Improve comparisons of PolyDicts

[embray]

The PolyDict class has a __richcmp__ method which just implements the default rich comparison between the underlying dicts of the PolyDicts being compared, which in of itself is not terrible useful, especially given that it's not a meaningful comparison for relational operators.

However there is also an undocumented PolyDict.rich_compare method which i the same as __richcmp__ but takes a key argument (previously required, but I made it optional and renamed it to sortkey) which can be used to provide meaningful ordering between polynomials. This is used currently in Sage in exactly one place. But this ticket shows one other example where it was useful to have. We've also polished up the documentation just a little bit.

[chapoton]

branch is red

[git]

Branch pushed to git repo; I updated commit sha1. This was a forced push. New commits:

​4fd9277	some cleanup for PolyDict.rich_compare and __richcmp__
​5fb5c76	demonstrate an example where it makes sense to use PolyDict.rich_compare

[embray]

Trivial rebase at first. But I also fixed a test that was failing on Python 2 (the p1 < p2 test) due to the differences in default comparison behavior between Python 2 and 3. Not 100% satisfied with that, and wondering if we should instead explicitly enforce == and != only.

[chapoton]

I have corrected a typo and changed 2 http to https.

This allows many doctests to pass in Iwahori-Hecke algebras with python3.

---

New commits:

​ad058d6	Merge branch 'u/embray/polydict/richcmp' in 8.5.b2
​1cfdd4c	trac 26279 some http details and one typo

[embray]

Thanks!