elfpen has asked for the wisdom of the Perl Monks concerning the following question:

Greetings and humble thanksgiving in advance of your excellent advice.
Every few years I must expiate for my sin of invoking Math::BigNum in an important script by wrestling again with the install of Math::Peri.
This time around I am on an HP-UX v11.31 Itanium box. And all seems to be so close to right. But one test.

t/55_intnum.t 

# eval: sumnum($n,1,[2,log(2)], sub{2**(-$n)},intnumstep()+1,1)-1

not ok 46 # in='sumnum($n,1,[2,log(2)], sub{2**(-$n)},intnumstep()+1,1
+)-1', err='PARI:   ***   the PARI stack overflows !
        current stack size: 20000000 (19.073 Mbytes)
  [hint] you can increase GP stack with allocatemem()
'

[download]

I don't know if this is a terrible thing or an ignorable thing or somewhere in between.
In the end, Math-Pari-2.010808 declares FAIL and am dejected and downcast.

Replies are listed 'Best First'.
Re: Math::Peri test: FAIL but hari-kari not an option
by syphilis (Archbishop) on Feb 18, 2016 at 23:18 UTC
    I don't know if this is a terrible thing or an ignorable thing or somewhere in between

    I would 'make install' it and see how it goes.
    If the Pari stack keeps overflowing for you, then you'll need to take some action - either increase the allowable size with allocatemem (as suggested) or take some steps to find out more about what is going wrong.

    For the latter, filing a bug report is probably a good start as the Math::Pari author is still quite responsive to such bug reports (I think).

    Is 20 MB a reasonable maximum setting for the Pari stack ?

    Cheers,
    Rob
[reply]

Back to Seekers of Perl Wisdom