Re: [stack] small additions to interp.c
John Cowan — 2001-05-17 15:50:32
Manfred von Thun wrote:
> The first new operator is "rotate" for interchanging the
> first and third element of the stack. The other new operators
> are variants of the (now 6) ordinary stack shuffling operators.
> For X = (pop,dup,swap,rollup,rolldown,rotate), a variant
> is introduced called "Xd", with the "d" for "dipped".
Integrating these will be trivial. However, I do not have C
code for rollup, rolldown yet! Nor is it on the Web site.
--
There is / one art || John Cowan <
jcowan@...>
no more / no less ||
http://www.reutershealth.com
to do / all things ||
http://www.ccil.org/~cowan
with art- / lessness \\ -- Piet Hein
Manfred von Thun — 2001-05-22 01:02:18
On Thu, 17 May 2001, John Cowan wrote:
> Manfred von Thun wrote:
>
> > The first new operator is "rotate" for interchanging the
> > first and third element of the stack. The other new operators
> > are variants of the (now 6) ordinary stack shuffling operators.
> > For X = (pop,dup,swap,rollup,rolldown,rotate), a variant
> > is introduced called "Xd", with the "d" for "dipped".
>
>
> Integrating these will be trivial. However, I do not have C
> code for rollup, rolldown yet! Nor is it on the Web site.
Oops. So sorry. Here it is:
--- the three stack shuffling operators affecting the top three
--- items on the stack. Insert this just after "swap"
PRIVATE void rollup_()
{
THREEPARAMS("rollup");
SAVESTACK;
TERNARY(SAVED1->op,SAVED1->u.num);
NULLARY(SAVED3->op,SAVED3->u.num);
NULLARY(SAVED2->op,SAVED2->u.num);
POP(dump);
}
PRIVATE void rolldown_()
{
THREEPARAMS("rolldown");
SAVESTACK;
TERNARY(SAVED2->op,SAVED2->u.num);
NULLARY(SAVED1->op,SAVED1->u.num);
NULLARY(SAVED3->op,SAVED3->u.num);
POP(dump);
}
PRIVATE void rotate_()
{
THREEPARAMS("rotate");
SAVESTACK;
TERNARY(SAVED1->op,SAVED1->u.num);
NULLARY(SAVED2->op,SAVED2->u.num);
NULLARY(SAVED3->op,SAVED3->u.num);
POP(dump);
}
--- The online manual for the above three operators.
--- Insert this just after the online entries for "swap".
{"rollup", rollup_, "X Y Z -> Z X Y",
"Moves X and Y up, moves Z down"},
{"rolldown", rolldown_, "X Y Z -> Y Z X",
"Moves Y and Z down, moves X up"},
{"rotate", rotate_, "X Y Z -> Z Y X",
"Interchanges X and Z"},
--- You already have the additions for the "dipped" versions.
Very sorry about wasting your time.
- Manfred
John Cowan — 2001-05-22 02:57:42
Manfred von Thun scripsit:
> On Thu, 17 May 2001, John Cowan wrote:
>
> > Manfred von Thun wrote:
> >
> > > The first new operator is "rotate" for interchanging the
> > > first and third element of the stack. The other new operators
> > > are variants of the (now 6) ordinary stack shuffling operators.
> > > For X = (pop,dup,swap,rollup,rolldown,rotate), a variant
> > > is introduced called "Xd", with the "d" for "dipped".
> >
> > Integrating these will be trivial. However, I do not have C
> > code for rollup, rolldown yet! Nor is it on the Web site.
>
> Oops. So sorry. Here it is:
Integrated. The new version is available at the usual place:
http://www.ccil.org/~cowan/joy.tar.gz .
--
John Cowan
cowan@...
One art/there is/no less/no more/All things/to do/with sparks/galore
--Douglas Hofstadter
Manfred von Thun — 2001-05-22 04:30:04
On Mon, 21 May 2001, John Cowan wrote:
> Manfred von Thun scripsit:
> > Oops. So sorry. Here it is:
>
> Integrated. The new version is available at the usual place:
> http://www.ccil.org/~cowan/joy.tar.gz .
(Tuesday, about an hour after I received your message above)
I think you left out interp.c in the tar-file.
I did:
gunzip joy.tar.gz no complaints
tar -xf joy.tar no complaints
make the usual very long messages
wizzing by, and finally:
gcc: interp.o No such file or directory
make[1]: *** [joy] Error 1
make[1]: Leaving directory `home/staff/manfred/cowan'
make: ***[all] Error 2
ls no interp.c no interp.o
more joy.tar
looking for interp.c or for
stuff inside that - no luck
Just an oversight, I suppose.
- Manfred
Manfred von Thun — 2001-05-22 07:11:02
On Mon, 21 May 2001, John Cowan wrote:
>
> Integrated. The new version is available at the usual place:
> http://www.ccil.org/~cowan/joy.tar.gz .
>
I know you already have implemented "fold", but I have not seen
it yet. Before you fix the tar-file with te missing interp.c,
would you please consider the following implemenation of "fold".
It really mimicks the definition
fold == [swap] dip step
that I had given earlier. Now we have swapd, we might as well use it.
This can go anywhere after all the DIPPED definitions, perhaps
just after step.
PRIVATE void fold_()
{
swapd_(); step_();
}
The online manual entry, also just after the "step" entry:
{"fold", fold_, "A V0 BIN -> V1",
"Starting with value V0, sequentially puts members of aggregate A onto stack andcombines with binary operator BIN to finally prodice value V1"},
I hope you are not getting tired of all this, but that is the
end of my wishlist that I foreshadowed a while back.
Thanks for all your work.
- Manfred
John Cowan — 2001-05-22 11:04:41
Manfred von Thun scripsit:
> I know you already have implemented "fold", but I have not seen
> it yet. Before you fix the tar-file with the missing interp.c,
> would you please consider the following implemenation of "fold".
> It really mimicks the definition
> fold == [swap] dip step
> that I had given earlier. Now we have swapd, we might as well use it.
Done.
> I hope you are not getting tired of all this,
Not at all! Working on Joy continues to be a joy.
--
John Cowan
cowan@...
One art/there is/no less/no more/All things/to do/with sparks/galore
--Douglas Hofstadter
Manfred von Thun — 2001-05-23 07:56:04
On Tue, 22 May 2001, John Cowan wrote:
> Manfred von Thun scripsit:
> > I hope you are not getting tired of all this,
> Not at all! Working on Joy continues to be a joy.
> --
> John Cowan cowan@...
I unpacked your latest version of joy.tar.gz
Congratulations. I tried:
1. Sun under Linux, "make" got joy with BDW seems A1 OK
2. Alpha under DEC-unix, "make" got joy with BDW seems A1 OK
3. Alpha under DEC VMS, I only tried manual CC for the joy files,
got joy with NOBDW seems A1 OK
I'll have to try to understand the three makefiles that you
supplied to translate into VMS. That will be hard for me.
(Incidentally, your joy1.exe is now 36% bigger than my joy0.exe)
So, Joymaster, well done, and thank you very very much.
I hope that others in the group will unpack your latest version
to see how it works on their machines.
I shall draft a notice about your joy1 to go to the comp.lang.functional
newsgroup. I'll send it to concatenative first to hear comments
from the group and especially from you.
- Manfred
John Cowan — 2001-05-23 11:22:01
Manfred von Thun scripsit:
> 3. Alpha under DEC VMS, I only tried manual CC for the joy files,
> got joy with NOBDW seems A1 OK
> I'll have to try to understand the three makefiles that you
> supplied to translate into VMS. That will be hard for me.
The general logic is:
makefile does a "make test" in the gc directory. If that works, it
creates gc.succ as a sentinel so it doesn't have to do it again,
and invokes make.gc to compile the joy files with GC_BDW defined
and the gc library linked in. If it doesn't work, then it creates
gc.fail as a sentinel, and invokes make.nogc to compile the
joy files with GC_BDW not defined. There is probably a cleverer
way to do this, but I am not a "make" maven.
Anyway, the gc library is not supported under VMS, so a sensible
VMS build procedure would just ignore the gc library and compile
the joy files directly, as you did. If and when there is
VMS support, there will presumably be a VMS build procedure
for gc as well. The gc/README.alpha file implies that VMS support
would not be trivial.
> (Incidentally, your joy1.exe is now 36% bigger than my joy0.exe)
Sounds about right. It will require more memory at run time, too, because
Nodes are bigger.
> So, Joymaster, well done, and thank you very very much.
> I hope that others in the group will unpack your latest version
> to see how it works on their machines.
Yes, please!
> I shall draft a notice about your joy1 to go to the comp.lang.functional
> newsgroup. I'll send it to concatenative first to hear comments
> from the group and especially from you.
Great.
--
John Cowan
cowan@...
One art/there is/no less/no more/All things/to do/with sparks/galore
--Douglas Hofstadter
Manfred von Thun — 2001-05-24 06:28:00
On Wed, 23 May 2001, John Cowan wrote:
> Manfred von Thun scripsit:
[Stuff about gcc under VMS deleted, thanks John]
> > I shall draft a notice about your joy1 to go to the comp.lang.functional
> > newsgroup. I'll send it to concatenative first to hear comments
> > from the group and especially from you.
>
> Great.
Here is my draft:
newsgroup: comp.lang.functional
subject: ANNOUNCE: enhancements to Joy
I am please to announce John Cowan's substantial enhancements to
the prototype C-implemenation of Joy, a purely functional language.
John's additions are:
1. Floats (real numbers) as a new datatype, together with
numerous new functions over floats.
2. Files as a datatype, together with some common functions
for manipulation files and IO.
3. Some new string functions. Also, for unix?/gcc? only, an option
to include garbage collection for strings.
4. Some bugs in my prototype have been fixed. Portability
has been further increased.
It is intended that the new implementation becomes known as Joy1,
and that my earlier prototype, now Joy0, should gradually fade away.
I am enormously grateful to John for his work.
- Manfred von Thun
The new Joy1 is available
1. directly from John Cowan:
http://www.ccil.org/~cowan/joy.tar.gz
or 2. from the Joy site:
http://www.latrobe.edu.au/www/philosophy/phimvt/??? (not yet!)
Note: The main page of the Joy site is
http://www.latrobe.edu.au/www/philosophy/phimvt/j00syn.html
From the main page:
The language Joy is a purely functional programming language.
Whereas all other functional programming languages are based
on the application of functions to arguments, Joy is based on
the composition of functions. All such functions take a
stack as argument and produce a stack as value. Consequently
much of Joy looks like ordinary postfix notation. However,
in Joy a function can consume any number of parameters from
the stack and leave any number of results on the stack. The
concatenation of appropriate programs denotes the composition
of the functions which the programs denote. One of the
datatypes of Joy is that of quoted programs, of which lists
are a special case. Some functions expect quoted programs
on top of the stack and execute them in many different ways,
effectively by dequoting. So, where other functional languages
use abstraction and application, Joy uses quotation and
combinators -- functions which perform dequotation. As a
result, there are no named formal parameters, no substitution
of actual for formal parameters, and no environment of
name-value pairs. Combinators in Joy behave much like
functionals or higher order functions in other languages,
they minimise the need for recursive and non-recursive
definitions. Joy has a rich but simple algebra, and its
programs are easily manipulated by hand and by other programs.
John Cowan — 2001-05-24 14:29:05
Manfred von Thun wrote:
> Here is my draft:
Here's my slightly edited version.
Newsgroups: comp.lang.functional
Subject: ANNOUNCE: enhancements to Joy
I am pleased to announce John Cowan's substantial enhancements to
the prototype C implementation of Joy, a purely functional language.
John's additions are:
1. Floats (real numbers) as a new datatype, together with
numerous new functions over floats.
2. Files as a datatype, together with some common functions
for manipulating files and doing I/O.
3. Some new string functions. Also, on most platforms, an option
to include garbage collection for strings.
4. Some bugs in my prototype have been fixed. Portability
has been further increased. Joy1 will still compile and run
with any strictly conforming ISO C compiler.
The currently supported platforms for the general garbage collector
include Unix/Linux/*BSD on i386/Alpha/SPARC/SGI/HP, OS/2,
Win32 native/Cygwin/DJGPP, and perhaps Macintosh. One system
known *not* to work is OpenVMS. Joy1 has not yet been built on all
these systems.
It is intended that the new implementation become known as Joy1,
and that my earlier prototype, now Joy0, should gradually fade away.
I am enormously grateful to John for his work.
- Manfred von Thun
The new Joy1 is available:
1. directly from John Cowan:
http://www.ccil.org/~cowan/joy.tar.gz
or 2. from the Joy site:
http://www.latrobe.edu.au/www/philosophy/phimvt/??? (not yet!)
Note: The main page of the Joy site is
http://www.latrobe.edu.au/www/philosophy/phimvt/j00syn.html
--
There is / one art || John Cowan <
jcowan@...>
no more / no less ||
http://www.reutershealth.com
to do / all things ||
http://www.ccil.org/~cowan
with art- / lessness \\ -- Piet Hein
Manfred von Thun — 2001-05-29 04:07:36
On Thu, 24 May 2001, John Cowan wrote:
>
> Here's my slightly edited version.
>
Thanks, John. I am currently snowed under with end-of-semester marking,
so it will probably be next week till I really get aroung to it.
- Manfred
Manfred von Thun — 2001-05-31 09:19:39
On Tue, 29 May 2001, Manfred von Thun wrote:
>
> On Thu, 24 May 2001, John Cowan wrote:
> >
> > Here's my slightly edited version.
> >
> Thanks, John. I am currently snowed under with end-of-semester marking,
> so it will probably be next week till I really get aroung to it.
I have now made the changes to the main page of Joy -
http://www.latrobe.edu,au/www/philos/phimvt/j00syn.html
(except that the link to the joy.tar.gz currently still points
to John's site. I'll keep a local copy when advised to.)
Please have a look, let me know what you think.
I will wait with the announcement to comp.lang.functional
and I think I should also send one to comp.lang.misc
But I'll wait for a week or so, in case there are any
changes I should make first to the Joy web page.
- Manfred