From 4e4f136ec6ec0c849d4ba8ac37b313d126141911 Mon Sep 17 00:00:00 2001
From: SeongChan Lee <foriequal@gmail.com>
Date: Mon, 25 Apr 2022 11:11:51 +0900
Subject: [PATCH] Fix tokenizer for uppercase UUID

Microsoft toolchains usually emit UUID/GUID in UPPERCASE
---
 crates/typos/src/tokens.rs | 48 ++++++++++++++++++++++++++++++--------
 1 file changed, 38 insertions(+), 10 deletions(-)

diff --git a/crates/typos/src/tokens.rs b/crates/typos/src/tokens.rs
index 83241ae8..734f3a5b 100644
--- a/crates/typos/src/tokens.rs
+++ b/crates/typos/src/tokens.rs
@@ -319,16 +319,29 @@ mod parser {
         <T as nom::InputTakeAtPosition>::Item: AsChar + Copy,
         <T as nom::InputIter>::Item: AsChar + Copy,
     {
-        recognize(tuple((
-            take_while_m_n(8, 8, is_lower_hex_digit),
-            char('-'),
-            take_while_m_n(4, 4, is_lower_hex_digit),
-            char('-'),
-            take_while_m_n(4, 4, is_lower_hex_digit),
-            char('-'),
-            take_while_m_n(4, 4, is_lower_hex_digit),
-            char('-'),
-            take_while_m_n(12, 12, is_lower_hex_digit),
+        recognize(alt((
+            tuple((
+                take_while_m_n(8, 8, is_lower_hex_digit),
+                char('-'),
+                take_while_m_n(4, 4, is_lower_hex_digit),
+                char('-'),
+                take_while_m_n(4, 4, is_lower_hex_digit),
+                char('-'),
+                take_while_m_n(4, 4, is_lower_hex_digit),
+                char('-'),
+                take_while_m_n(12, 12, is_lower_hex_digit),
+            )),
+            tuple((
+                take_while_m_n(8, 8, is_upper_hex_digit),
+                char('-'),
+                take_while_m_n(4, 4, is_upper_hex_digit),
+                char('-'),
+                take_while_m_n(4, 4, is_upper_hex_digit),
+                char('-'),
+                take_while_m_n(4, 4, is_upper_hex_digit),
+                char('-'),
+                take_while_m_n(12, 12, is_upper_hex_digit),
+            )),
         )))(input)
     }
 
@@ -1087,6 +1100,21 @@ mod test {
         assert_eq!(expected, actual);
     }
 
+    #[test]
+    fn tokenize_ignore_uuid_uppercase() {
+        let parser = TokenizerBuilder::new().build();
+
+        let input = "Hello 123E4567-E89B-12D3-A456-426652340000 World";
+        let expected: Vec<Identifier> = vec![
+            Identifier::new_unchecked("Hello", Case::None, 0),
+            Identifier::new_unchecked("World", Case::None, 43),
+        ];
+        let actual: Vec<_> = parser.parse_bytes(input.as_bytes()).collect();
+        assert_eq!(expected, actual);
+        let actual: Vec<_> = parser.parse_str(input).collect();
+        assert_eq!(expected, actual);
+    }
+
     #[test]
     fn tokenize_ignore_hash() {
         let parser = TokenizerBuilder::new().build();