Block closures
Stephan Rudlof
sr at evolgo.de
Thu Jul 27 21:46:16 UTC 2000
Mats Nygren wrote:
>
> Stephan Rudlof <sr at evolgo.de> wrote:
> > Mats,
> >
> > Mats Nygren wrote:
> > >
> > > Hi,
> > >
> > > I've had a look at the compiler and the VM.
> > >
> > > I can do the block closures.
> >
> > Very good! It's one of the things at my wish list and I've thought about
> > to put it onto my toDo list...
>
> Why didnt you ;->
Possibly I'll *try* to do this stuff (it's an interesting task!), but my
time for Squeak is very limited in the next time (as for you).
> > > But don't expect to see anything before the
> > > middle of October as I have other things that has to be done. I have the
> > > straighforard way of doing it in mind,
> >
> > I'v just read the purple book chapter about interpreter stuff, so for me
> > it wouldn't be straightforward with my current knowledge...
> >
> > > allocating block contexts.
> >
> > Do you really mean BlockContexts or rather BlockClosures?
>
> I don't understand the question. (Perhaps my reply to Lex is
> what your thinking about)
There are BlockContexts in the current implementation, but not
BlockClosures as in VisualWorks® 5i.1b Non-Commercial Jan 19 2000.
> > > No fancy
> > > optimizations.
> >
> > But blocks shouldn't be much slower than before then.
>
> They will use more memory.
Not an issue.
> I'm not into the minute details
> yet. See reply to Lex.
>
> > > The semantics I have in mind is as in ANSI Smalltalk, is that what is
> > > wanted?
> >
> > To be compatible to the standard seems to be reasonable for me.
> >
> > > In particular it will be possible to shadow variables in
> > > enclosing scopes.
> >
> > BlockClosures?
>
> Dont understand.
See VisualWorks® 5i.1b Non-Commercial (or commercial).
>
> >
> > > While at it I will restructure the compiler a little.
> >
> > Seems as if you're really experienced in compiler construction...
>
> Unfortunately not, but this isn't that big a thing. A working compiler
> is already there. The handling of variable names need to be changed
different scopes...
> and
> the details of how to manage blocks need to be changed a little,
The 'details' probably affect the VM.
BTW: Please post when you effectively start with this task.
Greetings,
Stephan
> that's
> all. And see my reply to Dwight Hughes.
>
> > Greetings,
> >
> > Stephan
> >
> > >
> > > /Mats
> >
> > --
> > Stephan Rudlof (sr at evolgo.de)
> > "Genius doesn't work on an assembly line basis.
> > You can't simply say, 'Today I will be brilliant.'"
> > -- Kirk, "The Ultimate Computer", stardate 4731.3
--
Stephan Rudlof (sr at evolgo.de)
"Genius doesn't work on an assembly line basis.
You can't simply say, 'Today I will be brilliant.'"
-- Kirk, "The Ultimate Computer", stardate 4731.3
More information about the Squeak-dev
mailing list
|