chore: refactor test contracts to fixtures

[tserg]

What I did

Convert the following types of test contracts in the test suite to fixtures:

• does not take in parametrized values and are compiled without any changes for each parametrized value in the test function;

@pytest.mark.parametrize(...)
def test_foo():
  code = """
  ...
  """
  c = get_contract(code) # <-- gets run N times
  # do something with c

• takes in parametrized values, but the parametrized values are not also used in the test function, and the test function contains other parametrized values

@pytest.mark.parametrize(a, [...])
@pytest.mark.parametrize(b, [...])
def test_foo():
  code = f"""
  ...{a}
  """
  c = get_contract(code) # <-- gets run N times
  # do something with c and b, but not a

How I did it

Extract the test contract as a fixture, and pass the fixture instead to the test function.

How to verify it

Run tests.

Description for the changelog

Refactor test contracts to fixtures

Cute Animal Picture

[codecov-commenter]

Codecov Report

Merging #2682 (279b374) into master (ea4eaee) will not change coverage. The diff coverage is n/a.

@@           Coverage Diff           @@
##           master    #2682   +/-   ##
=======================================
  Coverage   87.27%   87.27%           
=======================================
  Files          92       92           
  Lines        9895     9895           
  Branches     2479     2479           
=======================================
  Hits         8636     8636           
  Misses        774      774           
  Partials      485      485           

---

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 ea4eaee...279b374. Read the comment docs.

[charles-cooper]

note: to compare timings between https://github.com/vyperlang/vyper/actions/runs/2042620929 vs https://github.com/vyperlang/vyper/actions/runs/2044843484

[charles-cooper]

looks like each test comes down by an average of 15 seconds

[charles-cooper]

closing on account of the merge conflicts. we can return to this on a case-by-case basis later