Closed marsell closed 3 years ago
The switch from @smaller/tap to tape happened with commit 26304bde0d77211f03a48972f46e55d9f00a38fa by @kusor . I wish to make sure the deepEqual behavior that concerns you isn't actually an improvement. My own JS knowledge is less than the two people I've tagged as reviewers, so I trust their opinions on this.
@marsell Hey, thanks for doing this! I'm in favor of these changes, but I'd like to do the following:
Some notes:
package.json uses tape 5.0.0 and above, but tape switched deep-equal library versions between 4.x and 5.x, and the new deep-equal has some strange behaviour: deepEqual() failing with objects that are semantically identical, but have a different ordering of attributes. For example:
I could fix the tests themselves, but since this new deepEqual() behaviour is nonsense, switch to using the older tape 4.13.2 instead.
Separately, there is a nics.test.js test failing due to a double-callback:
Separately, on my COAL instance, this test in machines.test.js is failing:
This is because two constraints (cpu and quota) are violated, not just one (quota):
I've extended the test so it handles this case as well.
Various other minor fixes to get tests passing.
make check
does not pass due to copyright check, but since I work for Spearhead Systems, I don't think my changing Joyent copyright headers is appropriate.