From 3d1507c3cf11de3a2f5dbf8571aa9c6c6020b529 Mon Sep 17 00:00:00 2001
From: Alexandre Bury <alexandre.bury@gmail.com>
Date: Wed, 27 May 2015 22:13:51 -0700
Subject: [PATCH] Add UTF-8 input support

Didn't the proper methods in the standard lib, so I implemented a simple
utf-8 char reader.
---
 examples/key_codes.rs |  6 +++---
 src/lib.rs            |  5 +++--
 src/utf8.rs           | 50 +++++++++++++++++++++++++++++++++++++++++++
 3 files changed, 56 insertions(+), 5 deletions(-)
 create mode 100644 src/utf8.rs

diff --git a/examples/key_codes.rs b/examples/key_codes.rs
index eeeed3d..61cdd30 100644
--- a/examples/key_codes.rs
+++ b/examples/key_codes.rs
@@ -4,12 +4,12 @@ use cursive::Cursive;
 
 use cursive::view::{View,BoxView};
 use cursive::printer::Printer;
-use cursive::event::EventResult;
+use cursive::event::{EventResult,Event};
 
 fn main() {
     let mut siv = Cursive::new();
 
-    siv.add_layer(BoxView::new((10,4), KeyCodeView::new(4)));
+    siv.add_layer(BoxView::new((30,10), KeyCodeView::new(10)));
 
     siv.run();
 }
@@ -39,7 +39,7 @@ impl View for KeyCodeView {
         let line = match event {
             Event::CharEvent(c) => format!("Char: {}", c),
             Event::KeyEvent(key) => format!("Key: {}", key),
-        }
+        };
         self.history.push(line);
 
         while self.history.len() > self.size {
diff --git a/src/lib.rs b/src/lib.rs
index 7ffc072..c066761 100644
--- a/src/lib.rs
+++ b/src/lib.rs
@@ -30,6 +30,7 @@ pub mod vec;
 pub mod color;
 
 mod div;
+mod utf8;
 
 use std::any::Any;
 use std::rc::Rc;
@@ -201,8 +202,8 @@ impl Cursive {
         let ch = ncurses::getch();
 
         // Is it a UTF-8 starting point?
-        if 32 <= ch && ch < 127 {
-            Event::CharEvent(ch as u8 as char)
+        if 32 <= ch && ch < 0x100 {
+            Event::CharEvent(utf8::read_char(ch as u8, || ncurses::getch() as u8).unwrap())
         } else {
             Event::KeyEvent(Key::from_ncurses(ch))
         }
diff --git a/src/utf8.rs b/src/utf8.rs
new file mode 100644
index 0000000..2f1e9a9
--- /dev/null
+++ b/src/utf8.rs
@@ -0,0 +1,50 @@
+use std::char::from_u32;
+
+/// Reads a char from a first byte, and a function to fetch next bytes as required.
+///
+/// Returns an error if the stream is invalid utf-8.
+pub fn read_char<F>(first: u8, next: F) -> Result<char,String>
+    where F: Fn() -> u8
+{
+    if first < 0x80 {
+        return Ok(first as char)
+    }
+
+    // Number of leading 1s determines the number of bytes we'll have to read
+    let n_bytes = match (!first).leading_zeros() {
+        n @ 2 ... 6 => n as usize,
+        1 => return Err("First byte is continuation byte.".to_string()),
+        7 ... 8 => return Err("WTF is this byte??".to_string()),
+        _ => unreachable!(),
+    };
+
+    let mut res = 0u32;
+
+    // First, get the data - only the few last bits
+    res |= (first & make_mask(7 - n_bytes)) as u32;
+
+    // We already have one byte, now read the others.
+    for _ in 1..n_bytes {
+        let byte = next();
+        if byte & 0xC0 != 0x80 {
+            return Err(format!("Found non-continuation byte after leading: {}", byte));
+        }
+        // We have 6 fresh new bits to read, make room.
+        res <<= 6;
+        // 0x3F is 00111111, so we keep the last 6 bits
+        res |= (byte & 0x3F) as u32;
+    }
+
+    // from_u32 could return an error if we gave it invalid utf-8.
+    // But we're probably safe since we respected the rules when building it.
+    Ok(from_u32(res).unwrap())
+}
+
+// Returns a simple bitmask with n 1s to the right.
+fn make_mask(n: usize) -> u8 {
+    let mut r = 0u8;
+    for i in 0..n {
+        r |= 1 << i;
+    }
+    r
+}