Avoid use of inspect.signature on CommandContribution class

[tlambert03]

fixes #144

This fixes the breakage of napari used with PySide2 in npe2=0.2.2, just by "avoiding" using inspect.signature ... this is super annoying bug in PySide2, that affects the ability to modify signatures of any python object. See napari/napari#2265 for details

[codecov]

Codecov Report

Merging #146 (8406043) into main (d86c2c4) will not change coverage.
The diff coverage is 100.00%.

@@            Coverage Diff            @@
##              main      #146   +/-   ##
=========================================
  Coverage   100.00%   100.00%           
=========================================
  Files           26        26           
  Lines         1719      1717    -2     
=========================================
- Hits          1719      1717    -2     

Impacted Files	Coverage Δ
npe2/_dynamic_plugin.py	100.00% <100.00%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update d86c2c4...8406043. Read the comment docs.

[tlambert03]

pre-commit failure is unrelated... upstream problem

[Carreau]

Trying closing and reopening for precommit.

[Czaki]

Suggested change

	cmd_kwargs = {
	k: kwargs.pop(k)
	for k in list(kwargs)
	if k in CommandContribution.__fields__
	}
	kwargs_set = set(kwargs)
	cmd_kwargs = {
	k: kwargs.pop(k)
	for k in kwargs_set
	if k in CommandContribution.__fields__
	}

I know that this is only a few positions, but creating a list every time? Some less experienced users may use this as inspiration in places where this condition is not satisfied.

[Carreau]

I mean you don't even need the set/list, as 'k in kwargs' will iterate over the dict keys, and if you use the set, then why not kwargs_set = set(kwargs).union(set(CommandContribution.__fields__)) and avoid the test.

[Czaki]

not union but intersection

[Czaki]

ahh, maybe to early. I replace in the head if check with iteration. Ignore this change request.

[Carreau]

not union but intersection

Ghhhhaaaa. Pfrrlrlrlflfr. Yes.

[Carreau]

... How many programmers do you need to get a dict comprehension right ...
Today is not a functioning brain day.

[tlambert03]

I mean you don't even need the set/list, as 'k in kwargs' will iterate over the dict keys,

we do in this case, since kwargs is getting changed on each iteration by kwargs.pop...

but creating a list every time?

it's not creating a list on each iteration either way (the list would only get called on the first evaluation of the comprehension)

In [13]: class T:
    ...:     def __iter__(self):
    ...:         print('iter')
    ...:         yield from ['a','b','c']
    ...:

In [14]: kwargs = T()

In [15]: list(kwargs)
iter
Out[15]: ['a', 'b', 'c']

In [16]: {k for k in list(kwargs) if k in ['a','b']}
iter
Out[16]: {'a', 'b'}

In [17]: {k for k in kwargs if k in ['a','b']}
iter
Out[17]: {'a', 'b'}

[Czaki]

pre-commit failure is unrelated... upstream problem

see #147