assertFalse has wrong name

[lionel-rowe]

Is your feature request related to a problem? Please describe.

Not sure what the prospects of changing this are given that assert is 1.x, but assertFalse has the wrong name (should be assertFalsy):

type Falsy = false | 0 | 0n | "" | null | undefined;
declare function assertFalse(expr: unknown, msg?: string): asserts expr is Falsy;

Describe the solution you'd like

Maybe rename to assertFalsy and deprecate assertFalse?

Describe alternatives you've considered

Do nothing

[kt3k]

assertFalsy was also suggested when we first discussed this feature, but we decided on this name because there are many prior examples in other languages and libraries #2045

I personally think assertFalse is acceptable as it's conventional name in software ecosystem in general

[kt3k]

There were also suggestions of renames like:

• assertStrictEquals -> assertIs
• assertEquals -> assertAlike

(I guess these are in similar direction)

But they didn't get much attention from the community #3513 (comment)

[lionel-rowe]

From the examples in other languages/frameworks listed in #2045 (comment), it looks like all of them except Python's unittest strictly mean "false" rather than "falsy". Tbh doesn't seem like there's much utility in having functions that only check one specific value, given that assertEquals(val, false) works just as well. I think the only way such a function is useful at all is if it means "falsy", and even then that's only preferable over assert(!val) because it enables type narrowing in TS Actually, it looks looks like TS is now able to narrow on assert(!val), so maybe there's not much utility to this function regardless of exact semantics.

In any case, using "false" to mean "falsy" strikes me as confusing, especially given the precedent in other testing frameworks to include such strictly-equal-to-a-specific-value functions.

Happy to close this issue if the chances of changing it are slim-to-none.