implement ASCIIHexDecode filter
add a fixme re packed/endian in ccitt fax filter
add "xref" eval (prints xref table); add "#<offset>" eval (prints an object at the offset within the file)
@<id> eval: back to chain eval
fix printing of objects by id
xref reading: workaround for broken PDFs with startxref pointint to garbage (not "xref" table)
reduce pdfobj recursion on number parsing
Since only the next object needs to be looked ahead (it has to be a number)
we should not recurse further because otherwise parsing integer-only arrays
is going to be rather slow and most likely produce a stack overflow.
Add an explicit "flags" to pdfobj function, with Norecurse set when looking
ahead.
ops: fix a few unused variables
bump hardcoded font width
parse: ignore escaped newlines in strings
parse: unrecognized escapes should drop the backslash
render: cropbox is inheritable
render: simplify glyph rendering
parse: I'm a dumbass sometimes