#15278: Hash and equality for graphs
-------------------------------------+-------------------------------------
Reporter: SimonKing | Owner:
Type: defect | Status: needs_info
Priority: major | Milestone: sage-6.1
Component: graph theory | Resolution:
Keywords: | Merged in:
Authors: Simon King | Reviewers:
Report Upstream: N/A | Work issues:
Branch: | Commit:
u/SimonKing/ticket/15278 | 2fc8a772ee12fce7ac6abc4ecf9916f4746f5ee2
Dependencies: #12601, #15491 | Stopgaps:
-------------------------------------+-------------------------------------
Comment (by SimonKing):
Replying to [comment:47 ncohen]:
> > Problem: If I am not mistaken, people currently ''do'' mess with the
attribute `._immutable`, if I am not mistaken.
>
> Yes yes indeed, that's what Nicolas told me they did at the moment, and
that's what Python uses too know whether it should consider the object as
mutable or not. I just wondered if it checked whether the variable
existed, or whether the variable was set to True too.
The usage is `if getattr(self,'_immutable',False/True): raise
TypeError("...")` (false or true depending on whether we want to allow
something for all graphs except those that are known to be mutable, or
allow something for all graphs except those that are known to be
immutable).
> > Would it be feasible to try to remove the `._immutable` attribute (and
testing the type of the backend instead), see how much fails, and fix the
failing code by using "proper" immutable graphs?
>
> Hmmmm, I thought that Python really needed this `_immutable` variable
somewhere `O_o`
I've never heard before that Python uses it.
> > Would a flag in the "copy" function be supported by Python?
>
> Well not in `.__copy__` but in `.copy()` no problem I guess.
Now we talk about three things. I was talking about Python's (or Sage's?)
`copy()` function, which dispatches to `__copy__` or to pickling,
depending on what's available. You talk about the `.__copy__()` method,
which exists for graphs, and the `.copy()` method, which is set to
`__copy__` for graphs.
The `copy()` ''function'' does not use any argument except the object that
is being copied. For that reason, I think the `__copy__` method ''should''
not accept any arguments except `self`. But currently, the `.__copy__()`
method of graphs uses a plethora of optional arguments.
Instead, I suggest that one should introduce a separate `.copy()` method:
Here, it is no problem to add arguments. In particular, you can use such
method to create a mutable copy resp. an immutable copy of an immutable
resp. a mutable (di)graph. And it would be available in the docs (unlike
`.__copy__()`).
On the other hand (thinking after writing, as usual...): The optional
arguments of `__copy__` are not used, but when one defines `copy=__copy__`
then the method and its documentation do appear in the docs and can be
used by people. Hm. In the end, it might be better to keep it.
But one question on copying needs to be addressed: I think in some places
in Sage, `copy(X)` returns `X` if `X` is immutable, but in other places in
Sage, `copy(X)` always returns a ''mutable'' copy of `X`, regardless
whether `X` is mutable or not. What would you prefer for graphs?
In any case, I should add a test showing what happens with copies of
immutable graphs.
> I usually settle this by setting the keywords to "None" by default,
instead of True/False. Then if the user manually sets two keywords to
contradicting values, either ignore one of the two or scream, as usual.
I'll do that in another patch when this one will be reviewed, it will be
an easier syntax.
Do I understand correctly: You say I shouldn't try to add the "nicer"
syntax now, since you would take care of it in a review patch/other
ticket?
--
Ticket URL: <http://trac.sagemath.org/ticket/15278#comment:50>
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/groups/opt_out.
