Re: [Beacon] Review of current spec

On Fri, Nov 15, 2013 at 10:33 AM, Jonas Sicking <jonas@sicking.cc> wrote:

> On Fri, Nov 15, 2013 at 2:00 AM, Ilya Grigorik <igrigorik@google.com>
> wrote:
> > On Thu, Nov 14, 2013 at 11:32 PM, Jonas Sicking <jonas@sicking.cc>
> wrote:
> >> Mozilla has proposed APIs to enable this.
> >
> > ^ Jonas, do you have any links to past proposals? Was looking for exactly
> > that earlier, but no luck...
>
> Hrm.. I can't find it either. I guess we never made one publicly but
> only internally.
>
> We have an implementation of the API at [1] though. Think of
> "ArchiveRequest" as "Promise".
>

That looks to be a read-only API, no? Take a (DOMString) flename > return
contents... But in either case, not hard to imagine extending that.


> An alternative approach would be to expose a read-only version of the
> Directory interface defined in [2]. That's definitely more complex,
> but has the advantage that it in the future could allow code sharing
> with filesystems.
>

+1. It would be great to have a generic interface to handle all of these
cases.

Even if performance is ~similar to pure JS (don't have any benchmarks), I
believe this is a common enough use case to warrant a native API. E.g. most
recently I bumped into this with WS compression: Chrome M32 is shipping WS
compression support, but at the moment, we blindly compress any/all
client-generated messages, even if you're sending an already compressed
Blob.. which is a big waste of cycles on both sides.

ig

Received on Friday, 15 November 2013 11:27:26 UTC