Question for all using this library - is the hierarchical error structure useful?

[deankarn]

Hello All!

Since implementing the "dive" tag, I have been questioning whether the hierarchical data structure is the best option, mainly because of the extra overhead and placeholder errors needed.

I could store the struct name, field name and name-spaced name on the field error and could just return a map of map[string]*FieldError where the key is the name-spaced name; with this approach I should be able to make validator even faster and more efficient.

my only concern is getting rid of the hierarchical functionality in case someone is using it or it would be beneficial to keep.

so should I just return a map of errors?

[deankarn]

@manucorporat I would definitly be interested in your opinion

[deankarn]

After a bunch of discussions it has been decided, in v6 I will drop the error hierarchy and just return the field errors. see #74