I've been following this thread with interest. I also find the documentation about async methods in vala to be a little thin. For example, in the tutorial, the first line of description says: "With asynchronous methods it is possible to do programming without any blocking." The following examples in the tutorial are much more enlightening...so it's not that bad. But, maybe it could be improved. I always thought async methods: a) are always single threaded b) uses the glib runtime scheduler to simulate multi-processes c) cannot take advantage of multi-core or multi-processor computers Is that correct? Also, are there any plans to ease multi-core development in Vala? thanks, best -august.
Actually, this is Luca's code (the coroutine-based generator). (I am the guilty party to blame for the naive idea to try to add threads to it : I had the Goroutine/channel concept of Go in mind and was trying to emulate it by attempting to combine Luca's generator with threads). :) Serge. On Wed, Jul 13, 2011 at 3:02 PM, Jim Peters <jim uazu net> wrote:Serge Hulne wrote:Here is a further development of the idea of Luca Bruno about a Vala implementation for Generators: Simulating Go's goroutines and channels in Vala: Basically the idea is to start as many threads as needed (which play the role of Go' goroutines) and to recuperate their output from a "Generator" (which plays the role of the "Go" channel form which the result from agiventhread can be pulled):To clarify after all the confusion (partly my fault -- sorry): - Luco's Generator doesn't need threads, and it doesn't need a main loop either, even though it is using 'async' - The whole thing runs in a single thread A few thoughts: - 'async' is really powerful in Vala, but also quite hard to understand. (I wish I could find a high-level design document explaining async in GIO.) - Its implementation of coroutines doesn't require another stack, it holds its state in a private data structure. When it does a 'yield' it returns to the caller, and when it is next given control, it resumes the previous point of execution using a switch and 'goto'. - Probably it could all go a lot faster if less housekeeping was going on, but I haven't seen Go's implementation either, so who knows how it compares. Anyway, Vala always puts functionality first and optimisation second, which suits people who want to get the job done now instead of waiting for the ultimate language to arrive, so that seems okay. Attached is a version of Serge's code with all the thread stuff cut out, which still works. Jim ///////////////////////////////// using Posix; abstract class Generator<G> { private bool consumed; private unowned G value; private SourceFunc callback; public Generator () { helper (); } private async void helper () { yield generate (); consumed = true; } protected abstract async void generate (); protected async void feed (G value) { this.value = value; this.callback = feed.callback; yield; } public bool next () { return !consumed; } public G get () { var result = value; callback (); return result; } public Generator<G> iterator () { return this; } } class IntGenerator : Generator<int> { protected override async void generate () { for (int i=0; i < 10; i++) { if (i%2 ==0) yield feed (i); } } } class IntGenerator_1 : Generator<int> { protected override async void generate () { for (int i=0; i < 10; i++) { if (i%2 !=0) yield feed (i); } } } int main(string[] args) { var gen = new IntGenerator(); var gen_1 = new IntGenerator_1(); print("\n\nResults computed in first generator\n"); var i=0; foreach (var item in gen) { if (i<10) Posix.stdout.printf("%i\n", item); i++; } print("\n\nResults computed in the second generator\n\n"); i=0; foreach (var item in gen_1) { if (i<10) Posix.stdout.printf("%i\n", item); i++; } return 0; } ///////////////////////////////// -- Jim Peters (_)/=\~/_(_) jim uazu net (_) /=\ ~/_ (_) UazĂș (_) /=\ ~/_ (_) http:// in Peru (_) ____ /=\ ____ ~/_ ____ (_) uazu.net
_______________________________________________ vala-list mailing list vala-list gnome org http://mail.gnome.org/mailman/listinfo/vala-list
-- ------------------- http://aug.ment.org
Attachment:
signature.asc
Description: Digital signature