summaryrefslogtreecommitdiffstats
path: root/compiler/rustc_parse_format/src
diff options
context:
space:
mode:
authorDaniel Baumann <daniel.baumann@progress-linux.org>2024-04-17 12:18:32 +0000
committerDaniel Baumann <daniel.baumann@progress-linux.org>2024-04-17 12:18:32 +0000
commit4547b622d8d29df964fa2914213088b148c498fc (patch)
tree9fc6b25f3c3add6b745be9a2400a6e96140046e9 /compiler/rustc_parse_format/src
parentReleasing progress-linux version 1.66.0+dfsg1-1~progress7.99u1. (diff)
downloadrustc-4547b622d8d29df964fa2914213088b148c498fc.tar.xz
rustc-4547b622d8d29df964fa2914213088b148c498fc.zip
Merging upstream version 1.67.1+dfsg1.
Signed-off-by: Daniel Baumann <daniel.baumann@progress-linux.org>
Diffstat (limited to 'compiler/rustc_parse_format/src')
-rw-r--r--compiler/rustc_parse_format/src/lib.rs141
1 files changed, 75 insertions, 66 deletions
diff --git a/compiler/rustc_parse_format/src/lib.rs b/compiler/rustc_parse_format/src/lib.rs
index 1394993ab..0113eb4e3 100644
--- a/compiler/rustc_parse_format/src/lib.rs
+++ b/compiler/rustc_parse_format/src/lib.rs
@@ -818,85 +818,94 @@ fn find_skips_from_snippet(
_ => return (vec![], false),
};
- fn find_skips(snippet: &str, is_raw: bool) -> Vec<usize> {
- let mut s = snippet.char_indices().peekable();
- let mut skips = vec![];
- while let Some((pos, c)) = s.next() {
- match (c, s.peek()) {
- // skip whitespace and empty lines ending in '\\'
- ('\\', Some((next_pos, '\n'))) if !is_raw => {
- skips.push(pos);
- skips.push(*next_pos);
- let _ = s.next();
+ if str_style.is_some() {
+ return (vec![], true);
+ }
- while let Some((pos, c)) = s.peek() {
- if matches!(c, ' ' | '\n' | '\t') {
- skips.push(*pos);
- let _ = s.next();
- } else {
- break;
- }
- }
- }
- ('\\', Some((next_pos, 'n' | 't' | 'r' | '0' | '\\' | '\'' | '\"'))) => {
- skips.push(*next_pos);
- let _ = s.next();
- }
- ('\\', Some((_, 'x'))) if !is_raw => {
- for _ in 0..3 {
- // consume `\xAB` literal
- if let Some((pos, _)) = s.next() {
- skips.push(pos);
- } else {
- break;
- }
+ let snippet = &snippet[1..snippet.len() - 1];
+
+ let mut s = snippet.char_indices();
+ let mut skips = vec![];
+ while let Some((pos, c)) = s.next() {
+ match (c, s.clone().next()) {
+ // skip whitespace and empty lines ending in '\\'
+ ('\\', Some((next_pos, '\n'))) => {
+ skips.push(pos);
+ skips.push(next_pos);
+ let _ = s.next();
+
+ while let Some((pos, c)) = s.clone().next() {
+ if matches!(c, ' ' | '\n' | '\t') {
+ skips.push(pos);
+ let _ = s.next();
+ } else {
+ break;
}
}
- ('\\', Some((_, 'u'))) if !is_raw => {
+ }
+ ('\\', Some((next_pos, 'n' | 't' | 'r' | '0' | '\\' | '\'' | '\"'))) => {
+ skips.push(next_pos);
+ let _ = s.next();
+ }
+ ('\\', Some((_, 'x'))) => {
+ for _ in 0..3 {
+ // consume `\xAB` literal
if let Some((pos, _)) = s.next() {
skips.push(pos);
+ } else {
+ break;
}
- if let Some((next_pos, next_c)) = s.next() {
- if next_c == '{' {
- skips.push(next_pos);
- let mut i = 0; // consume up to 6 hexanumeric chars + closing `}`
- while let (Some((next_pos, c)), true) = (s.next(), i < 7) {
- if c.is_digit(16) {
- skips.push(next_pos);
- } else if c == '}' {
- skips.push(next_pos);
- break;
- } else {
- break;
- }
- i += 1;
- }
- } else if next_c.is_digit(16) {
- skips.push(next_pos);
- // We suggest adding `{` and `}` when appropriate, accept it here as if
- // it were correct
- let mut i = 0; // consume up to 6 hexanumeric chars
- while let (Some((next_pos, c)), _) = (s.next(), i < 6) {
- if c.is_digit(16) {
- skips.push(next_pos);
- } else {
- break;
- }
- i += 1;
+ }
+ }
+ ('\\', Some((_, 'u'))) => {
+ if let Some((pos, _)) = s.next() {
+ skips.push(pos);
+ }
+ if let Some((next_pos, next_c)) = s.next() {
+ if next_c == '{' {
+ // consume up to 6 hexanumeric chars
+ let digits_len =
+ s.clone().take(6).take_while(|(_, c)| c.is_digit(16)).count();
+
+ let len_utf8 = s
+ .as_str()
+ .get(..digits_len)
+ .and_then(|digits| u32::from_str_radix(digits, 16).ok())
+ .and_then(char::from_u32)
+ .map_or(1, char::len_utf8);
+
+ // Skip the digits, for chars that encode to more than 1 utf-8 byte
+ // exclude as many digits as it is greater than 1 byte
+ //
+ // So for a 3 byte character, exclude 2 digits
+ let required_skips = digits_len.saturating_sub(len_utf8.saturating_sub(1));
+
+ // skip '{' and '}' also
+ for pos in (next_pos..).take(required_skips + 2) {
+ skips.push(pos)
+ }
+
+ s.nth(digits_len);
+ } else if next_c.is_digit(16) {
+ skips.push(next_pos);
+ // We suggest adding `{` and `}` when appropriate, accept it here as if
+ // it were correct
+ let mut i = 0; // consume up to 6 hexanumeric chars
+ while let (Some((next_pos, c)), _) = (s.next(), i < 6) {
+ if c.is_digit(16) {
+ skips.push(next_pos);
+ } else {
+ break;
}
+ i += 1;
}
}
}
- _ => {}
}
+ _ => {}
}
- skips
}
-
- let r_start = str_style.map_or(0, |r| r + 1);
- let r_end = str_style.unwrap_or(0);
- let s = &snippet[r_start + 1..snippet.len() - r_end - 1];
- (find_skips(s, str_style.is_some()), true)
+ (skips, true)
}
#[cfg(test)]