Bug 64207 - IDNA or URL parsing error with mixed escapings of CJK
Summary: IDNA or URL parsing error with mixed escapings of CJK
Status: UNCONFIRMED
Alias: None
Product: WebKit
Classification: Unclassified
Component: Page Loading (show other bugs)
Version: 528+ (Nightly build)
Hardware: PC Windows 7
: P2 Normal
Assignee: Nobody
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2011-07-08 14:42 PDT by Chris Weber
Modified: 2011-07-11 16:33 PDT (History)
2 users (show)

See Also:


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Chris Weber 2011-07-08 14:42:41 PDT
Running Safari 5.0.5 on Windows 7 x64.  I was looking a little closer at the following test case from Webkit's
http://trac.webkit.org/browser/trunk/LayoutTests/fast/url/script-tests/host.js

// Mixed UTF-8 and escaped UTF-8 (narrow case) and UTF-16 and escaped
// UTF-8 (wide case). The output should be equivalent to the true wide
// character input above).
["%E4%BD%A0%E5%A5%BD\u4f60\u597d", "xn--6qqa088eba"],

My test suite is slightly different but I reuse this test case as in:

<a href='http://%E4%BD%A0%E5%A5%BD&#x4f60;&#x597d;' id='77'>77</a><img src='http://%E4%BD%A0%E5%A5%BD&#x4f60;&#x597d;' /><br />

Running a sniffer like Wireshark and testing on Windows 7, a DNS query hits the wire with the following name:

xn--\xc3\xa4\xc2\xbd\xc2\xa0\xc3\xa5\xc2\xa5\xc2\xbd-5g72ap39l

In my test setup, my DNS server responds with an IP address, so Safari continues to make the HTTP request for the <img> with the following Host header:

Host: xn--??-5g72ap39l

Is this a problem with IDNA characters or the mixed escaping/encoding presented in the <img> src?