Incomplete coverage in test.py is acceptable; the unexercised statements are due to:
The presence of uritemplate on my local system, so I don't excercise the ImportError fallback. The fallback allows the tests to run on systems without uritemplate, so we want to keep it.
The lack of buggy double entries, so I don't exercise the “entries in both” failure. This guard protects devs from adding buggy double entries, so we want to keep it.
The absence of any unexpected failures, so I don't exercise the raise. We want to keep it to get good messages if/when an implementation does a worse job of implementing the RFC.
The presence of all expected failures, so I don't exercise the “no longer raises an exception” failure. We want to keep it to get good messages if/when an implementation does a better job of implementing the RFC.
This gives us complete coverage for the
uri_template
package, as measured by:Incomplete coverage in
test.py
is acceptable; the unexercised statements are due to:The presence of
uritemplate
on my local system, so I don't excercise theImportError
fallback. The fallback allows the tests to run on systems withouturitemplate
, so we want to keep it.The lack of buggy double entries, so I don't exercise the “entries in both” failure. This guard protects devs from adding buggy double entries, so we want to keep it.
The absence of any unexpected failures, so I don't exercise the
raise
. We want to keep it to get good messages if/when an implementation does a worse job of implementing the RFC.The presence of all expected failures, so I don't exercise the “no longer raises an exception” failure. We want to keep it to get good messages if/when an implementation does a better job of implementing the RFC.