On Monday, August 28, 2017 at 3:45:23 AM UTC-7, Nicolas M. Thiéry wrote:
>
> sage: from __future__ import unicode_literals
> sage: QQ['x,y']
> ...
> ValueError: variable name 'x,y' is not alphanumeric
>
PolynomialRing(QQ,'x,y')
and
PolynomialRing(QQ,u'x,y')
both work just fine, whereas
PolynomialRing(QQ,name="x,y")
does not. So I expect that QQ[...] does some analysis on its argument to
decide if it needs to pass a "name" or a "names" argument. However, finding
out where that code might live seems to be nigh impossible thanks to the
category stuff:
sage: QQ.__getitem__??
...
try:
meth = super(Parent, self).__getitem__
except AttributeError:
# needed when self is a Cython object
try:
meth = self.getattr_from_category('__getitem__')
except AttributeError:
return self.list()[n]
return meth(n)
so finding where this code needs to be made more robust for str/unicode is
very difficult.
MAINTAINABILITY REQUEST: make methods better introspectable in the face of
the category mro extensions.
It might be in sage.cagtegories.rings.py:999
def normalize_arg(arg):
if isinstance(arg, (tuple, list)):
# Allowing arbitrary iterables would create confusion,
but we
# may want to support a few more.
return tuple(arg)
elif isinstance(arg, str):
return tuple(arg.split(','))
else:
return (arg,)
We need a typecheck there that handles str and unicode the same.
(incidentally, because str == unicode in Py3, this issue would not occur in
Py3)
--
You received this message because you are subscribed to the Google Groups
"sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To post to this group, send email to [email protected].
Visit this group at https://groups.google.com/group/sage-devel.
For more options, visit https://groups.google.com/d/optout.
