aboutsummaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorAleksey Kladov <[email protected]>2019-07-22 15:56:19 +0100
committerAleksey Kladov <[email protected]>2019-07-22 15:56:19 +0100
commit700669bbd0ab3ae0c5a56985ce13ca896d342a3a (patch)
tree20ee49ed4ee94e463cd81f3f8142d64cde0ca134
parent75761c0e47d8c20a490a3d61ea64d2413d3c3570 (diff)
kill old lexer
Diffstat
-rw-r--r--crates/ra_syntax/src/parsing/lexer.rs165
-rw-r--r--crates/ra_syntax/src/parsing/lexer/classes.rs26
-rw-r--r--crates/ra_syntax/src/parsing/lexer/comments.rs57
-rw-r--r--crates/ra_syntax/src/parsing/lexer/numbers.rs66
-rw-r--r--crates/ra_syntax/src/parsing/lexer/ptr.rs162
-rw-r--r--crates/ra_syntax/src/parsing/lexer/strings.rs112
6 files changed, 17 insertions, 571 deletions
diff --git a/crates/ra_syntax/src/parsing/lexer.rs b/crates/ra_syntax/src/parsing/lexer.rs
index 1c818fdf4..2a4343b0a 100644
--- a/crates/ra_syntax/src/parsing/lexer.rs
+++ b/crates/ra_syntax/src/parsing/lexer.rs
@@ -1,22 +1,6 @@
1mod classes;
2mod comments;
3mod numbers;
4mod ptr;
5mod strings;
6
7use crate::{ 1use crate::{
8 SyntaxKind::{self, *}, 2 SyntaxKind::{self, *},
9 TextUnit, T, 3 TextUnit,
10};
11
12use self::{
13 classes::*,
14 comments::{scan_comment, scan_shebang},
15 numbers::scan_number,
16 ptr::Ptr,
17 strings::{
18 is_string_literal_start, scan_byte_char_or_string, scan_char, scan_raw_string, scan_string,
19 },
20}; 4};
21 5
22/// A token of Rust source. 6/// A token of Rust source.
@@ -141,138 +125,23 @@ pub fn tokenize(text: &str) -> Vec<Token> {
141 acc 125 acc
142} 126}
143 127
144/// Get the next token from a string
145fn next_token(text: &str) -> Token {
146 assert!(!text.is_empty());
147 let mut ptr = Ptr::new(text);
148 let c = ptr.bump().unwrap();
149 let kind = next_token_inner(c, &mut ptr);
150 let len = ptr.into_len();
151 Token { kind, len }
152}
153
154fn next_token_inner(c: char, ptr: &mut Ptr) -> SyntaxKind {
155 if is_whitespace(c) {
156 ptr.bump_while(is_whitespace);
157 return WHITESPACE;
158 }
159
160 match c {
161 '#' => {
162 if scan_shebang(ptr) {
163 return SHEBANG;
164 }
165 }
166 '/' => {
167 if let Some(kind) = scan_comment(ptr) {
168 return kind;
169 }
170 }
171 _ => (),
172 }
173
174 let ident_start = is_ident_start(c) && !is_string_literal_start(c, ptr.current(), ptr.nth(1));
175 if ident_start {
176 return scan_ident(c, ptr);
177 }
178
179 if is_dec_digit(c) {
180 let kind = scan_number(c, ptr);
181 scan_literal_suffix(ptr);
182 return kind;
183 }
184
185 // One-byte tokens.
186 if let Some(kind) = SyntaxKind::from_char(c) {
187 return kind;
188 }
189
190 match c {
191 // Possiblily multi-byte tokens,
192 // but we only produce single byte token now
193 // T![...], T![..], T![..=], T![.]
194 '.' => return T![.],
195 // T![::] T![:]
196 ':' => return T![:],
197 // T![==] FATARROW T![=]
198 '=' => return T![=],
199 // T![!=] T![!]
200 '!' => return T![!],
201 // T![->] T![-]
202 '-' => return T![-],
203
204 // If the character is an ident start not followed by another single
205 // quote, then this is a lifetime name:
206 '\'' => {
207 return if ptr.at_p(is_ident_start) && !ptr.at_str("''") {
208 ptr.bump();
209 while ptr.at_p(is_ident_continue) {
210 ptr.bump();
211 }
212 // lifetimes shouldn't end with a single quote
213 // if we find one, then this is an invalid character literal
214 if ptr.at('\'') {
215 ptr.bump();
216 return CHAR;
217 }
218 LIFETIME
219 } else {
220 scan_char(ptr);
221 scan_literal_suffix(ptr);
222 CHAR
223 };
224 }
225 'b' => {
226 let kind = scan_byte_char_or_string(ptr);
227 scan_literal_suffix(ptr);
228 return kind;
229 }
230 '"' => {
231 scan_string(ptr);
232 scan_literal_suffix(ptr);
233 return STRING;
234 }
235 'r' => {
236 scan_raw_string(ptr);
237 scan_literal_suffix(ptr);
238 return RAW_STRING;
239 }
240 _ => (),
241 }
242 ERROR
243}
244
245fn scan_ident(c: char, ptr: &mut Ptr) -> SyntaxKind {
246 let is_raw = match (c, ptr.current()) {
247 ('r', Some('#')) => {
248 ptr.bump();
249 true
250 }
251 ('_', None) => return T![_],
252 ('_', Some(c)) if !is_ident_continue(c) => return T![_],
253 _ => false,
254 };
255 ptr.bump_while(is_ident_continue);
256 if !is_raw {
257 if let Some(kind) = SyntaxKind::from_keyword(ptr.current_token_text()) {
258 return kind;
259 }
260 }
261 IDENT
262}
263
264fn scan_literal_suffix(ptr: &mut Ptr) {
265 if ptr.at_p(is_ident_start) {
266 ptr.bump();
267 }
268 ptr.bump_while(is_ident_continue);
269}
270
271pub fn classify_literal(text: &str) -> Option<Token> { 128pub fn classify_literal(text: &str) -> Option<Token> {
272 let tkn = next_token(text); 129 let t = ra_rustc_lexer::first_token(text);
273 if !tkn.kind.is_literal() || tkn.len.to_usize() != text.len() { 130 if t.len != text.len() {
274 return None; 131 return None;
275 } 132 }
276 133 let kind = match t.kind {
277 Some(tkn) 134 ra_rustc_lexer::TokenKind::Literal { kind, .. } => match kind {
135 ra_rustc_lexer::LiteralKind::Int { .. } => INT_NUMBER,
136 ra_rustc_lexer::LiteralKind::Float { .. } => FLOAT_NUMBER,
137 ra_rustc_lexer::LiteralKind::Char { .. } => CHAR,
138 ra_rustc_lexer::LiteralKind::Byte { .. } => BYTE,
139 ra_rustc_lexer::LiteralKind::Str { .. } => STRING,
140 ra_rustc_lexer::LiteralKind::ByteStr { .. } => BYTE_STRING,
141 ra_rustc_lexer::LiteralKind::RawStr { .. } => RAW_STRING,
142 ra_rustc_lexer::LiteralKind::RawByteStr { .. } => RAW_BYTE_STRING,
143 },
144 _ => return None,
145 };
146 Some(Token { kind, len: TextUnit::from_usize(t.len) })
278} 147}
diff --git a/crates/ra_syntax/src/parsing/lexer/classes.rs b/crates/ra_syntax/src/parsing/lexer/classes.rs
deleted file mode 100644
index 4235d2648..000000000
--- a/crates/ra_syntax/src/parsing/lexer/classes.rs
+++ /dev/null
@@ -1,26 +0,0 @@
1use unicode_xid::UnicodeXID;
2
3pub fn is_ident_start(c: char) -> bool {
4 (c >= 'a' && c <= 'z')
5 || (c >= 'A' && c <= 'Z')
6 || c == '_'
7 || (c > '\x7f' && UnicodeXID::is_xid_start(c))
8}
9
10pub fn is_ident_continue(c: char) -> bool {
11 (c >= 'a' && c <= 'z')
12 || (c >= 'A' && c <= 'Z')
13 || (c >= '0' && c <= '9')
14 || c == '_'
15 || (c > '\x7f' && UnicodeXID::is_xid_continue(c))
16}
17
18pub fn is_whitespace(c: char) -> bool {
19 //FIXME: use is_pattern_whitespace
20 //https://github.com/behnam/rust-unic/issues/192
21 c.is_whitespace()
22}
23
24pub fn is_dec_digit(c: char) -> bool {
25 '0' <= c && c <= '9'
26}
diff --git a/crates/ra_syntax/src/parsing/lexer/comments.rs b/crates/ra_syntax/src/parsing/lexer/comments.rs
deleted file mode 100644
index 8bbbe659b..000000000
--- a/crates/ra_syntax/src/parsing/lexer/comments.rs
+++ /dev/null
@@ -1,57 +0,0 @@
1use crate::parsing::lexer::ptr::Ptr;
2
3use crate::SyntaxKind::{self, *};
4
5pub(crate) fn scan_shebang(ptr: &mut Ptr) -> bool {
6 if ptr.at_str("!/") {
7 ptr.bump();
8 ptr.bump();
9 bump_until_eol(ptr);
10 true
11 } else {
12 false
13 }
14}
15
16fn scan_block_comment(ptr: &mut Ptr) -> Option<SyntaxKind> {
17 if ptr.at('*') {
18 ptr.bump();
19 let mut depth: u32 = 1;
20 while depth > 0 {
21 if ptr.at_str("*/") {
22 depth -= 1;
23 ptr.bump();
24 ptr.bump();
25 } else if ptr.at_str("/*") {
26 depth += 1;
27 ptr.bump();
28 ptr.bump();
29 } else if ptr.bump().is_none() {
30 break;
31 }
32 }
33 Some(COMMENT)
34 } else {
35 None
36 }
37}
38
39pub(crate) fn scan_comment(ptr: &mut Ptr) -> Option<SyntaxKind> {
40 if ptr.at('/') {
41 bump_until_eol(ptr);
42 Some(COMMENT)
43 } else {
44 scan_block_comment(ptr)
45 }
46}
47
48fn bump_until_eol(ptr: &mut Ptr) {
49 loop {
50 if ptr.at('\n') || ptr.at_str("\r\n") {
51 return;
52 }
53 if ptr.bump().is_none() {
54 break;
55 }
56 }
57}
diff --git a/crates/ra_syntax/src/parsing/lexer/numbers.rs b/crates/ra_syntax/src/parsing/lexer/numbers.rs
deleted file mode 100644
index e53ae231b..000000000
--- a/crates/ra_syntax/src/parsing/lexer/numbers.rs
+++ /dev/null
@@ -1,66 +0,0 @@
1use crate::parsing::lexer::{classes::*, ptr::Ptr};
2
3use crate::SyntaxKind::{self, *};
4
5pub(crate) fn scan_number(c: char, ptr: &mut Ptr) -> SyntaxKind {
6 if c == '0' {
7 match ptr.current().unwrap_or('\0') {
8 'b' | 'o' => {
9 ptr.bump();
10 scan_digits(ptr, false);
11 }
12 'x' => {
13 ptr.bump();
14 scan_digits(ptr, true);
15 }
16 '0'..='9' | '_' | '.' | 'e' | 'E' => {
17 scan_digits(ptr, true);
18 }
19 _ => return INT_NUMBER,
20 }
21 } else {
22 scan_digits(ptr, false);
23 }
24
25 // might be a float, but don't be greedy if this is actually an
26 // integer literal followed by field/method access or a range pattern
27 // (`0..2` and `12.foo()`)
28 if ptr.at('.') && !(ptr.at_str("..") || ptr.nth_is_p(1, is_ident_start)) {
29 // might have stuff after the ., and if it does, it needs to start
30 // with a number
31 ptr.bump();
32 scan_digits(ptr, false);
33 scan_float_exponent(ptr);
34 return FLOAT_NUMBER;
35 }
36 // it might be a float if it has an exponent
37 if ptr.at('e') || ptr.at('E') {
38 scan_float_exponent(ptr);
39 return FLOAT_NUMBER;
40 }
41 INT_NUMBER
42}
43
44fn scan_digits(ptr: &mut Ptr, allow_hex: bool) {
45 while let Some(c) = ptr.current() {
46 match c {
47 '_' | '0'..='9' => {
48 ptr.bump();
49 }
50 'a'..='f' | 'A'..='F' if allow_hex => {
51 ptr.bump();
52 }
53 _ => return,
54 }
55 }
56}
57
58fn scan_float_exponent(ptr: &mut Ptr) {
59 if ptr.at('e') || ptr.at('E') {
60 ptr.bump();
61 if ptr.at('-') || ptr.at('+') {
62 ptr.bump();
63 }
64 scan_digits(ptr, false);
65 }
66}
diff --git a/crates/ra_syntax/src/parsing/lexer/ptr.rs b/crates/ra_syntax/src/parsing/lexer/ptr.rs
deleted file mode 100644
index c341c4176..000000000
--- a/crates/ra_syntax/src/parsing/lexer/ptr.rs
+++ /dev/null
@@ -1,162 +0,0 @@
1use crate::TextUnit;
2
3use std::str::Chars;
4
5/// A simple view into the characters of a string.
6pub(crate) struct Ptr<'s> {
7 text: &'s str,
8 len: TextUnit,
9}
10
11impl<'s> Ptr<'s> {
12 /// Creates a new `Ptr` from a string.
13 pub fn new(text: &'s str) -> Ptr<'s> {
14 Ptr { text, len: 0.into() }
15 }
16
17 /// Gets the length of the remaining string.
18 pub fn into_len(self) -> TextUnit {
19 self.len
20 }
21
22 /// Gets the current character, if one exists.
23 pub fn current(&self) -> Option<char> {
24 self.chars().next()
25 }
26
27 /// Gets the nth character from the current.
28 /// For example, 0 will return the current character, 1 will return the next, etc.
29 pub fn nth(&self, n: u32) -> Option<char> {
30 self.chars().nth(n as usize)
31 }
32
33 /// Checks whether the current character is `c`.
34 pub fn at(&self, c: char) -> bool {
35 self.current() == Some(c)
36 }
37
38 /// Checks whether the next characters match `s`.
39 pub fn at_str(&self, s: &str) -> bool {
40 let chars = self.chars();
41 chars.as_str().starts_with(s)
42 }
43
44 /// Checks whether the current character satisfies the predicate `p`.
45 pub fn at_p<P: Fn(char) -> bool>(&self, p: P) -> bool {
46 self.current().map(p) == Some(true)
47 }
48
49 /// Checks whether the nth character satisfies the predicate `p`.
50 pub fn nth_is_p<P: Fn(char) -> bool>(&self, n: u32, p: P) -> bool {
51 self.nth(n).map(p) == Some(true)
52 }
53
54 /// Moves to the next character.
55 pub fn bump(&mut self) -> Option<char> {
56 let ch = self.chars().next()?;
57 self.len += TextUnit::of_char(ch);
58 Some(ch)
59 }
60
61 /// Moves to the next character as long as `pred` is satisfied.
62 pub fn bump_while<F: Fn(char) -> bool>(&mut self, pred: F) {
63 loop {
64 match self.current() {
65 Some(c) if pred(c) => {
66 self.bump();
67 }
68 _ => return,
69 }
70 }
71 }
72
73 /// Returns the text up to the current point.
74 pub fn current_token_text(&self) -> &str {
75 let len: u32 = self.len.into();
76 &self.text[..len as usize]
77 }
78
79 /// Returns an iterator over the remaining characters.
80 fn chars(&self) -> Chars {
81 let len: u32 = self.len.into();
82 self.text[len as usize..].chars()
83 }
84}
85
86#[cfg(test)]
87mod tests {
88 use super::*;
89
90 #[test]
91 fn test_current() {
92 let ptr = Ptr::new("test");
93 assert_eq!(ptr.current(), Some('t'));
94 }
95
96 #[test]
97 fn test_nth() {
98 let ptr = Ptr::new("test");
99 assert_eq!(ptr.nth(0), Some('t'));
100 assert_eq!(ptr.nth(1), Some('e'));
101 assert_eq!(ptr.nth(2), Some('s'));
102 assert_eq!(ptr.nth(3), Some('t'));
103 assert_eq!(ptr.nth(4), None);
104 }
105
106 #[test]
107 fn test_at() {
108 let ptr = Ptr::new("test");
109 assert!(ptr.at('t'));
110 assert!(!ptr.at('a'));
111 }
112
113 #[test]
114 fn test_at_str() {
115 let ptr = Ptr::new("test");
116 assert!(ptr.at_str("t"));
117 assert!(ptr.at_str("te"));
118 assert!(ptr.at_str("test"));
119 assert!(!ptr.at_str("tests"));
120 assert!(!ptr.at_str("rust"));
121 }
122
123 #[test]
124 fn test_at_p() {
125 let ptr = Ptr::new("test");
126 assert!(ptr.at_p(|c| c == 't'));
127 assert!(!ptr.at_p(|c| c == 'e'));
128 }
129
130 #[test]
131 fn test_nth_is_p() {
132 let ptr = Ptr::new("test");
133 assert!(ptr.nth_is_p(0, |c| c == 't'));
134 assert!(!ptr.nth_is_p(1, |c| c == 't'));
135 assert!(ptr.nth_is_p(3, |c| c == 't'));
136 assert!(!ptr.nth_is_p(150, |c| c == 't'));
137 }
138
139 #[test]
140 fn test_bump() {
141 let mut ptr = Ptr::new("test");
142 assert_eq!(ptr.current(), Some('t'));
143 ptr.bump();
144 assert_eq!(ptr.current(), Some('e'));
145 ptr.bump();
146 assert_eq!(ptr.current(), Some('s'));
147 ptr.bump();
148 assert_eq!(ptr.current(), Some('t'));
149 ptr.bump();
150 assert_eq!(ptr.current(), None);
151 ptr.bump();
152 assert_eq!(ptr.current(), None);
153 }
154
155 #[test]
156 fn test_bump_while() {
157 let mut ptr = Ptr::new("test");
158 assert_eq!(ptr.current(), Some('t'));
159 ptr.bump_while(|c| c != 's');
160 assert_eq!(ptr.current(), Some('s'));
161 }
162}
diff --git a/crates/ra_syntax/src/parsing/lexer/strings.rs b/crates/ra_syntax/src/parsing/lexer/strings.rs
deleted file mode 100644
index f74acff9e..000000000
--- a/crates/ra_syntax/src/parsing/lexer/strings.rs
+++ /dev/null
@@ -1,112 +0,0 @@
1use crate::{
2 parsing::lexer::ptr::Ptr,
3 SyntaxKind::{self, *},
4};
5
6pub(crate) fn is_string_literal_start(c: char, c1: Option<char>, c2: Option<char>) -> bool {
7 match (c, c1, c2) {
8 ('r', Some('"'), _)
9 | ('r', Some('#'), Some('"'))
10 | ('r', Some('#'), Some('#'))
11 | ('b', Some('"'), _)
12 | ('b', Some('\''), _)
13 | ('b', Some('r'), Some('"'))
14 | ('b', Some('r'), Some('#')) => true,
15 _ => false,
16 }
17}
18
19pub(crate) fn scan_char(ptr: &mut Ptr) {
20 while let Some(c) = ptr.current() {
21 match c {
22 '\\' => {
23 ptr.bump();
24 if ptr.at('\\') || ptr.at('\'') {
25 ptr.bump();
26 }
27 }
28 '\'' => {
29 ptr.bump();
30 return;
31 }
32 '\n' => return,
33 _ => {
34 ptr.bump();
35 }
36 }
37 }
38}
39
40pub(crate) fn scan_byte_char_or_string(ptr: &mut Ptr) -> SyntaxKind {
41 // unwrapping and not-exhaustive match are ok
42 // because of string_literal_start
43 let c = ptr.bump().unwrap();
44 match c {
45 '\'' => {
46 scan_byte(ptr);
47 BYTE
48 }
49 '"' => {
50 scan_byte_string(ptr);
51 BYTE_STRING
52 }
53 'r' => {
54 scan_raw_string(ptr);
55 RAW_BYTE_STRING
56 }
57 _ => unreachable!(),
58 }
59}
60
61pub(crate) fn scan_string(ptr: &mut Ptr) {
62 while let Some(c) = ptr.current() {
63 match c {
64 '\\' => {
65 ptr.bump();
66 if ptr.at('\\') || ptr.at('"') {
67 ptr.bump();
68 }
69 }
70 '"' => {
71 ptr.bump();
72 return;
73 }
74 _ => {
75 ptr.bump();
76 }
77 }
78 }
79}
80
81pub(crate) fn scan_raw_string(ptr: &mut Ptr) {
82 let mut hashes = 0;
83 while ptr.at('#') {
84 hashes += 1;
85 ptr.bump();
86 }
87 if !ptr.at('"') {
88 return;
89 }
90 ptr.bump();
91
92 while let Some(c) = ptr.bump() {
93 if c == '"' {
94 let mut hashes_left = hashes;
95 while ptr.at('#') && hashes_left > 0 {
96 hashes_left -= 1;
97 ptr.bump();
98 }
99 if hashes_left == 0 {
100 return;
101 }
102 }
103 }
104}
105
106fn scan_byte(ptr: &mut Ptr) {
107 scan_char(ptr)
108}
109
110fn scan_byte_string(ptr: &mut Ptr) {
111 scan_string(ptr)
112}
generated by cgit v1.2.3 (git 2.25.1) at 2024-07-08 09:30:02 +0100