Known nath bugs?

Known nath bugs?

[Sebastian Sturm]

Hello all,

is there a list of know nath bugs somewhere on the web? I tried to find 
one using Google, but unfortunately without success.

The reason I'm asking this is that I'm encountering some problems with 
the module and would like to know if they are common or if it's related 
to my configuration. First of all, using the units module together with 
nath doesn't seem to work - for instance, $1\Micro\Coulomb$ works as 
expected without nath, but produces the error message "Please don't say 
\def cs{...}, say \def\cs{...}" when used together with nath. I also 
thought that delimiters like ( or [ should automatically adjust to the 
formula they contain, but (\sum_{something} 
\frac{something}{something}) only produces standard parentheses. 
Commands like \lbrace or \lbrack work as expected, though.

If these problems are unknown, I can post a small example file and my 
logfile.

Best regards,

Sebastian

Re: Known nath bugs?

[Sebastian Sturm]

Hi,

the delimiter sizing is working now; i had to load the module directly 
before \starttext. However, the units module still isn't working when 
nath is loaded. Is it supposed to?

Best regards,
Sebastian

Re[2]: Known nath bugs?

[Giuseppe Bilotta]

Saturday, May 29, 2004 Sebastian Sturm wrote:

> Hi,

> the delimiter sizing is working now; i had to load the module directly
> before \starttext.

Yes, this is generally true. All modules should get loaded
before \starttext.

> However, the units module still isn't working when nath is
> loaded. Is it supposed to?

Well, I had never tested them together, and given the very
extensive changes nath enacts, I'm not surprised it doesn't
work.

Anyway, it looks like the fix is easy: the problem is that
dimension commands are not protected (unexpandable).

In m-units.tex, look for the lines

\def\dimensionprefix#1%
  {\gdef\thedimensionprefix{#1}}

\def\dimensionaddfix#1%
  {\unskip
  %\mathematics{\umathematicstypeface#1}%
   \ustartmathmode\umathematicstypeface#1\ustopmathmode
   \nobreak
   \hskip\dimensionaddfixsignal}

\def\dimensionnopfix#1%
  {\dontbreakdimension
   \ifdim\scratchdimen=\dimensionpowersignal\relax
     \ustartmathmode
   \else
     \ustartmathmode
     \udimensionhalfspace
     \nobreak
   \fi
   \umathematicstypeface#1%
   \ustopmathmode
   \nobreak
   \hskip\dimensionsignal}

\def\dimensionmidfix#1%
  {\dontbreakdimension
   \ifdim\scratchdimen=\dimensionpowersignal\relax
     \ustartmathmode
     \udimensionbackspace
     \nobreak
   \else
     \ustartmathmode
   \fi
   \umathematicstypeface#1%
   \ustopmathmode
   \nobreak
   \hskip\dimensionmidfixsignal}

\def\dimensionpower#1%
  {\gdef\thedimensionpower{#1}}

and change each

\def\dimension...

into

\unexpanded\def\dimension....

This should solve the problem.

I'm not 100% sure this is the best solution, though. An
alternative could be to keep those definition as they are and
instead make unexpandable the actual commands (\Micro etc).

Hans, what do you think?

-- 
Giuseppe "Oblomov" Bilotta

Re: Re[2]: Known nath bugs?

[Sebastian Sturm]

Hi,

> Yes, this is generally true. All modules should get loaded
> before \starttext.

Well, I had loaded it before \starttext in both cases. But when I 
loaded it in the first line of my file, it didn't adjust the 
delimiters; after moving the \usemodule[nath] line directly above the 
\starttext line, it worked like a charm. Maybe I produced some garbage 
inbetween. :-)

>> However, the units module still isn't working when nath is
>> loaded. Is it supposed to?
>
> [...]
> Anyway, it looks like the fix is easy: the problem is that
> dimension commands are not protected (unexpandable).
>
> [...]

Thank you very much, I'll try that. If it doesn't work, I can still get 
along without units, though it would have been nice to have that too. 
BTW, is there some kind of manual on aligning nath equations? I tried 
to align multiline equations using \startalign, but that only worked 
(in display mode) as long as no delimiters were used;
\[ \startalign a &= (b \\
b &= c) \stopalign \]
always produced an error message saying that I had left out an "}". I 
probably made some mistake, so a tutorial on aligning equations with 
nath would be great. The nathguide.pdf mentions eqnarray, but I guess 
that's not available in ConTeXt, or is it?

Best regards,
Sebastian

Re[4]: Known nath bugs?

[Giuseppe Bilotta]

Saturday, May 29, 2004 Sebastian Sturm wrote:

>> Yes, this is generally true. All modules should get loaded
>> before \starttext.

> Well, I had loaded it before \starttext in both cases. But when I
> loaded it in the first line of my file, it didn't adjust the 
> delimiters; after moving the \usemodule[nath] line directly above the
> \starttext line, it worked like a charm. Maybe I produced some garbage
> inbetween. :-)

Nath is a very complex module, easily broken. If you can
privately send me your preamble we might be able to find what
was breaking it. Or you can try it yourself: move it up line by
line until delimiters don't work anymore: you'll find what
makes it bomb :)

>> Anyway, it looks like the fix is easy: the problem is that
>> dimension commands are not protected (unexpandable).

> Thank you very much, I'll try that. If it doesn't work, I can still get
> along without units, though it would have been nice to have that too.

> BTW, is there some kind of manual on aligning nath equations? I tried
> to align multiline equations using \startalign, but that only worked
> (in display mode) as long as no delimiters were used;
> \[ \startalign a &= (b \\
> b &= c) \stopalign \]
> always produced an error message saying that I had left out an "}". I
> probably made some mistake, so a tutorial on aligning equations with
> nath would be great. The nathguide.pdf mentions eqnarray, but I guess
> that's not available in ConTeXt, or is it?

Not yet. Those multi-equation environments are somewhat more
difficult to convert to ConTeXt. And I still haven't got a
complete grasp on how things work in ConTeXt (esp wrt
numbering), so it's a double effort. It's on my TODO list
though.

-- 
Giuseppe "Oblomov" Bilotta

Re: Re[4]: Known nath bugs?

[Sebastian Sturm]

Hi,

> Nath is a very complex module, easily broken. If you can
> privately send me your preamble we might be able to find what
> was breaking it. Or you can try it yourself: move it up line by
> line until delimiters don't work anymore: you'll find what
> makes it bomb :)

OK, I'll try that. At the moment I'm having all kinds of issues with 
nath (even without any preamble at all), like \binom{\frac 12}{a} not 
working, \to, \ot and \otto arrows not adjusting to the lengths of 
their subscripts, etc... and I can hardly imagine that all of these are 
nath bugs, so maybe something's wrong with my TeX installation. I'm 
using the latest i-Installer teTeX for MacOS X package (with non-beta 
ConTeXt version), configuration files unchanged.
So if someone else is using nath with Gerben Wieda's teTeX 
installation, it would be helpful to know if he/she is experiencing 
similar problems.

Anyway, the nath module looks very promising; I'm really looking 
forward to its next revision!

Best regards,
Sebastian

Re: Re[4]: Known nath bugs?

[Gary Pajer]

> Hi,
>
> > Nath is a very complex module, easily broken. If you can
> > privately send me your preamble we might be able to find what
> > was breaking it. Or you can try it yourself: move it up line by
> > line until delimiters don't work anymore: you'll find what
> > makes it bomb :)
>
> OK, I'll try that. At the moment I'm having all kinds of issues with

Let me point out, mainly for the record, that ConTeXt / nath does not seem
to work properly under MikTeX.

The following script works fine with Linux / teTeX, but with MikTeX, the
in-line versions of the math is wrong.


-gary

------------------------------------------
\usemodule[nath]
\starttext

$ y = (1 + \frac{}{a}{b}) $

$ y = (1 + \frac{a}{b}) $

$ y = (1 + a\frac{b}{c}{d})$

$$ y = (1 + \frac{}{a}{b}) $$

$$ y = (1 + \frac{a}{b}) $$

$$ y = (1 + a\frac{b}{c}{d})$$

\stoptext

Re[6]: Known nath bugs?

[Giuseppe Bilotta]

Tuesday, June 1, 2004 Gary Pajer wrote:

> Let me point out, mainly for the record, that ConTeXt / nath does not seem
> to work properly under MikTeX.

> The following script works fine with Linux / teTeX, but with MikTeX, the
> in-line versions of the math is wrong.


> -gary

> ------------------------------------------
> \usemodule[nath]
> \starttext

> $ y = (1 + \frac{}{a}{b}) $

> $ y = (1 + \frac{a}{b}) $

> $ y = (1 + a\frac{b}{c}{d})$

> $$ y = (1 + \frac{}{a}{b}) $$

> $$ y = (1 + \frac{a}{b}) $$

> $$ y = (1 + a\frac{b}{c}{d})$$

> \stoptext

The latest nath (submitted to CTAN, you can retrieve it from
there) works correctly on my MiKTeX installation, both inline
and displayed obey the structure. Please try updating nath. If it
still doesn't work: do you have the same ConTeXt version on the
two system?

-- 
Giuseppe "Oblomov" Bilotta

Re: Re[6]: Known nath bugs?

[Gary Pajer]

> The latest nath (submitted to CTAN, you can retrieve it from
> there) works correctly on my MiKTeX installation, both inline
> and displayed obey the structure. Please try updating nath. If it
> still doesn't work: do you have the same ConTeXt version on the
> two system?
>
> -- 
> Giuseppe "Oblomov" Bilotta

Thank you very much, yes, the new version does work.

Giuseppe, please note:  I waited a few days to make sure CTAN was up to
date.  When I first looked for the file, there were different versions on
different mirrors.  One was dated 2003.12.07, the other 2003.12.08.    The
version that I downloaded today, and works properly, is dated 2003.12.08.
Typo?

-gary

Re[8]: Known nath bugs?

[Giuseppe Bilotta]

Tuesday, June 8, 2004 Gary Pajer wrote:

> Thank you very much, yes, the new version does work.

> Giuseppe, please note:  I waited a few days to make sure CTAN was up to
> date.  When I first looked for the file, there were different versions on
> different mirrors.  One was dated 2003.12.07, the other 2003.12.08.    The
> version that I downloaded today, and works properly, is dated 2003.12.08.
> Typo?

I simply often forget to update the date, if you're talking about the
date *in* the file (as opposed to the date *of* the file) :\

-- 
Giuseppe "Oblomov" Bilotta

Re[6]: Known nath bugs?

[Giuseppe Bilotta]

Sunday, May 30, 2004 Sebastian Sturm wrote:

> OK, I'll try that. At the moment I'm having all kinds of issues with
> nath (even without any preamble at all), like \binom{\frac 12}{a} not
> working, \to, \ot and \otto arrows not adjusting to the lengths of
> their subscripts, etc... and I can hardly imagine that all of these are
> nath bugs, so maybe something's wrong with my TeX installation. I'm
> using the latest i-Installer teTeX for MacOS X package (with non-beta
> ConTeXt version), configuration files unchanged.
> So if someone else is using nath with Gerben Wieda's teTeX 
> installation, it would be helpful to know if he/she is experiencing
> similar problems.

If you could please provide some example file showing the
problems, I'll see if I can nail them down.

-- 
Giuseppe "Oblomov" Bilotta

Re: Re[6]: Known nath bugs?

[Sebastian Sturm]

[-- Attachment #1: Type: text/plain, Size: 395 bytes --]

Hi,

> If you could please provide some example file showing the
> problems, I'll see if I can nail them down.

the problem with delimiter autosizing seems to be caused by 
MPenvironment; however, the problem with subscripts still seems very 
strange to me (since superscripts work as expected). I've included a 
test file and the PDF generated by my TeX installation.

Best regards,

Sebastian

[-- Attachment #2: nath_test.tex --]
[-- Type: application/x-tex, Size: 497 bytes --]

[-- Attachment #3: nath_test.log --]
[-- Type: application/octet-stream, Size: 3970 bytes --]

This is pdfeTeXk, Version 3.141592-1.11b-2.1 (Web2C 7.5.2) (format=cont-en 2004.5.14)  6 JUN 2004 11:40
entering extended mode
 %&-line parsing enabled.
 (/usr/local/teTeX/share/texmf.tetex/web2c/cp8bit.tcx)
**&cont-en nath_test.tex
(./nath_test.tex{/usr/local/teTeX/share/texmf.local/pdftex/config/pdftex.cfg}

ConTeXt  ver: 2004.1.13  fmt: 2004.5.14  int: english  mes: english

language       : language en is active
<protectionstate 0>
system         : cont-new loaded
(/usr/local/teTeX/share/texmf.local/tex/context/base/cont-new.tex
systems        : beware: some patches loaded from cont-new.tex!
system (E-TEX) : [line 99] 
system (E-TEX) : [line 105] \lastnodetype 
system (E-TEX) : [line 151] 
system (E-TEX) : [line 237] \lastnodetype 
)
system         : cont-old loaded
(/usr/local/teTeX/share/texmf.local/tex/context/base/cont-old.tex
loading        : Context Old Macros
)
system         : cont-fil loaded
(/usr/local/teTeX/share/texmf.local/tex/context/base/cont-fil.tex
loading        : Context File Synonyms
)
bodyfont       : 12pt rm is loaded
language       : patterns 2:2-en-2 3:2-uk-2 4:2-de-2 5:2-fr-2 6:2-es-2 7:2-it-2
 8:2-nl-2 loaded
specials       : tex,postscript,rokicki loaded
\openout2 = `nath_test.tui'.

system         : nath_test.top loaded
(./nath_test.top
specials       : loading definition file tpd
(/usr/local/teTeX/share/texmf.local/tex/context/base/spec-tpd.tex
specials       : loading definition file fdf
(/usr/local/teTeX/share/texmf.local/tex/context/base/spec-fdf.tex
<unprotect 3> <unprotect 4>
system (E-TEX) : [line 1907] \ifcsname 
<protect 4> <protect 3>)
specials       : fdf loaded
<unprotect 3> <protect 3>)
specials       : fdf,tpd loaded
)
\openout0 = `mpgraph.mp'.

\openout0 = `mprun.mp'.

\openout0 = `mpgraph.mp'.

\openout0 = `mprun.mp'.


pdftex         : needs map file: original-context-symbol.map
pdftex         : needs map file: pl0-ams-cmr.map
pdftex         : needs map file: original-vogel-symbol.map
system         : macros of module nath loaded
(/usr/local/teTeX/share/texmf.local/tex/context/user/t-nath.tex
system         : macros of module amsl loaded
(/usr/local/teTeX/share/texmf.local/tex/context/user/t-amsl.tex
system         : macros of module newmat loaded
(/usr/local/teTeX/share/texmf.local/tex/context/base/m-newmat.tex)))
(./nath_test.tuo) (./nath_test.tuo) (./nath_test.tuo) (./nath_test.tuo)
(./nath_test.tuo) (./nath_test.tuo) (./nath_test.tuo) (./nath_test.tuo)
(./nath_test.tuo) (./nath_test.tuo) (./nath_test.tuo) (./nath_test.tuo)
(./nath_test.tuo) (./nath_test.tuo)
systems        : begin file nath_test at line 14
[1.1{/usr/local/teTeX/share/texmf.local/fonts/map/pdftex/updmap/pdftex.map}]
systems        : end file nath_test at line 23
 ) 
Here is how much of TeX's memory you used:
 1458 strings out of 66961
 19259 string characters out of 725323
 4435645 words of memory out of 5534819
 32259 multiletter control sequences out of 10000+50000
 17900 words of font info for 60 fonts, out of 1000000 for 2000
 62 hyphenation exceptions out of 1000
 44i,16n,49p,189b,310s stack positions out of 5000i,500n,6000p,200000b,40000s
 25 PDF objects out of 300000
 0 named destinations out of 131072
 1 words of extra memory for PDF output out of 65536
{/usr/local/teTeX/share/texmf.tetex/dvips/tetex/bbad153f.enc}</usr/local/teTe
X/share/texmf.tetex/fonts/type1/bluesky/cm/cmsy7.pfb></usr/local/teTeX/share/te
xmf.tetex/fonts/type1/bluesky/cm/cmex10.pfb>{/usr/local/teTeX/share/texmf.tetex
/dvips/tetex/f7b6d320.enc}</usr/local/teTeX/share/texmf.tetex/fonts/type1/blues
ky/cm/cmr7.pfb></usr/local/teTeX/share/texmf.tetex/fonts/type1/bluesky/cm/cmsy1
0.pfb>{/usr/local/teTeX/share/texmf.tetex/dvips/tetex/aae443f0.enc}</usr/local/
teTeX/share/texmf.tetex/fonts/type1/bluesky/cm/cmmi7.pfb></usr/local/teTeX/shar
e/texmf.tetex/fonts/type1/bluesky/cm/cmmi10.pfb></usr/local/teTeX/share/texmf.t
etex/fonts/type1/bluesky/cm/cmr10.pfb>
Output written on nath_test.pdf (1 page, 28021 bytes).

[-- Attachment #4: nath_test.pdf --]
[-- Type: application/pdf, Size: 28021 bytes --]

[-- Attachment #5: Type: text/plain, Size: 1 bytes --]

Re[8]: Known nath bugs?

[Giuseppe Bilotta]

Sunday, June 6, 2004 Sebastian Sturm wrote:

> Hi,

>> If you could please provide some example file showing the
>> problems, I'll see if I can nail them down.

> the problem with delimiter autosizing seems to be caused by 
> MPenvironment; however, the problem with subscripts still seems very
> strange to me (since superscripts work as expected). I've included a
> test file and the PDF generated by my TeX installation.

I will look into the interaction with MPenvironment. To solve
the subscript problem, please change line 2738 of t-nath.tex
from

 \ifdim\wd\!!boxa>\er@wd \er@wd=\wd\nathbox\fi

to

 \ifdim\wd\!!boxa>\er@wd \er@wd=\wd\!!boxa\fi

(in the definition of \axis@)

and let me know how it works.

-- 
Giuseppe "Oblomov" Bilotta

Re: Re[8]: Known nath bugs?

[Sebastian Sturm]

> I will look into the interaction with MPenvironment. To solve
> the subscript problem, please change line 2738 of t-nath.tex
> from
>
>  \ifdim\wd\!!boxa>\er@wd \er@wd=\wd\nathbox\fi
>
> to
>
>  \ifdim\wd\!!boxa>\er@wd \er@wd=\wd\!!boxa\fi
>
> (in the definition of \axis@)
>
> and let me know how it works.

Hi,

works perfectly, thanks!

Best regards,
Sebastian

Re[10]: Known nath bugs?

[Giuseppe Bilotta]

Sunday, July 4, 2004 Sebastian Sturm wrote:

>> I will look into the interaction with MPenvironment. To solve
>> the subscript problem, please change line 2738 of t-nath.tex
>> from
>>
>>  \ifdim\wd\!!boxa>\er@wd \er@wd=\wd\nathbox\fi
>>
>> to
>>
>>  \ifdim\wd\!!boxa>\er@wd \er@wd=\wd\!!boxa\fi
>>
>> (in the definition of \axis@)
>>
>> and let me know how it works.

> Hi,

> works perfectly, thanks!


Good. I'll submit the new version to CTAN.

-- 
Giuseppe "Oblomov" Bilotta

Re[4]: Known nath bugs?

[Hans Hagen]

At 11:47 30/05/2004, Giuseppe Bilotta wrote:

>Nath is a very complex module, easily broken. If you can
>privately send me your preamble we might be able to find what
>was breaking it. Or you can try it yourself: move it up line by
>line until delimiters don't work anymore: you'll find what
>makes it bomb :)

one problem is that nath redefines and has no real namespace; once you got 
things properly working, i'll go over the module and see what needs to be 
fixed

>Not yet. Those multi-equation environments are somewhat more
>difficult to convert to ConTeXt. And I still haven't got a
>complete grasp on how things work in ConTeXt (esp wrt
>numbering), so it's a double effort. It's on my TODO list
>though.

just make things so that they work acceptable and i'll handle the nasty 
details -)

Hans