forked from RustPython/RustPython
-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathstring_parser.rs
More file actions
287 lines (248 loc) · 10.2 KB
/
string_parser.rs
File metadata and controls
287 lines (248 loc) · 10.2 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
//! A stripped-down version of ruff's string literal parser, modified to
//! handle surrogates in string literals and output WTF-8.
//!
//! Any `unreachable!()` statements in this file are because we only get here
//! after ruff has already successfully parsed the string literal, meaning
//! we don't need to do any validation or error handling.
use std::convert::Infallible;
use ruff_python_ast::{AnyStringFlags, StringFlags};
use rustpython_wtf8::{CodePoint, Wtf8, Wtf8Buf};
// use ruff_python_parser::{LexicalError, LexicalErrorType};
type LexicalError = Infallible;
enum EscapedChar {
Literal(CodePoint),
Escape(char),
}
struct StringParser {
/// The raw content of the string e.g., the `foo` part in `"foo"`.
source: Box<str>,
/// Current position of the parser in the source.
cursor: usize,
/// Flags that can be used to query information about the string.
flags: AnyStringFlags,
}
impl StringParser {
fn new(source: Box<str>, flags: AnyStringFlags) -> Self {
Self {
source,
cursor: 0,
flags,
}
}
#[inline]
fn skip_bytes(&mut self, bytes: usize) -> &str {
let skipped_str = &self.source[self.cursor..self.cursor + bytes];
self.cursor += bytes;
skipped_str
}
/// Returns the next byte in the string, if there is one.
///
/// # Panics
///
/// When the next byte is a part of a multi-byte character.
#[inline]
fn next_byte(&mut self) -> Option<u8> {
self.source[self.cursor..].as_bytes().first().map(|&byte| {
self.cursor += 1;
byte
})
}
#[inline]
fn next_char(&mut self) -> Option<char> {
self.source[self.cursor..].chars().next().inspect(|c| {
self.cursor += c.len_utf8();
})
}
#[inline]
fn peek_byte(&self) -> Option<u8> {
self.source[self.cursor..].as_bytes().first().copied()
}
fn parse_unicode_literal(&mut self, literal_number: usize) -> Result<CodePoint, LexicalError> {
let mut p: u32 = 0u32;
for i in 1..=literal_number {
match self.next_char() {
Some(c) => match c.to_digit(16) {
Some(d) => p += d << ((literal_number - i) * 4),
None => unreachable!(),
},
None => unreachable!(),
}
}
Ok(CodePoint::from_u32(p).unwrap())
}
fn parse_octet(&mut self, o: u8) -> char {
let mut radix_bytes = [o, 0, 0];
let mut len = 1;
while len < 3 {
let Some(b'0'..=b'7') = self.peek_byte() else {
break;
};
radix_bytes[len] = self.next_byte().unwrap();
len += 1;
}
// OK because radix_bytes is always going to be in the ASCII range.
let radix_str = std::str::from_utf8(&radix_bytes[..len]).expect("ASCII bytes");
let value = u32::from_str_radix(radix_str, 8).unwrap();
char::from_u32(value).unwrap()
}
fn parse_unicode_name(&mut self) -> Result<char, LexicalError> {
let Some('{') = self.next_char() else {
unreachable!()
};
let Some(close_idx) = self.source[self.cursor..].find('}') else {
unreachable!()
};
let name_and_ending = self.skip_bytes(close_idx + 1);
let name = &name_and_ending[..name_and_ending.len() - 1];
unicode_names2::character(name).ok_or_else(|| unreachable!())
}
/// Parse an escaped character, returning the new character.
fn parse_escaped_char(&mut self) -> Result<Option<EscapedChar>, LexicalError> {
let Some(first_char) = self.next_char() else {
unreachable!()
};
let new_char = match first_char {
'\\' => '\\'.into(),
'\'' => '\''.into(),
'\"' => '"'.into(),
'a' => '\x07'.into(),
'b' => '\x08'.into(),
'f' => '\x0c'.into(),
'n' => '\n'.into(),
'r' => '\r'.into(),
't' => '\t'.into(),
'v' => '\x0b'.into(),
o @ '0'..='7' => self.parse_octet(o as u8).into(),
'x' => self.parse_unicode_literal(2)?,
'u' if !self.flags.is_byte_string() => self.parse_unicode_literal(4)?,
'U' if !self.flags.is_byte_string() => self.parse_unicode_literal(8)?,
'N' if !self.flags.is_byte_string() => self.parse_unicode_name()?.into(),
// Special cases where the escape sequence is not a single character
'\n' => return Ok(None),
'\r' => {
if self.peek_byte() == Some(b'\n') {
self.next_byte();
}
return Ok(None);
}
_ => return Ok(Some(EscapedChar::Escape(first_char))),
};
Ok(Some(EscapedChar::Literal(new_char)))
}
fn parse_fstring_middle(mut self) -> Result<Box<Wtf8>, LexicalError> {
// Fast-path: if the f-string doesn't contain any escape sequences, return the literal.
let Some(mut index) = memchr::memchr3(b'{', b'}', b'\\', self.source.as_bytes()) else {
return Ok(self.source.into());
};
let mut value = Wtf8Buf::with_capacity(self.source.len());
loop {
// Add the characters before the escape sequence (or curly brace) to the string.
let before_with_slash_or_brace = self.skip_bytes(index + 1);
let before = &before_with_slash_or_brace[..before_with_slash_or_brace.len() - 1];
value.push_str(before);
// Add the escaped character to the string.
match &self.source.as_bytes()[self.cursor - 1] {
// If there are any curly braces inside a `FStringMiddle` token,
// then they were escaped (i.e. `{{` or `}}`). This means that
// we need increase the location by 2 instead of 1.
b'{' => value.push_char('{'),
b'}' => value.push_char('}'),
// We can encounter a `\` as the last character in a `FStringMiddle`
// token which is valid in this context. For example,
//
// ```python
// f"\{foo} \{bar:\}"
// # ^ ^^ ^
// ```
//
// Here, the `FStringMiddle` token content will be "\" and " \"
// which is invalid if we look at the content in isolation:
//
// ```python
// "\"
// ```
//
// However, the content is syntactically valid in the context of
// the f-string because it's a substring of the entire f-string.
// This is still an invalid escape sequence, but we don't want to
// raise a syntax error as is done by the CPython parser. It might
// be supported in the future, refer to point 3: https://peps.python.org/pep-0701/#rejected-ideas
b'\\' => {
if !self.flags.is_raw_string() && self.peek_byte().is_some() {
match self.parse_escaped_char()? {
None => {}
Some(EscapedChar::Literal(c)) => value.push(c),
Some(EscapedChar::Escape(c)) => {
value.push_char('\\');
value.push_char(c);
}
}
} else {
value.push_char('\\');
}
}
ch => {
unreachable!("Expected '{{', '}}', or '\\' but got {:?}", ch);
}
}
let Some(next_index) =
memchr::memchr3(b'{', b'}', b'\\', self.source[self.cursor..].as_bytes())
else {
// Add the rest of the string to the value.
let rest = &self.source[self.cursor..];
value.push_str(rest);
break;
};
index = next_index;
}
Ok(value.into())
}
fn parse_string(mut self) -> Result<Box<Wtf8>, LexicalError> {
if self.flags.is_raw_string() {
// For raw strings, no escaping is necessary.
return Ok(self.source.into());
}
let Some(mut escape) = memchr::memchr(b'\\', self.source.as_bytes()) else {
// If the string doesn't contain any escape sequences, return the owned string.
return Ok(self.source.into());
};
// If the string contains escape sequences, we need to parse them.
let mut value = Wtf8Buf::with_capacity(self.source.len());
loop {
// Add the characters before the escape sequence to the string.
let before_with_slash = self.skip_bytes(escape + 1);
let before = &before_with_slash[..before_with_slash.len() - 1];
value.push_str(before);
// Add the escaped character to the string.
match self.parse_escaped_char()? {
None => {}
Some(EscapedChar::Literal(c)) => value.push(c),
Some(EscapedChar::Escape(c)) => {
value.push_char('\\');
value.push_char(c);
}
}
let Some(next_escape) = self.source[self.cursor..].find('\\') else {
// Add the rest of the string to the value.
let rest = &self.source[self.cursor..];
value.push_str(rest);
break;
};
// Update the position of the next escape sequence.
escape = next_escape;
}
Ok(value.into())
}
}
pub(crate) fn parse_string_literal(source: &str, flags: AnyStringFlags) -> Box<Wtf8> {
let source = &source[flags.opener_len().to_usize()..];
let source = &source[..source.len() - flags.quote_len().to_usize()];
StringParser::new(source.into(), flags)
.parse_string()
.unwrap_or_else(|x| match x {})
}
pub(crate) fn parse_fstring_literal_element(source: Box<str>, flags: AnyStringFlags) -> Box<Wtf8> {
StringParser::new(source, flags)
.parse_fstring_middle()
.unwrap_or_else(|x| match x {})
}