diff options
-rw-r--r-- | crates/mbe/src/subtree_source.rs | 22 | ||||
-rw-r--r-- | crates/syntax/src/parsing/lexer.rs | 23 |
2 files changed, 23 insertions, 22 deletions
diff --git a/crates/mbe/src/subtree_source.rs b/crates/mbe/src/subtree_source.rs index 226dc3bec..396ce8b16 100644 --- a/crates/mbe/src/subtree_source.rs +++ b/crates/mbe/src/subtree_source.rs | |||
@@ -2,7 +2,7 @@ | |||
2 | 2 | ||
3 | use parser::{Token, TokenSource}; | 3 | use parser::{Token, TokenSource}; |
4 | use std::cell::{Cell, Ref, RefCell}; | 4 | use std::cell::{Cell, Ref, RefCell}; |
5 | use syntax::{tokenize, SmolStr, SyntaxKind, SyntaxKind::*, T}; | 5 | use syntax::{lex_single_syntax_kind, SmolStr, SyntaxKind, SyntaxKind::*, T}; |
6 | use tt::buffer::{Cursor, TokenBuffer}; | 6 | use tt::buffer::{Cursor, TokenBuffer}; |
7 | 7 | ||
8 | #[derive(Debug, Clone, Eq, PartialEq)] | 8 | #[derive(Debug, Clone, Eq, PartialEq)] |
@@ -155,17 +155,15 @@ fn convert_delim(d: Option<tt::DelimiterKind>, closing: bool) -> TtToken { | |||
155 | } | 155 | } |
156 | 156 | ||
157 | fn convert_literal(l: &tt::Literal) -> TtToken { | 157 | fn convert_literal(l: &tt::Literal) -> TtToken { |
158 | let mut kinds = tokenize(&l.text).0.into_iter().map(|token| token.kind); | 158 | let is_negated = l.text.starts_with('-'); |
159 | 159 | let inner_text = &l.text[if is_negated { 1 } else { 0 }..]; | |
160 | let kind = match kinds.next() { | 160 | |
161 | Some(kind) if kind.is_literal() => Some(kind), | 161 | let kind = lex_single_syntax_kind(inner_text) |
162 | Some(SyntaxKind::MINUS) => match kinds.next() { | 162 | .map(|(kind, _error)| kind) |
163 | Some(kind) if kind.is_literal() => Some(kind), | 163 | .filter(|kind| { |
164 | _ => None, | 164 | kind.is_literal() && (!is_negated || matches!(kind, FLOAT_NUMBER | INT_NUMBER)) |
165 | }, | 165 | }) |
166 | _ => None, | 166 | .unwrap_or_else(|| panic!("Fail to convert given literal {:#?}", &l)); |
167 | } | ||
168 | .unwrap_or_else(|| panic!("Fail to convert given literal {:#?}", &l)); | ||
169 | 167 | ||
170 | TtToken { kind, is_joint_to_next: false, text: l.text.clone() } | 168 | TtToken { kind, is_joint_to_next: false, text: l.text.clone() } |
171 | } | 169 | } |
diff --git a/crates/syntax/src/parsing/lexer.rs b/crates/syntax/src/parsing/lexer.rs index f1202113b..7e38c32cc 100644 --- a/crates/syntax/src/parsing/lexer.rs +++ b/crates/syntax/src/parsing/lexer.rs | |||
@@ -1,10 +1,10 @@ | |||
1 | //! Lexer analyzes raw input string and produces lexemes (tokens). | 1 | //! Lexer analyzes raw input string and produces lexemes (tokens). |
2 | //! It is just a bridge to `rustc_lexer`. | 2 | //! It is just a bridge to `rustc_lexer`. |
3 | 3 | ||
4 | use rustc_lexer::{LiteralKind as LK, RawStrError}; | ||
5 | |||
6 | use std::convert::TryInto; | 4 | use std::convert::TryInto; |
7 | 5 | ||
6 | use rustc_lexer::{LiteralKind as LK, RawStrError}; | ||
7 | |||
8 | use crate::{ | 8 | use crate::{ |
9 | SyntaxError, | 9 | SyntaxError, |
10 | SyntaxKind::{self, *}, | 10 | SyntaxKind::{self, *}, |
@@ -61,17 +61,18 @@ pub fn tokenize(text: &str) -> (Vec<Token>, Vec<SyntaxError>) { | |||
61 | (tokens, errors) | 61 | (tokens, errors) |
62 | } | 62 | } |
63 | 63 | ||
64 | /// Returns `SyntaxKind` and `Option<SyntaxError>` of the first token | 64 | /// Returns `SyntaxKind` and `Option<SyntaxError>` if `text` parses as a single token. |
65 | /// encountered at the beginning of the string. | ||
66 | /// | 65 | /// |
67 | /// Returns `None` if the string contains zero *or two or more* tokens. | 66 | /// Returns `None` if the string contains zero *or two or more* tokens. |
68 | /// The token is malformed if the returned error is not `None`. | 67 | /// The token is malformed if the returned error is not `None`. |
69 | /// | 68 | /// |
70 | /// Beware that unescape errors are not checked at tokenization time. | 69 | /// Beware that unescape errors are not checked at tokenization time. |
71 | pub fn lex_single_syntax_kind(text: &str) -> Option<(SyntaxKind, Option<SyntaxError>)> { | 70 | pub fn lex_single_syntax_kind(text: &str) -> Option<(SyntaxKind, Option<SyntaxError>)> { |
72 | lex_first_token(text) | 71 | let (first_token, err) = lex_first_token(text)?; |
73 | .filter(|(token, _)| token.len == TextSize::of(text)) | 72 | if first_token.len != TextSize::of(text) { |
74 | .map(|(token, error)| (token.kind, error)) | 73 | return None; |
74 | } | ||
75 | Some((first_token.kind, err)) | ||
75 | } | 76 | } |
76 | 77 | ||
77 | /// The same as `lex_single_syntax_kind()` but returns only `SyntaxKind` and | 78 | /// The same as `lex_single_syntax_kind()` but returns only `SyntaxKind` and |
@@ -79,9 +80,11 @@ pub fn lex_single_syntax_kind(text: &str) -> Option<(SyntaxKind, Option<SyntaxEr | |||
79 | /// | 80 | /// |
80 | /// Beware that unescape errors are not checked at tokenization time. | 81 | /// Beware that unescape errors are not checked at tokenization time. |
81 | pub fn lex_single_valid_syntax_kind(text: &str) -> Option<SyntaxKind> { | 82 | pub fn lex_single_valid_syntax_kind(text: &str) -> Option<SyntaxKind> { |
82 | lex_first_token(text) | 83 | let (single_token, err) = lex_single_syntax_kind(text)?; |
83 | .filter(|(token, error)| !error.is_some() && token.len == TextSize::of(text)) | 84 | if err.is_some() { |
84 | .map(|(token, _error)| token.kind) | 85 | return None; |
86 | } | ||
87 | Some(single_token) | ||
85 | } | 88 | } |
86 | 89 | ||
87 | /// Returns `SyntaxKind` and `Option<SyntaxError>` of the first token | 90 | /// Returns `SyntaxKind` and `Option<SyntaxError>` of the first token |