Hi,
On Mac, I am using gsed of MacPorts rather than /usr/bin/sed in extract_categories.sh and avoid the problems.
I forgot the reason, I wonder some incompatibility about regular expression.
By the way, if warnings of empty argument for @opencatbox and @closecatbox annoy you, you can add the following lines right after the line of python execution.
--
# fixed warings about @opencatbox and @closecatbox
for f in *.texi; do
gsed 's/^@opencatbox/@opencatbox {}/; s/^@closecatbox/@closecatbox {}/' "$f" > tmp.texi
mv tmp.texi "$f"
done
--
-
ICHIKAWA, Yuji
On 2011/11/05, at 10:10, Michel Talon wrote:
> Dieter Kaiser wrote:
>> Most messages are generated from the building of the categories. I have
>> always ignored all these messages. I have cut out the categories for the
>> German manual, because the categories do not work for the German manual.
>> Without the categories I get the following with the command `make html':
>>
>> dieter at dieter:~/Maxima/maxima/doc/info/de$ make html
>> perl texi2html -split_chapter --lang=de --output=.
>> --css-include=../manual.css --init-file ../texi2html.init maxima.texi
>> *** Undefined node `iframe_flag' in @ref (in ./Itensor.de.texi l. 1254
>> in mref)
>> *** Undefined node `vect_coords' in @ref (in ./Itensor.de.texi l. 1411
>> in mref)
>> *** Undefined node `iframe_flag' in @ref (in ./Itensor.de.texi l. 1467
>> in mref)
>> dieter at dieter:~/Maxima/maxima/doc/info/de$
>>
>> At this time the definition of three nodes is missing. There are no more
>> problems in the German manual. Without the categories in the English
>> manual there are no problems too.
>>
>
>
> Indeed all the messages are produced running the command
> sh extract_categories.sh maxima
>
> This is for a large part because the shell is marked to trace execution of
> commands, but of course there are errors. The first error is produced when
> filtering Affine.texi through various sed and awk filters.
>
> The following
> @deffn {Function} fast_linsolve ([@var{expr_1}, ..., @var{expr_m}],
> [@var{x_1}, ..., @var{x_n}])
>
> is filtered incorrectly and gets verbatim in the python file
> tmp-make-categories.py which appears in a temporary directory.
>
> It thus begins like:
>
> categories = {}
>
> foo = []
> for x in items: foo.append ([items[0], x])
> try: categories ["Polynomials"] . extend (foo)
> except KeyError: categories ["Polynomials"] = foo
>
> foo = []
> for x in items: foo.append ([items[0], x])
> try: categories ["Groebner bases"] . extend (foo)
> except KeyError: categories ["Groebner bases"] = foo
>
> foo = []
> for x in items: foo.append ([items[0], x])
> try: categories ["Share packages"] . extend (foo)
> except KeyError: categories ["Share packages"] = foo
>
> foo = []
> for x in items: foo.append ([items[0], x])
> try: categories ["Package affine"] . extend (foo)
> except KeyError: categories ["Package affine"] = foo
>
> @deffn {Function} fast_linsolve ([@var{expr_1}, ..., @var{expr_m}],
> [@var{x_1}, ..., @var{x_n}])
>
> ?..
>
>
> The first statements are intended to enter the various categories in an
> associative array and are OK. But the @deffn if of course syntactically
> incorrect for python which stops immediately here.
>
> Then no categories are defined and all the rest of the scripts is bound to
> fail horribly.
>
> So to solve the problem one has to understand why the above line is not
> filtered by sed and awk like other lines are. How is it possible that it works
> on some installations, apparently? Perhaps a different version of sed and awk
> like gawk? The regular expressions here are too messy for my understanding at
> the moment.
>
>
>
> --
> Michel Talon
>
> _______________________________________________
> Maxima mailing list
> Maxima at math.utexas.edu
> http://www.math.utexas.edu/mailman/listinfo/maxima