W3C home > Mailing lists > Public > public-webapps@w3.org > July to September 2011

Re: [XHR2] Blobs, names and FormData

From: Jonas Sicking <jonas@sicking.cc>
Date: Wed, 24 Aug 2011 23:36:13 -0700
Message-ID: <CA+c2ei8y2c758CZvLwF5=JH8Vz7Qj+kyt47LWm23Sigez1uLyA@mail.gmail.com>
To: Charles Pritchard <chuck@jumis.com>
Cc: Anne van Kesteren <annevk@opera.com>, WebApps WG <public-webapps@w3.org>
On Wed, Aug 24, 2011 at 12:57 PM, Charles Pritchard <chuck@jumis.com> wrote:
>>> Prpoposed:
>>>
>>> FormData output with the x-www-form-urlencoded mime type:
>>> formData.toUrlEncodedBlob(xhr.send)
>>>
>>> If going down the blob path, these two would have the same end-result:
>>> formData.toMultipartBlob(xhr.send)
>>> xhr.send(formData);
>>
>> What kind of API-style is this?
>>
> [Supplemental] FormData
> void toMultipartBlob(in callback)
> void toUrlEncodedBlob(in callback)
>
> The first would create a multipart mime message, in a blob, and run the
> callback with the blob as the first argument,
> the second would create a urlencoded message, in a blob, and also run the
> callback.
> They'd set the appropriate content type on generated blob.

The syntax you've written above wouldn't work in JS. You're only
passing in a reference to the send function, not a reference to the
XHR object on which to call .send on. So

formData.toMultipartBlob(xhr.send)

is equivalent to

formData.toMultipartBlob(XMLHttpRequest.prototype.send)

So in this case you'd have to pass in two argument, the function and
the 'this' object. Or require people to use .bind.

In general I'm not a fan of this syntax.

/ Jonas
Received on Thursday, 25 August 2011 06:37:18 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 18:49:47 GMT