Re: doing exercises for starting forth, kinda discouraging, some are easy and some seem text not prepare me for....

Newsgroups	comp.lang.forth
From	Albert van der Horst <albert@spenarnc.xs4all.nl>
Subject	Re: doing exercises for starting forth, kinda discouraging, some are easy and some seem text not prepare me for....
Date	2012-07-19 19:15 +0000
Message-ID	<m7f9hk.6iz@spenarnc.xs4all.nl> (permalink)
Organization	Dutch Forth Workshop
References	<2a5baeb8-2c72-41a6-8648-d52842405f18@n16g2000vbn.googlegroups.com> <wbOdnWwHX40ca5rNnZ2dnUVZ8lSdnZ2d@supernews.com> <2012Jul19.144859@mips.complang.tuwien.ac.at> <xdGdnToPTLAKjpXNnZ2dnUVZ8nWdnZ2d@supernews.com>

Show all headers | View raw

In article <xdGdnToPTLAKjpXNnZ2dnUVZ8nWdnZ2d@supernews.com>,
Andrew Haley  <andrew29@littlepinkcloud.invalid> wrote:
>Anton Ertl <anton@mips.complang.tuwien.ac.at> wrote:
>> Andrew Haley <andrew29@littlepinkcloud.invalid> writes:
>>>Albert van der Horst <albert@spenarnc.xs4all.nl> wrote:
>>>> In article <ju7ifu02cet@enews2.newsguy.com>, WJ <w_a_x_man@yahoo.com> wrote:
>>>>>Anton Ertl wrote:
>>>>>
>>>>>> BTW, the source code for ;S and ?BRANCH is:
>>>>>>
>>>>>> ;s   ( R:w -- )              gforth  semis
>>>>>> ""The primitive compiled by @code{EXIT}.""
>>>>>> #ifdef NO_IP
>>>>>> INST_TAIL;
>>>>>> goto *(void *)w;
>>>>>> #else
>>>>>> SET_IP((Xt *)w);
>>>>>> #endif
>>>>>>
>>>>>> condbranch(?branch,f --,f83  question_branch,
>>>>>> ,if (f==0) {
>>>>>> ,:
>>>>>>  0= dup 0=          \ !f f
>>>>>>  r> tuck cell+      \ !f branchoffset f IP+
>>>>>>  and -rot @ and or  \ f&IP+|!f&branch
>>>>>>  >r ;)
>>>>>>
>>>>>> There's some m4 in this, some Vmgen, some C, and some Forth.
>>>>>
>>>>>KISS
>>>>>
>>>>>(Keep it simple, stupid.)
>>>>
>>>> I can confirm that m4 is sometimes just the best solution for
>>>> a problem. To generate ciforth's I use m4 to massage assembler
>>>> files.
>>>
>>>Sure, but if your Forth was written in Forth you wouldn't have to do
>>>any of this stuff.
>>
>> If my Forth was written in Forth (with built-in Forth assembler), it
>> would not be portable.
>
>I accept that.  You have a particular set of problems that have to be
>solved in some way.  That doesn't apply to Albert, though.

It would be weird to say I have problems.
I create portability opportunities.

All intel ciforth's are generated from one file.
The difference between generating a linux 32 and a 64 bit Forth
is
1. use a configuration file that specifies _BITS32_ instead of _BITS64_
2. include a width32.m4 instead of the widht64.m4 file.
  (Done automatically by using _BITS32_)
  There are about 10 lines different in the width file,
  pretty trivially different like

diff width*.
>define({AX},{EAX})dnl
<define({AX},{RAX})dnl

>define({CELL_M4},{LONG})dnl
<define({CELL_M4},{QUAD})dnl

>define({LODSW},{LODSD})dnl
<define({LODSW},{LODSQ})dnl

The difference between DPMI MS-windows and DLL MS-windows and linux
is similar.

Building a system includes making tailor made documentation, so
you read " you have a 64 bits system." not "depending on the
CELL-SIZE of your system FM/MOD works on a 128 or a 64 bit ..."

The portability of ciforth and gforth is so different that it is
hard to compare. Both are totally different from the kind of
portability obtained using a Forth compiled in Forth.

Let's define the challenge to build a new Forth system with
e.g. the words 2/ and 2* removed. This is to be executed by
some one with no prior knowledge.
This challenge will be won by ciforth, hands down.
The fact alone that there is just one file to change, and
that that person need not involve the generic system
to build ciforth's makes me sure of that.

I'm also sure that one can make up a challenge that is easy
for gForth and near-impossible for ciforth or for a meta-compiled
Forth.

>
>Andrew.

Groetjes Albert

--
-- 
Albert van der Horst, UTRECHT,THE NETHERLANDS
Economic growth -- being exponential -- ultimately falters.
albert@spe&ar&c.xs4all.nl &=n http://home.hccnet.nl/a.w.m.van.der.horst

Back to comp.lang.forth | Previous | Next — Previous in thread | Next in thread | Find similar

Thread

csiph-web