We briefly discussed in the maintainers' meeting of 2022-04-07, the deepEqual file in the harness has some problems in its semantics and we should at least encourage new tests not to use it, while working out a better way to provide that functionality.
I've taken a look at what uses it and identified three categories:
- Comparing two arrays with primitive values in them — it seems to me this can be replaced wholesale by
compareArray, which is clearer and more fit for purpose anyway.
- Comparing two objects with primitive values as their keys, where order of assignment
doesn't does matter.
- Comparing two arrays with non-primitive values in them.
Here are the places where it's currently used:
Providing alternatives for the remaining uses could be an interesting first test of our RFC process once I have come up with a proposal for that.
We briefly discussed in the maintainers' meeting of 2022-04-07, the
deepEqualfile in the harness has some problems in its semantics and we should at least encourage new tests not to use it, while working out a better way to provide that functionality.I've taken a look at what uses it and identified three categories:
compareArray, which is clearer and more fit for purpose anyway.doesn'tdoes matter.Here are the places where it's currently used:
Providing alternatives for the remaining uses could be an interesting first test of our RFC process once I have come up with a proposal for that.