[Webkit-unassigned] [Bug 39083] Refactor FormData and Blob for better support of Blobs synthesized by BlobBuilder

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=39083





--- Comment #51 from Darin Fisher (:fishd, Google) <fishd at chromium.org>  2010-06-02 22:35:13 PST ---
(In reply to comment #50)
> I'm still somewhat confused by the desire to make FormData share code with
> Blob.  It seems to me that FormData should just be a list of strings and file 
> paths (plus range and timestamp).  What is the advantage of making the internal 
> representation of a Blob share code with FormData?

I should clarify.  What I really mean is that it seems like the structure used
by the networking stack (which is currently called WebCore::FormData) should 
really just be a list of strings and file paths (plus range and timestamp).  It 
should just be the input parameters to the network stack.  This is why I have
previously recommended renaming that structure to be named HTTPBody.  Another
good name might be ResourceRequestBody.

Also, I agree with the idea of sharing code between the BlobBuilder and FormData 
web platform interfaces.  It makes sense for those to have a common backend.
Perhaps they should not be building on top of WebCore::FormData, but instead we
should generate one of those when it comes time to submit a Blob or a FormData
to the networking stack.

(It is confusing to talk about this since the web platform API named FormData is not the same thing as WebCore::FormData.)

Are you trying to avoid extra copies between a Blob and the network stack?

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.