Re: Bug testing 'nfields'

[Bill Allombert <allomber@math.u-bordeaux.fr>]

On Thu, Mar 09, 2006 at 12:18:39PM -0800, Justin C. Walker wrote:
> Hi, all,
> 
> I have built the following versions of Pari/GP, for Mac OS X, 10.4.5:
>   Dual G5/PowerPC: (1.1502)
>   Core Duo/ix86:   (1.1500)

Hello Justin,

Which gcc build are you using ?

> In both cases, I get a 'make bench' failure with 'nfields': checking  
> the full output, I see this:
> 
> G5:
> ? rnfinit(nf2,x^5-x-2)
>   *** rnfinit: bug in GP (Bus Error), please report
> ? nfisincl(x^2+1,x^4+1)
>   *** nfisincl: bug in GP (Bus Error), please report
> 
> Core Duo:
> ? rnfinit(nf2,x^5-x-2)
>   *** rnfinit: bug in GP (Bus Error), please report
> ? factornf(x^3+x^2-2*x-1,t^3+t^2-2*t-1)
>   *** factornf: bug in GP (Segmentation Fault), please report
> ? nfisincl(x^2+1,x^4+1)
>   *** nfisincl: bug in GP (Segmentation Fault), please report
> ? nfisisom(x^3+x^2-2*x-1,x^3+x^2-2*x-1)
>   *** nfisisom: bug in GP (Segmentation Fault), please report
> 
> The two Core Duo seg faults that don't appear in the G5 case show up  
> in the latter as complaints about a shortage of primes.
> 
> Are these known problems?
> 
> I ran the same check with 1.1277; it finished without any  
> complaints.  Well, just one, which is repeated in all three cases:
> ? setrand(1);bnfclassunit(x^2-x-100000)
>   *** bnfclassunit: Warning: insufficient precision for fundamental  
> units, not given.

I don't know if it is related we received bug #383 about problems
on Mac OS X that we could not reproduce:

? fordiv(4,X,print(X))
   *** fordiv: bug in GP (Segmentation Fault), please report

Can you reproduce that as well ?

Cheers,
Bill.