LaurentPolynomial_mpair.__init__ modifies input

[etn40ff]

Creating mutlivariate Laurent polynomials from other mutlivariate Laurent polynomials sometimes fails.

sage: LQ = LaurentPolynomialRing(QQ, 'x0, x1, x2, y0, y1, y2, y3, y4, y5')
sage: LZ = LaurentPolynomialRing(ZZ, 'x0, x1, x2, y0, y1, y2, y3, y4, y5')
sage: LQ.inject_variables()
Defining x0, x1, x2, y0, y1, y2, y3, y4, y5
sage: x2^-1*y0*y1*y2*y3*y4*y5 + x1^-1*x2^-1*y0*y1*y3*y4 + x0^-1 in LZ
True
sage: x2^-1*y0*y1*y2*y3*y4*y5 + x1^-1*x2^-1*y0*y1*y3*y4 + x0^-1*x1^-1*y0*y3 + x0^-1 in LZ
Traceback (most recent call last):
...
AttributeError: 'tuple' object has no attribute 'esub'

Apparently this is due to the fact that LaurentPolynomial_mpair.__init__ changes the keys of the input dictionary.

[etn40ff]

Your patch seems to work. Thanks

---

New commits:

​ba66e9c	fix for loop whose changing its keys inside
​500205f	Added trac reference

[git]

Branch pushed to git repo; I updated commit sha1. New commits:

​c930e8c

Trac #22398: minor rewrite to use tuple(...) instead of copy(...)

[jdemeyer]

The correct solution is to not modify the input data (instead of working around the fact that the input data is modified).

It would be good to add an explicit doctest showing that the input is not modified. Something like

sage: LQ.<x,y> = LaurentPolynomialRing(QQ)
sage: D = {(-1, 1): 1}
sage: type(D.keys()[0])
<type 'tuple'>
sage: LQ(D)
x^-1*y
sage: type(D.keys()[0])
<type 'tuple'>

[dkrenn]

Replying to jdemeyer:

The correct solution is to not modify the input data (instead of working around the fact that the input data is modified).

I completely agree. I'll adapt the patch during the day (once my 7.6.beta4 has recompiled again...).

[git]

Branch pushed to git repo; I updated commit sha1. New commits:

​a4e1504	Trac #22398: make __init__ not modify its input
​1ca887a	Trac #22398: simplify factoring out _mon
​4a31db6	Trac #22398: doctest non-modifying input

[dkrenn]

Replying to jdemeyer:

The correct solution is to not modify the input data (instead of working around the fact that the input data is modified).

Done. Needs_review (...and let's see what the patchbot says).

[jdemeyer]

In the doctests, you can simplify code of the form

id(x) == id(y)

by

x is y

[git]

Branch pushed to git repo; I updated commit sha1. New commits:

​1f2d7f4

Trac #22398: simplify id(...) = id(...) in doctest

[dkrenn]

Replying to jdemeyer:

In the doctests, you can simplify code of the form

id(x) == id(y)

by

x is y

Oh, indeed :) ...done.

[jdemeyer]

If tests pass, you can set this to positive_review.

[git]

Branch pushed to git repo; I updated commit sha1. New commits:

​b56984c

Trac #22398: py3: fix <type 'tuple'>