PMap order changes in Python 3.11a3, causing a doctest failure

[Ben Beasley]

One doctest in pyrsistent/_pmap.py fails on Python 3.11a3 because the order of PMap entries is not as expected.

I suspect this is due to the change from siphash24 to siphash13 for str/bytes.

I’m happy to submit a PR if you can suggest a preferred approach to dealing with this. One way to make these doctests order-insensitive would be to compare against a dict and expect a particular Boolean result, e.g. instead of

    >>> m1
    pmap({'b': 3, 'a': 1})

write

    >>> m1 == {'a': 1, 'b': 2}
    True

but of course this has slightly less value as documentation.

[Ben Beasley]

The failing doctest is:

___________________________________ [doctest] pyrsistent._pmap.PMap.update ____________________________________
211 
212         Return a new PMap with the items in Mappings inserted. If the same key is present in multiple
213         maps the rightmost (last) value is inserted.
214 
215         >>> m1 = m(a=1, b=2)
216         >>> m1.update(m(a=2, c=3), {'a': 17, 'd': 35})
Expected:
    pmap({'c': 3, 'b': 2, 'a': 17, 'd': 35})
Got:
    pmap({'d': 35, 'c': 3, 'b': 2, 'a': 17})

/home/ben/src/forks/pyrsistent/pyrsistent/_pmap.py:216: DocTestFailure

(Test command was PYTHONHASHSEED=0 PYTHONPATH="$PWD" pytest --doctest-modules pyrsistent.)

[Tobias Gustafsson]

Nice find! I'm good with the equality comparison you suggest above even though it's slightly contrived as doc. I haven't found any better way after some very quick research. Please file a PR!