#17814: Make calling a cached method independent of source code inspection
-------------------------------------+-------------------------------------
Reporter: SimonKing | Owner:
Type: defect | Status: needs_review
Priority: major | Milestone: sage-6.6
Component: distribution | Resolution:
Keywords: | Merged in:
Authors: Simon King | Reviewers:
Report Upstream: N/A | Work issues:
Branch: | Commit:
u/SimonKing/make_calling_a_cached_method_independent_of_source_code_inspection|
a0b829dc10b4ccb00b99e073de8a1c0006aa1e51
Dependencies: | Stopgaps:
-------------------------------------+-------------------------------------
Comment (by jdemeyer):
Replying to [comment:51 SimonKing]:
> I think that the current branch does not have a high danger of creating
subtle caching problems. First of all, with the branch, a change in
behaviour can only occur when we have a cached method-with-default-
arguments-and/or-args/kwds of a cdef class whose source file is not
available. Methods without arguments will be totally fine if the source
file is gone (so, the branch does fix a problem).
Proposal [comment:15] also has that property.
> A change in behaviour would most likely be a problem where coercion is
involved, i.e., in `UniqueRepresentation`. But there, we are safe, since
(a) sub-classes of `UniqueRepresentation` must be Python classes anyway
(and Python classes aren't affected by the bug), and (b)
`UniqueRepresentation.__classcall__` has generic arguments `*args,**kwds`.
Python classes would be unaffected by proposal [comment:15].
> If you think that it is too dangerous to silently change behaviour
rather than raise a straight forward error, we could instead raise a
warning (only once, similar to deprecation warnings) when a cached method
is called on a method whose argspec can't be determined.
I think we need to agree first on whether the uniquess property for Cython
cached methods is important.
1. If uniqueness is important, then it is indeed "too dangerous to
silently change behaviour". ==> Go with the solution of Nils to somehow
put the argspec information in the docstring and use that.
2. If uniqueness is not important, then why bother reading the argspec
anyway? ==> Proposal [comment:15]
--
Ticket URL: <http://trac.sagemath.org/ticket/17814#comment:53>
Sage <http://www.sagemath.org>
Sage: Creating a Viable Open Source Alternative to Magma, Maple, Mathematica,
and MATLAB
--
You received this message because you are subscribed to the Google Groups
"sage-trac" 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 http://groups.google.com/group/sage-trac.
For more options, visit https://groups.google.com/d/optout.
