Add unicode-based grapheme width estimation method

CHANGELOG.md

@@ -16,6 +16,12 @@ Procedure when bumping the version number:
 
 ## Unreleased
 
+### Added
+
+- Unicode-based grapheme width estimation method
+  - `width_estimation_method` config option
+  - `--width-estimation-method` option
+
 ### Changed
 
 - Updated documentation for `time_zone` config option

Cargo.lock

@@ -1647,7 +1647,7 @@ dependencies = [
 [[package]]
 name = "toss"
 version = "0.3.1"
-source = "git+https://github.com/Garmelon/toss.git?tag=v0.3.1#be7eff0979e0e95d070e7c9cea42c328ffd04cc4"
+source = "git+https://github.com/Garmelon/toss.git?rev=423dd100c1360decffc5107ea4757d751ac0f4db#423dd100c1360decffc5107ea4757d751ac0f4db"
 dependencies = [
  "async-trait",
  "crossterm",

Cargo.toml

@@ -39,7 +39,7 @@ features = ["bot"]
 
 [workspace.dependencies.toss]
 git = "https://github.com/Garmelon/toss.git"
-tag = "v0.3.1"
+rev = "423dd100c1360decffc5107ea4757d751ac0f4db"
 
 [workspace.dependencies.vault]
 git = "https://github.com/Garmelon/vault.git"

cove-config/src/lib.rs

@@ -5,7 +5,7 @@ use std::{
 };
 
 use doc::Document;
-use serde::Deserialize;
+use serde::{Deserialize, Serialize};
 
 pub use crate::{euph::*, keys::*};
 
@@ -21,6 +21,14 @@ pub enum Error {
     Toml(#[from] toml::de::Error),
 }
 
+#[derive(Debug, Clone, Copy, Default, Serialize, Deserialize, Document)]
+#[serde(rename_all = "snake_case")]
+pub enum WidthEstimationMethod {
+    #[default]
+    Legacy,
+    Unicode,
+}
+
 #[derive(Debug, Default, Deserialize, Document)]
 pub struct Config {
     /// The directory that cove stores its data in when not running in ephemeral
@@ -41,12 +49,26 @@ pub struct Config {
     #[serde(default)]
     pub ephemeral: bool,
 
-    /// Whether to measure the width of characters as displayed by the terminal
+    /// How to estimate the width of graphemes (i.e. characters) as displayed by
-    /// emulator instead of guessing the width.
+    /// the terminal emulator.
+    ///
+    /// `"legacy"`: Use a legacy method that should mostly work on most terminal
+    /// emulators. This method will never be correct in all cases since every
+    /// terminal emulator handles grapheme widths slightly differently. However,
+    /// those cases are usually rare (unless you view a lot of emoji).
+    ///
+    /// `"unicode"`: Use the unicode standard in a best-effort manner to
+    /// determine grapheme widths.
+    ///
+    /// This method is used when `measure_widths` is set to `false`.
+    #[serde(default)]
+    pub width_estimation_method: WidthEstimationMethod,
+
+    /// Whether to measure the width of graphemes (i.e. characters) as displayed
+    /// by the terminal emulator instead of estimating the width.
     ///
     /// Enabling this makes rendering a bit slower but more accurate. The screen
-    /// might also flash when encountering new characters (or, more accurately,
+    /// might also flash when encountering new graphemes.
-    /// graphemes).
     ///
     /// See also the `--measure-widths` command line option.
     #[serde(default)]

cove/src/main.rs

@@ -46,6 +46,12 @@ enum Command {
     HelpConfig,
 }
 
+#[derive(Debug, Clone, Copy, clap::ValueEnum)]
+enum WidthEstimationMethod {
+    Legacy,
+    Unicode,
+}
+
 impl Default for Command {
     fn default() -> Self {
         Self::Run
@@ -79,6 +85,11 @@ struct Args {
     #[arg(long, short)]
     offline: bool,
 
+    /// Method for estimating the width of characters as displayed by the
+    /// terminal emulator.
+    #[arg(long, short)]
+    width_estimation_method: Option<WidthEstimationMethod>,
+
     /// Measure the width of characters as displayed by the terminal emulator
     /// instead of guessing the width.
     #[arg(long, short)]
@@ -114,6 +125,12 @@ fn update_config_with_args(config: &mut Config, args: &Args) {
     }
 
     config.ephemeral |= args.ephemeral;
+    if let Some(method) = args.width_estimation_method {
+        config.width_estimation_method = match method {
+            WidthEstimationMethod::Legacy => cove_config::WidthEstimationMethod::Legacy,
+            WidthEstimationMethod::Unicode => cove_config::WidthEstimationMethod::Unicode,
+        }
+    }
     config.measure_widths |= args.measure_widths;
     config.offline |= args.offline;
 }
@@ -182,6 +199,10 @@ async fn run(
 
     let mut terminal = Terminal::new()?;
     terminal.set_measuring(config.measure_widths);
+    terminal.set_width_estimation_method(match config.width_estimation_method {
+        cove_config::WidthEstimationMethod::Legacy => toss::WidthEstimationMethod::Legacy,
+        cove_config::WidthEstimationMethod::Unicode => toss::WidthEstimationMethod::Unicode,
+    });
     Ui::run(config, tz, &mut terminal, vault.clone(), logger, logger_rx).await?;
     drop(terminal);