What could perhaps be added is some kind of a normalization routine (to be implemented by subclasses) that is called before performing a hash. This might be quite expensive. There's also some fuzziness about what equality should be used, e.g. is this OK?
sage: S = set([(x+1)^2]) sage: x^2 + 2*x + 1 in S False That really depends on what you want. On Wed, Aug 28, 2013 at 2:38 PM, Stefan <stefanvanz...@gmail.com> wrote: > I ran into the following issue: > > sage: H5.<a,b,c> = FractionField(PolynomialRing(ZZ, ['a', 'b', 'c'])) > sage: S = set([1/(1-c)]) > sage: (-1)/(c-1) in S > False > > I know Sage doesn't promise that equal objects have equal hashes, but surely > this should be valid for objects from *a single ring*? > > I looked at the hash function implementation, and it's not straightforward to > fix this, since it's the generic FractionField method. In particular, we can > assume nothing about the ring on which the FractionField is based. I can come > up with some examples of rings, and try to do something for them, but my > imagination is kind of limited. Subclassing FractionField feels like > overkill, but a long list of try... Except clauses inside __hash__ is bad too. > > Thoughts? > > --Stefan > > -- > 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/groups/opt_out. -- 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/groups/opt_out.
