Streams don't need whitespace between dict and token

Added check for severity level in VIOLATION action to avoid segfaults
Added relaxation/logging for extra << endobj

Added check for severity level in VIOLATION action to avoid segfaults

Added relaxation/logging for stream length malformations (short-by-1, short-by-more, missing endstream)
amended endobj parsing to allow whitespace before it

TODO: would like to clean up so it doesn't get triggered twice (one of which has the object byte offset included)

Added relaxation/violation for missing %%EOF and/or startxref after internal (updated) xref-table + trailer