The Gray parser generator, written by Anton Ertl, takes the
grammar specification for a target language and generates a
parser for that language by directly compiling executable code
for the parser. This parser will then compile a program written
in the target language. As such Gray is an on-line tool as it
must be present whenever a target language program is compiled
and executed.

I've modified the code generation part of Gray to generate a
parser as Forth source code thereby making Gray an off-line tool
i.e. the generated parser may then be run independently of Gray.
I've tested it with the simple examples contained in the Gray
distibution and with simple examples of my own.

Does anyone have or know of a moderately complex Gray
application that I can feed into this tool in order to give it a
more severe test and improve confidence in its operation?

Gerry

Gerry jackson9000.fsnet.co.uk> writes Re: Modified version of the Gray parser generator
[..]

I could send you my PL1 language with (i)Forth backend. It uses
Gray, but I don't know if it tracks with the latest version.

It might not be easy to make it run, but it will test everything :-)

Here's a piece of PL1:

-- The standard Sieve ---------------------------------------------

anew -work PL/1

(* PL/1; the infamous Sieve *)

PROGRAM Sieve;

CONST
size=8190;

On 22 May, 21:09, m...@iae.nl (Marcel Hendrix) wrote:

[snip]

Yes that seems to be what I am looking for. I'd be grateful if you
could send me the files and I'll see what I can do with them.

Gerry jackson9000.fsnet.co.uk> writes:

Why do you want to do that?

I think that the possibility to generate executable code is one of the
cool features of Forth, and Gray demonstrates this nicely.

mhx@iae.nl (Marcel Hendrix) writes:

Note that Marcel's PL1 is not IBM's PL/I
<http://en.wikipedia.org/wiki/PL/I>, but is an extended version of
Wirth's PL/0 <http://en.wikipedia.org/wiki/PL/0>, which is a subset of
Pascal.

Gray has remained very stable in the last 15 years or so (basically
ever since it was converted to Forth-94). Of course, if you
iForthized it, all bets are off.

On 23 May, 18:38, an...@mips.complang.tuwien.ac.at (Anton Ertl) wrote:

First let me say that I have no wish to denigrate Gray at all, its a
nice application that works well. I can think of several reasons for
doing what I have (not all of which apply to me):

anton@mips.complang.tuwien.ac.at (Anton Ertl) writes Re: Modified version of the Gray parser generator

Of course not, I'm not a masochist :-)

Gerry jackson9000.fsnet.co.uk> writes:

...

I think it's harder. Now you have an extra generate step that you
have to manage.

The same is true for the generate-on-startup parser.

mhx@iae.nl (Marcel Hendrix) writes:

Yes, the first version was developed on a 68020-based machine, so I
could not test for alignment issues and did not program for them.
However, looking in the RCS, I see that I checked in an
alignment-clean version in 1993, so that has been fixed since then.

anton@mips.complang.tuwien.ac.at (Anton Ertl) writes:

And actually it already says in the CHANGES file of Release 2:

|Gray now runs on computers with memory alignment restrictions, too.

The dates of all files in that directory are June 28 1991.

- anton