Compilation error for alpha0

[Nathann Cohen]

Hello everybody !!!

I just noticed the compilation of alpha0 on my computer failed... And
I am afraid I have no idea how to inquire where it comes from... I can
give you the last few lines of the install log, and you but have to
ask me which other information you may need ( and tell me how to
obtain it, just in case ) :-)

I am running a Fedora distribution.

/////////////////////////////////////////////////////////
gcc -pthread -shared
build/temp.linux-i686-2.6/sage/rings/polynomial/polynomial_real_mpfr_dense.o
-L/usr/local/sage/local//lib -lcsage -lmpfr -lgmp -lstdc++ -lntl -o
build/lib.linux-i686-2.6/sage/rings/polynomial/polynomial_real_mpfr_dense.so
gcc -fno-strict-aliasing -DNDEBUG -g -fwrapv -O3 -Wall
-Wstrict-prototypes -fPIC
-I/usr/local/sage/local/lib/python/site-packages/numpy/core/include
-I/usr/local/sage/local//include
-I/usr/local/sage/local//include/csage
-I/usr/local/sage/devel//sage/sage/ext
-I/usr/local/sage/local/include/python2.6 -c
sage/rings/polynomial/real_roots.c -o
build/temp.linux-i686-2.6/sage/rings/polynomial/real_roots.o -w
gcc -pthread -shared
build/temp.linux-i686-2.6/sage/rings/polynomial/real_roots.o
-L/usr/local/sage/local//lib -lcsage -lmpfr -lstdc++ -lntl -o
build/lib.linux-i686-2.6/sage/rings/polynomial/real_roots.so
gcc -fno-strict-aliasing -DNDEBUG -g -fwrapv -O3 -Wall
-Wstrict-prototypes -fPIC -I/usr/local/sage/local//include
-I/usr/local/sage/local//include/csage
-I/usr/local/sage/devel//sage/sage/ext
-I/usr/local/sage/local/include/python2.6 -c
sage/rings/polynomial/symmetric_reduction.c -o
build/temp.linux-i686-2.6/sage/rings/polynomial/symmetric_reduction.o
-w
gcc -pthread -shared
build/temp.linux-i686-2.6/sage/rings/polynomial/polynomial_element.o
-L/usr/local/sage/local//lib -lcsage -lstdc++ -lntl -o
build/lib.linux-i686-2.6/sage/rings/polynomial/polynomial_element.so
gcc -fno-strict-aliasing -DNDEBUG -g -fwrapv -O3 -Wall
-Wstrict-prototypes -fPIC -I/usr/local/sage/local/include/FLINT/
-I/usr/local/sage/local//include
-I/usr/local/sage/local//include/csage
-I/usr/local/sage/devel//sage/sage/ext
-I/usr/local/sage/local/include/python2.6 -c
sage/schemes/elliptic_curves/descent_two_isogeny.c -o
build/temp.linux-i686-2.6/sage/schemes/elliptic_curves/descent_two_isogeny.o
-std=c99 -w
error: command 'gcc' failed with exit status 1
sage: There was an error installing modified sage library code.

ERROR installing SAGE

real 10m49.092s
user 15m46.387s
sys 1m4.440s
sage: An error occurred while installing sage-4.3.1.alpha0
Please email sage-devel http://groups.google.com/group/sage-devel
explaining the problem and send the relevant part of
of /usr/local/sage/install.log. Describe your computer, operating system, etc.
If you want to try to fix the problem yourself, *don't* just cd to
/usr/local/sage/spkg/build/sage-4.3.1.alpha0 and type 'make check' or
whatever is appropriate.
Instead, the following commands setup all environment variables
correctly and load a subshell for you to debug the error:
(cd '/usr/local/sage/spkg/build/sage-4.3.1.alpha0' &&
'/usr/local/sage/sage' -sh)
When you are done debugging, you can type "exit" to leave the
subshell.
make[1]: *** [installed/sage-4.3.1.alpha0] Error 1
make[1]: Leaving directory `/usr/local/sage/spkg'

real 113m54.424s
user 53m3.836s
sys 28m37.585s
Error building Sage.

/////////////////////////////////////////////////////////////

Nathann

[John Cremona]

It seems to be failing while compiling the cython code introduced by
#6583, so perhaps Robert Miller can help?

John

2010/1/4 Nathann Cohen <nathan...@gmail.com>:

> --
> To post to this group, send an email to sage-...@googlegroups.com
> To unsubscribe from this group, send an email to sage-devel+...@googlegroups.com
> For more options, visit this group at http://groups.google.com/group/sage-devel
> URL: http://www.sagemath.org
>

[Robert Miller]

It does look like "descent_two_isogeny" is what it is failing to
compile, but I really can't help without more information. Is there a
way to get gcc to print out why it's failing?

It occurs to me that this depends on ratpoints, which has some issues
compiling with earlier versions of gcc. What version are you using?

On Mon, Jan 4, 2010 at 3:41 AM, John Cremona <john.c...@gmail.com> wrote:
> It seems to be failing while compiling the cython code introduced by
> #6583, so perhaps Robert Miller can help?

> ...

>> gcc -fno-strict-aliasing -DNDEBUG -g -fwrapv -O3 -Wall
>> -Wstrict-prototypes -fPIC -I/usr/local/sage/local/include/FLINT/
>> -I/usr/local/sage/local//include
>> -I/usr/local/sage/local//include/csage
>> -I/usr/local/sage/devel//sage/sage/ext
>> -I/usr/local/sage/local/include/python2.6 -c
>> sage/schemes/elliptic_curves/descent_two_isogeny.c -o
>> build/temp.linux-i686-2.6/sage/schemes/elliptic_curves/descent_two_isogeny.o
>> -std=c99 -w
>> error: command 'gcc' failed with exit status 1

--
Robert L. Miller
http://www.rlmiller.org/

[Nathann Cohen]

Here is mine :

gcc version 4.3.2 20081105 (Red Hat 4.3.2-7) (GCC)

Nathann

[Dr. David Kirkby]

Robert Miller wrote:
> It does look like "descent_two_isogeny" is what it is failing to
> compile, but I really can't help without more information. Is there a
> way to get gcc to print out why it's failing?
>
> It occurs to me that this depends on ratpoints, which has some issues
> compiling with earlier versions of gcc. What version are you using?
>

GCC 4.2.1 on a SPARC is showing a slightly more helpful error message. See

http://trac.sagemath.org/sage_trac/ticket/7867

Dave

[Dr. David Kirkby]

This will come as no great surprise, since the setups are not very different
(same gcc version, though different versions of the SPARC processor), but this
failed on my own Sun Blade 2000 too. So it seems Fedora Linux and Solaris 10 on
SPARC are affected by this.

However, there is something a bit odd. In the post Nathann Cohen on his Fedora
system, we can see that the final compile line ends in a '-w' which is supposed
to suppress all warnings (much to my annoyance). The same '-w' is seen on 't2',
which has a SPARC processor.

For reasons unknown, the '-w' option is *not* included on my build on the Sun
Blade 2000, so the warnings are displayed.

Why an option to suppress warnings is seed on Fedora and one SPARC machine, but
not on the build of another SPARC machine is anyones guess!

Dave