Re: Handling of CryptoOperationData needs to be defined

Boris,

Would you mind filing bugs in the bug tracker to make sure we never miss
comments?


On Wed, Apr 23, 2014 at 12:40 PM, Boris Zbarsky <bzbarsky@mit.edu> wrote:

> As far as I can tell, nothing actually defines the handling of
> CryptoOperationData arguments.
>
> For example, looking at https://dvcs.w3.org/hg/webcrypto-api/raw-file/tip/
> spec/Overview.html#dfn-SubtleCrypto-method-encrypt it just says:
>
>   Let ciphertext be the result of performing the encrypt operation
>   specified by normalizedAlgorithm using algorithm and key and with
>   data as plaintext.
>
> but nothing ever defines the type of "plaintext" here.  I assume it's
> meant to be some sort of byte array, right?
>
> If so, the following need to be clearly defined:
>
> 1)  How do extract a byte array from an ArrayBufferView.  Presumably this
> is a matter of looking at the underlying buffer and taking byteLength bytes
> starting at byteOffset.  And presumably this uses the canonical getters for
> those properties, not whatever is on the prototype (so the
> [[ViewedArrayBuffer]], [[ByteLength]], and [[ByteOffset]] internal slots of
> the object).  Note that at least one implementor just used all the bytes of
> the underlying ArrayBuffer instead, so this clearly does need defining.
>

Agreed


>
> 2) The exact point when the snapshot of the data is taken.  This matters,
> because the data is not immutable.  Presumably there's some sort of data
> copy before step 3 of https://dvcs.w3.org/hg/webcrypto-api/raw-file/tip/
> spec/Overview.html#dfn-SubtleCrypto-method-encrypt for example?


Agreed. The explicit copy step was called out in the pre-Promises version
(under the 'processing data' model), but I hadn't updated that section for
promises.


>
>
> -Boris
>
>