dir(RIF) contains duplicates

[saraedum]

Currently, dir(RIF), and probably a few others, contains duplicates:

sage: A = dir(RIF)
sage: B = set(A)
sage: for b in B: A.remove(b)
sage: A
['__doc__', '_sage_src_lines_', 'is_field', 'is_integrally_closed']

The attached patch fixes this problem.

[dsm]

I don't seem to see the patch.

[saraedum]

hehe, you've been too fast. I needed the ticket number first to mention it in the TESTS section. I'm running the full doctests now, after that I'll set it to needs review; but feel free to have a look already.

[roed]

The problem with using the previous approach was that various inputs to unique_merge weren't sorted. Rather than this fix I would prefer fixing dir(self.class) to return a sorted list (currently 'Hom' is first), and to sort self.dict.keys() before passing it to unique_merge (note that all of the duplicated keys in this case come from self.dict.keys(), since dir(cls) is almost sorted).

[mhansen]

Replying to roed:

Rather than this fix I would prefer fixing dir(self.class) to return a sorted list (currently 'Hom' is first), and to sort self.dict.keys() before passing it to unique_merge (note that all of the duplicated keys in this case come from self.dict.keys(), since dir(cls) is almost sorted).

Is this just for performance reasons?

[roed]

Yeah. But I didn't mark it as needs work since I wanted another opinion. And since I haven't compared the performance of the two options.

[mhansen]

Is dir() actually used in a performance critical spot? I had just assumed it was basically for introspection.

[roed]

It probably is just for introspection. If someone else wants to give Julian's solution a positive review I'll be totally happy. :-)

[mhansen]

I looked at it and think that it's fine.