture a new `Ident` with a leading tick to ensure it parses
correctly.
Closes#50942
lexer: Fix span override for the first token in a string
Previously due to peculiarities of `StringReader` construction something like `"a b c d".parse::<TokenStream>()` gave you one non-overridden span for `a` and then three correctly overridden spans for `b`, `c` and `d`.
Now all the spans are overridden.
rustc: introduce {ast,hir}::AnonConst to consolidate so-called "embedded constants".
Previously, constants in array lengths and enum variant discriminants were "merely an expression", and had no separate ID for, e.g. type-checking or const-eval, instead reusing the expression's.
That complicated code working with bodies, because such constants were the only special case where the "owner" of the body wasn't the HIR parent, but rather the same node as the body itself.
Also, if the body happened to be a closure, we had no way to allocate a `DefId` for both the constant *and* the closure, leading to *several* bugs (mostly ICEs where type errors were expected).
This PR rectifies the situation by adding another (`{ast,hir}::AnonConst`) node around every such constant. Also, const generics are expected to rely on the new `AnonConst` nodes, as well (cc @varkor).
* fixes#48838
* fixes#50600
* fixes#50688
* fixes#50689
* obsoletes #50623
r? @nikomatsakis
Consider this a down payment on #50723. To recap, an `Applicability`
enum was recently (#50204) added, to convey to Rustfix and other tools
whether we think it's OK for them to blindly apply the suggestion, or
whether to prompt a human for guidance (because the suggestion might
contain placeholders that we can't infer, or because we think it has a
sufficiently high probability of being wrong even though it's—
presumably—right often enough to be worth emitting in the first place).
When a suggestion is marked as `MaybeIncorrect`, we try to use comments
to indicate precisely why (although there are a few places where we just
say `// speculative` because the present author's subjective judgement
balked at the idea that the suggestion has no false positives).
The `run-rustfix` directive is opporunistically set on some relevant UI
tests (and a couple tests that were in the `test/ui/suggestions`
directory, even if the suggestions didn't originate in librustc or
libsyntax). This is less trivial than it sounds, because a surprising
number of test files aren't equipped to be tested as fixed even when
they contain successfully fixable errors, because, e.g., there are more,
not-directly-related errors after fixing. Some test files need an
attribute or underscore to avoid unused warnings tripping up the "fixed
code is still producing diagnostics" check despite the fixes being
correct; this is an interesting contrast-to/inconsistency-with the
behavior of UI tests (which secretly pass `-A unused`), a behavior which
we probably ought to resolve one way or the other (filed issue #50926).
A few suggestion labels are reworded (e.g., to avoid phrasing it as a
question, which which is discouraged by the style guidelines listed in
`.span_suggestion`'s doc-comment).
Speed up the macro parser
These three commits reduce the number of allocations done by the macro parser, in some cases dramatically. For example, for a clean check builds of html5ever, the number of allocations is reduced by 40%.
Here are the rustc-benchmarks that are sped up by at least 1%.
```
html5ever-check
avg: -6.6% min: -10.3% max: -4.1%
html5ever
avg: -5.2% min: -9.5% max: -2.8%
html5ever-opt
avg: -4.3% min: -9.3% max: -1.6%
crates.io-check
avg: -1.8% min: -2.9% max: -0.6%
crates.io-opt
avg: -1.0% min: -2.2% max: -0.1%
crates.io
avg: -1.1% min: -2.2% max: -0.2%
```
rustc: Disallow modules and macros in expansions
This commit feature gates generating modules and macro definitions in procedural
macro expansions. Custom derive is exempt from this check as it would be a large
retroactive breaking change (#50587). It's hoped that we can hopefully stem the
bleeding to figure out a better solution here before opening up the floodgates.
The restriction here is specifically targeted at surprising hygiene results [1]
that result in non-"copy/paste" behavior. Hygiene and procedural macros is
intended to be avoided as much as possible for Macros 1.2 by saying everything
is "as if you copy/pasted the code", but modules and macros are sort of weird
exceptions to this rule that aren't fully fleshed out.
[1]: https://github.com/rust-lang/rust/issues/50504#issuecomment-387734625
cc #50504
This commit feature gates generating modules and macro definitions in procedural
macro expansions. Custom derive is exempt from this check as it would be a large
retroactive breaking change (#50587). It's hoped that we can hopefully stem the
bleeding to figure out a better solution here before opening up the floodgates.
The restriction here is specifically targeted at surprising hygiene results [1]
that result in non-"copy/paste" behavior. Hygiene and procedural macros is
intended to be avoided as much as possible for Macros 1.2 by saying everything
is "as if you copy/pasted the code", but modules and macros are sort of weird
exceptions to this rule that aren't fully fleshed out.
[1]: https://github.com/rust-lang/rust/issues/50504#issuecomment-387734625
cc #50504
This commit fixes `StringReader`'s parsing of tokens which have been stringified
through procedural macros. Whether or not a token tree is joint is defined by
span information, but when working with procedural macros these spans are often
dummy and/or overridden which means that they end up considering all operators
joint if they can!
The fix here is to track the raw source span as opposed to the overridden span.
With this information we can more accurately classify `Punct` structs as either
joint or not.
Closes#50700
Implement edition hygiene for keywords
Determine "keywordness" of an identifier in its hygienic context.
cc https://github.com/rust-lang/rust/pull/49611
I've resurrected `proc` as an Edition-2015-only keyword for testing purposes, but it should probably be buried again. EDIT: `proc` is removed again.
Review proc macro API 1.2
cc https://github.com/rust-lang/rust/issues/38356
Summary of applied changes:
- Documentation for proc macro API 1.2 is expanded.
- Renamed APIs: `Term` -> `Ident`, `TokenTree::Term` -> `TokenTree::Ident`, `Op` -> `Punct`, `TokenTree::Op` -> `TokenTree::Punct`, `Op::op` -> `Punct::as_char`.
- Removed APIs: `Ident::as_str`, use `Display` impl for `Ident` instead.
- New APIs (not stabilized in 1.2): `Ident::new_raw` for creating a raw identifier (I'm not sure `new_x` it's a very idiomatic name though).
- Runtime changes:
- `Punct::new` now ensures that the input `char` is a valid punctuation character in Rust.
- `Ident::new` ensures that the input `str` is a valid identifier in Rust.
- Lifetimes in proc macros are now represented as two joint tokens - `Punct('\'', Spacing::Joint)` and `Ident("lifetime_name_without_quote")` similarly to multi-character operators.
- Stabilized APIs: None yet.
A bit of motivation for renaming (although it was already stated in the review comments):
- With my compiler frontend glasses on `Ident` is the single most appropriate name for this thing, *especially* if we are doing input validation on construction. `TokenTree::Ident` effectively wraps `token::Ident` or `ast::Ident + is_raw`, its meaning is "identifier" and it's already named `ident` in declarative macros.
- Regarding `Punct`, the motivation is that `Op` is actively misleading. The thing doesn't mean an operator, it's neither a subset of operators (there is non-operator punctuation in the language), nor superset (operators can be multicharacter while this thing is always a single character). So I named it `Punct` (first proposed in [the original RFC](https://github.com/rust-lang/rfcs/pull/1566), then [by @SimonSapin](https://github.com/rust-lang/rust/issues/38356#issuecomment-276676526)) , together with input validation it's now a subset of ASCII punctuation character category (`u8::is_ascii_punctuation`).
It only has a single use, within code handling indented block comments.
We can replace that with the new `FileMap::col_pos()`, which computes
the col position (BytePos instead of CharPos) based on the record of the
last newline char (which we already record).
This is actually an improvement, because
`trim_whitespace_prefix_and_push_line()` was using `col`, which is a
`CharPos`, as a slice index, which is a byte/char confusion.
It's silly for a hot function like `bump()` to have such an expensive
bounds check. This patch replaces terminator with `end_src_index`.
Note that the `self.terminator` check in `is_eof()` wasn't necessary
because of the way `StringReader` is initialized.
- `source_text` becomes `src`, matching `FileMap::src`.
- `byte_offset()` becomes `src_index()`, which makes it clearer that
it's an index into `src`. (Likewise for variables containing
`byte_offset` in their name.) This function also now returns a `usize`
instead of a `BytePos`, because every callsite immediately converted
the `BytePos` to a `usize`.
This patch removes the "old"/"new" names in favour of "foo"/"next_foo",
which matches the field names.
It also moves the setting of `self.{ch,pos,next_pos}` in the common case
to the end, so that the meaning of "foo"/"next_foo" is consistent until
the end.
This commit fixes a hard error where the `#![feature(rust_2018_preview)]`
feature was forbidden to be mentioned when the `--edition 2018` flag was passed.
This instead silently accepts that feature gate despite it not being necessary.
It's intended that this will help ease the transition into the 2018 edition as
users will, for the time being, start off with the `rust_2018_preview` feature
and no longer immediately need to remove it.
Closes#50662
Rename the 2018 edition lint names
* `rust_2018_breakage` -> `rust_2018_compatibility` - the lint for ensuring
that your code, in the 2015 edition, is compatible with the 2018 edition's
semantics. This is required to pass *before* you enable the 2018 edition.
* `rust_2018_migration` -> `rust_2018_idioms` - the lint for writing idiomatic
code after you've already enabled the 2018 edition
* `rust_2018_breakage` -> `rust_2018_compatibility` - the lint for ensuring
that your code, in the 2015 edition, is compatible with the 2018 edition's
semantics. This is required to pass *before* you enable the 2018 edition.
* `rust_2018_migration` -> `rust_2018_idioms` - the lint for writing idiomatic
code after you've already enabled the 2018 edition
Optimize string handling in lit_token().
In the common case, the string value in a string literal Token is the
same as the string value in a string literal LitKind. (The exception is
when escapes or \r are involved.) This patch takes advantage of that to
avoid calling str_lit() and re-interning the string in that case. This
speeds up incremental builds for a few of the rustc-benchmarks, the best
by 3%.
Benchmarks that got a speedup of 1% or more:
```
coercions
avg: -1.1% min: -3.5% max: 0.4%
regex-check
avg: -1.2% min: -1.5% max: -0.6%
futures-check
avg: -0.9% min: -1.4% max: -0.3%
futures
avg: -0.8% min: -1.3% max: -0.3%
futures-opt
avg: -0.7% min: -1.2% max: -0.1%
regex
avg: -0.5% min: -1.2% max: -0.1%
regex-opt
avg: -0.5% min: -1.1% max: -0.1%
hyper-check
avg: -0.7% min: -1.0% max: -0.3%
```
In the common case, the string value in a string literal Token is the
same as the string value in a string literal LitKind. (The exception is
when escapes or \r are involved.) This patch takes advantage of that to
avoid calling str_lit() and re-interning the string in that case. This
speeds up incremental builds for a few of the rustc-benchmarks, the best
by 3%.
Rollup of 12 pull requests
Successful merges:
- #50302 (Add query search order check)
- #50320 (Fix invalid path generation in rustdoc search)
- #50349 (Rename "show type declaration" to "show declaration")
- #50360 (Clarify wordings of the `unstable_name_collision` lint.)
- #50365 (Use two vectors in nearest_common_ancestor.)
- #50393 (Allow unaligned reads in constants)
- #50401 (Revert "Implement FromStr for PathBuf")
- #50406 (Forbid constructing empty identifiers from concat_idents)
- #50407 (Always inline simple BytePos and CharPos methods.)
- #50416 (check if the token is a lifetime before parsing)
- #50417 (Update Cargo)
- #50421 (Fix ICE when using a..=b in a closure.)
Failed merges:
Implement tool_attributes feature (RFC 2103)
cc #44690
This is currently just a rebased and compiling (hopefully) version of #47773.
Let's see if travis likes this. I will add the implementation for `tool_lints` this week.
Reduce maximum repr(align(N)) to 2^29
The current maximum `repr(align(N))` alignment is larger than the maximum alignment accepted by LLVM, which can cause issues for huge values of `N`, as seen in #49492. Fixes#49492.
r? @rkruppe
This avoids converting every char to \u{...} form, which bloats the
resulting strings unnecessarily. It also provides consistency with the
existing escape_default() calls in LitKind::token() used for raw
string literals, char literals, and raw byte char literals.
There are two benefits from this change.
- Compilation is faster. Most of the rustc-perf benchmarks see a
non-trivial speedup, particularly for incremental rebuilds, with the
best speedup over 13%, and multiple others over 10%.
- Generated rlibs are smaller. An extreme example is libfutures.rlib,
which shrinks from 2073306 bytes to 1765927 bytes, a 15% reduction.
Module experiments: Add one more prelude layer for extern crate names passed with `--extern`
Implements one item from https://internals.rust-lang.org/t/the-great-module-adventure-continues/6678/183
When some name is looked up in lexical scope (`name`, i.e. not module-relative scope `some_mod::name` or `::name`), it's searched roughly in the next order:
- local variables
- items in unnamed blocks
- items in the current module
- ✨ NEW! ✨ crate names passed with `--extern` ("extern prelude")
- standard library prelude (`Vec`, `drop`)
- language prelude (built-in types like `u8`, `str`, etc)
The last two layers contain a limited set of names controlled by us and not arbitrary user-defined names like upper layers. We want to be able to add new names into these two layers without breaking user code, so "extern prelude" names have higher priority than std prelude and built-in types.
This is a one-time breaking change, that's why it would be nice to run this through crater.
Practical impact is expected to be minimal though due to stylistic reasons (there are not many `Uppercase` crates) and due to the way how primitive types are resolved (https://github.com/rust-lang/rust/pull/32131).
Warn on pointless #[derive] in more places
This fixes the regression in #49934 and ensures that unused `#[derive]` invocations on statements, expressions and generic type parameters survive to trip the `unused_attributes` lint. There is a separate warning hardcoded for `#[derive]` on macro invocations since linting (even the early-lint pass) occurs after expansion. This also adds regression tests for some nodes that were already warning properly.
closes#49934
This fixes the regression in #49934 and ensures that unused `#[derive]`s on statements, expressions and generic type parameters survive to trip the `unused_attributes` lint. For `#[derive]` on macro invocations it has a hardcoded warning since linting occurs after expansion. This also adds regression testing for some nodes that were already warning properly.
closes#49934
don't see issue #0
The unstable-feature attribute requires an issue (neglecting it is
E0547), which gets used in the error messages. Unfortunately, there are
some cases where "0" is apparently used a placeholder where no issue
exists, directing the user to see the (nonexistent) issue #0. (It would
have been better to either let `issue` be optional—compare to how issue
is an `Option<u32>` in the feature-gate declarations in
libsyntax/feature-gate.rs—or actually require that an issue be created.)
Rather than endeavoring to change how `#[unstable]` works at this time
(given competing contributor and reviewer priorities), this simple patch
proposes the less-ambitious solution of just not adding the "(see
issue)" note when the number is zero.
Resolves#49983.
Revert stabilization of never_type (!) et al
Fix#49691
I *think* this correctly adopts @nikomatsakis 's desired fix of:
* reverting stabilization of `!` and `TryFrom`, and
* returning to the previous fallback semantics (i.e. it is once again dependent on whether the crate has opted into `#[feature(never_type)]`,
* **without** attempting to put back in the previous future-proofing warnings regarding the change in fallback semantics.
(I'll be away from computers for a week starting now, so any updates to this PR should be either pushed into it, or someone else should adopt the task of polishing this fix and put up their own PR.)
rustc: Tweak custom attribute capabilities
This commit starts to lay some groundwork for the stabilization of custom
attribute invocations and general procedural macros. It applies a number of
changes discussed on [internals] as well as a [recent issue][issue], namely:
* The path used to specify a custom attribute must be of length one and cannot
be a global path. This'll help future-proof us against any ambiguities and
give us more time to settle the precise syntax. In the meantime though a bare
identifier can be used and imported to invoke a custom attribute macro. A new
feature gate, `proc_macro_path_invoc`, was added to gate multi-segment paths
and absolute paths.
* The set of items which can be annotated by a custom procedural attribute has
been restricted. Statements, expressions, and modules are disallowed behind
two new feature gates: `proc_macro_expr` and `proc_macro_mod`.
* The input to procedural macro attributes has been restricted and adjusted.
Today an invocation like `#[foo(bar)]` will receive `(bar)` as the input token
stream, but after this PR it will only receive `bar` (the delimiters were
removed). Invocations like `#[foo]` are still allowed and will be invoked in
the same way as `#[foo()]`. This is a **breaking change** for all nightly
users as the syntax coming in to procedural macros will be tweaked slightly.
* Procedural macros (`foo!()` style) can only be expanded to item-like items by
default. A separate feature gate, `proc_macro_non_items`, is required to
expand to items like expressions, statements, etc.
Closes#50038
[internals]: https://internals.rust-lang.org/t/help-stabilize-a-subset-of-macros-2-0/7252
[issue]: https://github.com/rust-lang/rust/issues/50038
This commit starts to lay some groundwork for the stabilization of custom
attribute invocations and general procedural macros. It applies a number of
changes discussed on [internals] as well as a [recent issue][issue], namely:
* The path used to specify a custom attribute must be of length one and cannot
be a global path. This'll help future-proof us against any ambiguities and
give us more time to settle the precise syntax. In the meantime though a bare
identifier can be used and imported to invoke a custom attribute macro. A new
feature gate, `proc_macro_path_invoc`, was added to gate multi-segment paths
and absolute paths.
* The set of items which can be annotated by a custom procedural attribute has
been restricted. Statements, expressions, and modules are disallowed behind
two new feature gates: `proc_macro_expr` and `proc_macro_mod`.
* The input to procedural macro attributes has been restricted and adjusted.
Today an invocation like `#[foo(bar)]` will receive `(bar)` as the input token
stream, but after this PR it will only receive `bar` (the delimiters were
removed). Invocations like `#[foo]` are still allowed and will be invoked in
the same way as `#[foo()]`. This is a **breaking change** for all nightly
users as the syntax coming in to procedural macros will be tweaked slightly.
* Procedural macros (`foo!()` style) can only be expanded to item-like items by
default. A separate feature gate, `proc_macro_non_items`, is required to
expand to items like expressions, statements, etc.
Closes#50038
[internals]: https://internals.rust-lang.org/t/help-stabilize-a-subset-of-macros-2-0/7252
[issue]: https://github.com/rust-lang/rust/issues/50038
This commit is just covering the feature gate itself and the tests
that made direct use of `!` and thus need to opt back into the
feature.
A follow on commit brings back the other change that motivates the
revert: Namely, going back to the old rules for falling back to `()`.
This commit transitions the `target_feature` attribute from `Normal` to
`Whitelisted`. Discovered in #50095 the fact of whether this attribute is used
or not is dependent on typechecking running and executing `check_name`, but
incremental compilation doesn't currently account for this, meaning that the
attribute ends up being flagged as unused when it shouldn't be.
I was a little too ambitious it seems hoping that `Normal` could be used, so
instead this transitions to `Whitelisted` to be the same as other codegen
attributes like `#[inline]`
Closes#50095
Discovered in #50061 we're falling off the "happy path" of using a stringified
token stream more often than we should. This was due to the fact that a
user-written token like `0xf` is equality-different from the stringified token
of `15` (despite being semantically equivalent).
This patch updates the call to `eq_unspanned` with an even more awful solution,
`probably_equal_for_proc_macro`, which ignores the value of each token and
basically only compares the structure of the token stream, assuming that the AST
doesn't change just one token at a time.
While this is a step towards fixing #50061 there is still one regression
from #49154 which needs to be fixed.
`char_lit` uses an allocation in order to ignore '_' chars in \u{...}
literals. This patch changes it to not do that by processing the chars
more directly.
This improves various rustc-perf benchmark measurements by up to 6%,
particularly regex, futures, clap, coercions, hyper, and encoding.
Stabilize x86/x86_64 SIMD
This commit stabilizes the SIMD in Rust for the x86/x86_64 platforms. Notably
this commit is stabilizing:
* The `std::arch::{x86, x86_64}` modules and the intrinsics contained inside.
* The `is_x86_feature_detected!` macro in the standard library
* The `#[target_feature(enable = "...")]` attribute
* The `#[cfg(target_feature = "...")]` matcher
Stabilization of the module and intrinsics were primarily done in
rust-lang-nursery/stdsimd#414 and the two attribute stabilizations are done in
this commit. The standard library is also tweaked a bit with the new way that
stdsimd is integrated.
Note that other architectures like `std::arch::arm` are not stabilized as part
of this commit, they will likely stabilize in the future after they've been
implemented and fleshed out. Similarly the `std::simd` module is also not being
stabilized in this commit, only `std::arch`. Finally, nothing related to `__m64`
is stabilized in this commit either (MMX), only SSE and up types and intrinsics
are stabilized.
Closes#29717Closes#44839Closes#48556
This commit stabilizes the SIMD in Rust for the x86/x86_64 platforms. Notably
this commit is stabilizing:
* The `std::arch::{x86, x86_64}` modules and the intrinsics contained inside.
* The `is_x86_feature_detected!` macro in the standard library
* The `#[target_feature(enable = "...")]` attribute
* The `#[cfg(target_feature = "...")]` matcher
Stabilization of the module and intrinsics were primarily done in
rust-lang-nursery/stdsimd#414 and the two attribute stabilizations are done in
this commit. The standard library is also tweaked a bit with the new way that
stdsimd is integrated.
Note that other architectures like `std::arch::arm` are not stabilized as part
of this commit, they will likely stabilize in the future after they've been
implemented and fleshed out. Similarly the `std::simd` module is also not being
stabilized in this commit, only `std::arch`. Finally, nothing related to `__m64`
is stabilized in this commit either (MMX), only SSE and up types and intrinsics
are stabilized.
Closes#29717Closes#44839Closes#48556
Update `?` repetition disambiguation.
**Do not merge** (yet)
This is a test implementation of some ideas from discussion in https://github.com/rust-lang/rust/issues/48075 . This PR
- disallows `?` repetition from taking a separator, since the separator is never used.
- disallows the use of `?` as a separator. This allows patterns like `$(a)?+` to match `+` and `a+` rather than `a?a?a`. This is a _breaking change_, but maybe that's ok? Perhaps a crater run is the right approach?
cc @durka @alexreg @nikomatsakis
The unstable-feature attribute requires an issue (neglecting it is
E0547), which gets used in the error messages. Unfortunately, there are
some cases where "0" is apparently used a placeholder where no issue
exists, directing the user to see the (nonexistent) issue #0. (It would
have been better to either let `issue` be optional—compare to how issue
is an `Option<u32>` in the feature-gate declarations in
libsyntax/feature-gate.rs—or actually require that an issue be created.)
Rather than endeavoring to change how `#[unstable]` works at this time
(given competing contributor and reviewer priorities), this simple patch
proposes the less-ambitious solution of just not adding the "(see
issue)" note when the number is zero.
Resolves#49983.
Add error codes for libsyntax_ext
I intend to add error codes for `libsyntax_ext` as well. However, they cannot be used at stage 0 directly so I thought it might be possible to enable them at the stage 1 only so we can have access to the macros. However, the error code registration seems to not work this way. Currently I get the following error:
```
error: used diagnostic code E0660 not registered
--> libsyntax_ext/asm.rs:93:25
|
93 | span_err!(cx, sp, E0660, "malformed inline assembly");
| ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
|
= note: this error originates in a macro outside of the current crate (in Nightly builds, run with -Z external-macro-backtrace for more info)
error: used diagnostic code E0661 not registered
--> libsyntax_ext/asm.rs:151:33
|
151 | / span_err!(cx, sp, E0661,
152 | | "output operand constraint lacks '=' or '+'");
| |________________________________________________________________________________________^
|
= note: this error originates in a macro outside of the current crate (in Nightly builds, run with -Z external-macro-backtrace for more info)
error: aborting due to 2 previous errors
error: Could not compile `syntax_ext`.
```
If anyone has an idea, I'd gladly take it. I'm trying to figure this out on my side as well. I also opened this PR to know if it was worth it to continue (maybe we don't want this?).
Anyway, any answer for both questions is very welcome!
cc @rust-lang/compiler
proc_macro: Avoid cached TokenStream more often
This commit adds even more pessimization to use the cached `TokenStream` inside
of an AST node. As a reminder the `proc_macro` API requires taking an arbitrary
AST node and transforming it back into a `TokenStream` to hand off to a
procedural macro. Such functionality isn't actually implemented in rustc today,
so the way `proc_macro` works today is that it stringifies an AST node and then
reparses for a list of tokens.
This strategy unfortunately loses all span information, so we try to avoid it
whenever possible. Implemented in #43230 some AST nodes have a `TokenStream`
cache representing the tokens they were originally parsed from. This
`TokenStream` cache, however, has turned out to not always reflect the current
state of the item when it's being tokenized. For example `#[cfg]` processing or
macro expansion could modify the state of an item. Consequently we've seen a
number of bugs (#48644 and #49846) related to using this stale cache.
This commit tweaks the usage of the cached `TokenStream` to compare it to our
lossy stringification of the token stream. If the tokens that make up the cache
and the stringified token stream are the same then we return the cached version
(which has correct span information). If they differ, however, then we will
return the stringified version as the cache has been invalidated and we just
haven't figured that out.
Closes#48644Closes#49846
Hygiene 2.0: Avoid comparing fields by name
There are two separate commits here (not counting tests):
- The first one unifies named (`obj.name`) and numeric (`obj.0`) field access expressions in AST and HIR. Before field references in these expressions are resolved it doesn't matter whether the field is named or numeric (it's just a symbol) and 99% of code is common. After field references are resolved we work with
them by index for all fields (see the second commit), so it's again not important whether the field was named or numeric (this includes MIR where all fields were already by index).
(This refactoring actually fixed some bugs in HIR-based borrow checker where borrows through names (`S {
0: ref x }`) and indices (`&s.0`) weren't considered overlapping.)
- The second commit removes all by-name field comparison and instead resolves field references to their indices once, and then uses those resolutions. (There are still a few name comparisons in save-analysis, because save-analysis is weird, but they are made correctly hygienic).
Thus we are fixing a bunch of "secondary" field hygiene bugs (in borrow checker, lints).
Fixes https://github.com/rust-lang/rust/issues/46314
