Better restrictions on manifolds

[gh-FlorentinJ]

This ticket implements two new attributes of tensor fields on manifolds used to improve the restriction operation, by using rudimentary graph exploration to check for existing restriction and extensions.

Previously, defining the restriction of a tensor field on a subset V of a subset U before the restriction on the subset U would not register the first tensor as a restriction of the second, resulting in the creation of two python objects, despite being mathematically identical.

An example of code reproducing this behavior can be seen on this ​notebook. With this ticket, the order of creation no longer matters.

This is part of the ​​SageManifolds project; see the metaticket #18528 for an overview.

[git]

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

​9700c8b

Merge branch with 8.3.beta2

[Eric Gourgoulhon]

LGTM. Simply, could you add some Python comments in lines 426-427 of tensorfield.py regarding the meaning of the new attributes _extensions_graph and _restrictions_graph? This would be helpfull, especially since, technically speaking, these attributes are not graphs, but dictionaries.

[git]

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

​db9cd16

Added some comments to explain the purpose of new attributes

[Eric Gourgoulhon]

Thanks for the changes. LGTM. I think the failures in the plugins startup_time and startup_modules reported by the patchbot are irrelevant as far as this ticket is concerned.

[git]

Branch pushed to git repo; I updated commit sha1 and set ticket back to needs_review. New commits:

​cbbbf47

missing 's' in a un-doctested piece of code

[Eric Gourgoulhon]

OK, so everything is green now, including the patchbot.