Support ANSI-escaped color codes and hyperlinks.

Cargo.toml

@@ -42,3 +42,4 @@ lazy_static = "1.4"
 is-terminal = "0.4"
 encode_unicode = "1.0"
 csv = { version = "1.1", optional = true }
+regex = "1"

README.md

@@ -249,6 +249,14 @@ Uppercase letters stand for **bright** counterparts of the above colors:
 * **B** : Bright Blue
 * ... and so on ...
 
+## ANSI hyperlinks
+
+In most modern terminal emulators, it is possible to embed hyperlinks using ANSI escape codes. The following string field would display as a clickable link:
+
+```rust
+"\u{1b}]8;;http://example.com\u{1b}\\example.com\u{1b}]8;;\u{1b}\\"
+```
+
 ## Slicing
 
 Tables can be sliced into immutable borrowed subtables.

src/utils.rs

@@ -3,6 +3,7 @@ use std::fmt;
 use std::io::{Error, ErrorKind, Write};
 use std::str;
 
+use regex::Regex;
 use unicode_width::{UnicodeWidthChar, UnicodeWidthStr};
 
 use super::format::Alignment;
@@ -82,45 +83,20 @@ pub fn print_align<T: Write + ?Sized>(
 }
 
 /// Return the display width of a unicode string.
-/// This functions takes ANSI-escaped color codes into account.
+/// This functions takes ANSI-escaped color codes and hyperlinks into account.
 pub fn display_width(text: &str) -> usize {
-    #[derive(PartialEq, Eq, Clone, Copy)]
-    enum State {
-        /// We are not inside any terminal escape.
-        Normal,
-        /// We have just seen a \u{1b}
-        EscapeChar,
-        /// We have just seen a [
-        OpenBracket,
-        /// We just ended the escape by seeing an m
-        AfterEscape,
-    }
-
     let width = UnicodeWidthStr::width(text);
-    let mut state = State::Normal;
     let mut hidden = 0;
 
-    for c in text.chars() {
-        state = match (state, c) {
-            (State::Normal, '\u{1b}') => State::EscapeChar,
-            (State::EscapeChar, '[') => State::OpenBracket,
-            (State::EscapeChar, _) => State::Normal,
-            (State::OpenBracket, 'm') => State::AfterEscape,
-            _ => state,
-        };
-
-        // We don't count escape characters as hidden as
-        // UnicodeWidthStr::width already considers them.
-        if matches!(state, State::OpenBracket | State::AfterEscape) {
-            // but if we see an escape char *inside* the ANSI escape, we should ignore it.
-            if UnicodeWidthChar::width(c).unwrap_or(0) > 0 {
-                hidden += 1;
-            }
-        }
-
-        if state == State::AfterEscape {
-            state = State::Normal;
-        }
+    lazy_static! {
+        static ref COLOR_RE: Regex = Regex::new(r"\u{1b}(?P<colors>\[[^m]+?)m").unwrap();
+        static ref HYPERLINK_RE: Regex = Regex::new(r"\u{1b}]8;;(?P<url>[^\u{1b}]+?)\u{1b}\\(?P<text>[^\u{1b}]+?)\u{1b}]8;;\u{1b}\\").unwrap();
+    }
+    for caps in COLOR_RE.captures_iter(text) {
+        hidden += UnicodeWidthStr::width(&caps["colors"])
+    }
+    for caps in HYPERLINK_RE.captures_iter(text) {
+        hidden += 10 + UnicodeWidthStr::width(&caps["url"])
     }
 
     assert!(
@@ -225,6 +201,17 @@ mod tests {
         assert_eq!(out.as_string(), "foo");
     }
 
+    #[test]
+    fn ansi_escapes() {
+        let mut out = StringWriter::new();
+        print_align(&mut out, Alignment::LEFT, "\u{1b}[31;40mred\u{1b}[0m", ' ', 10, false).unwrap();
+        assert_eq!(display_width(out.as_string()), 10);
+
+        let mut out = StringWriter::new();
+        print_align(&mut out, Alignment::LEFT, "\u{1b}]8;;http://example.com\u{1b}\\example\u{1b}]8;;\u{1b}\\", ' ', 10, false).unwrap();
+        assert_eq!(display_width(out.as_string()), 10);
+    }
+
     #[test]
     fn utf8_error() {
         let mut out = StringWriter::new();