Re: Re : Re : Re : [Chicken-users] hacking the mailbox egg

[Kon Lovett]

> >
<snip>
> > >
> > > If the above is true, what are the real primitives of chicken (that
> > > can thus be used without 'getting out' of the file) ?
> >
> > I do not understand "... without 'getting out' of the file".
> Well, I meant "without running code that re-enables interrupts". And
> I guessed the "re-enabling" is done when executing code from
> "somewhere else". But I don't see what is the "somewhere else".
> I.e. I don't see what triggers the "re-enabling".

And you won't, unless you either disassemble or tell the compiler to  
keep the generated .c file(s). The declarations are file global &  
apply to every procedure - the compiler generates the interrupt stuff  
automatically. The compiler knows when a proc "calls out" & is  
"called in".

There is a manual wrapper, 'critical-section', but it is deprecated.

Mailbox is special. It could have been implemented using explicit  
synchronization but for such a high-performance & low computation  
facility it makes more sense to just treat every procedure as if, as  
you said, it were Java proc w/ synchronized.

There is an egg that provides syntactic-sugar for explicit  
synchronization, "synch.egg". It has a small user community & makes  
writing such code much easier.

<snip>