W3C home > Mailing lists > Public > public-test-infra@w3.org > January to March 2014

Re: /html/syntax/parsing not working?

From: Koji Ishii <kojiishi@gluesoft.co.jp>
Date: Thu, 20 Mar 2014 00:12:14 +0000
To: James Graham <james@hoppipolla.co.uk>
CC: "public-test-infra@w3.org" <public-test-infra@w3.org>
Message-ID: <59093D83-8B6C-4F98-A71D-21F3882EF26F@gluesoft.co.jp>
>> I believe this is because Blink/WebKit incorrectly prevent accessing resources loaded into data: urls cross-origin. Unfortunately I think working around this bug will require rewriting the tests to load the test data into an isolated browsing context in which to perform the parsing using some other mechanism.
> 
> Fails on IE11 too:
> Access is denied. (stack: Error: Access is denied. at Anonymous function (http://w3c-test.org/html/syntax/parsing/test.js:162:16) at step (http://w3c-test.org/resources/testharness.js:1159:13) at onload (http://w3c-test.org/html/syntax/parsing/test.js:160:7))
> 
> So it looks like the only platform the tests support is Firefox. It doesn’t look like a great test for interoperability…
> 
> Is it hard to rewrite the test as you suggested?

After a bit more thoughts…

Would it be a simple fix to add CORS header to serve.py?

If it’s not a good way to go, what do you think for the python script to create an HTML file for each test and load them using the regular iframe? Does it sound like a reasonable way to go?

/koji
Received on Thursday, 20 March 2014 00:12:49 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:20:34 UTC