Re: Request bodies alteration and character encoding issue

On Tue, Mar 18, 2008 at 7:07 AM, Francois Daoust <fd@w3.org> wrote:
>  We identified a case where the CT-proxy typically alters request bodies
>  when it splits forms in multiple pages.

Right, this is a good one.

>  You mentioned in a previous call [2] that there exists another case
>  where the CT-proxy actually may change the encoding of posted data, but
>  we couldn't quite figure out when that may be needed in practice.
>
>  Would you have an example of this?

Sure.  There are a few situations that come to mind:

1. The mobile doesn't support multipart/form-data but the page has
requested this encoding for the form.  In this case, we send a form
with encoding="application/x-www-form-urlencoded" to the mobile and do
the transformation before sending the result to the server.

2. The mobile doesn't support the character set that the form uses.
In this case, we convert the content from the mobile's preferred
character set into whatever the form orginally required.

3. Forms that use the GET method are often modified to be POST from
the mobile for a variety of reasons.

Aaron

Received on Tuesday, 18 March 2008 14:56:13 UTC