QPDF Manual

For QPDF Version 6.0.0, November 10, 2015

Jay Berkenbilt

QPDF Manual: For QPDF Version 6.0.0, November 10, 2015

Jay Berkenbilt
Copyright 20052015 Jay Berkenbilt

Table of Contents
General Information ............................................................................................................................ iv
1. What is QPDF? ............................................................................................................................... 1
2. Building and Installing QPDF ............................................................................................................ 2
2.1. System Requirements ............................................................................................................. 2
2.2. Build Instructions .................................................................................................................. 2
3. Running QPDF ............................................................................................................................... 4
3.1. Basic Invocation ................................................................................................................... 4
3.2. Basic Options ....................................................................................................................... 4
3.3. Encryption Options ............................................................................................................... 5
3.4. Page Selection Options .......................................................................................................... 6
3.5. Advanced Transformation Options ........................................................................................... 8
3.6. Testing, Inspection, and Debugging Options ............................................................................ 10
4. QDF Mode ................................................................................................................................... 12
5. Using the QPDF Library ................................................................................................................. 14
6. Design and Library Notes ............................................................................................................... 15
6.1. Introduction ........................................................................................................................ 15
6.2. Design Goals ...................................................................................................................... 15
6.3. Casting Policy .................................................................................................................... 17
6.4. Encryption ......................................................................................................................... 18
6.5. Random Number Generation ................................................................................................. 19
6.6. Adding and Removing Pages ................................................................................................. 19
6.7. Reserving Object Numbers ................................................................................................... 19
6.8. Copying Objects From Other PDF Files .................................................................................. 20
6.9. Writing PDF Files ............................................................................................................... 20
6.10. Filtered Streams ................................................................................................................ 21
7. Linearization ................................................................................................................................. 22
7.1. Basic Strategy for Linearization ............................................................................................. 22
7.2. Preparing For Linearization ................................................................................................... 22
7.3. Optimization ....................................................................................................................... 22
7.4. Writing Linearized Files ....................................................................................................... 23
7.5. Calculating Linearization Data ............................................................................................... 23
7.6. Known Issues with Linearization ........................................................................................... 23
7.7. Debugging Note .................................................................................................................. 24
8. Object and Cross-Reference Streams ................................................................................................. 25
8.1. Object Streams .................................................................................................................... 25
8.2. Cross-Reference Streams ...................................................................................................... 25
8.2.1. Cross-Reference Stream Data ...................................................................................... 26
8.3. Implications for Linearized Files ............................................................................................ 26
8.4. Implementation Notes .......................................................................................................... 27
A. Release Notes ............................................................................................................................... 28
B. Upgrading from 2.0 to 2.1 .............................................................................................................. 38
C. Upgrading to 3.0 ........................................................................................................................... 39
D. Upgrading to 4.0 ........................................................................................................................... 40

iii

General Information
QPDF is a program that does structural, content-preserving transformations on PDF files. QPDF's website is located
at http://qpdf.sourceforge.net/. QPDF's source code is hosted on github at https://github.com/qpdf/qpdf.
QPDF has been released under the terms of Version 2.0 of the Artistic License [http://www.opensource.org/licenses/
artistic-license-2.0.php], a copy of which appears in the file Artistic-2.0 in the source distribution.
QPDF was originally created in 2001 and modified periodically between 2001 and 2005 during my employment at
Apex CoVantage [http://www.apexcovantage.com]. 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.

iv

Chapter 1. What is QPDF?

QPDF is a program that does structural, content-preserving transformations on PDF files. It could have been called
something like pdf-to-pdf. 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.
With QPDF, it is possible to copy objects from one PDF file into another and to manipulate the list of pages in a PDF
file. This makes it possible to merge and split PDF files. The QPDF library also makes it possible for you to create
PDF files from scratch. In this mode, you are responsible for supplying all the contents of the file, while the QPDF
library takes care off all the syntactical representation of the objects, creation of cross references tables and, if you
use them, object streams, encryption, linearization, and other syntactic details. You are still responsible for generating
PDF content on your own.
QPDF has been designed with very few external dependencies, and it is intentionally very lightweight. QPDF is not 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, many programs generate simple PDF files but can't
password-protect them, web-optimize them, or perform other transformations of that type.

Chapter 2. Building and Installing

QPDF
This chapter describes how to build and install qpdf. Please see also the README and INSTALL files in the source
distribution.

2.1. System Requirements

The qpdf package has relatively few external dependencies. In order to build qpdf, the following packages are required:
zlib: http://www.zlib.net/
pcre: http://www.pcre.org/
gnu make 3.81 or newer: http://www.gnu.org/software/make
perl version 5.8 or newer: http://www.perl.org/; required for fix-qdf and the test suite.
GNU diffutils (any version): http://www.gnu.org/software/diffutils/ 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 diff -u.
A C++ compiler that works well with STL and has the long long type. Most modern C++ compilers should fit
the bill fine. QPDF is tested with gcc and Microsoft Visual C++.
Part of qpdf's test suite does comparisons of the contents PDF files by converting them images and comparing the
images. The image comparison tests are disabled by default. Those tests are not required for determining correctness of
a qpdf build if you have not modified the code since the test suite also contains expected output files that are compared
literally. The image comparison tests provide an extra check to make sure that any content transformations don't break
the rendering of pages. Transformations that affect the content streams themselves are off by default and are only
provided to help developers look into the contents of PDF files. If you are making deep changes to the library that cause
changes in the contents of the files that qpdf generates, then you should enable the image comparison tests. Enable
them by running configure with the --enable-test-compare-images flag. If you enable this, the following additional
requirements are required by the test suite. Note that in no case are these items required to use qpdf.
libtiff: http://www.remotesensing.org/libtiff/
GhostScript version 8.60 or newer: http://www.ghostscript.com
If you do not enable this, then you do not need to have tiff and ghostscript.
If Adobe Reader is installed as acroread, 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.
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 (http://
downloads.sourceforge.net/docbook/). To build the PDF version of the documentation, you need Apache fop (http://
xml.apache.org/fop/) version 0.94 or higher.

2.2. Build Instructions

Building qpdf on UNIX is generally just a matter of running

Building and Installing QPDF

./configure
make
You can also run make check to run the test suite and make install to install. Please run ./configure --help for
options on what can be configured. You can also set the value of DESTDIR during installation to install to a temporary
location, as is common with many open source packages. Please see also the README and INSTALL files in the
source distribution.
Building on Windows is a little bit more complicated. For details, please see README-windows.txt in the source
distribution. You can also download a binary distribution for Windows. There is a port of qpdf to Visual C++ version 6
in the contrib area generously contributed by Jian Ma. This is also discussed in more detail in README-windows.txt.
There are some other things you can do with the build. Although qpdf uses autoconf, it does not use automake 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 Makefile.

Chapter 3. Running QPDF

This chapter describes how to run the qpdf program from the command line.

3.1. Basic Invocation

When running qpdf, the basic invocation is as follows:
qpdf [ options ] infilename [ outfilename ]
This converts PDF file infilename to PDF file outfilename. 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. In place of infilename, the parameter --empty may be specified. This
causes qpdf to use a dummy input file that contains zero pages. The only normal use case for using --empty would be if
you were going to add pages from another source, as discussed in Section 3.4, Page Selection Options, page 6.
outfilename does not have to be seekable, even when generating linearized files. Specifying -- as outfilename
means to write to standard output. However, you can't specify the same file as both the input and the output because
qpdf reads data from the input file as it writes to the output file.
Most options require an output file, but some testing or inspection commands do not. These are specifically noted.

3.2. Basic Options

The following options are the most common ones and perform commonly needed transformations.
--password=password
Specifies a password for accessing encrypted files.
--linearize
Causes generation of a linearized (web-optimized) output file.
--copy-encryption=file
Encrypt the file using the same encryption parameters, including user and owner password, as the specified file.
Use --encrypt-file-password to specify a password if one is needed to open this file. Note that copying the
encryption parameters from a file also copies the first half of /ID from the file since this is part of the encryption
parameters.
--encrypt-file-password=password
If the file specified with --copy-encryption requires a password, specify the password using this option. Note
that only one of the user or owner password is required. Both passwords will be preserved since QPDF does
not distinguish between the two passwords. It is possible to preserve encryption parameters, including the owner
password, from a file even if you don't know the file's owner password.
--encrypt options -Causes generation an encrypted output file. Please see Section 3.3, Encryption Options, page 5 for details
on how to specify encryption parameters.
--decrypt
Removes any encryption on the file. A password must be supplied if the file is password protected.
--pages options -Select specific pages from one or more input files. See Section 3.4, Page Selection Options, page 6 for
details on how to do page selection (splitting and merging).

Running QPDF

Password-protected files may be opened by specifying a password. By default, qpdf will preserve any encryption data
associated with a file. If --decrypt is specified, qpdf will attempt to remove any encryption information. If --encrypt
is specified, qpdf will replace the document's encryption parameters with whatever is specified.
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.
In all cases where qpdf allows specification of a password, care must be taken if the password contains characters
that fall outside of the 7-bit US-ASCII character range to ensure that the exact correct byte sequence is provided. It
is possible that a future version of qpdf may handle this more gracefully. For example, if a password was encrypted
using a password that was encoded in ISO-8859-1 and your terminal is configured to use UTF-8, the password you
supply may not work properly. There are various approaches to handling this. For example, if you are using Linux and
have the iconv executable (part of the ICU package) installed, you could pass --password=`echo password | iconv
-t iso-8859-1` to qpdf where password is a password specified in your terminal's locale. A detailed discussion of
this is out of scope for this manual, but just be aware of this issue if you have trouble with a password that contains
8-bit characters.

3.3. Encryption Options

To change the encryption parameters of a file, use the --encrypt flag. The syntax is
--encrypt user-password owner-password key-length [ restrictions ] -Note that -- terminates parsing of encryption flags and must be present even if no restrictions are present.
Either or both of the user password and the owner password may be empty strings.
The value for key-length may be 40, 128, or 256. The restriction flags are dependent upon key length. When no
additional restrictions are given, the default is to be fully permissive.
If key-length is 40, the following restriction options are available:
--print=[yn]
Determines whether or not to allow printing.
--modify=[yn]
Determines whether or not to allow document modification.
--extract=[yn]
Determines whether or not to allow text/image extraction.
--annotate=[yn]
Determines whether or not to allow comments and form fill-in and signing.
If key-length is 128, the following restriction options are available:
--accessibility=[yn]
Determines whether or not to allow accessibility to visually impaired.
--extract=[yn]
Determines whether or not to allow text/graphic extraction.
--print=print-opt
Controls printing access. print-opt may be one of the following:
full: allow full printing

Running QPDF

low: allow low-resolution printing only

none: disallow printing
--modify=modify-opt
Controls modify access. modify-opt may be one of the following, each of which implies all the options that
follow it:
all: allow full document modification
annotate: allow comment authoring and form operations
form: allow form field fill-in and signing
assembly: allow document assembly only
none: allow no modifications
--cleartext-metadata
If specified, any metadata stream in the document will be left unencrypted even if the rest of the document is
encrypted. This also forces the PDF version to be at least 1.5.
--use-aes=[yn]
If --use-aes=y is specified, AES encryption will be used instead of RC4 encryption. This forces the PDF version
to be at least 1.6.
--force-V4
Use of this option forces the /V and /R parameters in the document's encryption dictionary to be set to the value
4. As qpdf will automatically do this when required, there is no reason to ever use this option. It exists primarily
for use in testing qpdf itself. This option also forces the PDF version to be at least 1.5.
If key-length is 256, the minimum PDF version is 1.7 with extension level 8, and the AES-based encryption format
used is the PDF 2.0 encryption method supported by Acrobat X. the same options are available as with 128 bits with
the following exceptions:
--use-aes
This option is not available with 256-bit keys. AES is always used with 256-bit encryption keys.
--force-V4
This option is not available with 256 keys.
--force-R5
If specified, qpdf sets the minimum version to 1.7 at extension level 3 and writes the deprecated encryption format
used by Acrobat version IX. This option should not be used in practice to generate PDF files that will be in general
use, but it can be useful to generate files if you are trying to test proper support in another application for PDF
files encrypted in this way.
The default for each permission option is to be fully permissive.

3.4. Page Selection Options

Starting with qpdf 3.0, it is possible to split and merge PDF files by selecting pages from one or more input files.
Whatever file is given as the primary input file is used as the starting point, but its pages are replaced with pages as
specified.
--pages input-file [ --password=password ] [ page-range ] [ ... ] --

Running QPDF

Multiple input files may be specified. Each one is given as the name of the input file, an optional password (if required
to open the file), and the range of pages. Note that -- terminates parsing of page selection flags.
For each file that pages should be taken from, specify the file, a password needed to open the file (if any), and a page
range. The password needs to be given only once per file. If any of the input files are the same as the primary input
file or the file used to copy encryption parameters (if specified), you do not need to repeat the password here. The
same file can be repeated multiple times. If a file that is repeated has a password, the password only has to be given
the first time. All non-page data (info, outlines, page numbers, etc.) are taken from the primary input file. To discard
these, use --empty as the primary input.
Starting with qpdf 5.0.0, it is possible to omit the page range. If qpdf sees a value in the place where it expects a page
range and that value is not a valid range but is a valid file name, qpdf will implicitly use the range 1-z, meaning that
it will include all pages in the file. This makes it possible to easily combine all pages in a set of files with a command
like qpdf --empty out.pdf --pages *.pdf --.
It is not presently possible to specify the same page from the same file directly more than once, but you can make this
work by specifying two different paths to the same file (such as by putting ./ somewhere in the path). This can also be
used if you want to repeat a page from one of the input files in the output file. This may be made more convenient in
a future version of qpdf if there is enough demand for this feature.
The page range is a set of numbers separated by commas, ranges of numbers separated dashes, or combinations of
those. The character z represents the last page. Pages can appear in any order. Ranges can appear with a high number
followed by a low number, which causes the pages to appear in reverse. Repeating a number will cause an error, but
you can use the workaround discussed above should you really want to include the same page twice.
Example page ranges:
1,3,5-9,15-12: pages 1, 2, 3, 5, 6, 7, 8, 9, 15, 14, 13, and 12.
z-1: all pages in the document in reverse
Note that qpdf doesn't presently do anything special about other constructs in a PDF file that may know about pages,
so semantics of splitting and merging vary across features. For example, the document's outlines (bookmarks) point
to actual page objects, so if you select some pages and not others, bookmarks that point to pages that are in the output
file will work, and remaining bookmarks will not work. On the other hand, page labels (page numbers specified in the
file) are just sequential, so page labels will be messed up in the output file. A future version of qpdf may do a better
job at handling these issues. (Note that the qpdf library already contains all of the APIs required in order to implement
this in your own application if you need it.) In the mean time, you can always use --empty as the primary input file to
avoid copying all of that from the first file. For example, to take pages 1 through 5 from a infile.pdf while preserving
all metadata associated with that file, you could use
qpdf infile.pdf --pages infile.pdf 1-5 -- outfile.pdf
If you wanted pages 1 through 5 from infile.pdf but you wanted the rest of the metadata to be dropped, you could
instead run
qpdf --empty --pages infile.pdf 1-5 -- outfile.pdf
If you wanted to take pages 15 from file1.pdf and pages 1115 from file2.pdf in reverse, you would run
qpdf file1.pdf --pages file1.pdf 1-5 file2.pdf 15-11 -- outfile.pdf
If, for some reason, you wanted to take the first page of an encrypted file called encrypted.pdf with password pass
and repeat it twice in an output file, and if you wanted to drop metadata (like page numbers and outlines) but preserve
encryption, you would use

Running QPDF

qpdf --empty --copy-encryption=encrypted.pdf --encryption-file-password=pass

--pages encrypted.pdf --password=pass 1 ./encrypted.pdf --password=pass 1 -outfile.pdf
Note that we had to specify the password all three times because giving a password as --encryption-file-password
doesn't count for page selection, and as far as qpdf is concerned, encrypted.pdf and ./encrypted.pdf are separated files.
These are all corner cases that most users should hopefully never have to be bothered with.

3.5. Advanced Transformation Options

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:
--stream-data=option
Controls transformation of stream data. The value of option may be one of the following:
compress: recompress stream data when possible (default)
preserve: leave all stream data as is
uncompress: uncompress stream data when possible
--normalize-content=[yn]
Enables or disables normalization of content streams.
--suppress-recovery
Prevents qpdf from attempting to recover damaged files.
--object-streams=mode
Controls handling of object streams. The value of mode may be one of the following:
preserve: preserve original object streams (default)
disable: don't write any object streams
generate: use object streams wherever possible
--ignore-xref-streams
Tells qpdf to ignore any cross-reference streams.
--qdf
Turns on QDF mode. For additional information on QDF, please see Chapter 4, QDF Mode, page 12.
--min-version=version
Forces the PDF version of the output file to be at least version. In other words, if the input file has a lower
version than the specified version, the specified version will be used. If the input file has a higher version, the
input file's original version will be used. It is seldom necessary to use this option since qpdf will automatically
increase the version as needed when adding features that require newer PDF readers.
The version number may be expressed in the form major.minor.extension-level, in which case the
version is interpreted as major.minor at extension level extension-level. For example, version 1.7.8
represents version 1.7 at extension level 8. Note that minimal syntax checking is done on the command line.
--force-version=version
This option forces the PDF version to be the exact version specified even when the file may have content that
is not supported in that version. The version number is interpreted in the same way as with --min-version so

Running QPDF

that extension levels can be set. In some cases, forcing the output file's PDF version to be lower than that of the
input file will cause qpdf to disable certain features of the document. Specifically, 256-bit keys are disabled if
the version is less than 1.7 with extension level 8 (except R5 is disabled if less than 1.7 with extension level 3),
AES encryption is disabled if the version is less than 1.6, cleartext metadata and object streams are disabled if less
than 1.5, 128-bit encryption keys are disabled if less than 1.4, and all encryption is disabled if less than 1.3. Even
with these precautions, qpdf won't be able to do things like eliminate use of newer image compression schemes,
transparency groups, or other features that may have been added in more recent versions of PDF.
As a general rule, with the exception of big structural things like the use of object streams or AES encryption,
PDF viewers are supposed to ignore features in files that they don't support from newer versions. This means that
forcing the version to a lower version may make it possible to open your PDF file with an older version, though
bear in mind that some of the original document's functionality may be lost.
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 --streamdata=compress.
When --stream-data=preserve is specified, qpdf will never attempt to change the filtering of any stream data.
When --stream-data=uncompress is specified, qpdf will attempt to remove any non-lossy filters that it supports. This
includes /FlateDecode, /LZWDecode, /ASCII85Decode, and /ASCIIHexDecode. This can be very useful
for inspecting the contents of various streams.
When --normalize-content=y 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 production
PDF files.
Ordinarily, qpdf will attempt to recover from certain types of errors in PDF files. These include errors in the crossreference 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 --suppress-recovery option will prevent qpdf from attempting recovery. In this case,
it will fail on the first error that it encounters.
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: preserve, disable, and generate.
In preserve mode, the relationship to objects and the streams that contain them is preserved from the original file. In
disable 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 generate 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.
Ordinarily, qpdf reads cross-reference streams when they are present in a PDF file. If --ignore-xref-streams 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.
The --qdf 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

Running QPDF

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 fix-qdf 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 Chapter 4, QDF Mode, page 12.

3.6. Testing, Inspection, and Debugging

Options
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:
--deterministic-id
Causes generation of a deterministic value for /ID. This prevents use of timestamp and output file name
information in the /ID generation. Instead, at some slight additional runtime cost, the /ID field is generated to
include a digest of the significant parts of the content of the output PDF file. This means that a given qpdf operation
should generate the same /ID each time it is run, which can be useful when caching results or for generation of
some test data. Use of this flag is not compatible with creation of encrypted files.
--static-id
Causes generation of a fixed value for /ID. This is intended for testing only. Never use it for production files. If
you are trying to get the same /ID each time for a given file and you are not generating encrypted files, consider
using the --deterministic-id option.
--static-aes-iv
Causes use of a static initialization vector for AES-CBC. This is intended for testing only so that output files
can be reproducible. Never use it for production files. This option in particular is not secure since it significantly
weakens the encryption.
--no-original-object-ids
Suppresses inclusion of original object ID comments in QDF files. This can be useful when generating QDF files
for test purposes, particularly when comparing them to determine whether two PDF files have identical content.
--show-encryption
Shows document encryption parameters. Also shows the document's user password if the owner password is given.
--check-linearization
Checks file integrity and linearization status.
--show-linearization
Checks and displays all data in the linearization hint tables.
--show-xref
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.
--show-object=obj[,gen]
Show the contents of the given object. This is especially useful for inspecting objects that are inside of object
streams (also known as compressed objects).
--raw-stream-data
When used along with the --show-object option, if the object is a stream, shows the raw stream data instead of
object's contents.

10

Running QPDF

--filtered-stream-data
When used along with the --show-object 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.
--show-npages
Prints the number of pages in the input file on a line by itself. Since the number of pages appears by itself on a
line, this option can be useful for scripting if you need to know the number of pages in a file.
--show-pages
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.
--with-images
When used along with --show-pages, 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.)
--check
Checks file structure and well as encryption, linearization, and encoding of stream data. A file for which --check
reports no errors may still have errors in stream data content but should otherwise be structurally sound. If --check
any errors, qpdf will exit with a status of 2. There are some recoverable conditions that --check detects. These
are issued as warnings instead of errors. If qpdf finds no errors but finds warnings, it will exit with a status of
3 (as of version 2.0.4).
The --raw-stream-data and --filtered-stream-data options are ignored unless --show-object 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.
If --filtered-stream-data is given and --normalize-content=y 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 unusable results.

11

Chapter 4. QDF Mode

In QDF mode, qpdf creates PDF files in what we call QDF form. A PDF file in QDF form, sometimes called a QDF
file, is a completely valid PDF file that has %QDF-1.0 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).
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 fix-qdf program, distributed with
qpdf, is able to restore edited files to a correct state. The fix-qdf program takes no command-line arguments. It reads
a possibly edited QDF file from standard input and writes a repaired file to standard output.
The following attributes characterize a QDF file:
All objects appear in numerical order in the PDF file, including when objects appear in object streams.
Objects are printed in an easy-to-read format, and all line endings are normalized to UNIX line endings.
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).
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.
If the PDF file contains object streams, if object stream n contains k objects, those objects are numbered from n
+1 through n+k, 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.
All beginnings of objects, stream tokens, endstream tokens, and endobj tokens appear on lines by themselves.
A blank line follows every endobj token.
If there is a cross-reference stream, it is unfiltered.
Page dictionaries and page content streams are marked with special comments that make them easy to find.
Comments precede each object indicating the object number of the corresponding object in the original file.
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.
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 fix-qdf), and qpdf will omit the now-orphaned
object.
When fix-qdf is run, it goes through the file and recomputes the following parts of the file:
the /N, /W, and /First keys of all object stream dictionaries

12

QDF Mode

the pairs of numbers representing object numbers and offsets of objects in object streams
all stream lengths
the cross-reference table or cross-reference stream
the offset to the cross-reference table or cross-reference stream following the startxref token

13

Chapter 5. Using the QPDF Library

The source tree for the qpdf package has an examples directory that contains a few example programs. The qpdf/
qpdf.cc 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 include/qpdf/QPDF.hh, include/qpdf/
QDFWriter.hh, and include/qpdf/QPDFObjectHandle.hh.
All header files are installed in the include/qpdf directory. It is recommend that you use #include <qpdf/
QPDF.hh> rather than adding include/qpdf to your include path.
When linking against the qpdf static library, you may also need to specify -lpcre -lz on your link command. If
your system understands how to read libtool .la files, this may not be necessary.
The qpdf library is safe to use in a multithreaded program, but no individual QPDF object instance (including QPDF,
QPDFObjectHandle, or QPDFWriter) can be used in more than one thread at a time. Multiple threads may
simultaneously work with different instances of these and all other QPDF objects.

14

Chapter 6. Design and Library Notes

6.1. Introduction
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.
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 setSuppressWarnings(true). If you want to fail on the first error, you can call
setAttemptRecovery(false). 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 This file is damaged,, 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.)

6.2. Design Goals

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.
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.
The PointerHolder 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 shared_ptr 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 QPDFObjectHandle::Factory as an example.
The top-level qpdf class is QPDF. A QPDF object represents a PDF file. The library provides methods for both
accessing and mutating PDF files.
QPDFObject 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 QPDFObjectHandle.
QPDFObjectHandle contains PointerHolder<QPDFObject> 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 QPDFObjectHandle always
contain a reference back to the QPDF object from which they were created. A QPDFObjectHandle may be direct
or indirect. If indirect, the QPDFObject the PointerHolder initially points to is a null pointer. In this case, the
first attempt to access the underlying QPDFObject will result in the QPDFObject being resolved via a call to the
referenced QPDF 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.

15

Design and Library Notes

Instances of QPDFObjectHandle can be directly created and modified using static factory methods in the
QPDFObjectHandle class. There are factory methods for each type of object as well as a convenience method
QPDFObjectHandle::parse that creates an object from a string representation of the object. Existing instances of
QPDFObjectHandle can also be modified in several ways. See comments in QPDFObjectHandle.hh for details.
When the QPDF class creates a new object, it dynamically allocates the appropriate type of QPDFObject and
immediately hands the pointer to an instance of QPDFObjectHandle. 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 R keyword is recognized and an indirect
QPDFObjectHandle may be constructed.
The QPDF::resolve() method, which is used to resolve an indirect object, may be invoked from the
QPDFObjectHandle 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 QPDFObjectHandle. The calling object handle
then replaces its PointerHolder<QDFObject> with the one from the newly returned QPDFObjectHandle. 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.
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.
An instance of QPDF is constructed by using the class's default constructor. If desired, the QPDF object may be
configured with various methods that change its default behavior. Then the QPDF::processFile() 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. 1 QPDF will allow recovery of a
user password given an owner password. The input PDF file must be seekable. (Output files written by QPDFWriter
need not be seekable, even when creating linearized files.) During construction, QPDF validates the PDF file's header,
and then reads the cross reference tables and trailer dictionaries. The QPDF class keeps only the first trailer dictionary
though it does read all of them so it can check the /Prev key. QPDF 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.
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.
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 QPDF.hh.
The following example should clarify how QPDF processes a simple file.
Client constructs QPDF pdf and calls pdf.processFile("a.pdf");.
The QPDF class checks the beginning of a.pdf for %!PDF-1.[0-9]+. It then reads the cross reference table
mentioned at the end of the file, ensuring that it is looking before the last %%EOF. After getting to trailer
keyword, it invokes the parser.
1

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.

16

Design and Library Notes

The parser sees <<, so it calls itself recursively in dictionary creation mode.
In dictionary creation mode, the parser keeps accumulating objects until it encounters >>. Each object that is
read is pushed onto a stack. If R 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 >> is finally read, the stack is converted into a QPDF_Dictionary which is placed in a
QPDFObjectHandle and returned.
The resulting dictionary is saved as the trailer dictionary.
The /Prev key is searched. If present, QPDF seeks to that point and repeats except that the new trailer dictionary
is not saved. If /Prev is not present, the initial parsing process is complete.
If there is an encryption dictionary, the document's encryption parameters are initialized.
The client requests root object. The QPDF class gets the value of root key from trailer dictionary and returns it. It
is an unresolved indirect QPDFObjectHandle.
The client requests the /Pages key from root QPDFObjectHandle. The QPDFObjectHandle notices that it
is indirect so it asks QPDF to resolve it. QPDF 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
QPDFObjectHandle replaces its object pointer with the one from the resolved QPDFObjectHandle, verifies
that it a valid dictionary object, and returns the (unresolved indirect) QPDFObject handle to the top of the Pages
hierarchy.
As the client continues to request objects, the same process is followed for each new requested object.

6.3. Casting Policy

This section describes the casting policy followed by qpdf's implementation. This is no concern to qpdf's end users
and largely of no concern to people writing code that uses qpdf, but it could be of interest to people who are porting
qpdf to a new platform or who are making modifications to the code.
The C++ code in qpdf is free of old-style casts except where unavoidable (e.g. where the old-style cast is in a macro
provided by a third-party header file). When there is a need for a cast, it is handled, in order of preference, by rewriting
the code to avoid the need for a cast, calling const_cast, calling static_cast, calling reinterpret_cast, or calling some
combination of the above. As a last resort, a compiler-specific #pragma may be used to suppress a warning that we
don't want to fix. Examples may include suppressing warnings about the use of old-style casts in code that is shared
between C and C++ code.
The casting policy explicitly prohibits casting between integer sizes for no purpose other than to quiet a compiler
warning when there is no reasonable chance of a problem resulting. The reason for this exclusion is that the practice
of adding these additional casts precludes future use of additional compiler warnings as a tool for making future
improvements to this aspect of the code, and it also damages the readability of the code.
There are a few significant areas where casting is common in the qpdf sources or where casting would be required to
quiet higher levels of compiler warnings but is omitted at present:
char vs. unsigned char. For historical reasons, there are a lot of places in qpdf's internals that deal with
unsigned char, which means that a lot of casting is required to interoperate with standard library calls and
std::string. In retrospect, qpdf should have probably used regular (signed) char and char* everywhere and
just cast to unsigned char when needed, but it's too late to make that change now. There are reinterpret_cast
calls to go between char* and unsigned char*, and there are static_cast calls to go between char and
unsigned char. These should always be safe.

17

Design and Library Notes

Non-const unsigned char* used in the Pipeline interface. The pipeline interface has a write call that uses
unsigned char* without a const qualifier. The main reason for this is to support pipelines that make calls to
third-party libraries, such as zlib, that don't include const in their interfaces. Unfortunately, there are many places
in the code where it is desirable to have const char* with pipelines. None of the pipeline implementations in
qpdf currently modify the data passed to write, and doing so would be counter to the intent of Pipeline, but there
is nothing in the code to prevent this from being done. There are places in the code where const_cast is used to
remove the const-ness of pointers going into Pipelines. This could theoretically be unsafe, but there is adequate
testing to assert that it is safe and will remain safe in qpdf's code.
size_t vs. qpdf_offset_t. This is pretty much unavoidable since sizes are unsigned types and offsets are
signed types. Whenever it is necessary to seek by an amount given by a size_t, it becomes necessary to mix and
match between size_t and qpdf_offset_t. Additionally, qpdf sometimes treats memory buffers like files (as
with BufferInputSource, and those seek interfaces have to be consistent with file-based input sources. Neither
gcc nor MSVC give warnings for this case by default, but both have warning flags that can enable this. (MSVC:
/W14267 or /W3, which also enables some additional warnings that we ignore; gcc: -Wconversion -Wsignconversion). This could matter for files whose sizes are larger than 263 bytes, but it is reasonable to expect that a
world where such files are common would also have larger size_t and qpdf_offset_t types in it. On most
64-bit systems at the time of this writing (the release of version 4.1.0 of qpdf), both size_t and qpdf_offset_t
are 64-bit integer types, while on many current 32-bit systems, size_t is a 32-bit type while qpdf_offset_t is
a 64-bit type. I am not aware of any cases where 32-bit systems that have size_t smaller than qpdf_offset_t
could run into problems. Although I can't conclusively rule out the possibility of such problems existing, I suspect
any cases would be pretty contrived. In the event that someone should produce a file that qpdf can't handle because
of what is suspected to be issues involving the handling of size_t vs. qpdf_offset_t (such files may behave
properly on 64-bit systems but not on 32-bit systems because they have very large embedded files or streams, for
example), the above mentioned warning flags could be enabled and all those implicit conversions could be carefully
scrutinized. (I have already gone through that exercise once in adding support for files larger than 4 GB in size.) I
continue to be committed to supporting large files on 32-bit systems, but I would not go to any lengths to support
corner cases involving large embedded files or large streams that work on 64-bit systems but not on 32-bit systems
because of size_t being too small. It is reasonable to assume that anyone working with such files would be using
a 64-bit system anyway since many 32-bit applications would have similar difficulties.
size_t vs. int or long. There are some cases where size_t and int or long or size_t and unsigned
int or unsigned long are used interchangeably. These cases occur when working with very small amounts of
memory, such as with the bit readers (where we're working with just a few bytes at a time), some cases of strlen, and
a few other cases. I have scrutinized all of these cases and determined them to be safe, but there is no mechanism in
the code to ensure that new unsafe conversions between int and size_t aren't introduced short of good testing
and strong awareness of the issues. Again, if any such bugs are suspected in the future, enabling the additional
warning flags and scrutinizing the warnings would be in order.
To be clear, I believe qpdf to be well-behaved with respect to sizes and offsets, and qpdf's test suite includes actual
generation and full processing of files larger than 4 GB in size. The issues raised here are largely academic and should
not in any way be interpreted to mean that qpdf has practical problems involving sloppiness with integer types. I also
believe that appropriate measures have been taken in the code to avoid problems with signed vs. unsigned integers
from resulting in memory overwrites or other issues with potential security implications, though there are never any
absolute guarantees.

6.4. Encryption
Encryption is supported transparently by qpdf. When opening a PDF file, if an encryption dictionary exists, the QPDF
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.

18

Design and Library Notes

An interface is also provided for writing encrypted streams and strings given an encryption key. This is used by
QPDFWriter when it rewrites encrypted files.
When copying encrypted files, unless otherwise directed, qpdf will preserve any encryption in force in the original
file. qpdf can do this with either the user or the owner password. There is no difference in capability based on which
password is used. When 40 or 128 bit encryption keys are used, the user password can be recovered with the owner
password. With 256 keys, the user and owner passwords are used independently to encrypt the actual encryption key,
so while either can be used, the owner password can no longer be used to recover the user password.
Starting with version 4.0.0, qpdf can read files that are not encrypted but that contain encrypted attachments, but it
cannot write such files. qpdf also requires the password to be specified in order to open the file, not just to extract
attachments, since once the file is open, all decryption is handled transparently. When copying files like this while
preserving encryption, qpdf will apply the file's encryption to everything in the file, not just to the attachments. When
decrypting the file, qpdf will decrypt the attachments. In general, when copying PDF files with multiple encryption
formats, qpdf will choose the newest format. The only exception to this is that clear-text metadata will be preserved
as clear-text if it is that way in the original file.

6.5. Random Number Generation

QPDF generates random numbers to support generation of encrypted data. Versions prior to 5.0.1 used random or rand
from stdlib to generate random numbers. Version 5.0.1, if available, used operating system-provided secure random
number generation instead, enabling use of stdlib random number generation only if enabled by a compile-time option.
Starting in version 5.1.0, use of insecure random numbers was disabled unless enabled at compile time. Starting in
version 5.1.0, it is also possible for you to disable use of OS-provided secure random numbers. This is especially
useful on Windows if you want to avoid a dependency on Microsoft's cryptography API. In this case, you must provide
your own random data provider. Regardless of how you compile qpdf, starting in version 5.1.0, it is possible for you
to provide your own random data provider at runtime. This would enable you to use some software-based secure
pseudorandom number generator and to avoid use of whatever the operating system provides. For details on how to
do this, please refer to the top-level README file in the source distribution and to comments in QUtil.hh.

6.6. Adding and Removing Pages

While qpdf's API has supported adding and modifying objects for some time, version 3.0 introduces specific methods
for adding and removing pages. These are largely convenience routines that handle two tricky issues: pushing
inheritable resources from the /Pages tree down to individual pages and manipulation of the /Pages tree itself. For
details, see addPage and surrounding methods in QPDF.hh.

6.7. Reserving Object Numbers

Version 3.0 of qpdf introduced the concept of reserved objects. These are seldom needed for ordinary operations, but
there are cases in which you may want to add a series of indirect objects with references to each other to a QPDF object.
This causes a problem because you can't determine the object ID that a new indirect object will have until you add it to
the QPDF object with QPDF::makeIndirectObject. The only way to add two mutually referential objects to a QPDF
object prior to version 3.0 would be to add the new objects first and then make them refer to each other after adding
them. Now it is possible to create a reserved object using QPDFObjectHandle::newReserved. This is an indirect object
that stays unresolved even if it is queried for its type. So now, if you want to create a set of mutually referential
objects, you can create reservations for each one of them and use those reservations to construct the references. When
finished, you can call QPDF::replaceReserved to replace the reserved objects with the real ones. This functionality
will never be needed by most applications, but it is used internally by QPDF when copying objects from other PDF
files, as discussed in Section 6.8, Copying Objects From Other PDF Files, page 20. For an example of how to
use reserved objects, search for newReserved in test_driver.cc in qpdf's sources.

19

Design and Library Notes

6.8. Copying Objects From Other PDF Files

Version 3.0 of qpdf introduced the ability to copy objects into a QPDF object from a different QPDF object, which
we refer to as foreign objects. This allows arbitrary merging of PDF files. The qpdf command-line tool provides
limited support for basic page selection, including merging in pages from other files, but the library's API makes
it possible to implement arbitrarily complex merging operations. The main method for copying foreign objects is
QPDF::copyForeignObject. This takes an indirect object from another QPDF and copies it recursively into this object
while preserving all object structure, including circular references. This means you can add a direct object that you
create from scratch to a QPDF object with QPDF::makeIndirectObject, and you can add an indirect object from
another file with QPDF::copyForeignObject. The fact that QPDF::makeIndirectObject does not automatically detect
a foreign object and copy it is an explicit design decision. Copying a foreign object seems like a sufficiently significant
thing to do that it should be done explicitly.
The other way to copy foreign objects is by passing a page from one QPDF to another by calling QPDF::addPage.
In contrast to QPDF::makeIndirectObject, this method automatically distinguishes between indirect objects in the
current file, foreign objects, and direct objects.

6.9. Writing PDF Files

The qpdf library supports file writing of QPDF objects to PDF files through the QPDFWriter class. The
QPDFWriter class has two writing modes: one for non-linearized files, and one for linearized files. See Chapter 7,
Linearization, page 22 for a description of linearization is implemented. This section describes how we write nonlinearized files including the creation of QDF files (see Chapter 4, QDF Mode, page 12.
This outline was written prior to implementation and is not exactly accurate, but it provides a correct notional idea
of how writing works. Look at the code in QPDFWriter for exact details.
Initialize state:
next object number = 1
object queue = empty
renumber table: old object id/generation to new id/0 = empty
xref table: new id -> offset = empty
Create a QPDF object from a file.
Write header for new PDF file.
Request the trailer dictionary.
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.
While there are more objects on the queue:
Pop queue.
Look up object's new number n in the renumbering table.
Store current offset into xref table.
Write n 0 obj.

20

Design and Library Notes

If object is null, whether direct or indirect, write out null, thus eliminating unresolvable indirect object references.
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.
If object is not a stream, array, or dictionary, write out its contents.
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.
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.
If the object is a stream, write stream\n, the stream contents (from the memory buffer), and \nendstream\n.
When done, write endobj.
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 %%EOF.

6.10. Filtered Streams

Support for streams is implemented through the Pipeline interface which was designed for this package.
When reading streams, create a series of Pipeline objects. The Pipeline abstract base requires implementation write()
and finish() and provides an implementation of getNext(). 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 endof-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 Pipeline.hh.
QPDF can read raw or filtered streams. When reading a filtered stream, the QPDF class creates a Pipeline 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 QPDF class has an interface to write raw or filtered stream contents to a given pipeline.

21

Chapter 7. Linearization
This chapter describes how QPDF and QPDFWriter implement creation and processing of linearized PDFS.

7.1. Basic Strategy for Linearization

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 qpdf --check-linearization (or qpdf --check). 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 QPDFWriter itself.

7.2. Preparing For Linearization

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 /Pages 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 optimization, discussed in Section 7.3, Optimization, page 22. Note the, in
this context, the term optimization is a qpdf term, and the term linearization 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.
When creating linearized PDF files from optimized PDF files, there are really only a few issues that need to be dealt
with:
Creation of hints tables
Placing objects in the correct order
Filling in offsets and byte sizes

7.3. Optimization
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.
We refer to the process of enforcing these constraints as optimization. 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.
PDF file optimization is implemented in the QPDF_optimization.cc source file. That file is richly commented and
serves as the primary reference for the optimization process.
After optimization has been completed, the private member variables obj_user_to_objects and object_to_obj_users in
QPDF have been populated. Any object that has more than one value in the object_to_obj_users table is shared. Any
object that has exactly one value in the object_to_obj_users 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 obj_user_to_objects
table for the given page or key.
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.

22

Linearization

7.4. Writing Linearized Files

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, QPDFWriter in linearization mode invokes exactly the same code twice to write the file to a pipeline.
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.
At the end of the first pass, this information is passed to the QPDF class which constructs a compressed hint stream
in a memory buffer and returns it. QPDFWriter 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.
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.
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.

7.5. Calculating Linearization Data

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 Linearized PDF Document Structure in the PDF specification)
each object is contained within. This tells us the exact order in which objects are written. The QPDFWriter 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.)

7.6. Known Issues with Linearization

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.
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.
We are not calculating numerators for shared object positions within content streams or interleaving them within
content streams.
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.

23

Linearization

7.7. Debugging Note

The qpdf --show-linearization 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 qpdf --show-object=n --filteredstream-data. 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:
use bytes;
binmode STDIN;
undef $/;
my $a = <STDIN>;
my @ch = split(//, $a);
map { printf("%08b", ord($_)) } @ch;
print "\n";

24

Chapter 8. Object and Cross-Reference

Streams
This chapter provides information about the implementation of object stream and cross-reference stream support in
qpdf.

8.1. Object Streams

Object streams can contain any regular object except the following:
stream objects
objects with generation > 0
the encryption dictionary
objects containing the /Length of another stream
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.
There are additional restrictions for linearized files. See Section 8.3, Implications for Linearized Files, page 26for
details.
The PDF specification refers to objects in object streams as compressed objects regardless of whether the object
stream is compressed.
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.
The object stream dictionary has the following keys:
/N: number of objects
/First: byte offset of first object
/Extends: indirect reference to stream that this extends
Stream collections are formed with /Extends. 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.
The specification recommends limiting the number of objects in object stream for efficiency in reading and decoding.
Acrobat 6 uses no more than 100 objects per object stream for linearized files and no more 200 objects per stream for
non-linearized files. QPDFWriter, in object stream generation mode, never puts more than 100 objects in an object
stream.
Object stream contents consists of N 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.

8.2. Cross-Reference Streams

For non-hybrid files, the value following startxref is the byte offset to the xref stream rather than the word xref.

25

Object and Cross-Reference Streams

For hybrid files (files containing both xref tables and cross-reference streams), the xref table's trailer dictionary contains
the key /XRefStm 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 /Prev pointer in the original xref table's trailer
dictionary. The specification is not clear about what should be done, if anything, with a /Prev pointer in the xref
stream referenced by an xref table. The QPDF 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 /XRefStm pointer
of its own. For example, if a hybrid file were appended, the appended section would have its own xref table and /
XRefStm. The appended xref table would point to the previous xref table which would point the /XRefStm, meaning
that the new /XRefStm doesn't have to point to it.
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:
/Type: value /XRef
/Size: value n+1: where n is highest object number (same as /Size in the trailer dictionary)
/Index (optional): value [n count ...] used to determine which objects' information is stored in this stream.
The default is [0 /Size].
/Prev: value offset: byte offset of previous xref stream (same as /Prev in the trailer dictionary)
/W [...]: sizes of each field in the xref table
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.

8.2.1. Cross-Reference Stream Data

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 /W 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 /W above. A 0 in /W indicates that the field is omitted and
has the default value. The default value for the field type is 1. All other default values are 0.
PDF 1.5 has three field types:
0: for free objects. Format: 0 obj next-generation, same as the free table in a traditional cross-reference table
1: regular non-compressed object. Format: 1 offset generation
2: for objects in object streams. Format: 2 object-stream-number index, the number of object stream
containing the object and the index within the object stream of the object.
It seems standard to have the first entry in the table be 0 0 0 instead of 0 0 ffff if there are no deleted objects.

8.3. Implications for Linearized Files

For linearized files, the linearization dictionary, document catalog, and page objects may not be contained in object
streams.
Objects stored within object streams are given the highest range of object numbers within the main and first-page
cross-reference sections.
It is okay to use cross-reference streams in place of regular xref tables. There are on special considerations.

26

Object and Cross-Reference Streams

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).
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.

8.4. Implementation Notes

There are three modes for writing object streams: disable, preserve, and generate. 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 /Extends 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 preserve.
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.

27

Appendix A. Release Notes

For a detailed list of changes, please see the file ChangeLog in the source distribution.
6.0.0: November 10, 2015
Implement --deterministic-id command-line option and QPDFWriter::setDeterministicID as well as C API
function qpdf_set_deterministic_ID for generating a deterministic ID for non-encrypted files. When this option
is selected, the ID of the file depends on the contents of the output file, and not on transient items such as the
timestamp or output file name.
Make qpdf more tolerant of files whose xref table entries are not the correct length.
5.1.3: May 24, 2015
Bug fix: fix-qdf was not properly handling files that contained object streams with more than 255 objects in
them.
Bug fix: qpdf was not properly initializing Microsoft's secure crypto provider on fresh Windows installations
that had not had any keys created yet.
Fix a few errors found by Gynvael Coldwind and Mateusz Jurczyk of the Google Security Team. Please see
the ChangeLog for details.
Properly handle pages that have no contents at all. There were many cases in which qpdf handled this fine, but
a few methods blindly obtained page contents with handling the possibility that there were no contents.
Make qpdf more robust for a few more kinds of problems that may occur in invalid PDF files.
5.1.2: June 7, 2014
Bug fix: linearizing files could create a corrupted output file under extremely unlikely file size circumstances.
See ChangeLog for details. The odds of getting hit by this are very low, though one person did.
Bug fix: qpdf would fail to write files that had streams with decode parameters referencing other streams.
New example program: pdf-split-pages: efficiently split PDF files into individual pages. The example program
does this more efficiently than using qpdf --pages to do it.
Packaging fix: Visual C++ binaries did not support Windows XP. This has been rectified by updating the
compilers used to generate the release binaries.
5.1.1: January 14, 2014
Performance fix: copying foreign objects could be very slow with certain types of files. This was most likely
to be visible during page splitting and was due to traversing the same objects multiple times in some cases.
5.1.0: December 17, 2013
Added runtime option (QUtil::setRandomDataProvider) to supply your own random data provider. You can
use this if you want to avoid using the OS-provided secure random number generation facility or stdlib's less
secure version. See comments in include/qpdf/QUtil.hh for details.
Fixed image comparison tests to not create 12-bit-per-pixel images since some versions of tiffcmp have bugs
in comparing them in some cases. This increases the disk space required by the image comparison tests, which
are off by default anyway.

28

Release Notes

Introduce a number of small fixes for compilation on the latest clang in MacOS and the latest Visual C++ in
Windows.
Be able to handle broken files that end the xref table header with a space instead of a newline.
5.0.1: October 18, 2013
Thanks to a detailed review by Florian Weimer and the Red Hat Product Security Team, this release includes a
number of non-user-visible security hardening changes. Please see the ChangeLog file in the source distribution
for the complete list.
When available, operating system-specific secure random number generation is used for generating
initialization vectors and other random values used during encryption or file creation. For the Windows build,
this results in an added dependency on Microsoft's cryptography API. To disable the OS-specific cryptography
and use the old version, pass the --enable-insecure-random option to ./configure.
The qpdf command-line tool now issues a warning when -accessibility=n is specified for newer encryption
versions stating that the option is ignored. qpdf, per the spec, has always ignored this flag, but it previously
did so silently. This warning is issued only by the command-line tool, not by the library. The library's handling
of this flag is unchanged.
5.0.0: July 10, 2013
Bug fix: previous versions of qpdf would lose objects with generation != 0 when generating object streams.
Fixing this required changes to the public API.
Removed methods from public API that were only supposed to be called by QPDFWriter and couldn't
realistically be called anywhere else. See ChangeLog for details.
New QPDFObjGen class added to represent an object ID/generation pair. QPDFObjectHandle::getObjGen() is
now preferred over QPDFObjectHandle::getObjectID() and QPDFObjectHandle::getGeneration() as it makes
it less likely for people to accidentally write code that ignores the generation number. See QPDF.hh and
QPDFObjectHandle.hh for additional notes.
Add --show-npages command-line option to the qpdf command to show the number of pages in a file.
Allow omission of the page range within --pages for the qpdf command. When omitted, the page range is
implicitly taken to be all the pages in the file.
Various enhancements were made to support different types of broken files or broken readers. Details can be
found in ChangeLog.
4.1.0: April 14, 2013
Note to people including qpdf in distributions: the .la files generated by libtool are now installed by qpdf's make
install target. Before, they were not installed. This means that if your distribution does not want to include .la
files, you must remove them as part of your packaging process.
Major enhancement: API enhancements have been made to support parsing of content streams. This
enhancement includes the following changes:
QPDFObjectHandle::parseContentStream method parses objects in a content stream and calls handlers in a
callback class. The example examples/pdf-parse-content.cc illustrates how this may be used.
QPDFObjectHandle can now represent operators and inline images, object types that may only appear
in content streams.

29

Release Notes

Method QPDFObjectHandle::getTypeCode() returns an enumerated type value representing the underlying

object type. Method QPDFObjectHandle::getTypeName() returns a text string describing the name of
the type of a QPDFObjectHandle object. These methods can be used for more efficient parsing and
debugging/diagnostic messages.
qpdf --check now parses all pages' content streams in addition to doing other checks. While there are still many
types of errors that cannot be detected, syntactic errors in content streams will now be reported.
Minor compilation enhancements have been made to facilitate easier for support for a broader range of compilers
and compiler versions.
Warning flags have been moved into a separate variable in autoconf.mk
The configure flag --enable-werror work for Microsoft compilers
All MSVC CRT security warnings have been resolved.
All C-style casts in C++ Code have been replaced by C++ casts, and many casts that had been included
to suppress higher warning levels for some compilers have been removed, primarily for clarity. Places
where integer type coercion occurs have been scrutinized. A new casting policy has been documented in the
manual. This is of concern mainly to people porting qpdf to new platforms or compilers. It is not visible to
programmers writing code that uses the library
Some internal limits have been removed in code that converts numbers to strings. This is largely invisible to
users, but it does trigger a bug in some older versions of mingw-w64's C++ library. See README-windows.txt
in the source distribution if you think this may affect you. The copy of the DLL distributed with qpdf's binary
distribution is not affected by this problem.
The RPM spec file previously included with qpdf has been removed. This is because virtually all Linux
distributions include qpdf now that it is a dependency of CUPS filters.
A few bug fixes are included:
Overridden compressed objects are properly handled. Before, there were certain constructs that could cause
qpdf to see old versions of some objects. The most usual manifestation of this was loss of filled in form
values for certain files.
Installation no longer uses GNU/Linux-specific versions of some commands, so make install works on
Solaris with native tools.
The 64-bit mingw Windows binary package no longer includes a 32-bit DLL.
4.0.1: January 17, 2013
Fix detection of binary attachments in test suite to avoid false test failures on some platforms.
Add clarifying comment in QPDF.hh to methods that return the user password explaining that it is no longer
possible with newer encryption formats to recover the user password knowing the owner password. In earlier
encryption formats, the user password was encrypted in the file using the owner password. In newer encryption
formats, a separate encryption key is used on the file, and that key is independently encrypted using both the
user password and the owner password.
4.0.0: December 31, 2012
Major enhancement: support has been added for newer encryption schemes supported by version X of Adobe
Acrobat. This includes use of 127-character passwords, 256-bit encryption keys, and the encryption scheme

30

Release Notes

specified in ISO 32000-2, the PDF 2.0 specification. This scheme can be chosen from the command line by
specifying use of 256-bit keys. qpdf also supports the deprecated encryption method used by Acrobat IX. This
encryption style has known security weaknesses and should not be used in practice. However, such files exist in
the wild, so support for this scheme is still useful. New methods QPDFWriter::setR6EncryptionParameters
(for the PDF 2.0 scheme) and QPDFWriter::setR5EncryptionParameters (for the deprecated scheme) have
been added to enable these new encryption schemes. Corresponding functions have been added to the C API
as well.
Full support for Adobe extension levels in PDF version information. Starting with PDF version 1.7,
corresponding to ISO 32000, Adobe adds new functionality by increasing the extension level rather
than increasing the version. This support includes addition of the QPDF::getExtensionLevel method for
retrieving the document's extension level, addition of versions of QPDFWriter::setMinimumPDFVersion
and QPDFWriter::forcePDFVersion that accept an extension level, and extended syntax for specifying
forced and minimum versions on the command line as described in Section 3.5, Advanced Transformation
Options, page 8. Corresponding functions have been added to the C API as well.
Minor fixes to prevent qpdf from referencing objects in the file that are not referenced in the file's overall
structure. Most files don't have any such objects, but some files have contain unreferenced objects with errors,
so these fixes prevent qpdf from needlessly rejecting or complaining about such objects.
Add new generalized methods for reading and writing files from/to programmer-defined sources. The
method QPDF::processInputSource allows the programmer to use any input source for the input file, and
QPDFWriter::setOutputPipeline allows the programmer to write the output file through any pipeline. These
methods would make it possible to perform any number of specialized operations, such as accessing external
storage systems, creating bindings for qpdf in other programming languages that have their own I/O systems,
etc.
Add new method QPDF::getEncryptionKey for retrieving the underlying encryption key used in the file.
This release includes a small handful of non-compatible API changes. While effort is made to avoid such
changes, all the non-compatible API changes in this version were to parts of the API that would likely never
be used outside the library itself. In all cases, the altered methods or structures were parts of the QPDF that
were public to enable them to be called from either QPDFWriter or were part of validation code that was overzealous in reporting problems in parts of the file that would not ordinarily be referenced. In no case did any of
the removed methods do anything worse that falsely report error conditions in files that were broken in ways
that didn't matter. The following public parts of the QPDF class were changed in a non-compatible way:
Updated nested QPDF::EncryptionData class to add fields needed by the newer encryption formats,
member variables changed to private so that future changes will not require breaking backward compatibility.
Added additional parameters to compute_data_key, which is used by QPDFWriter to compute the
encryption key used to encrypt a specific object.
Removed the method flattenScalarReferences. This method was previously used prior to writing a new PDF
file, but it has the undesired side effect of causing qpdf to read objects in the file that were not referenced.
Some otherwise files have unreferenced objects with errors in them, so this could cause qpdf to reject files
that would be accepted by virtually all other PDF readers. In fact, qpdf relied on only a very small part of
what flattenScalarReferences did, so only this part has been preserved, and it is now done directly inside
QPDFWriter.
Removed the method decodeStreams. This method was used by the --check option of the qpdf command-line
tool to force all streams in the file to be decoded, but it also suffered from the problem of opening otherwise
unreferenced streams and thus could report false positive. The --check option now causes qpdf to go through
all the motions of writing a new file based on the original one, so it will always reference and check exactly
those parts of a file that any ordinary viewer would check.
31

Release Notes

Removed the method trimTrailerForWrite. This method was used by QPDFWriter to modify the original
QPDF object by removing fields from the trailer dictionary that wouldn't apply to the newly written file.
This functionality, though generally harmless, was a poor implementation and has been replaced by having
QPDFWriter filter these out when copying the trailer rather than modifying the original QPDF object. (Note
that qpdf never modifies the original file itself.)
Allow the PDF header to appear anywhere in the first 1024 bytes of the file. This is consistent with what other
readers do.
Fix the pkg-config files to list zlib and pcre in Requires.private to better support static linking using pkg-config.
3.0.2: September 6, 2012
Bug fix: QPDFWriter::setOutputMemory did not work when not used with QPDFWriter::setStaticID, which
made it pretty much useless. This has been fixed.
New API call QPDFWriter::setExtraHeaderText inserts additional text near the header of the PDF file. The
intended use case is to insert comments that may be consumed by a downstream application, though other use
cases may exist.
3.0.1: August 11, 2012
Version 3.0.0 included addition of files for pkg-config, but this was not mentioned in the release notes. The
release notes for 3.0.0 were updated to mention this.
Bug fix: if an object stream ended with a scalar object not followed by space, qpdf would incorrectly report that
it encountered a premature EOF. This bug has been in qpdf since version 2.0.
3.0.0: August 2, 2012
Acknowledgment: I would like to express gratitude for the contributions of Tobias Hoffmann toward the release
of qpdf version 3.0. He is responsible for most of the implementation and design of the new API for manipulating
pages, and contributed code and ideas for many of the improvements made in version 3.0. Without his work,
this release would certainly not have happened as soon as it did, if at all.
Non-compatible API change: The version of QPDFObjectHandle::replaceStreamData that uses a
StreamDataProvider no longer requires (or accepts) a length parameter. See Appendix C, Upgrading to
3.0, page 39 for an explanation. While care is taken to avoid non-compatible API changes in general, an
exception was made this time because the new interface offers an opportunity to significantly simplify calling
code.
Support has been added for large files. The test suite verifies support for files larger than 4 gigabytes, and
manual testing has verified support for files larger than 10 gigabytes. Large file support is available for both
32-bit and 64-bit platforms as long as the compiler and underlying platforms support it.
Support for page selection (splitting and merging PDF files) has been added to the qpdf command-line tool.
See Section 3.4, Page Selection Options, page 6.
Options have been added to the qpdf command-line tool for copying encryption parameters from another file.
See Section 3.2, Basic Options, page 4.
New methods have been added to the QPDF object for adding and removing pages. See Section 6.6, Adding
and Removing Pages, page 19.
New methods have been added to the QPDF object for copying objects from other PDF files. See Section 6.8,
Copying Objects From Other PDF Files, page 20

32

Release Notes

A new method QPDFObjectHandle::parse has been added for constructing QPDFObjectHandle objects
from a string description.
Methods have been added to QPDFWriter to allow writing to an already open stdio FILE* addition to writing
to standard output or a named file. Methods have been added to QPDF to be able to process a file from an
already open stdio FILE*. This makes it possible to read and write PDF from secure temporary files that have
been unlinked prior to being fully read or written.
The QPDF::emptyPDF can be used to allow creation of PDF files from scratch. The example examples/pdfcreate.cc illustrates how it can be used.
Several methods to take PointerHolder<Buffer> can now also accept std::string arguments.
Many new convenience methods have been added to the library, most in QPDFObjectHandle. See
ChangeLog for a full list.
When building on a platform that supports ELF shared libraries (such as Linux), symbol versions are enabled
by default. They can be disabled by passing --disable-ld-version-script to ./configure.
The file libqpdf.pc is now installed to support pkg-config.
Image comparison tests are off by default now since they are not needed to verify a correct build or port of
qpdf. They are needed only when changing the actual PDF output generated by qpdf. You should enable them
if you are making deep changes to qpdf itself. See README for details.
Large file tests are off by default but can be turned on with ./configure or by setting an environment variable
before running the test suite. See README for details.
When qpdf's test suite fails, failures are not printed to the terminal anymore by default. Instead, find them in
build/qtest.log. For packagers who are building with an autobuilder, you can add the --enable-show-failedtest-output option to ./configure to restore the old behavior.
2.3.1: December 28, 2011
Fix thread-safety problem resulting from non-thread-safe use of the PCRE library.
Made a few minor documentation fixes.
Add workaround for a bug that appears in some versions of ghostscript to the test suite
Fix minor build issue for Visual C++ 2010.
2.3.0: August 11, 2011
Bug fix: when preserving existing encryption on encrypted files with cleartext metadata, older qpdf versions
would generate password-protected files with no valid password. This operation now works. This bug only
affected files created by copying existing encryption parameters; explicit encryption with specification of
cleartext metadata worked before and continues to work.
Enhance QPDFWriter with a new constructor that allows you to delay the specification of the output file.
When using this constructor, you may now call QPDFWriter::setOutputFilename to specify the output file,
or you may use QPDFWriter::setOutputMemory to cause QPDFWriter to write the resulting PDF file to a
memory buffer. You may then use QPDFWriter::getBuffer to retrieve the memory buffer.
Add new API call QPDF::replaceObject for replacing objects by object ID
Add new API call QPDF::swapObjects for swapping two objects by object ID

33

Release Notes

Add QPDFObjectHandle::getDictAsMap and QPDFObjectHandle::getArrayAsVector to allow retrieval of

dictionary objects as maps and array objects as vectors.
Add functions qpdf_get_info_key and qpdf_set_info_key to the C API for manipulating string fields of the
document's /Info dictionary.
Add functions qpdf_init_write_memory, qpdf_get_buffer_length, and qpdf_get_buffer to the C API for writing
PDF files to a memory buffer instead of a file.
2.2.4: June 25, 2011
Fix installation and compilation issues; no functionality changes.
2.2.3: April 30, 2011
Handle some damaged streams with incorrect characters following the stream keyword.
Improve handling of inline images when normalizing content streams.
Enhance error recovery to properly handle files that use object 0 as a regular object, which is specifically
disallowed by the spec.
2.2.2: October 4, 2010
Add new function qpdf_read_memory to the C API to call QPDF::processMemoryFile. This was an omission
in qpdf 2.2.1.
2.2.1: October 1, 2010
Add new method QPDF::setOutputStreams to replace std::cout and std::cerr with other streams for generation
of diagnostic messages and error messages. This can be useful for GUIs or other applications that want to
capture any output generated by the library to present to the user in some other way. Note that QPDF does
not write to std::cout (or the specified output stream) except where explicitly mentioned in QPDF.hh, and
that the only use of the error stream is for warnings. Note also that output of warnings is suppressed when
setSuppressWarnings(true) is called.
Add new method QPDF::processMemoryFile for operating on PDF files that are loaded into memory rather
than in a file on disk.
Give a warning but otherwise ignore empty PDF objects by treating them as null. Empty object are not permitted
by the PDF specification but have been known to appear in some actual PDF files.
Handle inline image filter abbreviations when the appear as stream filter abbreviations. The PDF specification
does not allow use of stream filter abbreviations in this way, but Adobe Reader and some other PDF readers
accept them since they sometimes appear incorrectly in actual PDF files.
Implement miscellaneous enhancements to PointerHolder and Buffer to support other changes.
2.2.0: August 14, 2010
Add new methods to QPDFObjectHandle (newStream and replaceStreamData for creating new streams and
replacing stream data. This makes it possible to perform a wide range of operations that were not previously
possible.
Add new helper method in QPDFObjectHandle (addPageContents) for appending or prepending new content
streams to a page. This method makes it possible to manipulate content streams without having to be concerned
whether a page's contents are a single stream or an array of streams.

34

Release Notes

Add new method in QPDFObjectHandle: replaceOrRemoveKey, which replaces a dictionary key with a
given value unless the value is null, in which case it removes the key instead.
Add new method in QPDFObjectHandle: getRawStreamData, which returns the raw (unfiltered) stream data
into a buffer. This complements the getStreamData method, which returns the filtered (uncompressed) stream
data and can only be used when the stream's data is filterable.
Provide two new examples: pdf-double-page-size and pdf-invert-images that illustrate the newly added
interfaces.
Fix a memory leak that would cause loss of a few bytes for every object involved in a cycle of object references.
Thanks to Jian Ma for calling my attention to the leak.
2.1.5: April 25, 2010
Remove restriction of file identifier strings to 16 bytes. This unnecessary restriction was preventing qpdf from
being able to encrypt or decrypt files with identifier strings that were not exactly 16 bytes long. The specification
imposes no such restriction.
2.1.4: April 18, 2010
Apply the same padding calculation fix from version 2.1.2 to the main cross reference stream as well.
Since qpdf --check only performs limited checks, clarify the output to make it clear that there still may be
errors that qpdf can't check. This should make it less surprising to people when another PDF reader is unable
to read a file that qpdf thinks is okay.
2.1.3: March 27, 2010
Fix bug that could cause a failure when rewriting PDF files that contain object streams with unreferenced objects
that in turn reference indirect scalars.
Don't complain about (invalid) AES streams that aren't a multiple of 16 bytes. Instead, pad them before
decrypting.
2.1.2: January 24, 2010
Fix bug in padding around first half cross reference stream in linearized files. The bug could cause an assertion
failure when linearizing certain unlucky files.
2.1.1: December 14, 2009
No changes in functionality; insert missing include in an internal library header file to support gcc 4.4, and
update test suite to ignore broken Adobe Reader installations.
2.1: October 30, 2009
This is the first version of qpdf to include Windows support. On Windows, it is possible to build a DLL.
Additionally, a partial C-language API has been introduced, which makes it possible to call qpdf functions
from non-C++ environments. I am very grateful to Zarko Gagic (http://delphi.about.com/) for tirelessly testing
numerous pre-release versions of this DLL and providing many excellent suggestions on improving the
interface.
For programming to the C interface, please see the header file qpdf/qpdf-c.h and the example examples/pdflinearize.c.
Zarko Gajic has written a Delphi wrapper for qpdf, which can be downloaded from qpdf's download side.
Zarko's Delphi wrapper is released with the same licensing terms as qpdf itself and comes with this disclaimer:

35

Release Notes

Delphi wrapper unit qpdf.pas created by Zarko Gajic (http://delphi.about.com/). Use at your own risk and for
whatever purpose you want. No support is provided. Sample code is provided.
Support has been added for AES encryption and crypt filters. Although qpdf does not presently support files
that use PKI-based encryption, with the addition of AES and crypt filters, qpdf is now be able to open most
encrypted files created with newer versions of Acrobat or other PDF creation software. Note that I have not
been able to get very many files encrypted in this way, so it's possible there could still be some cases that qpdf
can't handle. Please report them if you find them.
Many error messages have been improved to include more information in hopes of making qpdf a more useful
tool for PDF experts to use in manually recovering damaged PDF files.
Attempt to avoid compressing metadata streams if possible. This is consistent with other PDF creation
applications.
Provide new command-line options for AES encrypt, cleartext metadata, and setting the minimum and forced
PDF versions of output files.
Add additional methods to the QPDF object for querying the document's permissions. Although qpdf does
not enforce these permissions, it does make them available so that applications that use qpdf can enforce
permissions.
The --check option to qpdf has been extended to include some additional information.
There have been a handful of non-compatible API changes. For details, see Appendix B, Upgrading from 2.0
to 2.1, page 38.
2.0.6: May 3, 2009
Do not attempt to uncompress streams that have decode parameters we don't recognize. Earlier versions of qpdf
would have rejected files with such streams.
2.0.5: March 10, 2009
Improve error handling in the LZW decoder, and fix a small error introduced in the previous version with regard
to handling full tables. The LZW decoder has been more strongly verified in this release.
2.0.4: February 21, 2009
Include proper support for LZW streams encoded without the early code change flag. Special thanks to Atom
Smasher who reported the problem and provided an input file compressed in this way, which I did not previously
have.
Implement some improvements to file recovery logic.
2.0.3: February 15, 2009
Compile cleanly with gcc 4.4.
Handle strings encoded as UTF-16BE properly.
2.0.2: June 30, 2008
Update test suite to work properly with a non-bash /bin/sh and with Perl 5.10. No changes were made to the
actual qpdf source code itself for this release.

36

Release Notes

2.0.1: May 6, 2008

No changes in functionality or interface. This release includes fixes to the source code so that qpdf compiles
properly and passes its test suite on a broader range of platforms. See ChangeLog in the source distribution
for details.
2.0: April 29, 2008
First public release.

37

Appendix B. Upgrading from 2.0 to 2.1

Although, as a general rule, we like to avoid introducing source-level incompatibilities in qpdf's interface, there were a
few non-compatible changes made in this version. A considerable amount of source code that uses qpdf will probably
compile without any changes, but in some cases, you may have to update your code. The changes are enumerated here.
There are also some new interfaces; for those, please refer to the header files.
QPDF's exception handling mechanism now uses std::logic_error for internal errors and std::runtime_error
for runtime errors in favor of the now removed QEXC classes used in previous versions. The QEXC exception
classes predated the addition of the <stdexcept> header file to the C++ standard library. Most of the exceptions
thrown by the qpdf library itself are still of type QPDFExc which is now derived from std::runtime_error.
Programs that caught an instance of std::exception and displayed it by calling the what() method will not need
to be changed.
The QPDFExc class now internally represents various fields of the error condition and provides interfaces for
querying them. Among the fields is a numeric error code that can help applications act differently on (a small number
of) different error conditions. See QPDFExc.hh for details.
Warnings can be retrieved from qpdf as instances of QPDFExc instead of strings.
The nested QPDF::EncryptionData class's constructor takes an additional argument. This class is primarily
intended to be used by QPDFWriter. There's not really anything useful an end-user application could do with it. It
probably shouldn't really be part of the public interface to begin with. Likewise, some of the methods for computing
internal encryption dictionary parameters have changed to support /R=4 encryption.
The method QPDF::getUserPassword has been removed since it didn't do what people would think it did. There
are now two new methods: QPDF::getPaddedUserPassword and QPDF::getTrimmedUserPassword. The first one
does what the old QPDF::getUserPassword method used to do, which is to return the password with possible binary
padding as specified by the PDF specification. The second one returns a human-readable password string.
The enumerated types that used to be nested in QPDFWriter have moved to top-level enumerated types and are
now defined in the file qpdf/Constants.h. This enables them to be shared by both the C and C++ interfaces.

38

Appendix C. Upgrading to 3.0

For the most part, the API for qpdf version 3.0 is backward compatible with versions 2.1 and later. There are two
exceptions:
The method QPDFObjectHandle::replaceStreamData that uses a StreamDataProvider to provide the stream
data no longer takes a length parameter. While it would have been easy enough to keep the parameter for backward
compatibility, in this case, the parameter was removed since this provides the user an opportunity to simplify the
calling code. This method was introduced in version 2.2. At the time, the length parameter was required in order
to ensure that calls to the stream data provider returned the same length for a specific stream every time they were
invoked. In particular, the linearization code depends on this. Instead, qpdf 3.0 and newer check for that constraint
explicitly. The first time the stream data provider is called for a specific stream, the actual length is saved, and
subsequent calls are required to return the same number of bytes. This means the calling code no longer has to
compute the length in advance, which can be a significant simplification. If your code fails to compile because of
the extra argument and you don't want to make other changes to your code, just omit the argument.
Many methods take long long instead of other integer types. Most if not all existing code should compile fine
with this change since such parameters had always previously been smaller types. This change was required to
support files larger than two gigabytes in size.

39

Appendix D. Upgrading to 4.0

While version 4.0 includes a few non-compatible API changes, it is very unlikely that anyone's code would have
used any of those parts of the API since they generally required information that would only be available inside the
library. In the unlikely event that you should run into trouble, please see the ChangeLog. See also Appendix A, Release
Notes, page 28 for a complete list of the non-compatible API changes made in this version.

40