On Thu, Aug 27, 2009 at 1:55 PM, Martin Rubey wrote:
>
> Bill Page writes:
>
>> What version of Sage are you using? What patches applied
>> (if any)?
>
> 4.0.2, combinat branch, patches 6318 applied.
>
Ok, thanks. I am currently testing with Sage 4.1.1. Without patch
6318, I get the error that I reported. As Adam said, 6318 patch
corrects this error. But if I install clisp and fricas-1.0.3.p0 as
some other people suggested in this thread I still see some errors in
the doc tests. Specifically:
wsp...@debian:~/sage-4.1.1$ ./sage
----------------------------------------------------------------------
| Sage Version 4.1.1, Release Date: 2009-08-14 |
| Type notebook() for the GUI, and license() for information. |
----------------------------------------------------------------------
sage: b=fricas('x^2 + y^2 + 1/2')
sage: b.type()
Polynomial Fraction Integer
sage: b.sage()
ERROR: An unexpected error occurred while tokenizing input
The following traceback may be corrupted or invalid
The error message is: ('EOF in multi-line statement', (6, 0))
---------------------------------------------------------------------------
NotImplementedError Traceback (most recent call last)
/home/wspage/.sage/temp/debian/477/_home_wspage__sage_init_sage_0.py
in <module>()
/home/wspage/sage-4.1.1/local/lib/python2.6/site-packages/sage/interfaces/expect.pyc
in sage(self)
1576 Rational Field
1577 """
-> 1578 return self._sage_()
1579
1580 def __repr__(self):
/home/wspage/sage-4.1.1/local/lib/python2.6/site-packages/sage/interfaces/axiom.pyc
in _sage_(self)
878 elif type.startswith('Polynomial'):
879 from sage.rings.all import PolynomialRing
--> 880 base_ring = P(type.lstrip('Polynomial '))._sage_domain()
881 vars = str(self.variables())[1:-1]
882 R = PolynomialRing(base_ring, vars)
/home/wspage/sage-4.1.1/local/lib/python2.6/site-packages/sage/interfaces/fricas.pyc
in _sage_domain(self)
284 return P(name)._sage_domain().fraction_field()
285
--> 286 raise NotImplementedError
287
288 class FriCASFunctionElement(PanAxiomFunctionElement):
NotImplementedError:
sage:
--
Looking at the code I can see this is because the fricas.py interface
(which depends on the axiom.py interface) assumes that the type will
be expressed as a fully parenthesized string like this:
Polynomial(Fraction(Integer))
but this only applies in later versions of FriCAS.
So this patch really does depend on installing a later version of
FriCAS than what is available at the Sage web site.
Also, if I try to install the newer version of FriCAS included with patch 6517:
$ wget http://www.mediafire.com/file/im5zd201mh0/fricas-1.0.7.p0.spkg
$ ./sage -f fricas-1.0.7.p0.spkg
I get an error because by default the fricas configure and make files
trt to build the Aldor interface and it fails.
...
touch -t 199901010000 gendeps/.dir
DAASE=/home/wspage/sage-4.1.1/spkg/build/fricas-1.0.7.p0/build-dir/target/i686-pc-linux
/home/wspage/sage-4.1.1/spkg/build/fricas-1.0.7.p0/build-dir/target/i686-pc-linux/bin/AXIOMsys
< mkinit.lsp > tmp/mkinit.log
test -f initlist.lst
sed -e 's/^/initaxiom /' initlist.lst > libaxiom.lst
for i in lang; do echo "lang $i" >> libaxiom.lst; done
for i in minimach boolean0 attrib; do echo "base $i" >>
libaxiom.lst; done
for i in subsetc; do echo "base $i" >> libaxiom.lst; done
for i in ; do echo "axiom $i" >> libaxiom.lst; done
for i in axextend axlit; do echo "aldorext $i" >> libaxiom.lst; done
aldor -mno-abbrev -mno-mactext -laldor -fx -dTRACE
../../../src/src/aldor/cliques.as
/home/wspage/aldor-src/aldor/install/aldor/linux/1.1.0/lib/libfoam.a(bigint.o):
In function `bintRadixScanFrString':
/home/wspage/aldor-src/aldor/aldor/lib/libfoam/links/bigint.c:567:
undefined reference to `log'
collect2: ld returned 1 exit status
#1 (Fatal Error) Linker failed. Command was: gcc cliques.o
aldormain.o -L.
-L/home/wspage/aldor-src/aldor/install/aldor/linux/1.1.0/share/lib
-L/home/wspage/aldor-src/aldor/install/aldor/linux/1.1.0/lib -o
cliques -laldor -lfoam
#1 (Warning) Removing file `cliques.o'.
#2 (Warning) Removing file `aldormain.o'.
make[2]: *** [cliques] Error 1
make[2]: Leaving directory
`/home/wspage/sage-4.1.1/spkg/build/fricas-1.0.7.p0/build-dir/src/aldor'
make[1]: *** [all-aldor] Error 2
make[1]: Leaving directory
`/home/wspage/sage-4.1.1/spkg/build/fricas-1.0.7.p0/build-dir/src'
make: *** [all-src] Error 2
***********************************************************
Failed to make FriCAS.
***********************************************************
real 11m24.582s
user 10m15.414s
sys 0m50.943s
sage: An error occurred while installing fricas-1.0.7.p0
Please email sage-devel http://groups.google.com/group/sage-devel
explaining the problem and send the relevant part of
of /home/wspage/sage-4.1.1/install.log. Describe your computer,
operating system, etc.
If you want to try to fix the problem, yourself *don't* just cd to
/home/wspage/sage-4.1.1/spkg/build/fricas-1.0.7.p0 and type 'make'.
Instead type "/home/wspage/sage-4.1.1/sage -sh"
in order to set all environment variables correctly, then cd to
/home/wspage/sage-4.1.1/spkg/build/fricas-1.0.7.p0
(When you are done debugging, you can type "exit" to leave the
subshell.)
wsp...@debian:~/sage-4.1.1$
--
In fact I have not been able to build the Aldor interface on this
system with this version of FriCAS for the last several revisions.
Unfortunately I have not had time to try to figure out why it is
failing.
Maybe for now the fricas-1.0.7.p0.spkg script should explicitly disable aldor?
>>> It might make sense to expand FriCASElement a little. Currently we have:
>>>
>>> P = self._check_valid()
>>> name = str(self)
>>> if name == 'Integer':
>>> from sage.rings.all import ZZ
>>> return ZZ
>>> elif name == 'DoubleFloat':
>>> from sage.rings.all import RDF
>>> return RDF
>>> elif name.startswith('Fraction('):
>>> name = name.lstrip('Fraction(')
>>> name = name.rstrip(')')
>>> return P(name)._sage_domain().fraction_field()
>>>
>>> raise NotImplementedError
>>>
>>> Types that often occur in FriCAS, might have an equivalent in sage, and
>>> come to my mind right now are:
>>>
>>> Polynomial(YourFavoriteRing) [polynomials with arbitrary many
>>> variables]
>>> Expression(Integer) [probably SymbolicRing]
>>> PrimeField(yourFavouritePrime) [probably
>>> Integers(yourFavouritePrime)]
>>> IntegerMod(yourFavouriteNaturalNumber) [probably
>>> Integers(yourFavouriteNaturalNumber)]
>>> AlgebraicNumber [probably QQbar]
>>> Float [RR with given precision]
>>> Complex(YourFavoriteCommutativeRing) [no idea]
>
> ...
>
>> Yes, I think that would be an excellent idea.
>
> Well, I just realized that the axiom interface was done differently:
> you used INFORM from axiom there. That worked reasonably well,
> as far as I remember.
I think it originally parsed the Type: part of the output. But that
was inadequate because at least in the original Axiom this was
sometimes truncated. At some point I changed the axiom.py interface
to use unparse(x::INFORM) that why it does now. But the output of
unparse has also changed recently in FriCAS and OpenAxiom to include
parenthesis and other minor enhancements. I think the axiom.py and
fricas.py interfaces do not quite properly take account of this. As I
said above the interpretation of the unparse string in the fricas.py
interface is not very robust and assumes that these parenthesis are
present.
>
> And now I also remember that some time ago I thought that the best
> way to have a (fricas) package SageForm, that converts InputForm
> into something appropriate for sage.
That might be nice. For sure this has to be standardized and
maintained in either in the FriCAS project or in Sage.
Personally, I was in favor of passing the SExpression version of
INFORM to Sage and then using some kind of lisp-like parser in the
axiom.py interface. In principle this would have a slightly lower
parsing overhead.
>
> Doing the conversion separately for the type (as above) and the
> element itself seems a bit like doing things twice, doesn't it?
>
Well given that things are passed as strings, parsing and
re-interpretation is unavoidable.
Regards,
Bill Page.
--~--~---------~--~----~------------~-------~--~----~
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to
sage-devel-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URLs: http://www.sagemath.org
-~----------~----~----~----~------~----~------~--~---
