Bug in Chart.__init__ regarding the determination of top charts

[egourgoulhon]

As noticed in #31901 comment:28, there is an issue when constructing two charts sharing the same coordinate symbols but not being otherwise related:

sage: M = Manifold(2, 'M', structure='topological')
sage: U = M.open_subset('U')
sage: V = M.open_subset('V')
sage: XU = U.chart('x y')
sage: XV = V.chart('x y')
sage: M.top_charts()
[Chart (U, (x, y))]

The chart XV should also appear as a top chart on M.

CC: @mkoeppe @tscrim @mjungmath @vbraun

Component: manifolds

Keywords: coordinate chart

Author: Eric Gourgoulhon

Branch/Commit: public/manifolds/top_charts_32112 @ 4e316e9

Reviewer: Matthias Koeppe

Issue created by migration from https://trac.sagemath.org/ticket/32112

[egourgoulhon]

New commits:

4e316e9

Fix bug in Chart.__init__ regarding the determination of top charts (Trac #32112)

[egourgoulhon]

Author: Eric Gourgoulhon

[egourgoulhon]

Commit: 4e316e9

[egourgoulhon]

Branch: public/manifolds/top_charts_32112

[mkoeppe]

Reviewer: Matthias Koeppe

[mkoeppe]

comment:2

Thanks! This is working well.

[mkoeppe]

comment:3

It would have been good to do this on top of #32009 though

[egourgoulhon]

comment:4

Thanks for the review.

Replying to @mkoeppe:

It would have been good to do this on top of #32009 though

Could you explain why? I mean, why _domain should become a Cython attribute?

[mkoeppe]

comment:5

In #31894 I hope to make Chart a subclass of sage.categories.map.Map so that all functionalities of morphisms such as composition etc. become available. Map is a Cython class in which _domain happens to be a Cython attribute.

[mkoeppe]

comment:6

As #32009 came back with a bug, I have made the present ticket a dependency of #32009 instead.

[egourgoulhon]

comment:7

Replying to @mkoeppe:

In #31894 I hope to make Chart a subclass of sage.categories.map.Map so that all functionalities of morphisms such as composition etc. become available. Map is a Cython class in which _domain happens to be a Cython attribute.

Thanks for the explanation.

[mkoeppe]

comment:8

Apparently this has been merged as part of #32089.