Cache section data for scroll indicator computation

[TonicArtos]

Caching section data for use in the scroll indicator computation will allow for people to only need to reference SLMs in the first items in a section.

However, there aren't any significant advantages to this change. The API would be just a little more flexible for reusing xml layouts. In some cases there might be a little less coding.