[Webkit-unassigned] [Bug 147445] Validation in isValidHTTPHeaderValue is too sensitive.

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=147445

--- Comment #16 from youenn fablet <youennf at gmail.com> ---
(In reply to comment #15)
> > Is it considered as a CFNetwork bug?
> 
> That's actually in WebKit, as calling HTTPHeaderMap::add twice combines the
> values. I don't know why the same didn't use to happen with a space, that's
> probably just another bug somewhere down the line.

I read your previous message too fast, sorry about that.

It seems strange that the combine rule applies to "Content-Type".
That seems like a different bug in the XHR implementation.
Either the web app sets it and the browser uses it as is.
Or the web app does not set it and the browser will set it instead.

If we split the sub-issues, that would give:
1. Add support for trimming whitespaces in XHR.setRequestHeader
2. Ensure XHR can set a header with an empty value
3. Ensure XHR can set "Content-Type" with an empty value
4. Decide what to do about empty values in the context of XHR combine rule.

Are we converging?

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.webkit.org/pipermail/webkit-unassigned/attachments/20150805/8f34904d/attachment-0001.html>