I see. From the traceback, it looks like there are other places that ints/longs are being passed to mpmath. We could probably more carefully fix this in SymPy, but it might be better to fix this upstream in gmpy2 or mpmath. Anyway, I would be happy to have a PR against SymPy to fix this issue (I think the same workaround should work, wrap the value in MPZ).
Regarding the other comment, it was fixed in this pull request https://github.com/sympy/sympy/pull/12088/files#diff-6cd5274897d73197c25bc41e9afe63e4 Aaron Meurer On Tue, Jul 4, 2017 at 11:34 AM, Antoine Falaize <[email protected]> wrote: > Ok, thank you for your answer. > > Unfortunately, the current release candidate 1.1rc1 does not solve the > problem we are facing (see this recent test). > > Could this behavior be related to the second part of this commit message? > "" > There is another case at the end of Float.__new__ that is attempting to do > some normalization on unknown input types using mpmath.mpf(num)._mpf_, but I > am unclear when that case would actually be hit, so I've added a comment and > left it alone for now. > "" > > Can we help in a way or another? > > > > Le jeudi 29 juin 2017 16:56:57 UTC+2, Isuru Fernando a écrit : >> >> >> >> On Thu, Jun 29, 2017 at 5:43 PM, Antoine Falaize <[email protected]> >> wrote: >>> >>> Thank you for your investigations! >>> >>> So, as far as I understand: this error is related to the use of the >>> mpmath gmpy backend, and has been corrected. >>> >>> But, currently, this makes a large part of pyphs unusable for a small set >>> of platforms, and I do not understand: >>> (i) the reason for certain platforms to raise this error, while other >>> won't (we never explicitely specify any backend in pyphs, and do not point >>> any specific version of sympy); >> >> >> This happens when gmpy is installed only. (SymPy detects whether gmpy is >> installed and uses gmpy) >> >>> >>> (ii) if this fix is already merged in the current sympy master, or if it >>> will be in a future release. >> >> >> This is fixed in sympy master and will be in the next release > > > It seems the 1.1.rc1 > > >> >> Isuru Fernando >>> >>> >>> Can we do anything to help? >>> >>> >>> Le mercredi 28 juin 2017 21:57:54 UTC+2, Aaron Meurer a écrit : >>>> >>>> I don't think the fix was accidental. See >>>> >>>> https://github.com/sympy/sympy/pull/11862/commits/8a7e07992035f05780ebc33b4ffbf72d32a83069. >>>> From the commit message, "Also modify mpf_norm to handle non-mpz input >>>> when the mpmath gmpy backend is >>>> used." >>>> >>>> Your Point example doesn't fail for me in master. >>>> >>>> Aaron Meurer >>>> >>>> On Wed, Jun 28, 2017 at 8:02 AM, Isuru Fernando <[email protected]> >>>> wrote: >>>> > From reproduce I meant a similar error (didn't check pyphs) >>>> > >>>> > from sympy import Point >>>> > import pickle >>>> > pickle.loads(pickle.dumps(Point(1.1, 2.1).evalf())) >>>> > >>>> > >>>> > Isuru Fernando >>>> > >>>> > On Wed, Jun 28, 2017 at 5:28 PM, Isuru Fernando <[email protected]> >>>> > wrote: >>>> >> >>>> >> I can reproduce with gmpy2 and sympy=1.0, but goes away if I remove >>>> >> gmpy2 >>>> >> or update to sympy master. This was fixed accidentally in >>>> >> https://github.com/sympy/sympy/pull/11862 >>>> >> >>>> >> >>>> >> Isuru Fernando. >>>> > >>>> > >>>> > -- >>>> > You received this message because you are subscribed to the Google >>>> > Groups >>>> > "sympy" group. >>>> > To unsubscribe from this group and stop receiving emails from it, send >>>> > an >>>> > email to [email protected]. >>>> > To post to this group, send email to [email protected]. >>>> > Visit this group at https://groups.google.com/group/sympy. >>>> > To view this discussion on the web visit >>>> > >>>> > https://groups.google.com/d/msgid/sympy/CA%2B01voM9aDxDcb5HP8vRbj8jFpDXT3wm7KMo0RV2SqeHB7bm5w%40mail.gmail.com. >>>> > >>>> > For more options, visit https://groups.google.com/d/optout. >>> >>> -- >>> You received this message because you are subscribed to the Google Groups >>> "sympy" group. >>> To unsubscribe from this group and stop receiving emails from it, send an >>> email to [email protected]. >>> To post to this group, send email to [email protected]. >>> Visit this group at https://groups.google.com/group/sympy. >>> To view this discussion on the web visit >>> https://groups.google.com/d/msgid/sympy/c3651984-89af-42e7-bf5c-b6e0ce367dc1%40googlegroups.com. >>> >>> For more options, visit https://groups.google.com/d/optout. >> >> > -- > You received this message because you are subscribed to the Google Groups > "sympy" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected]. > To post to this group, send email to [email protected]. > Visit this group at https://groups.google.com/group/sympy. > To view this discussion on the web visit > https://groups.google.com/d/msgid/sympy/7d12409c-8619-480e-8886-ac0a12e5e278%40googlegroups.com. > > For more options, visit https://groups.google.com/d/optout. -- You received this message because you are subscribed to the Google Groups "sympy" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. Visit this group at https://groups.google.com/group/sympy. To view this discussion on the web visit https://groups.google.com/d/msgid/sympy/CAKgW%3D6L%2BFORzWRvTRjrt28NgfjkLwADYbSNrv3MrGxqT1X4uzw%40mail.gmail.com. For more options, visit https://groups.google.com/d/optout.
