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.
