actions/typos
1
0
Fork 0
mirror of https://github.com/crate-ci/typos.git synced 2024-11-22 09:01:04 -05:00
Code Issues Projects Releases Packages Wiki Activity Actions

Merge pull request #60 from epage/lazy

Browse source 
perf: Speed up identifier splitting
This commit is contained in:
Ed Page 2019-10-25 15:01:42 -06:00 committed by GitHub
commit af49b6af86
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
1 changed files with 89 additions and 61 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

150
typos/src/tokens.rs
View file

@ -238,72 +238,100 @@ impl WordMode {
} }
} }
fn split_ident(ident: &str, offset: usize) -> impl Iterator<Item = Word<'_>> { struct SplitIdent<'s> {
let mut result = vec![]; ident: &'s str,
offset: usize,
let mut char_indices = ident.char_indices().peekable(); char_indices: std::iter::Peekable<std::str::CharIndices<'s>>,
let mut start = 0; start: usize,
let mut start_mode = WordMode::Boundary; start_mode: WordMode,
let mut last_mode = WordMode::Boundary; last_mode: WordMode,
while let Some((i, c)) = char_indices.next() { }
let cur_mode = WordMode::classify(c);
if cur_mode == WordMode::Boundary {
assert!(start_mode == WordMode::Boundary);
continue;
}
if start_mode == WordMode::Boundary {
start_mode = cur_mode;
start = i;
}
if let Some(&(next_i, next)) = char_indices.peek() { impl<'s> SplitIdent<'s> {
// The mode including the current character, assuming the current character does fn new(ident: &'s str, offset: usize) -> Self {
// not result in a word boundary. Self {
let next_mode = WordMode::classify(next); ident,
offset,
match (last_mode, cur_mode, next_mode) { char_indices: ident.char_indices().peekable(),
// cur_mode is last of current word start: 0,
(_, _, WordMode::Boundary) start_mode: WordMode::Boundary,
| (_, WordMode::Lowercase, WordMode::Number) last_mode: WordMode::Boundary,
| (_, WordMode::Uppercase, WordMode::Number)
| (_, WordMode::Number, WordMode::Lowercase)
| (_, WordMode::Number, WordMode::Uppercase)
| (_, WordMode::Lowercase, WordMode::Uppercase) => {
let case = start_mode.case(cur_mode);
result.push(Word::new_unchecked(
&ident[start..next_i],
case,
start + offset,
));
start = next_i;
start_mode = WordMode::Boundary;
last_mode = WordMode::Boundary;
}
// cur_mode is start of next word
(WordMode::Uppercase, WordMode::Uppercase, WordMode::Lowercase) => {
result.push(Word::new_unchecked(
&ident[start..i],
Case::Scream,
start + offset,
));
start = i;
start_mode = cur_mode;
last_mode = WordMode::Boundary;
}
// No word boundary
(_, _, _) => {
last_mode = cur_mode;
}
}
} else {
// Collect trailing characters as a word
let case = start_mode.case(cur_mode);
result.push(Word::new_unchecked(&ident[start..], case, start + offset));
break;
} }
} }
}
result.into_iter() impl<'s> Iterator for SplitIdent<'s> {
type Item = Word<'s>;
fn next(&mut self) -> Option<Word<'s>> {
while let Some((i, c)) = self.char_indices.next() {
let cur_mode = WordMode::classify(c);
if cur_mode == WordMode::Boundary {
assert!(self.start_mode == WordMode::Boundary);
continue;
}
if self.start_mode == WordMode::Boundary {
self.start_mode = cur_mode;
self.start = i;
}
if let Some(&(next_i, next)) = self.char_indices.peek() {
// The mode including the current character, assuming the current character does
// not result in a word boundary.
let next_mode = WordMode::classify(next);
match (self.last_mode, cur_mode, next_mode) {
// cur_mode is last of current word
(_, _, WordMode::Boundary)
| (_, WordMode::Lowercase, WordMode::Number)
| (_, WordMode::Uppercase, WordMode::Number)
| (_, WordMode::Number, WordMode::Lowercase)
| (_, WordMode::Number, WordMode::Uppercase)
| (_, WordMode::Lowercase, WordMode::Uppercase) => {
let case = self.start_mode.case(cur_mode);
let result = Word::new_unchecked(
&self.ident[self.start..next_i],
case,
self.start + self.offset,
);
self.start = next_i;
self.start_mode = WordMode::Boundary;
self.last_mode = WordMode::Boundary;
return Some(result);
}
// cur_mode is start of next word
(WordMode::Uppercase, WordMode::Uppercase, WordMode::Lowercase) => {
let result = Word::new_unchecked(
&self.ident[self.start..i],
Case::Scream,
self.start + self.offset,
);
self.start = i;
self.start_mode = cur_mode;
self.last_mode = WordMode::Boundary;
return Some(result);
}
// No word boundary
(_, _, _) => {
self.last_mode = cur_mode;
}
}
} else {
// Collect trailing characters as a word
let case = self.start_mode.case(cur_mode);
let result =
Word::new_unchecked(&self.ident[self.start..], case, self.start + self.offset);
return Some(result);
}
}
None
}
}
fn split_ident(ident: &str, offset: usize) -> impl Iterator<Item = Word<'_>> {
SplitIdent::new(ident, offset)
} }
#[cfg(test)] #[cfg(test)]