Bug 99690 – [XMLHttpRequest] When String, Document or Blob is sent, Content-Type and encoding is set incorrectly

• Status: RESOLVED
• Resolution: DUPLICATE
• Component: WebCore Misc.

[Alexander Shalamov]

According to http://www.w3.org/TR/XMLHttpRequest/#the-send-method, content type and encoding should be:

String:   text/plain and encoding UTF-8
Document: For html document text/html and document encoding (UTF-8 if document encoding is not set).
          Other documents, application/xml and document encoding (UTF-8 if document encoding is not set).
Blob:     MIME type of the blob should be used if type is not empty.

[Alexander Shalamov]

Created attachment 169362 [details]
Patch 1

patch for initial review

[Build Bot]

Comment on attachment 169362 [details]
Patch 1

Attachment 169362 [details] did not pass mac-ews (mac):
Output: http://queues.webkit.org/results/14384895

[Alexander Shalamov]

Created attachment 169366 [details]
Patch 2

Fix mac-ews.
#if ENABLE(DASHBOARD_SUPPORT) was not handled properly.

[Alexey Proskuryakov]

How did you decide which result is correct, and which is not?

This is really a duplicate of bug 11049, and discussion should continue there.

[Alexey Proskuryakov]

*** This bug has been marked as a duplicate of bug 11049 ***

[Alexey Proskuryakov]

Comment on attachment 169366 [details]
Patch 2

Formally marking r- due to lack of rationale, but primarily to more discussion to original.

[Alexey Proskuryakov]

move*

[Alexander Shalamov]

(In reply to comment #7)
> move*

Sorry, I missed original bug when I searched for similar issues.

Do you want me to continue working on bug 11049 and reassign it to myself or I should stop working on this issue?

[Alexey Proskuryakov]

> Do you want me to continue working on bug 11049 and reassign it to myself

Yes, please! I haven't researched this any time recently, it's possible that WebKit is an outlier now, and needs to be fixed. But if the majority of browsers (by installed base, not by brand) still disagree with the spec, then it's a spec bug.

Also, I'd suggest doing this in small steps if possible - that way it wold be easier to find the cause if a site breaks (we would just pin that down to a specific revision).