Boost logo

Boost :

Subject: Re: [boost] Boost candidate similar to asyncplusplus
From: Vicente J. Botet Escriba (vicente.botet_at_[hidden])
Date: 2014-01-24 02:03:35

Le 23/01/14 22:02, Rodrigo Madera a écrit :
> Hello,
> I was wondering if there is anything similar to asyncplusplus [1] here
> in Boost land.
> I just found out about the library and I specially liked the task
> chaining syntax [2]:
> ---
> async::spawn([] {
> return 42;
> }).then([](int result) {
> return result;
> }).then([](task<int> parent) {
> std::cout << parent.get() << std::endl;
> });
> ---

I'm working on something like that [3] base on [4]. I expect to be ready
for release it for the next release, but I have yet some test cases that
are failing on some platforms, and some limitations on the interface.

There are some examples/tests in [5].

Your example would be something like that

async(ex,[] {
     return 42;
}).then(ex, [](int result) {
     return result;
}).then(ex, [](task<int> parent) {
     std::cout << parent.get() << std::endl;

Note the ex parameter. ex is an executor/scheduler. The library provides
some of the executors proposed in [4], as a basic_thread_pool. The main
difference between N3785 and my library is that the executors are
models of an Executor concept instead of inheriting from a base executor
class. I'm yet working on a scheduled_executor class that wraps an
Executor to provide time base scheduler operations.

In order to be able to do something closer to your example we will need
a default executor/scheduler which I have not yet and I don't know if I
would provide one in a near future.

> This is something I would love to see in Boost.
> So I ask:
> 1) Are there any similar Boost candidates for task scheduling?
> 2) If not, is ASIO up to this with some adaptation?
I don't master ASIO so I let others respond here.
> I felt specially inclined to ask this here because of all the
> discussion taking place over the recently reviewed Fibers library and
> the fundamentals of asynchronous processing. Very interesting thoughts
> being discussed there. And something more higher level (but not
> necessarily IO bound) might be helpful.
> Best regards,
> Rodrigo Madera
> [1]
> [2]
> _______________________________________________
> Unsubscribe & other changes:



Boost list run by bdawes at, gregod at, cpdaniel at, john at