nonempty is deprecated

[FabioZan86]

Hello, what reason to make it obsolete? How am I going to display this message now that the field is mandatory?

password: z.string()
    .nonempty('The Password field is required')
    .min(6, 'Password must be longer than 6 characters')
    .max(8, 'The password must not be a maximum of 8 characters'),

[renovatt]

I have the same question, what I have to use now?

[Pizle0210]

Just make sure it has a .min() which covers for .nonempty().

[samchungy]

I'm confused by your example, doesn't .min(6) cover your .nonempty case? I would assume Password must be longer than 6 characters covers that it's required. .min(1) would cover .nonempty() in other cases

[iway1]

No it doesn't because the error message is different, .min(1) would work though. It's a bit subjective but some would argue we should be showing a separate error message when empty versus not enough characters, pretty common I would say

[FabioZan86]

@samchungy I didn't understand what you meant. I just want to inform you that the field is mandatory. I don't want to inform you that the minimum is 6 to be mandatory. Every time the lib is being updated it becomes difficult to stay secure. I don't know why 'not empty' became obsolete!

[FabioZan86]

@samchungy Thank you very much! It worked perfectly for my case. I hadn't thought about that perspective, it may seem redundant, but it worked! The other 'codesandbox' example was missing z.coerce. Already ok.

[idlethumbs]

I'm using i18n translations for my errors, which means i have error mapping which previously had invalid_type_received_undefined mapped to 'field is required', this worked with non-empty,
using @samchungy's solution would work in situ but i would have to manually apply translations per zod schema in order to ensure that 'field is required' would be translated, whereas before it would apply globally,
not sure what the best solution is as 'field should be at least 1 character(s)' while accurate is a bit strange (and long!) from a UX point of view for a required field error.

[idlethumbs]

to answer my own comment - it seems that omitting min() and nonEmpty() - e.g. firstName: z.string() still results in a required field, and as long as i don't add .optional() to the field i can get the original required error to trigger.

[almeidapaulooliveira]

Use Yup !

[iway1]

@idlethumbs z.string() doesn't result in a required field because it will accept the empty string

[dacmarcell]

What about strings? Which method would assume nonempty?

[netojose]

What about strings? Which method would assume nonempty?

min(1)

[dacmarcell]

Tks!

[netojose]

You're welcome, @marcelldac

[oscarvedo]

What about strings? Which method would assume nonempty?

min(1)

Strings filled with spaces will pass the checks.

[jeremy-habit]

.trim().min(1)

[ACmaster7]

Why aren't the people feeling smart by answering "just do .min(1)" see how stupid this is?!
There should be a dedicated method for checking empty values and a separate one for minimum length, end of story. Doing .min(1, "Field is required") is not only unintuitive but also stupid. No validation or form library operates like this; empty-value validation and minimum-length validation are always separate concerns.

[samchungy]

Why aren't the people feeling smart by answering "just do .min(1)" see how stupid this is?!

My advice and opinions are free and apparently live rent free in some minds 😆.Take it or leave it 🤷

ps. JSON Schema does this 🤷

[kylecarhart]

For those coming to this discussion from google, nonempty for string has been undeprecated as of 0c6cbbd.

[iway1]

the deprecation here seems kind of random and counterproductive.

It's really subjective whether or not nonempty() is better than .min(1), both work fine and have identical behavior. But since it is subjective and already implemented, why not just leave it in and make upgrading versions easier for people who already use .nonempty()?

Edit: Oh lol just saw it's undeprecated now, good call