The qpdf package has few external dependencies. In order to build qpdf, the following packages are required:

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.

If you do not enable this, then you do not need to have tiff and ghostscript.

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.

Building qpdf on UNIX is generally just a matter of running

./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.md and INSTALL files in the source distribution.

Building on Windows is a little bit more complicated. For details, please see README-windows.md 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.md.

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.

Starting with qpdf 9.1.0, the qpdf library can be built with multiple implementations of providers of cryptographic functions, which we refer to as “crypto providers.” At the time of writing, a crypto implementation must provide MD5 and SHA2 (256, 384, and 512-bit) hashes and RC4 and AES256 with and without CBC encryption. In the future, if digital signature is added to qpdf, there may be additional requirements beyond this.

Starting with qpdf version 9.1.0, the available implementations are native and gnutls. Additional implementations may be added if needed. It is also possible for a developer to provide their own implementation without modifying the qpdf library.

If you are packaging qpdf for an operating system distribution, here are some things you may want to keep in mind:

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 3.5, “Page Selection Options”.

If @filename appears anywhere in the command-line, it will be read line by line, and each line will be treated as a command-line argument. The @- option allows arguments to be read from standard input. This allows qpdf to be invoked with an arbitrary number of arbitrarily long arguments. It is also very useful for avoiding having to pass passwords on the command line.

outfilename does not have to be seekable, even when generating linearized files. Specifying “-” as outfilename means to write to standard output. If you want to overwrite the input file with the output, use the option --replace-input and omit the output file name. You can't specify the same file as both the input and the output. If you do this, qpdf will tell you about the --replace-input option.

Most options require an output file, but some testing or inspection commands do not. These are specifically noted.

Starting in qpdf version 8.3.0, qpdf provides its own completion support for zsh and bash. You can enable bash completion with eval $(qpdf --completion-bash) and zsh completion with eval $(qpdf --completion-zsh). If qpdf is not in your path, you should invoke it above with an absolute path. If you invoke it with a relative path, it will warn you, and the completion won't work if you're in a different directory.

The following options are the most common ones and perform commonly needed transformations.

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.

Prior to 8.4.0, in the case of passwords that contain characters that fall outside of 7-bit US-ASCII, qpdf left the burden of supplying properly encoded encryption and decryption passwords to the user. Starting in qpdf 8.4.0, qpdf does this automatically in most cases. For an in-depth discussion, please see Section 3.10, “Unicode Passwords”. Previous versions of this manual described workarounds using the iconv command. Such workarounds are no longer required or recommended with qpdf 8.4.0. However, for backward compatibility, qpdf attempts to detect those workarounds and do the right thing in most cases.

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:

If key-length is 128, the following restriction options are available:

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:

The default for each permission option is to be fully permissive.

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 ] [ ... ] --

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.

Starting with qpf 8.4, the special input file name “.” can be used shortcut for the primary input filename.

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 --.

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. A number preceded by an “r” indicates to count from the end, so r3-r1 would be the last three pages of the document. 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. Numbers may be repeated in a page range. A page range may be optionally appended with :even or :odd to indicate only the even or odd pages in the given range. Note that even and odd refer to the positions within the specified, range, not whether the original number is even or odd.

Example page ranges:

Starting in qpdf version 8.3, you can specify the --collate option. Note that this option is specified outside of --pages ... --. When --collate is specified, it changes the meaning of --pages so that the specified files, as modified by page ranges, are collated rather than concatenated. For example, if you add the files odd.pdf and even.pdf containing odd and even pages of a document respectively, you could run qpdf --collate odd.pdf --pages odd.pdf even.pdf -- all.pdf to collate the pages. This would pick page 1 from odd, page 1 from even, page 2 from odd, page 2 from even, etc. until all pages have been included. Any number of files and page ranges can be specified. If any file has fewer pages, that file is just skipped when its pages have all been included. For example, if you ran qpdf --collate --empty --pages a.pdf 1-5 b.pdf 6-4 c.pdf r1 -- out.pdf, you would get the following pages in this order:

Starting in qpdf version 8.3, when you split and merge files, any page labels (page numbers) are preserved in the final file. It is expected that more document features will be preserved by splitting and merging. In the mean time, 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. 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 . 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 1–5 from file1.pdf and pages 11–15 from file2.pdf in reverse, you would run

qpdf file1.pdf --pages file1.pdf 1-5 . 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 document-level metadata but preserve encryption, you would use

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.

Prior to version 8.4, it was not possible to specify the same page from the same file directly more than once, and the workaround of specifying the same file in more than one way was required. Version 8.4 removes this limitation.

Starting with qpdf 8.4, it is possible to overlay or underlay pages from other files onto the output generated by qpdf. Specify overlay or underlay as follows:

{ --overlay | --underlay } file [ options ] --

Overlay and underlay options are processed late, so they can be combined with other like merging and will apply to the final output. The --overlay and --underlay options work the same way, except underlay pages are drawn underneath the page to which they are applied, possibly obscured by the original page, and overlay files are drawn on top of the page to which they are applied, possibly obscuring the page. You can combine overlay and underlay.

The default behavior of overlay and underlay is that pages are taken from the overlay/underlay file in sequence and applied to corresponding pages in the output until there are no more output pages. If the overlay or underlay file runs out of pages, remaining output pages are left alone. This behavior can be modified by options, which are provided between the --overlay or --underlay flag and the -- option. The following options are supported:

Here are some examples.

These options control aspects of how qpdf reads PDF files. Mostly these are of use to people who are working with damaged files. There is little reason to use these options unless you are trying to solve specific problems. The following options are available:

Ordinarily, qpdf will attempt to recover from certain types of errors in PDF files. These include errors in the cross-reference table, certain types of object numbering errors, and certain types of stream length errors. Sometimes, qpdf may think it has recovered but may not have actually recovered, so care should be taken when using this option as some data loss is possible. The --suppress-recovery option will prevent qpdf from attempting recovery. In this case, it will fail on the first error that it encounters.

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.

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:

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 --stream-data=compress.

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.

This paragraph discusses edge cases of content normalization that are not of concern to most users and are not relevant when content normalization is not enabled. When normalizing content, if qpdf runs into any lexical errors, it will print a warning indicating that content may be damaged. The only situation in which qpdf is known to cause damage during content normalization is when a page's contents are split across multiple streams and streams are split in the middle of a lexical token such as a string, name, or inline image. There may be some pathological cases in which qpdf could damage content without noticing this, such as if the partial tokens at the end of one stream and the beginning of the next stream are both valid, but usually qpdf will be able to detect this case. For slightly increased safety, you can specify --coalesce-contents in addition to --normalize-content or --qdf. This will cause qpdf to combine all the content streams into one, thus recombining any split tokens. However doing this will prevent you from being able to see the original layout of the content streams. If you must inspect the original content streams in an uncompressed format, you can always run with --qdf --normalize-content=n for a QDF file without content normalization, or alternatively --stream-data=uncompress for a regular non-QDF mode file with uncompressed streams. These will both uncompress all the streams but will not attempt to normalize content. Please note that if you are using content normalization or QDF mode for the purpose of manually inspecting files, you don't have to care about this.

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.

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 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.

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:

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.

At the library API level, all methods that perform encryption and decryption interpret passwords as strings of bytes. It is up to the caller to ensure that they are appropriately encoded. Starting with qpdf version 8.4.0, qpdf will attempt to make this easier for you when interact with qpdf via its command line interface. The PDF specification requires passwords used to encrypt files with 40-bit or 128-bit encryption to be encoded with PDF Doc encoding. This encoding is a single-byte encoding that supports ISO-Latin-1 and a handful of other commonly used characters. It has a large overlap with Windows ANSI but is not exactly the same. There is generally not a way to provide PDF Doc encoded strings on the command line. As such, qpdf versions prior to 8.4.0 would often create PDF files that couldn't be opened with other software when given a password with non-ASCII characters to encrypt a file with 40-bit or 128-bit encryption. Starting with qpdf 8.4.0, qpdf recognizes the encoding of the parameter and transcodes it as needed. The rest of this section provides the details about exactly how qpdf behaves. Most users will not need to know this information, but it might be useful if you have been working around qpdf's old behavior or if you are using qpdf to generate encrypted files for testing other PDF software.

A note about Windows: when qpdf builds, it attempts to determine what it has to do to use wmain instead of main on Windows. The wmain function is an alternative entry point that receives all arguments as UTF-16-encoded strings. When qpdf starts up this way, it converts all the strings to UTF-8 encoding and then invokes the regular main. This means that, as far as qpdf is concerned, it receives its command-line arguments with UTF-8 encoding, just as it would in any modern Linux or UNIX environment.

If a file is being encrypted with 40-bit or 128-bit encryption and the supplied password is not a valid UTF-8 string, qpdf will fall back to the behavior of interpreting the password as a string of bytes. If you have old scripts that encrypt files by passing the output of iconv to qpdf, you no longer need to do that, but if you do, qpdf should still work. The only exception would be for the extremely unlikely case of a password that is encoded with a single-byte encoding but also happens to be valid UTF-8. Such a password would contain strings of even numbers of characters that alternate between accented letters and symbols. In the extremely unlikely event that you are intentionally using such passwords and qpdf is thwarting you by interpreting them as UTF-8, you can use --password-mode=bytes to suppress qpdf's automatic behavior.

The --password-mode option, as described earlier in this chapter, can be used to change qpdf's interpretation of supplied passwords. There are very few reasons to use this option. One would be the unlikely case described in the previous paragraph in which the supplied password happens to be valid UTF-8 but isn't supposed to be UTF-8. Your best bet would be just to provide the password as a valid UTF-8 string, but you could also use --password-mode=bytes. Another reason to use --password-mode=bytes would be to intentionally generate PDF files encrypted with passwords that are not properly encoded. The qpdf test suite does this to generate invalid files for the purpose of testing its password recovery capability. If you were trying to create intentionally incorrect files for a similar purposes, the bytes password mode can enable you to do this.

When qpdf attempts to decrypt a file with a password that contains non-ASCII characters, it will generate a list of alternative passwords by attempting to interpret the password as each of a handful of different coding systems and then transcode them to the required format. This helps to compensate for the supplied password being given in the wrong coding system, such as would happen if you used the iconv workaround that was previously needed. It also generates passwords by doing the reverse operation: translating from correct in incorrect encoding of the password. This would enable qpdf to decrypt files using passwords that were improperly encoded by whatever software encrypted the files, including older versions of qpdf invoked without properly encoded passwords. The combination of these two recovery methods should make qpdf transparently open most encrypted files with the password supplied correctly but in the wrong coding system. There are no real downsides to this behavior, but if you don't want qpdf to do this, you can use the --suppress-password-recovery option. One reason to do that is to ensure that you know the exact password that was used to encrypt the file.

With these changes, qpdf now generates compliant passwords in most cases. There are still some exceptions. In particular, the PDF specification directs compliant writers to normalize Unicode passwords and to perform certain transformations on passwords with bidirectional text. Implementing this functionality requires using a real Unicode library like ICU. If a client application that uses qpdf wants to do this, the qpdf library will accept the resulting passwords, but qpdf will not perform these transformations itself. It is possible that this will be addressed in a future version of qpdf. The QPDFWriter methods that enable encryption on the output file accept passwords as strings of bytes.

Please note that the --password-is-hex-key option is unrelated to all this. This flag bypasses the normal process of going from password to encryption string entirely, allowing the raw encryption key to be specified directly. This is useful for forensic purposes or for brute-force recovery of files with unknown passwords.

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/QPDFWriter.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 -lz -ljpeg 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.

The qpdf library is implemented in C++, which makes it hard to use directly in other languages. There are a few things that can help.

When strings are passed to qpdf library routines either as char* or as std::string, they are treated as byte arrays except where otherwise noted. When Unicode is desired, qpdf wants UTF-8 unless otherwise noted in comments in header files. In modern UNIX/Linux environments, this generally does the right thing. In Windows, it's a bit more complicated. Starting in qpdf 8.4.0, passwords that contain Unicode characters are handled much better, and starting in qpdf 8.4.1, the library attempts to properly handle Unicode characters in filenames. In particular, in Windows, if a UTF-8 encoded string is used as a filename in either QPDF or QPDFWriter, it is internally converted to wchar_t*, and Unicode-aware Windows APIs are used. As such, qpdf will generally operate properly on files with non-ASCII characters in their names as long as the filenames are UTF-8 encoded for passing into the qpdf library API, but there are still some rough edges, such as the encoding of the filenames in error messages our CLI output messages. Patches or bug reports are welcome for any continuing issues with Unicode file names in Windows.

Beginning with qpdf version 8.3.0, the qpdf command-line program can produce a JSON representation of the non-content data in a PDF file. It includes a dump in JSON format of all objects in the PDF file excluding the content of streams. This JSON representation makes it very easy to look in detail at the structure of a given PDF file, and it also provides a great way to work with PDF files programmatically from the command-line in languages that can't call or link with the qpdf library directly. Note that stream data can be extracted from PDF files using other qpdf command-line options.

The qpdf JSON representation includes a JSON serialization of the raw objects in the PDF file as well as some computed information in a more easily extracted format. QPDF provides some guarantees about its JSON format. These guarantees are designed to simplify the experience of a developer working with the JSON format.

There are a few limitations to be aware of with the JSON structure:

For the most part, the built-in JSON help tells you everything you need to know about the JSON format, but there are a few non-obvious things to be aware of:

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.

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, though it is possible to determine whether an object is direct or not if this information is needed. 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 C++-11's std::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.

The primary class for interacting with PDF objects is QPDFObjectHandle. Instances of this class can be passed around by value, copied, stored in containers, etc. with very low overhead. Instances of QPDFObjectHandle created by reading from a file will 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.

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.

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 files QPDF.hh and QPDFObjectHandle.hh. There are also some additional helper classes that provide higher level API functions for certain document constructions. These are discussed in Section 7.3, “Helper Classes”.

QPDF version 8.1 introduced the concept of helper classes. Helper classes are intended to contain higher level APIs that allow developers to work with certain document constructs at an abstraction level above that of QPDFObjectHandle while staying true to qpdf's philosophy of not hiding document structure from the developer. As with qpdf in general, the goal is take away some of the more tedious bookkeeping aspects of working with PDF files, not to remove the need for the developer to understand how the PDF construction in question works. The driving factor behind the creation of helper classes was to allow the evolution of higher level interfaces in qpdf without polluting the interfaces of the main top-level classes QPDF and QPDFObjectHandle.

There are two kinds of helper classes: document helpers and object helpers. Document helpers are constructed with a reference to a QPDF object and provide methods for working with structures that are at the document level. Object helpers are constructed with an instance of a QPDFObjectHandle and provide methods for working with specific types of objects.

Examples of document helpers include QPDFPageDocumentHelper, which contains methods for operating on the document's page trees, such as enumerating all pages of a document and adding and removing pages; and QPDFAcroFormDocumentHelper, which contains document-level methods related to interactive forms, such as enumerating form fields and creating mappings between form fields and annotations.

Examples of object helpers include QPDFPageObjectHelper for performing operations on pages such as page rotation and some operations on content streams, QPDFFormFieldObjectHelper for performing operations related to interactive form fields, and QPDFAnnotationObjectHelper for working with annotations.

It is always possible to retrieve the underlying QPDF reference from a document helper and the underlying QPDFObjectHandle reference from an object helper. Helpers are designed to be helpers, not wrappers. The intention is that, in general, it is safe to freely intermix operations that use helpers with operations that use the underlying objects. Document and object helpers do not attempt to provide a complete interface for working with the things they are helping with, nor do they attempt to encapsulate underlying structures. They just provide a few methods to help with error-prone, repetitive, or complex tasks. In some cases, a helper object may cache some information that is expensive to gather. In such cases, the helper classes are implemented so that their own methods keep the cache consistent, and the header file will provide a method to invalidate the cache and a description of what kinds of operations would make the cache invalid. If in doubt, you can always discard a helper class and create a new one with the same underlying objects, which will ensure that you have discarded any stale information.

By Convention, document helpers are called QPDFSomethingDocumentHelper and are derived from QPDFDocumentHelper, and object helpers are called QPDFSomethingObjectHelper and are derived from QPDFObjectHelper. For details on specific helpers, please see their header files. You can find them by looking at include/qpdf/QPDF*DocumentHelper.hh and include/qpdf/QPDF*ObjectHelper.hh.

In order to avoid creation of circular dependencies, the following general guidelines are followed with helper classes:

Prior to qpdf version 8.1, higher level interfaces were added as “convenience functions” in either QPDF or QPDFObjectHandle. For compatibility, older convenience functions for operating with pages will remain in those classes even as alternatives are provided in helper classes. Going forward, new higher level interfaces will be provided using helper classes.

This section contains a few notes about QPDF's internal implementation, particularly around what it does when it first processes a file. This section is a bit of a simplification of what it actually does, but it could serve as a starting point to someone trying to understand the implementation. There is nothing in this section that you need to know to use the qpdf library.

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.

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 iterates 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.

The following example should clarify how QPDF processes a simple file.

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 QIntC namespace, provided by include/qpdf/QIntC.hh, implements safe functions for converting between integer types. These functions do range checking and throw a std::range_error, which is subclass of std::runtime_error, if conversion from one integer type to another results in loss of information. There are many cases in which we have to move between different integer types because of incompatible integer types used in interoperable interfaces. Some are unavoidable, such as moving between sizes and offsets, and others are there because of old code that is too in entrenched to be fixable without breaking source compatibility and causing pain for users. QPDF is compiled with extra warnings to detect conversions with potential data loss, and all such cases should be fixed by either using a function from QIntC or a static_cast.

When the intention is just to switch the type because of exchanging data between incompatible interfaces, use QIntC. This is the usual case. However, there are some cases in which we are explicitly intending to use the exact same bit pattern with a different type. This is most common when switching between signed and unsigned characters. A lot of qpdf's code uses unsigned characters internally, but std::string and char are signed. Using QIntC::to_char would be wrong for converting from unsigned to signed characters because a negative char value and the corresponding unsigned char value greater than 127 mean the same thing. There are also cases in which we use static_cast when working with bit fields where we are not representing a numerical value but rather a bunch of bits packed together in some integer type. Also note that size_t and long both typically differ between 32-bit and 64-bit environments, so sometimes an explicit cast may not be needed to avoid warnings on one platform but may be needed on another. A conversion with QIntC should always be used when the types are different even if the underlying size is the same. QPDF's CI build builds on 32-bit and 64-bit platforms, and the test suite is very thorough, so it is hard to make any of the potential errors here without being caught in build or test.

Non-const unsigned char* is 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.

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.

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.

One point of confusion some people have about encrypted PDF files is that encryption is not the same as password protection. Password protected files are always encrypted, but it is also possible to create encrypted files that do not have passwords. Internally, such files use the empty string as a password, and most readers try the empty string first to see if it works and prompt for a password only if the empty string doesn't work. Normally such files have an empty user password and a non-empty owner password. In that way, if the file is opened by an ordinary reader without specification of password, the restrictions specified in the encryption dictionary can be enforced. Most users wouldn't even realize such a file was encrypted. Since qpdf always ignores the restrictions (except for the purpose of reporting what they are), qpdf doesn't care which password you use. QPDF will allow you to create PDF files with non-empty user passwords and empty owner passwords. Some readers will require a password when you open these files, and others will open the files without a password and not enforce restrictions. Having a non-empty user password and an empty owner password doesn't really make sense because it would mean that opening the file with the user password would be more restrictive than not supplying a password at all. QPDF also allows you to create PDF files with the same password as both the user and owner password. Some readers will not ever allow such files to be accessed without restrictions because they never try the password as the owner password if it works as the user password. Nonetheless, one of the powerful aspects of qpdf is that it allows you to finely specify the way encrypted files are created, even if the results are not useful to some readers. One use case for this would be for testing a PDF reader to ensure that it handles odd configurations of input files.

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.md file in the source distribution and to comments in QUtil.hh.

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.

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 7.10, “Copying Objects From Other PDF Files”. For an example of how to use reserved objects, search for newReserved in test_driver.cc in qpdf's sources.

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 “from” QPDF object must remain valid after the copy as discussed in the note below. 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.

Please note: when you copy objects from one QPDF to another, the source QPDF object must remain valid until you have finished with the destination object. This is because the original object is still used to retrieve any referenced stream data from the copied object.

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 8, Linearization for a description of linearization is implemented. This section describes how we write non-linearized files including the creation of QDF files (see Chapter 4, QDF Mode.

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.

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.

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 end-of-the-line pipeline that does something like store its output to a file or a memory buffer ignoring a successor. For additional details, look at 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.

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.

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 8.3, “Optimization”. 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:

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.

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.

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.)

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.

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 --filtered-stream-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";

Object streams can contain any regular object except the following:

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 9.3, “Implications for Linearized Files”for 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:

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.

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

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:

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.

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.

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.

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.