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

Re: HmacKeyAlgorithm missing length?

From: Richard Barnes <rlb@ipv.sx>
Date: Tue, 18 Mar 2014 18:32:02 -0400
Message-ID: <CAL02cgRbZ2zGH9xeMg35qbzeKKA8a32M8vxR4+3GqH1f+nJZtA@mail.gmail.com>
To: Ryan Sleevi <sleevi@google.com>
Cc: "public-webcrypto@w3.org" <public-webcrypto@w3.org>
On Tue, Mar 18, 2014 at 5:13 PM, Ryan Sleevi <sleevi@google.com> wrote:

> The NoInterfaceObject is largely syntactic sugar, since interface objects
> do get created once the object is exposed as an attribute. Arguably, we
> could remove them all and have the exact same external behaviour.
>
> I'm assuming you mean to say "readonly unsigned long length" - since this
> is an interface, not a dictionary.
>

Yep.  Copy/paste error.

--Richard


>
> I can buy that for a dollar. It matches the same justification for the
> other keys - allow an app to determine (effective) security strength of the
> key in question.
>
>
> On Tue, Mar 18, 2014 at 10:40 AM, Richard Barnes <rlb@ipv.sx> wrote:
>
>> It seems like we expose length attributes for all of the *KeyAlgorithm
>> interfaces besides HmacKeyAlgorithm.  Also, the other *KeyAlgorithm
>> interfaces have the [NoInterfaceObject] directive.
>>
>> Proposed micro-patch:
>> +[NoInterfaceObject]
>>  interface HmacKeyAlgorithm : KeyAlgorithm {
>>    // The inner hash function to use.
>>    readonly attribute KeyAlgorithm hash;
>> +  // The length (in bits) of the key
>> +  [EnforceRange] unsigned long length;
>>  };
>>
>
>
Received on Tuesday, 18 March 2014 22:32:29 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:17:22 UTC