sideshowbarker

OK, I found that the existing code had effectively been using, as the internal canonical name, names with all dashes and underscores removed. That results in a conformance violation, because...

OK, I’ve now pushed a large number of commits to this branch that together make the parser’s handling of encodings conformant with current requirements in the Encoding spec and HTML...

OK, I’m now a bit embarrassed to admit: it hadn’t dawned on me until now that the changes to the `Encoding` class in this patch are completely superseded by the...

> do you see value in the `common`/`base` split that @anthonyvdotbe suggests? I don’t personally see value in it. But I’m not a domain expert around the fine points of...

OK, #48 has a fix for the following. > https://github.com/html5lib/html5lib-tests/blob/master/encoding/tests1.dat#L65 > > ``` > Failure. Expected: windows-1252 got ISO-8859-2 > > > " > #encoding > windows-1252 > ```

OK, #49 has a fix for the following. > https://github.com/html5lib/html5lib-tests/blob/master/encoding/tests1.dat#L314-L347 (6 cases; 2048–8193 characters) > > ``` > Failure. Expected: ISO-8859-2 got windows-1252. > > > #encoding > iso-8859-2 >...

With the fixes in #48 and #49 applied — and with the changes in html5lib/html5lib-tests#130 merged, we will be passing all the html5lib-tests encoding cases.

Support for it was recently added to the W3C CSS Validator. See https://github.com/w3c/css-validator/pull/134 & https://github.com/w3c/css-validator/commit/ab64237ce24499bb0f8c23631135d18ee58cb96f — with a warning emitted to indicate that it’s a non-standard vendor extension. So yeah...

Friendly ping to reviewers

I’ve opened PR #10 with a patch that fixes this problem.