OpenSystemsDevelopment / qpdf

<?xml version="1.0" encoding="utf-8"?>
<!DOCTYPE book [
<!ENTITY ldquo "&#x201C;">
<!ENTITY rdquo "&#x201D;">
<!ENTITY mdash "&#x2014;">
<!ENTITY ndash "&#x2013;">
<!ENTITY nbsp "&#xA0;">
<!ENTITY swversion "2.0">
<!ENTITY lastreleased "April 29, 2008">
]>
<book>
 <bookinfo>
  <title>QPDF Manual</title>
  <subtitle>For QPDF Version &swversion;, &lastreleased;</subtitle>
  <author>
   <firstname>Jay</firstname><surname>Berkenbilt</surname>
  </author>
  <copyright>
   <year>2005&ndash;2008</year>
   <holder>Jay Berkenbilt</holder>
  </copyright>
 </bookinfo>
 <preface id="acknowledgments">
  <title>General Information</title>
  <para>
   QPDF is a program that does structural, content-preserving
   transformations on PDF files.  QPDF's website is located at <ulink
   url="http://qpdf.qbilt.org/">http://qpdf.qbilt.org/</ulink>.
  </para>
  <para>
   QPDF has been released under the terms of <ulink
   url="http://www.opensource.org/licenses/artistic-license-2.0.php">Version
   2.0 of the Artistic License</ulink>, a copy of which appears in the
   file <filename>Artistic-2.0</filename> in the source distribution.
  </para>
  <para>
   QPDF was originally created in 2001 and modified periodically
   between 2001 and 2005 during my employment at <ulink
   url="http://www.apexcovantage.com">Apex CoVantage</ulink>.  Upon my
   departure from Apex, the company graciously allowed me to take
   ownership of the software and continue maintaining as an open
   source project, a decision for which I am very grateful.  I have
   made considerable enhancements to it since that time.  I feel
   fortunate to have worked for people who would make such a decision.
   This work would not have been possible without their support.
  </para>
 </preface>
 <chapter id="ref.overview">
  <title>What is QPDF?</title>
  <para>
   QPDF is a program that does structural, content-preserving
   transformations on PDF files.  It could have been called something
   like <emphasis>pdf-to-pdf</emphasis>.  It also provides many useful
   capabilities to developers of PDF-producing software or for people
   who just want to look at the innards of a PDF file to learn more
   about how they work.
  </para>
  <para>
   QPDF is <emphasis>not</emphasis> a PDF content creation library, a
   PDF viewer, or a program capable of converting PDF into other
   formats.  In particular, QPDF knows nothing about the semantics of
   PDF content streams.  If you are looking for something that can do
   that, you should look elsewhere.  However, once you have a valid
   PDF file, QPDF can be used to transform that file in ways perhaps
   your original PDF creation can't handle.  For example, programs
   generate simple PDF files but can't password-protect them,
   web-optimize them, or perform other transformations of that type.
  </para>
 </chapter>
 <chapter id="ref.installing">
  <title>Building and Installing QPDF</title>
  <para>
   This chapter describes how to build and install qpdf.  Please see
   also the <filename>README</filename> and
   <filename>INSTALL</filename> files in the source distribution.
  </para>
  <sect1 id="ref.prerequisites">
   <title>System Requirements</title>
   <para>
    The qpdf package has relatively few external dependencies.  In
    order to build qpdf, the following packages are required:
    <itemizedlist>
     <listitem>
      <para>
       zlib: <ulink url="http://www.zlib.net/">http://www.zlib.net/</ulink>
      </para>
     </listitem>
     <listitem>
      <para>
       pcre: <ulink url="http://www.pcre.org/">http://www.pcre.org/</ulink>
      </para>
     </listitem>
     <listitem>
      <para>
       gnu make 3.81 or newer: <ulink url="http://www.gnu.org/software/make">http://www.gnu.org/software/make</ulink>
      </para>
     </listitem>
     <listitem>
      <para>
       perl version 5.8 or newer:
       <ulink url="http://www.perl.org/">http://www.perl.org/</ulink>;
       required for <command>fix-qdf</command> and the test suite.
      </para>
     </listitem>
     <listitem>
      <para>
       GNU diffutils (any version): <ulink
       url="http://www.gnu.org/software/diffutils/">http://www.gnu.org/software/diffutils/</ulink>
       is required to run the test suite.  Note that this is the
       version of diff present on virtually all GNU/Linux systems.
       This is required because the test suite uses <command>diff
       -u</command>.
      </para>
     </listitem>
    </itemizedlist>
   </para>
   <para>
    Part of qpdf's test suite does comparisons of the contents PDF
    files by converting them images and comparing the images.  You can
    optionally disable this part of the test suite by running
    <command>configure</command> with the
    <option>--disable-test-compare-images</option> flag.  If you leave
    this enabled, the following additional requirements are required
    by the test suite.  Note that in no case are these items required
    to use qpdf.
    <itemizedlist>
     <listitem>
      <para>
       libtiff: <ulink url="http://www.remotesensing.org/libtiff/">http://www.remotesensing.org/libtiff/</ulink>
      </para>
     </listitem>
     <listitem>
      <para>
       GhostScript version 8.60 or newer: <ulink
       url="http://pages.cs.wisc.edu/~ghost/">http://pages.cs.wisc.edu/~ghost/</ulink>
      </para>
     </listitem>
    </itemizedlist>
    This option is primarily intended for use by packagers of qpdf so
    that they can avoid having the qpdf packages depend on tiff and
    ghostscript software.
   </para>
   <para>
    If Adobe Reader is installed as <command>acroread</command>, some
    additional test cases will be enabled.  These test cases simply
    verify that Adobe Reader can open the files that qpdf creates.
    They require version 8.0 or newer to pass.  However, in order to
    avoid having qpdf depend on non-free (as in liberty) software, the
    test suite will still pass without Adobe reader, and the test
    suite still exercises the full functionality of the software.
   </para>
   <para>
    Pre-built documentation is distributed with qpdf, so you should
    generally not need to rebuild the documentation.  In order to
    build the documentation from its docbook sources, you need the
    docbook XML style sheets (<ulink
    url="http://downloads.sourceforge.net/docbook/">http://downloads.sourceforge.net/docbook/</ulink>).
    To build the PDF version of the documentation, you need Apache fop
    (<ulink
    url="http://xml.apache.org/fop/">http://xml.apache.org/fop/</ulink>)
    version 0.94 of higher.
   </para>
  </sect1>
  <sect1 id="ref.building">
   <title>Build Instructions</title>
   <para>
    Building qpdf on UNIX is generally just a matter of running

    <programlisting>./configure
make
</programlisting>
    You can also run <command>make check</command> to run the test
    suite and <command>make install</command> to install.  Please run
    <command>./configure --help</command> for options on what can be
    configured.  You can also set the value of
    <varname>DESTDIR</varname> during installation to install to a
    temporary location, as is common with many open source packages.
    Please see also the <filename>README</filename> and
    <filename>INSTALL</filename> files in the source distribution.
   </para>
   <para>
    There is currently no support for building qpdf on Windows.  The
    code is reasonably portable, however, and making it work on
    Windows would probably be reasonably straightforward.  A
    significant amount of the code in QPDF has been known to work on
    Windows in the past.
   </para>
   <para>
    There are some other things you can do with the build.  Although
    qpdf uses <application>autoconf</application>, it does not use
    <application>automake</application> but instead uses a
    hand-crafted non-recursive Makefile that requires gnu make.  If
    you're really interested, please read the comments in the
    top-level <filename>Makefile</filename>.
   </para>
  </sect1>
 </chapter>
 <chapter id="ref.using">
  <title>Running QPDF</title>
  <para>
   This chapter describes how to run the qpdf program from the command
   line.
  </para>
  <sect1 id="ref.invocation">
   <title>Basic Invocation</title>
   <para>
    When running qpdf, the basic invocation is as follows:

    <programlisting><command>qpdf</command><option> [ <replaceable>options</replaceable> ] <replaceable>infilename</replaceable> [ <replaceable>outfilename</replaceable> ]</option>
</programlisting>
    This converts PDF file <option>infilename</option> to PDF file
    <option>outfilename</option>.  The output file is functionally
    identical to the input file but may have been structurally
    reorganized.  Also, orphaned objects will be removed from the
    file.  Many transformations are available as controlled by the
    options below.
   </para>
   <para>
    <option>outfilename</option> does not have to be seekable, even
    when generating linearized files.  Specifying
    &ldquo;<option>-</option>&rdquo; as <option>outfilename</option>
    means to write to standard output.
   </para>
   <para>
    Most options require an output file, but some testing or
    inspection commands do not.  These are specifically noted.
   </para>
  </sect1>
  <sect1 id="ref.basic-options">
   <title>Basic Options</title>
   <para>
    The following options are the most common ones and perform
    commonly needed transformations.
    <variablelist>
     <varlistentry>
      <term><option>--password=password</option></term>
      <listitem>
       <para>
        Specifies a password for accessing encrypted files.
       </para>
      </listitem>
     </varlistentry>
     <varlistentry>
      <term><option>--linearize</option></term>
      <listitem>
       <para>
        Causes generation of a linearized (web optimized) output file.
       </para>
      </listitem>
     </varlistentry>
     <varlistentry>
      <term><option>--encrypt options --</option></term>
      <listitem>
       <para>
        Causes generation an encrypted output file.  Please see <xref
        linkend="ref.encryption-options"/> for details on how to
        specify encryption parameters.
       </para>
      </listitem>
     </varlistentry>
     <varlistentry>
      <term><option>--decrypt</option></term>
      <listitem>
       <para>
        Removes any encryption on the file.  A password must be
        supplied if the file is password protected.
       </para>
      </listitem>
     </varlistentry>
    </variablelist>
   </para>
   <para>
    Password-protected files may be opened by specifying a password.
    By default, qpdf will preserve any encryption data associated with
    a file.  If <option>--decrypt</option> is specified, qpdf will
    attempt to remove any encryption information.  If
    <option>--encrypt</option> is specified, qpdf will replace the
    document's encryption parameters with whatever is specified.
   </para>
   <para>
    Note that qpdf does not obey encryption restrictions already
    imposed on the file.  Doing so would be meaningless since qpdf can
    be used to remove encryption from the file entirely.  This
    functionality is not intended to be used for bypassing copyright
    restrictions or other restrictions placed on files by their
    producers.
   </para>
  </sect1>
  <sect1 id="ref.encryption-options">
   <title>Encryption Options</title>
   <para>
    To change the encryption parameters of a file, use the --encrypt
    flag.  The syntax is

    <programlisting><option>--encrypt <replaceable>user-password</replaceable> <replaceable>owner-password</replaceable> <replaceable>key-length</replaceable> [ <replaceable>restrictions</replaceable> ] --</option>
</programlisting>
    Note that &ldquo;<option>--</option>&rdquo; terminates parsing of
    encryption flags and must be present even if no restrictions are
    present.
   </para>
   <para>
    Either or both of the user password and the owner password may be
    empty strings.
   </para>
   <para>
    The value for
    <option><replaceable>key-length</replaceable></option> may be 40
    or 128.  The restriction flags are dependent upon key length.
    When no additional restrictions are given, the default is to be
    fully permissive.
   </para>
   <para>
    If <option><replaceable>key-length</replaceable></option> is 40,
    the following restriction options are available:
    <variablelist>
     <varlistentry>
      <term><option>--print=[yn]</option></term>
      <listitem>
       <para>
        Determines whether or not to allow printing.
       </para>
      </listitem>
     </varlistentry>
     <varlistentry>
      <term><option>--modify=[yn]</option></term>
      <listitem>
       <para>
        Determines whether or not to allow document modification.
       </para>
      </listitem>
     </varlistentry>
     <varlistentry>
      <term><option>--extract=[yn]</option></term>
      <listitem>
       <para>
        Determines whether or not to allow text/image extraction.
       </para>
      </listitem>
     </varlistentry>
     <varlistentry>
      <term><option>--annotate=[yn]</option></term>
      <listitem>
       <para>
        Determines whether or not to allow comments and form fill-in
        and signing.
       </para>
      </listitem>
     </varlistentry>
    </variablelist>
    If <option><replaceable>key-length</replaceable></option> is 128,
    the following restriction options are available:
    <variablelist>
     <varlistentry>
      <term><option>--accessibility=[yn]</option></term>
      <listitem>
       <para>
        Determines whether or not to allow accessibility to visually
        impaired.
       </para>
      </listitem>
     </varlistentry>
     <varlistentry>
      <term><option>--extract=[yn]</option></term>
      <listitem>
       <para>
        Determines whether or not to allow text/graphic extraction.
       </para>
      </listitem>
     </varlistentry>
     <varlistentry>
      <term><option>--print=<replaceable>print-opt</replaceable></option></term>
      <listitem>
       <para>
        Controls printing access.
        <option><replaceable>print-opt</replaceable></option> may be
        one of the following:
        <itemizedlist>
         <listitem>
          <para>
           <option>full</option>: allow full printing
          </para>
         </listitem>
         <listitem>
          <para>
           <option>low</option>: allow low-resolution printing only
          </para>
         </listitem>
         <listitem>
          <para>
           <option>none</option>: disallow printing
          </para>
         </listitem>
        </itemizedlist>
       </para>
      </listitem>
     </varlistentry>
     <varlistentry>
      <term><option>--modify=<replaceable>modify-opt</replaceable></option></term>
      <listitem>
       <para>
        Controls modify access.
        <option><replaceable>modify-opt</replaceable></option> may be
        one of the following:
        <itemizedlist>
         <listitem>
          <para>
           <option>all</option>: allow full document modification
          </para>
         </listitem>
         <listitem>
          <para>
           <option>annotate</option>: allow comment authoring and form operations
          </para>
         </listitem>
         <listitem>
          <para>
           <option>form</option>: allow form field fill-in and signing
          </para>
         </listitem>
         <listitem>
          <para>
           <option>assembly</option>: allow document assembly only
          </para>
         </listitem>
         <listitem>
          <para>
           <option>none</option>: allow no modifications
          </para>
         </listitem>
        </itemizedlist>
       </para>
      </listitem>
     </varlistentry>
    </variablelist>
    The default for each permission option is to be fully permissive.
   </para>
  </sect1>
  <sect1 id="ref.advanced-transformation">
   <title>Advanced Transformation Options</title>
   <para>
    These transformation options control fine points of how qpdf
    creates the output file.  Mostly these are of use only to people
    who are very familiar with the PDF file format or who are PDF
    developers.  The following options are available:
    <variablelist>
     <varlistentry>
      <term><option>--stream-data=<replaceable>option</replaceable></option></term>
      <listitem>
       <para>
        Controls transformation of stream data.  The value of
        <option><replaceable>option</replaceable></option> may be one
        of the following:
        <itemizedlist>
         <listitem>
          <para>
           <option>compress</option>: recompress stream data when
           possible (default)
          </para>
         </listitem>
         <listitem>
          <para>
           <option>preserve</option>: leave all stream data as is
          </para>
         </listitem>
         <listitem>
          <para>
           <option>uncompress</option>: uncompress stream data when
           possible
          </para>
         </listitem>
        </itemizedlist>
       </para>
      </listitem>
     </varlistentry>
     <varlistentry>
      <term><option>--normalize-content=[yn]</option></term>
      <listitem>
       <para>
        Enables or disables normalization of content streams.
       </para>
      </listitem>
     </varlistentry>
     <varlistentry>
      <term><option>--suppress-recovery</option></term>
      <listitem>
       <para>
        Prevents qpdf from attempting to recover damaged files.
       </para>
      </listitem>
     </varlistentry>
     <varlistentry>
      <term><option>--object-streams=<replaceable>mode</replaceable></option></term>
      <listitem>
       <para>
        Controls handing of object streams.  The value of
        <option><replaceable>mode</replaceable></option> may be one of
        the following:
        <itemizedlist>
         <listitem>
          <para>
           <option>preserve</option>: preserve original object streams
           (default)
          </para>
         </listitem>
         <listitem>
          <para>
           <option>disable</option>: don't write any object streams
          </para>
         </listitem>
         <listitem>
          <para>
           <option>generate</option>: use object streams wherever
           possible
          </para>
         </listitem>
        </itemizedlist>
       </para>
      </listitem>
     </varlistentry>
     <varlistentry>
      <term><option>--ignore-xref-streams</option></term>
      <listitem>
       <para>
        Tells qpdf to ignore any cross-reference streams.
       </para>
      </listitem>
     </varlistentry>
     <varlistentry>
      <term><option>--qdf</option></term>
      <listitem>
       <para>
        Turns on QDF mode.  For additional information on QDF, please
        see <xref linkend="ref.qdf"/>.
       </para>
      </listitem>
     </varlistentry>
    </variablelist>
   </para>
   <para>
    By default, when a stream is encoded using non-lossy filters that
    qpdf understands and is not already compressed using a good
    compression scheme, qpdf will uncompress and recompress streams.
    Assuming proper filter implements, this is safe and generally
    results in smaller files.  This behavior may also be explicitly
    requested with <option>--stream-data=compress</option>.
   </para>
   <para>
    When <option>--stream-data=preserve</option> is specified, qpdf
    will never attempt to change the filtering of any stream data.
   </para>
   <para>
    When <option>--stream-data=uncompress</option> is specified, qpdf
    will attempt to remove any non-lossy filters that it supports.
    This includes <literal>/FlateDecode</literal>,
    <literal>/LZWDecode</literal>, <literal>/ASCII85Decode</literal>,
    and <literal>/ASCIIHexDecode</literal>.  This can be very useful
    for inspecting the contents of various streams.
   </para>
   <para>
    When <option>--normalize-content=y</option> is specified, qpdf
    will attempt to normalize whitespace and newlines in page content
    streams.  This is generally safe but could, in some cases, cause
    damage to the content streams.  This option is intended for people
    who wish to study PDF content streams or to debug PDF content.
    You should not use this for &ldquo;production&rdquo; PDF files.
   </para>
   <para>
    Ordinarily, qpdf will attempt to recover from certain types of
    errors in PDF files.  These include errors in the cross-reference
    table, certain types of object numbering errors, and certain types
    of stream length errors.  Sometimes, qpdf may think it has
    recovered but may not have actually recovered, so care should be
    taken when using this option as some data loss is possible.  The
    <option>--suppress-recovery</option> option will prevent qpdf from
    attempting recovery.  In this case, it will fail on the first
    error that it encounters.
   </para>
   <para>
    Object streams, also known as compressed objects, were introduced
    into the PDF specification at version 1.5, corresponding to
    Acrobat 6.  Some older PDF viewers may not support files with
    object streams.  qpdf can be used to transform files with object
    streams to files without object streams or vice versa.  As
    mentioned above, there are three object stream modes:
    <option>preserve</option>, <option>disable</option>, and
    <option>generate</option>.
   </para>
   <para>
    In <option>preserve</option> mode, the relationship to objects and
    the streams that contain them is preserved from the original file.
    In <option>disable</option> mode, all objects are written as
    regular, uncompressed objects.  The resulting file should be
    readable by older PDF viewers.  (Of course, the content of the
    files may include features not supported by older viewers, but at
    least the structure will be supported.)  In
    <option>generate</option> mode, qpdf will create its own object
    streams.  This will usually result in more compact PDF files,
    though they may not be readable by older viewers.  In this mode,
    qpdf will also make sure the PDF version number in the header is
    at least 1.5.
   </para>
   <para>
    Ordinarily, qpdf reads cross-reference streams when they are
    present in a PDF file.  If <option>--ignore-xref-streams</option>
    is specified, qpdf will ignore any cross-reference streams for
    hybrid PDF files.  The purpose of hybrid files is to make some
    content available to viewers that are not aware of cross-reference
    streams.  It is almost never desirable to ignore them.  The only
    time when you might want to use this feature is if you are testing
    creation of hybrid PDF files and wish to see how a PDF consumer
    that doesn't understand object and cross-reference streams would
    interpret such a file.
   </para>
   <para>
    The <option>--qdf</option> flag turns on QDF mode, which changes
    some of the defaults described above.  Specifically, in QDF mode,
    by default, stream data is uncompressed, content streams are
    normalized, and encryption is removed.  These defaults can still
    be overridden by specifying the appropriate options as described
    above.  Additionally, in QDF mode, stream lengths are stored as
    indirect objects, objects are laid out in a less efficient but
    more readable fashion, and the documents are interspersed with
    comments that make it easier for the user to find things and also
    make it possible for <command>fix-qdf</command> to work properly.
    QDF mode is intended for people, mostly developers, who wish to
    inspect or modify PDF files in a text editor.  For details, please
    see <xref linkend="ref.qdf"/>.
   </para>
  </sect1>
  <sect1 id="ref.testing-options">
   <title>Testing, Inspection, and Debugging Options</title>
   <para>
    These options can be useful for digging into PDF files or for use
    in automated test suites for software that uses the qpdf library.
    When any of the options in this section are specified, no output
    file should be given.  The following options are available:
    <variablelist>
     <varlistentry>
      <term><option>--static-id</option></term>
      <listitem>
       <para>
        Causes generation of a fixed value for /ID.  This is intended
        for testing only.  Never use it for production files.
       </para>
      </listitem>
     </varlistentry>
     <varlistentry>
      <term><option>-show-encryption</option></term>
      <listitem>
       <para>
        Shows document encryption parameters.  Also shows the
        document's user password if the owner password is given.
       </para>
      </listitem>
     </varlistentry>
     <varlistentry>
      <term><option>-check-linearization</option></term>
      <listitem>
       <para>
        Checks file integrity and linearization status.
       </para>
      </listitem>
     </varlistentry>
     <varlistentry>
      <term><option>-show-linearization</option></term>
      <listitem>
       <para>
        Checks and displays all data in the linearization hint tables.
       </para>
      </listitem>
     </varlistentry>
     <varlistentry>
      <term><option>-show-xref</option></term>
      <listitem>
       <para>
        Shows the contents of the cross-reference table in a
        human-readable form.  This is especially useful for files with
        cross-reference streams which are stored in a binary format.
       </para>
      </listitem>
     </varlistentry>
     <varlistentry>
      <term><option>-show-object=obj[,gen]</option></term>
      <listitem>
       <para>
        Show the contents of the given object.  This is especially
        useful for inspecting objects that are inside of object
        streams (also known as &ldquo;compressed objects&rdquo;).
       </para>
      </listitem>
     </varlistentry>
     <varlistentry>
      <term><option>-raw-stream-data</option></term>
      <listitem>
       <para>
        When used along with the <option>--show-object</option>
        option, if the object is a stream, shows the raw stream data
        instead of object's contents.
       </para>
      </listitem>
     </varlistentry>
     <varlistentry>
      <term><option>-filtered-stream-data</option></term>
      <listitem>
       <para>
        When used along with the <option>--show-object</option>
        option, if the object is a stream, shows the filtered stream
        data instead of object's contents.  If the stream is filtered
        using filters that qpdf does not support, an error will be
        issued.
       </para>
      </listitem>
     </varlistentry>
     <varlistentry>
      <term><option>-show-pages</option></term>
      <listitem>
       <para>
        Shows the object and generation number for each page
        dictionary object and for each content stream associated with
        the page.  Having this information makes it more convenient to
        inspect objects from a particular page.
       </para>
      </listitem>
     </varlistentry>
     <varlistentry>
      <term><option>-with-images</option></term>
      <listitem>
       <para>
        When used along with <option>--show-pages</option>, also shows
        the object and generation numbers for the image objects on
        each page.  (At present, information about images in shared
        resource dictionaries are not output by this command.  This is
        discussed in a comment in the source code.)
       </para>
      </listitem>
     </varlistentry>
     <varlistentry>
      <term><option>-check</option></term>
      <listitem>
       <para>
        Checks file structure and well as encryption and
        linearization.  A file for which <option>--check</option>
        reports no errors may still have errors in stream data but
        should otherwise be otherwise structurally sound.
       </para>
      </listitem>
     </varlistentry>
    </variablelist>
   </para>
   <para>
    The <option>--raw-stream-data</option> and
    <option>--filtered-stream-data</option> options are ignored unless
    <option>--show-object</option> is given.  Either of these options
    will cause the stream data to be written to standard output.  In
    order to avoid commingling of stream data with other output, it is
    recommend that these objects not be combined with other
    test/inspection options.
   </para>
   <para>
    If <option>--filtered-stream-data</option> is given and
    <option>--normalize-content=y</option> is also given, qpdf will
    attempt to normalize the stream data as if it is a page content
    stream.  This attempt will be made even if it is not a page
    content stream, in which case it will produce unusuable results.
   </para>
  </sect1>
 </chapter>
 <chapter id="ref.qdf">
  <title>QDF Mode</title>
  <para>
   In QDF mode, qpdf creates PDF files in what we call <firstterm>QDF
   form</firstterm>.  A PDF file in QDF form, sometimes called a QDF
   file, is a completely valid PDF file that has
   <literal>%QDF-1.0</literal> as its third line (after the pdf header
   and binary characters) and has certain other characteristics.  The
   purpose of QDF form is to make it possible to edit PDF files, with
   some restrictions, in an ordinary text editor.  This can be very
   useful for experimenting with different PDF constructs or for
   making one-off edits to PDF files (though there are other reasons
   why this may not always work).
  </para>
  <para>
   It is ordinarily very difficult to edit PDF files in a text editor
   for two reasons: most meaningful data in PDF files is compressed,
   and PDF files are full of offset and length information that makes
   it hard to add or remove data.  A QDF file is organized in a manner
   such that, if edits are kept within certain constraints, the
   <command>fix-qdf</command> program, distributed with qpdf, is able
   to restore edited files to a correct state.  The
   <command>fix-qdf</command> program takes no command-line
   arguments.  It reads a possibly edited QDF file from standard input
   and writes a repaired file to standard output.
  </para>
  <para>
   The following attributes characterize a QDF file:
   <itemizedlist>
    <listitem>
     <para>
      All objects appear in numerical order in the PDF file, including
      when objects appear in object streams.
     </para>
    </listitem>
    <listitem>
     <para>
      Objects are printed in an easy-to-read format, and all line
      endings are normalized to UNIX line endings.
     </para>
    </listitem>
    <listitem>
     <para>
      Unless specifically overridden, streams appear uncompressed
      (when qpdf supports the filters and they are compressed with a
      non-lossy compression scheme), and most content streams are
      normalized (line endings are converted to just a UNIX-style
      linefeeds).
     </para>
    </listitem>
    <listitem>
     <para>
      All streams lengths are represented as indirect objects, and the
      stream length object is always the next object after the stream.
      If the stream data does not end with a newline, an extra newline
      is inserted, and a special comment appears after the stream
      indicating that this has been done.
     </para>
    </listitem>
    <listitem>
     <para>
      If the PDF file contains object streams, if object stream
      <emphasis>n</emphasis> contains <emphasis>k</emphasis> objects,
      those objects are numbered from <emphasis>n+1</emphasis> through
      <emphasis>n+k</emphasis>, and the object number/offset pairs
      appear on a separate line for each object.  Additionally, each
      object in the object stream is preceded by a comment indicating
      its object number and index.  This makes it very easy to find
      objects in object streams.
     </para>
    </listitem>
    <listitem>
     <para>
      All beginnings of objects, <literal>stream</literal> tokens,
      <literal>endstream</literal> tokens, and
      <literal>endobj</literal> tokens appear on lines by themselves.
      A blank line follows every <literal>endobj</literal> token.
     </para>
    </listitem>
    <listitem>
     <para>
      If there is a cross-reference stream, it is unfiltered.
     </para>
    </listitem>
    <listitem>
     <para>
      Page dictionaries and page content streams are marked with
      special comments that make them easy to find.
     </para>
    </listitem>
   </itemizedlist>
  </para>
  <para>
   When editing a QDF file, any edits can be made as long as the above
   constraints are maintained.  This means that you can freely edit a
   page's content without worrying about messing up the QDF file.  It
   is also possible to add new objects so long as those objects are
   added after the last object in the file or subsequent objects are
   renumbered.  If a QDF file has object streams in it, you can always
   add the new objects before the xref stream and then change the
   number of the xref stream, since nothing generally ever references
   it by number.
  </para>
  <para>
   It is not generally practical to remove objects from QDF files
   without messing up object numbering, but if you remove all
   references to an object, you can run qpdf on the file (after
   running <command>fix-qdf</command>), and qpdf will omit the
   now-orphaned object.
  </para>
  <para>
   When <command>fix-qdf</command> is run, it goes through the file
   and recomputes the following parts of the file:
   <itemizedlist>
    <listitem>
     <para>
      the <literal>/N</literal>, <literal>/W</literal>, and
      <literal>/First</literal> keys of all object stream dictionaries
     </para>
    </listitem>
    <listitem>
     <para>
      the pairs of numbers representing object numbers and offsets of
      objects in object streams
     </para>
    </listitem>
    <listitem>
     <para>
      all stream lengths
     </para>
    </listitem>
    <listitem>
     <para>
      the cross-reference table or cross-reference stream
     </para>
    </listitem>
    <listitem>
     <para>
      the offset to the cross-reference table or cross-reference
      stream following the <literal>startxref</literal> token
     </para>
    </listitem>
   </itemizedlist>
  </para>
 </chapter>
 <chapter id="ref.using-library">
  <title>Using the QPDF Library</title>
   <para>
    The source tree for the qpdf package has an
    <filename>examples</filename> directory that contains a few
    example programs.  The <filename>qpdf/qpdf.cc</filename> source
    file also serves as a useful example since it exercises almost all
    of the qpdf library's public interface.  The best source of
    documentation on the library itself is reading comments in
    <filename>include/qpdf/QPDF.hh</filename>,
    <filename>include/qpdf/QDFWriter.hh</filename>, and
    <filename>include/qpdf/QPDFObjectHandle.hh</filename>.
   </para>
   <para>
    All header files are installed in the <filename>include/qpdf</filename> directory.  It
    is recommend that you use <literal>#include
    &lt;qpdf/QPDF.hh&gt;</literal> rather than adding
    <filename>include/qpdf</filename> to your include path.
   </para>
   <para>
    When linking against the qpdf library, you may also need to
    specify <literal>-lpcre -lz</literal> on your link command.  If
    your system understands how to read libtool
    <filename>.la</filename> files, this may not be necessary.
   </para>
 </chapter>
 <chapter id="ref.design">
  <title>Design and Library Notes</title>
  <sect1 id="ref.design.intro">
   <title>Introduction</title>
   <para>
    This section was written prior to the implementation of the qpdf
    package and was subsequently modified to reflect the
    implementation.  In some cases, for purposes of explanation, it
    may differ slightly from the actual implementation.  As always,
    the source code and test suite are authoritative.  Even if there
    are some errors, this document should serve as a road map to
    understanding how this code works.
   </para>
   <para>
    In general, one should adhere strictly to a specification when
    writing but be liberal in reading.  This way, the product of our
    software will be accepted by the widest range of other programs,
    and we will accept the widest range of input files.  This library
    attempts to conform to that philosophy whenever possible but also
    aims to provide strict checking for people who want to validate
    PDF files.  If you don't want to see warnings and are trying to
    write something that is tolerant, you can call
    <literal>setSuppressWarnings(true)</literal>.  If you want to fail
    on the first error, you can call
    <literal>setAttemptRecovery(false)</literal>.  The default
    behavior is to generating warnings for recoverable problems.  Note
    that recovery will not always produce the desired results even if
    it is able to get through the file.  Unlike most other PDF files
    that produce generic warnings such as &ldquo;This file is
    damaged,&rdquo;, qpdf generally issues a detailed error message
    that would be most useful to a PDF developer.  This is by design
    as there seems to be a shortage of PDF validation tools out
    there.  (This was, in fact, one of the major motivations behind
    the initial creation of qpdf.)
   </para>
  </sect1>
  <sect1 id="ref.design-goals">
   <title>Design Goals</title>
   <para>
    The QPDF package includes support for reading and rewriting PDF
    files.  It aims to hide from the user details involving object
    locations, modified (appended) PDF files, the
    directness/indirectness of objects, and stream filters including
    encryption.  It does not aim to hide knowledge of the object
    hierarchy or content stream contents.  Put another way, a user of
    the qpdf library is expected to have knowledge about how PDF files
    work, but is not expected to have to keep track of bookkeeping
    details such as file positions.
   </para>
   <para>
    A user of the library never has to care whether an object is
    direct or indirect.  All access to objects deals with this
    transparently.  All memory management details are also handled by
    the library.
   </para>
   <para>
    The <classname>PointerHolder</classname> object is used internally
    by the library to deal with memory management.  This is basically
    a smart pointer object very similar in spirit to the Boost
    library's <classname>shared_ptr</classname> object, but predating
    it by several years.  This library also makes use of a technique
    for giving fine-grained access to methods in one class to other
    classes by using public subclasses with friends and only private
    members that in turn call private methods of the containing class.
    See <classname>QPDFObjectHandle::Factory</classname> as an
    example.
   </para>
   <para>
    The top-level qpdf class is <classname>QPDF</classname>.  A
    <classname>QPDF</classname> object represents a PDF file.  The
    library provides methods for both accessing and mutating PDF
    files.
   </para>
   <para>
    <classname>QPDFObject</classname> is the basic PDF Object class.
    It is an abstract base class from which are derived classes for
    each type of PDF object.  Clients do not interact with Objects
    directly but instead interact with
    <classname>QPDFObjectHandle</classname>.
   </para>
   <para>
    <classname>QPDFObjectHandle</classname> contains
    <classname>PointerHolder&lt;QPDFObject&gt;</classname> and
    includes accessor methods that are type-safe proxies to the
    methods of the derived object classes as well as methods for
    querying object types.  They can be passed around by value,
    copied, stored in containers, etc. with very low overhead.
    Instances of <classname>QPDFObjectHandle</classname> always
    contain a reference back to the <classname>QPDF</classname> object
    from which they were created.  A
    <classname>QPDFObjectHandle</classname> may be direct or indirect.
    If indirect, the <classname>QPDFObject</classname> the
    <classname>PointerHolder</classname> initially points to is a null
    pointer.  In this case, the first attempt to access the underlying
    <classname>QPDFObject</classname> will result in the
    <classname>QPDFObject</classname> being resolved via a call to the
    referenced <classname>QPDF</classname> instance.  This makes it
    essentially impossible to make coding errors in which certain
    things will work for some PDF files and not for others based on
    which objects are direct and which objects are indirect.
   </para>
   <para>
    There is no public interface for creating instances of
    QPDFObjectHandle.  They can be created only inside the QPDF
    library.  This is generally done through a call to the private
    method <function>QPDF::readObject</function> which uses
    <classname>QPDFTokenizer</classname> to read an indirect object at
    a given file position and return a
    <classname>QPDFObjectHandle</classname> that encapsulates it.
    There are also internal methods to create fabricated indirect
    objects from existing direct objects or to change an indirect
    object into a direct object, though these steps are not performed
    except to support rewriting.
   </para>
   <para>
    When the <classname>QPDF</classname> class creates a new object,
    it dynamically allocates the appropriate type of
    <classname>QPDFObject</classname> and immediately hands the
    pointer to an instance of <classname>QPDFObjectHandle</classname>.
    The parser reads a token from the current file position.  If the
    token is a not either a dictionary or array opener, an object is
    immediately constructed from the single token and the parser
    returns.  Otherwise, the parser is invoked recursively in a
    special mode in which it accumulates objects until it finds a
    balancing closer.  During this process, the
    &ldquo;<literal>R</literal>&rdquo; keyword is recognized and an
    indirect <classname>QPDFObjectHandle</classname> may be
    constructed.
   </para>
   <para>
    The <function>QPDF::resolve()</function> method, which is used to
    resolve an indirect object, may be invoked from the
    <classname>QPDFObjectHandle</classname> class.  It first checks a
    cache to see whether this object has already been read.  If not,
    it reads the object from the PDF file and caches it.  It the
    returns the resulting <classname>QPDFObjectHandle</classname>.
    The calling object handle then replaces its
    <classname>PointerHolder&lt;QDFObject&gt;</classname> with the one
    from the newly returned <classname>QPDFObjectHandle</classname>.
    In this way, only a single copy of any direct object need exist
    and clients can access objects transparently without knowing
    caring whether they are direct or indirect objects.  Additionally,
    no object is ever read from the file more than once.  That means
    that only the portions of the PDF file that are actually needed
    are ever read from the input file, thus allowing the qpdf package
    to take advantage of this important design goal of PDF files.
   </para>
   <para>
    If the requested object is inside of an object stream, the object
    stream itself is first read into memory.  Then the tokenizer reads
    objects from the memory stream based on the offset information
    stored in the stream.  Those individual objects are cached, after
    which the temporary buffer holding the object stream contents are
    discarded.  In this way, the first time an object in an object
    stream is requested, all objects in the stream are cached.
   </para>
   <para>
    An instance of <classname>QPDF</classname> is constructed by using
    the class's default constructor.  If desired, the
    <classname>QPDF</classname> object may be configured with various
    methods that change its default behavior.  Then the
    <function>QPDF::processFile()</function> method is passed the name
    of a PDF file, which permanently associates the file with that
    QPDF object.  A password may also be given for access to
    password-protected files.  QPDF does not enforce encryption
    parameters and will treat user and owner passwords equivalently.
    Either password may be used to access an encrypted file.
    <footnote>
     <para>
      As pointed out earlier, the intention is not for qpdf to be used
      to bypass security on files. but as any open source PDF consumer
      may be easily modified to bypass basic PDF document security,
      and qpdf offers may transformations that can do this as well,
      there seems to be little point in the added complexity of
      conditionally enforcing document security.
     </para>
    </footnote>
    <classname>QPDF</classname> will allow recovery of a user password
    given an owner password.  The input PDF file must be seekable.
    (Output files written by <classname>QPDFWriter</classname> need
    not be seekable, even when creating linearized files.)  During
    construction, <classname>QPDF</classname> validates the PDF file's
    header, and then reads the cross reference tables and trailer
    dictionaries.  The <classname>QPDF</classname> class keeps only
    the first trailer dictionary though it does read all of them so it
    can check the <literal>/Prev</literal> key.
    <classname>QPDF</classname> class users may request the root
    object and the trailer dictionary specifically.  The cross
    reference table is kept private.  Objects may then be requested by
    number of by walking the object tree.
   </para>
   <para>
    When a PDF file has a cross-reference stream instead of a
    cross-reference table and trailer, requesting the document's
    trailer dictionary returns the stream dictionary from the
    cross-reference stream instead.
   </para>
   <para>
    There are some convenience routines for very common operations
    such as walking the page tree and returning a vector of all page
    objects.  For full details, please see the header file
    <filename>QPDF.hh</filename>.
   </para>
   <para>
    The following example should clarify how
    <classname>QPDF</classname> processes a simple file.
    <itemizedlist>
     <listitem>
      <para>
       Client constructs <classname>QPDF</classname>
       <varname>pdf</varname> and calls
       <function>pdf.processFile("a.pdf");</function>.
      </para>
     </listitem>
     <listitem>
      <para>
       The <classname>QPDF</classname> class checks the beginning of
       <filename>a.pdf</filename> for
       <literal>%!PDF-1.[0-9]+</literal>.  It then reads the cross
       reference table mentioned at the end of the file, ensuring that
       it is looking before the last <literal>%%EOF</literal>.  After
       getting to <literal>trailer</literal> keyword, it invokes the
       parser.
      </para>
     </listitem>
     <listitem>
      <para>
       The parser sees &ldquo;<literal>&lt;&lt;</literal>&rdquo;, so
       it calls itself recursively in dictionary creation mode.
      </para>
     </listitem>
     <listitem>
      <para>
       In dictionary creation mode, the parser keeps accumulating
       objects until it encounters
       &ldquo;<literal>&gt;&gt;</literal>&rdquo;.  Each object that is
       read is pushed onto a stack.  If
       &ldquo;<literal>R</literal>&rdquo; is read, the last two
       objects on the stack are inspected.  If they are integers, they
       are popped off the stack and their values are used to construct
       an indirect object handle which is then pushed onto the stack.
       When &ldquo;<literal>&gt;&gt;</literal>&rdquo; is finally read,
       the stack is converted into a
       <classname>QPDF_Dictionary</classname> which is placed in a
       <classname>QPDFObjectHandle</classname> and returned.
      </para>
     </listitem>
     <listitem>
      <para>
       The resulting dictionary is saved as the trailer dictionary.
      </para>
     </listitem>
     <listitem>
      <para>
       The <literal>/Prev</literal> key is searched.  If present,
       <classname>QPDF</classname> seeks to that point and repeats
       except that the new trailer dictionary is not saved.  If
       <literal>/Prev</literal> is not present, the initial parsing
       process is complete.
      </para>
      <para>
       If there is an encryption dictionary, the document's encryption
       parameters are initialized.
      </para>
     </listitem>
     <listitem>
      <para>
       The client requests root object.  The
       <classname>QPDF</classname> class gets the value of root key
       from trailer dictionary and returns it.  It is an unresolved
       indirect <classname>QPDFObjectHandle</classname>.
      </para>
     </listitem>
     <listitem>
      <para>
       The client requests the <literal>/Pages</literal> key from root
       <classname>QPDFObjectHandle</classname>.  The
       <classname>QPDFObjectHandle</classname> notices that it is
       indirect so it asks <classname>QPDF</classname> to resolve it.
       <classname>QPDF</classname> looks in the object cache for an
       object with the root dictionary's object ID and generation
       number.  Upon not seeing it, it checks the cross reference
       table, gets the offset, and reads the object present at that
       offset.  It stores the result in the object cache and returns
       the cached result.  The calling
       <classname>QPDFObjectHandle</classname> replaces its object
       pointer with the one from the resolved
       <classname>QPDFObjectHandle</classname>, verifies that it a
       valid dictionary object, and returns the (unresolved indirect)
       <classname>QPDFObject</classname> handle to the top of the
       Pages hierarchy.
      </para>
      <para>
       As the client continues to request objects, the same process is
       followed for each new requested object.
      </para>
     </listitem>
    </itemizedlist>
   </para>
  </sect1>
  <sect1 id="ref.encryption">
   <title>Encryption</title>
   <para>
    Encryption is supported transparently by qpdf.  When opening a PDF
    file, if an encryption dictionary exists, the
    <classname>QPDF</classname> object processes this dictionary using
    the password (if any) provided.  The primary decryption key is
    computed and cached.  No further access is made to the encryption
    dictionary after that time.  When an object is read from a file,
    the object ID and generation of the object in which it is
    contained is always known.  Using this information along with the
    stored encryption key, all stream and string objects are
    transparently decrypted.  Raw encrypted objects are never stored
    in memory.  This way, nothing in the library ever has to know or
    care whether it is reading an encrypted file.
   </para>
   <para>
    An interface is also provided for writing encrypted streams and
    strings given an encryption key.  This is used by
    <classname>QPDFWriter</classname> when it rewrites encrypted
    files.
   </para>
  </sect1>
  <sect1 id="ref.rewriting">
   <title>Writing PDF Files</title>
   <para>
    The qpdf library supports file writing of
    <classname>QPDF</classname> objects to PDF files through the
    <classname>QPDFWriter</classname> class.  The
    <classname>QPDFWriter</classname> class has two writing modes: one
    for non-linearized files, and one for linearized files.  See <xref
    linkend="ref.linearization"/> for a description of linearization
    is implemented.  This section describes how we write
    non-linearized files including the creation of QDF files (see
    <xref linkend="ref.qdf"/>.
   </para>
   <para>
    This outline was written prior to implementation and is not
    exactly accurate, but it provides a correct &ldquo;notional&rdquo;
    idea of how writing works.  Look at the code in
    <classname>QPDFWriter</classname> for exact details.
    <itemizedlist>
     <listitem>
      <para>
       Initialize state:
       <itemizedlist>
        <listitem>
         <para>
          next object number = 1
         </para>
        </listitem>
        <listitem>
         <para>
          object queue = empty
         </para>
        </listitem>
        <listitem>
         <para>
          renumber table: old object id/generation to new id/0 = empty
         </para>
        </listitem>
        <listitem>
         <para>
          xref table: new id -> offset = empty
         </para>
        </listitem>
       </itemizedlist>
      </para>
     </listitem>
     <listitem>
      <para>
       Create a QPDF object from a file.
      </para>
     </listitem>
     <listitem>
      <para>
       Write header for new PDF file.
      </para>
     </listitem>
     <listitem>
      <para>
       Request the trailer dictionary.
      </para>
     </listitem>
     <listitem>
      <para>
       For each value that is an indirect object, grab the next object
       number (via an operation that returns and increments the
       number).  Map object to new number in renumber table.  Push
       object onto queue.
      </para>
     </listitem>
     <listitem>
      <para>
       While there are more objects on the queue:
       <itemizedlist>
        <listitem>
         <para>
          Pop queue.
         </para>
        </listitem>
        <listitem>
         <para>
          Look up object's new number <emphasis>n</emphasis> in the
          renumbering table.
         </para>
        </listitem>
        <listitem>
         <para>
          Store current offset into xref table.
         </para>
        </listitem>
        <listitem>
         <para>
          Write <literal><replaceable>n</replaceable> 0 obj</literal>.
         </para>
        </listitem>
        <listitem>
         <para>
          If object is null, whether direct or indirect, write out
          null, thus eliminating unresolvable indirect object
          references.
         </para>
        </listitem>
        <listitem>
         <para>
          If the object is a stream stream, write stream contents,
          piped through any filters as required, to a memory buffer.
          Use this buffer to determine the stream length.
         </para>
        </listitem>
        <listitem>
         <para>
          If object is not a stream, array, or dictionary, write out
          its contents.
         </para>
        </listitem>
        <listitem>
         <para>
          If object is an array or dictionary (including stream),
          traverse its elements (for array) or values (for
          dictionaries), handling recursive dictionaries and arrays,
          looking for indirect objects.  When an indirect object is
          found, if it is not resolvable, ignore.  (This case is
          handled when writing it out.)  Otherwise, look it up in the
          renumbering table.  If not found, grab the next available
          object number, assign to the referenced object in the
          renumbering table, and push the referenced object onto the
          queue.  As a special case, when writing out a stream
          dictionary, replace length, filters, and decode parameters
          as required.
         </para>
         <para>
          Write out dictionary or array, replacing any unresolvable
          indirect object references with null (pdf spec says
          reference to non-existent object is legal and resolves to
          null) and any resolvable ones with references to the
          renumbered objects.
         </para>
        </listitem>
        <listitem>
         <para>
          If the object is a stream, write
          <literal>stream\n</literal>, the stream contents (from the
          memory buffer), and <literal>\nendstream\n</literal>.
         </para>
        </listitem>
        <listitem>
         <para>
          When done, write <literal>endobj</literal>.
         </para>
        </listitem>
       </itemizedlist>
      </para>
     </listitem>
    </itemizedlist>
   </para>
   <para>
    Once we have finished the queue, all referenced objects will have
    been written out and all deleted objects or unreferenced objects
    will have been skipped.  The new cross-reference table will
    contain an offset for every new object number from 1 up to the
    number of objects written.  This can be used to write out a new
    xref table.  Finally we can write out the trailer dictionary with
    appropriately computed /ID (see spec, 8.3, File Identifiers), the
    cross reference table offset, and <literal>%%EOF</literal>.
   </para>
  </sect1>
  <sect1 id="ref.filtered-streams">
   <title>Filtered Streams</title>
   <para>
    Support for streams is implemented through the
    <classname>Pipeline</classname> interface which was designed for
    this package.
   </para>
   <para>
    When reading streams, create a series of
    <classname>Pipeline</classname> objects.  The
    <classname>Pipeline</classname> abstract base requires
    implementation <function>write()</function> and
    <function>finish()</function> and provides an implementation of
    <function>getNext()</function>.  Each pipeline object, upon
    receiving data, does whatever it is going to do and then writes
    the data (possibly modified) to its successor.  Alternatively, a
    pipeline may be an end-of-the-line pipeline that does something
    like store its output to a file or a memory buffer ignoring a
    successor.  For additional details, look at
    <filename>Pipeline.hh</filename>.
   </para>
   <para>
    <classname>QPDF</classname> can read raw or filtered streams.
    When reading a filtered stream, the <classname>QPDF</classname>
    class creates a <classname>Pipeline</classname> object for one of
    each appropriate filter object and chains them together.  The last
    filter should write to whatever type of output is required.  The
    <classname>QPDF</classname> class has an interface to write raw or
    filtered stream contents to a given pipeline.
   </para>
  </sect1>
 </chapter>
 <chapter id="ref.linearization">
  <title>Linearization</title>
  <para>
   This chapter describes how <classname>QPDF</classname> and
   <classname>QPDFWriter</classname> implement creation and processing
   of linearized PDFS.
  </para>
  <sect1 id="ref.linearization-strategy">
   <title>Basic Strategy for Linearization</title>
   <para>
    To avoid the incestuous problem of having the qpdf library
    validate its own linearized files, we have a special linearized
    file checking mode which can be invoked via <command>qpdf
    --check-linearization</command> (or <command>qpdf
    --check</command>).  This mode reads the linearization parameter
    dictionary and the hint streams and validates that object
    ordering, parameters, and hint stream contents are correct.  The
    validation code was first tested against linearized files created
    by external tools (Acrobat and pdlin) and then used to validate
    files created by <classname>QPDFWriter</classname> itself.
   </para>
  </sect1>
  <sect1 id="ref.linearized.preparation">
   <title>Preparing For Linearization</title>
   <para>
    Before creating a linearized PDF file from any other PDF file, the
    PDF file must be altered such that all page attributes are
    propagated down to the page level (and not inherited from parents
    in the <literal>/Pages</literal> tree).  We also have to know
    which objects refer to which other objects, being concerned with
    page boundaries and a few other cases.  We refer to this part of
    preparing the PDF file as <firstterm>optimization</firstterm>,
    discussed in <xref linkend="ref.optimization"/>.  Note the, in
    this context, the term <firstterm>optimization</firstterm> is a
    qpdf term, and the term <firstterm>linearization</firstterm> is a
    term from the PDF specification.  Do not be confused by the fact
    that many applications refer to linearization as optimization or
    web optimization.
   </para>
   <para>
    When creating linearized PDF files from optimized PDF files, there
    are really only a few issues that need to be dealt with:
    <itemizedlist>
     <listitem>
      <para>
       Creation of hints tables
      </para>
     </listitem>
     <listitem>
      <para>
       Placing objects in the correct order
      </para>
     </listitem>
     <listitem>
      <para>
       Filling in offsets and byte sizes
      </para>
     </listitem>
    </itemizedlist>
   </para>
  </sect1>
  <sect1 id="ref.optimization">
   <title>Optimization</title>
   <para>
    In order to perform various operations such as linearization and
    splitting files into pages, it is necessary to know which objects
    are referenced by which pages, page thumbnails, and root and
    trailer dictionary keys.  It is also necessary to ensure that all
    page-level attributes appear directly at the page level and are
    not inherited from parents in the pages tree.
   </para>
   <para>
    We refer to the process of enforcing these constraints as
    <firstterm>optimization</firstterm>.  As mentioned above, note
    that some applications refer to linearization as optimization.
    Although this optimization was initially motivated by the need to
    create linearized files, we are using these terms separately.
   </para>
   <para>
    PDF file optimization is implemented in the
    <filename>QPDF_optimization.cc</filename> source file.  That file
    is richly commented and serves as the primary reference for the
    optimization process.
   </para>
   <para>
    After optimization has been completed, the private member
    variables <varname>obj_user_to_objects</varname> and
    <varname>object_to_obj_users</varname> in
    <classname>QPDF</classname> have been populated.  Any object that
    has more than one value in the
    <varname>object_to_obj_users</varname> table is shared.  Any
    object that has exactly one value in the
    <varname>object_to_obj_users</varname> table is private.  To find
    all the private objects in a page or a trailer or root dictionary
    key, one merely has make this determination for each element in
    the <varname>obj_user_to_objects</varname> table for the given
    page or key.
   </para>
   <para>
    Note that pages and thumbnails have different object user types,
    so the above test on a page will not include objects referenced by
    the page's thumbnail dictionary and nothing else.
   </para>
  </sect1>
  <sect1 id="ref.linearization.writing">
   <title>Writing Linearized Files</title>
   <para>
    We will create files with only primary hint streams.  We will
    never write overflow hint streams.  (As of PDF version 1.4,
    Acrobat doesn't either, and they are never necessary.)  The hint
    streams contain offset information to objects that point to where
    they would be if the hint stream were not present.  This means
    that we have to calculate all object positions before we can
    generate and write the hint table.  This means that we have to
    generate the file in two passes.  To make this reliable,
    <classname>QPDFWriter</classname> in linearization mode invokes
    exactly the same code twice to write the file to a pipeline.
   </para>
   <para>
    In the first pass, the target pipeline is a count pipeline chained
    to a discard pipeline.  The count pipeline simply passes its data
    through to the next pipeline in the chain but can return the
    number of bytes passed through it at any intermediate point.  The
    discard pipeline is an end of line pipeline that just throws its
    data away.  The hint stream is not written and dummy values with
    adequate padding are stored in the first cross reference table,
    linearization parameter dictionary, and /Prev key of the first
    trailer dictionary.  All the offset, length, object renumbering
    information, and anything else we need for the second pass is
    stored.
   </para>
   <para>
    At the end of the first pass, this information is passed to the
    <classname>QPDF</classname> class which constructs a compressed
    hint stream in a memory buffer and returns it.
    <classname>QPDFWriter</classname> uses this information to write a
    complete hint stream object into a memory buffer.  At this point,
    the length of the hint stream is known.
   </para>
   <para>
    In the second pass, the end of the pipeline chain is a regular
    file instead of a discard pipeline, and we have known values for
    all the offsets and lengths that we didn't have in the first pass.
    We have to adjust offsets that appear after the start of the hint
    stream by the length of the hint stream, which is known.  Anything
    that is of variable length is padded, with the padding code
    surrounding any writing code that differs in the two passes.  This
    ensures that changes to the way things are represented never
    results in offsets that were gathered during the first pass
    becoming incorrect for the second pass.
   </para>
   <para>
    Using this strategy, we can write linearized files to a
    non-seekable output stream with only a single pass to disk or
    wherever the output is going.
   </para>
  </sect1>
  <sect1 id="ref.linearization-data">
   <title>Calculating Linearization Data</title>
   <para>
    Once a file is optimized, we have information about which objects
    access which other objects.  We can then process these tables to
    decide which part (as described in &ldquo;Linearized PDF Document
    Structure&rdquo; in the PDF specification) each object is
    contained within.  This tells us the exact order in which objects
    are written.  The <classname>QPDFWriter</classname> class asks for
    this information and enqueues objects for writing in the proper
    order.  It also turns on a check that causes an exception to be
    thrown if an object is encountered that has not already been
    queued.  (This could happen only if there were a bug in the
    traversal code used to calculate the linearization data.)
   </para>
  </sect1>
  <sect1 id="ref.linearization-issues">
   <title>Known Issues with Linearization</title>
   <para>
    There are a handful of known issues with this linearization code.
    These issues do not appear to impact the behavior of linearized
    files which still work as intended: it is possible for a web
    browser to begin to display them before they are fully
    downloaded.  In fact, it seems that various other programs that
    create linearized files have many of these same issues.  These
    items make reference to terminology used in the linearization
    appendix of the PDF specification.
    <itemizedlist>
     <listitem>
      <para>
       Thread Dictionary information keys appear in part 4 with the
       rest of Threads instead of in part 9.  Objects in part 9 are
       not grouped together functionally.
      </para>
     </listitem>
     <listitem>
      <para>
       We are not calculating numerators for shared object positions
       within content streams or interleaving them within content
       streams.
      </para>
     </listitem>
     <listitem>
      <para>
       We generate only page offset, shared object, and outline hint
       tables.  It would be relatively easy to add some additional
       tables.  We gather most of the information needed to create
       thumbnail hint tables.  There are comments in the code about
       this.
      </para>
     </listitem>
    </itemizedlist>
   </para>
  </sect1>
  <sect1 id="ref.linearization-debugging">
   <title>Debugging Note</title>
   <para>
    The <command>qpdf --show-linearization</command> command can show
    the complete contents of linearization hint streams.  To look at
    the raw data, you can extract the filtered contents of the
    linearization hint tables using <command>qpdf --show-object=n
    --filtered-stream-data</command>.  Then, to convert this into a
    bit stream (since linearization tables are bit streams written
    without regard to byte boundaries), you can pipe the resulting
    data through the following perl code:

    <programlisting>use bytes;
binmode STDIN;
undef $/;
my $a = &lt;STDIN&gt;;
my @ch = split(//, $a);
map { printf("%08b", ord($_)) } @ch;
print "\n";
</programlisting>
   </para>
  </sect1>
 </chapter>
 <chapter id="ref.object-and-xref-streams">
  <title>Object and Cross-Reference Streams</title>
  <para>
   This chapter provides information about the implementation of
   object stream and cross-reference stream support in qpdf.
  </para>
  <sect1 id="ref.object-streams">
   <title>Object Streams</title>
   <para>
    Object streams can contain any regular object except the
    following:
    <itemizedlist>
     <listitem>
      <para>
       stream objects
      </para>
     </listitem>
     <listitem>
      <para>
       objects with generation &gt; 0
      </para>
     </listitem>
     <listitem>
      <para>
       the encryption dictionary
      </para>
     </listitem>
     <listitem>
      <para>
       objects containing the /Length of another stream
      </para>
     </listitem>
    </itemizedlist>
    In addition, Adobe reader (at least as of version 8.0.0) appears
    to not be able to handle having the document catalog appear in an
    object stream if the file is encrypted, though this is not
    specifically disallowed by the specification.
   </para>
   <para>
    There are additional restrictions for linearized files.  See <xref
    linkend="ref.object-streams-linearization"/>for details.
   </para>
   <para>
    The PDF specification refers to objects in object streams as
    &ldquo;compressed objects&rdquo; regardless of whether the object
    stream is compressed.
   </para>
   <para>
    The generation number of every object in an object stream must be
    zero.  It is possible to delete and replace an object in an object
    stream with a regular object.
   </para>
   <para>
    The object stream dictionary has the following keys:
    <itemizedlist>
     <listitem>
      <para>
       <literal>/N</literal>: number of objects
      </para>
     </listitem>
     <listitem>
      <para>
       <literal>/First</literal>: byte offset of first object
      </para>
     </listitem>
     <listitem>
      <para>
       <literal>/Extends</literal>: indirect reference to stream that
       this extends
      </para>
     </listitem>
    </itemizedlist>
   </para>
   <para>
    Stream collections are formed with <literal>/Extends</literal>.
    They must form a directed acyclic graph.  These can be used for
    semantic information and are not meaningful to the PDF document's
    syntactic structure.  Although qpdf preserves stream collections,
    it never generates them and doesn't make use of this information
    in any way.
   </para>
   <para>
    The specification recommends limiting the number of objects in
    object stream for efficiency in reading and decoding.  Acrobat 6
    uses no more than objects per object stream for linearized files
    and no more 200 objects per stream for non-linearized files.
    <classname>QPDFWriter</classname>, in object stream generation
    mode, never puts more than 100 objects in an object stream.
   </para>
   <para>
    Object stream contents consists of <emphasis>N</emphasis> pairs of
    integers, each of which is the object number and the byte offset
    of the object relative to the first object in the stream, followed
    by the objects themselves, concatenated.
   </para>
  </sect1>
  <sect1 id="ref.xref-streams">
   <title>Cross-Reference Streams</title>
   <para>
    For non-hybrid files, the value following
    <literal>startxref</literal> is the byte offset to the xref stream
    rather than the word <literal>xref</literal>.
   </para>
   <para>
    For hybrid files (files containing both xref tables and
    cross-reference streams), the xref table's trailer dictionary
    contains the key <literal>/XRefStm</literal> whose value is the
    byte offset to a cross-reference stream that supplements the xref
    table.  A PDF 1.5-compliant application should read the xref table
    first.  Then it should replace any object that it has already seen
    with any defined in the xref stream.  Then it should follow any
    <literal>/Prev</literal> pointer in the original xref table's
    trailer dictionary.  The specification is not clear about what
    should be done, if anything, with a <literal>/Prev</literal>
    pointer in the xref stream referenced by an xref table.  The
    <classname>QPDF</classname> class ignores it, which is probably
    reasonable since, if this case were to appear for any sensible PDF
    file, the previous xref table would probably have a corresponding
    <literal>/XRefStm</literal> pointer of its own.  For example, if a
    hybrid file were appended, the appended section would have its own
    xref table and <literal>/XRefStm</literal>.  The appended xref
    table would point to the previous xref table which would point the
    <literal>/XRefStm</literal>, meaning that the new
    <literal>/XRefStm</literal> doesn't have to point to it.
   </para>
   <para>
    Since xref streams must be read very early, they may not be
    encrypted, and the may not contain indirect objects for keys
    required to read them, which are these:
    <itemizedlist>
     <listitem>
      <para>
       <literal>/Type</literal>: value <literal>/XRef</literal>
      </para>
     </listitem>
     <listitem>
      <para>
       <literal>/Size</literal>: value <emphasis>n+1</emphasis>: where
       <emphasis>n</emphasis> is highest object number (same as
       <literal>/Size</literal> in the trailer dictionary)
      </para>
     </listitem>
     <listitem>
      <para>
       <literal>/Index</literal> (optional): value
       <literal>[<replaceable>n count</replaceable> ...]</literal>
       used to determine which objects' information is stored in this
       stream.  The default is <literal>[0 /Size]</literal>.
      </para>
     </listitem>
     <listitem>
      <para>
       <literal>/Prev</literal>: value
       <replaceable>offset</replaceable>: byte offset of previous xref
       stream (same as <literal>/Prev</literal> in the trailer
       dictionary)
      </para>
     </listitem>
     <listitem>
      <para>
       <literal>/W [...]</literal>: sizes of each field in the xref
       table
      </para>
     </listitem>
    </itemizedlist>
   </para>
   <para>
    The other fields in the xref stream, which may be indirect if
    desired, are the union of those from the xref table's trailer
    dictionary.
   </para>
   <sect2 id="ref.xref-stream-data">
    <title>Cross-Reference Stream Data</title>
    <para>
     The stream data is binary and encoded in big-endian byte order.
     Entries are concatenated, and each entry has a length equal to
     the total of the entries in <literal>/W</literal> above.  Each
     entry consists of one or more fields, the first of which is the
     type of the field.  The number of bytes for each field is given
     by <literal>/W</literal> above.  A 0 in <literal>/W</literal>
     indicates that the field is omitted and has the default value.
     The default value for the field type is
     &ldquo;<literal>1</literal>&rdquo;.  All other default values are
     &ldquo;<literal>0</literal>&rdquo;.
    </para>
    <para>
     PDF 1.5 has three field types:
     <itemizedlist>
      <listitem>
       <para>
        0: for free objects.  Format: <literal>0 obj
        next-generation</literal>, same as the free table in a
        traditional cross-reference table
      </para>
     </listitem>
     <listitem>
      <para>
       1: regular non-compressed object.  Format: <literal>1 offset
       generation</literal>
      </para>
     </listitem>
     <listitem>
      <para>
       2: for objects in object streams.  Format: <literal>2
       object-stream-number index</literal>, the number of object
       stream containing the object and the index within the object
       stream of the object.
      </para>
     </listitem>
     </itemizedlist>
    </para>
    <para>
     It seems standard to have the first entry in the table be
     <literal>0 0 0</literal> instead of <literal>0 0 ffff</literal>
     if there are no deleted objects.
    </para>
   </sect2>
  </sect1>
  <sect1 id="ref.object-streams-linearization">
   <title>Implications for Linearized Files</title>
   <para>
    For linearized files, the linearization dictionary, document
    catalog, and page objects may not be contained in object streams.
   </para>
   <para>
    Objects stored within object streams are given the highest range
    of object numbers within the main and first-page cross-reference
    sections.
   </para>
   <para>
    It is okay to use cross-reference streams in place of regular xref
    tables.  There are on special considerations.
   </para>
   <para>
    Hint data refers to object streams themselves, not the objects in
    the streams.  Shared object references should also be made to the
    object streams.  There are no reference in any hint tables to the
    object numbers of compressed objects (objects within object
    streams).
   </para>
   <para>
    When numbering objects, all shared objects within both the first
    and second halves of the linearized files must be numbered
    consecutively after all normal uncompressed objects in that half.
   </para>
  </sect1>
  <sect1 id="ref.object-stream-implementation">
   <title>Implementation Notes</title>
   <para>
    There are three modes for writing object streams:
    <option>disable</option>, <option>preserve</option>, and
    <option>generate</option>.  In disable mode, we do not generate
    any object streams, and we also generate an xref table rather than
    xref streams.  This can be used to generate PDF files that are
    viewable with older readers.  In preserve mode, we write object
    streams such that written object streams contain the same objects
    and <literal>/Extends</literal> relationships as in the original
    file.  This is equal to disable if the file has no object streams.
    In generate, we create object streams ourselves by grouping
    objects that are allowed in object streams together in sets of no
    more than 100 objects.  We also ensure that the PDF version is at
    least 1.5 in generate mode, but we preserve the version header in
    the other modes.  The default is <option>preserve</option>.
   </para>
   <para>
    We do not support creation of hybrid files.  When we write files,
    even in preserve mode, we will lose any xref tables and merge any
    appended sections.
   </para>
  </sect1>
 </chapter>
</book>