Vijay and Dave,

Thanks for the clarification. I am currently taking fresh look at formalizing 
hb-relation with clocks, and wanted to check some things against the 
implementation. As long as I am clear on the expected behavior, the 
implementation of X10 is not on my critical path.

I created awake/dormant state of the clocks to express what happens in the 
nested case, but the stack description is much better.

Regards,
Tomofumi Yuki

----- Original Message -----
> From: "Vijay Saraswat" <vi...@saraswat.org>
> To: x10-users@lists.sourceforge.net
> Sent: Thursday, December 4, 2014 5:38:42 PM
> Subject: Re: [X10-users] Nested Implicit Clock Behavior
> 
> Tomo -- We just discussed this internally. The semantics is consistent
> with what you know it to be, but it appears the implementation has some
> problems which will be fixed (very soon I hope).
> 
> On 12/4/14, 11:30 AM, David P Grove wrote:
> >
> > Tomofumi Yuki <tomofumi.y...@inria.fr> wrote on 12/04/2014 10:34:01 AM:
> > >
> > > That is different from what I understood.
> > >
> > > In fact, there is only one line in the language specification about
> > > nested clocked finishes:
> > >   "Clocked finishes may be nested. The inner clocked finish operates
> > > in a single phase of the outer one."
> > >
> > > and your description of Clock.advanceAll() is colliding with this line.
> > >
> > > Personally, I see little utility for nested clocked finishes if
> > > inner scopes are still tied to outer clocks.
> > >
> >
> > The description earlier in the chapter is fairly clear on what
> > Clock.advanceAll() does.  It advances all clocks that the activity is
> > registered on.
> >
> > I agree that the line in the language specification you are quoting is
> > out of synch with this.
> >
> > We've discussed and the intent was that implicit clocks should act
> > like a stack of clocks with clocked finish pushing/popping a clock
> > from this stack when it is entered/exited.  Only the implicit clock on
> > the top of the implicit stack of clocks is advanced by
> > Clock.advanceAll().  All of the dynamically enclosing implicit clocks
> > are hidden by the newly entered clocked finish.
> >
> > We'll need to add a clarifying note of this masking effect of clocked
> > finish to the language specification and change the implementation to
> > comply.
> >
> > --dave
> >
> >
> > ------------------------------------------------------------------------------
> > Download BIRT iHub F-Type - The Free Enterprise-Grade BIRT Server
> > from Actuate! Instantly Supercharge Your Business Reports and Dashboards
> > with Interactivity, Sharing, Native Excel Exports, App Integration & more
> > Get technology previously reserved for billion-dollar corporations, FREE
> > http://pubads.g.doubleclick.net/gampad/clk?id=164703151&iu=/4140/ostg.clktrk
> >
> >
> > _______________________________________________
> > X10-users mailing list
> > X10-users@lists.sourceforge.net
> > https://lists.sourceforge.net/lists/listinfo/x10-users
> 
> 
> ------------------------------------------------------------------------------
> Download BIRT iHub F-Type - The Free Enterprise-Grade BIRT Server
> from Actuate! Instantly Supercharge Your Business Reports and Dashboards
> with Interactivity, Sharing, Native Excel Exports, App Integration & more
> Get technology previously reserved for billion-dollar corporations, FREE
> http://pubads.g.doubleclick.net/gampad/clk?id=164703151&iu=/4140/ostg.clktrk
> _______________________________________________
> X10-users mailing list
> X10-users@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/x10-users
> 

------------------------------------------------------------------------------
Download BIRT iHub F-Type - The Free Enterprise-Grade BIRT Server
from Actuate! Instantly Supercharge Your Business Reports and Dashboards
with Interactivity, Sharing, Native Excel Exports, App Integration & more
Get technology previously reserved for billion-dollar corporations, FREE
http://pubads.g.doubleclick.net/gampad/clk?id=164703151&iu=/4140/ostg.clktrk
_______________________________________________
X10-users mailing list
X10-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/x10-users

Reply via email to