On 2/17/2014, 4:32 PM, Brendan Eich wrote:
The problem is one V8 principal (among others in what I think is a clear
minority) do not agree with the current consensus. The previous
consensus was actually fractured, but no one worked through it and some
amount of miscommunication, perhaps combined with "expansive" risk
tolerance by some on the committee, and some off, led to things
diverging. V8's implementation thus differs from ES6.
I see, thanks for the explanation.
The current ES6 consensus needs to be nailed down harder, but I think it
will stick. That it isn't compositional won't stop this. Promises were a
library de-facto standard from CommonJS and other ecosystems; the
committee erred in trying to redesign them, the long email threads both
before and after make this clear.
So, what happens if we end up shipping what we have now, and then TC39
propose another change to the spec? Right now Blink is in that same
situation where they have something shipped which no longer conforms to
the spec, and they don't know for a fact whether the current ES6
consensus is going to stick for sure, so they're unwilling to update
their implementation out of the fear of having to adjust it again later
on for further changes. We're going to be in the same vulnerable
position very soon, and I'm trying to understand how we're going to deal
with this possibility.
Thanks!
Ehsan
SpiderMonkey still needs to nativize the DOM/XPCOM-based implementation,
both to follow the spec (including subclassability) and to avoid
dependencies on the DOM or XPCOM.
/be
David Bruant <mailto:[email protected]>
February 17, 2014 at 1:27 PM
Telling Brendan can work I guess :-)
The topic came out in this (and maybe other) threads :
https://mail.mozilla.org/pipermail/es-discuss/2014-January/035976.html
I haven't read every single es-discuss email recently, but I have the
feeling that TC39 is ok to standardize the current spec and not make
changes at this point anymore.
That said, it'd be excellent if Blink and Moz completed the current
test suite (Promise/A+), or at least backport tests from one another
to be sure they implement the same thing to the finest detail (and
report spec issues if such a thing arise).
David
_______________________________________________
dev-tech-js-engine-internals mailing list
[email protected]
https://lists.mozilla.org/listinfo/dev-tech-js-engine-internals
Ehsan Akhgari <mailto:[email protected]>
February 17, 2014 at 12:56 PM
Hi everyone,
We are on track to ship our Promise implementation in Firefox 29. Blink
has already shipped their implementation on the stable channel of Chrome
and they are facing difficulties determining whether they should change
what they have shipped based on the recent ES6 changes to the Promise API
in TC39. I'm not very familiar with the TC39 processes, but is there any
way for us to indicate to TC39 that we should avoid further changes to
the
API as the second engine is getting very close to shipping Promise?
Please see this thread <
https://groups.google.com/a/chromium.org/forum/#!topic/blink-dev/RNphv8dgJ0g>
for more context.
Thanks!
--
Ehsan
<http://ehsanakhgari.org/>
_______________________________________________
dev-tech-js-engine-internals mailing list
[email protected]
https://lists.mozilla.org/listinfo/dev-tech-js-engine-internals
_______________________________________________
dev-tech-js-engine-internals mailing list
[email protected]
https://lists.mozilla.org/listinfo/dev-tech-js-engine-internals
