zero vs zero_element / one vs one_element

[vdelecroix]

We deprecate the usage of zero_element/one_element in favor of zero/one.

To avoid duplication we also upgrade slightly the category of polyhedra.

see also: #5891

follow up: #17692

[vdelecroix]

New commits:

​d3e1ad5	trac #17694: upgrade the category of polyhedra
​c0a0bbb	trac #17694: one_element->one and zero_element->zero

[ncohen]

Hello,

There are some zero_element that you rename to zero without adding a deprecated copy. I wondered if perhaps the deprecated copy was inherited, but found something weird:

sage: L = LazyPowerSeriesRing(QQ)
sage: L.zero # zero_element appears when you press tab
L.zero          L.zero_element  L.zero_ideal    
sage: L.zero_element()
...
AttributeError: 'LazyPowerSeriesRing' object has no attribute 'zero_element'

So it seems that this function, for instance, has no deprecated zero_element. And also that the function still appears in the tab expansion, which I cannot explain.

Nathann

[jdemeyer]

Replying to vdelecroix:

To avoid duplication we also upgrade slightly the category of polyhedra.

Please explain...

[vdelecroix]

Replying to ncohen:

Right. I guess that the problem is that Rings is not a sub category of additive monoid and multiplicative monoid! I really do not want to fix it within categories as there is an interplay with Nicolas axioms.

I will add a commit that takes care of LazyPowerSeries and have a look at other rings which do not inherit from sage.rings.ring.Ring.

Replying to jdemeyer:

Replying to vdelecroix:

To avoid duplication we also upgrade slightly the category of polyhedra.

Please explain...

The polyhedra do have a sum (Minkowski sum) and a product (cartesian product). The method zero_element was implemented but the category was AdditiveMagmas. So I switched to AdditiveMonoids (i.e. Magma + unit) and only rename the method zero_element() to zero(). And the ParentMethods of the category takes care of the deprecation.

[jdemeyer]

The polyhedra change worries me (as potential reviewer) because I don't fully understand how categories work.

[vdelecroix]

Replying to jdemeyer:

The polyhedra change worries me (as potential reviewer) because I don't fully understand how categories work.

For the purpose of this ticket, you only need to know that ParentMethods and ElementMethods in the categories are respectively inherited to the concrete Parent and Element (via a dirty mechanism of dynamic classes created on the fly).

An other option is to revert the changes and add a deprecated_function_alias inside the polyhedron class.

Vincent

[git]

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

​e3aa703	trac #17692: rewrite __invert__ (and delete some)
​a1a9db3	trac #17694: remove the zero test
​cccd10c	trac #17694: place deprecated_function_alias where appropriate

[git]

Branch pushed to git repo; I updated commit sha1. This was a forced push. New commits:

​9ac0f9a

trac #17694: place deprecated_function_alias where appropriate

[vdelecroix]

Sorry... wrong commit. I messed up with #17692. Fixed now.

[jdemeyer]

deprecation(17694, ".one() is deprecated. Please use .one() instead.")

[git]

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

​0ba31f7

trac #17694: fix deprecation message

[mmezzarobba]

Lgtm, but I'd prefer waiting for a second opinion as Nathann and Jeroen appeared to have concerns.

[jdemeyer]

I don't have further concerns.

[ncohen]

Hello,

Lgtm, but I'd prefer waiting for a second opinion as Nathann and Jeroen appeared to have concerns.

Well, all I wanted to do with this branch is make sure that all functions that have been removed are now deprecated. If this is done I have no objection at all.

Nathann

[mmezzarobba]

Turns out I didn't notice a merge conflict with f1a086ae (which also reformats the example in a way that will hopefully make such conflicts less frequent).

[vdelecroix]

Replying to mmezzarobba:

Turns out I didn't notice a merge conflict with f1a086ae (which also reformats the example in a way that will hopefully make such conflicts less frequent).

Which ticket?

[mmezzarobba]

Replying to vdelecroix:

Replying to mmezzarobba:

Turns out I didn't notice a merge conflict with f1a086ae (which also reformats the example in a way that will hopefully make such conflicts less frequent).

Which ticket?

#12600

[git]

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

​5e83b95	Add epsilon() method to rings
​f1a086a	Fix doctest
​e0ec6ec	trac #17694: merge #12600