IDNA support

Cargo.toml

@@ -1,7 +1,7 @@
 [package]
 
 name = "url"
-version = "0.5.1"
+version = "0.5.2"
 authors = [ "Simon Sapin <simon.sapin@exyr.org>" ]
 
 description = "URL library for Rust, based on the WHATWG URL Standard"
@@ -35,4 +35,6 @@ optional = true
 [dependencies]
 uuid = "0.1.17"
 rustc-serialize = "0.3"
+unicode-bidi = "0.2.3"
+unicode-normalization = "0.1.1"
 matches = "0.1"

make_idna_table.py

@@ -0,0 +1,60 @@
+# Copyright 2013-2014 Valentin Gosu.
+#
+# Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or
+# http://www.apache.org/licenses/LICENSE-2.0> or the MIT license
+# <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your
+# option. This file may not be copied, modified, or distributed
+# except according to those terms.
+
+
+# Run as: python make_idna_table.py idna_table.txt > src/idna_table.rs
+# You can get the latest idna table from
+# http://www.unicode.org/Public/idna/latest/IdnaMappingTable.txt
+
+print('''\
+// Copyright 2013-2014 Valentin Gosu.
+//
+// Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or
+// http://www.apache.org/licenses/LICENSE-2.0> or the MIT license
+// <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your
+// option. This file may not be copied, modified, or distributed
+// except according to those terms.
+
+// Generated by make_idna_table.py
+
+use idna::Mapping::*;
+use idna::Range;
+
+pub static TABLE: &'static [Range] = &[
+''')
+
+txt = open("IdnaMappingTable.txt")
+
+def char(s):
+    return (unichr(int(s, 16))
+        .encode('utf8')
+        .replace('\\', '\\\\')
+        .replace('"', '\\"')
+        .replace('\0', '\\0'))
+
+for line in txt:
+    # remove comments
+    line, _, _ = line.partition('#')
+    # skip empty lines
+    if len(line.strip()) == 0:
+        continue
+    fields = line.split(';')
+    if fields[0].strip() == 'D800..DFFF':
+        continue  # Surrogates don't occur in Rust strings.
+    first, _, last = fields[0].strip().partition('..')
+    if not last:
+        last = first
+    mapping = fields[1].strip().replace('_', ' ').title().replace(' ', '')
+    if len(fields) > 2:
+        if fields[2].strip():
+            mapping += '("%s")' % ''.join(char(c) for c in fields[2].strip().split(' '))
+        elif mapping == "Deviation":
+            mapping += '("")'
+    print("    Range { from: '%s', to: '%s', mapping: %s }," % (char(first), char(last), mapping))
+
+print("];")

src/host.rs

@@ -12,6 +12,7 @@ use std::fmt::{self, Formatter};
 use std::net::{Ipv4Addr, Ipv6Addr};
 use parser::{ParseResult, ParseError};
 use percent_encoding::{from_hex, percent_decode};
+use idna;
 
 
 /// The host name of an URL.
@@ -34,8 +35,6 @@ impl Host {
     ///
     /// Returns `Err` for an empty host, an invalid IPv6 address,
     /// or a or invalid non-ASCII domain.
-    ///
-    /// FIXME: Add IDNA support for non-ASCII domains.
     pub fn parse(input: &str) -> ParseResult<Host> {
         if input.len() == 0 {
             return Err(ParseError::EmptyHost)
@@ -48,10 +47,13 @@ impl Host {
         }
         let decoded = percent_decode(input.as_bytes());
         let domain = String::from_utf8_lossy(&decoded);
-        // TODO: Remove this check and use IDNA "domain to ASCII"
-        if !domain.is_ascii() {
-            return Err(ParseError::NonAsciiDomainsNotSupportedYet)
-        } else if domain.find(&[
+
+        let domain = match idna::domain_to_ascii(&domain) {
+            Ok(s) => s,
+            Err(_) => return Err(ParseError::InvalidDomainCharacter)
+        };
+
+        if domain.find(&[
             '\0', '\t', '\n', '\r', ' ', '#', '%', '/', ':', '?', '@', '[', '\\', ']'
         ][..]).is_some() {
             return Err(ParseError::InvalidDomainCharacter)