Struct validation should return ValidationErrors instead of ValidationError

[wackazong]

Thanks for putting this library together. I find it very easy and intuitive.

One question: are there plans to return ValidationErrors from struct validation instead of a single error? I think this would make much more sense since there can be multiple validation problems when validating the whole struct, probably related to specific fields.

[Keats]

Thanks!

So usually when doing the struct validation you stop at the first error. At least that is what I always do, no need to do further checks/work if we know it's not good.

[wackazong]

I think from a user perspective it would be helpful to not get caught in a trial/error loop. Especially if each transaction has an implied cost (I am working on a blockchain project).

From what I can see it should be possible to offer both options?

[Keats]

But you have individual field validation which behave the same way and you can hack it with ValidationError::params so I'm not too keen on that right now myself.

Let's see what other people think.

[theduke]

I agree with @wackazong here.

Concrete examples:

• validation of API request data:
  An API should report all errors at once, instead of just a single error

• validation of a configuration file before starting an application.
  As a user, I really want to see all the validation errors at once, instead of having to do a "fix one error, re-run the binary, fix the next error, re-run...

[Keats]

I think that makes sense to change for the next version.

[mseele]

@Keats do you mind if I do a PR for this feature?

[Keats]

Sure but make sure to do it on #249
I'm also wondering in general whether there are better/simpler ways to represent errors than those enums/structs, if you have ideas

[nouman2075]

Any update on this feature?