On 15 November 2013 13:20, Domenic Denicola <[email protected]> wrote:
> As of now, that changeset adds several non-standard methods and, more 
> importantly, fails a plethora of tests among those tested in the 
> promises-unwrapping spec repo. I'll put together a more detailed report this 
> weekend. But I wanted to make sure the appropriate people were aware of these 
> issues, and I couldn't figure out the code review interface.

Great, yes, please share your findings. We definitely will need to
have a more extensive test suite, and adapting existing Blink tests as
well as tests from your repo is a high priority. This CL is just the
first step. I fully expect that the implementation still needs some
tweaking. (I will also try to keep the github prototype in sync if I
can.)

Thanks,
/Andreas


>> -----Original Message-----
>> From: [email protected] [mailto:[email protected]] On Behalf
>> Of Andreas Rossberg
>> Sent: Friday, November 15, 2013 05:59
>> To: Yutaka Hirano
>> Cc: Erik Arvidsson; [email protected]; Dirk Pranke; blink-dev;
>> Petka Antonov; Adam Klein; Rafael Weinstein; [email protected]
>> Subject: Re: [blink-dev] Re: [v8-users] Intent to Implement Promises in V8
>>
>> Update on the V8/promise front:
>>
>> After Yutaka started this thread, I looked at the promise spec in more depth
>> and proposed some structural simplifications. As proof of concept for those I
>> quickly banged together a JavaScript prototype [1]. Most of that got adopted
>> by Domenic in the latest draft (modulo internal name changes)  [2].
>>
>> Since the prototype was already feature-complete, and a self-hosted V8
>> implementation shouldn't look much different, I simply went ahead and
>> turned it into a CL [3]. Comments are welcome.
>>
>> Of course, this is just an initial implementation, and I'm happy that Yutaka 
>> has
>> agreed to take over from there, even though I was jumping a bit ahead of his
>> own initiative.
>>
>> Thanks,
>> /Andreas
>>
>> [1] https://github.com/rossberg-chromium/js-promise
>> [2] https://github.com/domenic/promises-
>> unwrapping/blob/master/README.md
>> [3] https://codereview.chromium.org/64223010/
>>
>> On 15 October 2013 10:50, Yutaka Hirano <[email protected]> wrote:
>> > I've written a design document for Promises.
>> >
>> https://docs.google.com/a/chromium.org/document/d/1_aw7DInk0V1eBl4
>> OPaq
>> > XNMBgBPlN0hq-RC_OlteVkoY/edit# It would be appreciated if you could
>> > give me some comments / suggestions.
>> >
>> > Thanks,
>> >
>> >
>> > On Wed, Oct 9, 2013 at 10:03 AM, Yutaka Hirano <[email protected]>
>> wrote:
>> >>
>> >>
>> >>>
>> >>> There are tests for Object.observe in V8 (both mjusunit and cctest)
>> >>> and I believe the cctests tests the timing and ordering of the callbacks.
>> >>
>> >> Testing in cctest seems reasonable. Thank you.
>> >>
>> >>
>> >> On Tue, Oct 8, 2013 at 12:04 AM, Erik Arvidsson <[email protected]>
>> wrote:
>> >>>
>> >>>
>> >>> On Mon, Oct 7, 2013 at 5:33 AM, Yutaka Hirano <[email protected]>
>> wrote:
>> >>>>
>> >>>> Thank you for your comments.
>> >>>>
>> >>>>>
>> >>>>> - please clarify your testing plans, both w.r.t. conformance and
>> >>>>> performance
>> >>>>
>> >>>> We have LayoutTests in Blink and I plan to write tests something
>> >>>> like them.
>> >>>> IIUC we can't write async tests with mjsunit. Maybe we will modify
>> >>>> mjsunit.
>> >>>
>> >>>
>> >>> There are tests for Object.observe in V8 (both mjusunit and cctest)
>> >>> and I believe the cctests tests the timing and ordering of the callbacks.
>> >>>
>> >>>>
>> >>>> For performance, currently I have no specific idea.
>> >>>>
>> >>>>
>> >>>>> - as Jochen mentioned, let's work together to fit promises' needs
>> >>>>> into a larger work to unify threading between Blink and V8
>> >>>>
>> >>>> Sure. Can you provide some pointers if any?
>> >>>>
>> >>>>> - generally, it will be nice to have a design document describing
>> >>>>> promises implementation (maybe there is one already and I missed
>> >>>>> it?)
>> >>>>
>> >>>> I see. I will write one.
>> >>>>
>> >>>>
>> >>>>
>> >>>> On Mon, Oct 7, 2013 at 2:24 AM, Dmitry Lomov
>> <[email protected]>
>> >>>> wrote:
>> >>>>>
>> >>>>> If promises become a part of V8, we of course expect them to be
>> >>>>> quite usable on server-side too (node.js is our customer after
>> >>>>> all)
>> >>>>>
>> >>>>>
>> >>>>> On Sun, Oct 6, 2013 at 1:05 PM, Petka Antonov
>> >>>>> <[email protected]>
>> >>>>> wrote:
>> >>>>>>
>> >>>>>> The benchmark is emulating a server side workflow. I don't
>> >>>>>> imagine there is anything like that needed on client side.
>> >>>>>>
>> >>>>>> On client side you could be calculating fibonacci numbers for fun
>> >>>>>> in the promise internals and no-one would notice anything.
>> >>>>>> Although on client-side for some use-cases it's important to use
>> >>>>>> a fast scheduler like a MutationObserver.
>> >>>>>>
>> >>>>>> Petka
>> >>>>>>
>> >>>>>> --
>> >>>>>> --
>> >>>>>> v8-users mailing list
>> >>>>>> [email protected]
>> >>>>>> http://groups.google.com/group/v8-users
>> >>>>>> ---
>> >>>>>> You received this message because you are subscribed to the
>> >>>>>> Google Groups "v8-users" group.
>> >>>>>> To unsubscribe from this group and stop receiving emails from it,
>> >>>>>> send an email to [email protected].
>> >>>>>> For more options, visit https://groups.google.com/groups/opt_out.
>> >>>>>
>> >>>>>
>> >>>>> --
>> >>>>> --
>> >>>>> v8-users mailing list
>> >>>>> [email protected]
>> >>>>> http://groups.google.com/group/v8-users
>> >>>>> ---
>> >>>>> You received this message because you are subscribed to the Google
>> >>>>> Groups "v8-users" group.
>> >>>>> To unsubscribe from this group and stop receiving emails from it,
>> >>>>> send an email to [email protected].
>> >>>>> For more options, visit https://groups.google.com/groups/opt_out.
>> >>>>
>> >>>>
>> >>>
>> >>>
>> >>>
>> >>> --
>> >>> erik
>> >>>
>> >>>
>> >>
>> >

-- 
-- 
v8-users mailing list
[email protected]
http://groups.google.com/group/v8-users
--- 
You received this message because you are subscribed to the Google Groups 
"v8-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
For more options, visit https://groups.google.com/groups/opt_out.

Reply via email to