test: make sure linter descriptions are grammatically correct

Which helped me discover bugs in the `CompoundWords` linter.

demo.md

@@ -4,7 +4,7 @@ checkers don't cut it. That s where Harper comes in handy.
 Harper is an language checker for developers. it can detect
 improper capitalization and misspellled words,
 as well as a number of other issues.
-Like if you break up words you shouldn't.
+Like if you break up words you shoul dn't.
 
 Harper works everywhere, even offline. Since you r data
 never leaves your device, you don't ned to worry aout us

harper-comments/tests/language_support_sources/javadoc_complex.java

@@ -19,7 +19,7 @@ public static void main(String[] args) {
   /**
    * This doc has a link in it: {@link this sould b ignor} but not tis
    *
-   * @param name this is an other test.
+   * @param name this is anoher test.
    */
   public static void greet(String name) {
     System.out.println("Hello " + name + ".");

harper-core/dictionary.dict

@@ -28344,7 +28344,7 @@ inimitably/
 iniquitous/5Y
 iniquity/1SM
 initial/514SGMDY
-initialism/1
+initialism/1MS
 initialization/1
 initialize/4DSG
 initialized/4AU
@@ -49636,7 +49636,6 @@ scatterplot/14SMG
 Wikilink/MS1
 stacktrace/SM1
 scrollbar/1SM
-break-up/1SM
 sweetgrass/1SM
 PowerShell/SM
 WebSocket/SM

harper-core/src/lexing/hostname.rs

@@ -1,3 +1,30 @@
+use crate::TokenKind;
+
+use super::FoundToken;
+
+/// Lex a hostname token.
+pub fn lex_hostname_token(source: &[char]) -> Option<FoundToken> {
+    let len = lex_hostname(source)?;
+
+    // Might be word, just skip it.
+    if len <= 1 {
+        return None;
+    }
+
+    if !source.get(1..len - 1)?.contains(&'.') {
+        return None;
+    }
+
+    if source.get(len - 1) == Some(&'.') {
+        return None;
+    }
+
+    Some(FoundToken {
+        next_index: len,
+        token: TokenKind::Hostname,
+    })
+}
+
 pub fn lex_hostname(source: &[char]) -> Option<usize> {
     let mut passed_chars = 0;
 

harper-core/src/lexing/mod.rs

@@ -2,6 +2,7 @@ mod email_address;
 mod hostname;
 mod url;
 
+use hostname::lex_hostname_token;
 use url::lex_url;
 
 use self::email_address::lex_email_address;
@@ -27,6 +28,7 @@ pub fn lex_token(source: &[char]) -> Option<FoundToken> {
         lex_number,
         lex_url,
         lex_email_address,
+        lex_hostname_token,
         lex_word,
         lex_catch,
     ];
@@ -161,11 +163,25 @@ fn lex_catch(_source: &[char]) -> Option<FoundToken> {
 
 #[cfg(test)]
 mod tests {
+    use super::lex_token;
     use super::lex_word;
+    use super::{FoundToken, TokenKind};
 
     #[test]
     fn lexes_cjk_as_unlintable() {
         let source: Vec<_> = "世".chars().collect();
         assert!(lex_word(&source).is_none());
     }
+
+    #[test]
+    fn lexes_youtube_as_hostname() {
+        let source: Vec<_> = "YouTube.com".chars().collect();
+        assert!(matches!(
+            lex_token(&source),
+            Some(FoundToken {
+                token: TokenKind::Hostname,
+                ..
+            })
+        ));
+    }
 }

harper-core/src/linting/an_a.rs

@@ -64,7 +64,7 @@ impl Linter for AnA {
     }
 
     fn description(&self) -> &'static str {
-        "A rule that looks for incorrect indefinite articles. For example, \"this is an mule\" would be flagged as incorrect."
+        "A rule that looks for incorrect indefinite articles. For example, `this is an mule` would be flagged as incorrect."
     }
 }
 

harper-core/src/linting/compound_words.rs

@@ -51,18 +51,10 @@ impl Linter for CompoundWords {
             merged_word.extend_from_slice(b_chars);
 
             // Check for closed compound words
-            if self.dict.contains_word(&merged_word) {
-                potential_compounds.push(merged_word.clone());
-            }
-
-            // Check for hyphenated compound words
-            merged_word.clear();
-            merged_word.extend_from_slice(a_chars);
-            merged_word.push('-');
-            merged_word.extend_from_slice(b_chars);
-
-            // Check for closed compound words
-            if self.dict.contains_word(&merged_word) {
+            if self.dict.contains_word(&merged_word)
+                && !a.kind.is_common_word()
+                && !b.kind.is_common_word()
+            {
                 potential_compounds.push(merged_word.clone());
             }
 
@@ -134,15 +126,6 @@ mod tests {
         );
     }
 
-    #[test]
-    fn makeup() {
-        assert_lint_count(
-            "She spent a lot of time doing her make up this morning.",
-            CompoundWords::default(),
-            1,
-        );
-    }
-
     #[test]
     fn birthday() {
         assert_lint_count(
@@ -175,7 +158,7 @@ mod tests {
         assert_suggestion_count(
             "Like if you break up words you shouldn't.",
             CompoundWords::default(),
-            2,
+            0,
         );
     }
 }

harper-core/src/linting/correct_number_suffix.rs

@@ -33,7 +33,7 @@ impl Linter for CorrectNumberSuffix {
     }
 
     fn description(&self) -> &'static str {
-        "When making quick edits, it is common for authors to change the value of a number without changing its suffix. This rule looks for these cases, for example: \"2st\"."
+        "When making quick edits, it is common for authors to change the value of a number without changing its suffix. This rule looks for these cases, for example: `2st`."
     }
 }
 

harper-core/src/linting/linking_verbs.rs

@@ -36,7 +36,7 @@ impl Linter for LinkingVerbs {
     }
 
     fn description(&self) -> &'static str {
-        "Linking verbs connect nouns to other ideas. Make sure you do not accidentaly link words that aren't nouns."
+        "Linking verbs connect nouns to other ideas. Make sure you do not accidentally link words that aren't nouns."
     }
 }
 

harper-core/src/linting/lint_group.rs

@@ -45,6 +45,14 @@ macro_rules! create_lint_group_config {
                 pub spell_check: &'a str
             }
 
+
+            impl<'a>  LintGroupDescriptions<'a> {
+                /// Create a [`Vec`] containing the key-value pairs of this struct.
+                pub fn to_vec_pairs(self) -> Vec<(&'static str, &'a str)>{
+                    vec![$((stringify!([<$linter:snake>]), self.[<$linter:snake>],),)* ("spell_check", self.spell_check)]
+                }
+            }
+
             #[derive(Debug, Clone, Copy, Serialize, Deserialize, Default)]
             pub struct LintGroupConfig {
                 $(
@@ -175,7 +183,7 @@ create_lint_group_config!(
     AppleNames => true,
     AzureNames => true,
     CompoundWords => true,
-    PluralConjugate => true
+    PluralConjugate => false
 );
 
 impl<T: Dictionary + Default> Default for LintGroup<T> {
@@ -186,13 +194,30 @@ impl<T: Dictionary + Default> Default for LintGroup<T> {
 
 #[cfg(test)]
 mod tests {
-    use crate::FullDictionary;
+    use crate::{linting::Linter, Document, FstDictionary, FullDictionary};
 
-    use super::LintGroup;
+    use super::{LintGroup, LintGroupConfig};
 
     #[test]
     fn can_get_all_descriptions() {
         let group = LintGroup::<FullDictionary>::default();
         group.all_descriptions();
     }
+
+    #[test]
+    fn lint_descriptions_are_clean() {
+        let mut group = LintGroup::new(LintGroupConfig::default(), FstDictionary::curated());
+        let pairs: Vec<_> = group
+            .all_descriptions()
+            .to_vec_pairs()
+            .into_iter()
+            .map(|(a, b)| (a.to_string(), b.to_string()))
+            .collect();
+
+        for (key, value) in pairs {
+            let doc = Document::new_markdown_curated(&value);
+            eprintln!("{key}: {value}");
+            assert!(group.lint(&doc).is_empty())
+        }
+    }
 }

harper-core/src/linting/long_sentences.rs

@@ -27,7 +27,6 @@ impl Linter for LongSentences {
     }
 
     fn description(&self) -> &'static str {
-        "This rule looks for run-on sentences, which can make your work harder to grok.
-"
+        "This rule looks for run-on sentences, which can make your work harder to grok."
     }
 }

harper-core/src/linting/that_which.rs

@@ -58,7 +58,7 @@ impl PatternLinter for ThatWhich {
     }
 
     fn description(&self) -> &'static str {
-        "Repeating the word \"that\" twice is often redundent. \"That which\" is easier to read."
+        "Repeating the word \"that\" twice is often redundant. `That which` is easier to read."
     }
 }
 

harper-core/src/token.rs

@@ -327,6 +327,14 @@ impl TokenKind {
         metadata.is_not_plural_noun()
     }
 
+    pub fn is_common_word(&self) -> bool {
+        let TokenKind::Word(metadata) = self else {
+            return true;
+        };
+
+        metadata.common
+    }
+
     pub fn is_plural_noun(&self) -> bool {
         let TokenKind::Word(metadata) = self else {
             return false;

harper-core/tests/run_tests.rs

@@ -44,3 +44,4 @@ create_test!(chinese_lorem_ipsum.md, 2);
 create_test!(obsidian_links.md, 2);
 create_test!(issue_267.md, 0);
 create_test!(proper_noun_capitalization.md, 2);
+create_test!(amazon_hostname.md, 0);

harper-core/tests/test_sources/amazon_hostname.md

@@ -0,0 +1 @@
+This is a test of whether Amazon.com is considered a URI.

packages/harper.js/package.json

@@ -17,7 +17,7 @@
 	"scripts": {
 		"dev": "vite",
 		"build": "tsc && vite build",
-		"test": "vitest run --browser firefox && vitest run --browser chromium"
+		"test": "vitest run --browser chromium && vitest run --browser firefox"
 	},
 	"devDependencies": {
 		"wasm": "link:../../harper-wasm/pkg",