From 9cae074ff475186f494239499533c17b695766d7 Mon Sep 17 00:00:00 2001
From: Yorick Peterse <yorick@yorickpeterse.com>
Date: Sat, 21 Dec 2024 00:45:23 +0100
Subject: [PATCH] Wrap lines using the character width not count

Instead of counting the number of grapheme clusters per line, we now
count the character _widths_. This ensures that lines containing double
width characters are wrapped correctly.

Changelog: fixed
---
 Cargo.lock                                           | 10 ++++++++--
 compiler/Cargo.toml                                  |  2 +-
 compiler/src/format.rs                               |  6 +++---
 std/fixtures/fmt/double_width_characters/input.inko  |  1 +
 std/fixtures/fmt/double_width_characters/output.inko |  9 +++++++++
 5 files changed, 22 insertions(+), 6 deletions(-)
 create mode 100644 std/fixtures/fmt/double_width_characters/input.inko
 create mode 100644 std/fixtures/fmt/double_width_characters/output.inko

diff --git a/Cargo.lock b/Cargo.lock
index b287f4c3..328d298e 100644
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -137,7 +137,7 @@ dependencies = [
  "location",
  "similar-asserts",
  "types",
- "unicode-segmentation",
+ "unicode-width 0.2.0",
 ]
 
 [[package]]
@@ -280,7 +280,7 @@ version = "0.2.21"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "14dbbfd5c71d70241ecf9e6f13737f7b5ce823821063188d7e46c41d371eebd5"
 dependencies = [
- "unicode-width",
+ "unicode-width 0.1.14",
 ]
 
 [[package]]
@@ -1007,6 +1007,12 @@ version = "0.1.14"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "7dd6e30e90baa6f72411720665d41d89b9a3d039dc45b8faea1ddd07f617f6af"
 
+[[package]]
+name = "unicode-width"
+version = "0.2.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "1fc81956842c57dac11422a97c3b8195a1ff727f06e85c84ed2e8aa277c9a0fd"
+
 [[package]]
 name = "untrusted"
 version = "0.9.0"
diff --git a/compiler/Cargo.toml b/compiler/Cargo.toml
index 877e5ff8..b2c7c278 100644
--- a/compiler/Cargo.toml
+++ b/compiler/Cargo.toml
@@ -14,7 +14,7 @@ doctest = false
 ast = { path = "../ast" }
 types = { path = "../types" }
 location = { path = "../location" }
-unicode-segmentation = "^1.10"
+unicode-width = "^0.2"
 getopts = "^0.2"
 fnv = "^1.0"
 blake3 = "^1.5"
diff --git a/compiler/src/format.rs b/compiler/src/format.rs
index 30e122ed..dd3d6524 100644
--- a/compiler/src/format.rs
+++ b/compiler/src/format.rs
@@ -12,7 +12,7 @@ use std::collections::HashSet;
 use std::fs::{read, write};
 use std::io::{stdin, stdout, Error as IoError, Read as _, Write as _};
 use std::path::PathBuf;
-use unicode_segmentation::UnicodeSegmentation as _;
+use unicode_width::UnicodeWidthStr;
 
 /// The characters to use for indentation.
 const INDENT: char = ' ';
@@ -79,7 +79,7 @@ enum Node {
     /// The arguments are:
     ///
     /// 1. The text that may include Unicode characters
-    /// 2. The number of grapheme clusters in the string
+    /// 2. The width (in cells) of the string.
     Unicode(String, usize),
 
     /// A node to include if the code is to be wrapped across lines.
@@ -149,7 +149,7 @@ impl Node {
     }
 
     fn unicode(text: String) -> Node {
-        let width = text.graphemes(true).count();
+        let width = text.width();
 
         Node::Unicode(text, width)
     }
diff --git a/std/fixtures/fmt/double_width_characters/input.inko b/std/fixtures/fmt/double_width_characters/input.inko
new file mode 100644
index 00000000..bb03b8e1
--- /dev/null
+++ b/std/fixtures/fmt/double_width_characters/input.inko
@@ -0,0 +1 @@
+let FULL_WEEKDAYS = ['日曜日', '月曜日', '火曜日', '水曜日', '木曜日', '金曜日', '土曜日']
diff --git a/std/fixtures/fmt/double_width_characters/output.inko b/std/fixtures/fmt/double_width_characters/output.inko
new file mode 100644
index 00000000..ab142bb1
--- /dev/null
+++ b/std/fixtures/fmt/double_width_characters/output.inko
@@ -0,0 +1,9 @@
+let FULL_WEEKDAYS = [
+  '日曜日',
+  '月曜日',
+  '火曜日',
+  '水曜日',
+  '木曜日',
+  '金曜日',
+  '土曜日',
+]