explicitCharkey doesn't work for whitespace

[jcsahnwaldt]

With explicitCharkey: false (default), parsing results are predictable:

• <x>y</x> -> { x: 'y' }
• <x></x> -> { x: '' }
• <x> </x> -> { x: ' ' }
• <x>\n</x> (newline) -> { x: '\n' }

With explicitCharkey: true, parsing results can be surprising:

• <x>y</x> -> { x: { _: 'y' } } (ok)
• <x></x> -> { x: '' } (bad, no _)
• <x> </x> -> { x: ' ' } (bad, no _)
• <x>\n</x> (newline) -> { x: '\n' } (bad, no _)

When users set explicitCharkey: true, they probably expect an _ property in all objects. Currently, they have to distinguish the case where an element contains only whitespace from other cases and use different code to access the text content.

[jcsahnwaldt]

Very likely related to #466: "handling of whitespace content inconsistent"