From 9859e6007715ac0ef9d7c1b970c3e2041b685120 Mon Sep 17 00:00:00 2001
From: Ed Page <epage@duosecurity.com>
Date: Thu, 27 May 2021 13:10:42 -0500
Subject: [PATCH] fix(cli): Don't panic when finding column

When rendering typos, we look up what visual column the typoe starts on
but I mixed a raw byte offset with the offset into a lossy string.  This
caused panics when dealing with non-ascii content.

Fixes #258
---
 CHANGELOG.md                |  4 +++
 src/bin/typos-cli/report.rs |  7 ++---
 src/file.rs                 | 62 +++++++++++++++++++++++++++++++------
 3 files changed, 58 insertions(+), 15 deletions(-)

diff --git a/CHANGELOG.md b/CHANGELOG.md
index 0d0ad28..aa210cf 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -7,6 +7,10 @@ and this project adheres to [Semantic Versioning](http://semver.org/).
 <!-- next-header -->
 ## [Unreleased] - ReleaseDate
 
+#### Bug Fixes
+
+- Don't panic when rendering typos on lines with non-ASCII character
+
 ## [1.0.1] - 2021-05-27
 
 #### Bug Fixes
diff --git a/src/bin/typos-cli/report.rs b/src/bin/typos-cli/report.rs
index 53b337b..82c1e67 100644
--- a/src/bin/typos-cli/report.rs
+++ b/src/bin/typos-cli/report.rs
@@ -190,11 +190,8 @@ fn print_long_correction(msg: &Typo, palette: Palette) -> Result<(), std::io::Er
 
     let line = String::from_utf8_lossy(msg.buffer.as_ref());
     let line = line.replace("\t", " ");
-    let column = unicode_segmentation::UnicodeSegmentation::graphemes(
-        line.get(0..msg.byte_offset).unwrap(),
-        true,
-    )
-    .count();
+    let start = String::from_utf8_lossy(&msg.buffer[0..msg.byte_offset]);
+    let column = unicode_segmentation::UnicodeSegmentation::graphemes(start.as_ref(), true).count();
     match &msg.corrections {
         typos::Status::Valid => {}
         typos::Status::Invalid => {
diff --git a/src/file.rs b/src/file.rs
index 306f11c..54157f9 100644
--- a/src/file.rs
+++ b/src/file.rs
@@ -702,36 +702,78 @@ mod test {
 
     #[test]
     fn test_extract_line_single_line() {
-        let (line, offset) = extract_line(b"hello world", 6);
-        assert_eq!(line, b"hello world");
+        let buffer = b"hello world";
+        let buffer_offset = 6;
+        let expected_line = b"hello world";
+        let (line, offset) = extract_line(buffer, buffer_offset);
+        assert_eq!(line, expected_line);
         assert_eq!(offset, 6);
+        assert_eq!(line[offset], buffer[buffer_offset]);
     }
 
     #[test]
     fn test_extract_line_first() {
-        let (line, offset) = extract_line(b"1\n2\n3", 0);
-        assert_eq!(line, b"1");
+        let buffer = b"1\n2\n3";
+        let buffer_offset = 0;
+        let expected_line = b"1";
+        let (line, offset) = extract_line(buffer, buffer_offset);
+        assert_eq!(line, expected_line);
         assert_eq!(offset, 0);
+        assert_eq!(line[offset], buffer[buffer_offset]);
     }
 
     #[test]
     fn test_extract_line_middle() {
-        let (line, offset) = extract_line(b"1\n2\n3", 2);
-        assert_eq!(line, b"2");
+        let buffer = b"1\n2\n3";
+        let buffer_offset = 2;
+        let expected_line = b"2";
+        let (line, offset) = extract_line(buffer, buffer_offset);
+        assert_eq!(line, expected_line);
         assert_eq!(offset, 0);
+        assert_eq!(line[offset], buffer[buffer_offset]);
     }
 
     #[test]
     fn test_extract_line_end() {
-        let (line, offset) = extract_line(b"1\n2\n3", 4);
-        assert_eq!(line, b"3");
+        let buffer = b"1\n2\n3";
+        let buffer_offset = 4;
+        let expected_line = b"3";
+        let (line, offset) = extract_line(buffer, buffer_offset);
+        assert_eq!(line, expected_line);
         assert_eq!(offset, 0);
+        assert_eq!(line[offset], buffer[buffer_offset]);
     }
 
     #[test]
     fn test_extract_line_offset_change() {
-        let (line, offset) = extract_line(b"1\nhello world\n2", 8);
-        assert_eq!(line, b"hello world");
+        let buffer = b"1\nhello world\n2";
+        let buffer_offset = 8;
+        let expected_line = b"hello world";
+        let (line, offset) = extract_line(buffer, buffer_offset);
+        assert_eq!(line, expected_line);
         assert_eq!(offset, 6);
+        assert_eq!(line[offset], buffer[buffer_offset]);
+    }
+
+    #[test]
+    fn test_extract_line_windows() {
+        let buffer = b"1\r\nhello world\r\n2";
+        let buffer_offset = 9;
+        let expected_line = b"hello world";
+        let (line, offset) = extract_line(buffer, buffer_offset);
+        assert_eq!(line, expected_line);
+        assert_eq!(offset, 6);
+        assert_eq!(line[offset], buffer[buffer_offset]);
+    }
+
+    #[test]
+    fn test_extract_line_slovak() {
+        let buffer = b"LastErrorMessage=%1.%n%nChyba %2: %3\r\nSetupFileMissing=In\x9Atala\xE8n\xFD adres\xE1r neobsahuje s\xFAbor %1. Opravte, pros\xEDm, t\xFAto chybu alebo si zaobstarajte nov\xFA k\xF3piu tohto produktu.\r\nSetupFileCorrupt=S\xFAbory sprievodcu in\x9Atal\xE1ciou s\xFA po\x9Akoden\xE9. Zaobstarajte si, pros\xEDm, nov\xFA k\xF3piu tohto produktu.";
+        let buffer_offset = 66;
+        let expected_line = b"SetupFileMissing=In\x9Atala\xE8n\xFD adres\xE1r neobsahuje s\xFAbor %1. Opravte, pros\xEDm, t\xFAto chybu alebo si zaobstarajte nov\xFA k\xF3piu tohto produktu.";
+        let (line, offset) = extract_line(buffer, buffer_offset);
+        assert_eq!(line, expected_line);
+        assert_eq!(offset, 28);
+        assert_eq!(line[offset], buffer[buffer_offset]);
     }
 }