Map validation, contextual validation, validator tags for custom map/slice types

[kszafran]

This pull request offers a few simple, but very useful improvements:

1. Adds support for validating maps, including registering tags for map types (as well as slice types).
2. Adds support for contextual validation. Custom validators (e.g. those registered with (*Validator).RegisterValidationCtx or (*Validator).RegisterStructValidationCtx) will be passed the Gin context (as context.Context). Resolves Access gin request Context from custom validators #2741
3. Returns slice and map validation errors are regular validator.ValidationErrors, without any custom types. This way callers can handle all validation errors in the same manner. Related to: Export struct sliceValidateError to allow error casting #2777

A further improvement would be to merge this PR: #2825. That will enable getting the actual *gin.Context in custom validators, resolving #1489.

[codecov]

Codecov Report

Merging #2877 (b626f63) into master (efa3175) will increase coverage by 0.02%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master    #2877      +/-   ##
==========================================
+ Coverage   98.73%   98.76%   +0.02%     
==========================================
  Files          41       41              
  Lines        3080     3150      +70     
==========================================
+ Hits         3041     3111      +70     
  Misses         27       27              
  Partials       12       12              

Flag	Coverage Δ
go-1.13	98.76% <100.00%> (+0.02%)	⬆️
go-1.14	98.60% <100.00%> (+0.03%)	⬆️
go-1.15	98.60% <100.00%> (+0.03%)	⬆️
go-1.16	98.60% <100.00%> (+0.03%)	⬆️
go-1.17	98.50% <100.00%> (+0.03%)	⬆️
macos-latest	98.76% <100.00%> (+0.02%)	⬆️
nomsgpack	98.74% <100.00%> (+0.02%)	⬆️
ubuntu-latest	98.76% <100.00%> (+0.02%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
binding/binding.go	100.00% <100.00%> (ø)
binding/binding_nomsgpack.go	100.00% <100.00%> (ø)
binding/default_validator.go	100.00% <100.00%> (ø)
binding/form.go	100.00% <100.00%> (ø)
binding/header.go	100.00% <100.00%> (ø)
binding/json.go	100.00% <100.00%> (ø)
binding/msgpack.go	100.00% <100.00%> (ø)
binding/query.go	100.00% <100.00%> (ø)
binding/uri.go	100.00% <100.00%> (ø)
binding/xml.go	100.00% <100.00%> (ø)
... and 4 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update efa3175...b626f63. Read the comment docs.

[kszafran]

@appleboy @thinkerou Would you have the time to review this PR? I suggest looking at it commit by commit, since each commit is self-contained.

[kszafran]

I did some renaming (Ctx -> Context) and fixed my omission in ShouldBindWith and ShouldBindBodyWith, which should call the contextual binding methods, but didn't. I also rebased onto latest master.

[kszafran]

Turns out that this may break custom validator tags if they use FieldLevel.Top(). I'm working on a version which will avoid that.

[kszafran]

I have reworked my implementation. Now it does not rely on Validate.Var/Validate.VarCtx, except for validating custom slice/array/map types. It will behave similar to the existing implementation, except the returned error is always of type validator.ValidationErrors and can be inspected to get the failing element index (or key in the case of maps).

[SubhramRana]

how?

[kszafran]

how?

Hey, what exactly are you asking about?

[kszafran]

I have found an edge case when validating non-nil pointers with nil value. It can be fixed by changing

	case reflect.Ptr:
		return v.ValidateStructContext(ctx, value.Elem().Interface())

to

	case reflect.Ptr:
		if value.IsNil() {
			return nil
		}
		return v.ValidateStructContext(ctx, value.Elem().Interface())

in ValidateStructContext.

I'm not updating the commits, though, because the maintainers don't seem to be particularly interested in this PR anyway.

[cikay]

I think it will be more clear to reverse the condition and add the continue statement since the indentation level will be decreased.

[lsdch]

Great feature, that will allow me to refactor a bunch of validation code. Hope this gets merged soon! 🙏

[kszafran]

Great feature, that will allow me to refactor a bunch of validation code. Hope this gets merged soon! 🙏

This PR has been here for a long time. I kept it open just in case, but I'm not even updating it anymore. The chances of getting it merged seem slim. I could try to rebase it, though, if the maintainers are interested in merging it.