approx on Decimal value fails with default tolerance arguments #3247

>>> from pytest import approx
>>> from decimal import Decimal
>>> approx(Decimal('1'))
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/Users/mike/temp/env/lib/python3.6/site-packages/_pytest/python_api.py", line 175, in __repr__
    vetted_tolerance = '{:.1e}'.format(self.tolerance)
  File "/Users/mike/temp/env/lib/python3.6/site-packages/_pytest/python_api.py", line 244, in tolerance
    relative_tolerance = set_default(self.rel, 1e-6) * abs(self.expected)
TypeError: unsupported operand type(s) for *: 'float' and 'decimal.Decimal'
>>> import sys
>>> sys.version
'3.6.4 (v3.6.4:d48ecebad5, Dec 18 2017, 21:07:28) \n[GCC 4.2.1 (Apple Inc. build 5666) (dot 3)]'
>>> import pytest
>>> pytest.__version__
'3.4.1'

>>> approx(foo, rel=Decimal('1e-6'))
1 ± 1.0e-6
>>> approx(foo, abs=Decimal('1e-12'))
1 ± 1.0e-12

$ pip list
DEPRECATION: The default format will switch to columns in the future. You can use --format=(legacy|columns) (or define a format=(legacy|columns) in your pip.conf under the [list] section) to disable this warning.
attrs (17.4.0)
pip (9.0.1)
pluggy (0.6.0)
py (1.5.2)
pytest (3.4.1)
setuptools (38.5.1)
six (1.11.0)
tox (2.9.1)
virtualenv (15.1.0)
wheel (0.30.0)