Re: alive e/ken style projects?
Matt Rice <[email protected]> Thu, 15 Oct 2015 00:51:24 -0700
| Newsgroups | gmane.comp.capabilities.general |
|---|---|
| Message-ID | <CACTLOFpxrA5YqwKyidZh_yRFcf8YqAsE2VKuS-rHJ9CRzg=AZQ@mail.gmail.com> |
One of the things i've had difficulty with trying to use automagic promise pipelining in is unreliable protocols and worse mixed-reliability protocols like SCTP, say there is a promise whose reply gets dropped, is there a need for some reliable exception channel, or will generating an exception on timeout suffice, there really doesn't seem to be an automagic answer afaict... not sure that anybody has tried this sort of thing before, unreliable promises seem a bit counter intuitive, though I haven't used promise pipelining very much, the impression I get is that modeling unreliability makes things weird... On Wed, Oct 14, 2015 at 4:06 PM, Mark Miller <[email protected]> wrote: > I wonder whether you are using automagical on two different sides of the > abstraction boundary. In E/CapTP and Cap'n Proto, we platform builders put a > lot of work into building promise pipelining into the platform. The result > is that they are automagical to the users of these platforms. > > I know Waterken had the enabling mechanism for promise pipelining. But I > don't remember if it was completed and made the default mode of operation. > > > On Wed, Oct 14, 2015 at 5:28 PM, Marc Stiegler <[email protected]> wrote: >> >> pipelining promises is anything but automagical. Capper, my open source >> nodejs server that uses promises and semi-transparent checkpointing and >> webkeys, intended to bring some of the virtues of waterken to javascript, >> does not include pipelining for the promises. The pipelining is far more >> work to implement than are the promises themselves. >> >> --marcs >> >> On Wed, Oct 14, 2015 at 9:39 AM, Raoul Duke <[email protected]> wrote: >>> >>> >> they can be pipelined. In particular, they can be pipelined over the >>> >> network when invoking operations on remote objects, which means that a >>> > So I would say, yes. But the Javascript community largely hasn't >>> > discovered >>> > it yet. :-) >>> >>> But "can" doesn't mean it happens automagically by the underlying >>> framework. Maybe it does, but I suspect most likely it doesn't, yet? >>> _______________________________________________ >>> cap-talk mailing list >>> [email protected] >>> http://www.eros-os.org/mailman/listinfo/cap-talk >> >> >> >> _______________________________________________ >> cap-talk mailing list >> [email protected] >> http://www.eros-os.org/mailman/listinfo/cap-talk >> > > > > -- > Text by me above is hereby placed in the public domain > > Cheers, > --MarkM > > _______________________________________________ > cap-talk mailing list > [email protected] > http://www.eros-os.org/mailman/listinfo/cap-talk >