Skip to content

Offline, privacy-first grammar checker. Fast, open-source, Rust-powered

fix(core): out of bounds errors

Elijah Potter4167300e41b3
diff --git a/harper-core/src/parsers/markdown.rs b/harper-core/src/parsers/markdown.rs
index ea21e9c4e..16e40cb18 100644
--- a/harper-core/src/parsers/markdown.rs
+++ b/harper-core/src/parsers/markdown.rs
@@ -37,7 +37,7 @@ impl Markdown {
/// Remove hidden Wikilink target text.
///
- /// As in, the stuff to the left of the pipe operator:
+ /// As in the stuff to the left of the pipe operator:
///
/// ```markdown
/// [[Target text|Display Text]]
@@ -46,6 +46,10 @@ impl Markdown {
let mut to_remove = VecDeque::new();
for pipe_idx in tokens.iter_pipe_indices() {
+ if pipe_idx < 2 {
+ continue;
+ }
+
// Locate preceding `[[`
let mut cursor = pipe_idx - 2;
let mut open_bracket = None;
@@ -351,6 +355,35 @@ mod tests {
))
}
+ #[test]
+ fn just_pipe() {
+ let source = r"|";
+
+ let tokens = Markdown::default().parse_str(source);
+
+ let token_kinds = tokens.iter().map(|t| t.kind).collect::<Vec<_>>();
+
+ dbg!(&token_kinds);
+
+ assert!(matches!(
+ token_kinds.as_slice(),
+ &[TokenKind::Punctuation(Punctuation::Pipe)]
+ ))
+ }
+
+ #[test]
+ fn empty_wikilink_text() {
+ let source = r"[[|]]";
+
+ let tokens = Markdown::default().parse_str(source);
+
+ let token_kinds = tokens.iter().map(|t| t.kind).collect::<Vec<_>>();
+
+ dbg!(&token_kinds);
+
+ assert!(matches!(token_kinds.as_slice(), &[]))
+ }
+
#[test]
fn improper_wikilink_text() {
let source = r"this is shown|this is also shown]]";
diff --git a/harper-core/src/title_case.rs b/harper-core/src/title_case.rs
index 3dd262698..c215343eb 100644
--- a/harper-core/src/title_case.rs
+++ b/harper-core/src/title_case.rs
@@ -1,5 +1,6 @@
use crate::Lrc;
use crate::Token;
+use crate::TokenKind;
use hashbrown::HashSet;
use lazy_static::lazy_static;
@@ -33,33 +34,8 @@ pub fn make_title_case(toks: &[Token], source: &[char], dict: &impl Dictionary)
let mut word_likes = toks.iter_word_likes().enumerate().peekable();
let mut output = toks.span().unwrap().get_content(source).to_vec();
- // Only specific conjunctions are not capitalized.
- lazy_static! {
- static ref SPECIAL_CONJUNCTIONS: HashSet<Vec<char>> = ["and", "but", "for", "or", "nor"]
- .iter()
- .map(|v| v.chars().collect())
- .collect();
- }
-
while let Some((index, word)) = word_likes.next() {
- if !word.kind.is_word() {
- continue;
- }
-
- let chars = word.span.get_content(source);
- let chars_lower = chars.to_lower();
-
- let metadata = word
- .kind
- .as_word()
- .unwrap()
- .or(&dict.get_word_metadata(&chars_lower));
-
- let is_short_preposition = metadata.preposition && word.span.len() <= 4;
-
- let should_capitalize = (!is_short_preposition
- && !metadata.article
- && !SPECIAL_CONJUNCTIONS.contains(chars_lower.as_slice()))
+ let should_capitalize = should_capitalize_token(&word, source, dict)
|| index == 0
|| word_likes.peek().is_none();
@@ -82,10 +58,39 @@ pub fn make_title_case(toks: &[Token], source: &[char], dict: &impl Dictionary)
output
}
+/// Determines whether a token should be capitalized.
+/// Is not responsible for capitalization requirements that are dependent on token position.
+fn should_capitalize_token(tok: &Token, source: &[char], dict: &impl Dictionary) -> bool {
+ match tok.kind {
+ TokenKind::Word(mut metadata) => {
+ // Only specific conjunctions are not capitalized.
+ lazy_static! {
+ static ref SPECIAL_CONJUNCTIONS: HashSet<Vec<char>> =
+ ["and", "but", "for", "or", "nor"]
+ .iter()
+ .map(|v| v.chars().collect())
+ .collect();
+ }
+
+ let chars = tok.span.get_content(source);
+ let chars_lower = chars.to_lower();
+
+ metadata = metadata.or(&dict.get_word_metadata(&chars_lower));
+
+ let is_short_preposition = metadata.preposition && tok.span.len() <= 4;
+
+ !is_short_preposition
+ && !metadata.article
+ && !SPECIAL_CONJUNCTIONS.contains(chars_lower.as_slice())
+ }
+ _ => true,
+ }
+}
+
#[cfg(test)]
mod tests {
- use quickcheck::{Arbitrary, TestResult};
+ use quickcheck::TestResult;
use quickcheck_macros::quickcheck;
use super::make_title_case_str;
@@ -131,49 +136,28 @@ mod tests {
)
}
- #[derive(Debug, Clone)]
- struct Word(String);
-
- impl Arbitrary for Word {
- fn arbitrary(g: &mut quickcheck::Gen) -> Self {
- let mut s = String::new();
-
- for _ in 0..g.size() {
- let c = loop {
- let gen = char::arbitrary(g);
-
- if gen.is_ascii_alphanumeric() {
- break gen;
- }
- };
-
- s.push(c);
- }
-
- Self(s)
- }
+ #[test]
+ fn pipe_does_not_cause_crash() {
+ assert_eq!(
+ make_title_case_str("|", &Markdown::default(), &FstDictionary::curated()),
+ "|"
+ )
}
- #[derive(Debug, Clone)]
- struct Sentence(String);
-
- /// Builds a sentence out of words from the curated [`FullDictionary`].
- impl Arbitrary for Sentence {
- fn arbitrary(g: &mut quickcheck::Gen) -> Self {
- let mut s = String::new();
-
- let Word(first_word) = Word::arbitrary(g);
- s.push_str(&first_word);
-
- for _ in 0..g.size() {
- let Word(word) = Word::arbitrary(g);
-
- s.push(' ');
- s.push_str(&word);
- }
+ #[test]
+ fn a_paragraph_does_not_cause_crash() {
+ assert_eq!(
+ make_title_case_str("A\n", &Markdown::default(), &FstDictionary::curated()),
+ "A"
+ )
+ }
- Self(s)
- }
+ #[test]
+ fn tab_a_becomes_upcase() {
+ assert_eq!(
+ make_title_case_str("\ta", &PlainEnglish, &FstDictionary::curated()),
+ "\tA"
+ )
}
#[quickcheck]
@@ -195,8 +179,6 @@ mod tests {
.chars()
.collect();
- dbg!(&title_case);
-
TestResult::from_bool(title_case[prefix.chars().count() + 1] == 'a')
}
@@ -223,15 +205,15 @@ mod tests {
}
#[quickcheck]
- fn first_word_is_upcase(sentence: Sentence) -> TestResult {
+ fn first_word_is_upcase(text: String) -> TestResult {
let title_case: Vec<_> =
- make_title_case_str(&sentence.0, &Markdown::default(), &FstDictionary::curated())
+ make_title_case_str(&text, &PlainEnglish, &FstDictionary::curated())
.chars()
.collect();
if let Some(first) = title_case.first() {
- if first.is_alphabetic() {
- TestResult::from_bool(first.is_uppercase())
+ if first.is_ascii_alphabetic() {
+ TestResult::from_bool(first.is_ascii_uppercase())
} else {
TestResult::discard()
}
diff --git a/harper-core/src/token.rs b/harper-core/src/token.rs
index 06ab11ce9..f33c4a2b4 100644
--- a/harper-core/src/token.rs
+++ b/harper-core/src/token.rs
@@ -156,7 +156,16 @@ impl TokenStringExt for [Token] {
}
fn span(&self) -> Option<Span> {
- Some(Span::new(self.first()?.span.start, self.last()?.span.end))
+ let min_max = self
+ .iter()
+ .flat_map(|v| [v.span.start, v.span.end].into_iter())
+ .minmax();
+
+ match min_max {
+ itertools::MinMaxResult::NoElements => None,
+ itertools::MinMaxResult::OneElement(min) => Some(Span::new(min, min)),
+ itertools::MinMaxResult::MinMax(min, max) => Some(Span::new(min, max)),
+ }
}
fn iter_linking_verb_indices(&self) -> impl Iterator<Item = usize> + '_ {
diff --git a/justfile b/justfile
index 54f16bfc9..20f056b4b 100644
--- a/justfile
+++ b/justfile
@@ -270,12 +270,12 @@ bump-versions:
lazygit
# Enter an infinite loop of testing until a bug is found.
-fuzz mod:
+fuzz:
#!/usr/bin/bash
while true
do
- QUICKCHECK_TESTS=1000 cargo test {{mod}}
+ QUICKCHECK_TESTS=100000 cargo test
if [[ x$? != x0 ]] ; then
exit $?
fi