Add wildcard type support to go code generator

[ebeahan]

Summary

Add support for the wildcard data type for the Go code generator.

Why is this change needed?

When wildcard appears once or more in the field definitions under schemas, the gocodegen Makefile directive will cause a make run to fail with the exception:

cd scripts \
          && GO111MODULE=on go run cmd/gocodegen/gocodegen.go \
                -version=2.0.0-dev \
                -schema=../schemas \
                -out=../code/go/ecs
no translation for wildcard (field executable)
exit status 1
make: *** [gocodegen] Error 1

[webmat]

I think we should take the opportunity to also add version as well, here. WDYT?

[ebeahan]

Good idea!

Does using string work here for the version data type or is there a more appropriate type to use?

cc @andrewkroh

[ebeahan]

And perhaps constant_keyword could be another addition here.

[andrewkroh]

Yeah, string seems like the best option for version. Another option would be a struct that has parameters that match up with the version data type (but I've never seen the version data type in ES so I have no idea what it looks like).

[webmat]

version is another type in the keyword family. So you can safely cram all manners of weird/incorrect version strings in it. But well-formatted ones will be queryable in new interesting ways.

So I think the simple answer is to have it equivalent to a Go string.

I don't object to supporting constant_keyword here. But let me open a door for a moment🚪

constant_keyword will be tricky to add to "main ECS", because using it means any index not using it correctly will have ingestion screech to a halt when sending a second value for the field.

I'm thinking it will have to be some sort of explicit opt-in, for folks who know they are adopting that specific strategy to splitting indices. Its usage will work much better with composable templates.

But we can close that door 🚪 and add constant_keyword as well 😂

[ebeahan]

Both version and constant_keyword have been added.

We've had some recent discussions around version being a keyword specialization type instead of a member of the keyword family. If that changes feelings around including it here (for now), I'm ok to remove.

[webmat]

Let's leave it in for now. I'm not sure if the incompatibility we saw in the field caps API call is an oversight or if it's here to stay. If it's here to stay, it's likely because the query possibilities are significantly greater with version (pun intended).

But the structure of the data itself is best captured as a string in the various programming languages.

[webmat]

LGTM