Auto merge of #6614 - servo:replace-surrogates, r=Ms2ger

Add a `-Z replace-surrogates` command-line option.

See #6564.

r? @Ms2ger

<!-- Reviewable:start -->
[<img src="https://reviewable.io/review_button.png" height=40 alt="Review on Reviewable"/>](https://reviewable.io/reviews/servo/servo/6614)
<!-- Reviewable:end -->

components/script/dom/bindings/conversions.rs

@@ -352,10 +352,34 @@ pub fn jsstring_to_str(cx: *mut JSContext, s: *mut JSString) -> DOMString {
             JS_GetTwoByteStringCharsAndLength(cx, ptr::null(), s, &mut length)
         };
         assert!(!chars.is_null());
-        let char_vec = unsafe {
+        let potentially_ill_formed_utf16 = unsafe {
             slice::from_raw_parts(chars as *const u16, length as usize)
         };
-        String::from_utf16(char_vec).unwrap()
+        let mut s = String::with_capacity(length as usize);
+        for item in ::rustc_unicode::str::utf16_items(potentially_ill_formed_utf16) {
+            use ::rustc_unicode::str::Utf16Item::*;
+            match item {
+                ScalarValue(c) => s.push(c),
+                LoneSurrogate(_) => {
+                    // FIXME: Add more info like document URL in the message?
+                    macro_rules! message {
+                        () => {
+                            "Found an unpaired surrogate in a DOM string. \
+                             If you see this in real web content, \
+                             please comment on https://github.com/servo/servo/issues/6564"
+                        }
+                    }
+                    if ::util::opts::get().replace_surrogates {
+                        error!(message!());
+                        s.push('\u{FFFD}');
+                    } else {
+                        panic!(concat!(message!(), " Use `-Z replace-surrogates` \
+                            on the command line to make this non-fatal."));
+                    }
+                }
+            }
+        }
+        s
     }
 }
 

components/script/lib.rs

@@ -21,6 +21,7 @@
 #![feature(rc_unique)]
 #![feature(slice_chars)]
 #![feature(str_utf16)]
+#![feature(unicode)]
 #![feature(vec_push_all)]
 
 #![deny(unsafe_code)]
@@ -49,6 +50,7 @@ extern crate msg;
 extern crate net_traits;
 extern crate num;
 extern crate rustc_serialize;
+extern crate rustc_unicode;
 extern crate time;
 extern crate canvas;
 extern crate canvas_traits;

components/util/opts.rs

@@ -69,6 +69,11 @@ pub struct Opts {
     pub userscripts: Option<String>,
 
     pub output_file: Option<String>,
+
+    /// Replace unpaires surrogates in DOM strings with U+FFFD.
+    /// See https://github.com/servo/servo/issues/6564
+    pub replace_surrogates: bool,
+
     pub headless: bool,
     pub hard_fail: bool,
 
@@ -187,6 +192,8 @@ pub fn print_debug_usage(app: &str) -> ! {
     print_option("disable-share-style-cache",
                  "Disable the style sharing cache.");
     print_option("parallel-display-list-building", "Build display lists in parallel.");
+    print_option("replace-surrogates", "Replace unpaires surrogates in DOM strings with U+FFFD. \
+                                        See https://github.com/servo/servo/issues/6564");
 
     println!("");
 
@@ -223,6 +230,7 @@ pub fn default_opts() -> Opts {
         nossl: false,
         userscripts: None,
         output_file: None,
+        replace_surrogates: false,
         headless: true,
         hard_fail: true,
         bubble_inline_sizes_separately: false,
@@ -397,6 +405,7 @@ pub fn from_cmdline_args(args: &[String]) {
         nossl: nossl,
         userscripts: opt_match.opt_default("userscripts", ""),
         output_file: opt_match.opt_str("o"),
+        replace_surrogates: debug_options.contains(&"replace-surrogates"),
         headless: opt_match.opt_present("z"),
         hard_fail: opt_match.opt_present("f"),
         bubble_inline_sizes_separately: bubble_inline_sizes_separately,