Ged Haywood wrote:
>
> There's a file in the mod_perl directory called SUPPORT. (That bit
> about 'perl -V' was taken from there. :) SUPPORT contains detailed
> instructions about what to do when mod_perl crashes, including what
> information to provide and how to generate a stack backtrace.
>
> 73,
> Ged.
>
yeh, read that... I guess I shoudn't have used the word "crashes" as
it's really "dies" when it tries to do a numeric comparison on an
alphanumeric string, which is entirely reasonable, so I'm not getting a
core dump, and I don't think mod_perl itself is at fault. I suspect the
implementation of GP/Pari as I ended up with the "worst case" as
mentioned in the Math::Pari install and manually copied (as instructed)
what I guess is a 'C' library file of some description (paricfg.h - I am
by no stretch a C programmer..) to the proper location. It all seemed to
work ok after that and as I had had troubles prior I already had my test
scripts which all ran fine so I thought Hooray! Then I implemented the
real system and got the errors mentioned.
What I still don't understand is why the test scripts run ok, but the
mod_perl implementation doesn't. (But on my dev server everything works
as it should..)
I'm going to reinstall GP/Pari on the production server using the
src.rpm I found on the Pari site, then if that goes ok I might be able
to install/update Math::Pari without errors, then maybe it'll all work..
that's the plan anyway..
cheers,
--
J
Web Developer
Eight Degrees Off Centre
http://www.eightdegrees.com.au/