Offline, privacy-first grammar checker. Fast, open-source, Rust-powered
fix(core): avoid more allocation by using `Cow`
Elijah Potter9a018f49d112
diff --git a/harper-core/src/char_string.rs b/harper-core/src/char_string.rsindex f6df85487..ec65ac7fc 100644--- a/harper-core/src/char_string.rs+++ b/harper-core/src/char_string.rs@@ -1,3 +1,5 @@+use std::borrow::Cow;+use smallvec::SmallVec;/// A char sequence that improves cache locality.@@ -6,18 +8,23 @@ pub type CharString = SmallVec<[char; 16]>;/// Extensions to character sequences that make them easier to wrangle.pub trait CharStringExt {- fn to_lower(&self) -> CharString;+ fn to_lower(&self) -> Cow<[char]>;fn to_string(&self) -> String;}impl CharStringExt for [char] {- fn to_lower(&self) -> CharString {+ fn to_lower(&self) -> Cow<[char]> {+ if self.iter().all(|c| c.is_lowercase()) {+ return Cow::Borrowed(self);+ }+let mut out = CharString::with_capacity(self.len());out.extend(self.iter().flat_map(|v| v.to_lowercase()));- out+ Cow::Owned(out.to_vec())}+fn to_string(&self) -> String {self.iter().collect()}diff --git a/harper-core/src/linting/repeated_words.rs b/harper-core/src/linting/repeated_words.rsindex 8a73ebef8..79fb4ac39 100644--- a/harper-core/src/linting/repeated_words.rs+++ b/harper-core/src/linting/repeated_words.rs@@ -27,7 +27,7 @@ impl RepeatedWords {self.special_cases.iter()- .any(|v| v.as_slice() == lower.as_slice())+ .any(|v| v.as_slice() == lower.as_ref())}}diff --git a/harper-core/src/spell/mutable_dictionary.rs b/harper-core/src/spell/mutable_dictionary.rsindex 3822633d8..8b879165b 100644--- a/harper-core/src/spell/mutable_dictionary.rs+++ b/harper-core/src/spell/mutable_dictionary.rs@@ -64,7 +64,7 @@ fn uncached_inner_new() -> Arc<MutableDictionary> {let mut word_map_lowercase = HashMap::with_capacity(word_map.len());for key in word_map.keys() {- word_map_lowercase.insert(key.to_lower(), key.clone());+ word_map_lowercase.insert(key.to_lower().to_smallvec(), key.clone());}Arc::new(MutableDictionary {@@ -113,8 +113,11 @@ impl MutableDictionary {self.words.extend(pairs.iter().map(|(v, _)| v.clone()));self.words.sort_by_key(|w| w.len());self.word_len_starts = Self::create_len_starts(&self.words);- self.word_map_lowercase- .extend(pairs.iter().map(|(key, _)| (key.to_lower(), key.clone())));+ self.word_map_lowercase.extend(+ pairs+ .iter()+ .map(|(key, _)| (key.to_lower().to_smallvec(), key.clone())),+ );self.word_map.extend(pairs);}@@ -169,9 +172,9 @@ impl Dictionary for MutableDictionary {fn contains_word(&self, word: &[char]) -> bool {let normalized = seq_to_normalized(word);- let lowercase: CharString = normalized.to_lower();+ let lowercase = normalized.to_lower();- self.word_map_lowercase.contains_key(&lowercase)+ self.word_map_lowercase.contains_key(lowercase.as_ref())}fn contains_word_str(&self, word: &str) -> bool {@@ -186,10 +189,10 @@ impl Dictionary for MutableDictionary {fn get_correct_capitalization_of(&self, word: &[char]) -> Option<&'_ [char]> {let normalized = seq_to_normalized(word);- let lowercase: CharString = normalized.to_lower();+ let lowercase = normalized.to_lower();self.word_map_lowercase- .get(&lowercase)+ .get(lowercase.as_ref()).map(|v| v.as_slice())}diff --git a/harper-core/src/title_case.rs b/harper-core/src/title_case.rsindex 8c4b2b7da..b504151ac 100644--- a/harper-core/src/title_case.rs+++ b/harper-core/src/title_case.rs@@ -97,7 +97,7 @@ fn should_capitalize_token(tok: &Token, source: &[char], dict: &impl Dictionary)!is_short_preposition&& !metadata.article- && !SPECIAL_CONJUNCTIONS.contains(chars_lower.as_slice())+ && !SPECIAL_CONJUNCTIONS.contains(chars_lower.as_ref())}_ => true,}