Fixes #298

[bvulaj]

No description provided.

[bvulaj]

Worth noting that we can probably make this change for array fields as well. I can either make it part of this PR or just merge this and I'll make another.

[bserdar]

...else (if val==null) and if hidden_sub_path exists, remove the field from there

[bvulaj]

Good catch. We didn't have that happening before and this logic makes it quite a bit easier to implement. I think there's another issue out there that we can close.

[bserdar]

Lets make this for arrays as well

[bserdar]

I believe you're making this much more complicated than it needs to be. First, instead of strings for field names, use Path. Second, you don't need a Map for field and its corresponding hidden field. A list of fields
(paths) should suffice. And the algorithm would look like this:

populateCaseInsensitiveField(Object doc,Path field) {
   if field.numSegments==1
       if doc.get(field) is a list (must be primitive list)
           add hidden field to doc, and populate the list
       else
           add hidden field to doc, populate field
  else if field[0]=='*'
       doc is a List. For every element of doc, populateCaseInsensitiveField(element, field.suffix(-1)
  else
       populateCaseInsensitiveField(doc.get(field[0],field.suffix(-1)
}

[bvulaj]

@bserdar that's fair. I'm still spending some time cleaning this up. Let me take a look at the algo.

[bserdar]

Remove field from docObj.@hidden