W3C home > Mailing lists > Public > public-webcrypto@w3.org > June 2012

Re: Strawman proposal for the low-level API

From: Wan-Teh Chang <wtc@google.com>
Date: Wed, 20 Jun 2012 15:14:53 -0700
Message-ID: <CALTJjxE497oTidhm-zTspqrFC3fgXHXBhv5DqrOymGRZjSygzw@mail.gmail.com>
To: Ryan Sleevi <sleevi@google.com>
Cc: public-webcrypto@w3.org
Since I sit next to Ryan in the office, I've given my review comments
to him in person to maximize the communication bandwidth :-)

Overall I think Ryan's strawman proposal is well done and I appreciate
that he has researched and considered many issues.

I will save you guys my comments on minutiae such as typos.  My only
high-level complaint is the read-only 'result' attribute of
CryptoStream:

On Mon, Jun 18, 2012 at 10:53 AM, Ryan Sleevi <sleevi@google.com> wrote:
>
> [interface]
> interface CryptoStream : EventTarget {
>   void processData(ArrayBuffer buffer);
>   void processData(DOMString data);
>   void complete();
>
>   readonly attribute (DOMString or ArrayBuffer)? result;
>
>   attribute [TreatNonCallableAsNull] Function? onerror;
>   attribute [TreatNonCallableAsNull] Function? onprogress;
>   attribute [TreatNonCallableAsNull] Function? oncomplete;
> };

This forces us to keep the entire result, even if we have consumed it
chunk by chunk in the onprogress event handler.

I understand Ryan chose this design because it's used in other
approved or pending W3C API standards.  I hope we can do better in the
Web Crypto API.

Wan-Teh
Received on Wednesday, 20 June 2012 22:15:23 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 20 June 2012 22:15:23 GMT