nath catcode issues

nath catcode issues

[Nikolai Weibull]

It seems that the code to enable catcode handling in nath math mode
isn't working.  Currently, nath uses

\appendtoks
% \livechars@
\def({\delim@l 1\ch@lparenthesis}
\def){\delim@r 1\ch@rparenthesis}
\def[{\delim@l 1\o@lbrack}
\def]{\delim@r 1\o@rbrack}
\def>{\gt}
\def<{\lt}
\def!{\factorial}
\def,{\@comma}
\def;{\@semicolon}
\def`{\abbreviation}
% \killchars@
\to\everymath

to set it up, but as far as I can see this doesn't actually do anything
as \@comma doesn't get invoked for commas and \dots always renders like
\cdots.  Am I missing something perhaps?

Furthermore,

\def\over{\n@err{Disabled command \string\over }
 {No more a valid command. Replace {A \over B} with \frac A B.}}
\def\atop{\n@err{Disabled command \string\atop }
 {No more a valid command. Use array instead.}}
\def\choose{\n@err{Disabled command \string\choose }
 {No more a valid command. Replace {A \choose B} with \binom A B.}}

also messes up stuff in core-ref.tex for example:

! Disabled command \over.
\n@err ...rrhelp \expandafter {#2}\errmessage {#1}
                                                  \natherrormark
l.46 \indent\inchapter[definitions], \about
                                           [definitions], covers the definit...

?

Which I don't quite understand, but core-ref.tex has the line

\definecommand over {\dospecialover}

which I suppose could be troublesome.
	nikolai

--
::: name: Nikolai Weibull    :: aliases: pcp / lone-star / aka :::
::: born: Chicago, IL USA    :: loc atm: Gothenburg, Sweden    :::
::: page: www.pcppopper.org  :: fun atm: gf,lps,ruby,lisp,war3 :::
main(){printf(&linux["\021%six\012\0"],(linux)["have"]+"fun"-97);}

Re: nath catcode issues

[Giuseppe Bilotta]

Sunday, July 4, 2004 Nikolai Weibull wrote:

> It seems that the code to enable catcode handling in nath math mode
> isn't working.  Currently, nath uses

[snip]

> to set it up, but as far as I can see this doesn't actually do anything
> as \@comma doesn't get invoked for commas and \dots always renders like
> \cdots.  Am I missing something perhaps?

Well, the nath catcodes only apply in math mode, and indeed in
math mode , will trigger \@comma :)

About \dots: with the nath module, \dots will *in math mode*
use \cdots in all cases *except* after punctuation (e.g. after
a comma); indeed, the code

\[
a=b, \dots c=d \dots
\]

will give you \ldots in the first case and \cdots in the second
case. Do you have some example of when it doesn't work?

> Furthermore,

> \def\over{\n@err{Disabled command \string\over }
>  {No more a valid command. Replace {A \over B} with \frac A B.}}
> \def\atop{\n@err{Disabled command \string\atop }
>  {No more a valid command. Use array instead.}}
> \def\choose{\n@err{Disabled command \string\choose }
>  {No more a valid command. Replace {A \choose B} with \binom A B.}}

> also messes up stuff in core-ref.tex for example:

AHA! Good point. Try this fix:

\def\over{\ifmmode\n@err{Disabled command \string\over }
 {No more a valid command. Replace {A \over B} with \frac A
 B.}\else\o@over\fi}
\def\atop{\ifmmode\n@err{Disabled command \string\atop }
 {No more a valid command. Use array instead.}\else\o@atop\fi}
\def\choose{\ifmmode\n@err{Disabled command \string\choose }
 {No more a valid command. Replace {A \choose B} with \binom A
 B.}\else\o@choose\fi}
 
(hm. this is untested. I'm sure it'll bomb out because of
expansion, I'll have to design a cleaner one.)

-- 
Giuseppe "Oblomov" Bilotta

Re: nath catcode issues

[Nikolai Weibull]

* Giuseppe Bilotta <gip.bilotta@iol.it> [Jul 05, 2004 12:10]:
> Well, the nath catcodes only apply in math mode, and indeed in
> math mode , will trigger \@comma :)

Hm, not here...

> About \dots: with the nath module, \dots will *in math mode*
> use \cdots in all cases *except* after punctuation (e.g. after
> a comma); indeed, the code
>
> \[
> a=b, \dots c=d \dots
> \]
>
> will give you \ldots in the first case and \cdots in the second
> case. Do you have some example of when it doesn't work?

Your example above renders with \cdots for both \dots.

\usemodule[nath]
\starttext
$a, b, \dots, n$
\[ a=b, \dots c=d \dots \]
\stoptext

The first line is my old test case, the second is yours...both render
with \cdots everywhere.  Totally weird.  I'm using the latest context
from the pragma-ade page.

> \def\over{\ifmmode\n@err{Disabled command \string\over }
>  {No more a valid command. Replace {A \over B} with \frac A
>  B.}\else\o@over\fi}
> \def\atop{\ifmmode\n@err{Disabled command \string\atop }
>  {No more a valid command. Use array instead.}\else\o@atop\fi}
> \def\choose{\ifmmode\n@err{Disabled command \string\choose }
>  {No more a valid command. Replace {A \choose B} with \binom A
>  B.}\else\o@choose\fi}

Ah, of course...
	nikolai

--
::: name: Nikolai Weibull    :: aliases: pcp / lone-star / aka :::
::: born: Chicago, IL USA    :: loc atm: Gothenburg, Sweden    :::
::: page: www.pcppopper.org  :: fun atm: gf,lps,ruby,lisp,war3 :::
main(){printf(&linux["\021%six\012\0"],(linux)["have"]+"fun"-97);}

Re: Re: nath catcode issues

[Giuseppe Bilotta]

Monday, July 5, 2004 Nikolai Weibull wrote:

> * Giuseppe Bilotta <gip.bilotta@iol.it> [Jul 05, 2004 12:10]:
>> Well, the nath catcodes only apply in math mode, and indeed in
>> math mode , will trigger \@comma :)

> Hm, not here...

[snip]

> Your example above renders with \cdots for both \dots.

Weird indeed. I'm not using the latest ConTeXt. (Actually I'm
quite back with the updates ... gotta get down to it.) I'll see
if upgrading gives me the same problem and in case get a fix.

>> \def\over{\ifmmode\n@err{Disabled command \string\over }
>>  {No more a valid command. Replace {A \over B} with \frac A
>>  B.}\else\o@over\fi}
>> \def\atop{\ifmmode\n@err{Disabled command \string\atop }
>>  {No more a valid command. Use array instead.}\else\o@atop\fi}
>> \def\choose{\ifmmode\n@err{Disabled command \string\choose }
>>  {No more a valid command. Replace {A \choose B} with \binom A
>>  B.}\else\o@choose\fi}

> Ah, of course...
> 	nikolai

It works like that? Excellent.

-- 
Giuseppe "Oblomov" Bilotta

Re: Re: nath catcode issues

[Nikolai Weibull]

* Giuseppe Bilotta <gip.bilotta@iol.it> [Jul 05, 2004 14:00]:
> Weird indeed. I'm not using the latest ConTeXt. (Actually I'm
> quite back with the updates ... gotta get down to it.) I'll see
> if upgrading gives me the same problem and in case get a fix.

OK.

> >> \def\over{\ifmmode\n@err{Disabled command \string\over }
> >>  {No more a valid command. Replace {A \over B} with \frac A
> >>  B.}\else\o@over\fi}
> >> \def\atop{\ifmmode\n@err{Disabled command \string\atop }
> >>  {No more a valid command. Use array instead.}\else\o@atop\fi}
> >> \def\choose{\ifmmode\n@err{Disabled command \string\choose }
> >>  {No more a valid command. Replace {A \choose B} with \binom A
> >>  B.}\else\o@choose\fi}
>
> > Ah, of course...
> > 	nikolai
>
> It works like that? Excellent.

It seems I spoke too soon.  It does in fact not work.  \o@over can't be
put like that inside the if-statement.  A suggestion would be to re-work
the whole section and instead make it

\appendtoks
\def\over{\n@err{Disabled command \string\over }
 {No more a valid command. Replace {A \string\over B} with \string\frac A B.}}
\def\atop{\n@err{Disabled command \string\atop }
 {No more a valid command. Use array instead.}}
\def\choose{\n@err{Disabled command \string\choose }
 {No more a valid command. Replace {A \string\choose B} with \string\binom A B. >>> }}
\to\everymath

but I don't know if that's a good solution.  It forces you to use
\string\over in the second argument to \n@err which I don't understand
why, but otherwise TeX runs out of stack space.  I suppose it's being
expanded even though it shouldn't be.
	nikolai

--
::: name: Nikolai Weibull    :: aliases: pcp / lone-star / aka :::
::: born: Chicago, IL USA    :: loc atm: Gothenburg, Sweden    :::
::: page: www.pcppopper.org  :: fun atm: gf,lps,ruby,lisp,war3 :::
main(){printf(&linux["\021%six\012\0"],(linux)["have"]+"fun"-97);}

Re: Re: Re: nath catcode issues

[Giuseppe Bilotta]

Monday, July 5, 2004 Nikolai Weibull wrote:

> * Giuseppe Bilotta <gip.bilotta@iol.it> [Jul 05, 2004 14:00]:
>> Weird indeed. I'm not using the latest ConTeXt. (Actually I'm
>> quite back with the updates ... gotta get down to it.) I'll see
>> if upgrading gives me the same problem and in case get a fix.

> OK.

I'm now using

ConTeXt  ver: 2004.6.30  fmt: 2004.7.27  int: english  mes: english

and your example

\usemodule[nath]

\starttext

$a, b, \dots, n$
\[
a=b, \dots c=d \dots
\]
\stoptext

uses, in order, \ldots, \ldots, \cdots which sounds like the
correct behaviour to me. The engine is

This is pdfeTeX, Version 3.141592-1.20a-rc4-2.1 (MiKTeX 2.4)


>> >> \def\over{\ifmmode\n@err{Disabled command \string\over }
>> >>  {No more a valid command. Replace {A \over B} with \frac A
>> >>  B.}\else\o@over\fi}
>> >> \def\atop{\ifmmode\n@err{Disabled command \string\atop }
>> >>  {No more a valid command. Use array instead.}\else\o@atop\fi}
>> >> \def\choose{\ifmmode\n@err{Disabled command \string\choose }
>> >>  {No more a valid command. Replace {A \choose B} with \binom A
>> >>  B.}\else\o@choose\fi}
>>
>> > Ah, of course...
>> > 	nikolai
>>
>> It works like that? Excellent.

> It seems I spoke too soon.  It does in fact not work.  \o@over can't be
> put like that inside the if-statement.  A suggestion would be to re-work
> the whole section and instead make it

> \appendtoks
> \def\over{\n@err{Disabled command \string\over }
>  {No more a valid command. Replace {A \string\over B} with \string\frac A B.}}
> \def\atop{\n@err{Disabled command \string\atop }
>  {No more a valid command. Use array instead.}}
> \def\choose{\n@err{Disabled command \string\choose }
>  {No more a valid command. Replace {A \string\choose B}
> with \string\binom A B. >>> }}
> \to\everymath

> but I don't know if that's a good solution.  It forces you to use
> \string\over in the second argument to \n@err which I don't understand
> why, but otherwise TeX runs out of stack space.  I suppose it's being
> expanded even though it shouldn't be.
> 	nikolai

Also, I'm not sure that will work with text inside math. Can
you design a small test case that uses \over (directly or
indirectly) in normal text and in some \text inside a math
environment so that I can see what solution works fine?

In the meantime, I'm thinking along the lines of something like

\def\over{%
 \ifmmode\n@err{Disabled command \string\over }
  {No more a valid command. Replace {A \over B} with \frac A B.}%
  \let\next\relax
 \else\let\next\o@over\fi\next}
\def\atop{%
 \ifmmode\n@err{Disabled command \string\atop }
  {No more a valid command. Use array instead.}%
  \let\next\relax
 \else\let\next\o@atop\fi\next}
\def\choose{%
 \ifmmode\n@err{Disabled command \string\choose }
  {No more a valid command. Replace {A \choose B} with \binom A B.}%
  \lext\next\relax
 \else\let\next\o@choose\fi\next}

which should be robust enough. Can you try it in the context
where my previous solution busted?

-- 
Giuseppe "Oblomov" Bilotta

Re: Re: Re: nath catcode issues

[Nikolai Weibull]

* Giuseppe Bilotta <gip.bilotta@iol.it> [Aug 08, 2004 16:30]:
> ConTeXt  ver: 2004.6.30  fmt: 2004.7.27  int: english  mes: english

Me too.

> [Example]

Argh, that doesn't happen here...

> This is pdfeTeX, Version 3.141592-1.20a-rc4-2.1 (MiKTeX 2.4)

pdfeTeX (Web2C 7.4.5) 3.14159-1.10b-2.1

I run latest tetex.

> > > It works like that? Excellent.

> > It seems I spoke too soon.  It does in fact not work.

> In the meantime, I'm thinking along the lines of something like

> \def\over{%
>  \ifmmode\n@err{Disabled command \string\over }
>   {No more a valid command. Replace {A \over B} with \frac A B.}%
>   \let\next\relax
>  \else\let\next\o@over\fi\next}
> \def\atop{%
>  \ifmmode\n@err{Disabled command \string\atop }
>   {No more a valid command. Use array instead.}%
>   \let\next\relax
>  \else\let\next\o@atop\fi\next}
> \def\choose{%
>  \ifmmode\n@err{Disabled command \string\choose }
>   {No more a valid command. Replace {A \choose B} with \binom A B.}%
>   \lext\next\relax
>  \else\let\next\o@choose\fi\next}

> which should be robust enough. Can you try it in the context
> where my previous solution busted?

I don't remember what busted, but it looks good.  It works for my
thesis, which is weird because you misspelled a \let (\lext) and it
still worked fine.

Furthermore, nested parentheses and stuff that produces implicit
parentheses don't scale correctly.  In fact, \left and \right don't seem
to work either:

\usemodule[nath]
\starttext
\[ (\frac{a}{b})^2 \]
versus
\[ \left(\frac{a}{b}\right)^2 \]
\stoptext

...,
	nikolai

--
::: name: Nikolai Weibull    :: aliases: pcp / lone-star / aka :::
::: born: Chicago, IL USA    :: loc atm: Gothenburg, Sweden    :::
::: page: www.pcppopper.org  :: fun atm: gf,lps,ruby,lisp,war3 :::
main(){printf(&linux["\021%six\012\0"],(linux)["have"]+"fun"-97);}

Re: Re: Re: Re: nath catcode issues

[Giuseppe Bilotta]

Sunday, August 8, 2004 Nikolai Weibull wrote:

> * Giuseppe Bilotta <gip.bilotta@iol.it> [Aug 08, 2004 16:30]:
>> ConTeXt  ver: 2004.6.30  fmt: 2004.7.27  int: english  mes: english

> Me too.

>> [Example]

> Argh, that doesn't happen here...

>> This is pdfeTeX, Version 3.141592-1.20a-rc4-2.1 (MiKTeX 2.4)

> pdfeTeX (Web2C 7.4.5) 3.14159-1.10b-2.1

> I run latest tetex.


> I don't remember what busted, but it looks good.  It works for my
> thesis, which is weird because you misspelled a \let (\lext) and it
> still worked fine.

That part doesn't get reached by your thesis, evidently :)

> Furthermore, nested parentheses and stuff that produces implicit
> parentheses don't scale correctly.  In fact, \left and \right don't seem
> to work either:

> \usemodule[nath]
> \starttext
> \[ (\frac{a}{b})^2 \]
> versus
> \[ \left(\frac{a}{b}\right)^2 \]
> \stoptext

This works here :\

Ok, looks like my version of nath and yours are more different
than I thought at first. I'll try to upload a new version in the
next days.

-- 
Giuseppe "Oblomov" Bilotta