What version of ogen are you using?
v1.2.1
Can this issue be reproduced with the latest version?
Yes
What did you do?
I have a field in an API object with format: bytes and a length requirement:
key:
type: string
format: byte
minLength: 256
This is generating code that looks like this in oas_validators_gen.go
func (s *KeyInfo) Validate() error {
var failures []validate.FieldError
if err := func() error {
if err := (validate.String{
MinLength: 256,
MinLengthSet: true,
MaxLength: 0,
MaxLengthSet: false,
Email: false,
Hostname: false,
Regex: nil,
}).Validate(string(s.Key)); err != nil {
return errors.Wrap(err, "string")
}
return nil
}(); err != nil {
failures = append(failures, validate.FieldError{
Name: "key",
Error: err,
})
}
if len(failures) > 0 {
return &validate.Error{Fields: failures}
}
return nil
}
This converts the byte array to a string and uses the string validator, which counts runes for length, not bytes:
func (t String) Validate(v string) error {
if err := (Array{
MinLength: t.MinLength,
MinLengthSet: t.MinLengthSet,
MaxLength: t.MaxLength,
MaxLengthSet: t.MaxLengthSet,
}).ValidateLength(len([]rune(v))); err != nil {
return err
}
Thus, this code is incorrect: there is no guarantee that this byte array is valid utf-8. In fact it is not, so I consistently get errors that the key is too short because it is counting utf-8 runes rather than counting bytes.
What did you expect to see?
I would expect validation of format: bytes fields to be done over bytes, not converting to utf-8 runes
What did you see instead?
I see consistently incorrect length validation for format: bytes fields
What version of ogen are you using?
v1.2.1
Can this issue be reproduced with the latest version?
Yes
What did you do?
I have a field in an API object with format: bytes and a length requirement:
This is generating code that looks like this in oas_validators_gen.go
This converts the byte array to a string and uses the string validator, which counts runes for length, not bytes:
Thus, this code is incorrect: there is no guarantee that this byte array is valid utf-8. In fact it is not, so I consistently get errors that the key is too short because it is counting utf-8 runes rather than counting bytes.
What did you expect to see?
I would expect validation of format: bytes fields to be done over bytes, not converting to utf-8 runes
What did you see instead?
I see consistently incorrect length validation for format: bytes fields