I guess caching is causing that. Creating ProjectiveSpace twice with the same input creates two projective spaces that are equal (P==Q) but not identical (P is not Q). Caching will look at "==" so if you compute something on the second projective space then you might get back something from the first projective space. Haven't looked at the ticket, though.
On Wednesday, October 15, 2014 5:49:50 PM UTC+1, Joao Alberto de Faria wrote: > > In the midst reviewing ticket 16986, my reviewer came up with the > following example that fails: > > R.<z>=PolynomialRing(QQ) > K.<w>=NumberField(z^3+2) > R.<t>=PolynomialRing(K) > L.<v>=K.extension(t^2+t+1) > P.<x,y>=ProjectiveSpace(L,1) > H=End(P) > f=H([x^3-2*y^3,v*y^3]) > f.rational_preimages(P([0,1])) > > However, when I run it on my machine, I get it to run with no problems, > returning > > [(-w*v : 1), (-w : 1), (w*v + w : 1)] > > BUT, when I run it a second time, making no changes, I get the following > error: > Traceback (click to the left of this block for traceback) > ... > TypeError: unable to convert 1/2*w^2*v to a rational > > After restarting my notebook, I receive the first answer, and upon running it > a second time, I get the second error. > I've tried running on other copies of Sage, and the same thing happens, it > will work the first time, but fail the second time. > > On further investigation, using the .dumps function, it seems that f is > changing every time we compile. However, it seems that it only ever changes > the very > first time. What I mean to say by that is that when I dump f after running it > the first time it will be different than when I dump f after running it a > second time. > From the second time on, dumping f will always return the same string. > > This is weird, what is the underlying mechanics causing this, and how can we > trace it more explicitly. > > > > > -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.