Failure with Python int modulo a rational

[tscrim]

Some fun I came across:

sage: int(5) % QQ(2)
---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-4-8ec0f9994ded> in <module>()
----> 1 int(Integer(5)) % QQ(Integer(2))

TypeError: Argument 'self' has incorrect type (expected sage.rings.rational.Rational, got int)

Note that this works:

sage: 5 % QQ(2)
1

CC: @vbraun @roed314

Component: basic arithmetic

Author: Travis Scrimshaw

Reviewer: Beth Malmskog

Merged: sage-5.12.beta0

Issue created by migration from https://trac.sagemath.org/ticket/14870

[tscrim]

Author: Travis Scrimshaw

[tscrim]

comment:1

Fixed. Thanks Volker and David for your insight.

[vbraun]

comment:3

You shouldn't check the type exactly, derived classes of rational should be treated like a rational.

Also, the whole PY_* macro usage that you see in the sources is mostly historical, nowadays Cython understands isinstance(x, Rational) just fine and this is the preferred form of a type check.

[sagetrac-malmskog]

comment:4

We tried using decimals like 3.0, coerced rationals like QQ(3), non-reduced fractions like 4/2, complex-like values like 3+0*I, and absolute values of integers as moduli (second argument in function)--all of these were correctly interpreted as integers. We also found correct error messages for moduli that didn't make sense, for example mod(31/10, 10) produces a correct error message. Looks good!

[tscrim]

Attachment: trac_14870-fix_int_mod_QQ-ts.patch.gz

[tscrim]

comment:5

New version as per Volker's comment.

[tscrim]

comment:6

@sagetrac-malmskog, you'll need to write your real name as the reviewer.

[sagetrac-malmskog]

Reviewer: Beth Malmskog

[jdemeyer]

Merged: sage-5.12.beta0