Allow "Hidden" Buffers Used for Intermediate Decoding Steps

[rameshvarun]

Currently, it is always the case that the inspector Tree is decoded exactly from the user-provided data buffer. This doesn't work for all formats, particularly those that encapsulate their data in a lossless compression step (like GZIP or Huffman).

Examples:

• JPEG contains data compressed using Huffman encoding
• Archive files like ZIP contain compressed data.
• In the Ogg file format, a packet can be split across multiple pages, thereby representing a non-contiguous region

To decode these file formats, we need to be able to create an intermediate buffer, then create new Tree objects from the data in that intermediate buffer. This will have UI implications.