#16532: Weak morphism created with register_coercion can result in invalid
coercion
----------------------------+------------------------
Reporter: tscrim | Owner: tscrim
Type: defect | Status: new
Priority: critical | Milestone: sage-6.3
Component: coercion | Resolution:
Keywords: | Merged in:
Authors: | Reviewers:
Report Upstream: N/A | Work issues:
Branch: | Commit:
Dependencies: | Stopgaps:
----------------------------+------------------------
Comment (by SimonKing):
Replying to [ticket:16532 tscrim]:
> First noted on #15475. Given `UniqueRepresentation` parents `A, B` and
a coercion `phi : A -> B` that was constructed/registered in the
`__init__` of `A` by using `register_coercion`.
Is this a typo? Or is it really the case that a coerce map into B is
registered during ''creation of A'' by calling `register_coercion` ''on
B''? This would be a misuse.
> Next suppose that `B` becomes garbage collected, and then so does `phi`.
That's the expected/preferred behaviour, since otherwise we'd have memory
leaks.
> Then `B` is recreated. Expecting there to be a coercion map from `A ->
B`, the user becomes very surprised when Sage responds with an errors
saying there is no coercion map.
That's why the coerce map from A to B should either be registered ''during
creation of B'' using `register_coercion, or (as Nils pointed out) should
be registered using `register_embedding` during creation of A, which
assumes that A embeds into B and may be done only once.
> Right now I'm too tired to come up with a minimal example, but is this a
case of misuse
Misuse.
--
Ticket URL: <http://trac.sagemath.org/ticket/16532#comment:2>
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.
