PERL Beginners - Help: Math::Pari won't install

Author

Help: Math::Pari won't install

Richard Fernandez

2006-02-22, 6:56 pm

I'm trying to build Math::Pari along the way to building Net::SFTP, but
the make fails with the following error:

pariinl.h: In function `mulssmod':
pariinl.h:887: error: asm-specifier for variable `hiremainder' conflicts
with asm clobber list
pariinl.h:887:

by earlier errors, bailing out
make[1]: *** [es.o] Error 1
make[1]: Leaving directory
`/s01/.cpan588/build/Math-Pari-2.010703/libPARI'
make: *** [libPARI/libPARI.a] Error 2

Not sure what other information is relevant, but I'm doing this on a
Solaris 8 box using:

# gcc -v
Reading specs from
/usr/local/lib/gcc-lib/sparc-sun-solaris2.8/3.3.2/specs
Configured with: ../configure --with-as=3D/usr/ccs/bin/as
--with-ld=3D/usr/ccs/bin/ld --disable-nls
Thread model: posix
gcc version 3.3.2

Any help would be appreciated!

richf

usenet@DavidFilmer.com

2006-02-22, 6:56 pm

Richard Fernandez wrote:
> I'm trying to build Math::Pari ... on a Solaris 8 box

I don't know if this will help, but I had a confounded time building
Pari on AIX 5.3.3. After hours of Googling, I came across a post which
suggested:

./Configure --arch=AIX

Which worked perfectly (without any other switches/flags/settings).
Prehaps if you likewise stipulate your architecture you will have
better results. I posted my experiences building up this AIX machine in
comp.unix.aix, which you may view here: http://tinyurl.com/zvlxq. This
is, of course, AIX-specific, but some of what I learned may cross
platforms.

--
http://DavidFilmer.com

Wiggins d'Anconia

2006-02-28, 6:57 pm

RICHARD FERNANDEZ wrote:
> I'm trying to build Math::Pari along the way to building Net::SFTP, but
> the make fails with the following error:
>
> pariinl.h: In function `mulssmod':
> pariinl.h:887: error: asm-specifier for variable `hiremainder' conflicts
> with asm clobber list
> pariinl.h:887:

by earlier errors, bailing out
> make[1]: *** [es.o] Error 1
> make[1]: Leaving directory
> `/s01/.cpan588/build/Math-Pari-2.010703/libPARI'
> make: *** [libPARI/libPARI.a] Error 2
>
> Not sure what other information is relevant, but I'm doing this on a
> Solaris 8 box using:
>
> # gcc -v
> Reading specs from
> /usr/local/lib/gcc-lib/sparc-sun-solaris2.8/3.3.2/specs
> Configured with: ../configure --with-as=/usr/ccs/bin/as
> --with-ld=/usr/ccs/bin/ld --disable-nls
> Thread model: posix
> gcc version 3.3.2
>
> Any help would be appreciated!
>
> richf
>
>
>

This has lingered long enough without response. Sorry it took me a while
to get back to it after initially seeing it. I had this same problem a
couple of years ago. It does appear that Math::PARI has an active
maintainer again so you might try dropping him a line about this issue.

I solved it by testing each previous version (of libPARI) until I hit
one that would install cleanly. As much as it stinks to not have the
most recent version of something, a working version is better than no
version.

HTH,

http://danconia.org

Richard Fernandez

2006-02-28, 6:57 pm

This has lingered long enough without response. Sorry it took me a while
to get back to it after initially seeing it. I had this same problem a
couple of years ago. It does appear that Math::PARI has an active
maintainer again so you might try dropping him a line about this issue.

I solved it by testing each previous version (of libPARI) until I hit
one that would install cleanly. As much as it stinks to not have the
most recent version of something, a working version is better than no
version.

HTH,

http://danconia.org

Thanks for the response. I did manage to solve the Math::Pari install
problem by following Jeff Eggen's idea about configuring it with
machine=3Dnone. I have no idea about whether or not, as he mentions, =
it's
unusably slow because I proceeded with the rest of the Net::SFTP install
and got similarly hung up on the install for Math::GMP. (Un)fortunately
enough time has elapsed on this that the Powers That Be decided to
shelve this project and go a different way. So for now I'm off the hook.

Ironically, we have Net::SFTP running fine in other environments. It's
just this one box...

Thanks again for the help!

richf

Wiggins d'Anconia

2006-02-28, 6:57 pm

RICHARD FERNANDEZ wrote:

>
> Thanks for the response. I did manage to solve the Math::Pari install
> problem by following Jeff Eggen's idea about configuring it with
> machine=none. I have no idea about whether or not, as he mentions, it's
> unusably slow because I proceeded with the rest of the Net::SFTP install
> and got similarly hung up on the install for Math::GMP. (Un)fortunately
> enough time has elapsed on this that the Powers That Be decided to
> shelve this project and go a different way. So for now I'm off the hook.
>

Ah, hmm... I had the same problem with Math::GMP, though determined that
as long as you only want Protocol 2 it isn't really needed. So doing a
force install on Math::GMP will get it past the testing on Net::SFTP.
Sorry I couldn't be of much help sooner.

> Ironically, we have Net::SFTP running fine in other environments. It's
> just this one box...
>
> Thanks again for the help!
>

No worries.

> richf
>

http://danconia.org

Sponsored Links