Open jdemeyer opened 7 years ago
Author: Simon King, Ralf Stephan
I'm attaching a patch that I'm moving from #10483. One thing which is missing is the deprecation of the existing behaviour.
New commits:
acfead5 | Trac #23640: deprecate the misuse of symbolic variables as polynomial variable |
Minor nitpick about:
# Coerce alpha into L.
alpha = [L(a) for a in alpha]
Here you are doing a conversion, not a coercion. Either really do a coercion there or fix the comment.
It's also not clear why this coercion/conversion should be done only if indets
.
Reviewer: Daniel Krenn
Replying to @jdemeyer:
Minor nitpick about:
# Coerce alpha into L. alpha = [L(a) for a in alpha]
Here you are doing a conversion, not a coercion. Either really do a coercion there or fix the comment.
+1 for using .coerce
Replying to @jdemeyer:
It's also not clear why this coercion/conversion should be done only
if indets
.
+1 for doing this always.
if isinstance(a, str) and a not in K.variable_names():
indets.append(a)
There should be an error message, if some a
is ignored (i.e., when it is not a string and not an existing variable).
Currently,
FractionWithFactoredDenominator.smooth_critical_ideal
accepts its variable names as symbolic entries. This is really strange, especially since no symbolic computations are done.This should be deprecated and the variable names should be accepted as strings (or possibly variables from a pre-existing polynomial ring).
This is needed for #10483.
CC: @simon-king-jena @rwst @dkrenn
Component: asymptotic expansions
Author: Simon King, Ralf Stephan
Branch/Commit: u/jdemeyer/fractionwithfactoreddenominator_smooth_critical_ideal_should_accept_names_as_strings @
acfead5
Reviewer: Daniel Krenn
Issue created by migration from https://trac.sagemath.org/ticket/23640