Hi Yoav,
Thanks for your clarification, I missed out the initialization
requirement.
Jeeva P.
> Re: [X10-users] async at (here) and at (here) async
> Yoav Zibin
> Fri, 06 May 2011 15:44:15 -0700
>
> Please read carefully what I wrote:
> ... it must be *initialized *and declared final.
>
> You cannot just insert "at (here)" anywhere you want - the compiler doesn't
> treat "here" as a constant - so it assumes there might be a place shift as
> if you had written "at (...)".
>
> On Fri, May 6, 2011 at 2:08 PM, Jeeva Paudel <je...@ualberta.ca> wrote:
>
> > But, f1 and f2 are already declared as val. So, there should not be problem
> > in accessing them.
> >
> > Further, both the async statements just make it explicit what is already
> > implicit.
> >
> > async { f1 = fib(n-1); } implicitly means that async activity is created
> > here, where the root activity resides, and the enclosed statement is also
> > executed here.
> >
> > The statements at (here) async { f1 = fib(n-1); } and async at (here) {fi
> > = fib(n-1); } must also start new activity here and execute the enclosed
> > statement in the current place, i.e., here.
> >
> > So, shouldn't all these three forms be correct?
> >
> > Jeeva P.
> > Re: [X10-users] async at (here) and at (here) async
> > Yoav Zibin
> > Fri, 06 May 2011 10:22:37 -0700
> >
> >
> > This is a bug.
> >
> > In the upcoming 2.2 release both attempts will fail with the error:
> >
> > Local variable is accessed at a different place, and therefore it must be
> > initialized and declared final.
> >
> >
> > So you cannot access a local variable in a different place unless it is
> > initialized and "val".
> >
> >
> > In 2.3 there will be a new construct called "athome" that will allow you to
> > initialize a local at a remote place by going back to the original place.
> >
> >
> >
> > On Thu, May 5, 2011 at 11:45 PM, Jeeva Paudel <je...@ualberta.ca> wrote:
> >
> > > Hi all,
> > >
> > > This is in reference to the Fibonacci example available at
> > > http://dist.codehaus.org/x10/applications/samples/Fib.x10
> > >
> > > Although async at != at async always, but shouldn't the following two
> > > versions of async produce similar effect?
> > >
> > > at (here) async { f1 = fib(n-1); } must start a new activity
> > located
> > > here and executing S
> > >
> > > async at (here) {fi = fib(n-1); } must also start a new activity
> > > here and execute S here as well.
> > >
> > > Aren't both the same and should compile correctly when the program
> > > compiles correctly with async { f1 = fib(n-1)} ? However, the compiler
> > > reports this error : "f1" may not have been initialized
> > >
> > >
> > > Thanks,
> > > Jeeva P.
> >
------------------------------------------------------------------------------
WhatsUp Gold - Download Free Network Management Software
The most intuitive, comprehensive, and cost-effective network
management toolset available today. Delivers lowest initial
acquisition cost and overall TCO of any competing solution.
http://p.sf.net/sfu/whatsupgold-sd
_______________________________________________
X10-users mailing list
X10-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/x10-users
