refactor: improve "ident starts with number" error

Author: Ezrashaw

compiler/rustc_parse/src/errors.rs

@@ -986,7 +986,10 @@ impl<'a, G: EmissionGuarantee> IntoDiagnostic<'a, G> for ExpectedIdentifier {
 
 #[derive(Subdiagnostic)]
 #[help(parse_invalid_identifier_with_leading_number)]
-pub(crate) struct HelpIdentifierStartsWithNumber;
+pub(crate) struct HelpIdentifierStartsWithNumber {
+    #[primary_span]
+    pub num_span: Span,
+}
 
 pub(crate) struct ExpectedSemi {
     pub span: Span,

compiler/rustc_parse/src/parser/diagnostics.rs

@@ -38,7 +38,7 @@ use rustc_errors::{
 use rustc_session::errors::ExprParenthesesNeeded;
 use rustc_span::source_map::Spanned;
 use rustc_span::symbol::{kw, sym, Ident};
-use rustc_span::{Span, SpanSnippetError, DUMMY_SP};
+use rustc_span::{Span, SpanSnippetError, Symbol, DUMMY_SP};
 use std::mem::take;
 use std::ops::{Deref, DerefMut};
 use thin_vec::{thin_vec, ThinVec};
@@ -309,8 +309,11 @@ impl<'a> Parser<'a> {
             && self.look_ahead(1, |t| t.is_ident()))
         .then_some(SuggRemoveComma { span: self.token.span });
 
-        let help_cannot_start_number =
-            self.is_lit_bad_ident().then_some(HelpIdentifierStartsWithNumber);
+        let help_cannot_start_number = self.is_lit_bad_ident().map(|(len, _valid_portion)| {
+            let (invalid, _valid) = self.token.span.split_at(len as u32);
+
+            HelpIdentifierStartsWithNumber { num_span: invalid }
+        });
 
         let err = ExpectedIdentifier {
             span: self.token.span,
@@ -378,13 +381,24 @@ impl<'a> Parser<'a> {
 
     /// Checks if the current token is a integer or float literal and looks like
     /// it could be a invalid identifier with digits at the start.
-    pub(super) fn is_lit_bad_ident(&mut self) -> bool {
-        matches!(self.token.uninterpolate().kind, token::Literal(Lit { kind: token::LitKind::Integer | token::LitKind::Float, .. })
-            // ensure that the integer literal is followed by a *invalid*
-            // suffix: this is how we know that it is a identifier with an
-            // invalid beginning.
-            if rustc_ast::MetaItemLit::from_token(&self.token).is_none()
-        )
+    ///
+    /// Returns the number of characters (bytes) composing the invalid portion
+    /// of the identifier and the valid portion of the identifier.
+    pub(super) fn is_lit_bad_ident(&mut self) -> Option<(usize, Symbol)> {
+        // ensure that the integer literal is followed by a *invalid*
+        // suffix: this is how we know that it is a identifier with an
+        // invalid beginning.
+        if let token::Literal(Lit {
+            kind: token::LitKind::Integer | token::LitKind::Float,
+            symbol,
+            suffix,
+        }) = self.token.uninterpolate().kind
+            && rustc_ast::MetaItemLit::from_token(&self.token).is_none()
+        {
+            Some((symbol.as_str().len(), suffix.unwrap()))
+        } else {
+            None
+        }
     }
 
     pub(super) fn expected_one_of_not_found(

compiler/rustc_parse/src/parser/pat.rs

@@ -348,10 +348,6 @@ impl<'a> Parser<'a> {
             lo = self.token.span;
         }
 
-        if self.is_lit_bad_ident() {
-            return Err(self.expected_ident_found());
-        }
-
         let pat = if self.check(&token::BinOp(token::And)) || self.token.kind == token::AndAnd {
             self.parse_pat_deref(expected)?
         } else if self.check(&token::OpenDelim(Delimiter::Parenthesis)) {
@@ -395,7 +391,7 @@ impl<'a> Parser<'a> {
             } else {
                 PatKind::Lit(const_expr)
             }
-        } else if self.can_be_ident_pat() {
+        } else if self.can_be_ident_pat() || self.is_lit_bad_ident().is_some() {
             // Parse `ident @ pat`
             // This can give false positives and parse nullary enums,
             // they are dealt with later in resolve.

compiler/rustc_span/src/lib.rs

@@ -795,6 +795,18 @@ impl Span {
         })
     }
 
+    /// Splits a span into two composite spans around a certain position.
+    pub fn split_at(self, pos: u32) -> (Span, Span) {
+        let len = self.hi().0 - self.lo().0;
+        debug_assert!(pos <= len);
+
+        let split_pos = BytePos(self.lo().0 + pos);
+        (
+            Span::new(self.lo(), split_pos, self.ctxt(), self.parent()),
+            Span::new(split_pos, self.hi(), self.ctxt(), self.parent()),
+        )
+    }
+
     /// Returns a `Span` that would enclose both `self` and `end`.
     ///
     /// Note that this can also be used to extend the span "backwards":

tests/ui/parser/integer-literal-start-ident.stderr

@@ -4,7 +4,11 @@ error: expected identifier, found `1main`
 LL | fn 1main() {}
    |    ^^^^^ expected identifier
    |
-   = help: identifiers cannot start with a number
+help: identifiers cannot start with a number
+  --> $DIR/integer-literal-start-ident.rs:1:4
+   |
+LL | fn 1main() {}
+   |    ^
 
 error: aborting due to previous error
 

tests/ui/parser/issues/issue-104088.stderr

@@ -4,23 +4,35 @@ error: expected identifier, found `1x`
 LL |     let 1x = 123;
    |         ^^ expected identifier
    |
-   = help: identifiers cannot start with a number
+help: identifiers cannot start with a number
+  --> $DIR/issue-104088.rs:6:9
+   |
+LL |     let 1x = 123;
+   |         ^
 
 error: expected identifier, found `2x`
   --> $DIR/issue-104088.rs:11:9
    |
 LL |     let 2x: i32 = 123;
    |         ^^ expected identifier
    |
-   = help: identifiers cannot start with a number
+help: identifiers cannot start with a number
+  --> $DIR/issue-104088.rs:11:9
+   |
+LL |     let 2x: i32 = 123;
+   |         ^
 
 error: expected identifier, found `23name`
   --> $DIR/issue-104088.rs:22:9
    |
 LL |     let 23name = 123;
    |         ^^^^^^ expected identifier
    |
-   = help: identifiers cannot start with a number
+help: identifiers cannot start with a number
+  --> $DIR/issue-104088.rs:22:9
+   |
+LL |     let 23name = 123;
+   |         ^^
 
 error[E0308]: mismatched types
   --> $DIR/issue-104088.rs:16:12