feat: render fixed-width text for nex

gemtext/src/lib.rs

@@ -12,6 +12,7 @@
 use itsybitsy_core::Error;
 use itsybitsy_core::ir::{Block, Doc, Inline};
 use itsybitsy_core::render::{RenderCtx, Rendered, Renderer};
+use itsybitsy_core::wrap::{display_width, pad};
 
 pub struct Gemtext;
 
@@ -175,7 +176,7 @@ impl Writer {
         let widths: Vec<usize> = (0..columns)
             .map(|index| {
                 all.iter()
-                    .map(|row| row.get(index).map_or(0, |c| c.chars().count()))
+                    .map(|row| row.get(index).map_or(0, |cell| display_width(cell)))
                     .max()
                     .unwrap_or(0)
             })
@@ -183,8 +184,9 @@ impl Writer {
 
         self.line("```");
         for (index, row) in all.iter().enumerate() {
+            // Padded by display width, so a wide character does not skew its column.
             let cells: Vec<String> =
-                row.iter().zip(&widths).map(|(cell, width)| format!("{cell:<width$}")).collect();
+                row.iter().zip(&widths).map(|(cell, width)| pad(cell, *width)).collect();
             self.line(&cells.join("  "));
             // A rule under the header row, matching the padded column widths.
             if index == 0 && !head.is_empty() {
@@ -274,20 +276,29 @@ fn quoted(blocks: &[Block]) -> Vec<String> {
 }
 
 #[cfg(test)]
-mod tests {
+mod tests_support {
     use itsybitsy_core::config::PageSettings;
     use itsybitsy_core::parse;
 
     use super::*;
 
     /// Render a Markdown fragment the way the server would.
-    fn render(markdown: &str) -> String {
+    pub fn render(markdown: &str) -> String {
         let doc = parse::markdown(markdown);
         let settings = PageSettings::default();
         let ctx = RenderCtx { url: "/x", title: "T", settings: &settings, width: None };
         let out = Gemtext.render(&doc, &ctx).unwrap();
         String::from_utf8(out.body).unwrap()
     }
+}
+
+#[cfg(test)]
+mod tests {
+    use itsybitsy_core::config::PageSettings;
+    use itsybitsy_core::parse;
+
+    use super::tests_support::render;
+    use super::*;
 
     #[test]
     fn headings_flatten_onto_three_levels() {
@@ -387,3 +398,22 @@ mod tests {
         }
     }
 }
+
+#[cfg(test)]
+mod width_tests {
+    use itsybitsy_core::wrap::display_width;
+
+    use super::tests_support::render;
+
+    #[test]
+    fn table_columns_align_by_display_width_not_character_count() {
+        // A CJK ideograph is two columns wide; counting characters would leave
+        // the column a column short for each one.
+        let out = render("| a | b |\n| --- | --- |\n| 日本 | x |\n");
+        let lines: Vec<&str> = out.lines().collect();
+        // Compared in columns, not bytes: a multibyte cell shifts byte offsets
+        // without shifting where the text actually appears.
+        let column = |line: &str, ch: char| display_width(&line[..line.find(ch).unwrap()]);
+        assert_eq!(column(lines[1], 'b'), column(lines[3], 'x'), "{out}");
+    }
+}