Thanks. The large S-unit generator is indeed produced by pari (but I'm not 
certain whether these are incorrect generators, or whether the initial bug 
comes from testing them for being S-units, and whether this also explains 
other inconsistencies between coecions between L and OLSstar and back in 
other examples): 

L = bnfinit(x^6 - 68463*x^4 - 5120808*x^3 + 1250774892*x^2 + 192368273328*x 
+ 7520491439712,1)

S2 = idealprimedec(L,2)
S2 = [S2[3],S2[2],S2[1]]
S7 = idealprimedec(L,7)
S13 = idealprimedec(L,13)
S13 = [S13[1],S13[3],S13[2]]
S5 = idealprimedec(L,5)
S = concat([S2,S7,S13,S5]) #The order matters, this one is compatible with 
sage's S = L.primes_above(2*5*7*13)

sfu = bnfsunit(L,S)

In GP 2.11.2 (via Sage 9.0), the last two generators in sfu are indeed very 
large.
In GP 2.9.4, it runs for a minute before raising a stack overflow.


Running the same example with a different order of input primes raises 
another bug (perhaps they are related):

S2 = idealprimedec(L,2)
S2 = [S2[3],S2[2],S2[1]]
S5 = idealprimedec(L,5)
S7 = idealprimedec(L,7)
S13 = idealprimedec(L,13)
S13 = [S13[1],S13[3],S13[2]]
S = concat([S2,S5,S7,S13]) #same order as what L.primes_above(2),...,L.
primes_above(13) yields in sage.

sfu = bnfsunit(L,S)

In GP 2.11.2 (via Sage 9.0) this immediately raises an error:
  *** bnfsunit: impossible inverse in ZM_inv: [60, 40, 2, 0, 0, 0, 0, 54; 
0, 20, 2, 0, 0, 0, 0, 0; 0, 0, 2, 0, 0, 0, 0, 0; 0, 0, 0, 6, 1, 2, 0, 0; 0, 
0, 0, 0, 0, 6, 3, 0; 0, 0, 0, 0, 0, 0, 3, 0; 0, 0, 0, 0, 0, 0, 0, 42; 0, 0, 
0, 0, 0, 0, 0, 0].

In GP 2.9.4 it again runs for a minute before raising a stack overflow.

Benjamin



On Friday, February 28, 2020 at 8:24:36 PM UTC-5, Justin Walker wrote:
>
> This isn’t an answer, but FWIW, I checked this on sage versions back to 
> 8.3, and get the same result. 
>
> I also checked, on an older (“Yosemite”) system with sage 7.3, and that 
> worked as you expect. 
>
> I don’t have time to dig into this, though. 
>
> HTH, 
>
> Justin 
>
> PS: All these were on Macs. 
>
> > On Feb 27, 2020, at 23:02 , 'Benjamin Matschke' via sage-nt <
> [email protected] <javascript:>> wrote: 
> > 
> > Dear all, 
> > 
> > The following code raises a ValueError: [*some big element of L*] is not 
> an S-unit. 
> > 
> > L.<theta_L> = NumberField(x^6 - 68463*x^4 - 5120808*x^3 + 1250774892*x^2 
> + 192368273328*x + 7520491439712) 
> > OLSstar = UnitGroup(L,proof=False,S=tuple(L.primes_above(2*5*7*13))) 
> > u = OLSstar.gen(11) 
> > print(u)              # yields u11 
> > print(L(u))           # yields some very large element of L 
> > print(OLSstar(L(u)))  # raises a ValueError 
> > 
> > The last output should of course be u11 again. The above is the simplest 
> example that I could extract. I have other examples, where the composed 
> coersion from OLSstar to L and back to OLSstar is not the identity 
> (according to sage, sometimes a minus is falsely introduced), although it 
> should. I do not know where the error comes from. 
> > 
> > Sometimes UnitGroup() also raises a PariError, which comes from 
> bnfsunit() within UnitGroup.__init__(), which can be resolved by increasing 
> pari's precision. The above error however persists after increasing pari's 
> precision generously. 
> > 
> > This was run on Sage 9.0, Linux Mint 19.3. Any help is appreciated. 
> > 
> > Thanks, 
> > Benjamin 
> > 
> > -- 
> > You received this message because you are subscribed to the Google 
> Groups "sage-nt" group. 
> > To unsubscribe from this group and stop receiving emails from it, send 
> an email to [email protected] <javascript:>. 
> > To view this discussion on the web visit 
> https://groups.google.com/d/msgid/sage-nt/276bbea0-ca70-411b-b2f9-f8ced377c7c2%40googlegroups.com.
>  
>
>
> -- 
> Justin C. Walker 
> Curmudgeon at Large 
> Director 
> Institute for the Enhancement of the Director's Income 
> -- 
> Build a man a fire and he'll be warm 
>  for a night. 
> Set a man on fire and he'll be warm 
>  for the rest of his life. 
>
>
>
>

-- 
You received this message because you are subscribed to the Google Groups 
"sage-nt" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To view this discussion on the web visit 
https://groups.google.com/d/msgid/sage-nt/bea704f1-8822-4239-ba77-94728ce95ef3%40googlegroups.com.

Reply via email to