Il giorno lunedì 3 marzo 2014 03:52:15 UTC+11, skaller ha scritto:
>
> Now a refresher.
>
> Why can't you do sych I/O or async I/O in a function?
> I mean reading from an schannel or socket naturally
> IS a function (even if writing isn't). Or at least a generator!
>
> The reason again: these operations block fthreads. To do that,
> the fthread has yield, and a pointer to it put onto some list
> so another fthread can become current.
>
> The problem is you can't have a pointer to the machine stack,
> or anything closed over it, and swap it for another pointer,
> to exchange control.
>
> Actually you CAN do this: you use a pthread and a lock.
> Pthreads are the natural way to swap machine stacks.
>
> So actually you CAN use pchannels in generators.
>
> =============
>
> Note to self: pchannels BLOCK the whole pthread.
> Which means there's no way to send message between
> fibres on different pthreads.
>
> But clearly this can be done with async I/O machinery,
> which allows *suspending* a fibre while it waits for
> asynchronous service .. which is exactly how communication
> across pthread boundaries looks.
>
> So we really need to be able to do this. Interestingly this
> would work fine within the same pthread too.
>
Interesting!
------------------------------------------------------------------------------
Subversion Kills Productivity. Get off Subversion & Make the Move to Perforce.
With Perforce, you get hassle-free workflows. Merge that actually works.
Faster operations. Version large binaries. Built-in WAN optimization and the
freedom to use Git, Perforce or both. Make the move to Perforce.
http://pubads.g.doubleclick.net/gampad/clk?id=122218951&iu=/4140/ostg.clktrk
_______________________________________________
Felix-language mailing list
Felix-language@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/felix-language
