This is a big step, as most PDFs which are downloaded online will be
linearized. Pretty much the only difference is that the xref structure
is slightly different.
This commit adds a parser as well as the Reader class, which serves
as a utility to aid in reading the PDF both forwards and in reverse.
The parser currently is capable of reading xref tables, as well as
all values. We don't really do anything with any of this information,
however.