From Newsgroup: comp.lang.forth

Do you know practical cases of using "include-file" in programs?

I can't imagine how this word could be used in standard programs.

Only one idea: to skip BOM (byte-order mark) before include the file
contents.

Using other system-specific means, this word can probably be used to
organize inter-process communication: when a file descriptor (e.g. a
pipe) is passed from one process to another and used as the input
source. But why do you need to load Forth code this way?


--
Ruvim

--- Synchronet 3.20a-Linux NewsLink 1.114

From Newsgroup: comp.lang.forth

On Mon, 2 Sep 2024 20:11:25 +0000, Ruvim wrote:

Do you know practical cases of using "include-file" in programs?

[..]

Using other system-specific means, this word can probably be used to
organize inter-process communication: when a file descriptor (e.g. a
pipe) is passed from one process to another and used as the input
source. But why do you need to load Forth code this way?

My SPICE simulator INCLUDE's pre-processed netlists.
A pre-processor takes (again, INCLUDE's) the industrial-standard
netlist and rewrites it to Forth.
One can also write the netlist directly using Forth, but not
every engineer is able or willing to do that.

Debugging is easy: the original netlist runs on standard SPICE
programs and its results can be compared directly to iSPICE's
output when it uses the pre-processed netlist.

The principle works for other applications too. Here some
examples from my dfwforth/include directory:

FORTH> grep "INCLUDE" /mnt/d/dfwforth/include/*.frt
/assemble.frt:319: =ASM-NAME COUNT INCLUDED
/backtrac.frt:132: ['] INCLUDE-FILE OF
/dialogs.frt:116: R> COUNT INCLUDED
/eplot.frt:777: SWAP @ ?DUP IF COUNT INCLUDED ( header) ENDIF /eplot.frt:796: SWAP @ ?DUP IF COUNT INCLUDED ( header) ENDIF /eplot.frt:819: SWAP @ ?DUP IF COUNT INCLUDED ( header) ENDIF /eplot.frt:838: SWAP @+ SWAP >S ?DUP IF COUNT INCLUDED ( head /eplot.frt:882: :headname @ ?DUP IF COUNT INCLUDED
/eplot.frt:895: :dataname @ COUNT INCLUDED ;
/graphbench.frt:261:#63 #check [IF] INCLUDE gtk3/benchmarks.frt [T /iwords1.frt:4:CR .( FORTH.LOG : edit out garbage first, then INCLUDE i /lprof.frt:357:\ Instead of INCLUDE name , IN name , S" name" INCLUDED /lprof.frt:367: S" !!!!!!!!.$$$" INCLUDED
/lprofile.frt:198:\ Instead of INCLUDE name , IN name , S" name" INCLUD /lprofile.frt:207: S" !!!!!!!!.$$$" INCLUDED ; /playwav.frt:17:[IF] INCLUDE pwavlinu.frt
/playwav.frt:21:[IF] INCLUDE pwavlinu.frt
/playwav.frt:25:[IF] INCLUDE pwavewin.frt
/runsteady.frt:620: inpbuf #11 cnt MIN S" *| INCLUDE " COMP /syssim.frt:772: EVAL" ANEW __syssim__" S" spf_model.frt" INCLUDED ; /syssim.frt:812: CR ." SPF-INCLUDE -- load spf_model.frt control/simu

-marcel
--- Synchronet 3.20a-Linux NewsLink 1.114

From Newsgroup: comp.lang.forth

On Mon, 2 Sep 2024 21:22:14 +0000, mhx wrote:

On Mon, 2 Sep 2024 20:11:25 +0000, Ruvim wrote:

Do you know practical cases of using "include-file" in programs?

[..]

Sorry, I missed the point of your question.

No, I never have seen any use for the FILE wordset's
INCLUDE-FILE ( i*x fileid -- j*x ).

-marcel
--- Synchronet 3.20a-Linux NewsLink 1.114

From Newsgroup: comp.lang.forth

Ruvim <[email protected]> writes:

Do you know practical cases of using "include-file" in programs?

One way I am thinking about is to implement INCLUDED etc. Systems
deal differently with relative filenames for INCLUDED, with less
differences for OPEN-FILE. So if I want to implement a certain
handling of relative filenames for INCLUDED in order to get rid of
system differences, one way to do it is to do it for OPEN-FILE and
then perform INCLUDE-FILE.

Using other system-specific means, this word can probably be used to >organize inter-process communication: when a file descriptor (e.g. a
pipe) is passed from one process to another and used as the input
source. But why do you need to load Forth code this way?

Heinz Schnitter implemented Open Network Forth, a distributed system,
where the parts communicated by sending Forth source code to each
other. One could implement that by having network connections
(telnet-style or ssh-style) between the parts, and the receiver could INCLUDE-FILE the file-id that represents his end of the connection.
However, Open Network Forth was written around 1984, long before
INCLUDE-FILE was standardized. Bernd Paysan uses ideas from it in his
work, inclluding sending Forth commands between tasks, but AFAIK does
not use INCLUDE-FILE for that.

- anton
--
M. Anton Ertl http://www.complang.tuwien.ac.at/anton/home.html
comp.lang.forth FAQs: http://www.complang.tuwien.ac.at/forth/faq/toc.html
New standard: https://forth-standard.org/
EuroForth 2024: https://euro.theforth.net
--- Synchronet 3.20a-Linux NewsLink 1.114

From Newsgroup: comp.lang.forth

In article <[email protected]>,
Anton Ertl <[email protected]> wrote:

Ruvim <[email protected]> writes:

Do you know practical cases of using "include-file" in programs?

One way I am thinking about is to implement INCLUDED etc. Systems
deal differently with relative filenames for INCLUDED, with less
differences for OPEN-FILE. So if I want to implement a certain
handling of relative filenames for INCLUDED in order to get rid of
system differences, one way to do it is to do it for OPEN-FILE and
then perform INCLUDE-FILE.

Using other system-specific means, this word can probably be used to >>organize inter-process communication: when a file descriptor (e.g. a
pipe) is passed from one process to another and used as the input
source. But why do you need to load Forth code this way?

Heinz Schnitter implemented Open Network Forth, a distributed system,
where the parts communicated by sending Forth source code to each
other. One could implement that by having network connections
(telnet-style or ssh-style) between the parts, and the receiver could >INCLUDE-FILE the file-id that represents his end of the connection.
However, Open Network Forth was written around 1984, long before
INCLUDE-FILE was standardized. Bernd Paysan uses ideas from it in his
work, inclluding sending Forth commands between tasks, but AFAIK does
not use INCLUDE-FILE for that.

In other words. The use of INCLUDE-FILE in the wild has never been
seen, except for a factor of INCLUDE.

- anton

Groetjes Albert
--
Temu exploits Christians: (Disclaimer, only 10 apostles)
Last Supper Acrylic Suncatcher - 15Cm Round Stained Glass- Style Wall
Art For Home, Office And Garden Decor - Perfect For Windows, Bars,
And Gifts For Friends Family And Colleagues.
--- Synchronet 3.20a-Linux NewsLink 1.114

From Newsgroup: comp.lang.forth

off-topic:

Your SIG. Is Novabbs injecting ads, or is this a joke that
I don't get?

This is the third time I see it (from 2 different posters).

-marcel
--- Synchronet 3.20a-Linux NewsLink 1.114

From Newsgroup: comp.lang.forth

mhx <[email protected]> wrote:

off-topic:

Your SIG. Is Novabbs injecting ads, or is this a joke that

Perhaps it's the client, which the user should have some control.
In my client if signature is configured to point to a directory,
it will pull from there a random file to get text for the
signature.

Also found an option to allow the 'User-Agent' to place advertising
in headers and the default is "ON". But I haven't notice such
occuring.
--
me
--- Synchronet 3.20a-Linux NewsLink 1.114

From Newsgroup: comp.lang.forth

On 2024-09-03 10:50, Anton Ertl wrote:

Ruvim <[email protected]> writes:

Do you know practical cases of using "include-file" in programs?

One way I am thinking about is to implement INCLUDED etc. Systems
deal differently with relative filenames for INCLUDED, with less
differences for OPEN-FILE. So if I want to implement a certain
handling of relative filenames for INCLUDED in order to get rid of
system differences, one way to do it is to do it for OPEN-FILE and
then perform INCLUDE-FILE.

Yes, overriding the system's file search mechanism used by "included" is
a very interesting idea!

Using other system-specific means, this word can probably be used to
organize inter-process communication: when a file descriptor (e.g. a
pipe) is passed from one process to another and used as the input
source. But why do you need to load Forth code this way?

Heinz Schnitter implemented Open Network Forth, a distributed system,
where the parts communicated by sending Forth source code to each
other. One could implement that by having network connections
(telnet-style or ssh-style) between the parts, and the receiver could INCLUDE-FILE the file-id that represents his end of the connection.
However, Open Network Forth was written around 1984, long before
INCLUDE-FILE was standardized. Bernd Paysan uses ideas from it in his
work, inclluding sending Forth commands between tasks, but AFAIK does
not use INCLUDE-FILE for that.

Such a system must probably be multitasking. Otherwise, the Forth system instance that runs "read-file" is a slaver of the corresponding pipe
emitter until the pipe is closed (i.e., until the master wants).





Taking my idea of skipping a BOM (byte-order mark), a program can also
skip a shebang line, like this:

#! /usr/bin/env forth

I know Gforth has the word "#!" that comments out the rest of the line.

Using "open-file", "read-line", "reposition-file" and "include-file" is
an alternative way for a portable program. Yes, it seems far more
complex than a word "#!". But what if you need "#!" for something else (including recognizable things).



--
Ruvim

--- Synchronet 3.20a-Linux NewsLink 1.114

From Newsgroup: comp.lang.forth

On 4/09/2024 1:28 am, sjack wrote:

mhx <[email protected]> wrote:

off-topic:

Your SIG. Is Novabbs injecting ads, or is this a joke that

Perhaps it's the client, which the user should have some control.
In my client if signature is configured to point to a directory,
it will pull from there a random file to get text for the
signature.

First thing I find myself doing in any email/client is disabling the
default-on sig - typically an advertisement for the company that created
it. Were it possible to stop it showing sigs on received msgs, I'd do
that too.


--- Synchronet 3.20a-Linux NewsLink 1.114

From Newsgroup: comp.lang.forth

On 3/09/2024 6:23 pm, [email protected] wrote:

...
In other words. The use of INCLUDE-FILE in the wild has never been
seen, except for a factor of INCLUDE.

If the ANS file words were modelled after an existing implementation it
might explain INCLUDE-FILE's presence in the Standard. When text-file
handling was added to DX-Forth I had an equivalent function. It was necessarily a factor so as to be able to apply CATCH. Whether I was
influenced by ANS' INCLUDE-FILE I don't recall. All I know is I made
the factor headerless having apparently seen no further use for it.

--- Synchronet 3.20a-Linux NewsLink 1.114

From Newsgroup: comp.lang.forth

On 02/09/2024 21:11, Ruvim wrote:

Do you know practical cases of using "include-file" in programs?

I can't imagine how this word could be used in standard programs.

Only one idea: to skip BOM (byte-order mark) before include the file contents.

Using other system-specific means, this word can probably be used to organize inter-process communication: when a file descriptor (e.g. a
pipe) is passed from one process to another and used as the input
source.  But why do you need to load Forth code this way?

I've used it to redefine INCLUDED

: included ( ... caddr u -- ... )
r/o open-path-file throw include-file
;

where OPEN-PATH-FILE takes a string specifying a list of alternative
relative paths to Forth source files which it tries in turn to open. It
does this by creating an absolute directory path and calls OPEN-FILE. If
an open succeeds it returns the file-id to be included by INCLUDE-FILE.
If no relative path succeeds an exception is thrown.

I've used it to test a program on several different Forth systems, each
of which has its own way of handling relative directory paths but they
all work with an absolute path.
--
Gerry

--- Synchronet 3.20a-Linux NewsLink 1.114

From Newsgroup: comp.lang.forth

On 04/09/2024 10:54, Gerry Jackson wrote:

On 02/09/2024 21:11, Ruvim wrote:

Do you know practical cases of using "include-file" in programs?

I can't imagine how this word could be used in standard programs.

Only one idea: to skip BOM (byte-order mark) before include the file
contents.

Using other system-specific means, this word can probably be used to
organize inter-process communication: when a file descriptor (e.g. a
pipe) is passed from one process to another and used as the input
source.  But why do you need to load Forth code this way?

I've used it to redefine INCLUDED

: included  ( ... caddr u -- ... )
   r/o open-path-file throw include-file
;

where OPEN-PATH-FILE takes a string specifying a list of alternative relative paths to Forth source files which it tries in turn to open. It
does this by creating an absolute directory path and calls OPEN-FILE. If
an open succeeds it returns the file-id to be included by INCLUDE-FILE.
If no relative path succeeds an exception is thrown.

I've used it to test a program on several different Forth systems, each
of which has its own way of handling relative directory paths but they
all work with an absolute path.

I've just seen that Anton has suggested this as a possible use of OPEN-FILE
--
Gerry

--- Synchronet 3.20a-Linux NewsLink 1.114

From Newsgroup: comp.lang.forth

On 3/09/2024 6:11 am, Ruvim wrote:

Do you know practical cases of using "include-file" in programs?

I can't imagine how this word could be used in standard programs.
...

Here's a question. Is there a temporal dependency on the 'fileid' passed
to INCLUDE-FILE ?

On my system 'fileid' for INCLUDE-FILE would need to represent the most recently opened file. Otherwise should a compilation error occur the wrong statistics e.g. file-name will be printed.

AFAICS ANS only guarantees the contents of a file represented by 'fileid'.
The issue of filenames - if/when/how they're stored - is a detail left to the implementer. It's no problem for INCLUDED which binds opening of the file
with compilation. However INCLUDE-FILE potentially separates these two.

--- Synchronet 3.20a-Linux NewsLink 1.114

From Newsgroup: comp.lang.forth

On 2024-09-04 15:13, dxf wrote:

On 3/09/2024 6:11 am, Ruvim wrote:

Do you know practical cases of using "include-file" in programs?

I can't imagine how this word could be used in standard programs.
...

Here's a question. Is there a temporal dependency on the 'fileid' passed
to INCLUDE-FILE ?

The section "11.6.1.1717 INCLUDE-FILE" [1] does not impose any such restriction on a program.

A general rule is that a standard program is allowed to do anything that
is not explicitly prohibited for a program by the standard, a standard
system is only allowed to do anything that is explicitly allowed for a
system by the standard or undetectable by a standard program.

On my system 'fileid' for INCLUDE-FILE would need to represent the most recently opened file. Otherwise should a compilation error occur the wrong statistics e.g. file-name will be printed.

A compilation error is an ambiguous condition, so the system is allowed
to do whatever it wants. If the systems shows a wrong file name where an
error occurred, this does not make it non-compliant.

AFAICS ANS only guarantees the contents of a file represented by 'fileid'. The issue of filenames - if/when/how they're stored - is a detail left to the implementer. It's no problem for INCLUDED which binds opening of the file with compilation. However INCLUDE-FILE potentially separates these two.

Yes. OTOH, a system can obtain a file name from a file handle, see GetFileInformationByHandleEx in Windows, readlink on /proc/self/fd/NNN
in POSIX.


--
Ruvim

--- Synchronet 3.20a-Linux NewsLink 1.114

From Newsgroup: comp.lang.forth

On 2024-09-03 23:07, Ruvim wrote:

On 2024-09-03 10:50, Anton Ertl wrote:

Ruvim <[email protected]> writes:

Do you know practical cases of using "include-file" in programs?

One way I am thinking about is to implement INCLUDED etc.  Systems
deal differently with relative filenames for INCLUDED, with less
differences for OPEN-FILE.  So if I want to implement a certain
handling of relative filenames for INCLUDED in order to get rid of
system differences, one way to do it is to do it for OPEN-FILE and
then perform INCLUDE-FILE.

Yes, overriding the system's file search mechanism used by "included" is
a very interesting idea!

Another possible reason to not use `INCLUDED`:

| INCLUDED may allocate memory in data space before
| it starts interpreting the file.

So if this effect is unacceptable (for example, if you include part of a definition), then the solution is `INCLUDE-FILE`.


Also, `REQUIRED` does not load a file that was loaded by `INCLUDED`. `INCLUDE-FILE` is free of that too.


--
Ruvim

--- Synchronet 3.20a-Linux NewsLink 1.114

From Newsgroup: comp.lang.forth

Ruvim <[email protected]> writes:

Also, `REQUIRED` does not load a file that was loaded by `INCLUDED`. >`INCLUDE-FILE` is free of that too.

Yes, but why would anyone use REQUIRED or REQUIRE on a file that he
intends to include even if it has been INCLUDED before?

- anton
--
M. Anton Ertl http://www.complang.tuwien.ac.at/anton/home.html
comp.lang.forth FAQs: http://www.complang.tuwien.ac.at/forth/faq/toc.html
New standard: https://forth-standard.org/
EuroForth 2024: https://euro.theforth.net
--- Synchronet 3.20a-Linux NewsLink 1.114

From Newsgroup: comp.lang.forth

In article <[email protected]>,
Anton Ertl <[email protected]> wrote:

Ruvim <[email protected]> writes:

Also, `REQUIRED` does not load a file that was loaded by `INCLUDED`. >>`INCLUDE-FILE` is free of that too.

Yes, but why would anyone use REQUIRED or REQUIRE on a file that he
intends to include even if it has been INCLUDED before?

Making use of the idempotent behaviour of REQUIRE ?

- anton
--
M. Anton Ertl http://www.complang.tuwien.ac.at/anton/home.html >comp.lang.forth FAQs: http://www.complang.tuwien.ac.at/forth/faq/toc.html
New standard: https://forth-standard.org/
EuroForth 2024: https://euro.theforth.net

--
Temu exploits Christians: (Disclaimer, only 10 apostles)
Last Supper Acrylic Suncatcher - 15Cm Round Stained Glass- Style Wall
Art For Home, Office And Garden Decor - Perfect For Windows, Bars,
And Gifts For Friends Family And Colleagues.
--- Synchronet 3.20a-Linux NewsLink 1.114

From Newsgroup: comp.lang.forth

On Mon, 2 Sep 2024 20:11:25 +0000, Ruvim wrote:

Do you know practical cases of using "include-file" in programs?

I remember a script for reading csv data files with INCLUDE-FILE.
The interpreter (parser) was slightly modified for this.
--- Synchronet 3.20a-Linux NewsLink 1.114

From Newsgroup: comp.lang.forth

On 2024-10-24 10:32, Anton Ertl wrote:

Ruvim <[email protected]> writes:

Also, `REQUIRED` does not load a file that was loaded by `INCLUDED`.
`INCLUDE-FILE` is free of that too.

Yes, but why would anyone use REQUIRED or REQUIRE on a file that he
intends to include even if it has been INCLUDED before?

This can be in different modules, and one module does not know about
details of another.

If you include a file into a separate word list within a special
context, you might not want to prevent that file from being loaded by `require` into another word list (and in another context).


--
Ruvim

--- Synchronet 3.20a-Linux NewsLink 1.114

From Newsgroup: comp.lang.forth

On 2024-10-24 16:44, minforth wrote:

On Mon, 2 Sep 2024 20:11:25 +0000, Ruvim wrote:

Do you know practical cases of using "include-file" in programs?

I remember a script for reading csv data files with INCLUDE-FILE.
The interpreter (parser) was slightly modified for this.

Why wasn't `INCLUDED` suitable?


--
Ruvim

--- Synchronet 3.20a-Linux NewsLink 1.114

From Newsgroup: comp.lang.forth

It was opened in read/write mode to read or append new sections.
--- Synchronet 3.20a-Linux NewsLink 1.114

From Newsgroup: comp.lang.forth

Hi Marcel, the Mac Cocoa interface for i4 uses INCLUDE-FILE to deal with
code pasted in the OS clipboard. BTW same for VFX and VFX64. I needed a portable/standard word, 'et voila'.
--- Synchronet 3.20a-Linux NewsLink 1.114

From Newsgroup: comp.lang.forth

On Sat, 9 Nov 2024 8:29:44 +0000, rt wrote:

Hi Marcel, the Mac Cocoa interface for i4 uses INCLUDE-FILE
to deal with code pasted in the OS clipboard. BTW same for
VFX and VFX64. I needed a portable/standard word, 'et voila'.

Hi Roelf: not enough information. Isn't the file-id for the clipboard
acquired in a non-standard way then?

iForth's console already accepts multi-line text pasted in its
window. In SwiftForth the pasted text seems to have a fixed (small)
length. I don't receive VFX updates anymore (I probably said something offensive), so can't try.

There could be a few issues with a portable solution :--)

-marcel
--- Synchronet 3.20a-Linux NewsLink 1.114

From Newsgroup: comp.lang.forth

On 9 Nov 2024 at 13:06:19 CET, "mhx" <mhx> wrote:

I don't receive VFX updates anymore (I probably said something
offensive), so can't try.

Now that all products are freely downloadable for non-commercial use,
all you need to do is to download the latest versions from:
https://vfxforth.com/downloads/VfxCommunity/

Stephen
--
Stephen Pelc, [email protected]
MicroProcessor Engineering, Ltd. - More Real, Less Time
133 Hill Lane, Southampton SO15 5AF, England
tel: +44 (0)78 0390 3612, +34 649 662 974
http://www.mpeforth.com MPE website http://www.vfxforth.com/downloads/VfxCommunity/ downloads
--- Synchronet 3.20a-Linux NewsLink 1.114

From Newsgroup: comp.lang.forth

On 10/11/2024 5:13 am, Stephen Pelc wrote:

On 9 Nov 2024 at 13:06:19 CET, "mhx" <mhx> wrote:

I don't receive VFX updates anymore (I probably said something
offensive), so can't try.

Now that all products are freely downloadable for non-commercial use,
all you need to do is to download the latest versions from:
https://vfxforth.com/downloads/VfxCommunity/

Stephen

Any resolution to the problem of switching fp packs under Win x64
corrupting the GUI? A rebuild of the system using the desired fp
pack would likely work but that's a task in itself.


--- Synchronet 3.20a-Linux NewsLink 1.114

From Newsgroup: comp.lang.forth

<quote>
There could be a few issues with a portable solution :--)
</quote>

:-)

The Cocoa interface runs in CarbonMacForth(2008), iForth(2009), SwiftForth(2010) and VFX(2012).
Kind of recall, the Cocoa, not UNIX/Darwin, clipboard interface was
implemented in 2010, 14 years ago! At the time not seen any remarks
about the standard INCLUDE-FILE.
And it still works...

--
--- Synchronet 3.20a-Linux NewsLink 1.114