aboutsummaryrefslogtreecommitdiff
path: root/src/lexer
diff options
context:
space:
mode:
Diffstat (limited to 'src/lexer')
-rw-r--r--src/lexer/classes.rs26
-rw-r--r--src/lexer/comments.rs57
-rw-r--r--src/lexer/mod.rs209
-rw-r--r--src/lexer/numbers.rs67
-rw-r--r--src/lexer/ptr.rs74
-rw-r--r--src/lexer/strings.rs106
6 files changed, 0 insertions, 539 deletions
diff --git a/src/lexer/classes.rs b/src/lexer/classes.rs
deleted file mode 100644
index 4235d2648..000000000
--- a/src/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/src/lexer/comments.rs b/src/lexer/comments.rs
deleted file mode 100644
index 01acb6515..000000000
--- a/src/lexer/comments.rs
+++ /dev/null
@@ -1,57 +0,0 @@
1use lexer::ptr::Ptr;
2
3use SyntaxKind::{self, *};
4
5pub(crate) fn scan_shebang(ptr: &mut Ptr) -> bool {
6 if ptr.next_is('!') && ptr.nnext_is('/') {
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.next_is('*') {
18 ptr.bump();
19 let mut depth: u32 = 1;
20 while depth > 0 {
21 if ptr.next_is('*') && ptr.nnext_is('/') {
22 depth -= 1;
23 ptr.bump();
24 ptr.bump();
25 } else if ptr.next_is('/') && ptr.nnext_is('*') {
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.next_is('/') {
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.next_is('\n') || ptr.next_is('\r') && ptr.nnext_is('\n') {
51 return;
52 }
53 if ptr.bump().is_none() {
54 break;
55 }
56 }
57}
diff --git a/src/lexer/mod.rs b/src/lexer/mod.rs
deleted file mode 100644
index f8fdc41ac..000000000
--- a/src/lexer/mod.rs
+++ /dev/null
@@ -1,209 +0,0 @@
1mod classes;
2mod comments;
3mod numbers;
4mod ptr;
5mod strings;
6
7use {
8 SyntaxKind::{self, *},
9 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};
21
22/// A token of Rust source.
23#[derive(Debug, Clone, Copy, PartialEq, Eq, PartialOrd, Ord, Hash)]
24pub struct Token {
25 /// The kind of token.
26 pub kind: SyntaxKind,
27 /// The length of the token.
28 pub len: TextUnit,
29}
30
31/// Break a string up into its component tokens
32pub fn tokenize(text: &str) -> Vec<Token> {
33 let mut text = text;
34 let mut acc = Vec::new();
35 while !text.is_empty() {
36 let token = next_token(text);
37 acc.push(token);
38 let len: u32 = token.len.into();
39 text = &text[len as usize..];
40 }
41 acc
42}
43
44/// Get the next token from a string
45pub fn next_token(text: &str) -> Token {
46 assert!(!text.is_empty());
47 let mut ptr = Ptr::new(text);
48 let c = ptr.bump().unwrap();
49 let kind = next_token_inner(c, &mut ptr);
50 let len = ptr.into_len();
51 Token { kind, len }
52}
53
54fn next_token_inner(c: char, ptr: &mut Ptr) -> SyntaxKind {
55 if is_whitespace(c) {
56 ptr.bump_while(is_whitespace);
57 return WHITESPACE;
58 }
59
60 match c {
61 '#' => if scan_shebang(ptr) {
62 return SHEBANG;
63 },
64 '/' => if let Some(kind) = scan_comment(ptr) {
65 return kind;
66 },
67 _ => (),
68 }
69
70 let ident_start = is_ident_start(c) && !is_string_literal_start(c, ptr.next(), ptr.nnext());
71 if ident_start {
72 return scan_ident(c, ptr);
73 }
74
75 if is_dec_digit(c) {
76 let kind = scan_number(c, ptr);
77 scan_literal_suffix(ptr);
78 return kind;
79 }
80
81 // One-byte tokens.
82 if let Some(kind) = SyntaxKind::from_char(c) {
83 return kind;
84 }
85
86 match c {
87 // Multi-byte tokens.
88 '.' => {
89 return match (ptr.next(), ptr.nnext()) {
90 (Some('.'), Some('.')) => {
91 ptr.bump();
92 ptr.bump();
93 DOTDOTDOT
94 }
95 (Some('.'), Some('=')) => {
96 ptr.bump();
97 ptr.bump();
98 DOTDOTEQ
99 }
100 (Some('.'), _) => {
101 ptr.bump();
102 DOTDOT
103 }
104 _ => DOT,
105 };
106 }
107 ':' => {
108 return match ptr.next() {
109 Some(':') => {
110 ptr.bump();
111 COLONCOLON
112 }
113 _ => COLON,
114 };
115 }
116 '=' => {
117 return match ptr.next() {
118 Some('=') => {
119 ptr.bump();
120 EQEQ
121 }
122 Some('>') => {
123 ptr.bump();
124 FAT_ARROW
125 }
126 _ => EQ,
127 };
128 }
129 '!' => {
130 return match ptr.next() {
131 Some('=') => {
132 ptr.bump();
133 NEQ
134 }
135 _ => EXCL,
136 };
137 }
138 '-' => {
139 return if ptr.next_is('>') {
140 ptr.bump();
141 THIN_ARROW
142 } else {
143 MINUS
144 };
145 }
146
147 // If the character is an ident start not followed by another single
148 // quote, then this is a lifetime name:
149 '\'' => {
150 return if ptr.next_is_p(is_ident_start) && !ptr.nnext_is('\'') {
151 ptr.bump();
152 while ptr.next_is_p(is_ident_continue) {
153 ptr.bump();
154 }
155 // lifetimes shouldn't end with a single quote
156 // if we find one, then this is an invalid character literal
157 if ptr.next_is('\'') {
158 ptr.bump();
159 return CHAR; // TODO: error reporting
160 }
161 LIFETIME
162 } else {
163 scan_char(ptr);
164 scan_literal_suffix(ptr);
165 CHAR
166 };
167 }
168 'b' => {
169 let kind = scan_byte_char_or_string(ptr);
170 scan_literal_suffix(ptr);
171 return kind;
172 }
173 '"' => {
174 scan_string(ptr);
175 scan_literal_suffix(ptr);
176 return STRING;
177 }
178 'r' => {
179 scan_raw_string(ptr);
180 scan_literal_suffix(ptr);
181 return RAW_STRING;
182 }
183 _ => (),
184 }
185 ERROR
186}
187
188fn scan_ident(c: char, ptr: &mut Ptr) -> SyntaxKind {
189 let is_single_letter = match ptr.next() {
190 None => true,
191 Some(c) if !is_ident_continue(c) => true,
192 _ => false,
193 };
194 if is_single_letter {
195 return if c == '_' { UNDERSCORE } else { IDENT };
196 }
197 ptr.bump_while(is_ident_continue);
198 if let Some(kind) = SyntaxKind::from_keyword(ptr.current_token_text()) {
199 return kind;
200 }
201 IDENT
202}
203
204fn scan_literal_suffix(ptr: &mut Ptr) {
205 if ptr.next_is_p(is_ident_start) {
206 ptr.bump();
207 }
208 ptr.bump_while(is_ident_continue);
209}
diff --git a/src/lexer/numbers.rs b/src/lexer/numbers.rs
deleted file mode 100644
index 5c4641a2d..000000000
--- a/src/lexer/numbers.rs
+++ /dev/null
@@ -1,67 +0,0 @@
1use lexer::classes::*;
2use lexer::ptr::Ptr;
3
4use SyntaxKind::{self, *};
5
6pub(crate) fn scan_number(c: char, ptr: &mut Ptr) -> SyntaxKind {
7 if c == '0' {
8 match ptr.next().unwrap_or('\0') {
9 'b' | 'o' => {
10 ptr.bump();
11 scan_digits(ptr, false);
12 }
13 'x' => {
14 ptr.bump();
15 scan_digits(ptr, true);
16 }
17 '0'...'9' | '_' | '.' | 'e' | 'E' => {
18 scan_digits(ptr, true);
19 }
20 _ => return INT_NUMBER,
21 }
22 } else {
23 scan_digits(ptr, false);
24 }
25
26 // might be a float, but don't be greedy if this is actually an
27 // integer literal followed by field/method access or a range pattern
28 // (`0..2` and `12.foo()`)
29 if ptr.next_is('.') && !(ptr.nnext_is('.') || ptr.nnext_is_p(is_ident_start)) {
30 // might have stuff after the ., and if it does, it needs to start
31 // with a number
32 ptr.bump();
33 scan_digits(ptr, false);
34 scan_float_exponent(ptr);
35 return FLOAT_NUMBER;
36 }
37 // it might be a float if it has an exponent
38 if ptr.next_is('e') || ptr.next_is('E') {
39 scan_float_exponent(ptr);
40 return FLOAT_NUMBER;
41 }
42 INT_NUMBER
43}
44
45fn scan_digits(ptr: &mut Ptr, allow_hex: bool) {
46 while let Some(c) = ptr.next() {
47 match c {
48 '_' | '0'...'9' => {
49 ptr.bump();
50 }
51 'a'...'f' | 'A'...'F' if allow_hex => {
52 ptr.bump();
53 }
54 _ => return,
55 }
56 }
57}
58
59fn scan_float_exponent(ptr: &mut Ptr) {
60 if ptr.next_is('e') || ptr.next_is('E') {
61 ptr.bump();
62 if ptr.next_is('-') || ptr.next_is('+') {
63 ptr.bump();
64 }
65 scan_digits(ptr, false);
66 }
67}
diff --git a/src/lexer/ptr.rs b/src/lexer/ptr.rs
deleted file mode 100644
index d1391fd5f..000000000
--- a/src/lexer/ptr.rs
+++ /dev/null
@@ -1,74 +0,0 @@
1use TextUnit;
2
3use std::str::Chars;
4
5pub(crate) struct Ptr<'s> {
6 text: &'s str,
7 len: TextUnit,
8}
9
10impl<'s> Ptr<'s> {
11 pub fn new(text: &'s str) -> Ptr<'s> {
12 Ptr {
13 text,
14 len: 0.into(),
15 }
16 }
17
18 pub fn into_len(self) -> TextUnit {
19 self.len
20 }
21
22 pub fn next(&self) -> Option<char> {
23 self.chars().next()
24 }
25
26 pub fn nnext(&self) -> Option<char> {
27 let mut chars = self.chars();
28 chars.next()?;
29 chars.next()
30 }
31
32 pub fn next_is(&self, c: char) -> bool {
33 self.next() == Some(c)
34 }
35
36 pub fn nnext_is(&self, c: char) -> bool {
37 self.nnext() == Some(c)
38 }
39
40 pub fn next_is_p<P: Fn(char) -> bool>(&self, p: P) -> bool {
41 self.next().map(p) == Some(true)
42 }
43
44 pub fn nnext_is_p<P: Fn(char) -> bool>(&self, p: P) -> bool {
45 self.nnext().map(p) == Some(true)
46 }
47
48 pub fn bump(&mut self) -> Option<char> {
49 let ch = self.chars().next()?;
50 self.len += TextUnit::of_char(ch);
51 Some(ch)
52 }
53
54 pub fn bump_while<F: Fn(char) -> bool>(&mut self, pred: F) {
55 loop {
56 match self.next() {
57 Some(c) if pred(c) => {
58 self.bump();
59 }
60 _ => return,
61 }
62 }
63 }
64
65 pub fn current_token_text(&self) -> &str {
66 let len: u32 = self.len.into();
67 &self.text[..len as usize]
68 }
69
70 fn chars(&self) -> Chars {
71 let len: u32 = self.len.into();
72 self.text[len as usize..].chars()
73 }
74}
diff --git a/src/lexer/strings.rs b/src/lexer/strings.rs
deleted file mode 100644
index e3704fbb3..000000000
--- a/src/lexer/strings.rs
+++ /dev/null
@@ -1,106 +0,0 @@
1use SyntaxKind::{self, *};
2
3use lexer::ptr::Ptr;
4
5pub(crate) fn is_string_literal_start(c: char, c1: Option<char>, c2: Option<char>) -> bool {
6 match (c, c1, c2) {
7 ('r', Some('"'), _)
8 | ('r', Some('#'), _)
9 | ('b', Some('"'), _)
10 | ('b', Some('\''), _)
11 | ('b', Some('r'), Some('"'))
12 | ('b', Some('r'), Some('#')) => true,
13 _ => false,
14 }
15}
16
17pub(crate) fn scan_char(ptr: &mut Ptr) {
18 if ptr.bump().is_none() {
19 return; // TODO: error reporting is upper in the stack
20 }
21 scan_char_or_byte(ptr);
22 if !ptr.next_is('\'') {
23 return; // TODO: error reporting
24 }
25 ptr.bump();
26}
27
28pub(crate) fn scan_byte_char_or_string(ptr: &mut Ptr) -> SyntaxKind {
29 // unwrapping and not-exhaustive match are ok
30 // because of string_literal_start
31 let c = ptr.bump().unwrap();
32 match c {
33 '\'' => {
34 scan_byte(ptr);
35 BYTE
36 }
37 '"' => {
38 scan_byte_string(ptr);
39 BYTE_STRING
40 }
41 'r' => {
42 scan_raw_byte_string(ptr);
43 RAW_BYTE_STRING
44 }
45 _ => unreachable!(),
46 }
47}
48
49pub(crate) fn scan_string(ptr: &mut Ptr) {
50 while let Some(c) = ptr.bump() {
51 if c == '"' {
52 return;
53 }
54 }
55}
56
57pub(crate) fn scan_raw_string(ptr: &mut Ptr) {
58 if !ptr.next_is('"') {
59 return;
60 }
61 ptr.bump();
62
63 while let Some(c) = ptr.bump() {
64 if c == '"' {
65 return;
66 }
67 }
68}
69
70fn scan_byte(ptr: &mut Ptr) {
71 if ptr.next_is('\'') {
72 ptr.bump();
73 return;
74 }
75 ptr.bump();
76 if ptr.next_is('\'') {
77 ptr.bump();
78 return;
79 }
80}
81
82fn scan_byte_string(ptr: &mut Ptr) {
83 while let Some(c) = ptr.bump() {
84 if c == '"' {
85 return;
86 }
87 }
88}
89
90fn scan_raw_byte_string(ptr: &mut Ptr) {
91 if !ptr.next_is('"') {
92 return;
93 }
94 ptr.bump();
95
96 while let Some(c) = ptr.bump() {
97 if c == '"' {
98 return;
99 }
100 }
101}
102
103fn scan_char_or_byte(ptr: &mut Ptr) {
104 //FIXME: deal with escape sequencies
105 ptr.bump();
106}
generated by cgit v1.2.3 (git 2.25.1) at 2024-07-05 18:27:59 +0100