Coverage ignores .coveragerc

[lmshk]

Upgrading hypothesis beyond 3.29.0 broke our test setup: apparently, hypothesis now interferes with coverage and makes it ignore .coveragerc, causing lots of entries in the coverage report and violating --fail-under=100. (The entries now include missed paths in installed packages and in Python itself.)

(Also, the upgrade tripled test running time and caused deprecation warnings of the type "test took too long"; possibly related: #879.)

Versions:

• Python: 3.6.2
• hypothesis: 3.30.0
• coverage: 4.4.1
• (virtualenv: 15.1.0)

Setting use_coverage to False fixed these issues.

[DRMacIver]

Whoops, sorry! Ignoring coveragerc for our own internal usage is intended behaviour, but it certainly shouldn't be propagating excluded paths back to the enclosing coverage. That's pretty bad. I'll try to get a fix out in the next couple of days.