PowerSeriesRing should call Ring.__init__

[simon-king-jena]

We have

sage: A.<a,b> = PowerSeriesRing(QQ)
sage: A._is_category_initialized()
False
sage: isinstance(A,Ring)
True

Since Ring.__init__ initialises the category, it seems that the the base class' initialisation is not called. That also prevents us from removing the alias for _Hom_ in sage.rings.ring - see #12876.

CC: @nthiery @nilesjohnson

Component: categories

Keywords: power series

Author: Simon King

Reviewer: Travis Scrimshaw

Merged: sage-5.5.beta1

Issue created by migration from https://trac.sagemath.org/ticket/13412

[nthiery]

comment:1

+1

Just 2 cents: alternatively, it could possibly call Parent.init but passing down the appropriate category.

[simon-king-jena]

comment:2

I think I'll even be getting category and coercion framework right (there was __call__, there was no self.Element, there was no _coerce_map_from_). With the patch I'm preparing:

sage: A.<a> = PowerSeriesRing(QQ)
sage: TestSuite(A).run()
sage: A.<a,b> = PowerSeriesRing(QQ)
sage: TestSuite(A).run()

Need to document the new stuff, though.

[simon-king-jena]

comment:3

PS: I also made the power series rings UniqueRepresentation, getting rid of the custom cache.

[nthiery]

comment:4

Replying to @simon-king-jena:

PS: I also made the power series rings UniqueRepresentation, getting rid of the custom cache.

Neat!

It's kind of you to be going the extra mile!

[simon-king-jena]

comment:5

Replying to @nthiery:

Replying to @simon-king-jena:

PS: I also made the power series rings UniqueRepresentation, getting rid of the custom cache.

Neat!

It's kind of you to be going the extra mile!

No, that's more like going by car. At first, I tried to fix the TestSuite by implementing __reduce__, which didn't work well. Inheriting from UniqueRepresentation did the trick.

[nthiery]

comment:6

No, that's more like going by car. At first, I tried to fix the TestSuite by implementing __reduce__, which didn't work well. Inheriting from UniqueRepresentation did the trick.

:-)

Hopefuly the previous reduce was good enough that there won't be issues with the pickles from the pickle jar.

[simon-king-jena]

comment:7

Replying to @nthiery:

Hopefuly the previous reduce was good enough that there won't be issues with the pickles from the pickle jar.

Well, the pickle jar is doctested, plus it won't break unless I remove the unpickling function used in the old pickles (which I don't).

[simon-king-jena]

Author: Simon King

[simon-king-jena]

Changed keywords from none to power series

[simon-king-jena]

comment:8

Cc to the original author (hi Niles).

The patch implements using the category and coercion framework for uni- and multi-variate power series rings. For me, tests pass (but I have a lot of other patches applied as well, so, better you or the patchbot test it as well).