Fix URLs redirects

[ocefpaf]

Testing the URL breakages in #185

[ocefpaf]

@efiring while having accurate URLs in the docs is nice I'm tempted to skip the URL check for now b/c a correct fix should be in the upstream matlab docs, right?

[efiring]

@ocefpaf I think the corrections you have in gsw.rst and __init__.py are where they should be. for _wrapped_ufuncs the corrections should be made by adding entries to subs in fix_wrapped_ufuncs.py. It looks like maybe only one additional entry is needed there.

Getting corrections upstream is a reasonable goal for the future, but not something I think we should try right now.

[efiring]

Also, I don't understand why some of the teos-10 urls are failing in the test. Both the original http: the new https versions work in my browser. The test is apparently not allowing the redirect? In any case, fixing them via the sub mechanism makes sense, as far as I can see.

[ocefpaf]

Yeah, the redirection works fine in the browser and sporadically here when I re-run the tests. That also failed with both the buggy micromamba and a pure pip solution. I wonder if the 403 is b/c we are hitting that URL multiple times (it is everywhere in the docs).

[efiring]

Good, thank you.