Re: [Jgeneral] FF character in J source files on windows jconsole.

2021-07-18 Thread Eric Iverson
I just tried editing a file that has an FF with a few common editors. It
does not work properly in several of them. In particular in linux textedit
and in JHS codemirror. I would not be able to work with scripts that had FF
in them!

On Sun, Jul 18, 2021 at 6:38 PM Henry Rich  wrote:

> In the middle of a line FF would need to be whitespace which would mean
> you have to decide how to display it, etc.
>
> Treated as a line-end it is discarded as the file is read, and there are
> no further repercussions as the FF is not part of what gets interpreted.
>
> There would be the side effect that 5!:5 would not recover the FF.  But
> if the FF is useful to editors and doesn't hurt us, I can see allowing it.
>
> hhr
>
> On 7/18/2021 6:22 PM, Eric Iverson wrote:
> > Is it used only when alone on a line? Things get messier when it is
> allowed
> > anywhere,
> >
> > On Sun, Jul 18, 2021 at 6:10 PM Michal Wallace  >
> > wrote:
> >
> >> FF is actually pretty commonly used in source code... I use it all the
> time
> >> in python and rust.
> >> It's actually used as part of the GNU C coding standard, according to
> >> wikipedia (pretty sure that's where I picked it up...)
> >>
> >> https://en.wikipedia.org/wiki/Page_break#Semantic_use
> >>
> >>
> >> On Sun, Jul 18, 2021 at 5:55 PM Eric Iverson 
> >> wrote:
> >>
> >>> In linux jconsole input ctrl+l, is not treated as LF. Instead it
> scrolls
> >>> down the page.
> >>>
> >>> I think you will find that ctrl+l will behave differently on many
> >> different
> >>> hosts, terminal managers, J front ends, etc. I think making it behave
> in
> >> a
> >>> useful way and the same way in all cases is not possible. And including
> >> it
> >>> is a script introduces a whole new level of complication.
> >>>
> >>> Before changing the JE, the use case should be studied and probably
> >>> revisited.
> >>>
> >>>
> >>>
> >>> And it is unlikely to be consistent with other languages.
> >>>
> >>>
> >>> On Sun, Jul 18, 2021 at 5:37 PM Raul Miller 
> >> wrote:
>  On Sun, Jul 18, 2021 at 4:55 PM chris burke 
> >>> wrote:
> > I think FF is normally considered to be whitespace, so it can be
> >>> treated
> > like LF.
>  Yes.
> 
>  https://en.wikipedia.org/wiki/Whitespace_character
> 
>  Thanks,
> 
>  --
>  Raul
>  --
>  For information about J forums see
> http://www.jsoftware.com/forums.htm
> 
> >>> --
> >>> For information about J forums see http://www.jsoftware.com/forums.htm
> >>>
> >> --
> >> For information about J forums see http://www.jsoftware.com/forums.htm
> >>
> > --
> > For information about J forums see http://www.jsoftware.com/forums.htm
>
>
> --
> This email has been checked for viruses by AVG.
> https://www.avg.com
>
> --
> For information about J forums see http://www.jsoftware.com/forums.htm
>
--
For information about J forums see http://www.jsoftware.com/forums.htm


Re: [Jgeneral] FF character in J source files on windows jconsole.

2021-07-18 Thread Henry Rich
In the middle of a line FF would need to be whitespace which would mean 
you have to decide how to display it, etc.


Treated as a line-end it is discarded as the file is read, and there are 
no further repercussions as the FF is not part of what gets interpreted.


There would be the side effect that 5!:5 would not recover the FF.  But 
if the FF is useful to editors and doesn't hurt us, I can see allowing it.


hhr

On 7/18/2021 6:22 PM, Eric Iverson wrote:

Is it used only when alone on a line? Things get messier when it is allowed
anywhere,

On Sun, Jul 18, 2021 at 6:10 PM Michal Wallace 
wrote:


FF is actually pretty commonly used in source code... I use it all the time
in python and rust.
It's actually used as part of the GNU C coding standard, according to
wikipedia (pretty sure that's where I picked it up...)

https://en.wikipedia.org/wiki/Page_break#Semantic_use


On Sun, Jul 18, 2021 at 5:55 PM Eric Iverson 
wrote:


In linux jconsole input ctrl+l, is not treated as LF. Instead it scrolls
down the page.

I think you will find that ctrl+l will behave differently on many

different

hosts, terminal managers, J front ends, etc. I think making it behave in

a

useful way and the same way in all cases is not possible. And including

it

is a script introduces a whole new level of complication.

Before changing the JE, the use case should be studied and probably
revisited.



And it is unlikely to be consistent with other languages.


On Sun, Jul 18, 2021 at 5:37 PM Raul Miller 

wrote:

On Sun, Jul 18, 2021 at 4:55 PM chris burke 

wrote:

I think FF is normally considered to be whitespace, so it can be

treated

like LF.

Yes.

https://en.wikipedia.org/wiki/Whitespace_character

Thanks,

--
Raul
--
For information about J forums see http://www.jsoftware.com/forums.htm


--
For information about J forums see http://www.jsoftware.com/forums.htm


--
For information about J forums see http://www.jsoftware.com/forums.htm


--
For information about J forums see http://www.jsoftware.com/forums.htm



--
This email has been checked for viruses by AVG.
https://www.avg.com

--
For information about J forums see http://www.jsoftware.com/forums.htm


Re: [Jgeneral] FF character in J source files on windows jconsole.

2021-07-18 Thread Michal Wallace
Yeah. Or at least that's the only way I've ever seen it used.
It was originally literally to feed a new form into your printer when you
were printing your source code,
so it really doesn't make sense anywhere else.


On Sun, Jul 18, 2021 at 6:22 PM Eric Iverson 
wrote:

> Is it used only when alone on a line? Things get messier when it is allowed
> anywhere,
>
> On Sun, Jul 18, 2021 at 6:10 PM Michal Wallace 
> wrote:
>
> > FF is actually pretty commonly used in source code... I use it all the
> time
> > in python and rust.
> > It's actually used as part of the GNU C coding standard, according to
> > wikipedia (pretty sure that's where I picked it up...)
> >
> > https://en.wikipedia.org/wiki/Page_break#Semantic_use
> >
> >
> > On Sun, Jul 18, 2021 at 5:55 PM Eric Iverson 
> > wrote:
> >
> > > In linux jconsole input ctrl+l, is not treated as LF. Instead it
> scrolls
> > > down the page.
> > >
> > > I think you will find that ctrl+l will behave differently on many
> > different
> > > hosts, terminal managers, J front ends, etc. I think making it behave
> in
> > a
> > > useful way and the same way in all cases is not possible. And including
> > it
> > > is a script introduces a whole new level of complication.
> > >
> > > Before changing the JE, the use case should be studied and probably
> > > revisited.
> > >
> > >
> > >
> > > And it is unlikely to be consistent with other languages.
> > >
> > >
> > > On Sun, Jul 18, 2021 at 5:37 PM Raul Miller 
> > wrote:
> > >
> > > > On Sun, Jul 18, 2021 at 4:55 PM chris burke 
> > > wrote:
> > > > > I think FF is normally considered to be whitespace, so it can be
> > > treated
> > > > > like LF.
> > > >
> > > > Yes.
> > > >
> > > > https://en.wikipedia.org/wiki/Whitespace_character
> > > >
> > > > Thanks,
> > > >
> > > > --
> > > > Raul
> > > >
> --
> > > > For information about J forums see
> http://www.jsoftware.com/forums.htm
> > > >
> > > --
> > > For information about J forums see http://www.jsoftware.com/forums.htm
> > >
> > --
> > For information about J forums see http://www.jsoftware.com/forums.htm
> >
> --
> For information about J forums see http://www.jsoftware.com/forums.htm
>
--
For information about J forums see http://www.jsoftware.com/forums.htm


Re: [Jgeneral] FF character in J source files on windows jconsole.

2021-07-18 Thread Eric Iverson
Is it used only when alone on a line? Things get messier when it is allowed
anywhere,

On Sun, Jul 18, 2021 at 6:10 PM Michal Wallace 
wrote:

> FF is actually pretty commonly used in source code... I use it all the time
> in python and rust.
> It's actually used as part of the GNU C coding standard, according to
> wikipedia (pretty sure that's where I picked it up...)
>
> https://en.wikipedia.org/wiki/Page_break#Semantic_use
>
>
> On Sun, Jul 18, 2021 at 5:55 PM Eric Iverson 
> wrote:
>
> > In linux jconsole input ctrl+l, is not treated as LF. Instead it scrolls
> > down the page.
> >
> > I think you will find that ctrl+l will behave differently on many
> different
> > hosts, terminal managers, J front ends, etc. I think making it behave in
> a
> > useful way and the same way in all cases is not possible. And including
> it
> > is a script introduces a whole new level of complication.
> >
> > Before changing the JE, the use case should be studied and probably
> > revisited.
> >
> >
> >
> > And it is unlikely to be consistent with other languages.
> >
> >
> > On Sun, Jul 18, 2021 at 5:37 PM Raul Miller 
> wrote:
> >
> > > On Sun, Jul 18, 2021 at 4:55 PM chris burke 
> > wrote:
> > > > I think FF is normally considered to be whitespace, so it can be
> > treated
> > > > like LF.
> > >
> > > Yes.
> > >
> > > https://en.wikipedia.org/wiki/Whitespace_character
> > >
> > > Thanks,
> > >
> > > --
> > > Raul
> > > --
> > > For information about J forums see http://www.jsoftware.com/forums.htm
> > >
> > --
> > For information about J forums see http://www.jsoftware.com/forums.htm
> >
> --
> For information about J forums see http://www.jsoftware.com/forums.htm
>
--
For information about J forums see http://www.jsoftware.com/forums.htm


Re: [Jgeneral] FF character in J source files on windows jconsole.

2021-07-18 Thread Michal Wallace
FF is actually pretty commonly used in source code... I use it all the time
in python and rust.
It's actually used as part of the GNU C coding standard, according to
wikipedia (pretty sure that's where I picked it up...)

https://en.wikipedia.org/wiki/Page_break#Semantic_use


On Sun, Jul 18, 2021 at 5:55 PM Eric Iverson 
wrote:

> In linux jconsole input ctrl+l, is not treated as LF. Instead it scrolls
> down the page.
>
> I think you will find that ctrl+l will behave differently on many different
> hosts, terminal managers, J front ends, etc. I think making it behave in a
> useful way and the same way in all cases is not possible. And including it
> is a script introduces a whole new level of complication.
>
> Before changing the JE, the use case should be studied and probably
> revisited.
>
>
>
> And it is unlikely to be consistent with other languages.
>
>
> On Sun, Jul 18, 2021 at 5:37 PM Raul Miller  wrote:
>
> > On Sun, Jul 18, 2021 at 4:55 PM chris burke 
> wrote:
> > > I think FF is normally considered to be whitespace, so it can be
> treated
> > > like LF.
> >
> > Yes.
> >
> > https://en.wikipedia.org/wiki/Whitespace_character
> >
> > Thanks,
> >
> > --
> > Raul
> > --
> > For information about J forums see http://www.jsoftware.com/forums.htm
> >
> --
> For information about J forums see http://www.jsoftware.com/forums.htm
>
--
For information about J forums see http://www.jsoftware.com/forums.htm


Re: [Jgeneral] FF character in J source files on windows jconsole.

2021-07-18 Thread Eric Iverson
In linux jconsole input ctrl+l, is not treated as LF. Instead it scrolls
down the page.

I think you will find that ctrl+l will behave differently on many different
hosts, terminal managers, J front ends, etc. I think making it behave in a
useful way and the same way in all cases is not possible. And including it
is a script introduces a whole new level of complication.

Before changing the JE, the use case should be studied and probably
revisited.



And it is unlikely to be consistent with other languages.


On Sun, Jul 18, 2021 at 5:37 PM Raul Miller  wrote:

> On Sun, Jul 18, 2021 at 4:55 PM chris burke  wrote:
> > I think FF is normally considered to be whitespace, so it can be treated
> > like LF.
>
> Yes.
>
> https://en.wikipedia.org/wiki/Whitespace_character
>
> Thanks,
>
> --
> Raul
> --
> For information about J forums see http://www.jsoftware.com/forums.htm
>
--
For information about J forums see http://www.jsoftware.com/forums.htm


Re: [Jgeneral] FF character in J source files on windows jconsole.

2021-07-18 Thread Raul Miller
On Sun, Jul 18, 2021 at 4:55 PM chris burke  wrote:
> I think FF is normally considered to be whitespace, so it can be treated
> like LF.

Yes.

https://en.wikipedia.org/wiki/Whitespace_character

Thanks,

-- 
Raul
--
For information about J forums see http://www.jsoftware.com/forums.htm


Re: [Jgeneral] FF character in J source files on windows jconsole.

2021-07-18 Thread Henry Rich
Bingo!  When reading from a script, the code discards UP TO TWO trailing 
control characters at the end of a line.  Thus


   0!:1 (49 12 13 { a.)
   1
1
   0!:1 (49 12 13 10 { a.)  NB. normal Windows
   1♀
|spelling error
|   1♀
|    ^
|[-0]
   0!:1 (49 12 10 { a.)  NB. Normal Linux
   1
1
   0!:1 (49 5 10 { a.)   NB. Any control character is deleted
   1
1

That explains the mystery.  We still need to agree on whether FF should 
be a line-end.


Henry Rich




On 7/18/2021 5:21 PM, Michal Wallace wrote:

Hrm. I always place it on a blank line by itself, and it does not give a
spelling error on linux in this case.

Could it be that the code for platform-specific line endings ignores any
character c after an LF if  32>a.i.c ?

On Sun, Jul 18, 2021 at 4:55 PM chris burke  wrote:


It looks like FF is not supported in linux scripts, e.g.

9!:14''
j903/j64avx2/linux/beta-l/commercial/
www.jsoftware.com/2021-06-17T13:00:33/clang-10-0-0/SLEEF=1

('1',FF,'2') fwrite 't1.ijs'
3
0!:1 <'t1.ijs'
1
 2
|spelling error
|   1
  2
|^
|[-0] t1.ijs

I think FF is normally considered to be whitespace, so it can be treated
like LF.

On Sun, Jul 18, 2021 at 1:15 PM Henry Rich  wrote:


Every bytecode that is not an ASCII graphic or SP, CR, TAB, or LF will
cause a spelling error if encountered as a word:

 ". 12 { a.
|spelling error
|
The question here is whether FF is enough of an industry-standard to be
called whitespace, like TAB.

I don't know myself.  I have looked at the code & decided that making FF
look like LF for script purposes is not too hard, while making it look
like TAB takes more work.

But I am loath to make any change at all until I understand why FF works
sometimes and not others.  As I see it, any script containing FF would
fail when loaded, unless the FF were buried inside a string.  If you
have cases where this is not so I want to know where the FF is
disappearing to.

In the meantime you can use Eric's trick of stripping FF during load.

Henry Rich



On 7/18/2021 3:56 PM, Eric Iverson wrote:

FF char (12{a.) gets a spelling error if in a script that is loaded in
linux. I think this would be the same as in windows but have not

tested.

There are lots of values that will give spelling errors. I am not sure

it

makes sense to special case (i.e. ignore FF).

If the FF is desired for its behavior in an editor, that is OK. But it

and

perhaps other chars should be stripped out before being fed to load.

On Sun, Jul 18, 2021 at 3:39 PM Henry Rich 

wrote:

I will look into turning FF into LF when a script is read, & will do

it

if there are no problems.

Does Linux do something to the FF?  I don't see why you don't get the
same error there as on Windows.

Henry Rich

On 7/18/2021 7:58 AM, Michal Wallace wrote:

I sometimes use the ^L character (FF, or 12{a.) in source code files

to

divide my code into "pages".

(I have emacs set up so that I can page back and forth showing one

page

at

a time without scrolling, so the top line of the page is always at

the

top

of the screen when that page is showing, etc.)

I've done this with J source code with no problems in the past. It

works

fine in jqt and in jconsole on linux, but I just discovered that if I

try

to load a J script written this way in the windows version of

jconsole,

it

gives me a spelling error on that character.

Is this something that could be officially supported, or do I need to
change my ways? :D


--

For information about J forums see

http://www.jsoftware.com/forums.htm

--
This email has been checked for viruses by AVG.
https://www.avg.com

--
For information about J forums see

http://www.jsoftware.com/forums.htm

--
For information about J forums see http://www.jsoftware.com/forums.htm


--
This email has been checked for viruses by AVG.
https://www.avg.com

--
For information about J forums see http://www.jsoftware.com/forums.htm


--
For information about J forums see http://www.jsoftware.com/forums.htm


--
For information about J forums see http://www.jsoftware.com/forums.htm



--
This email has been checked for viruses by AVG.
https://www.avg.com

--
For information about J forums see http://www.jsoftware.com/forums.htm


Re: [Jgeneral] FF character in J source files on windows jconsole.

2021-07-18 Thread Henry Rich

Sound point.  To summarize:

* If FF is part of a broader markup language, that markup should be 
handled in a preprocessor.


* OTOH, if you use a preprocessor, you can't complain if everyone else 
uses one too, and if they are not compatible you have Babel. That is, 
calling for a preprocessor only postpones problems.


* Does treating FF like LF really rise to the level of change that 
deserves a preprocessor?
** We already have 2 line-ends - LF and CR - and this is just adding a 
3d - not breaking new ground
** FF is currently an error so there is no chance that making it a 
line-end will affect any existing cases


* Is FF useful and intuitive?  Intuitive for sure; useful perhaps.

If that ends the debate, I would vote to make FF a line-end, and I would 
vote against making it whitespace.


But I still want to know how you were able to load a script with a FF 
character.


Henry Rich

On 7/18/2021 5:11 PM, Eric Iverson wrote:

What about chars for vertical tab, bell, etc? What is whitespace and what
is an error is a tricky question. I would be against any change in this
area. It has been this way for 30 years.

Fancier scripts (literate programming) are great, but FF is the tip of that
iceberg. I suggest that a script preprocessor is the right solution.

On Sun, Jul 18, 2021 at 4:55 PM chris burke  wrote:


It looks like FF is not supported in linux scripts, e.g.

9!:14''
j903/j64avx2/linux/beta-l/commercial/
www.jsoftware.com/2021-06-17T13:00:33/clang-10-0-0/SLEEF=1

('1',FF,'2') fwrite 't1.ijs'
3
0!:1 <'t1.ijs'
1
 2
|spelling error
|   1
  2
|^
|[-0] t1.ijs

I think FF is normally considered to be whitespace, so it can be treated
like LF.

On Sun, Jul 18, 2021 at 1:15 PM Henry Rich  wrote:


Every bytecode that is not an ASCII graphic or SP, CR, TAB, or LF will
cause a spelling error if encountered as a word:

 ". 12 { a.
|spelling error
|
The question here is whether FF is enough of an industry-standard to be
called whitespace, like TAB.

I don't know myself.  I have looked at the code & decided that making FF
look like LF for script purposes is not too hard, while making it look
like TAB takes more work.

But I am loath to make any change at all until I understand why FF works
sometimes and not others.  As I see it, any script containing FF would
fail when loaded, unless the FF were buried inside a string.  If you
have cases where this is not so I want to know where the FF is
disappearing to.

In the meantime you can use Eric's trick of stripping FF during load.

Henry Rich



On 7/18/2021 3:56 PM, Eric Iverson wrote:

FF char (12{a.) gets a spelling error if in a script that is loaded in
linux. I think this would be the same as in windows but have not

tested.

There are lots of values that will give spelling errors. I am not sure

it

makes sense to special case (i.e. ignore FF).

If the FF is desired for its behavior in an editor, that is OK. But it

and

perhaps other chars should be stripped out before being fed to load.

On Sun, Jul 18, 2021 at 3:39 PM Henry Rich 

wrote:

I will look into turning FF into LF when a script is read, & will do

it

if there are no problems.

Does Linux do something to the FF?  I don't see why you don't get the
same error there as on Windows.

Henry Rich

On 7/18/2021 7:58 AM, Michal Wallace wrote:

I sometimes use the ^L character (FF, or 12{a.) in source code files

to

divide my code into "pages".

(I have emacs set up so that I can page back and forth showing one

page

at

a time without scrolling, so the top line of the page is always at

the

top

of the screen when that page is showing, etc.)

I've done this with J source code with no problems in the past. It

works

fine in jqt and in jconsole on linux, but I just discovered that if I

try

to load a J script written this way in the windows version of

jconsole,

it

gives me a spelling error on that character.

Is this something that could be officially supported, or do I need to
change my ways? :D


--

For information about J forums see

http://www.jsoftware.com/forums.htm

--
This email has been checked for viruses by AVG.
https://www.avg.com

--
For information about J forums see

http://www.jsoftware.com/forums.htm

--
For information about J forums see http://www.jsoftware.com/forums.htm


--
This email has been checked for viruses by AVG.
https://www.avg.com

--
For information about J forums see http://www.jsoftware.com/forums.htm


--
For information about J forums see http://www.jsoftware.com/forums.htm


--
For information about J forums see 

Re: [Jgeneral] FF character in J source files on windows jconsole.

2021-07-18 Thread Michal Wallace
Hrm. I always place it on a blank line by itself, and it does not give a
spelling error on linux in this case.

Could it be that the code for platform-specific line endings ignores any
character c after an LF if  32>a.i.c ?

On Sun, Jul 18, 2021 at 4:55 PM chris burke  wrote:

> It looks like FF is not supported in linux scripts, e.g.
>
>9!:14''
> j903/j64avx2/linux/beta-l/commercial/
> www.jsoftware.com/2021-06-17T13:00:33/clang-10-0-0/SLEEF=1
>
>('1',FF,'2') fwrite 't1.ijs'
> 3
>0!:1 <'t1.ijs'
>1
> 2
> |spelling error
> |   1
>  2
> |^
> |[-0] t1.ijs
>
> I think FF is normally considered to be whitespace, so it can be treated
> like LF.
>
> On Sun, Jul 18, 2021 at 1:15 PM Henry Rich  wrote:
>
> > Every bytecode that is not an ASCII graphic or SP, CR, TAB, or LF will
> > cause a spelling error if encountered as a word:
> >
> > ". 12 { a.
> > |spelling error
> > |
> > The question here is whether FF is enough of an industry-standard to be
> > called whitespace, like TAB.
> >
> > I don't know myself.  I have looked at the code & decided that making FF
> > look like LF for script purposes is not too hard, while making it look
> > like TAB takes more work.
> >
> > But I am loath to make any change at all until I understand why FF works
> > sometimes and not others.  As I see it, any script containing FF would
> > fail when loaded, unless the FF were buried inside a string.  If you
> > have cases where this is not so I want to know where the FF is
> > disappearing to.
> >
> > In the meantime you can use Eric's trick of stripping FF during load.
> >
> > Henry Rich
> >
> >
> >
> > On 7/18/2021 3:56 PM, Eric Iverson wrote:
> > > FF char (12{a.) gets a spelling error if in a script that is loaded in
> > > linux. I think this would be the same as in windows but have not
> tested.
> > >
> > > There are lots of values that will give spelling errors. I am not sure
> it
> > > makes sense to special case (i.e. ignore FF).
> > >
> > > If the FF is desired for its behavior in an editor, that is OK. But it
> > and
> > > perhaps other chars should be stripped out before being fed to load.
> > >
> > > On Sun, Jul 18, 2021 at 3:39 PM Henry Rich 
> wrote:
> > >
> > >> I will look into turning FF into LF when a script is read, & will do
> it
> > >> if there are no problems.
> > >>
> > >> Does Linux do something to the FF?  I don't see why you don't get the
> > >> same error there as on Windows.
> > >>
> > >> Henry Rich
> > >>
> > >> On 7/18/2021 7:58 AM, Michal Wallace wrote:
> > >>> I sometimes use the ^L character (FF, or 12{a.) in source code files
> to
> > >>> divide my code into "pages".
> > >>>
> > >>> (I have emacs set up so that I can page back and forth showing one
> page
> > >> at
> > >>> a time without scrolling, so the top line of the page is always at
> the
> > >> top
> > >>> of the screen when that page is showing, etc.)
> > >>>
> > >>> I've done this with J source code with no problems in the past. It
> > works
> > >>> fine in jqt and in jconsole on linux, but I just discovered that if I
> > try
> > >>> to load a J script written this way in the windows version of
> jconsole,
> > >> it
> > >>> gives me a spelling error on that character.
> > >>>
> > >>> Is this something that could be officially supported, or do I need to
> > >>> change my ways? :D
> > >>>
> --
> > >>> For information about J forums see
> http://www.jsoftware.com/forums.htm
> > >>
> > >> --
> > >> This email has been checked for viruses by AVG.
> > >> https://www.avg.com
> > >>
> > >> --
> > >> For information about J forums see
> http://www.jsoftware.com/forums.htm
> > >>
> > > --
> > > For information about J forums see http://www.jsoftware.com/forums.htm
> >
> >
> > --
> > This email has been checked for viruses by AVG.
> > https://www.avg.com
> >
> > --
> > For information about J forums see http://www.jsoftware.com/forums.htm
> >
> --
> For information about J forums see http://www.jsoftware.com/forums.htm
>
--
For information about J forums see http://www.jsoftware.com/forums.htm


Re: [Jgeneral] FF character in J source files on windows jconsole.

2021-07-18 Thread Eric Iverson
What about chars for vertical tab, bell, etc? What is whitespace and what
is an error is a tricky question. I would be against any change in this
area. It has been this way for 30 years.

Fancier scripts (literate programming) are great, but FF is the tip of that
iceberg. I suggest that a script preprocessor is the right solution.

On Sun, Jul 18, 2021 at 4:55 PM chris burke  wrote:

> It looks like FF is not supported in linux scripts, e.g.
>
>9!:14''
> j903/j64avx2/linux/beta-l/commercial/
> www.jsoftware.com/2021-06-17T13:00:33/clang-10-0-0/SLEEF=1
>
>('1',FF,'2') fwrite 't1.ijs'
> 3
>0!:1 <'t1.ijs'
>1
> 2
> |spelling error
> |   1
>  2
> |^
> |[-0] t1.ijs
>
> I think FF is normally considered to be whitespace, so it can be treated
> like LF.
>
> On Sun, Jul 18, 2021 at 1:15 PM Henry Rich  wrote:
>
> > Every bytecode that is not an ASCII graphic or SP, CR, TAB, or LF will
> > cause a spelling error if encountered as a word:
> >
> > ". 12 { a.
> > |spelling error
> > |
> > The question here is whether FF is enough of an industry-standard to be
> > called whitespace, like TAB.
> >
> > I don't know myself.  I have looked at the code & decided that making FF
> > look like LF for script purposes is not too hard, while making it look
> > like TAB takes more work.
> >
> > But I am loath to make any change at all until I understand why FF works
> > sometimes and not others.  As I see it, any script containing FF would
> > fail when loaded, unless the FF were buried inside a string.  If you
> > have cases where this is not so I want to know where the FF is
> > disappearing to.
> >
> > In the meantime you can use Eric's trick of stripping FF during load.
> >
> > Henry Rich
> >
> >
> >
> > On 7/18/2021 3:56 PM, Eric Iverson wrote:
> > > FF char (12{a.) gets a spelling error if in a script that is loaded in
> > > linux. I think this would be the same as in windows but have not
> tested.
> > >
> > > There are lots of values that will give spelling errors. I am not sure
> it
> > > makes sense to special case (i.e. ignore FF).
> > >
> > > If the FF is desired for its behavior in an editor, that is OK. But it
> > and
> > > perhaps other chars should be stripped out before being fed to load.
> > >
> > > On Sun, Jul 18, 2021 at 3:39 PM Henry Rich 
> wrote:
> > >
> > >> I will look into turning FF into LF when a script is read, & will do
> it
> > >> if there are no problems.
> > >>
> > >> Does Linux do something to the FF?  I don't see why you don't get the
> > >> same error there as on Windows.
> > >>
> > >> Henry Rich
> > >>
> > >> On 7/18/2021 7:58 AM, Michal Wallace wrote:
> > >>> I sometimes use the ^L character (FF, or 12{a.) in source code files
> to
> > >>> divide my code into "pages".
> > >>>
> > >>> (I have emacs set up so that I can page back and forth showing one
> page
> > >> at
> > >>> a time without scrolling, so the top line of the page is always at
> the
> > >> top
> > >>> of the screen when that page is showing, etc.)
> > >>>
> > >>> I've done this with J source code with no problems in the past. It
> > works
> > >>> fine in jqt and in jconsole on linux, but I just discovered that if I
> > try
> > >>> to load a J script written this way in the windows version of
> jconsole,
> > >> it
> > >>> gives me a spelling error on that character.
> > >>>
> > >>> Is this something that could be officially supported, or do I need to
> > >>> change my ways? :D
> > >>>
> --
> > >>> For information about J forums see
> http://www.jsoftware.com/forums.htm
> > >>
> > >> --
> > >> This email has been checked for viruses by AVG.
> > >> https://www.avg.com
> > >>
> > >> --
> > >> For information about J forums see
> http://www.jsoftware.com/forums.htm
> > >>
> > > --
> > > For information about J forums see http://www.jsoftware.com/forums.htm
> >
> >
> > --
> > This email has been checked for viruses by AVG.
> > https://www.avg.com
> >
> > --
> > For information about J forums see http://www.jsoftware.com/forums.htm
> >
> --
> For information about J forums see http://www.jsoftware.com/forums.htm
>
--
For information about J forums see http://www.jsoftware.com/forums.htm


Re: [Jgeneral] FF character in J source files on windows jconsole.

2021-07-18 Thread chris burke
It looks like FF is not supported in linux scripts, e.g.

   9!:14''
j903/j64avx2/linux/beta-l/commercial/
www.jsoftware.com/2021-06-17T13:00:33/clang-10-0-0/SLEEF=1

   ('1',FF,'2') fwrite 't1.ijs'
3
   0!:1 <'t1.ijs'
   1
2
|spelling error
|   1
 2
|^
|[-0] t1.ijs

I think FF is normally considered to be whitespace, so it can be treated
like LF.

On Sun, Jul 18, 2021 at 1:15 PM Henry Rich  wrote:

> Every bytecode that is not an ASCII graphic or SP, CR, TAB, or LF will
> cause a spelling error if encountered as a word:
>
> ". 12 { a.
> |spelling error
> |
> The question here is whether FF is enough of an industry-standard to be
> called whitespace, like TAB.
>
> I don't know myself.  I have looked at the code & decided that making FF
> look like LF for script purposes is not too hard, while making it look
> like TAB takes more work.
>
> But I am loath to make any change at all until I understand why FF works
> sometimes and not others.  As I see it, any script containing FF would
> fail when loaded, unless the FF were buried inside a string.  If you
> have cases where this is not so I want to know where the FF is
> disappearing to.
>
> In the meantime you can use Eric's trick of stripping FF during load.
>
> Henry Rich
>
>
>
> On 7/18/2021 3:56 PM, Eric Iverson wrote:
> > FF char (12{a.) gets a spelling error if in a script that is loaded in
> > linux. I think this would be the same as in windows but have not tested.
> >
> > There are lots of values that will give spelling errors. I am not sure it
> > makes sense to special case (i.e. ignore FF).
> >
> > If the FF is desired for its behavior in an editor, that is OK. But it
> and
> > perhaps other chars should be stripped out before being fed to load.
> >
> > On Sun, Jul 18, 2021 at 3:39 PM Henry Rich  wrote:
> >
> >> I will look into turning FF into LF when a script is read, & will do it
> >> if there are no problems.
> >>
> >> Does Linux do something to the FF?  I don't see why you don't get the
> >> same error there as on Windows.
> >>
> >> Henry Rich
> >>
> >> On 7/18/2021 7:58 AM, Michal Wallace wrote:
> >>> I sometimes use the ^L character (FF, or 12{a.) in source code files to
> >>> divide my code into "pages".
> >>>
> >>> (I have emacs set up so that I can page back and forth showing one page
> >> at
> >>> a time without scrolling, so the top line of the page is always at the
> >> top
> >>> of the screen when that page is showing, etc.)
> >>>
> >>> I've done this with J source code with no problems in the past. It
> works
> >>> fine in jqt and in jconsole on linux, but I just discovered that if I
> try
> >>> to load a J script written this way in the windows version of jconsole,
> >> it
> >>> gives me a spelling error on that character.
> >>>
> >>> Is this something that could be officially supported, or do I need to
> >>> change my ways? :D
> >>> --
> >>> For information about J forums see http://www.jsoftware.com/forums.htm
> >>
> >> --
> >> This email has been checked for viruses by AVG.
> >> https://www.avg.com
> >>
> >> --
> >> For information about J forums see http://www.jsoftware.com/forums.htm
> >>
> > --
> > For information about J forums see http://www.jsoftware.com/forums.htm
>
>
> --
> This email has been checked for viruses by AVG.
> https://www.avg.com
>
> --
> For information about J forums see http://www.jsoftware.com/forums.htm
>
--
For information about J forums see http://www.jsoftware.com/forums.htm


Re: [Jgeneral] FF character in J source files on windows jconsole.

2021-07-18 Thread Henry Rich
Every bytecode that is not an ASCII graphic or SP, CR, TAB, or LF will 
cause a spelling error if encountered as a word:


   ". 12 { a.
|spelling error
|
The question here is whether FF is enough of an industry-standard to be 
called whitespace, like TAB.


I don't know myself.  I have looked at the code & decided that making FF 
look like LF for script purposes is not too hard, while making it look 
like TAB takes more work.


But I am loath to make any change at all until I understand why FF works 
sometimes and not others.  As I see it, any script containing FF would 
fail when loaded, unless the FF were buried inside a string.  If you 
have cases where this is not so I want to know where the FF is 
disappearing to.


In the meantime you can use Eric's trick of stripping FF during load.

Henry Rich



On 7/18/2021 3:56 PM, Eric Iverson wrote:

FF char (12{a.) gets a spelling error if in a script that is loaded in
linux. I think this would be the same as in windows but have not tested.

There are lots of values that will give spelling errors. I am not sure it
makes sense to special case (i.e. ignore FF).

If the FF is desired for its behavior in an editor, that is OK. But it and
perhaps other chars should be stripped out before being fed to load.

On Sun, Jul 18, 2021 at 3:39 PM Henry Rich  wrote:


I will look into turning FF into LF when a script is read, & will do it
if there are no problems.

Does Linux do something to the FF?  I don't see why you don't get the
same error there as on Windows.

Henry Rich

On 7/18/2021 7:58 AM, Michal Wallace wrote:

I sometimes use the ^L character (FF, or 12{a.) in source code files to
divide my code into "pages".

(I have emacs set up so that I can page back and forth showing one page

at

a time without scrolling, so the top line of the page is always at the

top

of the screen when that page is showing, etc.)

I've done this with J source code with no problems in the past. It works
fine in jqt and in jconsole on linux, but I just discovered that if I try
to load a J script written this way in the windows version of jconsole,

it

gives me a spelling error on that character.

Is this something that could be officially supported, or do I need to
change my ways? :D
--
For information about J forums see http://www.jsoftware.com/forums.htm


--
This email has been checked for viruses by AVG.
https://www.avg.com

--
For information about J forums see http://www.jsoftware.com/forums.htm


--
For information about J forums see http://www.jsoftware.com/forums.htm



--
This email has been checked for viruses by AVG.
https://www.avg.com

--
For information about J forums see http://www.jsoftware.com/forums.htm


Re: [Jgeneral] FF character in J source files on windows jconsole.

2021-07-18 Thread Eric Iverson
More on FF. If there is to be preprocessing of a script to make it suitable
for load, it should be done in J, rather than the front end.

   load fix 'test.ijs'

On Sun, Jul 18, 2021 at 3:56 PM Eric Iverson 
wrote:

> FF char (12{a.) gets a spelling error if in a script that is loaded in
> linux. I think this would be the same as in windows but have not tested.
>
> There are lots of values that will give spelling errors. I am not sure it
> makes sense to special case (i.e. ignore FF).
>
> If the FF is desired for its behavior in an editor, that is OK. But it and
> perhaps other chars should be stripped out before being fed to load.
>
> On Sun, Jul 18, 2021 at 3:39 PM Henry Rich  wrote:
>
>> I will look into turning FF into LF when a script is read, & will do it
>> if there are no problems.
>>
>> Does Linux do something to the FF?  I don't see why you don't get the
>> same error there as on Windows.
>>
>> Henry Rich
>>
>> On 7/18/2021 7:58 AM, Michal Wallace wrote:
>> > I sometimes use the ^L character (FF, or 12{a.) in source code files to
>> > divide my code into "pages".
>> >
>> > (I have emacs set up so that I can page back and forth showing one page
>> at
>> > a time without scrolling, so the top line of the page is always at the
>> top
>> > of the screen when that page is showing, etc.)
>> >
>> > I've done this with J source code with no problems in the past. It works
>> > fine in jqt and in jconsole on linux, but I just discovered that if I
>> try
>> > to load a J script written this way in the windows version of jconsole,
>> it
>> > gives me a spelling error on that character.
>> >
>> > Is this something that could be officially supported, or do I need to
>> > change my ways? :D
>> > --
>> > For information about J forums see http://www.jsoftware.com/forums.htm
>>
>>
>> --
>> This email has been checked for viruses by AVG.
>> https://www.avg.com
>>
>> --
>> For information about J forums see http://www.jsoftware.com/forums.htm
>>
>
--
For information about J forums see http://www.jsoftware.com/forums.htm


Re: [Jgeneral] FF character in J source files on windows jconsole.

2021-07-18 Thread Eric Iverson
FF char (12{a.) gets a spelling error if in a script that is loaded in
linux. I think this would be the same as in windows but have not tested.

There are lots of values that will give spelling errors. I am not sure it
makes sense to special case (i.e. ignore FF).

If the FF is desired for its behavior in an editor, that is OK. But it and
perhaps other chars should be stripped out before being fed to load.

On Sun, Jul 18, 2021 at 3:39 PM Henry Rich  wrote:

> I will look into turning FF into LF when a script is read, & will do it
> if there are no problems.
>
> Does Linux do something to the FF?  I don't see why you don't get the
> same error there as on Windows.
>
> Henry Rich
>
> On 7/18/2021 7:58 AM, Michal Wallace wrote:
> > I sometimes use the ^L character (FF, or 12{a.) in source code files to
> > divide my code into "pages".
> >
> > (I have emacs set up so that I can page back and forth showing one page
> at
> > a time without scrolling, so the top line of the page is always at the
> top
> > of the screen when that page is showing, etc.)
> >
> > I've done this with J source code with no problems in the past. It works
> > fine in jqt and in jconsole on linux, but I just discovered that if I try
> > to load a J script written this way in the windows version of jconsole,
> it
> > gives me a spelling error on that character.
> >
> > Is this something that could be officially supported, or do I need to
> > change my ways? :D
> > --
> > For information about J forums see http://www.jsoftware.com/forums.htm
>
>
> --
> This email has been checked for viruses by AVG.
> https://www.avg.com
>
> --
> For information about J forums see http://www.jsoftware.com/forums.htm
>
--
For information about J forums see http://www.jsoftware.com/forums.htm


Re: [Jgeneral] FF character in J source files on windows jconsole.

2021-07-18 Thread Michal Wallace
no clue. as far as i know it's just a plain ascii character. as i said, it
works in jqt on windows, too.

On Sun, Jul 18, 2021 at 3:39 PM Henry Rich  wrote:

> I will look into turning FF into LF when a script is read, & will do it
> if there are no problems.
>
> Does Linux do something to the FF?  I don't see why you don't get the
> same error there as on Windows.
>
> Henry Rich
>
> On 7/18/2021 7:58 AM, Michal Wallace wrote:
> > I sometimes use the ^L character (FF, or 12{a.) in source code files to
> > divide my code into "pages".
> >
> > (I have emacs set up so that I can page back and forth showing one page
> at
> > a time without scrolling, so the top line of the page is always at the
> top
> > of the screen when that page is showing, etc.)
> >
> > I've done this with J source code with no problems in the past. It works
> > fine in jqt and in jconsole on linux, but I just discovered that if I try
> > to load a J script written this way in the windows version of jconsole,
> it
> > gives me a spelling error on that character.
> >
> > Is this something that could be officially supported, or do I need to
> > change my ways? :D
> > --
> > For information about J forums see http://www.jsoftware.com/forums.htm
>
>
> --
> This email has been checked for viruses by AVG.
> https://www.avg.com
>
> --
> For information about J forums see http://www.jsoftware.com/forums.htm
>
--
For information about J forums see http://www.jsoftware.com/forums.htm


Re: [Jgeneral] FF character in J source files on windows jconsole.

2021-07-18 Thread Henry Rich
I will look into turning FF into LF when a script is read, & will do it 
if there are no problems.


Does Linux do something to the FF?  I don't see why you don't get the 
same error there as on Windows.


Henry Rich

On 7/18/2021 7:58 AM, Michal Wallace wrote:

I sometimes use the ^L character (FF, or 12{a.) in source code files to
divide my code into "pages".

(I have emacs set up so that I can page back and forth showing one page at
a time without scrolling, so the top line of the page is always at the top
of the screen when that page is showing, etc.)

I've done this with J source code with no problems in the past. It works
fine in jqt and in jconsole on linux, but I just discovered that if I try
to load a J script written this way in the windows version of jconsole, it
gives me a spelling error on that character.

Is this something that could be officially supported, or do I need to
change my ways? :D
--
For information about J forums see http://www.jsoftware.com/forums.htm



--
This email has been checked for viruses by AVG.
https://www.avg.com

--
For information about J forums see http://www.jsoftware.com/forums.htm


[Jgeneral] FF character in J source files on windows jconsole.

2021-07-18 Thread Michal Wallace
I sometimes use the ^L character (FF, or 12{a.) in source code files to
divide my code into "pages".

(I have emacs set up so that I can page back and forth showing one page at
a time without scrolling, so the top line of the page is always at the top
of the screen when that page is showing, etc.)

I've done this with J source code with no problems in the past. It works
fine in jqt and in jconsole on linux, but I just discovered that if I try
to load a J script written this way in the windows version of jconsole, it
gives me a spelling error on that character.

Is this something that could be officially supported, or do I need to
change my ways? :D
--
For information about J forums see http://www.jsoftware.com/forums.htm