loads(dumps(G)) for G a graphics object

[kcrisman]

Dear Devel,

I am hoping to help the push to 75% by adding some doctests to some of
the plotting primitives. But for some reason, the following always
occurs:

sage: G = some graphics object
sage: G == loads(dumps(G))
False

Nonetheless, no matter how hard I try, I cannot actually find a
difference between G and loads(dumps(G)) when I view both of them,
look at xmin(), options(), etc. Why aren't they ==?

In any case, in plot.py and plot_field.py this is already worked
around by
sage: G = plot(something)
sage: H = loads(dumps(G))
so I'll go with that if there are no objections, but I would much
prefer == if possible. Incidentally, sage -coverage does not complain
about the current H=loads(dumps(G)) test.

Thanks,
- kcrisman

[Carl Witty]

On Thu, May 7, 2009 at 7:01 AM, kcrisman <kcri...@gmail.com> wrote:
> I am hoping to help the push to 75% by adding some doctests to some of
> the plotting primitives.  But for some reason, the following always
> occurs:
>
> sage: G = some graphics object
> sage: G == loads(dumps(G))
> False
>
> Nonetheless, no matter how hard I try, I cannot actually find a
> difference between G and loads(dumps(G)) when I view both of them,
> look at xmin(), options(), etc.  Why aren't they ==?

Because graphics objects don't implement equality (there are no __eq__
or __cmp__ methods defined), so the default implementation is used,
which is "is" (object identity, pointer equality).

Carl

[William Stein]

Ergo, you should implement __cmp__.

William

[kcrisman]

> > Because graphics objects don't implement equality (there are no __eq__
> > or __cmp__ methods defined), so the default implementation is used,
> > which is "is" (object identity, pointer equality).
>
> > Carl
>
> Ergo, you should implement __cmp__.

Umm... how would I do that? Or is "you" cwitty? Maybe if one of you
pointed me to a "good" implementation of __cmp__ for a similar type of
object, that would be helpful.

I don't know that I could do that in time for 4.0, though, so I'll
keep working on the other stuff with current loads(dumps()) behavior,
since it is already in two of the plot files.

Thanks for any other input,
- kcrisman

[William Stein]

On Thu, May 7, 2009 at 10:06 AM, kcrisman <kcri...@gmail.com> wrote:
>
>> > Because graphics objects don't implement equality (there are no __eq__
>> > or __cmp__ methods defined), so the default implementation is used,
>> > which is "is" (object identity, pointer equality).
>>
>> > Carl
>>
>> Ergo, you should implement __cmp__.
>
> Umm... how would I do that?  Or is "you" cwitty?  Maybe if one of you
> pointed me to a "good" implementation of __cmp__ for a similar type of
> object, that would be helpful.

You = somebody :-)

> I don't know that I could do that in time for 4.0, though, so I'll
> keep working on the other stuff with current loads(dumps()) behavior,
> since it is already in two of the plot files.

Yep, if you don't know how to implement __cmp__, then just don't
implement the loads(dumps(...))) doctests at all.

William