W3C home > Mailing lists > Public > public-webappsec@w3.org > March 2014

Re: [integrity] What should we hash?

From: Devdatta Akhawe <dev.akhawe@gmail.com>
Date: Wed, 12 Mar 2014 22:37:42 -0700
Message-ID: <CAPfop_1pibb=bdaKidWdNCtsjRWLM567CSaBFuOYjtcpRZjPDw@mail.gmail.com>
To: Mark Nottingham <mnot@mnot.net>
Cc: "public-webappsec@w3.org" <public-webappsec@w3.org>, Boris Zbarsky <bzbarsky@mit.edu>
Thanks!

Does "representation" or "the message payload before content codings
are applied." sound right to others? Boris?

I (or one of the editors) will update the spec soon unless someone
flags an issue.


~Dev

On 12 March 2014 22:28, Mark Nottingham <mnot@mnot.net> wrote:
> That would be the representation - see:
>   http://tools.ietf.org/html/draft-ietf-httpbis-p2-semantics-26#section-3.1.1.5
>
> You should still probably have a few words around it to clarify; this sort of thing is easy to mess up.
>
> Cheers,
>
>
> On 13 Mar 2014, at 4:24 pm, Devdatta Akhawe <dev.akhawe@gmail.com> wrote:
>
>> Thanks! Intuitively, it seems there should be a simpler way to refer
>> to "the message payload before content codings are applied.".
>>
>> What does the content-type refer to? Type of what? For example, if I
>> am not completely wrong, content-type of a text file with gzip
>> content-encoding is still text/plain and so presumably talks about
>> "before content codings are applied"
>>
>> thanks
>> Dev
>>
>> On 12 March 2014 17:47, Mark Nottingham <mnot@mnot.net> wrote:
>>> The HTTPbis docs are going to obsolete RFC2616 in a few weeks, so it's best to look at them.
>>>
>>> I think you want to say that integrity operates upon the payload of the message - see
>>>  http://tools.ietf.org/html/draft-ietf-httpbis-p1-messaging-26#section-3.3
>>>
>>> This is after chunk encoding, gzip transfer-codings, etc. have been removed. However, content-codings are still there, e.g., for gzip, deflate in the Content-Encoding header.
>>>
>>> That's because Content-Encoding is considered a property of the representation in HTTP, even though many people implement it as a separate layer.
>>>
>>> If you want to do it before content-encoding, you'd need to specify it explicitly; e.g., as "the message payload before content codings are applied."
>>>
>>> Hope this helps,
>>>
>>> On 12 Mar 2014, at 3:58 am, Devdatta Akhawe <dev.akhawe@gmail.com> wrote:
>>>
>>>> Hi
>>>>
>>>> One key question for integrity spec is "What should the browser hash?"
>>>> Boris mentioned this previously
>>>> http://lists.w3.org/Archives/Public/public-webappsec/2013Dec/0048.html
>>>>
>>>> Informally, I am leaning towards hashing content after undoing stuff
>>>> like gzip, deflate, chunked-encodings etc. Does that sound reasonable?
>>>>
>>>> Next, how do we formalize (spec) this? In an ideal world, just saying
>>>> "undo transfer-encoding" would be enough (i.e., spec would say "hash
>>>> entity body"). But, common behavior is to apply gzip via
>>>> Content-Encoding not transfer-encoding. And we want to hash after
>>>> undoing gzip. (see Boris' email above)
>>>>
>>>> Mark: Do you know good specification text for this? After looking at
>>>> the HTTP RFC, one wording that springs to my mind is: ""After decoding
>>>> the entity to the media-type referenced by the content-type header"
>>>>
>>>> Thanks
>>>> Dev
>>>
>>> --
>>> Mark Nottingham   http://www.mnot.net/
>>>
>>>
>>>
>
> --
> Mark Nottingham   http://www.mnot.net/
>
>
>
Received on Thursday, 13 March 2014 05:38:28 UTC

This archive was generated by hypermail 2.3.1 : Monday, 23 October 2017 14:54:05 UTC