Strictly return a word with given length

[Shinigami92]

Clear and concise description of the problem

#714 (comment)

MO it feels a bit weird if a word with given length is not found then use random length 🤔
IMO we should change behavior for all such cases to a min/max length and search between that. If nothing found return undefined or throw an error?

Suggested solution

We need to discuss whether we should return undefined, null or an error in such a case

Alternative

No response

Additional context

No response

[pkuczynski]

I think for some values it might be hard to keep the exact length (like when length > 10). So maybe we should simply call this param maxLength?

[Shinigami92]

It's not that simple, because then users would loose current behavior to generate an exact length word
So we need a minLength and a maxLength, but I also think we should do this via an options param, so maybe { minLength, maxLength } or length: { min, max }

[pkuczynski]

Well, the current behavior is actually maxLength and not length. The param name is misleading :)

[pkuczynski]

I agree those functions could have 2 signatures:

function xyz (maxLength: number)
function xyz (options: { minLength?: number, maxLength?: number})

[Shinigami92]

Well, the current behavior is actually maxLength and not length. The param name is misleading :)

No it's not. The current behavior is: if length match was found use exact length, if not fallback to random length.

[pkuczynski]

Which means that current length actually describes maxLength, no? Or at least kind of. More like desiredLength ;)

[ST-DDT]

As the word functions currently use desiredLength we should keep this as the default. Maybe adding other variants as an option.

I suggest this signature
word( length?: number | { min?: number, max?: number, fallbackToAnyLength: boolean = true})

That way it would be non breaking. v6.2 at the earliest.

[Shinigami92]

I suggest this signature word( length?: number | { min?: number, max?: number, fallbackToAnyLength: boolean = true})

On the first looks that looks good
We will see if this can successfully apply to everything

[xDivisionByZerox]

This got solved with #1407.