OpenSystemsDevelopment / qpdf

Embarcadero C++Builder doesn't support more than 50 files open at the same time for legacy 32 Bit apps, which makes a test fail trying to open more than that many files. This changes the number of open files for that test to far less to make the test succeed. Alternatively one could reduce the hard coded number of 200 in QPDF itself, which I didn't do currently because it needs adoption of manuals etc. and is something which needs to be discussed with the author of QPDF. I guess chances are better to get the test changed upstream.

This fixes #288: https://github.com/qpdf/qpdf/issues/288

Full parser context
    QPDF.cc(2): #include ..\..\..\..\src\include\qpdf\QPDF.hh
    QPDF.hh(48): class QPDF
    QPDF.hh(1380): decision to instantiate:  QPDF::ResolveRecorder::ResolveRecorder(QPDF *,const QPDFObjGen &)
    --- Resetting parser context for instantiation...
    QPDF.hh(799): parsing:  QPDF::ResolveRecorder::ResolveRecorder(QPDF *,const QPDFObjGen &)

Full parser context
    Pl_QPDFTokenizer.cc(1): #include ..\..\..\..\src\include\qpdf\Pl_QPDFTokenizer.hh
    Pl_QPDFTokenizer.hh(29): #include ..\..\..\..\src\include\qpdf/QPDFObjectHandle.hh
    QPDFObjectHandle.hh(51): class QPDFObjectHandle
    QPDFObjectHandle.hh(1052): decision to instantiate: PointerHolder<QPDFObject> QPDFObjectHandle::ObjAccessor::getObject(QPDFObjectHandle &)
    --- Resetting parser context for instantiation...
    QPDFObjectHandle.hh(909): parsing: PointerHolder<QPDFObject> QPDFObjectHandle::ObjAccessor::getObject(QPDFObjectHandle &)

Two operands must evaluate to the same value.

…lassic one nor newer ones based on CLANG.

Full parser context
    QPDF.cc(2): #include ..\..\..\..\src\include\qpdf\QPDF.hh
    QPDF.hh(47): class QPDF

Full parser context
    QPDF.cc(2): #include ..\..\..\..\src\include\qpdf\QPDF.hh
    QPDF.hh(46): class QPDF

JSON serialization was preserving the value as presented, but JSON
doesn't accept decimal values without a 0 before the decimal point.

The kfo-PDF files for testing need to be copied using "binmode"…

Full parser context
    test_driver.cc(208): parsing: void runtest(int,const char *,const char *)

Full parser context
    qpdf.cc(3803): parsing: PointerHolder<Pipeline> ImageOptimizer::makePipeline(const std::string &,Pipeline *)

qpdf: selecting --keep-open-files=n
qpdf: processing 001-kfo.pdf
WARNING: 001-kfo.pdf: file is damaged
WARNING: 001-kfo.pdf (offset 556): xref not found
WARNING: 001-kfo.pdf: Attempting to reconstruct cross-reference table

There have been issues reported where exceptions are not thrown
properly across shared library/DLL boundaries, so add a test
specifically to ensure that exceptions are caught as thrown.

Ensure zlib-flate doesn't link with an old libqpdf

Since autogenerated files are now committed, we no longer need to
rerun ./autogen.sh and actually shouldn't because it could mask
errors.

Also add test cases for additional coverage on image optimization.

Do not include the trailing EI, and handle cases where EI is not
preceded by a delimiter. Such cases have been seen in the wild.

This hasn't worked or been exercised in years since Adobe stopped
releasing a Linux version of reader.

We've actually seen a PDF file in the wild that contained EI
surrounded by delimiters inside the image data, which confused qpdf's
naive code. This significantly improves EI detection.

Add a version of expectInlineImage that takes an input source and
searches for EI. This is in preparation for improving the way EI is
found. This commit just refactors the code without changing the
functionality and adds tests to make sure the old and new code behave
identically.

The inline image token erroneously included the delimiter that
followed EI. The ObjectHandle created from it was correct.

When qpdf can't optimize an image because of an unsupported color
space, state this specifically. Recognize that many valid colorspaces
are not represented as name objects.

Make explicit that copyForeignObject can be used on page objects and
will copy them properly but not update the pages tree.

When linearizing a file or getting the list of all pages in a file,
detect if the pages tree contains a duplicated page object and, if so,
shallow copy it. This makes it possible to have a one to one mapping
of page positions to page objects.