I've been getting occasional messages about GForth failing the Forth
2012 test suite. Invariably this is because they are using GForth 0.7.3
which, I assume, is the latest 'official' release of GForth. The
cpoyright message I've seen is:

Gforth 0.7.3, Copyright (C) 1995-2008 Free Software Foundation, Inc.

It seems that people don't know about the 0.7.9 snapshots available that
do include Forth 2012 words and will pass the test programs.

Isn't it about time that another official version of GForth is released?
After all if people don't know GForth an ongoing development when they
see the date 2008 they might think the project is moribund.

--
Gerry

Gerry Jackson schrieb am Mittwoch, 9. März 2022 um 10:26:48 UTC+1:
> I've been getting occasional messages about GForth failing the Forth
> 2012 test suite. Invariably this is because they are using GForth 0.7.3
> which, I assume, is the latest 'official' release of GForth. The
> cpoyright message I've seen is:
>
> Gforth 0.7.3, Copyright (C) 1995-2008 Free Software Foundation, Inc.
>
> It seems that people don't know about the 0.7.9 snapshots available that
> do include Forth 2012 words and will pass the test programs.
>
> Isn't it about time that another official version of GForth is released?
> After all if people don't know GForth an ongoing development when they
> see the date 2008 they might think the project is moribund.
>

Version 0.x.y indicates that it is still in beta state. ;-)
IOW expect occasional troubles.

On 09/03/2022 12:29, minf...@arcor.de wrote:
> Gerry Jackson schrieb am Mittwoch, 9. März 2022 um 10:26:48 UTC+1:
>> I've been getting occasional messages about GForth failing the Forth
>> 2012 test suite. Invariably this is because they are using GForth 0.7.3
>> which, I assume, is the latest 'official' release of GForth. The
>> cpoyright message I've seen is:
>>
>> Gforth 0.7.3, Copyright (C) 1995-2008 Free Software Foundation, Inc.
>>
>> It seems that people don't know about the 0.7.9 snapshots available that
>> do include Forth 2012 words and will pass the test programs.
>>
>> Isn't it about time that another official version of GForth is released?
>> After all if people don't know GForth an ongoing development when they
>> see the date 2008 they might think the project is moribund.
>>
>
> Version 0.x.y indicates that it is still in beta state. ;-)
> IOW expect occasional troubles.

It is annoying that the test program gets blamed when a Forth 2012 word
like BUFFER: isn't available. Version 0.7.3 appears to be 13 or 14 years
old. I know that the wheels of Forth grind slowly but it's about time
that version 1.0.0 is released. How hard is it to release a stable
version every couple of years or so?

Anyway I've added something to the GitHub readme.md placing the blame on
GForth 0.7.3. Even if people don't read it I can easily dismiss an error
report by reference to the readme.

--
Gerry

Gerry Jackson <do-not-use@swldwa.uk> writes:
>Isn't it about time that another official version of GForth is released?

Certainly. But it's also necessary that the documentation reflects
what we release, and currently it is quite far behind. You can blame
it on me; too many other things to do.

- 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: http://www.forth200x.org/forth200x.html
EuroForth 2021: https://euro.theforth.net/2021

On Wednesday, March 9, 2022 at 7:29:35 AM UTC-5, minf...@arcor.de wrote:
> Gerry Jackson schrieb am Mittwoch, 9. März 2022 um 10:26:48 UTC+1:
> > I've been getting occasional messages about GForth failing the Forth
> > 2012 test suite. Invariably this is because they are using GForth 0.7.3
> > which, I assume, is the latest 'official' release of GForth. The
> > cpoyright message I've seen is:
> >
> > Gforth 0.7.3, Copyright (C) 1995-2008 Free Software Foundation, Inc.
> >
> > It seems that people don't know about the 0.7.9 snapshots available that
> > do include Forth 2012 words and will pass the test programs.
> >
> > Isn't it about time that another official version of GForth is released?
> > After all if people don't know GForth an ongoing development when they
> > see the date 2008 they might think the project is moribund.
> >
> Version 0.x.y indicates that it is still in beta state. ;-)
> IOW expect occasional troubles.

Version numbers mean exactly what the author wants them to mean, nothing less and nothing more. I think it was FreePCB that was incrementing 1.x over and over and someone asked why the 1 never changed. The reply was about some Linux aspect that was similar and one of the founders attitudes that it should never change. To me, that makes no sense. If it's not going to change, why include it? Same here. If it is always going to be zero, ignore that and it's now 7.3 or 7.9. Does that sound better to you?

--

Rick C.

- Get 1,000 miles of free Supercharging
- Tesla referral code - https://ts.la/richard11209

On Wednesday, March 9, 2022 at 11:48:36 AM UTC-5, Anton Ertl wrote:
> Gerry Jackson <do-no...@swldwa.uk> writes:
> >Isn't it about time that another official version of GForth is released?
> Certainly. But it's also necessary that the documentation reflects
> what we release, and currently it is quite far behind. You can blame
> it on me; too many other things to do.

I worked on a large government contracting job once where the code used the left N columns and the right M column were for comments which were fairly well specified so they would be useful. I worked on the documentation crew where we ripped out the right most M columns and partitioned it into units and sections and chapter. We also did pagination... BY HAND! Cut and paste with real scissors. The guy in charge was not computer savvy, so I came up with a scheme to automatically insert the page headers, complete with page numbers, into the text files... using DEC DCL. For a command line tool it was pretty powerful. I also created a DCL program to sound ships bells on my terminal. LOL I liked that job, well that portion of it. Beltway bandits are not so fun really. It's all about getting the next job number to charge your time to. That was actually the longest job I've ever had, until I worked for myself.

--

Rick C.

+ Get 1,000 miles of free Supercharging
+ Tesla referral code - https://ts.la/richard11209

On 11/03/2022 13:35, Rick C wrote:
> On Wednesday, March 9, 2022 at 7:29:35 AM UTC-5, minf...@arcor.de wrote:
>> Gerry Jackson schrieb am Mittwoch, 9. März 2022 um 10:26:48 UTC+1:
>> > I've been getting occasional messages about GForth failing the Forth
>> > 2012 test suite. Invariably this is because they are using GForth 0.7.3
>> > which, I assume, is the latest 'official' release of GForth. The
>> > cpoyright message I've seen is:
>> >
>> > Gforth 0.7.3, Copyright (C) 1995-2008 Free Software Foundation, Inc.
>> >
>> > It seems that people don't know about the 0.7.9 snapshots available that
>> > do include Forth 2012 words and will pass the test programs.
>> >
>> > Isn't it about time that another official version of GForth is released?
>> > After all if people don't know GForth an ongoing development when they
>> > see the date 2008 they might think the project is moribund.
>> >
>> Version 0.x.y indicates that it is still in beta state. ;-)
>> IOW expect occasional troubles.
>
> Version numbers mean exactly what the author wants them to mean, nothing less and nothing more. I think it was FreePCB that was incrementing 1.x over and over and someone asked why the 1 never changed. The reply was about some Linux aspect that was similar and one of the founders attitudes that it should never change. To me, that makes no sense. If it's not going to change, why include it? Same here. If it is always going to be zero, ignore that and it's now 7.3 or 7.9. Does that sound better to you?
>

"Can you ever trust a V1.0 ?". In never releasing a formal V1.0, authors
believe they can avoid criticism by using the defence of 'experimental'.
Needless to say it doesn't fool anyone. All it demonstrates is the author's
uncertainty and inability to commit.

https://drive.google.com/drive/folders/1kh2WcPUc3hQpLcz7TQ-YQiowrozvxfGw

On Thursday, March 10, 2022 at 11:12:28 PM UTC-5, dxforth wrote:
> On 11/03/2022 13:35, Rick C wrote:
> > On Wednesday, March 9, 2022 at 7:29:35 AM UTC-5, minf...@arcor.de wrote:
> >> Gerry Jackson schrieb am Mittwoch, 9. März 2022 um 10:26:48 UTC+1:
> >> > I've been getting occasional messages about GForth failing the Forth
> >> > 2012 test suite. Invariably this is because they are using GForth 0.7.3
> >> > which, I assume, is the latest 'official' release of GForth. The
> >> > cpoyright message I've seen is:
> >> >
> >> > Gforth 0.7.3, Copyright (C) 1995-2008 Free Software Foundation, Inc.
> >> >
> >> > It seems that people don't know about the 0.7.9 snapshots available that
> >> > do include Forth 2012 words and will pass the test programs.
> >> >
> >> > Isn't it about time that another official version of GForth is released?
> >> > After all if people don't know GForth an ongoing development when they
> >> > see the date 2008 they might think the project is moribund.
> >> >
> >> Version 0.x.y indicates that it is still in beta state. ;-)
> >> IOW expect occasional troubles.
> >
> > Version numbers mean exactly what the author wants them to mean, nothing less and nothing more. I think it was FreePCB that was incrementing 1.x over and over and someone asked why the 1 never changed. The reply was about some Linux aspect that was similar and one of the founders attitudes that it should never change. To me, that makes no sense. If it's not going to change, why include it? Same here. If it is always going to be zero, ignore that and it's now 7.3 or 7.9. Does that sound better to you?
> >
> "Can you ever trust a V1.0 ?". In never releasing a formal V1.0, authors
> believe they can avoid criticism by using the defence of 'experimental'.
> Needless to say it doesn't fool anyone. All it demonstrates is the author's
> uncertainty and inability to commit.
>
> https://drive.google.com/drive/folders/1kh2WcPUc3hQpLcz7TQ-YQiowrozvxfGw

I think the bigger issue is not that Gforth doesn't have a "1.0" release, it's more that they haven't had a stable release since 2014. Sure, you can build the git repo, but as a newbie like me it's pretty discouraging to see that a project hasn't released in a while.
Rant Time:
Part of the advantage of developing in a Linux-based OS is you can use a package manager to install things easily, but with Gforth I'm stuck with an old version if I want to use my package manager. However, I recently tried to install Gforth by building it from source, but because the script detected I was using Debian it told my package manager to install Gforth via apt instead. This means that by trying to install the newer Gforth, it just installed the old one that I didn't want in the first place.
When I get the time, I'll report this as a bug.

Troy Jacobs <tmj9001@gmail.com> writes:
>Rant Time:
> Part of the advantage of developing in a Linux-based OS is you can use a =
>package manager to install things easily, but with Gforth I'm stuck with an=
> old version if I want to use my package manager.

Actually Gforth's INSTALL.md contains a section about installing
binaries, which tells you about installing for Debian, or as Flatpack,
Docker container, or Snap. Here's what it says about Debian:

|## Packets for Debian GNU/Linux
| |You can use the following Debian repository to make it easy to install
|Gforth. For the following commands, you need to be root (or prepend all the
|`apt` commands with `sudo`):
| |If you don't have https transport for apt installed, do that first:
| | apt install apt-transport-https
| |Create a debian sources.list file pointing to the net2o repository,
|and add the key to the trust db so that Debian can verify the packets,
|update the repository data and install net2o, so enter:
| | cat >/etc/apt/sources.list.d/net2o.list <<EOF
| deb [arch=i386,amd64,armhf,armel,arm64,powerpc,mips,mipsel,all] https://net2o.de/debian testing main
| EOF
| wget -O - https://net2o.de/bernd@net2o.de-yubikey.pgp.asc | apt-key add -
| |Remove the architectures on the list above which you don't need; on Debian
|stable, the list is not necessary anymore. On older versions (oldstable), the
|“`all`” part is not searched if you don't have that list, then Gforth fails to
|install the “`gforth-common`” part (and others that are not architecture
|dependent).
| |Now you are ready to install:
| | apt update
| apt install gforth
| |There are actually four repositories: oldstable, stable, testing and unstable,
|compiled for the respective Debian systems.

>However, I recently trie=
>d to install Gforth by building it from source, but because the script dete=
>cted I was using Debian it told my package manager to install Gforth via ap=
>t instead. This means that by trying to install the newer Gforth, it just =
>installed the old one that I didn't want in the first place.
> When I get the time, I'll report this as a bug.

I guess you have found the time. If you build from the git, you need
an existing Gforth to build the new Gforth. If you build from the
tarball, that's not necessary.

- 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: http://www.forth200x.org/forth200x.html
EuroForth 2021: https://euro.theforth.net/2021

On Wednesday, March 23, 2022 at 6:19:13 PM UTC-4, Anton Ertl wrote:
> Troy Jacobs <tmj...@gmail.com> writes:
> >Rant Time:
> > Part of the advantage of developing in a Linux-based OS is you can use a =
> >package manager to install things easily, but with Gforth I'm stuck with an> > old version if I want to use my package manager.
> Actually Gforth's INSTALL.md contains a section about installing
> binaries, which tells you about installing for Debian, or as Flatpack,
> Docker container, or Snap. Here's what it says about Debian:
>
> |## Packets for Debian GNU/Linux
> |
> |You can use the following Debian repository to make it easy to install
> |Gforth. For the following commands, you need to be root (or prepend all the
> |`apt` commands with `sudo`):
> |
> |If you don't have https transport for apt installed, do that first:
> |
> | apt install apt-transport-https
> |
> |Create a debian sources.list file pointing to the net2o repository,
> |and add the key to the trust db so that Debian can verify the packets,
> |update the repository data and install net2o, so enter:
> |
> | cat >/etc/apt/sources.list.d/net2o.list <<EOF
> | deb [arch=i386,amd64,armhf,armel,arm64,powerpc,mips,mipsel,all] https://net2o.de/debian testing main
> | EOF
> | wget -O - https://net2o.de/be...@net2o.de-yubikey.pgp.asc | apt-key add -
> |
> |Remove the architectures on the list above which you don't need; on Debian
> |stable, the list is not necessary anymore. On older versions (oldstable), the
> |“`all`” part is not searched if you don't have that list, then Gforth fails to
> |install the “`gforth-common`” part (and others that are not architecture
> |dependent).
> |
> |Now you are ready to install:
> |
> | apt update
> | apt install gforth
> |
> |There are actually four repositories: oldstable, stable, testing and unstable,
> |compiled for the respective Debian systems.
>
> >However, I recently trie=
> >d to install Gforth by building it from source, but because the script dete=
> >cted I was using Debian it told my package manager to install Gforth via ap=
> >t instead. This means that by trying to install the newer Gforth, it just > >installed the old one that I didn't want in the first place.
> > When I get the time, I'll report this as a bug.
> I guess you have found the time. If you build from the git, you need
> an existing Gforth to build the new Gforth. If you build from the
> tarball, that's not necessary.
> - 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: http://www.forth200x.org/forth200x.html
> EuroForth 2021: https://euro.theforth.net/2021

Anton,
Thanks for the response, I hadn't noticed this post until now, but it's really cool how responsive and willing to help you Gforth folks are. I can't say enough how awesome that is, I'll be sticking with Gforth for a whlie! :D

On Wednesday, March 23, 2022 at 6:19:13 PM UTC-4, Anton Ertl wrote:
> Troy Jacobs <tmj...@gmail.com> writes:
> >Rant Time:
> > Part of the advantage of developing in a Linux-based OS is you can use a =
> >package manager to install things easily, but with Gforth I'm stuck with an> > old version if I want to use my package manager.
> Actually Gforth's INSTALL.md contains a section about installing
> binaries, which tells you about installing for Debian, or as Flatpack,
> Docker container, or Snap. Here's what it says about Debian:
>
> |## Packets for Debian GNU/Linux
> |
> |You can use the following Debian repository to make it easy to install
> |Gforth. For the following commands, you need to be root (or prepend all the
> |`apt` commands with `sudo`):
> |
> |If you don't have https transport for apt installed, do that first:
> |
> | apt install apt-transport-https
> |
> |Create a debian sources.list file pointing to the net2o repository,
> |and add the key to the trust db so that Debian can verify the packets,
> |update the repository data and install net2o, so enter:
> |
> | cat >/etc/apt/sources.list.d/net2o.list <<EOF
> | deb [arch=i386,amd64,armhf,armel,arm64,powerpc,mips,mipsel,all] https://net2o.de/debian testing main
> | EOF
> | wget -O - https://net2o.de/be...@net2o.de-yubikey.pgp.asc | apt-key add -
> |
> |Remove the architectures on the list above which you don't need; on Debian
> |stable, the list is not necessary anymore. On older versions (oldstable), the
> |“`all`” part is not searched if you don't have that list, then Gforth fails to
> |install the “`gforth-common`” part (and others that are not architecture
> |dependent).
> |
> |Now you are ready to install:
> |
> | apt update
> | apt install gforth
> |
> |There are actually four repositories: oldstable, stable, testing and unstable,
> |compiled for the respective Debian systems.
>
> >However, I recently trie=
> >d to install Gforth by building it from source, but because the script dete=
> >cted I was using Debian it told my package manager to install Gforth via ap=
> >t instead. This means that by trying to install the newer Gforth, it just > >installed the old one that I didn't want in the first place.
> > When I get the time, I'll report this as a bug.
> I guess you have found the time. If you build from the git, you need
> an existing Gforth to build the new Gforth. If you build from the
> tarball, that's not necessary.
> - 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: http://www.forth200x.org/forth200x.html
> EuroForth 2021: https://euro.theforth.net/2021

Anton,
Thanks for the response, I hadn't noticed this post until now, but it's really cool how responsive and willing to help you Gforth folks are. I can't say enough how awesome that is, I'll be sticking with Gforth for a whlie! :D

On Friday, March 11, 2022 at 5:12:28 AM UTC+1, dxforth wrote:
> "Can you ever trust a V1.0 ?"
Since I've had plenty of v1.0 releases (and NEVER v0.x) I'd say - it depends. Sometimes yes and sometimes I disappointed myself.

> In never releasing a formal V1.0, authors
> believe they can avoid criticism by using the defense of 'experimental'.
If you'd say "They are communicating 'experimental' I'd say yes - since that is the convention.
If they're doing that to "avoid criticism" - I don't know. I'm not in their heads.
To me it doesn't necessarily say it's "not production quality", but AT LEAST it's saying me it's not feature complete.
Expect stubs and stuff.

> Needless to say it doesn't fool anyone.
That depends in the authors intentions. If it consciously used to misrepresent its actual state - yes, it is.
If it is his honest belief there is something lacking in features or quality - no, it isn't.
I wouldn't call laziness "fooling", for example.

> All it demonstrates is the author's uncertainty and inability to commit.
To commit to what?
"Yeah, it's feature complete";
"Yeah, it's production ready";
"Yeah, it's an official release".

When I dump something on the web I always assume "it's ready". That's honestly where I'm going for.
So - most of the time it's safe to grab a SVN dump of 4tH and be at the cutting edge with no risk at all.
Which is logical, because I'm at the cutting edge of 4tH ALL THE TIME.

Does that mean an absolute guarantee it's "error free"? No, it doesn't. There are a few known bugs with minor impact.
Those will not be fixed quickly - mostly because there is a solid workaround. And then there are limitations.
All of them documented.

And then there are the errors that WILL pop up - because it's still early development (we're still talking SVN).

The releases mean a bit more. That means we've had serious, systematic testing and prolonged use. All the stuff
has been properly documented - and documentation and product are in sync. And since I have been able to come
up with a release just about every year, it means it will NEVER be "feature complete". That would be the end of a
nice hobby ;-)

If you've used a product for a long time, you will have formed an opinion about it. And IMHO Gforth is pretty
reliable - at least for my uses (which is mostly as a test bed for ANS-Forth 94 compatible programs that I've
ported from 4tH, a calculator - and *occasionally* a reference tool).

... and then IMHO your knowledge of the tool is much more detailed than a version number.

But may it put off people when SELECTING a tool? Absolutely!

Hans Bezemer

On 1/04/2022 01:27, Hans Bezemer wrote:
>
> The releases mean a bit more. That means we've had serious, systematic testing and prolonged use. All the stuff
> has been properly documented - and documentation and product are in sync. And since I have been able to come
> up with a release just about every year, it means it will NEVER be "feature complete". That would be the end of a
> nice hobby ;-)
>
> If you've used a product for a long time, you will have formed an opinion about it. And IMHO Gforth is pretty
> reliable - at least for my uses (which is mostly as a test bed for ANS-Forth 94 compatible programs that I've
> ported from 4tH, a calculator - and *occasionally* a reference tool).
>
> .. and then IMHO your knowledge of the tool is much more detailed than a version number.
>
> But may it put off people when SELECTING a tool? Absolutely!

Put that way, I'd agree. Looking back at DX-Forth 1.0 I definitely wouldn't say it was
complete or even well thought out - though it did feel like it at the time. At version
4.x and experience of a number of apps written with it, I'm far more confident it's
'done' - as it ought to be after so many years. Today it's more a case of moving the
deck chairs around to obtain the best light.