Primes()?? gets hung in len call; also bring coverage to 100% for primes.py

[saliola]

You don't need the first two lines below in cmp anymore.

        if isinstance(right, Primes_class):
            return 0
        return cmp(type(self), type(right))

Otherwise, it's a positive review.

[mabshoff]

This patch causes one doctest failure:

sage -t -long "devel/sage/sage/sets/set.py"                 
**********************************************************************
File "/scratch/mabshoff/sage-3.3.alpha4/devel/sage/sage/sets/set.py", line 278:
    sage: Primes() < Set(QQ)
Expected:
    True
Got:
    False
**********************************************************************

Cheers,

Michael

[saliola]

Replying to mabshoff:

This patch causes one doctest failure:

This is a weird test. I'm not even sure that it says anything meaningful. In fact, according to the documentation of cmp for Set, it doesn't:

        \note{If $X < Y$ is true this does \emph{not} necessarily mean
        that $X$ is a subset of $Y$.  Also, any two sets can be
        compared, which is a general Python philosophy.}

[AlexGhitza]

See #5933, which brings the coverage to 100% and was merged in 3.4.2.rc0.

[mabshoff]

Well, there is still a potential bug fix in here, so can someone take a look and extra a potential fix?

Cheers,

Michael

[mhansen]

I've rebased the patch of #5933.

[mhansen]

err, on top of #5933

[kcrisman]

Positive review. Unless you can provide an easily accessible example of where Primes(False) isn't Primes(True), but perhaps the first place that happens is waaay down the road...