Auto merge of #85359 - lrh2000:reserved-prefixes, r=nikomatsakis

Reserve prefixed identifiers and literals (RFC 3101)

This PR denies any identifiers immediately followed by one of three tokens `"`, `'` or `#`, which is stricter than the requirements of RFC 3101 but may be necessary according to the discussion at [Zulip].

[Zulip]: https://rust-lang.zulipchat.com/#narrow/stream/268952-edition-2021/topic/reserved.20prefixes/near/238470099

The tracking issue #84599 says we'll add a feature gate named `reserved_prefixes`, but I don't think I can do this because it is impossible for the lexer to know whether a feature is enabled or not. I guess determining the behavior by the edition information should be enough.

Fixes #84599

Author: bors

compiler/rustc_lexer/src/lib.rs

@@ -66,6 +66,13 @@ pub enum TokenKind {
     Ident,
     /// "r#ident"
     RawIdent,
+    /// An unknown prefix like `foo#`, `foo'`, `foo"`. Note that only the
+    /// prefix (`foo`) is included in the token, not the separator (which is
+    /// lexed as its own distinct token). In Rust 2021 and later, reserved
+    /// prefixes are reported as errors; in earlier editions, they result in a
+    /// (allowed by default) lint, and are treated as regular identifier
+    /// tokens.
+    UnknownPrefix,
     /// "12_u8", "1.0e-40", "b"123"". See `LiteralKind` for more details.
     Literal { kind: LiteralKind, suffix_start: usize },
     /// "'a"
@@ -323,7 +330,7 @@ impl Cursor<'_> {
                     let kind = RawStr { n_hashes, err };
                     Literal { kind, suffix_start }
                 }
-                _ => self.ident(),
+                _ => self.ident_or_unknown_prefix(),
             },
 
             // Byte literal, byte string literal, raw byte string literal or identifier.
@@ -358,12 +365,12 @@ impl Cursor<'_> {
                     let kind = RawByteStr { n_hashes, err };
                     Literal { kind, suffix_start }
                 }
-                _ => self.ident(),
+                _ => self.ident_or_unknown_prefix(),
             },
 
             // Identifier (this should be checked after other variant that can
             // start as identifier).
-            c if is_id_start(c) => self.ident(),
+            c if is_id_start(c) => self.ident_or_unknown_prefix(),
 
             // Numeric literal.
             c @ '0'..='9' => {
@@ -487,11 +494,16 @@ impl Cursor<'_> {
         RawIdent
     }
 
-    fn ident(&mut self) -> TokenKind {
+    fn ident_or_unknown_prefix(&mut self) -> TokenKind {
         debug_assert!(is_id_start(self.prev()));
         // Start is already eaten, eat the rest of identifier.
         self.eat_while(is_id_continue);
-        Ident
+        // Known prefixes must have been handled earlier. So if
+        // we see a prefix here, it is definitely a unknown prefix.
+        match self.first() {
+            '#' | '"' | '\'' => UnknownPrefix,
+            _ => Ident,
+        }
     }
 
     fn number(&mut self, first_digit: char) -> LiteralKind {

compiler/rustc_lint/src/context.rs

@@ -723,6 +723,15 @@ pub trait LintContext: Sized {
                 BuiltinLintDiagnostics::OrPatternsBackCompat(span,suggestion) => {
                     db.span_suggestion(span, "use pat_param to preserve semantics", suggestion, Applicability::MachineApplicable);
                 }
+                BuiltinLintDiagnostics::ReservedPrefix(span) => {
+                    db.span_label(span, "unknown prefix");
+                    db.span_suggestion_verbose(
+                        span.shrink_to_hi(),
+                        "insert whitespace here to avoid this being parsed as a prefix in Rust 2021",
+                        " ".into(),
+                        Applicability::MachineApplicable,
+                    );
+                }
             }
             // Rewrap `db`, and pass control to the user.
             decorate(LintDiagnosticBuilder::new(db));

compiler/rustc_lint_defs/src/builtin.rs

@@ -2973,6 +2973,7 @@ declare_lint_pass! {
         OR_PATTERNS_BACK_COMPAT,
         LARGE_ASSIGNMENTS,
         FUTURE_PRELUDE_COLLISION,
+        RESERVED_PREFIX,
     ]
 }
 
@@ -3263,3 +3264,39 @@ declare_lint! {
         reason: FutureIncompatibilityReason::EditionError(Edition::Edition2021),
     };
 }
+
+declare_lint! {
+    /// The `reserved_prefix` lint detects identifiers that will be parsed as a
+    /// prefix instead in Rust 2021.
+    ///
+    /// ### Example
+    ///
+    /// ```rust,compile_fail
+    /// #![deny(reserved_prefix)]
+    ///
+    /// macro_rules! m {
+    ///     (z $x:expr) => ();
+    /// }
+    ///
+    /// m!(z"hey");
+    /// ```
+    ///
+    /// {{produces}}
+    ///
+    /// ### Explanation
+    ///
+    /// In Rust 2015 and 2018, `z"hey"` is two tokens: the identifier `z`
+    /// followed by the string literal `"hey"`. In Rust 2021, the `z` is
+    /// considered a prefix for `"hey"`.
+    ///
+    /// This lint suggests to add whitespace between the `z` and `"hey"` tokens
+    /// to keep them separated in Rust 2021.
+    pub RESERVED_PREFIX,
+    Allow,
+    "identifiers that will be parsed as a prefix in Rust 2021",
+    @future_incompatible = FutureIncompatibleInfo {
+        reference: "issue #84978 <https://github.com/rust-lang/rust/issues/84978>",
+        reason: FutureIncompatibilityReason::EditionError(Edition::Edition2021),
+    };
+    crate_level_only
+}

compiler/rustc_lint_defs/src/lib.rs

@@ -300,6 +300,7 @@ pub enum BuiltinLintDiagnostics {
     ExternDepSpec(String, ExternDepSpec),
     ProcMacroBackCompat(String),
     OrPatternsBackCompat(Span, String),
+    ReservedPrefix(Span),
 }
 
 /// Lints that are buffered up early on in the `Session` before the

compiler/rustc_parse/src/lexer/mod.rs

@@ -1,12 +1,14 @@
-use rustc_ast::ast::AttrStyle;
+use rustc_ast::ast::{self, AttrStyle};
 use rustc_ast::token::{self, CommentKind, Token, TokenKind};
 use rustc_ast::tokenstream::{Spacing, TokenStream};
 use rustc_errors::{error_code, Applicability, DiagnosticBuilder, FatalError, PResult};
 use rustc_lexer::unescape::{self, Mode};
 use rustc_lexer::{Base, DocStyle, RawStrError};
+use rustc_session::lint::builtin::RESERVED_PREFIX;
+use rustc_session::lint::BuiltinLintDiagnostics;
 use rustc_session::parse::ParseSess;
 use rustc_span::symbol::{sym, Symbol};
-use rustc_span::{BytePos, Pos, Span};
+use rustc_span::{edition::Edition, BytePos, Pos, Span};
 
 use tracing::debug;
 
@@ -166,12 +168,18 @@ impl<'a> StringReader<'a> {
                 self.cook_doc_comment(content_start, content, CommentKind::Block, doc_style)
             }
             rustc_lexer::TokenKind::Whitespace => return None,
-            rustc_lexer::TokenKind::Ident | rustc_lexer::TokenKind::RawIdent => {
+            rustc_lexer::TokenKind::Ident
+            | rustc_lexer::TokenKind::RawIdent
+            | rustc_lexer::TokenKind::UnknownPrefix => {
                 let is_raw_ident = token == rustc_lexer::TokenKind::RawIdent;
+                let is_unknown_prefix = token == rustc_lexer::TokenKind::UnknownPrefix;
                 let mut ident_start = start;
                 if is_raw_ident {
                     ident_start = ident_start + BytePos(2);
                 }
+                if is_unknown_prefix {
+                    self.report_unknown_prefix(start);
+                }
                 let sym = nfc_normalize(self.str_from(ident_start));
                 let span = self.mk_sp(start, self.pos);
                 self.sess.symbol_gallery.insert(sym, span);
@@ -491,6 +499,42 @@ impl<'a> StringReader<'a> {
         FatalError.raise()
     }
 
+    // RFC 3101 introduced the idea of (reserved) prefixes. As of Rust 2021,
+    // using a (unknown) prefix is an error. In earlier editions, however, they
+    // only result in a (allowed by default) lint, and are treated as regular
+    // identifier tokens.
+    fn report_unknown_prefix(&self, start: BytePos) {
+        let prefix_span = self.mk_sp(start, self.pos);
+        let msg = format!("prefix `{}` is unknown", self.str_from_to(start, self.pos));
+
+        let expn_data = prefix_span.ctxt().outer_expn_data();
+
+        if expn_data.edition >= Edition::Edition2021 {
+            // In Rust 2021, this is a hard error.
+            let mut err = self.sess.span_diagnostic.struct_span_err(prefix_span, &msg);
+            err.span_label(prefix_span, "unknown prefix");
+            if expn_data.is_root() {
+                err.span_suggestion_verbose(
+                    prefix_span.shrink_to_hi(),
+                    "consider inserting whitespace here",
+                    " ".into(),
+                    Applicability::MachineApplicable,
+                );
+            }
+            err.note("prefixed identifiers and literals are reserved since Rust 2021");
+            err.emit();
+        } else {
+            // Before Rust 2021, only emit a lint for migration.
+            self.sess.buffer_lint_with_diagnostic(
+                &RESERVED_PREFIX,
+                prefix_span,
+                ast::CRATE_NODE_ID,
+                &msg,
+                BuiltinLintDiagnostics::ReservedPrefix(prefix_span),
+            );
+        }
+    }
+
     /// Note: It was decided to not add a test case, because it would be too big.
     /// <https://github.com/rust-lang/rust/pull/50296#issuecomment-392135180>
     fn report_too_many_hashes(&self, start: BytePos, found: usize) -> ! {

src/librustdoc/html/highlight.rs

@@ -413,7 +413,7 @@ impl<'a> Classifier<'a> {
                 },
                 c => c,
             },
-            TokenKind::RawIdent => Class::Ident,
+            TokenKind::RawIdent | TokenKind::UnknownPrefix => Class::Ident,
             TokenKind::Lifetime { .. } => Class::Lifetime,
         };
         // Anything that didn't return above is the simple case where we the

src/test/ui/rust-2021/auxiliary/reserved-prefixes-macro-2018.rs

@@ -0,0 +1,25 @@
+// force-host
+// edition:2018
+// no-prefer-dynamic
+
+#![crate_type = "proc-macro"]
+
+extern crate proc_macro;
+
+use proc_macro::TokenStream;
+use std::str::FromStr;
+
+#[proc_macro]
+pub fn number_of_tokens_in_a_prefixed_integer_literal(_: TokenStream) -> TokenStream {
+    TokenStream::from_str("hey#123").unwrap().into_iter().count().to_string().parse().unwrap()
+}
+
+#[proc_macro]
+pub fn number_of_tokens_in_a_prefixed_char_literal(_: TokenStream) -> TokenStream {
+    TokenStream::from_str("hey#'a'").unwrap().into_iter().count().to_string().parse().unwrap()
+}
+
+#[proc_macro]
+pub fn number_of_tokens_in_a_prefixed_string_literal(_: TokenStream) -> TokenStream {
+    TokenStream::from_str("hey#\"abc\"").unwrap().into_iter().count().to_string().parse().unwrap()
+}

src/test/ui/rust-2021/auxiliary/reserved-prefixes-macro-2021.rs

@@ -0,0 +1,25 @@
+// force-host
+// edition:2021
+// no-prefer-dynamic
+
+#![crate_type = "proc-macro"]
+
+extern crate proc_macro;
+
+use proc_macro::TokenStream;
+use std::str::FromStr;
+
+#[proc_macro]
+pub fn number_of_tokens_in_a_prefixed_integer_literal(_: TokenStream) -> TokenStream {
+    TokenStream::from_str("hey#123").unwrap().into_iter().count().to_string().parse().unwrap()
+}
+
+#[proc_macro]
+pub fn number_of_tokens_in_a_prefixed_char_literal(_: TokenStream) -> TokenStream {
+    TokenStream::from_str("hey#'a'").unwrap().into_iter().count().to_string().parse().unwrap()
+}
+
+#[proc_macro]
+pub fn number_of_tokens_in_a_prefixed_string_literal(_: TokenStream) -> TokenStream {
+    TokenStream::from_str("hey#\"abc\"").unwrap().into_iter().count().to_string().parse().unwrap()
+}

src/test/ui/rust-2021/reserved-prefixes-migration.fixed

@@ -0,0 +1,38 @@
+// check-pass
+// run-rustfix
+// compile-flags: -Z unstable-options --edition 2018
+
+#![warn(reserved_prefix)]
+
+macro_rules! m2 {
+    ($a:tt $b:tt) => {};
+}
+
+macro_rules! m3 {
+    ($a:tt $b:tt $c:tt) => {};
+}
+
+fn main() {
+    m2!(z "hey");
+    //~^ WARNING prefix `z` is unknown [reserved_prefix]
+    //~| WARNING hard error in Rust 2021
+    m2!(prefix "hey");
+    //~^ WARNING prefix `prefix` is unknown [reserved_prefix]
+    //~| WARNING hard error in Rust 2021
+    m3!(hey #123);
+    //~^ WARNING prefix `hey` is unknown [reserved_prefix]
+    //~| WARNING hard error in Rust 2021
+    m3!(hey #hey);
+    //~^ WARNING prefix `hey` is unknown [reserved_prefix]
+    //~| WARNING hard error in Rust 2021
+}
+
+macro_rules! quote {
+    (# name = # kind # value) => {};
+}
+
+quote! {
+    #name = #kind #value
+    //~^ WARNING prefix `kind` is unknown [reserved_prefix]
+    //~| WARNING hard error in Rust 2021
+}

src/test/ui/rust-2021/reserved-prefixes-migration.rs

@@ -0,0 +1,38 @@
+// check-pass
+// run-rustfix
+// compile-flags: -Z unstable-options --edition 2018
+
+#![warn(reserved_prefix)]
+
+macro_rules! m2 {
+    ($a:tt $b:tt) => {};
+}
+
+macro_rules! m3 {
+    ($a:tt $b:tt $c:tt) => {};
+}
+
+fn main() {
+    m2!(z"hey");
+    //~^ WARNING prefix `z` is unknown [reserved_prefix]
+    //~| WARNING hard error in Rust 2021
+    m2!(prefix"hey");
+    //~^ WARNING prefix `prefix` is unknown [reserved_prefix]
+    //~| WARNING hard error in Rust 2021
+    m3!(hey#123);
+    //~^ WARNING prefix `hey` is unknown [reserved_prefix]
+    //~| WARNING hard error in Rust 2021
+    m3!(hey#hey);
+    //~^ WARNING prefix `hey` is unknown [reserved_prefix]
+    //~| WARNING hard error in Rust 2021
+}
+
+macro_rules! quote {
+    (# name = # kind # value) => {};
+}
+
+quote! {
+    #name = #kind#value
+    //~^ WARNING prefix `kind` is unknown [reserved_prefix]
+    //~| WARNING hard error in Rust 2021
+}