Immutability of chart functions

[gh-mjungmath]

Immutability of chart functions, see #30261.

[gh-mjungmath]

New commits:

​286d3c6	FreeModuleTensor, Vector: Subclass a new class ModuleElementWithMutability
​9df3d22	sage.tensor.modules: Make all zero() and one() elements immutable
​4373ea2	FreeModuleTensor: Replace special casing of the immutable zero by is_immutable
​a2ee8be	ModuleElementWithMutability.is_immutable, is_mutable: Change def to cpdef
​8600c2d	Merge branch 't/30181/immutable_elements_of_freemoduletensor' into t/30310/immutability_of_chart_functions
​76229bd	Trac #30310: immutability of chart functions

[egourgoulhon]

Why is #30181 a dependency of this ticket? A priori chart functions are totally independent from tensor fields. They even exist on pure topological manifolds.

[gh-mjungmath]

It is because of

-from sage.structure.element import AlgebraElement
+from sage.structure.element import AlgebraElement, ModuleElementWithMutability
...
-class ChartFunction(AlgebraElement):
+class ChartFunction(AlgebraElement, ModuleElementWithMutability):

[gh-mjungmath]

ModuleElementWithMutability is first introduced in #30181.

[tscrim]

I feel like the MultiCoordFunction should inherit from Mutability rather than copy the same code.

[gh-mjungmath]

It'd be the same issue like we had for affine connections (#30280): __reduce__ (#30281).

[tscrim]

I know, but as a stop-gap, you could instead implement a __reduce__. I don't really like this duplication. Hopefully this week I can get around to properly doing #30281...

[egourgoulhon]

Replying to gh-mjungmath:

It is because of

-from sage.structure.element import AlgebraElement
+from sage.structure.element import AlgebraElement, ModuleElementWithMutability
...
-class ChartFunction(AlgebraElement):
+class ChartFunction(AlgebraElement, ModuleElementWithMutability):

Ah yes, thanks.

[gh-mjungmath]

I tried to remove the redundant code and inherit from Mutability. For some reason, I still get a pickling error during doctest.

Furthermore, the code seems broken: is_immutable does not return anything:

    def is_immutable(self):
        """
        ...
        """
        self._is_immutable

When did that happen?

[gh-mjungmath]

I opened a ticket for the missing return: #31181.

Furthermore, I've opened a ticket for the still remaining pickling test error: #31182.

[git]

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

​09086cf

Trac #30310: inherit from Mutability

[git]

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

​6c76b66

Trac #30310: AssertionError -> ValueError

[git]

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

​7247620	Merge branch 'develop' into t/30310/immutability_of_chart_functions
​5ecbf3a	Trac #31181: return added
​8b20fc4	Merge branch 't/31181/mutability_class_does_not_return_is_immutable' into t/30310/immutability_of_chart_functions
​4c33935	Trac #31196: minor code improvements, py3 compatibility, documentation improved
​e5228d3	Trac #31196: cpdef require methods + example added
​d957f73	Trac #31196: unnecessary line in docstring removed
​d6d6ba4	Trac #31182: __getstate__ and __setstate__
​6cbd1fd	Trac #31182: doctests added for __setstate__ and __getstate__
​848e96b	Merge branch 't/31182/mutability_class_and_pickling' into t/30310/immutability_of_chart_functions

[gh-mjungmath]

Green patchbot.

[tscrim]

I think you are better off explicitly calling

ModuleElementWithMutability.__init__(self, parent)

[gh-mjungmath]

Yes thanks, this should be better.

[gh-mjungmath]

Pushed.

[git]

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

​e855ebc

Trac #30310: init ModuleElementWithMutability directly

[tscrim]

Green bot => positive review.

[gh-mjungmath]

Thank you.