If you can reproduce the bug in the most recent version of pari, perhaps
submitting a bug report
<https://pari.math.u-bordeaux.fr/Bugs/Reporting.html> to Pari would be
good; they tend to be pretty responsive.
David

On Sun, Mar 1, 2020 at 7:15 PM 'Benjamin Matschke' via sage-nt <
[email protected]> wrote:

> 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]> 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].
>> > 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
> <https://groups.google.com/d/msgid/sage-nt/bea704f1-8822-4239-ba77-94728ce95ef3%40googlegroups.com?utm_medium=email&utm_source=footer>
> .
>

-- 
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/CAChs6_mgjrhXfWsZ-Uiv%2BGgUdLx_C-Dzd4Eb%2BHK3Y8sU%3DbEiqQ%40mail.gmail.com.

Reply via email to