SR.symbol: correct parent in inherting classes of SymbolicRing

[dkrenn]

sage: from sage.symbolic.ring import SymbolicRing
sage: class MySymbolicRing(SymbolicRing):
....:     def _repr_(self):
....:         return 'My Symbolic Ring'
sage: MySR = MySymbolicRing()
sage: MySR.var('x').parent()
Symbolic Ring

Component: symbolics

Author: Daniel Krenn

Branch/Commit: 9ac89ae

Reviewer: Jeroen Demeyer

Issue created by migration from https://trac.sagemath.org/ticket/19437

[dkrenn]

Branch: u/dkrenn/symbolic/sub-var

[dkrenn]

comment:2

make ptestlong is currently running...

---

New commits:

e1aca77

SR.symbol: set parent correctly (inheritance)

[dkrenn]

Commit: e1aca77

[dkrenn]

Author: Daniel Krenn

[dkrenn]

comment:3

Replying to @dkrenn:

make ptestlong is currently running...

Passed, thus, now really needs review :)

[jdemeyer]

comment:4

Can you tell me what's your use case for this branch?

I am a bit worried about

1. the impact on current code, in particular CallableSymbolicExpressionRing_class.
2. conflicts in case different parents use variables with the same name, see the global pynac_symbol_registry.

[jdemeyer]

Changed branch from u/dkrenn/symbolic/sub-var to u/jdemeyer/symbolic/sub-var

[jdemeyer]

comment:6

Have a look at my extra commit which makes symbol names specific to a parent. I'm not really sure that this is what you want since I don't know your use-case.

---

New commits:

9ac89ae

Move pynac_symbol_registry to cdef attribute SR.symbols

[jdemeyer]

Changed commit from e1aca77 to 9ac89ae

[dkrenn]

comment:7

Replying to @jdemeyer:

Can you tell me what's your use case for this branch?

Basically, I came long this with #19259, which implements subrings of the symbolic ring. A symbolic subring is inheriting from the symbolic ring class. The element_constructor is overridden in the following way: It calls the element_constructor of SR and then checks if the element's variables are "valid".

I had a short look at your changes and they seem to do what I need, but I'll have a more careful check later (I'm kind of busy right now).

Thanks

Daniel

[dkrenn]

comment:8

Replying to @jdemeyer:

Have a look at my extra commit which makes symbol names specific to a parent. I'm not really sure that this is what you want since I don't know your use-case.

Yout changes look good to me and do what this ticket claims. So from my side a positive review.

[jdemeyer]

Reviewer: Jeroen Demeyer

[vbraun]

Changed branch from u/jdemeyer/symbolic/sub-var to 9ac89ae