Merge pull request #469 from foriequal0/column-number

Change column number to 1-based index, and fix indenting for emojis and tabs

Cargo.lock

@@ -1574,6 +1574,7 @@ dependencies = [
  "typos",
  "typos-dict",
  "typos-vars",
+ "unic-emoji-char",
  "unicase",
  "unicode-segmentation",
  "unicode-width",
@@ -1642,6 +1643,47 @@ dependencies = [
  "varcon-core",
 ]
 
+[[package]]
+name = "unic-char-property"
+version = "0.9.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "a8c57a407d9b6fa02b4795eb81c5b6652060a15a7903ea981f3d723e6c0be221"
+dependencies = [
+ "unic-char-range",
+]
+
+[[package]]
+name = "unic-char-range"
+version = "0.9.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "0398022d5f700414f6b899e10b8348231abf9173fa93144cbc1a43b9793c1fbc"
+
+[[package]]
+name = "unic-common"
+version = "0.9.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "80d7ff825a6a654ee85a63e80f92f054f904f21e7d12da4e22f9834a4aaa35bc"
+
+[[package]]
+name = "unic-emoji-char"
+version = "0.9.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "0b07221e68897210270a38bde4babb655869637af0f69407f96053a34f76494d"
+dependencies = [
+ "unic-char-property",
+ "unic-char-range",
+ "unic-ucd-version",
+]
+
+[[package]]
+name = "unic-ucd-version"
+version = "0.9.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "96bd2f2237fe450fcd0a1d2f5f4e91711124f7857ba2e964247776ebeeb7b0c4"
+dependencies = [
+ "unic-common",
+]
+
 [[package]]
 name = "unicase"
 version = "2.6.0"

Cargo.toml

@@ -92,6 +92,7 @@ kstring = { version = "2.0.0", features = ["serde"] }
 typed-arena = "2.0.1"
 maplit = "1.0"
 unicode-width = "0.1.9"
+unic-emoji-char = "0.9.0"
 
 [dev-dependencies]
 assert_fs = "1.0"

src/bin/typos-cli/report.rs

@@ -181,7 +181,8 @@ 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 start = String::from_utf8_lossy(&msg.buffer[0..msg.byte_offset]);
-    let column = unicode_segmentation::UnicodeSegmentation::graphemes(start.as_ref(), true).count();
+    let column_number =
+        unicode_segmentation::UnicodeSegmentation::graphemes(start.as_ref(), true).count() + 1;
     match &msg.corrections {
         typos::Status::Valid => {}
         typos::Status::Invalid => {
@@ -213,15 +214,15 @@ fn print_long_correction(msg: &Typo, palette: Palette) -> Result<(), std::io::Er
         "  --> {}{}{}",
         palette.info.paint(context_display(&msg.context)),
         palette.info.paint(divider),
-        palette.info.paint(column)
+        palette.info.paint(column_number)
     )?;
 
     if let Some(Context::File(context)) = &msg.context {
         let line_num = context.line_num.to_string();
         let line_indent: String = itertools::repeat_n(" ", line_num.len()).collect();
 
-        let visible_column = UnicodeWidthStr::width(start.as_ref());
+        let visible_column = calculate_visible_column_width(start.as_ref());
-        let visible_len = UnicodeWidthStr::width(msg.typo);
+        let visible_len = calculate_visible_column_width(msg.typo);
 
         let hl_indent: String = itertools::repeat_n(" ", visible_column).collect();
         let hl: String = itertools::repeat_n("^", visible_len).collect();
@@ -246,6 +247,27 @@ fn print_long_correction(msg: &Typo, palette: Palette) -> Result<(), std::io::Er
     Ok(())
 }
 
+fn calculate_visible_column_width(str: &str) -> usize {
+    let mut result = 0;
+    let graphemes = unicode_segmentation::UnicodeSegmentation::graphemes(str, true);
+    for grapheme in graphemes {
+        result += if grapheme == "\t" {
+            // TODO: config tab width
+            1
+        } else if grapheme.chars().any(unic_emoji_char::is_emoji) {
+            // UnicodeWidthStr::width doesn't cover for emoji according to their README.
+            // See: https://github.com/unicode-rs/unicode-width#unicode-width
+            // Also, the actual rendered column width may differ from calculation, especially for emojis.
+            // In here, we expect emoji renderers should render this emoji properly.
+            2
+        } else {
+            UnicodeWidthStr::width(grapheme)
+        }
+    }
+
+    result
+}
+
 fn context_display<'c>(context: &'c Option<Context<'c>>) -> &'c dyn std::fmt::Display {
     context
         .as_ref()

tests/cmd/stdin-failure-multiwidth.stdout

@@ -1,29 +1,29 @@
 error: `Apropriate` should be `Appropriate`
-  --> -:5:2
+  --> -:5:3
   |
 5 | 한 Apropriate world
   |    ^^^^^^^^^^
   |
 error: `Apropriate` should be `Appropriate`
-  --> -:12:2
+  --> -:12:3
    |
 12 | 한 Apropriate world
    |    ^^^^^^^^^^
    |
 error: `Apropriate` should be `Appropriate`
-  --> -:19:2
+  --> -:19:3
    |
 19 | 👁️‍🗨️ Apropriate world
    |    ^^^^^^^^^^
    |
 error: `Apropriate` should be `Appropriate`
-  --> -:26:2
+  --> -:26:3
    |
 26 | 😵‍💫 Apropriate world
    |    ^^^^^^^^^^
    |
 error: `Apropriate` should be `Appropriate`
-  --> -:33:1
+  --> -:33:2
    |
 33 |  Apropriate world
    |  ^^^^^^^^^^