html Incorrect spans generated for HTML with higher-plane unicode characters

When parsing HTML that includes characters like "🍋", the start and end FileLocations are generated incorrectly.

Here's a short repo:

import 'package:html/dom.dart';
import "package:html/parser.dart";
import "package:source_span/source_span.dart";

void main() {
  final dom = parse(contents,generateSpans: true);
  final Element element = dom.querySelectorAll("link").single;
  final span = element.sourceSpan;
  final spanCopy = new SourceSpan(span.start, span.end, contents);
}

const contents = """
<head>
    <meta charset="UTF-8">
    <title></title>
    <link rel="alternate" type="application/rss+xml" title="ArtLung &raquo; Limones 🍋 Comments Feed" href="subdirectory/other.html" />
</head>
""";

This will throw the following error:

Unhandled exception:
Invalid argument(s): Text "<head>
    <meta charset="UTF-8">
    <title></title>
    <link rel="alternate" type="application/rss+xml" title="ArtLung &raquo; Limones 🍋 Comments Feed" href="subdirectory/other.html" />
</head>
" must be 130 characters long.
#0      new SourceSpanBase (package:source_span/src/span.dart:85:7)
#1      new SourceSpan (package:source_span/src/span.dart:34:11)
#2      main (file:///Users/filiph/dev/linkcheck/test/source_span_bug.dart:9:24)
#3      _startIsolate.<anonymous closure> (dart:isolate-patch/isolate_patch.dart:265)
#4      _RawReceivePortImpl._handleMessage (dart:isolate-patch/isolate_patch.dart:151)

This is not an issue with package:source_span — when I create the span manually, without parse(), copying it works okay.

Mar 21 '18 23:03 filiph

Hi, friendly nudge. This prevents package:html to be used with HTML that includes unicode chars in attributes. Which is an increasing portion of them (according to bugs reported to linkcheck).

May 27 '19 19:05 filiph

I've created a pull request with a fix: https://github.com/dart-lang/html/pull/109

Aug 30 '19 02:08 cvolzke4

Carriage returns also affect the file location start and end points.

Aug 30 '19 07:08 cvolzke4

Hey there, thanks for the great work. Now that the fix is merged, would it be possible to release a new version?

We're stuck with this issue downstream (there: https://github.com/filiph/linkcheck/issues/35)

Sep 27 '19 13:09 b4stien