I have no idea how we missed this before but test output was always kind of bad for now reason. It was a relatively quick fix to improve it. In this PR:
Added line numbers to each output (sadly can't be extracted due to $LINENO being line-sensitive)
We briefly explored building out our own assertions (e.g. fail, assert_equals) (benefit of seeing expected/actual values) but it felt like we were going to rebuild the wheel (e.g. shunit2)
That's fine but feels like unnecessary work when the existing "wheel" works fine, this is mostly polishing the wheel
Based on #91, update
base
before landingI have no idea how we missed this before but test output was always kind of bad for now reason. It was a relatively quick fix to improve it. In this PR:
$LINENO
being line-sensitive)shunit2
technique with${assert_equals}
but it evaluated single quotes strings (thus breaking all our backticks), https://github.com/kward/shunit2/blob/v2.1.8/shunit2#L192make test
output!==
to!=
Notes:
fail
,assert_equals
) (benefit of seeing expected/actual values) but it felt like we were going to rebuild the wheel (e.g.shunit2
)Before:
After:
/cc @rpdelaney