Closes #408

Co-authored-by: George Alverson <George.Alverson@cern.ch>

Co-authored-by: Matthew Peveler <matt.peveler@gmail.com>

See #1030

This allows adding options from command line. 

For example, run only one test in py39 env using an option -k:
$ tox -e py39 -- -k test_issue585

For example, run only one test in all envs using an option -k:
$ tox -- -k test_issue585

Co-authored-by: Harry Karvonen <harry.karvonen@onebyte.fi>

The Path constructor allows a variable amount of arguments to it which it joins together similar to os.path.join works, so it's not necessary to use os.path.join before passing the args to Path.

From the PDF 1.7 docs https://opensource.adobe.com/dc-acrobat-sdk-docs/standards/pdfstandards/pdf/PDF32000_2008.pdf:

Section 7.3.10 Indirect Objects:
An indirect reference to an undefined object shall not be considered an error by a conforming reader;
it shall be treated as a reference to the null object.

And section 7.5.4 Cross-Reference Table:
There are two ways an entry may be a member of the free entries list. Using the basic mechanism the free
entries in the cross-reference table may form a linked list, with each free entry containing the object number of
the next. The first entry in the table (object number 0) shall always be free and shall have a generation number
of 65,535; it is shall be the head of the linked list of free objects. The last free entry (the tail of the linked list)
links back to object number 0. Using the second mechanism, the table may contain other free entries that link
back to object number 0 and have a generation number of 65,535, even though these entries are not in the
linked list itself.

Those entries form a linked list. The correct way to handle this is to resolve the indirect reference to the NullObject.

See "3.4.3 Cross-Reference Table" in the PDF 1.7 standard for free cross-reference entries in general.

Co-authored-by: Harry Karvonen <harry.karvonen@onebyte.fi>

Closes #521
Closes #1034

Caused-by: #207

Why it wasn't detected by the tests: We don't have any tests that check
for the correct result of a merge. We just check for exceptions

How we prevent it in future: Unit test was added

Risk of the fix:
- We will have bigger file sizes again as #207 was effectively reverted
- We will need to adjust this test if we change the way we write PDFs

Closes: #1062

New Features (ENH):
-  Add PdfReader.xfa attribute (#1026)

Bug Fixes (BUG):
-  Wrong page inserted when PdfMerger.merge is done (#1063)
-  Resolve IndirectObject when it refers to a free entry (#1054)

Developer Experience (DEV):
-  Added {posargs} to tox.ini (#1055)

Maintenance (MAINT):
-  Remove PyPDF2._utils.bytes_type (#1053)

Testing (TST):
-  Scale page (indirect rect object) (#1057)
-  Simplify pathlib PdfReader test (#1056)
-  IndexError of VirtualList (#1052)
-  Invalid XML in xmp information (#1051)
-  No pycryptodome (#1050)
-  Increase test coverage (#1045)

Code Style (STY):
-  DOC of compress_content_streams (#1061)
-  Minimize diff for #879 (#1049)

Full Changelog: 2.4.1...2.4.2