OpenSystemsDevelopment / qpdf

Fix gratuitous calls to bad_subsections in Xref_table::subsections

Add inner class QPDF::Objects to encapsulate reading and managing of objects

Also, make obj_cache private and rename to table.

Move methods responsible for loading or keeping track of objects to
QPDF_objects.cc.

Add new protected inline method Pipeline::next

Also, tidy pipeline constructors and make subclasses final where possible.

Add additional xref and object stream sanity checks

Tweak #1287 comments

Add C API qpdf_oh_free_buffer to release memory allocated by stream data functions

Add new commands --remove-metadata and --remove-info

In QPDFWordTokenFinder::check limit the token length

Tokens longer than the target cannot be a match and therefore there is no
need to read to the end of token.

Fix bugs found during fuzzing

Return results rather than using reference parameters.

Fixes bug in #1272 where parameters were not reinitialized when calling
read_bad_entry from read_entry.

In  QPDFParser add a limit on total number of errors in one object

Currently, QPDFParser gives up attempting to parse an object if 5
near-consecutive bad tokens are encountered. Add a limit of a total of 15
bad tokens in a single object before giving up.

Refactor QPDF xref table

Also, delay adjustments for compressed objects until needed by
linearization checks.

Optimistically read subsection headers without reading individual object
entries, assuming that they are 20 bytes long as per the PDF spec. If
problems are encountered, fall back to calling bad_subsections.