Thanks, I did. For the reference: 
https://pari.math.u-bordeaux.fr/cgi-bin/bugreport.cgi?bug=2207

Benjamin

On Sunday, March 1, 2020 at 8:04:03 PM UTC-5, David Roe wrote:
>
> 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] <javascript:>> 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] <javascript:>.
>> 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/0748ea13-96cc-421a-9038-1da439b30ac5%40googlegroups.com.

Reply via email to