FriCAS X.Y.Z

[maxthemouse]

This is an update to fricas to use ecl and to update to the recent fricas release.

Adam

CC: @sagetrac-bpage @mwhansen @hemmecke

Component: packages: optional

Author: Ralf Hemmecke

Reviewer: Jaap Spies, Adam Webb, Mike Hansen

Merged: sage-4.3.alpha0

Issue created by migration from https://trac.sagemath.org/ticket/6517

[sagetrac-mvngu]

comment:1

This might depends on or related to ticket #6318.

[sagetrac-mvngu]

Author: Adam Webb

[maxthemouse]

comment:3

Ticket #6318 is related. The patches on that ticket were tested with this package.

[jaapspies]

comment:4

79 FriCAS installation finished.

real	46m21.303s
user	34m38.652s
sys	4m30.484s
Successfully installed fricas-1.0.7.p0

I tested the patches from #6318 related to fricas.

./sage -t --long --optional devel/sage/sage/interface/fricas.py

alls tests passed.

Positive review.

Jaap

[jaapspies]

Reviewer: jsp

[sagetrac-mvngu]

Changed reviewer from jsp to Jaap Spies

[sagetrac-mvngu]

comment:5

Replying to @maxthemouse:

Ticket #6318 is related. The patches on that ticket were tested with this package.

So what is the order of merging patches here? Is it

1. merge the patches at optional doctest failure -- axiom interface -- something doesn't work #6318;
2. then merge the spkg on this ticket in the optional spkg repository?
   In that case, optional doctest failure -- axiom interface -- something doesn't work #6318 must get positive review first. On the other hand, is this ticket and its spkg not at all dependent on optional doctest failure -- axiom interface -- something doesn't work #6318?

[jaapspies]

comment:6

Replying to @sagetrac-mvngu:

Replying to @maxthemouse:

Ticket #6318 is related. The patches on that ticket were tested with this package.

So what is the order of merging patches here? Is it

1. merge the patches at optional doctest failure -- axiom interface -- something doesn't work #6318;
2. then merge the spkg on this ticket in the optional spkg repository?
   In that case, optional doctest failure -- axiom interface -- something doesn't work #6318 must get positive review first. On the other hand, is this ticket and its spkg not at all dependent on optional doctest failure -- axiom interface -- something doesn't work #6318?

I think it is independent of #6318 in the sense that it updates the fricas optional spkg.

And than it fixes the fricas.py failures mentioned in #6318, Not the axiom issues.

Jaap

[sagetrac-mvngu]

comment:7

I've uploaded Adam's FriCAS 1.0.7 spkg to

http://sage.math.washington.edu/home/mvngu/patch/fricas-1.0.7.p0.spkg

I install it as in the following transcript, but received "configure: error: Unable to determine Lisp flavor":

[mvngu@sage sage-4.1.1.alpha0]$ ./sage -f /home/mvngu/patch/fricas-1.0.7.p0.spkg
<insall-messages>
axiom_build_bindir = /scratch/mvngu/release/sage-4.1.1.alpha0/spkg/build/fricas-1.0.7.p0/build-dir/build/x86_64-unknown-linux/bin
checking Lisp implementation... ECL (Embeddable Common-Lisp) 9.4.1
Copyright (C) 1984 Taiichi Yuasa and Masami Hagiya
Copyright (C) 1993 Giuseppe Attardi
Copyright (C) 2000 Juan J. Garcia-Ripoll
ECL is free software, and you are welcome to redistribute it
under certain conditions; see file 'Copyright' for details.
Type :h for Help.  Top level.
> 
;;; Loading "/scratch/mvngu/release/sage-4.1.1.alpha0/spkg/build/fricas-1.0.7.p0/src/config.lisp"
Filesystem error with pathname #P"SYS:cmp.NEWEST".
Either
 1) the file does not exist, or
 2) we are not allow to access the file, or
 3) the pathname points to a broken symbolic link.
Broken at SI:BYTECODES.No restarts available.
Broken at SI:BYTECODES. File: #P"/scratch/mvngu/release/sage-4.1.1.alpha0/spkg/build/fricas-1.0.7.p0/src/config.lisp" (Form #1)
>> 
configure: error: Unable to determine Lisp flavor
***********************************************************
Failed to configure FriCAS.
***********************************************************

real	0m0.577s
user	0m0.250s
sys	0m0.340s
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 /scratch/mvngu/release/sage-4.1.1.alpha0/install.log.  Describe your computer, operating system, etc.
If you want to try to fix the problem, yourself *don't* just cd to
/scratch/mvngu/release/sage-4.1.1.alpha0/spkg/build/fricas-1.0.7.p0 and type 'make'.
Instead type "/scratch/mvngu/release/sage-4.1.1.alpha0/sage -sh"
in order to set all environment variables correctly, then cd to
/scratch/mvngu/release/sage-4.1.1.alpha0/spkg/build/fricas-1.0.7.p0
(When you are done debugging, you can type "exit" to leave the
subshell.)

Am I doing something wrong or missing something?

[maxthemouse]

comment:9

This package depends on #6318 for the fricas interface to work therefore that patch needs to go in first.

adam

[maxthemouse]

comment:10

In previous versions there was an install option for the lisp flavour but according to the fricas mailing list, this is depreciated and fricas is supposed to detect which lisp is used. I have not yet tested with 4.1.1.alpha0 but it worked for me on 4.0 and 4.1.

[maxthemouse]

comment:11

update: Tested package with 4.1.1.alpha0 and it worked. This could be sensitive to the lisp install but I am not an expert on this. The long path name might also be a problem but this only speculation.

[burcin]

comment:12

Could someone comment on the relation between the two pairs of issues:

• this one, optional doctest failure -- axiom interface -- something doesn't work #6318
• fricas-1.0.5 update #4461, minor improvements to the Axiom interface #4036 ?

Is the package here built on the one at #4461? Are the changes on #4036 independent of the ones at #6318?

Thanks.

Burcin

[maxthemouse]

comment:13

Sorry,
I think there is a bit of cross-posting of tickets. My bad. I did not see the the earlier ones (#4461 and #4036) and with the new versions of fricas it is not clear to me which is now needed. I agree that the axiom and fricas should be separate interfaces although at the moment the fricas interface is a basically a subclass of the axiom one. They are still very similar but it means that to test axiom.py one needs both axiom and fricas installed.

Adam

[maxthemouse]

comment:14

Just a comment.
#4461 deals with an issue with clisp which is no longer used. It seems that it could just be closed now.
#4036 are independent of #6318. (I am not sure what they do.)
#6318 fix some interface issues but perhaps should be added after #4036?

As a guess: 1. apply #4036 to a new version (and rebase), 2. apply #6318 and test, 3. check/make a new fricas spkg.

Adam

PS I will not have computer access for at least a week but can look at it after that if no one else has.

[maxthemouse]

comment:16

Replying to @burcin:

Could someone comment on the relation between the two pairs of issues:

• this one, optional doctest failure -- axiom interface -- something doesn't work #6318
• fricas-1.0.5 update #4461, minor improvements to the Axiom interface #4036 ?

Is the package here built on the one at #4461? Are the changes on #4036 independent of the ones at #6318?

Thanks.

Burcin

The patches at #4036 are already in and I only added a few fricas tests. I would apply #6318 first, then #4036 (trac_4036-axiom_interface.patch only) and of course axiom and fricas need to be installed to test everything.

Adam

[hemmecke]

I've created an spkg builder at http://boxen.math.washington.edu:29792 .
This code is not an spkg itself but rather builds two spkg: fricas-VERSION.spkg and fricasaldor-VERSION.spkg. It requires a checkout of the FriCAS sources from SVN at sourceforge.

On sage.math the following should work.

WARNING, before you start compiling FriCAS, you should read the top of the Makefile, in particular the part connected to an X server. It's unwise to have the X server running on a different machine from where you compile the spkg!!! The X server is not needed for actually installing the .spkg into Sage.

Get the FriCAS sources:

svn export https://fricas.svn.sourceforge.net/svnroot/fricas/trunk fricas-sources

Use the installed Aldor from my HOME (or install aldor from http://www.aldor.org).

export ALDORROOT=/home/hemmecke/aldorroot
export PATH=$ALDORROOT/bin:$PATH

Make sure you have ecl 9.8.4.

export PATH=$SAGE_ROOT/local/bin:$PATH
export LD_LIBRARY_PATH=$SAGE_ROOT/local/lib:$LD_LIBRARY_PATH
hg clone http://boxen.math.washington.edu:29792 fricas-aldor-spkg
cd fricas-aldor-spkg
make

Installing FriCAS inside Sage

cp fricas-aldor-spkg/spkg/fricas*.spkg  $SAGE_ROOT/spkg/optional
sage -i fricasaldor-VERSION

The VERSION of the spkg will automatically be extracted from FriCAS configure script.
VERSION is appropriate if you use a proper release version, i.e. (should work for a release > 1.0.7)

svn export https://fricas.svn.sourceforge.net/svnroot/fricas/releases/1.0.8 fricas-sources

and running the fricas-aldor-spkg/Makefile will produce VERSION=1.0.8.

Ralf

[hemmecke]

comment:26

Replying to @maxthemouse:

http://sage.math.washington.edu/home/awebb/fricas-1.0.8.spkg. It installs and I can start the console.

If it works, someone should upload it to http://sagemath.org/packages/optional/ .
What is the workflow to get the package into the "optional" archive?

[maxthemouse]

comment:27

I think it just needs to be reviewed. If it works then it should be set to 'positive review' and then the current release manager looks at it. If they are happy than they put it in the optional packages. The package works for me and I ran a quick test from inside sage so I would give it a positive review. If the package also works for you than I guess the status can be changed. BTW I used sage 4.2.1.alpha0 for building the package and testing it. This includes ecl 9.10.2.

Adam

[williamstein]

comment:28

Replying to @maxthemouse:

I think it just needs to be reviewed. If it works then it should be set to 'positive review' and then the current release manager looks at it. If they are happy than they put it in the optional packages. The package works for me and I ran a quick test from inside sage so I would give it a positive review. If the package also works for you than I guess the status can be changed. BTW I used sage 4.2.1.alpha0 for building the package and testing it. This includes ecl 9.10.2.

Adam

Exactly right.

[sagetrac-drkirkby]

comment:29

I just tried it on Solaris, and it fails very quickly:

fricas-1.0.8/src/src/ChangeLog.bu
fricas-1.0.8/src/confignotes.tex
Finished extraction
****************************************************
Host system
uname -a:
SunOS swan 5.10 Generic_139555-08 sun4u sparc SUNW,Sun-Blade-1000
****************************************************
****************************************************
CC Version
/usr/local/gcc-4.4.1-sun-linker/bin/gcc -v
Using built-in specs.
Target: sparc-sun-solaris2.10
Configured with: ../gcc-4.4.1/configure --prefix=/usr/local/gcc-4.4.1-sun-linker/ --with-as=/usr/ccs/bin/as --without-gnu-as --with-ld=/usr/ccs/bin/ld --without-gnu-ld --enable-languages=c,c++,fortran --with-mpfr-include=/usr/local/include --with-mpfr-lib=/usr/local/lib --with-gmp-include=/usr/local/include --with-gmp-lib=/usr/local/lib CC=/usr/sfw/bin/gcc CXX=/usr/sfw/bin/g++ LDFLAGS='-R /usr/local/lib -L /usr/local/lib'
Thread model: posix
gcc version 4.4.1 (GCC) 
****************************************************
checking build system type... sparc-sun-solaris2.10
checking host system type... sparc-sun-solaris2.10
checking target system type... sparc-sun-solaris2.10
checking for gmake... make
GNU Make 3.80
checking for gcc... /usr/local/gcc-4.4.1-sun-linker/bin/gcc
checking for C compiler default output file name... a.out
checking whether the C compiler works... yes
checking whether we are cross compiling... no
checking for suffix of executables... 
checking for suffix of object files... o
checking whether we are using the GNU C compiler... yes
checking whether /usr/local/gcc-4.4.1-sun-linker/bin/gcc accepts -g... yes
checking for /usr/local/gcc-4.4.1-sun-linker/bin/gcc option to accept ISO C89... none needed
checking for a BSD-compatible install... /usr/local/bin/ginstall -c
checking for touch... touch
checking for mktemp... mktemp
checking for gawk... no
checking for mawk... no
checking for nawk... nawk
checking for gtar... gtar
checking for ranlib... ranlib
checking for ar... ar
checking for latex... no
checking for makeindex... no
checking for bash... /usr/bin/bash
checking for notangle... no
checking for noweave... no
axiom_build_bindir = /export/home/drkirkby/sage-4.2/spkg/build/fricas-1.0.8/build-dir/build/sparc-sun-solaris2.10/bin
checking Lisp implementation... ../src/configure: line 3955: ecl: command not found
Can't open config_cl.out
configure: error: Unable to determine Lisp flavor
***********************************************************
Failed to configure FriCAS.
***********************************************************

real    0m2.871s
user    0m0.924s
sys     0m1.561s
sage: An error occurred while installing fricas-1.0.8

[sagetrac-drkirkby]

comment:30

Oops, ignore that comment. I realised I downloaded the package into an incomplete build of Sage.

Dave

[hemmecke]

comment:31

Replying to @sagetrac-drkirkby:

I just tried it on Solaris, and it fails very quickly:

checking Lisp implementation... ../src/configure: line 3955: ecl: command not found

The error message says it all. And http://boxen.math.washington.edu:29792/file/25ab4f31480e/SPKG.txt.fricas#l1 also lists ecl as a dependency.

If there were a proper package management in Sage, I would have made ecl a formal dependency, but adding "sage -i ecl-9.8.4" to the install script is also not fool proof, since specifying a version number is bad. Giving no version number is bad as well. I'd rather wish to be able to specify a range of versions that should work for the fricas-1.0.8.
Try

sage -i http://sage.math.washington.edu/home/hemmecke/pub/ecl-9.8.4.spkg

[mwhansen]

Changed reviewer from Jaap Spies, Adam Webb to Jaap Spies, Adam Webb, Mike Hansen

[mwhansen]

comment:32

I've merged this into the optional packages.

[mwhansen]

Author: Adam Webb, Ralf Hemmecke

[sagetrac-mvngu]

Merged: sage-4.3.alpha0

[hemmecke]

comment:34

FriCAS 1.0.9 and FriCAS-Aldor is available from

http://sage.math.washington.edu/home/hemmecke/pub/fricas-1.0.9.spkg

http://sage.math.washington.edu/home/hemmecke/pub/fricasaldor-1.0.9.spkg

[maxthemouse]

comment:37

Thanks for the packages. I guess it is mentioned somewhere that installing fricasaldor depends on the source of fricas. This means that the following is needed:

sage -i -s fricas-1.0.9.spkg
sage -i fricasaldor-1.0.9.spkg

I had a problem with an example for fricasaldor.

(1) -> )co fib.as
   Compiling FriCAS source code from file /home/math/test/fib.as using
      AXIOM-XL compiler and options
-O -Fasy -Fao -Flsp -laxiom -Mno-ALDOR_W_WillObsolete -DAxiom -Y $AXIOM/algebra -I $AXIOM/algebra
      Use the system command )set compiler args to change these
      options.
   Compiling Lisp source code from file fib.lsp
   Issuing )library command for fib
   Reading /home/math/test/fib.asy
(1) -> fib 5

   >> System error:
   Error when using UNREAD-CHAR on stream #<string-input stream from "3(1) ->

where fib.as is

#include "axiom"
fib(n: Integer): Integer == {
        n < 0 => 0;
	n < 3 => 1;
	fib(n-1) + fib(n-2);
}

I get get this with ecl but not with sbcl but I think this is a known problem. Another example with a simple factorial function worked. I did some simple things and the fricas package seemed to work fine.

Adam

[maxthemouse]

comment:38

There seems to be an API change. A few tests fail in the Sage interface. This is now ticket #9258. It appears to be a change in the return string. Note that these tests still work for axiom.

Adam

[hemmecke]

comment:39

Replying to @maxthemouse:

Thanks for the packages. I guess it is mentioned somewhere that installing fricasaldor depends on the source of fricas. This means that the following is needed:

sage -i -s fricas-1.0.9.spkg
sage -i fricasaldor-1.0.9.spkg

That is not so! As explained above, the commands are simply:

cp fricas-aldor-spkg/spkg/fricas*.spkg  $SAGE_ROOT/spkg/optional
sage -i fricasaldor-X.Y.Z

Dependency is included in the package fricasaldor.

I had a problem with an example for fricasaldor.

(1) -> )co fib.as
   Compiling FriCAS source code from file /home/math/test/fib.as using
      AXIOM-XL compiler and options
-O -Fasy -Fao -Flsp -laxiom -Mno-ALDOR_W_WillObsolete -DAxiom -Y $AXIOM/algebra -I $AXIOM/algebra
      Use the system command )set compiler args to change these
      options.
   Compiling Lisp source code from file fib.lsp
   Issuing )library command for fib
   Reading /home/math/test/fib.asy
(1) -> fib 5

   >> System error:
   Error when using UNREAD-CHAR on stream #<string-input stream from "3(1) ->

I get get this with ecl but not with sbcl but I think this is a known problem.

Yes that seems to be an ecl problem and I have no idea how to track that. So unfortunately, I cannot fix it. I think the link to Aldor is not as stable as it should be. There might also be a problem on 64bit machines with respect to internal hash codes. But since the Aldor sources are not free, it's up to a person who agrees with the Aldor Public License 2.0 to sync with FriCAS.

Ralf

[maxthemouse]

comment:40

Replying to @hemmecke:

That is not so! As explained above, the commands are simply:

cp fricas-aldor-spkg/spkg/fricas*.spkg  $SAGE_ROOT/spkg/optional
sage -i fricasaldor-X.Y.Z

Dependency is included in the package fricasaldor.

I got the following when installing fricasaldor. The message says that the build directory is missing so I assumed that I had made a mistake.

****************************************************
Installing -s fricas-1.0.9
Calling sage-spkg on fricas-1.0.9
Warning: Attempted to overwrite SAGE_ROOT environment variable
fricas-1.0.9
Machine:
Linux haso220w 2.6.18-92.1.6.el5 #1 SMP Wed Jun 25 12:46:39 EDT 2008 i686 i686 i386 GNU/Linux
sage: /home/math/sage/spkg/build/fricasaldor-1.0.9/fricas-1.0.9 is already installed
grep: /home/math/sage/spkg/build/fricas-1.0.9/spkg-install: No such file or directory
cp: cannot create regular file `/home/math/sage/spkg/build/fricas-1.0.9/build-dir/src/aldor/': No such file or directory
mkdir: cannot create directory `/home/math/sage/spkg/build/fricas-1.0.9/build-dir/src/aldor/al': No such file or directory
cp: cannot create regular file `/home/math/sage/spkg/build/fricas-1.0.9/build-dir/src/aldor/al': No such file or directory
./spkg-install: line 60: cd: /home/math/sage/spkg/build/fricas-1.0.9/build-dir: No such file or directory
./spkg-install: line 61: /home/math/sage/spkg/build/fricas-1.0.9/src/configure: No such file or directory
***********************************************************
Failed to configure FriCAS-Aldor interface.
***********************************************************

Adam

[maxthemouse]

comment:41

Replying to @hemmecke:

Yes that seems to be an ecl problem and I have no idea how to track that. So unfortunately, I cannot fix it. I think the link to Aldor is not as stable as it should be. There might also be a problem on 64bit machines with respect to internal hash codes. But since the Aldor sources are not free, it's up to a person who agrees with the Aldor Public License 2.0 to sync with FriCAS.

Ralf

I agree

Adam

[hemmecke]

comment:42

Dependency is included in the package fricasaldor.

I got the following when installing fricasaldor. The message says that the build directory is missing so I assumed that I had made a mistake.

Yes. The whole problem is due to the Aldor Public License 2. :-( That's the only reason why there are two packages. The fricas spkg is mBSD and the fricasaldor spkg is APL2.
The other problem is that Sage does not provide a proper mechanism to deal with dependencies. (People should look at "zc.buildout".)

It's all not so easy since currently fricasaldor needs the build directory of the build of the fricas package. It basically copies copies some files into this directory and then starts the build.

So saying

sage -i fricas-X.Y.Z
sage -i fricasaldor-X.Y.Z

will not work, because the build directory of fricas is gone. I guess that might be the problem with your error.

The other thing is

****************************************************
Installing -s fricas-1.0.9
Calling sage-spkg on fricas-1.0.9
Warning: Attempted to overwrite SAGE_ROOT environment variable

Do you have an idea who is printing that warning?

[maxthemouse]

comment:43

Replying to @hemmecke:

It's all not so easy since currently fricasaldor needs the build directory of the build of the fricas package. It basically copies copies some files into this directory and then starts the build.

So saying

sage -i fricas-X.Y.Z
sage -i fricasaldor-X.Y.Z

will not work, because the build directory of fricas is gone. I guess that might be the problem with your error.

Exactly,

sage -i -s fricas-X.Y.Z
sage -i fricasaldor-X.Y.Z

This worked for me since the -s keeps the build directory in the first step.

The other thing is

****************************************************
Installing -s fricas-1.0.9
Calling sage-spkg on fricas-1.0.9
Warning: Attempted to overwrite SAGE_ROOT environment variable

Do you have an idea who is printing that warning?

I think I see it everytime I use sage -i to install something. I assume that it is part of the Sage installing process somewhere. I don't know why the SAGE_ROOT would need to be changed though.

[maxthemouse]

Changed author from Adam Webb, Ralf Hemmecke to Ralf Hemmecke

[maxthemouse]

comment:44

I am not sure of the procedure of who should give the positive review and in any case several people should look at it. As I have not been very active lately I will just give a comment. This seems to be a simple upgrade of two optional packages and so I don't have any major concerns. I installed the packages on two Linux systems (32 bit SL 5.1, 64 bit Ubuntu 10.04) and ran some tests without severe problems (small issues noted above). I am aware that there are problems with the Sage interface but I think that is a separate issue. Therefore, I would give the packages a positive review.

Adam

[mwhansen]

comment:45

These look good to me. I've made a new ticket at #9354 -- please don't reopen this ticket once it has been closed. I'll post a fix to the issues at #9258.