Opened 9 years ago

Closed 8 years ago

#36 closed defect (wontfix)

Some HTTP implementations send UTF8 path directly

Reported by: duerst@… Owned by:
Priority: major Milestone:
Component: 4395bis Version:
Severity: - Keywords:
Cc: masinter@…

Description

(from the "Open Issues" section in draft -01)
This may not really be an IRI issue, but some HTTP implementations send UTF8 path directly. This needs some review.

Attachments (2)

IRI Testing Results.xls (515.5 KB) - added by chris@… 8 years ago.
A collection of IRI test results from Quirks-mode HTML pages loaded into popular Web browsers, with the UTF-8 charset set in the HTTP header.
Capture.PNG (49.0 KB) - added by masinter@… 8 years ago.
Internet Options in Windows/IE9 showing option to use UTF-8 URLs

Download all attachments as: .zip

Change History (5)

Changed 8 years ago by chris@…

A collection of IRI test results from Quirks-mode HTML pages loaded into popular Web browsers, with the UTF-8 charset set in the HTTP header.

comment:1 Changed 8 years ago by chris@…

I've completed some initial testing of 5 major browsers and found that only MSIE sends raw UTF-8 bytes in an HTTP request. This only happens in the query component and not in the path, which is always percent-encoded. What more HTTP implementations should be tested?

Changed 8 years ago by masinter@…

Internet Options in Windows/IE9 showing option to use UTF-8 URLs

comment:2 Changed 8 years ago by masinter@…

  • Cc masinter@… added

See Internet Options on Windows 7 / IE9 for setting this.

I think the IRI spec now explicitly allows scheme definitions to talk about how they process "parsed IRI components". So I believe that this means the HTTP spec can/should be updated to allow HTTP implementations to do that, rather than first converting the path to ASCII with hex encoding.

So I would claim this might be a problem with the definition of the HTTP protocol and also the HTTP URI spec. I'm not sure I would change the IRI document, except perhaps to point this out as a possibility?

comment:3 Changed 8 years ago by masinter@…

  • Component changed from 3987bis to 4395bis
  • Resolution set to wontfix
  • Status changed from new to closed

I don't see what change to 3987bis would be helpful.
I might see 4395bis urging updates to existing IRI specs?

Would the HTTP working group will declare this out of scope?

Note: See TracTickets for help on using tickets.