path: root/Master/texmf-dist/doc/pdftex/manual/pdftex-t.tex

% interface=english modes=letter,screen output=pdftex
% vim: tw=79

% $Id: pdftex-t.tex 790 2018-01-02 00:12:04Z karl $

% The number of lines on the title page depends on exactly
% what \PDF\ code is generated.
\setvariables[pdftex][titlepagelines=64]

\setupoutput
  [pdftex]

%D ConTeXt defaults to 1.5; we want to be maximally compatible (and we don't
%D use any features from 1.4++ anyway, do we?)
\pdfminorversion=3

%D First we define ourselves some abbreviations, if only to force
%D consistency and to save typing. We use real small capitals, not pseudo.

\setupsynonyms
  [abbreviation]
  [textstyle=smallcaps,
   style=smallcaps,
   location=left,
   width=broad,
   sample=\POSTSCRIPT]

\setupcapitals
  [title=no]

\def\svnscan $#1 #2 #3 #4-#5-#6 #7${
  \def\rcsrevision{#3}
  \def\rcsyear{#4}
  \def\rcsmonth{#5}
  \def\rcsday{{\count0=#6\relax\the\count0}}
  \def\rcsmonthname{\ifcase\rcsmonth ERROR\or
    January\or February\or March\or April\or May\or June\or July\or
    August\or September\or October\or November\or December\else ERROR\fi}
}

\svnscan $Id: pdftex-t.tex 790 2018-01-02 00:12:04Z karl $

\def\currentpdftex{1.40.18}

%***********************************************************************

\def\eTeX{{$\varepsilon$}-\kern-.125em\TeX}
\def\eg{e.g.}
\def\Eg{E.g.}
\def\PDFReference{{\sl PDF Reference}} % PDF with capital letters

\abbreviation [AFM]        {afm}        {Adobe Font Metrics}
\abbreviation [BIBTEX]     {Bib\TeX}    {Handles bibliographies}
\abbreviation [ASCII]      {ascii}      {American Standard Code for Information Interchange}
\abbreviation [CONTEXT]    {\ConTeXt}   {general purpose macro package}
\abbreviation [CTAN]       {ctan}       {global \TEX\ archive server}
\abbreviation [DVI]        {dvi}        {native \TEX\ Device Independent file format}
\abbreviation [ENCTEX]     {enc\TeX}    {enc\TeX\ extension to \TEX}
\abbreviation [EPSTOPDF]   {epstopdf}   {\EPS\ to \PDF\ conversion tool}
\abbreviation [EPS]        {eps}        {Encapsulated PostScript}
\abbreviation [ETEX]       {\eTeX}      {an extension to \TEX}
\abbreviation [EXIF]       {exif}       {Exchangeable Image File format (JPEG file variant)}
\abbreviation [FPTEX]      {fp\TeX}     {\WIN\ \WEBC\ distribution}
\abbreviation [GHOSTSCRIPT] {Ghostscript} {\PS\ and \PDF\ language interpreter}
\abbreviation [GNU]        {gnu}        {GNU's Not Unix}
\abbreviation [HZ]         {hz}         {Hermann Zapf optimization}
\abbreviation [ISO]        {iso}        {International Organization for Standardization}
\abbreviation [JBIG]       {jbig}       {Joint Bi-level Image Experts Group}
\abbreviation [JBIGTWO]    {jbig2}      {Joint Bi-level Image Experts Group}
\abbreviation [JFIF]       {jfif}       {JPEG File Interchange Format}
\abbreviation [JPEG]       {jpeg}       {Joint Photographic Experts Group}
\abbreviation [JPEG]       {jpeg}       {Joint Photographic Experts Group}
\abbreviation [LATEX]      {\LaTeX}     {general purpose macro package}
\abbreviation [MAC]        {Macintosh}  {Macintosh hardware platform}
\abbreviation [MACOSX]     {Mac\,OS\,X} {Macintosh operating system version 10}
\abbreviation [MACTEX]     {Mac\TeX}   {\MAC\ \WEBC\ distribution}
\abbreviation [MDFIVE]     {md5}        {MD5 message-digest algorithm}
\abbreviation [METAFONT]   {\MetaFont}  {graphic programming environment, bitmap output}
\abbreviation [METAPOST]   {\MetaPost}  {graphic programming environment, vector output}
\abbreviation [MIKTEX]     {MiK\TeX}    {\WIN\ distribution}
\abbreviation [MLTEX]      {ml\TeX}     {ML\TeX\ extension to \TEX}
\abbreviation [MPTOPDF]    {mptopdf}    {\METAPOST\ to \PDF\ conversion tool}
\abbreviation [MSDOS]      {ms-dos}     {Microsoft DOS platform (Intel)}
\abbreviation [PDFETEX]    {pdfe\TeX}   {\ETEX\ extension producing \PDF\ output}
\abbreviation [PDFLATEX]   {pdf\LaTeX}  {\TEX\ extension producing \PDF\ output (\LATEX\ format loaded)}
\abbreviation [PDFTEX]     {pdf\TeX}    {\TEX\ extension producing \PDF\ output}
\abbreviation [PDF]        {pdf}        {Portable Document Format}
\abbreviation [PERL]       {Perl}       {Perl programming environment}
\abbreviation [PFA]        {PFA}        {Adobe PostScript Font format (ASCII)}
\abbreviation [PFB]        {PFB}        {Adobe PostScript Font format (Binary)}
\abbreviation [PK]         {pk}         {Packed bitmap font}
\abbreviation [PNG]        {png}        {Portable Network Graphics}
\abbreviation [POSIX]      {posix}      {Portable Operating System Interface}
\abbreviation [PROTEXT]    {pro\TeX t}  {\WIN\ \WEBC\ distribution based on \MIKTEX}
\abbreviation [PS]         {ps}         {PostScript}
\abbreviation [POSTSCRIPT] {PostScript} {PostScript}
\abbreviation [PSTOPDF]    {PStoPDF}    {PostScript to \PDF\ converter (on top of \GHOSTSCRIPT)}
\abbreviation [RGB]        {rgb}        {Red Green Blue color specification}
\abbreviation [TCX]        {tcx}        {\TEX\ Character Translation}
\abbreviation [TDS]        {tds}        {\TEX\ Directory Standard}
\abbreviation [TETEX]      {te\TeX}     {\TEX\ distribution for \UNIX\ (based on \WEBC)}
\abbreviation [TEXEXEC]    {\TeX exec}  {\CONTEXT\ command line interface}
\abbreviation [TEXINFO]    {Texinfo}    {generate typeset documentation from info pages}
\abbreviation [TEXUTIL]    {\TeX util}  {\CONTEXT\ utility tool}
\abbreviation [TEX]        {\TeX}       {typographic language and program}
\abbreviation [TEXLIVE]    {\TeX\ Live} {\TeX\ Live distribution (multiple platform)}
\abbreviation [TFM]        {tfm}        {\TEX\ Font Metrics}
\abbreviation [TIF]        {tiff}       {Tagged Interchange File format}
\abbreviation [TUG]        {tug}        {\TEX\ Users Group}
\abbreviation [UNIX]       {Unix}       {Unix platform}
\abbreviation [URL]        {url}        {Uniform Resource Locator}
\abbreviation [WEBC]       {Web2c}      {official multi||platform \WEB\ environment}
\abbreviation [WEB]        {web}        {literate programming environment}
\abbreviation [WIN]        {Windows}    {Microsoft Windows platform}
\abbreviation [XEMTEX]     {XEm\TeX}    {\WIN\ \WEBC\ distribution}
\abbreviation [ZIP]        {zip}        {compressed file format}

%D It makes sense to predefine the name of the author of \PDFTEX, doesn't it?

\def\THANH{H\`an Th\^e\llap{\raise 0.5ex\hbox{\'{}}} Th\`anh}

%D Because they are subjected to change and tend to spoil the visual
%D appearance of the \TEX\ source, \URL's are defined here.

\useURL [ptex_org]      [http://www.pdftex.org] % links to ptex_examples
\useURL [ptex_ctan]     [http://ctan.org/pkg/pdftex]
\useURL [ptex_devel]    [http://foundry.supelec.fr/projects/pdftex]

% where bug reports should go:
\useURL [ptex_bugs]     [mailto:pdftex@tug.org] [] [pdftex@tug.org]
\useURL [ptex_listinfo] [http://lists.tug.org/pdftex]

\useURL [kpathsea]      [http://tug.org/kpathsea]
\useURL [texlive]       [http://tug.org/texlive]
\useURL [web2c]         [http://tug.org/web2c]
\useURL [ctan_systems]  [http://mirror.ctan.org/systems]
\useURL [win32]         [http://mirror.ctan.org/systems/win32]
\useURL [context]       [http://www.pragma-ade.com]
\useURL [tex_showcase]  [http://tug.org/texshowcase]

\useURL [pdfreference] [http://www.adobe.com/devnet/pdf/pdf_reference.html]
\useURL [thanh_truetype_tub] [http://tug.org/TUGboat/tb30-1/tb94thanh.pdf]
\useURL [jbig2enc]           [https://github.com/agl/jbig2enc]

%D The primitive definitions are specified a bit fuzzy using the next set of
%D commands. Some day I'll write some proper macros to deal with this.

% keep next 2 lines as temporary kludge for a while to make \type{} of
% older ConTeXt versions handle these two new primitives; the original
% problem with \type{} is already solved in ConTeXt as of 2006-02-14.
\let\ifpdfabsnum\relax
\let\ifpdfabsdim\relax

\def\Syntax   #1{\strut\kern-.25em{#1}\kern-.25em}
\def\Next       {\crlf\hbox to 2em{}\nobreak}
\def\Sugar    #1{\unskip\unskip\unskip\kern.25em{#1}\kern.25em\ignorespaces}
%
\def\Lbrace     {\Sugar{\tttf\leftargument}}
\def\Literal  #1{\Sugar{\type{#1}}}
\def\Means      {\Sugar{\mathematics{\rightarrow}}}
\def\Modelist #1{\Sugar{\mathematics{(\hbox{#1})}}}
\def\Optional #1{\Sugar{\mathematics{[\hbox{#1}]}}}
\def\Or         {\Sugar{\mathematics{\vert}}}
\def\Rbrace     {\Sugar{\tttf\rightargument}}
\def\Tex      #1{\Sugar{\type{#1}}}
\def\Whatever #1{\Sugar{\mathematics{(\hbox{#1})}}}

% hyphenates
\def\Something#1{\Sugar{\mathematics{\langle}\prewordbreak
                        {#1}\prewordbreak\mathematics{\rangle}}}

\hbadness=10000 % don't care

% Break after these chars in urls, not before.
\sethyphenatedurlafter /
\sethyphenatedurlafter .
\sethyphenatedurlafter _

% introduced
\def\introduced#1{The primitive was introduced in \PDFTEX\ #1.}

% to get bookmarks for primitives like \ifpdfprimitive
\appendtoks\def\tex#1{\string\\#1}\to\simplifiedcommands

%D We typeset the \URL's in a monospaced font.

\setupurl
  [style=type]

%D The basic layout is pretty simple. Because we want non european readers to
%D read the whole text as well, a letter size based alternative is offered
%D too. Mode switching is taken care of at the command line when running
%D \TEXEXEC.

\startmode[letter]

  \setuppapersize
    [letter][letter]

\stopmode

\setuplayout
  [topspace=1.5cm,
   backspace=2.5cm,
   leftmargin=2.5cm,
   width=middle,
   footer=1.5cm,
   header=1.5cm,
   height=middle]

%D For the moment we use the description mechanism to typeset keywords etc.
%D Well, I should have used capitals.

\definedescription
  [description]
  [location=serried,
   width=broad]

\definedescription
  [PathDescription]
  [location=left,
   sample=TEXPSHEADERS,
   width=broad,
   headstyle=type]

\definehead
  [pdftexprimitive]
  [subsubsection]

\setuphead
  [pdftexprimitive]
  [style=,
   before=\blank,
   after=\blank,
   numbercommand=\SubSub]

%D This is typically a document where we want to save pages,
%D so we don't start any matter (part) on a new page.

\setupsectionblock [frontpart] [page=]
\setupsectionblock [bodypart]  [page=]
\setupsectionblock [backpart]  [page=]

%D The \type{\SubSub} command is rather simple and generates a triangle.
%D This makes the primitives stand out a bit more.

\def\SubSub#1{\mathematics{\blacktriangleright}}

%D But, for non Lucida users, the next one is more safe:

\def\SubSub#1{\goforwardcharacter}

%D I could have said:
%D
%D \starttyping
%D \setupsection[section-5][previousnumber=no,bodypartconversion=empty]
%D \setuplabeltext[subsubsection=\mathematics{\blacktriangleright}]
%D \stoptyping
%D
%D but this is less clear.

%D We use white space, but not too much.

\setupwhitespace
  [medium]

\setupblank
  [medium]

%D Verbatim things get a small margin.

\setuptyping
  [margin=standard]

\definetyping
  [esctyping]
\setuptyping
  [esctyping]
  [margin=standard,option=commands,escape=@]

%D Due to the lots of verbatim we will be a bit more tolerant in breaking
%D paragraphs into lines.

\setuptolerance
  [verytolerant,stretch]

%D We put the chapter and section numbers in the margin and use bold fonts.

\setupheads
  [alternative=margin]

\setuphead
  [section]
  [style=\bfb]

\setuphead
  [subsection]
  [style=\bfa]

%D The small table of contents is limited to section titles and is fully
%D interactive.

\setuplist
  [section]
  [criterium=all,
   interaction=all,
   width=2em,
   aligntitle=yes,
   alternative=a]

%D Ah, this manual is in english!

\mainlanguage
  [en]

%D We use Palatino fonts, because they look so well on the screen. The
%D version generated at \PRAGMA\ uses Optima Nova instead. Both fonts are
%D designed by Hermann Zapf.

\setupfonthandling [hz] [min=25,max=25,step=5]

\setupalign
  [hz,hanging]

\startnotmode[atpragma]

  \setupfontsynonym [Serif]            [handling=quality]
  \setupfontsynonym [SerifBold]        [handling=quality]
  \setupfontsynonym [SerifSlanted]     [handling=quality]
  \setupfontsynonym [SerifItalic]      [handling=quality]
  \setupfontsynonym [SerifBoldSlanted] [handling=quality]
  \setupfontsynonym [SerifBoldItalic]  [handling=quality]

  %setupfontsynonym [Mono]             [handling=quality] % sloooow

  % We use adobe metrics instead of urw metrics because tetex only
  % ships the former. Beware, these metrics differ!

  \loadmapfile[context-base.map]
  \usetypescript [adobekb] [\defaultencoding]
  \usetypescript [palatino][\defaultencoding]

  \setupbodyfont
    [palatino,10pt]

  \definefontsynonym[TitleFont][SerifBold]

\stopnotmode

\startmode[atpragma]

  \usetypescriptfile[type-ghz]

  \setupfontsynonym [Sans]            [handling=quality]
  \setupfontsynonym [SansBold]        [handling=quality]
  \setupfontsynonym [SansSlanted]     [handling=quality]
  \setupfontsynonym [SansItalic]      [handling=quality]
  \setupfontsynonym [SansBoldSlanted] [handling=quality]
  \setupfontsynonym [SansBoldItalic]  [handling=quality]

  %setupfontsynonym [Mono]             [handling=quality] % sloooow

  \definetypeface[optima][ss][sans][optima-nova] [default][encoding=\defaultencoding]
  \definetypeface[optima][tt][mono][latin-modern][default][encoding=\defaultencoding,rscale=1.1]

  \setupbodyfont[optima,10pt,ss]

  \definefontsynonym[TitleFont][SansBold]

\stopmode

%D This document is mildly interactive. Yes, Thanh's name will end up ok in
%D the document information data.

\setupinteraction
  [state=start,
   style=normal,
   color=,
   page=yes,
   openaction=firstpage,
   title=pdfTeX users manual,
   author={H\`an Th\^e Th\`anh, Sebastian Rahtz, Hans Hagen, Hartmut Henkel,
   Pawe\l\ Jackowski, Martin Schr\"oder, Karl Berry}]

\setupinteractionscreen % still needed?

\startnotmode[screen]

\setupinteractionscreen
  [option=bookmark]

\stopnotmode

%D Some headers and footers will complete the layout.

\setupheadertexts
  [The pdf\TeX\ user manual]

\setupfootertexts
  [pagenumber]

%D For Tobias Burnus, who loves bookmarks, I've enabled them.

\placebookmarks
  [section,subsection,pdftexprimitive]

%D Let's also define a screen layout:

\startmode[screen] \environment pdftex-i \stopmode

%D We auto-cross link the paper and screen version:

\startnotmode[screen]

%\coupledocument
%  [screenversion]
%  [pdftex-s]
%  [section,subsection,subsubsection,pdftexprimitive]
%  [the screen version]

\setuphead
  [section,subsection,subsubsection,pdftexprimitive]
  [file=screenversion]

\setuplist
  [section]
  [alternative=c]

\stopnotmode

%D The text starts here!

\starttext

%D Being lazy, I don't split the file, so paper and screen get
%D mixed in one document.

\startmode[screen] \getbuffer[screen] \stopmode

\startnotmode[screen]

%D But first we typeset the title page. It has a background. This
%D background, showing a piece of \PDF\ code, will be referred to
%D later on.

%D We can use more modern \CONTEXT\ features, but for the moment
%D stick to the old style and methods.

\NormalizeFontWidth
  \TitleFont
  {\hskip.5mm The pdf\TeX\ user manual} % \hskip is fake offset
  \paperheight
  {TitleFont}

\setupbackgrounds
  [page]
  [background={title,joke,names,content}]

\defineoverlay
  [title]
  [\hbox to \paperwidth
     {\hfill
      \TitleFont\setstrut
      \rotate{The pdf\TeX\ user manual}%
      \hskip.5cm}] % \dp\strutbox}]

% \defineoverlay
%   [joke]
%   [\hbox to \paperwidth
%      {\TitleFont\setstrut
%       \dimen0=\overlaywidth
%       \advance\dimen0 by -\ht\strutbox
%       \advance\dimen0 by -\dp\strutbox
%       \advance\dimen0 by -1cm
%       \dimen2=\overlayheight
%       \advance\dimen2 by -1cm
%       \hskip.5cm\expanded{\externalfigure
%          [pdftex-z.pdf]
%          [width=\the\dimen0,height=\the\dimen2]}%
%       \hfill}]

% \defineoverlay
%   [names]
%   [\vbox to \paperheight
%      {\dontcomplain
%       \TitleFont\setstrut
%       \setbox0=\hbox{\TeX}%
%       \hsize\paperwidth
%       \rightskip\ht0
%       \advance\rightskip\dp\strutbox
%       \advance\rightskip\dp\strutbox
%       \bfa \setupinterlinespace
%       \vfill
%       \hfill \THANH \endgraf
%       \hfill Sebastian Rahtz \endgraf
%       \hfill Hans Hagen
%       \hfill Hartmut Henkel
%       \hfill Pawe\l\ Jackowski
%       \vskip 1ex
%       \hfill \currentdate
%       \vskip 2ex}]

\defineoverlay
  [content]
  [\overlaybutton{contents}]

% title page

\definelayout
  [titlepage]
  [backspace=.5cm,
   cutspace=3.5cm,
   topspace=.5cm,
   bottomspace=.5cm,
   header=0pt,
   footer=0pt,
   lines=\getvariable{pdftex}{titlepagelines},
   grid=yes,
   width=middle]

\definecolumnset
  [titlepage]
  [n=2,distance=0.2cm]

\start

    \chardef\fontdigits=2
    \switchtobodyfont[10pt,tt]
    \setupinterlinespace
      [line=\dimexpr((\paperheight-1cm)/\getvariable{pdftex}{titlepagelines})]
    \setuptyping[margin=,color=]
    \setuplayout[titlepage]

    \startcolumnset[titlepage]
      \typefile{pdftex-w.txt}
    \stopcolumnset

    \page
    \setuplayout

\stop

% \startstandardmakeup
%
% %D The titlepage is generated using the background overlay mechanism. This
% %D saves me the trouble of determining funny skips and alignments. So no text
% %D goes here.
%
% \stopstandardmakeup

\setupbackgrounds
  [page]
  [background=]

%D The inside title page is as follows.

\startstandardmakeup[footerstate=none]

  \dontcomplain

  \setupalign[left]

  \start

    \bfd The pdf\TeX\ user manual

    \bfa \setupinterlinespace

    \vskip4ex

    \THANH\par
    Sebastian Rahtz\par
    Hans Hagen\par
    Hartmut Henkel\par
    Pawe\l\ Jackowski\par
    Martin Schr\"oder\par

    \vskip3ex

    \rcsmonthname\ \rcsday, \rcsyear\par
    \vskip1ex
    Rev.\ \rcsrevision

  \stop

  \vfill

  \startlines
    The title page is the result of
    this plain \TeX\ text:
  \stoplines

  \blank[2*big] \setuptyping[after=]

  \starttyping
    \pdfoutput=1
    \pdfcompresslevel=0
    \pdfobjcompresslevel=0
    \pdfmapline{ptmr8r Times-Roman 2 <8r.enc}
    \font\tenrm=ptmr8r
    \tenrm
    Welcome to pdf\TeX!
    \bye
  \stoptyping

\stopstandardmakeup

\setuppagenumber[number=1] % added in case of single sided usage

%D So far for non screen mode.

\stopnotmode

%D We start with a small table of contents, typeset in double column format.

\startfrontmatter

\subject[contents]{Contents}

\startcolumns[distance=3em]
  \placelist[section]
\stopcolumns

\stopfrontmatter

%D And here it is: the main piece of text.

\startbodymatter

%***********************************************************************

\section{Introduction}

The main purpose of the \PDFTEX\ project is to create and maintain
an extension of \TEX\ that can produce \PDF\ directly from \TEX\
source files and improve|/|enhance the result of \TEX\ typesetting
with the help of \PDF. When \PDF\ output is not selected, \PDFTEX\
produces standard \DVI\ output, otherwise it generates \PDF\ output
that looks identical to the \DVI\ output. An important aspect of this
project is to investigate alternative justification algorithms (\eg\
a font expansion algorithm akin to the \HZ\ micro||typography algorithm by
Prof.~Hermann Zapf).

\PDFTEX\ is based on the original \TEX\ sources and \WEBC, and has been
successfully compiled on \UNIX, \WIN\ and \MSDOS\ systems. It is
actively maintained, with new features trickling in. Great care is taken
to keep new \PDFTEX\ versions backward compatible with earlier ones.

For some years there has been a \quote {conservative} successor to \TEX\
available, called \ETEX. Because mainstream macro packages such as
\LATEX\ have started supporting this welcome extension, the \ETEX\
functionality has also been integrated into the \PDFTEX\ code. For a
while (\TEXLIVE~2004 and~2005) \PDFTEX\ therefore came in two flavours:
the \ETEX\ enabled \PDFETEX\ engine and the standard one, \PDFTEX. The
ability to produce both \PDF\ and \DVI\ output made \PDFETEX\ the
primary \TEX\ engine in these distributions. Since \PDFTEX\ version 1.40
now the \ETEX\ extensions are part already of the \PDFTEX\ engine, so
there is no longer any need to ship \PDFETEX. The \ETEX\ functionality
of \PDFTEX\ can be disabled if not required. Other extensions are
\MLTEX\ and \ENCTEX; these are also included in the current \PDFTEX\
code.

\PDFTEX\ is maintained by \THANH, Martin Schr\"oder, and others. The
\PDFTEX\ homepage is \from [ptex_org]. Please send \PDFTEX\ comments and
bug reports to the mailing list \from [ptex_bugs] (\from [ptex_listinfo]).

%***********************************************************************

\subsection{About this manual}

This manual revision (\rcsrevision) is intended to cover
\PDFTEX\ development up to version \currentpdftex.  The primary
repository for the manual and its sources is at \from[ptex_devel].
Copies in \PDF\ format can also be found on \CTAN\ via \from[ptex_ctan].

Thanks to the many people who have contributed to the manual.
New errors might have slipped in afterwards by the editor.
Please send questions or suggestions by email to \from[ptex_bugs].

%***********************************************************************

\subsection{Legal Notice}

Copyright \copyright\ 1996||2017  \THANH.
Permission is granted to copy, distribute and/or modify this document
under the terms of the GNU Free Documentation License, Version 1.2
or any later version published by the Free Software Foundation;
with no Invariant Sections, no Front-Cover Texts, and no Back-Cover Texts.
A copy of the license is included in the section entitled ``GNU
Free Documentation License''.

%***********************************************************************

\section{About \PDF}

The cover of this manual lists an almost minimal \PDF\ file generated by
\PDFTEX, from the corresponding source on the next page.  Since
compression is not enabled, such a \PDF\ file is rather verbose and
readable.  The first line specifies the version used.  \PDF\ viewers are
supposed to silently skip over all elements they cannot handle.

A \PDF\ file consists of objects. These objects can be recognized by their
number and keywords:

\starttyping
9 0 obj << /Type /Catalog /Pages 5 0 R >> endobj
\stoptyping

Here \typ{9 0 obj ... endobj} is the object capsule. The first number
is the object number. The sequence \type{5 0 R} is an object reference,
a pointer to another object (no.~5). The second number (here a zero)
is currently not used in \PDFTEX; it is the version number of the
object. It is for instance used by \PDF\ editors, when they replace
objects by new ones.

When a viewer opens a \PDF\ file, it goes to the end of the file,
looking for the keyword \type{startxref}. The number after
\type{startxref} gives the absolute position (byte offset from the file
start) of the so||called \quote {object cross-reference table} that
begins with the keyword \type{xref}. This table in turn tells the byte
offsets of all objects that make up the \PDF\ file, providing fast
random access to the individual objects (here the \type{xref} table
shows 11~objects, numbered from~0 to~10; the object no.~0 is always
unused). The actual starting point of the file's object structure is
defined after the \type{trailer}: the \type{/Root} entry points to the
\type{/Catalog} object (no.~9). In this object the viewer can find the
pointer \type{/Pages} to the page list object (no.~5). In our example we
have only one page. The trailer also holds an \type{/Info} entry, which
points to an object (no.~10) with a bit more about the document. Just
follow the thread:

\startnarrower
\type{/Root}     $\longrightarrow$ object~9 $\longrightarrow$
\type{/Pages}    $\longrightarrow$ object~5 $\longrightarrow$
\type{/Kids}     $\longrightarrow$ object~2 $\longrightarrow$
\type{/Contents} $\longrightarrow$ object~3
\stopnarrower

As soon as we add annotations, a fancy word for hyperlinks and the like,
some more entries will be present in the catalog. We invite users to
take a look at the \PDF\ code of this file to get an impression of that.

The page content is a stream of drawing operations. Such a stream
can be compressed, where the level of compression can be set with
\type{\pdfcompresslevel} (compression is switched off for the title
page). Let's take a closer look at this stream in object~3. Often
(but not in our example) there is a transformation matrix, six numbers
followed by \type{cm}. As in \POSTSCRIPT, the operator comes after the
operands. Between \type{BT} and \type{ET} comes the text. A font is
selected by a \type{Tf} operator, which is given a font resource name
\type{/F..} and the font size. The actual text goes into \type{()}
bracket pairs so that it creates a \POSTSCRIPT\ string. The numbers
in bracket pairs provide horizontal movements like spaces and
fine glyph positioning (kerning). When one analyzes a file produced by
a less sophisticated typesetting engine, whole sequences of words can
be recognized. In \PDF\ files generated by \PDFTEX\ however, many words
come out rather fragmented, mainly because a lot of kerning takes place;
in our example the \type{80} moves the text \type{(elcome)} left towards
the letter \type{(W)} by 80/1000 of the font size. \PDF\ viewers in search
mode simply ignore the kerning information in these text streams. When
a document is searched, the search engine reconstructs the text from these
(string) snippets.

Every \type{/Page} object points also to a \type{/Resources} object
(no.~1) that gives all ingredients needed to assemble the page. In our
example only a \type{/Font} object (no.~4) is referenced, which in turn
tells that the text is typeset in \type{/Font} \type{/Times-Roman}. The
\type{/Font} object points also to a \type{/Widths} array (object no.~7)
that tells for each character by how much the viewer must move forward
horizontally after typesetting a glyph. More details about the font
can be found in the \type{/FontDescriptor} object (no.~8); if a font
file is embedded, this object points to the font program stream. But as
the Times-Roman font used for our example is one of the 14 so||called
standard fonts that should always be present in any \PDF\ viewer and
therefore need not be embedded in the \PDF\ file, it is left out here
for brevity. However, when we use for instance a Computer Modern Roman
font, we have to make sure that this font is later available to the \PDF\
viewer, and the best way to do this is to embed the font.
It's highly recommended nowadays to embed even the standard fonts;
you can't know how it looks exactly at the viewer side unless you embed
every font.

In this simple file we don't specify in what way the file should be opened,
for instance full screen or clipped. A closer look at the page object no.~2
(\typ{/Type /Page}) shows that a mediabox (\typ{/MediaBox}) is part of the
page description. A mediabox acts like the (high-resolution) bounding box
in a \POSTSCRIPT\ file. \PDFTEX\ users can add dictionary entries to page
objects with the \type{\pdfpageattr} primitive.

Although in most cases macro packages will shield users from these
internals, \PDFTEX\ provides access to many of the entries described
here, either automatically by translating the \TEX\ data structures into
\PDF\ ones, or manually by pushing entries to the catalog, page, info or
self-created objects. One can for instance create an object by using
\type{\pdfobj}, after which \type{\pdflastobj} returns its number. So

\starttyping
\pdfobj{<< /Type/ExtGState /LW 2 >>}
\stoptyping

inserts an object into the \PDF\ file (it creates a ``graphics state''
object setting the line width to 2~units), and \type{\pdflastobj} now
returns the number \PDFTEX\ assigned to this object. Unless objects are
referenced by others, they will just end up as isolated entities, not
doing any real harm but bloating the \PDF\ file.

In general this rather direct way of pushing objects in the \PDF\ files
by primitives like \type{\pdfobj} is not very useful, and only makes
sense when implementing, say, fill||in field support or annotation
content reuse. We will come to that later.

Of course, this is just the barest introduction to \PDF\ format.  For
those who want to learn more about the gory \PDF\ details, the best bet
is to read the \PDFReference. You can download this book as a big \PDF\
file from Adobe's \PDF\ Technology Center, \from[pdfreference] --- or
get the heavy paper version.

We now turn to specifics of \PDFTEX.

%***********************************************************************

\section{Getting started}

This section describes the steps needed to get \PDFTEX\ running on
a system where \PDFTEX\ is not yet installed. Nowadays virtually all
\TEX\ distributions have \PDFTEX\ as a component, such as \TEXLIVE,
\MIKTEX, \TETEX, \XEMTEX, \PROTEXT, and \MACTEX. The ready to run
\TEXLIVE\ distribution comes with \PDFTEX\ versions for many \UNIX,
\WIN, and \MACOSX\ systems; more information can be found at
\hbox{\from[texlive].}  There are also \WIN-specific distributions which
contain \PDFTEX, under \from[win32]: \MIKTEX\ by Christian Schenk, and
\PROTEXT\ (based on \MIKTEX) by Thomas Feuerstack. When you use any
of these distributions, you don't need to bother with the \PDFTEX\
installation procedure in the next sections.

If there is no precompiled \PDFTEX\ binary for your system, or the version
coming with a distribution is not the current one and you would like to
try out a fresh \PDFTEX\ immediately, you will need to build \PDFTEX\
from sources; read on. You should already have a working \TEX\ system,
\eg\ \TEXLIVE\ or \TETEX, into which the freshly compiled \PDFTEX\ will
be integrated. Note that the installation description in this manual
is \WEBC||specific.

%***********************************************************************

\subsection{Getting sources and binaries}

The latest sources of \PDFTEX\ are distributed for compilation on \UNIX\
systems (including \GNU/Linux), and \WIN\ systems.  The primary home
page is \from[ptex_org], where you also find bug tracking information.
Development sources are at \from[ptex_devel].  Precompiled \PDFTEX\
binaries for various systems might be available in subdirectories below
\from[ctan_systems], or via \TEX\ distribution web pages.

%***********************************************************************

\subsection{Compiling}

The compilation is expected to be easy on \UNIX||like systems and
can be described best by example. Assuming that the file \filename
{pdftex.zip} is downloaded to some working directory, \eg\
\filename {\$HOME/pdftex}, on a \UNIX\ system the following steps are
needed to compile \PDFTEX:

\startesctyping
cd pdftex.../source
./build-pdftex.sh
\stopesctyping

The binary \filename{pdftex} is then built in the subdirectory
\filename{build/texk/web2c}. 

The obsolescent binary \filename{pdfetex} is still generated for backward
compatibility, but since version 1.40 it is just a file copy
of the file \filename{pdftex}.

As well as the main \filename{pdftex} binary, binaries for the utilities
\filename{pdftosrc} and \filename{ttf2afm} are generated.

Incidentally, for \PDFTEX\ maintains, a sibling script to
\type{build-pdftex.sh} is included, namely \type{sync-pdftex.sh}, which
syncs changes from a \TEXLIVE\ source repository to a \PDFTEX\ source
repository.  Read the script before using it.  And don't use it unless
you understand what you read.

%***********************************************************************

\subsection{Placing files}

The next step is to put the freshly compiled \filename{pdftex},
\filename{pdftosrc}, and \filename{ttf2afm} binaries into the binary
directory (\eg\ for a typical \TEXLIVE\ system, and on the appropriate
platform) \filename{/usr/local/texlive/\rcsyear/bin/x86_64-linux}.

If you're doing this into a live hierarchy, don't forget to do a
\type{texconfig-sys init} afterwards, so that all formats are
regenerated system-wide with the fresh \filename{pdftex} binary.

%***********************************************************************

\subsection{Setting search paths}

\WEBC||based programs, including \PDFTEX, use the \WEBC\ run||time
configuration file called \filename {texmf.cnf}.  The location
of this file is the appropriate position within the \TDS\ tree
relative to the place of the \PDFTEX\ binary; on a \TEXLIVE\ system,
\filename{texmf.cnf} is typically located either in directory
\filename{texmf/web2c} or \filename{texmf-local/web2c}.  The path to
file \filename{texmf.cnf} can also be set up by the environment variable
\type{TEXMFCNF}.

The \filename{texmf.cnf} files coming with the major \TEX\ distributions
should already be set up for normal use, so you shouldn't need to edit
it. You might still like to read it to see where the various bits and
pieces are going.

\PDFTEX\ uses the search path variables shown in
\in{table}[tbl:spathvar], among others.

\startbuffer
\starttabulate[|l|l|]
\HL
\NC \bf used for        \NC \bf texmf.cnf        \NC\NR
\HL
\NC output files        \NC \type{TEXMFOUTPUT}   \NC\NR
\NC input files, images \NC \type{TEXINPUTS}     \NC\NR
\NC format files        \NC \type{TEXFORMATS}    \NC\NR
\NC \TeX\ pool files    \NC \type{TEXPOOL}       \NC\NR
\NC encoding files      \NC \type{ENCFONTS}      \NC\NR
\NC font map files      \NC \type{TEXFONTMAPS}   \NC\NR
\NC \TFM\ files         \NC \type{TFMFONTS}      \NC\NR
\NC virtual fonts       \NC \type{VFFONTS}       \NC\NR
\NC Type~1 fonts        \NC \type{T1FONTS}       \NC\NR
\NC TrueType fonts      \NC \type{TTFONTS}       \NC\NR
\NC OpenType fonts      \NC \type{OPENTYPEFONTS} \NC\NR
\NC bitmap fonts        \NC \type{PKFONTS}       \NC\NR
\HL
\stoptabulate
\stopbuffer

\placetable[here][tbl:spathvar]
  {The principal \WEBC\ variables.}
  {\getbuffer}

\PathDescription {TEXMFOUTPUT} Normally, \PDFTEX\ puts its output files
in the current directory, overridden by the \type{-output-directory}
option. If any output file cannot be opened there, it tries to open it
in the environment variable \type{TEXMFOUTPUT}, if that is set. There is
no default value for that variable. For example, if \type{TEXMFOUTPUT}
has the value \type{/tmp}, and you run \type{pdftex paper} when the
current directory is not writable, \PDFTEX\ attempts to create
\type{/tmp/paper.log} (and \type{/tmp/paper.pdf}, etc.)

\PathDescription {TEXINPUTS} This variable specifies where \PDFTEX\ finds
its input files. Image files are considered
input files and searched for along this path.

\PathDescription {TEXFORMATS} Search path for format (\type{.fmt}) files.

\PathDescription {TEXPOOL} Search path for pool (\type{.pool}) files; no
longer used, since the pool file (program strings) are compiled into
the binary.

\PathDescription {ENCFONTS} Search path for encoding (\type{.enc}) files.

\PathDescription {TEXFONTMAPS} Search path for font map (\type{.map}) files.

\PathDescription {TFMFONTS} Search path for font metric (\type{.tfm}) files.

\PathDescription {VFFONTS} Search path for virtual font (\type{.vf})
files.  Virtual fonts are fonts made up of other fonts.
Because \PDFTEX\ produces the
final output code, it must consult those files.

\PathDescription {T1FONTS} Search path for Type~1 font files (\type{.pfa}
and \type{.pfb}).  These outline (vector) fonts are to be preferred over
bitmap \PK\ fonts. In most cases Type~1 fonts are used and this variable
tells \PDFTEX\ where to find them.

\PathDescription {TTFONTS,\hfil\break \hbox{OPENTYPEFONTS}} Search paths
for TrueType (\type{.ttf}) and OpenType (\type{.otf}) font files.  Like
Type~1 fonts, TrueType and OpenType fonts are also outlines.

\PathDescription {PKFONTS} Search path for packed (bitmap) font
(\type{.pk}) files.
Unfortunately bitmap fonts are still displayed poorly by some \PDF\
viewers, so when possible one should use outline fonts. When no outline
is available, \PDFTEX\ tries to locate a suitable \PK\ font (or invoke
a process that generates it).

\PathDescription{TEXFONTS} Fallback for all the font paths, so that if
you want to look in a particular directory for fonts on a given run, you
can set that one variable.

Many more variables may be consulted, and there are many details to
file name lookups.  See the Kpathsea manual (\from [kpathsea]).

%***********************************************************************

\subsection[cfg]{The \PDFTEX\ configuration}

We must keep in mind that, as opposed to \TEX\ with its \DVI\ output,
the \PDFTEX\ program does not have a separate postprocessing stage to
transform the \TEX\ input into final \PDF. As a consequence, all data
needed for building a ready \PDF\ page must be available during the
\PDFTEX\ run, in particular information on media dimensions and offsets,
graphics files for embedding, and font information (font files,
encodings).

When \TEX\ builds a page, it places items relative to the top left page
corner (the \DVI\ reference point). Separate \DVI\ postprocessors allow
specifying the paper size (\eg\ \quote {A4} or \quote{letter}), so
that this reference point is moved to the correct position on the paper,
and the text ends up at the right place.

In \PDF, the paper dimensions are part of the page definition, and
\PDFTEX\ therefore requires that they be defined at the beginning of
the \PDFTEX\ run. As with pages described by \POSTSCRIPT, the \PDF\
reference point is in the lower||left corner.

Formerly, these dimensions and other \PDFTEX\ parameters were read
in from a configuration file named \filename{pdftex.cfg}, which had
a special (non-\TEX) format, at the start of processing.  Nowadays such
a file is ignored by \PDFTEX.  Instead, the page dimensions and offsets,
as well as many other parameters, can be set by \PDFTEX\ primitives
during the \PDFTEX\ format building process, so that the settings are
dumped into the fresh format and consequently will be used when \PDFTEX\
is later called with that format. All settings from the format can
still be overridden during a \PDFTEX\ run by using the same primitives.
This new configuration concept is a more unified approach, as it avoids
the configuration file with a special format.

A list of \PDFTEX\ primitives relevant to setting up the \PDFTEX\ engine
is given in \in{table}[tbl:configparms]. All primitives are described in
detail within later sections. \in{Figure}[in:pdftexconfig] shows a recent
configuration file (\type{pdftexconfig.tex}) in \TEX\ format, using the
primitives from \in{table}[tbl:configparms], which typically is read
in during the format building process. It enables \PDF\ output, sets paper
dimensions and the default pixel density for \PK\ font inclusion. The default
values are chosen so that \PDFTEX\ often can be used (\eg\ in \type{-ini} mode)
even without setting any parameters.

\startbuffer
\starttabulate[|l|l|l|l|l|]
\HL
\NC \bf internal name              \NC \bf type  \NC\bf default\NC\bf comment\NC\NR
\HL
\NC \type{\pdfoutput}              \NC integer   \NC      0 \NC \DVI                 \NC\NR
\NC \type{\pdfadjustspacing}       \NC integer   \NC      0 \NC off                  \NC\NR
\NC \type{\pdfcompresslevel}       \NC integer   \NC      9 \NC best                 \NC\NR
\NC \type{\pdfobjcompresslevel}    \NC integer   \NC      0 \NC no object streams    \NC\NR
\NC \type{\pdfdecimaldigits}       \NC integer   \NC      4 \NC max.                 \NC\NR
\NC \type{\pdfimageresolution}     \NC integer   \NC     72 \NC dpi                  \NC\NR
\NC \type{\pdfpkresolution}        \NC integer   \NC      0 \NC 72\,dpi              \NC\NR
\NC \type{\pdfpkmode}              \NC token reg.\NC  empty \NC mode set in \type{mktex.cnf} \NC\NR
\NC \type{\pdfuniqueresname}       \NC integer   \NC      0 \NC                      \NC\NR
\NC \type{\pdfprotrudechars}       \NC integer   \NC      0 \NC                      \NC\NR
\NC \type{\pdfgentounicode}        \NC integer   \NC      0 \NC                      \NC\NR
\NC \type{\pdfminorversion}        \NC integer   \NC      4 \NC \PDF\ 1.4            \NC\NR
\NC \type{\pdfpagebox}             \NC integer   \NC      0 \NC                      \NC\NR
\NC \type{\pdfforcepagebox}        \NC integer   \NC      0 \NC                      \NC\NR
\NC \type{\pdfinclusionerrorlevel} \NC integer   \NC      0 \NC                      \NC\NR
%-----------------------------------------------------------------------
\NC \type{\pdfhorigin}             \NC dimension \NC  1\,in \NC                      \NC\NR
\NC \type{\pdfvorigin}             \NC dimension \NC  1\,in \NC                      \NC\NR
\NC \type{\pdfpagewidth}           \NC dimension \NC  0\,pt \NC                      \NC\NR
\NC \type{\pdfpageheight}          \NC dimension \NC  0\,pt \NC                      \NC\NR
%\NC \type{\pdffirstlineheight}     \NC dimention \NC  -1000\,pt \NC                  \NC\NR
%\NC \type{\pdflastlinedepth}       \NC dimention \NC  -1000\,pt \NC                  \NC\NR
%\NC \type{\pdfeachlineheight}      \NC dimention \NC  -1000\,pt \NC                  \NC\NR
%\NC \type{\pdfeachlinedepth}       \NC dimention \NC  -1000\,pt \NC                  \NC\NR
\NC \type{\pdflinkmargin}          \NC dimension \NC  0\,pt \NC                      \NC\NR
\NC \type{\pdfdestmargin}          \NC dimension \NC  0\,pt \NC                      \NC\NR
\NC \type{\pdfthreadmargin}        \NC dimension \NC  0\,pt \NC                      \NC\NR
\NC \type{\pdfmapfile}             \NC text      \NC \filename{pdftex.map} \NC not dumped\NC\NR
\HL
\stoptabulate
\stopbuffer

\placetable[here][tbl:configparms]
  {The set of \PDFTEX\ configuration parameters.}
  {\getbuffer}

\startbuffer
\tx\setupinterlinespace
\startframedtext
\starttyping
% tex-ini-files 2016-04-15: pdftexconfig.tex

% Load shared (PDF) settings in pdfTeX

% Enable PDF output
\pdfoutput           = 1

% Paper size: dimensions given in absolute terms
\pdfpageheight       = 11 true in
\pdfpagewidth        = 8.5 true in

% Enable PDF 1.5 output and thus more compression
\pdfminorversion     = 5
\pdfobjcompresslevel = 2

% Low-level settings unlikely ever to need to change
\pdfcompresslevel    = 9
\pdfdecimaldigits    = 3
\pdfpkresolution     = 600
\pdfhorigin          = 1 true in
\pdfvorigin          = 1 true in
\stoptyping
\stopframedtext
\stopbuffer

\placefigure[here][in:pdftexconfig]
  {\PDFTEX\ configuration file for \TEXLIVE\ (\filename{pdftexconfig.tex}).}
  {\getbuffer}

Independent of whether such a configuration file is read or not, the
first action in a \PDFTEX\ run is that the program reads the global \WEBC\
configuration file (\filename{texmf.cnf}), which is common to all programs
in the \WEBC\ system. This file mainly defines file search paths, the
memory layout (\eg\ string pool and hash size), and other general parameters.

%***********************************************************************

\subsection{Creating format files}

\startbuffer
\tx\setupinterlinespace
\startframedtext
\starttyping
% Thomas Esser, 1998. public domain.
\input etex.src
\dump
\endinput
\stoptyping
\stopframedtext
\stopbuffer

\placefigure[here][in:etexini]
  {File \type{etex.ini} for the plain \ETEX\ format with \DVI\ output.}
  {\getbuffer}

\startbuffer
\tx\setupinterlinespace
\startframedtext
\starttyping
% Thomas Esser, 1998. public domain.
% This is used for pdftex and pdfetex, which are now identical: both
% with e-TeX extensions, both with pdf output.
\input pdftexconfig.tex
\input etex.src
\input pdftexmagfix.tex
\dump
\endinput
\stoptyping
\stopframedtext
\stopbuffer

\placefigure[here][in:pdfetexini]
  {File \type{pdfetex.ini} for plain \ETEX\ with \PDF\ output.}
  {\getbuffer}

\startbuffer
\tx\setupinterlinespace
\startframedtext
\starttyping
% Thomas Esser, 1998. public domain.
\input pdftexconfig.tex
\scrollmode
\input latex.ltx
\endinput
\stoptyping
\stopframedtext
\stopbuffer

\placefigure[here][in:pdflatexini]
  {File \type{pdflatex.ini} for the \LATEX\ format with \PDF\ output.}
  {\getbuffer}

The \PDFTEX\ engine supports building formats for \DVI\ and \PDF\ output
in the same way as the classical \TEX\ engine does for \DVI.  Format
generation is enabled by the \type{-ini} option. The default mode (\DVI\
or \PDF) can be chosen either on the command line by setting the option
\type{-output-format} to \type{dvi} or \type{pdf}, or by setting the
\type{\pdfoutput} parameter. The format file then inherits this setting,
so that a later invocation of \PDFTEX\ with this format starts in the
preselected mode (which still can be overridden). A format file can be
read in only by the engine that has generated it; a format incompatible
with an engine leads to a fatal error.

It is customary to package the configuration and macro file input
into a \type{.ini} file.  \Eg, the file \type{etex.ini} in
\in{figure}[in:etexini] is for generating an \ETEX\ format with \DVI\
output.  It has been traditional for many years to generate
\type{etex.fmt} with \PDFTEX\ rather than the original \ETEX, because
\PDFTEX\ contains a few additional programming and other
non-\PDF-related features on which people have come to rely.

The \type{pdfetex.ini} file \in{figure}[in:etexini] shows the
corresponding format with \PDF\ output by default; this is what creates
the format file read when \type{pdftex} is normally invoked.

Finally, \type{pdflatex.ini} \in{figure}[in:pdflatexini] shows how the
\LATEX\ format with \PDF\ output by default is generated.

The corresponding \PDFTEX\ calls for format generation are:

\starttyping
pdftex -ini *etex.ini
pdftex -ini *pdfetex.ini
pdftex -ini *pdflatex.ini
\stoptyping

These calls produce format files \filename{etex.fmt},
\filename{pdfetex.fmt}, and \filename{pdflatex.fmt}, as the default
format file name is taken from the input file name. You can overrule
this with the \type{-jobname} option. The asterisk \type{*} before the
file name is a unusual feature, only in \type{-ini} mode, which causes
the \PDFTEX\ engine to enable \ETEX\ features.

\subsection{Testing the installation}

When everything is set up, you can test the installation.  A simple test
of plain \PDFTEX\ is:

\starttyping
pdftex story \\end
\stoptyping

This should typeset the famous one-page short story by A.U. Thor.

A more thorough and descriptive test is the plain \TEX\ test file
\filename{samplepdf.tex}, available in the distribution in the
\type{samplepdftex/} directory. Process this file by typing:

\starttyping
pdftex samplepdf
\stoptyping

If the installation is ok, this should produce a file called
\filename{samplepdf.pdf}. The file \filename {samplepdf.tex} is a good
place to look for examples of how to use \PDFTEX's primitives.

%***********************************************************************

\subsection{Common problems}

The most common problem with installations is that \PDFTEX\ complains
that something cannot be found. In such cases make sure that
\type{TEXMFCNF} is set correctly, so \PDFTEX\ can find \filename
{texmf.cnf}. The next best place to look|/|edit is the file
\type{texmf.cnf}. When still in deep trouble, set
\type{KPATHSEA_DEBUG=255} before running \PDFTEX\ or use the option
\type{-kpathsea-debug 255}. This will cause \PDFTEX\ to write a lot of
debugging information that can be useful to trace problems. More options
can be found in the \WEBC\ documentation.

Variables in \filename {texmf.cnf} can be overwritten by environment
variables. Here are some of the most common problems you can encounter when
getting started:

\startitemize

\head  \type{I can't find the format file `pdftex.fmt'!} \crlf
       \type{I can't find the format file `pdflatex.fmt'!}

       The format file is not created (see above how to do that) or
       is not properly placed. Make sure that \type{TEXFORMATS} in
       \filename {texmf.cnf} contains the path to \filename {pdftex.fmt}
       or \filename {pdflatex.fmt}.

\head \type{Fatal format file error; I'm stymied}

       This appears \eg\ if you forgot to regenerate the \type{.fmt}
       files after installing a new version of the \PDFTEX\ binary.
       The first line tells by which engine the offending format was generated.

\head  \PDFTEX\ cannot find one or more map files (\type{*.map}),
       encoding vectors (\type{*.enc}), virtual fonts, Type~1 fonts,
       TrueType or OpenType fonts, or some image file.

       Make sure that the required file exists and the corresponding variable
       in \filename {texmf.cnf} contains a path to the file. See above which
       variables \PDFTEX\ needs apart from the ones \TEX\ uses.

       When you have installed new fonts, and your \PDF\ viewer complains
       about missing fonts, you should take a look at the log file produced
       by \PDFTEX. Missing fonts, map files, encoding vectors as well as
       missing characters (glyphs) are reported there.

\stopitemize

Normally the page content takes one object. This means that one seldom
finds more than a few hundred objects in a simple file. This \PDFTEX\
manual for instance uses approx.~750 objects. In more complex applications
this number can grow quite rapidly, especially when one uses a lot of
widget annotations, shared annotations or other shared things. In any
case \PDFTEX's internal object table size will automatically grow to the
required size (the parameter \type{obj_tab_size} for manual control of
the object table size is now obsolete and ignored).

%***********************************************************************

\section{Invoking \PDFTEX}

\PDFTEX\ has many command line options.  With the exception of the
simple and rarely-used \type{-draftmode} and \type{-output-format}
options, they are all inherited from the common framework for \TeX\
engines as implemented in \WEBC\ (\from [web2c] has the manual).

The same commonality holds for environment variables; see the section
``Setting search paths'' above for an overview.  Two additional
environment variables need more description here: first,
\type{SOURCE_DATE_EPOCH} (introduced in version 1.40.17, 2016).  If this
is set, it must be a positive integer (with one trivial exception: if it
is set but empty, that is equivalent to 0).  Non-integer values cause a
fatal error.  The value is used as the current time in seconds since the
usual Unix ``epoch'': the beginning of 1970-01-01, UTC.  Thus, a value
of \type{32} would result in a \type{/CreationDate} and \type{/ModDate}
values of \type{19700101000032Z}.  This is useful for reproducible
builds of documents.  (See also \type{\pdfinfoomitdate},
\type{\pdfsuppressptexinfo}, et al.)

The second, related, environment variable is \type{FORCE_SOURCE_DATE}.
If this is set to~\type{1}, \TEX's time-related primitives are also
initialized from the value of \type{SOURCE_DATE_EPOCH}.  These
primitives are \type{\year}, \type{\month}, \type{\day}, and
\type{\time}.  If \type{SOURCE_DATE_EPOCH} is not set, setting
\type{FORCE_SOURCE_DATE} has no effect.  If \type{FORCE_SOURCE_DATE} is
unset, set to the empty string, or set to~\type{0}, the primitives
reflect the current time as usual.  Any other value elicits a warning,
and the current time is used.  (This is useful only if one wants to make
reproducible \PDF{}s for a set of documents without changing them in any
way, e.g., an operating system distribution with manuals that use
\type{\today}.  Except in such unusual circumstances, it is better not
to set this, and let the \TEX\ primitives retain the meaning they have
always had.)

Finally, just to have the list of options and basic invocation at hand,
here is a verbatim listing of the \type{-}\type{-help} output.  All
options can be specified with one or two dashes and unambiguously
abbreviated.

\startnotmode[screen]
  \switchtobodyfont[9pt] % squeeze everything on one page
\stopnotmode

\typefile{pdftex-help.txt}

\startnotmode[screen]
  \switchtobodyfont[10pt] % squeeze everything on one page
\stopnotmode

%***********************************************************************

\section{Macro packages supporting \PDFTEX}

As \PDFTEX\ generates the final \PDF\ output without help of
a postprocessor, macro packages that take care of these \PDF\ features
have to be set up properly.  Typical tasks are handling color,
graphics, hyperlink support, threading, font||inclusion, as well as
page imposition and manipulation.  All these \PDF||specific tasks can be
controlled by \PDFTEX's own primitives (a few also by a \PDFTEX||specific
\type{\special{pdf: ...}} primitive). Any other \type{\special{}} commands,
like the ones defined for various \DVI\ postprocessors, are simply
ignored by \PDFTEX\ when in \PDF\ output mode; a warning is given only
for non||empty \type{\special{}} commands.

When a macro package already written for classical \TEX\ with \DVI\ output
is to be modified for use with \PDFTEX, it is very helpful to get some
insight to what extent \PDFTEX||specific support is needed. This info can
be gathered \eg\ by outputting the various \type{\special} commands
as \type{\message}. Simply type

\starttyping
\pdfoutput=1 \let\special\message
\stoptyping

or, if this leads to confusion,

\starttyping
\pdfoutput=1 \def\special#1{\write16{special: #1}}
\stoptyping

and see what happens. As soon as one \quote {special} message turns up, one
knows for sure that some kind of \PDFTEX\ specific support is needed, and
often the message itself gives a indication of what is needed.

Currently all mainstream macro packages offer \PDFTEX\ support, with
automatic detection of \PDFTEX\ as engine. So there is normally no need
to turn on \PDFTEX\ support explicitly.

\startitemize

\item  For \LATEX\ users, Sebastian Rahtz and Heiko Oberdiek's \type
       {hyperref} package has substantial support for \PDFTEX\ and
       provides access to most of its features.  In the simplest and
       most common case, the user merely needs to load \type{hyperref},
       and all cross||references will be converted to \PDF\ hypertext
       links. \PDF\ output is automatically selected, compression is
       turned on, and the page size is set up correctly. Bookmarks are
       created to match the table of contents.

\item  The standard \LATEX\ \type{graphics}, \type{graphicx}, and
       \type{color} packages also have automatic \pdfTeX\ support, which
       allow use of color, text rotation, and graphics inclusion commands.

\item  The \CONTEXT\ macro package by Hans Hagen has very full support
       for \PDFTEX\ in its generalized hypertext features. Support for
       \PDFTEX\ is implemented as a special driver, and is invoked by
       typing \type{\setupoutput[pdftex]} or feeding \TEXEXEC\ with the
       \hbox{\tt -{}-pdf} option.

\item  \PDF\ from \TEXINFO\ documents can be created by running \PDFTEX\ on
       the \TEXINFO\ file, instead of \TEX.  Alternatively, run the shell
       command \type{texi2pdf} instead of \type{texi2dvi}.

\item  A small modification of \filename {webmac.tex}, called \filename
       {pdfwebmac.tex}, allows production of hyperlinked \PDF\ versions
       of the program code written in \WEB.

\stopitemize

Some nice samples of \PDFTEX\ output can be found at
\from[ptex_org], \from[context], and \from[tex_showcase].

%***********************************************************************

\section{Setting up fonts}

\PDFTEX\ can work with Type~1 and TrueType fonts (and to some extent
also with OpenType fonts). Font files should be available and embedded
for all fonts used in the generated PDF. It is possible to use
\METAFONT||generated fonts in \PDFTEX\ --- but it is strongly
recommended not to use these fonts if an equivalent is available in
Type~1 or TrueType format, if only because bitmap Type~3 fonts render
poorly under enlargement.

%***********************************************************************

\subsection[mapfile]{Map files}

Font map files provide the connection between \TEX\ \TFM\ font files
and outline font file names. They contain also information about
re||encoding arrays, partial font embedding (``subsetting''), and
character transformation parameters (like SlantFont and ExtendFont). Those
map files were first created for \DVI\ postprocessors. But, as \PDFTEX\
in \PDF\ output mode includes all \PDF\ processing steps, it also needs
to know about font mapping, and therefore reads in one or more map files.
Map files are not read in when \PDFTEX\ is in \DVI\ mode. Bitmap fonts
can (and normally should) be used without being listed in the map file.

By default, \PDFTEX\ reads the map file \filename{pdftex.map}.  In \WEBC,
map files are searched for using the \type{TEXFONTMAPS} config file value
and environment variable.  By default, the current directory and various
system directories are searched.

Within the map file, each font is listed on a single line.  The syntax of
each line is upward||compatible with \type{dvips} map files and can contain
the following fields (some are optional; explanations follow):

\startnarrower
{\em tfmname psname fontflags special encodingfile fontfile}
\stopnarrower

It is mandatory that {\em tfmname} is the first field. If a {\em
psname} is given, it must be the second field. Similarly if {\em
fontflags} is given it must be the third field (if {\em psname} is
present) or the second field (if {\em psname} is left out). The
positions of {\em special}, {\em encodingfile}, and {\em fontfile} can
be mixed.

\startdescription {tfmname}
sets the name of the \TFM\ file for a font --- the file name given in a
\TEX\ \type{\font} command.  This name must always be given.

\stopdescription

\startdescription {psname}
sets the (\POSTSCRIPT) base font name, which has two uses:

First, when a \PDF\ file is embedded by \type{\pdfximage}, the
\type{/BaseFont} names in the font dictionaries of Type~1 and Type~1C
(CFF) fonts from the embedded \PDF\ file are checked against this {\em
psname} field. If names match, the glyphs of that font will not be
copied from the embedded \PDF\ file, but instead a local font is opened,
and all needed glyphs will be taken from the Type~1 font file that is
mentioned in the map line (see {\em fontfile} below). By this collecting
mechanism Type~1 glyphs can be shared between several embedded \PDF\
files and with text that is typeset by \PDFTEX, which helps keeping the
resulting \PDF\ file size small, if many files with similar Type~1(C)
fonts are embedded. Replacing Type~1 fonts from embedded \PDF\ files
requires that also a Type~1 font file name is in the {\em fontfile} field
(see below).

Second, if a font file is not to be embedded into the \PDF\ output
({\em fontfile} field missing), then the {\em psname} field will be
copied to the \type{/BaseFont} and \type{/FontName} dictionary entries
in the \PDF\ file, so that the \POSTSCRIPT\ font name will be known to
reading applications (\eg\ viewers).

It is highly recommended to use the {\em psname} field, but
strictly speaking it is optional.

\stopdescription

\startdescription {fontflags}
optionally specify some characteristics of the font. The following description of
these flags is taken, with slight modification, from the \PDFReference\
(the section on font descriptor flags). Viewers can adapt their rendering
to these flags, especially when they substitute a non-embedded font by
some own approximation.

\startnarrower

The value of the flags key in a font descriptor is a 32||bit integer that
contains a collection of boolean attributes. These attributes are true if the
corresponding bit is set to~1. \in{Table}[flags] specifies the meanings of
the bits, with bit~1 being the least significant. Reserved bits must be set
to zero.

\startbuffer
\starttabulate[|c|l|]
\HL
\NC \bf bit position \NC \bf semantics                               \NC\NR
\HL
\NC 1                \NC Fixed||width font                           \NC\NR
\NC 2                \NC Serif font                                  \NC\NR
\NC 3                \NC Symbolic font                               \NC\NR
\NC 4                \NC Script font                                 \NC\NR
\NC 5                \NC Reserved                                    \NC\NR
\NC 6                \NC Uses the Adobe Standard Roman Character Set \NC\NR
\NC 7                \NC Italic                                      \NC\NR
\NC 8--16            \NC Reserved                                    \NC\NR
\NC 17               \NC All||cap font                               \NC\NR
\NC 18               \NC Small||cap font                             \NC\NR
\NC 19               \NC Force bold at small text sizes              \NC\NR
\NC 20--32           \NC Reserved                                    \NC\NR
\HL
\stoptabulate
\stopbuffer

\placetable
  [here][flags]
  {The meaning of flags in the font descriptor.}
  {\getbuffer}

All characters in a {\em fixed||width} font have the same width, while
characters in a proportional font have different widths. Characters in a {\em
serif font} have short strokes drawn at an angle on the top and bottom of
character stems, while sans serif fonts do not have such strokes. A {\em
symbolic font} contains symbols rather than letters and numbers. Characters
in a {\em script font} resemble cursive handwriting. An {\em all||cap} font,
which is typically used for display purposes such as titles or headlines,
contains no lowercase letters. It differs from a {\em small||cap} font in
that characters in the latter, while also capital letters, have been sized
and their proportions adjusted so that they have the same size and stroke
weight as lowercase characters in the same typeface family.

Bit~6 in the flags field indicates that the font's character set
conforms to the
Adobe Standard Roman Character Set, or a subset of that, and that it uses
the standard names for those characters.

Finally, bit~19 is used to determine whether or not bold characters are
drawn with extra pixels even at very small text sizes. Typically, when
characters are drawn at small sizes on very low resolution devices such
as display screens, features of bold characters may appear only one pixel
wide. Because this is the minimum feature width on a pixel||based device,
ordinary non||bold characters also appear with one||pixel wide features,
and thus cannot be distinguished from bold characters. If bit~19 is set,
features of bold characters may be thickened at small text sizes.

\stopnarrower

If the {\em fontflags} field is not given, and the font is embedded,
\PDFTEX\ treats it as the value~4 (decimal, that is, bit position 3 is set),
a symbolic font. For non-embedded fonts, the default value is
\type{0x22}, a non-symbolic serif font. If you do not know the correct
value, it is best not to specify it at all, as specifying a bad value of
font flags may cause trouble in viewers. On the other hand this option
is not absolutely useless because it provides backward compatibility
with older map files (see the {\em fontfile} description below).

\stopdescription

\startdescription {special}
specifies font manipulations in the same way as
\type{dvips}. Currently only the keywords \type{SlantFont}
and \type{ExtendFont} are interpreted, other instructions (notably
\type{ReEncodeFont} and its parameters, see {\em encoding} below) are
just ignored.  The permitted \type{SlantFont} range is $-$1..1;
for \type{ExtendFont} it's $-$2..2.  The block of {\em special}
instruction must be enclosed by double quote characters: \type{"}.

\stopdescription

\startdescription {encodingfile} specifies the name of the file
containing the external encoding vector to be used for the font. The
encoding file name must have the extension \type{.enc}, and the file
name including extension must be given with either a preceding~\type{<}
character or a preceding~\type{<[}. The format of the encoding vector is
identical to that used by \type{dvips}. If no encoding is specified, the
font's built||in default encoding is used. The {\em encodingfile} field
may be omitted if you are sure that the font resource has the correct
built||in encoding. In general this option is highly recommended, and it
is {\em required} when subsetting a TrueType font.

Starting with version 1.40.19, an encoding file can also be specified
for bitmap \PK\ fonts. In this case, it assigns the glyph names from the
given encoding vector, which can be used with the
\type{\pdfglyphtounicode} primitive (q.v.).  For example:

\starttyping
\pdfglyphtounicode{ffi}{0066 0066 0069} % normally: \input glyphtounicode
\pdfgentounicode=1
\pdfmapline{cmb10 <7t.enc}
\font\cmb=cmb10 \cmb ffi
\stoptyping

The result is a \PDF\ file with a correctly-labeled \type{/ffi}
character instead of the numeric character position in the
\type{cmb10.tfm} (decimal 14).

\stopdescription

\startdescription {fontfile}
sets the name of the font file to be embedded into the \PDF\ output for a
given \TeX\ font (the {\em tfmname}~$\leftrightarrow$~{\em fontfile}
mapping is the most prominent use of the \filename{pdftex.map} file).
The font file name must belong to a Type~1 or TrueType font file. If
the {\em fontfile} field is missing, no font embedding can take place;
in case the {\em psname} field does not contain one of the 14 standard
font names also a warning will be given. Not embedding a font into a \PDF\
file might be troublesome, as it requires that the font or some similar
looking replacement font is available within the \PDF\ viewer, so that
it can render the glyphs with its own font version.

The font file name should be preceded by one or two special characters,
specifying how to handle the font file:

\startitemize

\item  If the font file name is preceded by a \type{<} character, the
       font file will be only partially embedded in the \PDF\ output
       (``subsetted''), meaning that only used glyphs are written to
       the \PDF\ file. This is the most common use and is {\em strongly
       recommended} for any font, as it ensures the portability and
       reduces the size of the \PDF\ output. Subsetted fonts are included
       in such a way that name and cache clashes are minimized.

\item  If the font file name is preceded by a double \type{<<}, the font
       file will be included entirely --- all glyphs of the font are
       embedded, including even those not used in the document. Apart
       from causing large size \PDF\ output, this option may cause
       troubles with TrueType fonts, so it is normally not recommended
       for Type~1 or TrueType fonts. But this is currently the only mode
       that allows the use of OpenType fonts. This mode might also be
       useful in case the font is atypical and cannot be subsetted well
       by \PDFTEX. {\em Beware: proprietary font vendors typically
       forbid full font inclusion.}

\item  If no special character precedes the font file name, it is
       ignored, with a warning (this case was deprecated in \PDFTEX\
       version 1.40.0). You achieve exactly the same \PDF\ result if you
       just remove the font file name from the map entry. Then the glyph
       widths that go into the \PDF~file are extracted from the
       \TFM~file, and a font descriptor object is created that contains
       approximations of the font metrics for the selected font.

\item  Specifying the {\em psname} and no font file name is only useful
       as a last-ditch fallback when you do not want to embed the font
       (\eg\ due to font license restrictions), but wish to use the font
       metrics and let the \PDF\ viewer generate instances that look
       close to the used font in case the font resource is not installed
       on the system where the \PDF\ output will be viewed or
       printed. To use this feature, the font flags {\em must} be
       specified, and it must have the bit~6 set on, which means that
       only fonts with the Adobe Standard Roman Character Set can be
       simulated. The only exception is the case of a symbolic font,
       which is not very useful.

\stopitemize

If you encounter problematic lookups, for instance if \PDFTEX\ tries
to open a \type{.pfa} file instead of a \type{.pfb}, you can add
the suffix to the filename.  In this respect, \PDFTEX\ completely relies
on the \type{kpathsea} library.

\stopdescription

For Type~1 and TrueType fonts, the font file will be included only once
in the \PDF\ output, regardless of how many \TeX\ \type{\font} instances
are used in the document. For instance, given

\starttyping
\font\a = cmr12
\font\b = cmr12 at 11pt
\stoptyping

the outline file \type{cmr12.pfb} will only be included once in the
\PDF, and merely scaled down to create the instance for \type{\b}.

If a used font is not present in the map files, \PDFTEX\ will try to use
\PK~fonts as most \DVI\ drivers do, creating \PK~fonts on||the||fly if
needed. This is the normal, and recommended, way to use bitmap fonts.

To summarize this rather confusing story, we include some example map
lines.  The most common way is to embed only a subset of glyphs from a
font for a particular desired encoding, like this:

\starttyping
ptmri8r Times-Italic <8r.enc <ptmri8a.pfb
\stoptyping

Without re||encoding it looks like this:

\starttyping
cmr10 CMR10 <cmr10.pfb
\stoptyping

\type{SlantFont} and \type{ExtendFont} fields are specified as with
\type{dvips}. \type{SlantFont} and \type{ExtendFont} work only with
embedded Type~1 fonts:

\starttyping
psyro   StandardSymL ".167 SlantFont"         <usyr.pfb
pcrr8rn Courier      ".85 ExtendFont" <8r.enc <pcrr8a.pfb
\stoptyping

Entirely embed a font into the \PDF\ file without and with re||encoding
(not typically useful):

\starttyping
fmvr8x MarVoSym         <<marvosym.pfb
pgsr8r GillSans <8r.enc <<pgsr8a.pfb
\stoptyping

A TrueType font can be used in the same way as a Type~1 font:

\starttyping
verdana8r Verdana <8r.enc <verdana.ttf
\stoptyping

Finally, a few cases with non-embedded fonts. If the fontfile is
missing, the viewer application will have to use its own approximation
of the missing font (with and without re||encoding):

\starttyping
ptmr8r Times-Roman <8r.enc
psyr Symbol
\stoptyping

In the next example the numerical font flags give some rough hint what
general characteristics the GillSans font has, so \eg\ the Adobe Reader
might try an approximation, if it doesn't have the font resource nor
any clue how a font named GillSans should look like:

\starttyping
pgsr8r GillSans 32 <8r.enc
\stoptyping

Not embedding fonts is rather risky and should generally be avoided.
The recommendation these days is to embed all fonts, even the 14 standard ones.

%***********************************************************************

\subsection{Helper tools for TrueType fonts}

As mentioned above, \PDFTEX\ can work with TrueType fonts. Defining
TrueType fonts is similar to Type~1. The only extra thing to do
with TrueType is to create a \TFM\ file. There is a program called
\type{ttf2afm} in the \PDFTEX\ distribution which can be used to
extract \AFM\ from TrueType fonts (another conversion program is
\type{ttf2pt1}). Usage of \type{ttf2afm} is simple:

\starttyping
ttf2afm -e <encoding vector> -o <afm outputfile> <ttf input file>
\stoptyping

A TrueType file can be recognized by its suffix \filename {ttf}. The optional
{\em encoding} specifies the encoding, which is the same as the encoding
vector used in map files for \PDFTEX\ and \type{dvips}. If the encoding is
not given, all the glyphs of the \AFM\ output will be mapped to \type
{/.notdef}. \type{ttf2afm} writes the output \AFM\ to standard output. If we
need to know which glyphs are available in the font, we can run \type
{ttf2afm} without encoding to get all glyph names. The resulting \AFM\ file
can be used to generate a \TFM\ one by applying \filename {afm2tfm}.

To use a new TrueType font the minimal steps may look like below. We suppose
that \filename {test.map} is used.

\starttyping
ttf2afm -e 8r.enc -o times.afm times.ttf
afm2tfm times.afm -T 8r.enc
echo "times TimesNewRomanPSMT <8r.enc <times.ttf" >>test.map
\stoptyping

There are a few limitations with TrueType fonts in comparison with
Type~1 fonts:

\startitemize[a,packed]
\item  The special effects SlantFont|/|ExtendFont did not work
       before version 1.40.0.
\item  To subset a TrueType font, the font must be specified as re||encoded,
       therefore an encoding vector must be given.
\item  TrueType fonts coming with embedded \PDF\ files are kept
       untouched; they are not replaced by local ones.
\stopitemize

For much more about \PDFTEX\ and TrueType fonts, including many details
on handling glyph names, see ``A closer look at TrueType fonts and
\PDFTEX'', {\em TUGboat} 30:1 (2009), pp.~32||34, \from[thanh_truetype_tub]

%***********************************************************************

\section{Formal syntax specification}

This section formally specifies the \PDFTEX-specific extensions to the
\TEX\ macro programming language. Most primitives are prefixed by
\type{pdf}.  The general definitions and syntax rules follow after the
list of primitives.

Two new units of measure were introduced in \PDFTEX\ 1.30.0: the
new Didot (1\,nd~=~0.375\,mm) and the new Cicero (1\,nc~=~12\,nd)
(the former was proposed by \ISO\ in 1975).

% Generated list of primitives (see Makefile).
\input pdftex-syntax.tex

\subsubject{General definitions and syntax rules}

\startpacked

%%S NL
%%S Syntax rules:

\Syntax{
\Something{general text} \Means %
  \Lbrace \Something{balanced text} \Rbrace
}

\Syntax{
\Something{attr spec} \Means %
  \Literal{attr} \Something{general text}
}

\Syntax{
\Something{resources spec} \Means %
  \Literal{resources} \Something{general text}
}

\Syntax{
\Something{rule spec} \Means %
  (\Literal{width} \Or \Literal{height} \Or \Literal{depth}) %
  \Something{dimen} \Optional{\Something{rule spec}}
}

%\Syntax{\Something{object type spec} \Means
%  \Optional{\Literal{stream}
%  \Optional{\Something{attr spec}}} \Something{object contents}}

\Syntax{
\Something{object type spec} \Means %
  \Literal{reserveobjnum}
  \Or \Next %
  \Optional{\Literal{useobjnum} \Something{number}} \Next %
  \Optional{\Literal{stream} \Optional{\Something{attr spec}}} %
  \Something{object contents}
}

\Syntax{
\Something{annot type spec} \Means %
  \Literal{reserveobjnum}
  \Or \Next %
  \Optional{\Literal{useobjnum} \Something{number}} %
  \Optional{\Something{rule spec}} %
  \Something{general text}
}

\Syntax{
\Something{object contents} \Means %
  \Something{file spec}
  \Or \Something{general text}
}

\Syntax{
\Something{xform attr spec} \Means %
  \Optional{\Something{attr spec}} \Optional{\Something{resources spec}}
}

\Syntax{
\Something{image attr spec} \Means %
  \Optional{\Something{rule spec}} %
  \Optional{\Something{attr spec}} %
  \Optional{\Something{page spec}} %
  \Optional{\Something{colorspace spec}} %
  \Optional{\Something{pdf box spec}}
}

\Syntax{
\Something{outline spec} \Means %
  \Optional{\Something{attr spec}} %
  \Something{action spec} %
  \Optional{\Literal{count} \Something{number}} %
  \Something{general text}
}

\Syntax{
\Something{action spec} \Means %
  \Literal{user} \Something{user-action spec}
  \Or \Literal{goto} \Something{goto-action spec}
  \Or \Next \Literal{thread} \Something{thread-action spec}
}

\Syntax{
\Something{user-action spec} \Means %
  \Something{general text}
}

%HE Check:
\Syntax{
\Something{goto-action spec} \Means %
  \Something{numid}
  \Or \Next \Optional{\Something{file spec}} \Something{nameid}
  \Or \Next \Optional{\Something{file spec}} \Optional{\Something{page spec}} \Something{general text}
  \Or \Next \Something{file spec} \Something{nameid} \Something{newwindow spec}
  \Or \Next \Something{file spec} \Optional{\Something{page spec}} \Something{general text} \Something{newwindow spec}
}

\Syntax{
\Something{thread-action spec} \Means %
  \Optional{\Something{file spec}} \Something{numid}
  \Or \Optional{\Something{file spec}} \Something{nameid}
}

\Syntax{
\Something{colorspace spec} \Means %
  \Literal{colorspace} \Something{number}
}

\Syntax{
\Something{pdf box spec} \Means %
  \Literal{mediabox} %
  \Or \Literal{cropbox} %
  \Or \Literal{bleedbox} %
  \Or \Literal{trimbox} %
  \Or \Literal{artbox}
}

\Syntax{
\Something{map spec} \Means %
  \Lbrace \Optional{\Something{map modifier}} %
  \Something{balanced text} \Rbrace
}

\Syntax{
\Something{map modifier} \Means %
  \Literal{+} \Or \Literal{=} \Or \Literal{-}
}

\Syntax{
\Something{numid} \Means %
  \Literal{num} \Something{number}
}

\Syntax{
\Something{nameid} \Means %
  \Literal{name} \Something{general text}
}

\Syntax{
\Something{newwindow spec} \Means %
  \Literal{newwindow} \Or \Literal{nonewwindow}
}

\Syntax{
\Something{dest spec} \Means %
  \Something{numid} \Something{dest type}
  \Or \Something{nameid} \Something{dest type}
}

\Syntax{
\Something{dest type} \Means %
  \Literal{xyz} \Optional{\Literal{zoom} \Something{number}}
  \Or \Literal{fitr} \Something{rule spec}
  \Or \Next \Literal{fitbh}
  \Or \Literal{fitbv}
  \Or \Literal{fitb}
  \Or \Literal{fith}
  \Or \Literal{fitv}
  \Or \Literal{fit}
}

\Syntax{
\Something{thread spec} \Means %
  \Optional{\Something{rule spec}} %
  \Optional{\Something{attr spec}} %
  \Something{id spec}
}

\Syntax{
\Something{id spec} \Means %
  \Something{numid} \Or \Something{nameid}
}

\Syntax{
\Something{file spec} \Means %
  \Literal{file} \Something{general text}
}

\Syntax{
\Something{page spec} \Means %
  \Literal{page} \Something{number}
}

\Syntax{
\Something{expand spec} \Means %
  \Something{stretch} %
  \Something{shrink} %
  \Something{step} %
  \Optional{\Literal{autoexpand}}
}

\Syntax{
\Something{stretch} \Means %
  \Something{number}
}

\Syntax{
\Something{shrink} \Means %
  \Something{number}
}

\Syntax{
\Something{step} \Means %
  \Something{number}
}

\Syntax{
\Something{pre spec} \Means %
  \Literal{pre}
}

\Syntax{
\Something{pdfliteral spec} \Means %
  \Literal{direct} \Or \Literal{page}
}

\Syntax{
\Something{pdfspecial spec} \Means %
  \Lbrace \Optional{\Something{pdfspecial id} %
  \Optional{\Something{pdfspecial modifier}}} %
  \Something{balanced text} \Rbrace
}

\Syntax{
\Something{pdfspecial id} \Means %
  \Literal{pdf:} \Or \Literal{PDF:}
}

\Syntax{
\Something{pdfspecial modifier} \Means %
  \Literal{direct:}
}

\Syntax{
\Something{stack action} \Means %
  \Literal{set} \Or \Literal{push} \Or \Literal{pop} \Or \Literal{current}
}

\stoppacked

A \Something{general text} is expanded immediately, like \type{\special}
in traditional \TEX, unless explicitly mentioned otherwise.

Some of the object and image related primitives can be prefixed by
\type{\immediate}. More about that in the next sections.

%***********************************************************************

\section[primitives]{\PDFTEX\ primitives}

Here follows a description of the primitives added by \PDFTEX\ to the
original \TEX\ engine (other extensions by \ETEX, \MLTEX\ and \ENCTEX\
are not listed).  Many of these primitives are described further in the
\filename {samplepdf.tex} file in the \PDFTEX\ distribution.

If the output is \DVI, then the \PDFTEX-specific dimension parameters
are not used at all. However, some \PDFTEX\ integer parameters can
affect \DVI\ as well as \PDF\ output (specifically, \type{\pdfoutput} and
\type{\pdfadjustspacing}).

%A warning to macro writers: The \PDFTEX-team reserves the namespaces
%\type{\pdf*} and \type{\ptex*} for use by \PDFTEX; if you define macros
%whose names start with \type{\pdf} or \type{\ptex}, you risk nameclashes
%with new primitives introduced in future versions of \PDFTEX.

General warning: many of these new primitives, for example
\type{\pdfdest} and \type{\pdfoutline}, write their arguments directly
to the \PDF\ output file (when producing \PDF), as \PDF\ string
constants.  This means that {\em you} (or, more likely, the macros you
write) must escape characters as necessary (namely \type{\}, \type{(},
and \type{)}.  Otherwise, an invalid \PDF\ file may result.  The
\type{hyperref} and \TEXINFO\ packages have code which may serve as
a starting point for implementing
this, although it will certainly need to be adapted to any particular
situation.

%***********************************************************************

\subsection{Document setup}

\pdftexprimitive{\Syntax{\Tex{\pdfoutput} \Whatever{integer}}}
\bookmark{\tex{pdfoutput}}

This parameter specifies whether the output format should be \DVI\ or
\PDF. A positive value means \PDF\ output, otherwise (default 0) one gets
\DVI\ output.  This primitive is the only one that must be set to produce
\PDF\ output (unless the commandline option \type{-output-format=pdf}
is used); all other primitives are optional.  This parameter cannot be
specified {\em after} shipping out the first page.  In other words, if
we want \PDF\ output, we have to set \type{\pdfoutput} before \PDFTEX\
ships out the first page.

When \PDFTEX\ starts complaining about specials, one can be rather sure
that a macro package is not aware of the \PDF\ mode. A simple way of
making macros aware of \PDFTEX\ in \PDF\ or \DVI\ mode is:

\starttyping
\ifx\pdfoutput\undefined \csname newcount\endcsname\pdfoutput \fi
\ifcase\pdfoutput DVI CODE \else PDF CODE \fi
\stoptyping

Using the \type{ifpdf.sty} file, which works with both \LATEX\ and plain
\TeX, is a cleaner way of doing this.  Historically, the simple test
\type{\ifx\pdfoutput\undefined} was defined; but nowadays, the \PDFTEX\
engine is used in distributions also for non-\PDF\ formats (\eg\
\LATEX), so \type{\pdfoutput} may be defined even when the output format
is \DVI.

\pdftexprimitive{\Syntax{\Tex{\pdfminorversion} \Whatever{integer}}}
\bookmark{\tex{pdfminorversion}}

This primitive sets the \PDF\ version of the generated file and the
highest \PDF\ version of included \PDF{}s allowed without warning, by
default (see \type{\pdfinclusionerrorlevel}).  The default compiled into
the \PDFTEX\ program is \type{\pdfminorversion=4}, setting the \PDF\
version to~1.4 and allowing included \PDF\ files with versions up
to~1.4.  If specified, this primitive must appear before any data is to
be written to the generated \PDF\ file.

Distributions typically alter the engine's compiled default of~4 when
building formats; for example, as of 2010, \TEXLIVE\ sets
\type{\pdfminorversion=5} when formats are built.  This is so object
compression can be enabled (described below).

This was originally a shortened synonym of the
\type{\pdfoptionpdfminorversion} command, which is now obsolete.
\introduced{1.30.0}


\pdftexprimitive{\Syntax{\Tex{\pdfcompresslevel} \Whatever{integer}}}
\bookmark{\tex{pdfcompresslevel}}

This integer parameter specifies the level of {\em stream} compression
(text, inline graphics, and embedded \PNG\ images (only if they are un-
and re-compressed during the embedding process); all done by the
\type{zlib} library). Zero means no compression, 1~means fastest,
9~means best, $2..8$ means something in between. A value outside this
range will be adjusted to the nearest meaningful value. This parameter
is read each time \PDFTEX\ starts a stream. Setting
\type{\pdfcompresslevel=0} is useful for \PDF\ stream debugging.

\pdftexprimitive{\Syntax{\Tex{\pdfobjcompresslevel} \Whatever{integer}}}
\bookmark{\tex{pdfobjcompresslevel}}

This integer parameter controls the compression of {\em non-stream}
objects. In the \PDF-1.4 specification these objects still had to go into
the \PDF\ file as clear text, uncompressed. The \PDF-1.5 specification
now allows to collect non-stream objects as ``compressed objects'' into
``object stream'' objects (\type{/Type /ObjStm}, see \PDF\ Ref.\ 5th~ed.,
sect.~3.4.6). At the \PDF\ file end instead of the object table then
an \type{/XRef} cross-reference stream is written out. This results in
considerably smaller \PDF\ files, particularly if lots of annotations
and links are used.
\introduced{1.40.0}

The writing of compressed objects is enabled by setting
\type{\pdfobjcompresslevel} to a value between~1 and~3; it's
disabled by value~0 (default). Enabling requires that also
\type{\pdfminorversion}~$>$~4. If \type{\pdfobjcompresslevel}~$>$~0,
but \type{\pdfminorversion}~$<$~5, a warning is given and object stream
writing is disabled. The \type{\pdfobjcompresslevel} value is clipped
to the range $0..3$. Using values outside this range is not recommended
(for future extension).

The \type{\pdfobjcompresslevel} settings have the following effects:
When set to~0, no object streams are generated at all. When set to~1,
all non-stream objects are compressed with the exception of any objects
coming with embedded \PDF\ files (``paranoid'' mode, to avoid yet unknown
problems), and also the \type{/Info} dictionary is not compressed for
clear-text legibility. When set to~2, also all non-stream objects coming
with embedded \PDF\ files are compressed, but the \type{/Info} dictionary
is still not compressed. Finally, when set to~3, all non-stream objects
are compressed, including the \type{/Info} dictionary (this means that
the \type{/Info} can't be read as clear text any more). If object streams
are to be used, currently \type{\pdfobjcompresslevel=2} is recommended,
and is so specified in some distributions, including \TEXLIVE~2010 and later.

\description{Caveat:} \PDF\ files generated with object streams enabled
can't be read with (sufficiently old) \PDF\ viewers that don't
understand \PDF-1.5 files. For widest distribution and unknown audience,
don't activate object stream writing. The \PDF-1.5 standard describes
also a hybrid object compression mode that gives some backward
compatibility, but this is currently not implemented, as \PDF-1.5 was
rather quickly adopted by modern \PDF\ viewers. Also not implemented is
the optional \type{/Extends} key.

\pdftexprimitive{\Syntax{\Tex{\pdfdecimaldigits} \Whatever{integer}}}
\bookmark{\tex{pdfdecimaldigits}}

This integer parameter specifies the numeric accuracy of real
coordinates as written to the \PDF\ file. It gives the maximal number of
decimal digits after the decimal point. Valid values are in range
$0..4$. A higher value means more precise output, but also results in a
larger file size and more time to display or print. In most cases the
optimal value is~2. This parameter does not influence the precision of
numbers used in raw \PDF\ code, like that used in \type{\pdfliteral} and
annotation action specifications; also multiplication items (\eg\
scaling factors) are not affected and are always output with best
precision.  This parameter is read when \PDFTEX\ writes a real number to
the \PDF\ output.

When including huge \METAPOST\ images using \filename {supp-pdf.tex}, one can
limit the accuracy to two digits by typing: \type{\twodigitMPoutput}.

\pdftexprimitive{\Syntax{\Tex{\pdfhorigin} \Whatever{dimen}}}
\bookmark{\tex{pdfhorigin}}

This parameter can be used to set the horizontal offset the output box
from the top left corner of the page. A value of 1~inch corresponds to the
normal \TEX\ offset. This parameter is read when \PDFTEX\ starts shipping
out a page to the \PDF\ output.

For standard purposes, this parameter should always be kept at
1~true inch.  If you want to shift text on the page, use \TEX's
own \type{\hoffset} primitive.  To avoid surprises, after global
magnification has been changed by the \type{\mag} primitive, the
\type{\pdfhorigin} parameter should still be 1~true inch, \eg\
by typing \type{\pdfhorigin=1 true in} after issuing the \type{\mag}
command.  Or, you can preadjust the \type{\pdfhorigin} value before typing
\type{\mag}, so that its value after the \type{\mag} command ends up at
1~true inch again.

\pdftexprimitive{\Syntax{\Tex{\pdfvorigin} \Whatever{dimen}}}
\bookmark{\tex{pdfvorigin}}

This parameter is the vertical companion of \type{\pdfhorigin}, and the
notes above regarding \type{\mag} and true dimensions apply. Also keep
in mind that the \TEX\ coordinate system starts in the top left corner
(downward), while \PDF\ coordinates start at the bottom left corner
(upward).

\pdftexprimitive{\Syntax{\Tex{\pdfpagewidth} \Whatever{dimen}}}
\bookmark{\tex{pdfpagewidth}}

This dimension parameter specifies the page width of the \PDF\ output
(the screen, the paper, etc.). \PDFTEX\ reads this parameter when it
starts shipping out a page.  After magnification has been changed by
the \type{\mag} primitive, check that this parameter reflects the wished
true page width.

If the value is set to zero, the page width is calculated as
$w_{\hbox{\txx box being shipped out}} + 2 \times (\hbox{horigin} +
\hbox{\type{\hoffset}})$.  When part of the page falls off the paper or
screen, you can be rather sure that this parameter is set wrong.

\pdftexprimitive{\Syntax{\Tex{\pdfpageheight} \Whatever{dimen}}}
\bookmark{\tex{pdfpageheight}}

Similar to the previous item, this dimension parameter specifies the
page height of the \PDF\ output. If set to zero, the page height will
be calculated analogously to the above. After magnification has been
changed by the \type{\mag} primitive, check that this parameter reflects
the wished true page height.

%***********************************************************************

\subsection[sec.docinfocatalog]{The document info and catalog}

\pdftexprimitive{\Syntax{\Tex{\pdfinfo} \Something{general text}}}
\bookmark{\tex{pdfinfo}}

This primitive allows the user to specify information for the document
info dictionary; if this information is provided, it can be extracted,
\eg\ by the \type{pdfinfo} program.  The \Something{general text} is a
collection of key||value||pairs. The key names are preceded by a
\type{/}, and the values, being strings, are given between
parentheses. All keys, and the primitive itself, are optional. Possible
keys are:\crlf
\type{/Title},\crlf
\type{/Author},\crlf
\type{/Subject},\crlf
\type{/Keywords},\crlf
\type{/Producer} (defaults to \hbox{\tt pdfTeX-\currentpdftex}),\crlf
\type{/Creator} (defaults to \type{TeX}),\crlf
\type{/CreationDate} (defaults to current date and time, with time zone),\crlf
\type{/ModDate} (same default),\crlf
\type{/Trapped} (defaults to \type{/False},\crlf
\type{/PTEX.Fullbanner} (defaults to the \type{\pdftexbanner} string, q.v.).

\type{/CreationDate} and \type{/ModDate} are expressed in the form
\type{D:YYYYMMDDhhmmssTZ}, where \type{YYYY} is the year, \type{MM} is
the month, \type{DD} is the day, hh is the hour, \type{mm} is the
minutes, \type{ss} is the seconds, and \type{TZ} is an optional string
denoting the time zone, \type{Z} for universal time.  For example, this
is the Unix epoch, the beginning of 1970-01-01 UTC, in this format:
\type{19700101000000Z}. If the time zone is not UTC, it is given as
\type{+HH'mm'} or \type{-HH'mm'}, indicating an offset of the given
hours and minutes from UTC, with the given either later (\type{+}) or
earlier (\type{-}) than UTC. (This syntax is specified by the PDF
definition.)

Multiple appearances of \type{\pdfinfo} are concatenated. Usually if a
key is given more than once, the first appearance will be used, but this
is viewer||dependent. Except for standard \TeX\ expansion, \PDFTEX\ does
not perform any further operations in the \Something{general text}
provided by the user.

Here is an example of using \type{\pdfinfo} to include the
information not supplied by \PDFTEX:

\starttyping
\pdfinfo {
    /Title        (example.pdf)
    /Author       (Tom and Jerry)
    /Subject      (Example)
    /Keywords     (mouse, cat)
}
\stoptyping

For more details on all this, see the \PDFReference.

\pdftexprimitive{\Syntax{\Tex{\pdfinfoomitdate} \Whatever{integer}}}
\bookmark{\tex{pdfinfoomitdate}}

If nonzero, omit the \type{/CreationDate} and \type{/ModDate} keys from
the document info dictionary described above.  This can be useful in
making reproducible \PDF{}s.  \introduced{1.40.17}

\pdftexprimitive{\Syntax{\Tex{\pdfsuppressptexinfo} \Whatever{integer}}}
\bookmark{\tex{pdfsuppressptexinfo}}

Treated as a bitmask, specifying which \type{PTEX.*} keys to omit from
the output:

\startbuffer
\starttabulate[|l|l|]
\HL
\NC \bf value        \NC \bf suppresses  \NC\NR
\HL
\NC \tt 1     \NC \type{PTEX.Fullbanner} \NC\NR
\NC \tt 2     \NC \type{PTEX.FileName}   \NC\NR
\NC \tt 4     \NC \type{PTEX.PageNumber} \NC\NR
\NC \tt 8     \NC \type{PTEX.InfoDict}   \NC\NR
\HL
\stoptabulate
\stopbuffer

\placetable
  [here][suppressptexinfo]
  {\type{\pdfsuppressptexinfo} bit meanings.}
  {\getbuffer}

Thus, the value \type{-1}, or the sum of all defined bits, will suppress
everything.

\type{PTEX.Fullbanner} is included by default in the general document
info dictionary, as mentioned under \type{\pdfinfo} above.  The other
\type{PTEX.*} keys are included when a \PDF\ is included in the document
(and not otherwise), as described in \in{section}[sec.addpdfkeys].

This conditional suppression can be useful in making reproducible
\PDF{}s.  \introduced{1.40.17}

\pdftexprimitive{\Syntax{\Tex{\pdfcatalog} \Something{general text}
  \Optional{\Literal{openaction} \Something{action spec}}}}
\bookmark{\tex{pdfcatalog}}

Similar to the document info section is the document catalog, where
possible keys are \type{/URI}, which specifies the base \URL\ of the
document, and \type {/PageMode}, which determines how the \PDF\ viewer
displays the document on startup. The possibilities for the latter are
explained in \in {table} [pagemode]:

\startbuffer
\starttabulate[|l|l|]
\HL
\NC \bf value        \NC \bf meaning                  \NC\NR
\HL
\NC \tt /UseNone     \NC neither outline nor thumbnails visible \NC\NR
\NC \tt /UseOutlines \NC outline visible                        \NC\NR
\NC \tt /UseThumbs   \NC thumbnails visible                     \NC\NR
\NC \tt /FullScreen  \NC full||screen mode                      \NC\NR
\HL
\stoptabulate
\stopbuffer

\placetable
  [here][pagemode]
  {Supported \type{/PageMode} values.}
  {\getbuffer}

The default \type{/PageMode} setting is \type{/UseNone}.  In
full||screen mode, there is no menu bar, window controls, nor any other
window present.

After the \Something{general text}, a construct \Literal{openaction}
\Something{action spec} can be given, where \Literal{openaction} is a
\PDFTEX\ keyword, and \Something{action spec} specifies the action to be
taken when opening the document.  This \Something{action spec} is the
same as for internal links; see \in {section} [linking].  (Instead of
using this method, one can also write the open action directly into the
catalog.)

Several settings can be made in one \type{\pdfcatalog} call, for
example:

\starttyping
\pdfcatalog{
  /PageMode /FullScreen
} openaction goto page 2 {/Fit}
\stoptyping

\pdftexprimitive{\Syntax{\Tex{\pdfnames} \Something{general text}}}
\bookmark{\tex{pdfnames}}

This primitive inserts the \Something{general text} to the \type
{/Names} array. The text must
conform to the specifications as laid down in the \PDFReference,
otherwise the document can be invalid.

\pdftexprimitive{\Syntax{\Tex{\pdftrailer} \Something{general text}}}
\bookmark{\tex{pdftrailer}}

This command puts its argument text verbatim into the file trailer
dictionary. Example: \type{\pdftrailer {/mytrlrkey /mytrlrval}}.
\introduced{1.11a}

\pdftexprimitive{\Syntax{\Tex{\pdftrailerid} \Something{general text}}}
\bookmark{\tex{pdftrailerid}}

Use the \Something{general text} to seed the \type{/ID} value in the
trailer, instead of the default combination of the input file
name and starting time.  If the argument is empty, the \type{/ID} is
omitted entirely.  Example: \type{\pdftrailerid{}}.  This can be useful
in making reproducible \PDF{}s.  \introduced{1.40.17}

%***********************************************************************

\subsection{Fonts}

\pdftexprimitive{\Syntax{\Tex{\pdfpkresolution} \Whatever{integer}}}
\bookmark{\tex{pdfpkresolution}}

This integer parameter specifies the default resolution of embedded \PK\
fonts and is read when \PDFTEX\ embeds a \PK\ font during finishing the
\PDF\ output. As bitmap fonts are still rendered poorly by some \PDF\
viewers, it is best to use Type~1 fonts when available.

\pdftexprimitive{\Syntax{\Tex{\pdffontexpand}
  \Something{font} \Something{stretch} \Something{shrink}
  \Something{step} \Optional{\Literal{autoexpand}}}}
\bookmark{\tex{pdffontexpand}}

This extension to \TEX's font definitions controls a major \PDFTEX\ feature
called {\em font expansion}. We describe this by an example:

\starttyping
\font\somefont=sometfm at 10pt
\pdffontexpand\somefont 30 20 10 autoexpand
\pdfadjustspacing=2
\stoptyping

The \type{30 20 10} means this: \quotation {hey \TEX, when line breaking
is going badly, you may stretch the glyphs from this font as much as
3\,\% or shrink them as much as 2\,\%}. For practical reasons \PDFTEX\
uses discrete expansion steps, in this example, 1\,\%. Roughly spoken, the
trick is as follows. Consider a text typeset in triple column mode. When
\TEX\ cannot break a line in the appropriate way, the unbreakable parts
of the word will stick into the margin. When \PDFTEX\ notes this, it
will try to scale (shrink) the glyphs in that line using fixed steps,
until the line fits. When lines are too spacy, the opposite happens:
\PDFTEX\ starts scaling (stretching) the glyphs until the white space
gaps is acceptable. This glyph stretching and shrinking is called
{\em font expansion}. To enable font expansion, don't forget to set
\type{\pdfadjustspacing} to a value greater than zero.

There are two different modes for font expansion:

First, if the \type{autoexpand} option is given --- which is the
recommended mode --- only a single map entry is needed for all expanded
font versions, using the name of the unexpanded \TFM\ file ({\em
tfmname}). No expanded {\em tfmname} versions need to mentioned (they are
ignored), as \PDFTEX\ generates expanded copies of the unexpanded \TFM\
data structures and keeps them in its memory. Since \PDFTEX~1.40.0 the
\type{autoexpand} mode happens within the page stream by modification of
the text matrix (\PDF\ operator ``\type{Tm}''), and not anymore on font
file level, giving the advantage that it now works not only with Type~1
but also with TrueType and OpenType fonts (and even without embedding
a font file; but that's not recommended). In this mode \PDFTEX\ requires
only unexpanded font files.

Second, if the \type{autoexpand} option is missing, setting up font
expansion gets more tedious, as there must be map entries for \TFM\ files
in all required expansion values. The expanded {\em tfmname} variants
are constructed by adding the font expansion value to the {\em tfmname}
of the base font, \eg\ there must be a map entry with {\em tfmname}
\type{sometfm+10} for 1\,\% stretch or \type{sometfm-15} for 1.5\,\%
shrink. This also means, that for each expanded font variant a \TFM\
file with properly expanded metrics must exist. Having several map
entries for the various expansion values of a font requires to provide
for each expansion value an individually crafted font file with expanded
glyphs. Depending on how these glyphs are generated, this might give
slightly better glyph forms than the rather simple glyph stretching
used in \type{autoexpand mode}. The drawback is that several font
files will be embedded in the \PDF\ output for each expanded font,
leading to significantly larger \PDF\ files than in \type{autoexpand}
mode. For moderate expansion values going without \type{autoexpand}
mode is not worth the trouble.

The font expansion mechanism is inspired by an optimization first
introduced by Prof.~Hermann Zapf, which in itself goes back to
optimizations used in the early days of typesetting: use different
glyphs to optimize the grayness of a page. So, there are many, slightly
different~a's, e's, etc. For practical reasons \PDFTEX\ does not use
such huge glyph collections; it uses horizontal scaling instead. This is
sub||optimal, and for many fonts, possibly offensive to the design. But,
when using \PDF, it's not illogical: \PDF\ viewers use so||called
Multiple Master fonts when no fonts are embedded and|/|or can be found
on the target system. Such fonts are designed to adapt their design to
the different scaling parameters. It is up to the user to determine
to what extent mixing slightly remastered fonts can be used without
violating the design. Think of an O: when geometrically stretched, the
vertical part of the glyph becomes thicker, and looks incompatible with
an unscaled original.  With a Multiple Master situation, one can stretch
while keeping this thickness compatible.

\pdftexprimitive{\Syntax{\Tex{\pdfadjustspacing} \Whatever{integer}}}
\bookmark{\tex{pdfadjustspacing}}

This primitive provides a switch for enabling font expansion. By default,
\type{\pdfadjustspacing} is set to~0; then font expansion is disabled,
so that the \PDFTEX\ output is identical to that from the original \TEX\
engine.

Font expansion can be activated in two modes. When
\type{\pdfadjustspacing} is set to~1, font expansion is applied {\em
after} \TEX's normal paragraph breaking routines have broken the paragraph
into lines. In this case, line breaks are identical to standard \TEX\
behavior.

When set to~2, the width changes that are the result of stretching and
shrinking are taken into account {\em while} the paragraph is broken into
lines. In this case, line breaks are likely to be different from those of
standard \TEX. In fact, paragraphs may even become longer or shorter.

Both alternatives require a collection of \TFM\ files that are
related to the \Something{stretch} and \Something{shrink} settings
for the \type{\pdffontexpand} primitive, unless this is given with the
\type{autoexpand} option.

\pdftexprimitive{\Syntax{\Tex{\efcode} \Something{font}
  \Something{8-bit number} \Whatever{integer}}}
\bookmark{\tex{efcode}}

We haven't yet told the whole story. One can imagine that some glyphs are
visually more sensitive to stretching or shrinking than others. Then the
\type{\efcode} primitive can be used to influence the expandability of
individual glyphs within a given font, as a factor to the expansion
setting from the \type{\pdffontexpand} primitive. The syntax is similar
to \type{\sfcode} (but with the \Something{font} required), and it
defaults to~1000, meaning 100\,\% expandability. The given integer value
is clipped to the range $0..1000$, corresponding to a usable
expandability range of $0..100$\,\%. Example:

\starttyping
\efcode\somefont`A=800
\efcode\somefont`O=0
\stoptyping

Here an A may stretch or shrink only by 80\,\% of the current expansion
value for that font, and expansion for the~O is disabled. The actual
expansion is still bound to the steps as defined by \type{\pdffontexpand}
primitive, otherwise one would end up with more possible font inclusions
than would be comfortable.

Changes to this table are global, i.e., ignore \TeX's usual grouping,
and apply only to the given \Something{font}.

\pdftexprimitive{\Syntax{\Tex{\pdfprotrudechars} \Whatever{integer}}}
\bookmark{\tex{pdfprotrudechars}}

Yet another way of optimizing paragraph breaking is to let certain
characters move into the margin (`character protrusion'). When
\type{\pdfprotrudechars=1}, the glyphs qualified as such will make
this move when applicable, without changing the line-breaking. When
\type{\pdfprotrudechars=2} (or greater), character protrusion will
be taken into account while considering breakpoints, so line-breaking
might be changed. This qualification and the amount of shift are set by
the primitives \type{\rpcode} and \type{\lpcode}. Character protrusion
is disabled when \type{\pdfprotrudechars=0} (or negative).

If you want to protrude some item other than a character (\eg\
a \type{\hbox}), you can do so by padding the item with an invisible
zero||width character, for which protrusion is activated.

\pdftexprimitive{\Syntax{\Tex{\rpcode} \Something{font}
  \Something{8-bit number} \Whatever{integer}}}
\bookmark{\tex{rpcode}}

The amount that a character from a given font may shift into the right
margin (`character protrusion') is set by the primitive \type{\rpcode}.
The protrusion distance is the integer value given to \type{\rpcode},
multiplied with 0.001\,em from the current font. The given integer value
is clipped to the range $-1000..1000$, corresponding to a usable
protrusion range of $-$1\,em..1\,em. Example:

\starttyping
\rpcode\somefont`,=200
\rpcode\somefont`-=150
\stoptyping

Here the comma may shift by 0.2\,em into the margin and the hyphen by
0.15\,em. All these small bits and pieces will help \PDFTEX\ to give
you better paragraphs (use \type{\rpcode} judiciously; don't overdo it).

Remark: old versions of \PDFTEX\ use the character width as measure. This
was changed to a proportion of the em-width after \THANH\ finished his
master's thesis.

Changes to this table are global, i.e., ignore \TeX's usual grouping,
and apply only to the given \Something{font}.

\pdftexprimitive{\Syntax{\Tex{\lpcode} \Something{font}
  \Something{8-bit number} \Whatever{integer}}}
\bookmark{\tex{lpcode}}

This is similar to \type{\rpcode}, but affects the amount by which
characters may protrude into the left margin. Here also the given integer
value is clipped to the range $-1000..1000$.

Changes to this table are global, i.e., ignore \TeX's usual grouping,
and apply only to the given \Something{font}.

\pdftexprimitive{\Syntax{\Tex{\leftmarginkern} \Something{box number}
  \Whatever{expandable}}}
\bookmark{\tex{leftmarginkern}}

The \Tex{\leftmarginkern} \Something{box number} primitive expands to the
width of the margin kern at the left side of the horizontal list stored
in \Tex{\box} \Something{box number}. The expansion string includes the
unit \type{pt}. \Eg, when the left margin kern of \type{\box0} amounts
to $-$10\,pt, \type{\leftmarginkern0} will expand to \type{-10pt}.
A similar primitive \type{\rightmarginkern} exists for the right margin.
\introduced{1.30.0}

These are auxiliary primitives to make character protrusion
more versatile. When using the \TEX\ primitive \type{\unhbox} or
\type{\unhcopy}, the margin kerns at either end of the unpackaged
hbox will be removed (\eg\ to avoid weird effects if several
hboxes are unpackaged behind each other into the same horizontal
list). These \type{\unhbox} or \type{\unhcopy} are often used together
with \type{\vsplit} for dis- and re||assembling of paragraphs, \eg\ to
add line numbers. Paragraphs treated like this do not show character
protrusion by default, as the margin kerns have been removed during the
unpacking process.

The \type{\leftmarginkern} and \type{\rightmarginkern} primitives allow
to access the margin kerns and store them away before unpackaging the
hbox. \Eg\, the following code snipplet restores margin kerning of
a horizontal list stored in \type{\box\testline}, resulting in a hbox with
proper margin kerning (which is then done by ordinary kerns).

\starttyping
\dimen0=\leftmarginkern\testline
\dimen1=\rightmarginkern\testline
\hbox to\hsize{\kern\dimen0\unhcopy\testline\kern\dimen1}
\stoptyping

\pdftexprimitive{\Syntax{\Tex{\rightmarginkern} \Something{box number}
  \Whatever{expandable}}}
\bookmark{\tex{rightmarginkern}}

The \Tex{\rightmarginkern} \Something{box number} primitive expands to
the width of the margin kern at the right side of the horizontal list
stored in \Tex{\box} \Something{box number}. See \type{\leftmarginkern}
for more details.
\introduced{1.30.0}

\pdftexprimitive{\Syntax{\Tex{\letterspacefont} \Something{control sequence}
  \Something{font} \Something{integer}}}
\bookmark{\tex{letterspacefont}}

This primitive creates an instance of \Something{font} with the widths
of all glyphs increased by \Something{integer} thousandths of an em (as
set in \Something{font}).  The effect is letter spacing, but the glyphs
are actually larger (sidebearings are increased), so a single glyph will
take more space. For instance, the following creates a font
\type{\lsfont} whose glyphs are all 1.2\,pt larger than those of
\type{\normalfont}:

\starttyping
\font\normalfont=myfont at 12pt
\letterspacefont\lsfont\normalfont 100
\stoptyping

Negative values are allowed for \Something{integer}.
Letter spacing works natively in \PDF\ mode only, unless special fonts are
devised (in our example, a \type{myfont+100ls} font), as with font expansion.

\pdftexprimitive{\Syntax{\Tex{\pdfcopyfont} \Something{control sequence}
  \Something{font}}}
\bookmark{\tex{pdfcopyfont}}

This primitive defines \Something{control sequence} as a synonym for
\Something{font}.

\pdftexprimitive{\Syntax{\Tex{\pdffontattr} \Something{font}
  \Something{general text}}}
\bookmark{\tex{pdffontattr}}

This primitive inserts the \Something{general text} to the \type{/Font}
dictionary. The text must conform to the specifications as laid down in
the \PDFReference, otherwise the document can be invalid.  For examples,
see the \type{cmap} and \type{CJK} packages.

\pdftexprimitive{\Syntax{\Tex{\pdffontname} \Something{font}
  \Whatever{expandable}}}
\bookmark{\tex{pdffontname}}

In \PDF\ files produced by \PDFTEX\ one can recognize a font resource
by the prefix~\type{/F} followed by a number, for instance \type{/F12}
or \type{/F54}.  For a given \TEX\ \Something{font}, this primitive
expands to the number from the corresponding font resource name.
\Eg, if \type{/F12} corresponds to some \TEX\ font \type{\foo}, the
\type{\pdffontname\foo} expands to the number~\type{12}.

In the current implementation, when \type{\pdfuniqueresname} (see below)
is set to a positive value, the \type{\pdffontname} still returns only the
number from the font resource name, but not the appended random string.

\pdftexprimitive{\Syntax{\Tex{\pdffontobjnum} \Something{font}
  \Whatever{expandable}}}
\bookmark{\tex{pdffontobjnum}}

This command is similar to \type{\pdffontname}, but it returns the
\PDF\ object number of the font dictionary instead of the number from
the font resource name.  \Eg, if the font dictionary (\type{/Type
/Font}) in \PDF\ object~3 corresponds to some \TEX\ font \type{\foo},
the \type{\pdffontobjnum\foo} gives the number~\type{3}.

Use of \type{\pdffontname} and \type{\pdffontobjnum} allows users full
access to all the font resources used in the document.

\pdftexprimitive{\Syntax{\Tex{\pdffontsize} \Something{font}
  \Whatever{expandable}}}
\bookmark{\tex{pdffontsize}}

This primitive expands to the font size of the given font, with unit
\type{pt}.  \Eg, when using the plain \TeX\ macro package, the call
\type{\pdffontsize\tenrm} expands to \type{10.0pt}.

\pdftexprimitive{\Syntax{\Tex{\pdfincludechars} \Something{font}
  \Something{general text} \Whatever{expandable}}}
\bookmark{\tex{pdfincludechars}}

This command causes \PDFTEX\ to treat the characters in \Something{general
text} as if they were used with \Something{font}\unkern, which means that the
corresponding glyphs will be embedded into the font resources in the \PDF\
output. Nothing is appended to the list being built.

\pdftexprimitive{\Syntax{\Tex{\pdfuniqueresname} \Whatever{integer}}}
\bookmark{\tex{pdfuniqueresname}}

When this primitive is assigned a positive number, \PDF\ resource names
will be made reasonably unique by appending a random string consisting
of six \ASCII\ characters.

\pdftexprimitive{\Syntax{\Tex{\pdfmapfile} \Something{map spec}}}
\bookmark{\tex{pdfmapfile}}

This primitive is used for reading a font map file consisting of
one or more font map lines. The name of the map file is given in
the \Something{map spec} together with an optional leading modifier
character. If no \type{\pdfmapfile} primitive is given, the default map
file \filename{pdftex.map} will be read by \PDFTEX. There is a companion
primitive \type{\pdfmapline} that allows to scan single map lines;
its map line argument has the same syntax as the map lines from a map
file. Both primitives can be used concurrently. The \type{\pdfmapfile}
primitive is fast for reading external bulk font map information (many
map lines collected in a map file), whereas the \type{\pdfmapline} allows
to put the font map information for individual \TeX\ fonts right into
the \TeX\ source or a style file. In any case the map line information
is scanned by \PDFTEX, and in the most common case the data are put into
a fresh internal map entry data structure, which is then consulted once
a font is called.

Normally there is no need for the \PDFTEX\ user to bother about the
\type{\pdfmapfile} or \type{\pdfmapline} primitives, as the main \TEX\
distributions provide nice helper tools that automatically assemble
the default font map file. Prominent tool examples are the scripts
\type{updmap} and \type{updmap-sys} coming with \TEXLIVE\ and \TETEX.
If your map file isn't in the current directory (or a standard system
directory), you will need to set the \type{TEXFONTMAPS} variable (in
\WEBC) or give an explicit path so that it will be found.

When the \type{\pdfmapfile} or \type{\pdfmapline} primitive is read
by \PDFTEX, the argument (map file or map line) will be processed
{\em immediately}, and the internal map entry database is updated.
The operation mode of the \type{\pdfmapfile} and \type{\pdfmapline}
primitives is selected by an optional modifier character (\type{+},
\type{=}, \type{-}) in front of the {\em tfmname} field. This modifier
defines how the individual map lines are going to be handled, and how
a collision between an already registered map entry and a newer one is
resolved; either ignoring a later entry, or replacing or deleting an
existing entry. But in any case, map entries of fonts already in use
are kept untouched. Here are two examples:

\starttyping
\pdfmapfile{+myfont.map}
\pdfmapline{+ptmri8r Times-Italic <8r.enc <ptmri8a.pfb}
\stoptyping

When no modifier character is given (\eg\ \type{\pdfmapfile{foo.map}}
or \type{\pdfmapline{phvr8r Helvetica}}) and there hasn't already been
any call of one of these primitives before, then the default map file
\filename{pdftex.map} will {\em not} be read in. Apart from this the
given map file will be processed similarly as with a \type{+} modifier:
duplicate later map entries within the file are ignored and a warning is
issued. This means, that you can block reading of the default map file
also by an empty \type{\pdfmapfile{}} or \type{\pdfmapline{}} early in
the \TeX\ file. When the default map file is large but you don't need it
anyway, these command variants might considerably speed up the \PDFTEX\
startup process.

If a modifier is given, the mechanism is so that before reading the
items given as arguments to \type{\pdfmapfile} or \type{\pdfmapline}
the default map file will be read first --- if this hasn't already
been done or been prevented by the above blocking cases. This should
be mostly compatible with the traditional behavior. If you want to add
support for a new font through an additional font map file while keeping
all the existing mappings, don't use the primitive versions without
modifier, but instead type either \type{\pdfmapfile{+myfont.map}} or
\type{\pdfmapfile{=myfont.map}}, as described below.

\type{\pdfmapfile {+foo.map}} reads the file \filename{foo.map}; duplicate
later map entries within the file are ignored and a warning is issued.

\type{\pdfmapfile {=foo.map}} reads the file \filename{foo.map};
matching map entries in the database are replaced by new entries from
\filename{foo.map} (if they aren't already in use).

\type{\pdfmapfile {-foo.map}} reads the file \filename{foo.map}; matching
map entries are deleted from the database (if not yet in use).

If you want to use a base map file name other than \type{pdftex.map},
or change its processing options through a \PDFTEX\ format, you can do
this by appending the \type{\pdfmapfile} command to the \type{\everyjob{}}
token list for the \type{-ini} run, \eg:

\starttyping
\everyjob\expandafter{\the\everyjob\pdfmapfile{+myspecial.map}}
\dump
\stoptyping

This would always read the file \type{myspecial.map} after the default
\type{pdftex.map} file.


\pdftexprimitive{\Syntax{\Tex{\pdfmapline} \Something{map spec}}}
\bookmark{\tex{pdfmapline}}

Similar to \type{\pdfmapfile}, but here you can give a single
map line (like the ones in map files) as an argument. The optional modifiers
(\type{+-=}) have the same effect as with \type{\pdfmapfile}; see also
the description above.  Example:

\starttyping
\pdfmapline{+ptmri8r Times-Italic <8r.enc <ptmri8a.pfb}
\stoptyping

This primitive (especially the \type{\pdfmapline{=...}} variant) is useful
for temporary quick checks of a new font map entry during development,
before finally putting it into a map file.

\type{\pdfmapline {}} like \type{\pdfmapfile {}} blocks reading
of the default map file, if it comes early enough in the \TeX\
input. \introduced{1.20a}


\pdftexprimitive{\Syntax{\Tex{\pdfsuppresswarningdupmap} \Whatever{integer}}}
\bookmark{\tex{pdfsuppresswarningdupmap}}

Ordinarily, \PDFTEX\ gives a warning when a duplicate map entry for a
given font is read, whatever the mechanism.  However, sometimes it is
useful to include map information within the document, using
\type{\pdfmapfile} or \type{\pdfmapline}, even for fonts that happen to
be installed on the system.  Then seeing the warnings on every run is
just noise, and can be suppressed by setting this parameter to a
positive number.  \introduced{1.40.13}

\pdftexprimitive{\Syntax{\Tex{\pdftracingfonts} \Whatever{integer}}}
\bookmark{\tex{pdftracingfonts}}

This integer parameter specifies the level of verbosity for info about
expanded fonts given in the log, \eg\ when \type{\tracingoutput=1}.
If \type{\pdftracingfonts=0}, which is the default, the log shows the
actual non-zero signed expansion value for each expanded letter within
brackets, \eg:

\starttyping
...\xivtt (+20) t
\stoptyping

If \type{\pdftracingfonts=1}, also the name of the \TFM\ file is listed,
together with the font size, \eg:

\starttyping
...\xivtt (cmtt10+20@14.0pt) t
\stoptyping

Setting \type{\pdftracingfonts} to a value other than~0 or~1 is not
recommended, to allow future extensions.  \introduced{1.30.0}

\pdftexprimitive{\Syntax{\Tex{\pdfmovechars} \Whatever{integer}}}
\bookmark{\tex{pdfmovechars}}

Since \PDFTEX\ version 1.30.0 the primitive \type{\pdfmovechars} is obsolete,
and its use merely leads to a warning. (This primitive specified whether
\PDFTEX\ should try to move characters in range 0..31 to higher slots;
its sole purpose was to remedy certain bugs of early \PDF\ viewers.)

\pdftexprimitive{\Syntax{\Tex{\pdfpkmode} \Whatever{tokens}}}
\bookmark{\tex{pdfpkmode}}

The \type{\pdfpkmode} is a token register that sets the \METAFONT\ mode for
pixel font generation. The contents of this register is dumped into the
format, so one can (optionally) preset it \eg\ in \type{pdftexconfig.tex}.
\introduced{1.30.0}

\pdftexprimitive{\Syntax{\Tex{\pdfnoligatures} \Something{font}}}
\bookmark{\tex{pdfnoligatures}}

This disables all ligatures in the loaded font \Something{font}.
\introduced{1.30.0}

\pdftexprimitive{\Syntax{\Tex{\tagcode} \Something{font}
  \Something{8-bit number} \Whatever{integer}}}
\bookmark{\tex{tagcode}}

This primitive accesses a character's \type{char_tag} info. It is meant
to delete \type{lig_tag} (the character's ligature/kerning program),
\type{list_tag} (which indicates that the character belongs to a chain of
ascending sizes) and/or \type{ext_tag} (which indicates that the character
is extensible), with the following options: assigning \type{-7} or smaller
clears all tags, \type{-6} clears \type{ext_tag} and \type{list_tag},
\type{-5} clears \type{ext_tag} and \type{lig_tag}, \type{-4} clears
\type{ext_tag}, \type{-3} clears \type{list_tag} and \type{lig_tag},
\type{-2} clears \type{list_tag}, \type{-1} clears \type{lig_tag},
and \type{0} or larger does nothing. Changes are irreversible and global.

Conversely, when queried, the primitive returns \type{0} if the tag's
value is \type{no_tag}, \type{1} if \type{lig_tag} is set, \type{2} if
\type{list_tag} is set or \type{4} (not~3) if \type{ext_tag} is set.

\pdftexprimitive{\Syntax{\Tex{\pdfgentounicode} \Whatever{integer}}}

By default, \PDFTEX\ does not include a \type{/ToUnicode} resource when
including fonts in the output. Such a resource (also called a CMap
resource) maps glyph names to Unicode characters in a \PDF\ file.
Lacking such a resource, it is the \PDF\ reader which determines how and
whether searching in the \PDF\ file works.  In practice, searching for
basic ASCII characters generally works, but searching for anything
beyond those, including ligatures such as `fi', fails in some versions
of some PDF browsers (most notably Adobe Reader).

If \type{\pdfgentounicode} is set to \type{1} when the job ends, the
\type{/ToUnicode} resource will be included in the output, with mappings
for Type~1 fonts used in the documents. The mapping is created as
follows: for each glyph in the font, look for its \type{ToUnicode} value
in a global hash table.  By default that global hash table is empty;
entries are added to the table using the following command:

\pdftexprimitive{\Syntax{\Tex{\pdfglyphtounicode} \Something{general text}
  \Something{general text}}}

The first argument is the name of a glyph, the second is a string of Unicode
numeric values denoting characters, separated by spaces. For instance:

\starttyping
\pdfgentounicode=1
\pdfglyphtounicode{ff}{0066 0066}
\stoptyping

\noindent maps the \type{ff} ligature to a pair of \type{f}'s (whose
code is \type{U+0066}).

The \type{glyphtounicode.tex} file (distributed with \PDFTEX\ and other
software) contains thousands of such definitions, covering most common
glyph names.  So, for practical purposes, one would probably want:

\starttyping
\input glyphtounicode
\pdfgentounicode=1
\stoptyping

\LATEX\ users could load the \type{cmap} package to achieve the same
effect.


\pdftexprimitive{\Syntax{\Tex{\pdfnobuiltintounicode} \Something{font}}}

The primary purpose of this command is to prevent \PDFTEX\ from
generating the \type{ToUnicode}/CMap resource for the given font when
\type{\pdfgentounicode=1}, most likely because the CMap resource is
already generated by some other method (for example, the \LATEX\
\type{cmap} package uses \type{\pdffontattr} to generate a CMap
resource).

Minimal example: 
\starttyping
\font\f=cmb10
\pdfnobuiltintounicode\f
\f No unicode mappings for this output.
\stoptyping
\introduced{1.40.11}


\pdftexprimitive{\Syntax{\Tex{\pdfinterwordspaceon}}}
\pdftexprimitive{\Syntax{\Tex{\pdfinterwordspaceoff}}}

These commands create corresponding whatsit nodes which turn on/off
generation of faked interword spaces in the output.  This allows for
better reflowing of text on the fly by \PDF\ readers, better extraction
of textual content, and is required by \PDF/A.  It does not affect the
normal \TeX\ justification with glue.

This requires finding and reading font files \type{dummy-space.tfm} and
\type{pfb}; the font is included in the \PDF\ output and character 32 is
inserted from it as the ``fake'' space.

Example of usage (see also the \type{fake-interword-space.tex} test file):

\starttyping
Text with no interword spaces.

\pdfmapline{+dummy-space <dummy-space.pfb}
\pdfglyphtounicode{space}{0020}
\pdfinterwordspaceon

Switch to text with faked interword spaces.

\pdfinterwordspaceoff

Back to text with no interword spaces.
\stoptyping
\introduced{1.40.15}

\pdftexprimitive{\Syntax{\Tex{\pdffakespace}}}

Insert a faked interword space to the output, regardless of the value of
\type{\pdfinterwordspaceon} and \type{\pdfinterwordspaceoff}.  Example:

\starttyping
Text with no interword \pdffakespace\pdffakespace spaces.
\stoptyping
\introduced{1.40.15}

%***********************************************************************

\subsection{Spacing}

Controlling spacing before and after characters was introduced in
version 1.30, mostly to handle punctuation rules in different
languages. The \type{\...code} tables here, like those in the previous
section, operate globally, i.e., ignore \TeX's usual grouping, and apply
only to the given \Something{font}, not other instances of the
underlying font.

\pdftexprimitive{\Syntax{\Tex{\pdfadjustinterwordglue} \Whatever{integer}}}
\bookmark{\tex{pdfadjustinterwordglue}}

If positive, adjustment of interword glue is enabled and controlled by the
following three primitives.

\pdftexprimitive{\Syntax{\Tex{\knbscode} \Something{font}
  \Something{8-bit number} \Whatever{integer}}}
\bookmark{\tex{knbscode}}

The amount of space, in thousandths of an em, added to the natural width
of the glue following a character (the name stands for ``kern before
space'', although technically it is looking at glue items, not kern
items). This amounts is clipped to the range $-1000..1000$.  For
instance, the following example means that glues after periods will be
increased by .2\,em.

\starttyping
\pdfadjustinterwordglue=1
\knsbcode\font`\.=200
\stoptyping

\pdftexprimitive{\Syntax{\Tex{\stbscode} \Something{font}
  \Something{8-bit number} \Whatever{integer}}}
\bookmark{\tex{stbscode}}

This works like \type{\knbscode}, but applies to the stretch component of
the following glue.

\pdftexprimitive{\Syntax{\Tex{\shbscode} \Something{font}
  \Something{8-bit number} \Whatever{integer}}}
\bookmark{\tex{shbscode}}

Like \type{\stbscode}, but for the shrink component.

\pdftexprimitive{\Syntax{\Tex{\pdfprependkern} \Whatever{integer}}}
\bookmark{\tex{pdfprependkern}}

If positive, automatic insertion of kerns before characters is enabled.

\pdftexprimitive{\Syntax{\Tex{\knbccode} \Something{font}
  \Something{8-bit number} \Whatever{integer}}}
\bookmark{\tex{knbccode}}

The width of the kern, in thousandths of an em, inserted before a character.
It is clipped to the range $-1000..1000$. For instance, with the
following code, a .15\,em-kern will be inserted before all question marks
(useful for \eg\ French punctuation):

\starttyping
\pdfprependkern=1
\knbccode\font`\?=150
\stoptyping

\pdftexprimitive{\Syntax{\Tex{\pdfappendkern} \Whatever{integer}}}
\bookmark{\tex{pdfappendkern}}

Same as \type{\pdfprependkern}, but for kerns inserted after characters.

\pdftexprimitive{\Syntax{\Tex{\knaccode} \Something{font}
  \Something{8-bit number} \Whatever{integer}}}
\bookmark{\tex{knaccode}}

Same as \type{\knbccode}, except the kern is inserted after the
character.  Such a kern is required for instance after a left guillemet
in French punctuation.

%***********************************************************************

\subsection{Vertical adjustments}

\pdftexprimitive{\Syntax{\Tex{\pdfignoreddimen} \Whatever{dimen}}}
\bookmark{\tex{pdfignoreddimen}}

This is the dimension which must assigned to the following four primitives
so they are ignored. Default is \type{-1000pt}, and it should be modified
with care since it also influences when a previous paragraph's depth is ignored
(for instance, the plain \TEX\ macro \type{\nointerlineskip} should be
modified accordingly).

\pdftexprimitive{\Syntax{\Tex{\pdffirstlineheight} \Whatever{dimen}}}
\bookmark{\tex{pdffirstlineheight}}

This parameter specifies the height of the first line of a paragraph,
regardless of its content. It is read when the paragraph builder is
called, and ignored when set to \type{\pdfignoreddimen}. By default,
it is set to \type{-1000pt}, so it is ignored as long as the value
of \type{\pdfignoreddimen} is the same.

\pdftexprimitive{\Syntax{\Tex{\pdflastlinedepth} \Whatever{dimen}}}
\bookmark{\tex{pdflastlinedepth}}

This is similar to the previous parameter, but affects the last line's
depth of a paragraph.

\pdftexprimitive{\Syntax{\Tex{\pdfeachlineheight} \Whatever{dimen}}}
\bookmark{\tex{pdfeachlineheight}}

Similar to \type{\pdffirstlineheight}, but for all lines of a paragraph,
including the first one, unless \type{\pdffirstlineheight} is specified.

\pdftexprimitive{\Syntax{\Tex{\pdfeachlinedepth} \Whatever{dimen}}}
\bookmark{\tex{pdfeachlinedepth}}

Like the preceding parameter, but for depth.

%***********************************************************************

\subsection{\PDF\ objects}

\pdftexprimitive{\Syntax{\Tex{\pdfobj} \Something{object type spec}
  \Modelist{h, v, m}}}
\bookmark{\tex{pdfobj}}

This command creates a raw \PDF\ object that is written to the \PDF\
file as \type{1 0 obj} \unknown\ \type{endobj}. The object is written to
\PDF\ output as provided by the user. When \Something{object type spec}
is not given, \PDFTEX\ no longer creates a dictionary object with
contents \Something{general text}, as it did in the past.

When, however, \Something{object type spec} is given as \Something{attr
spec} \Literal{stream}, the object will be created as a stream with
contents \Something{general text} and additional attributes in
\Something{attr spec}\unkern.

When \Something{object type spec} is given as \Something{attr spec}
\Literal{file}, then the \Something{general text} will be treated as a file
name and its contents will be copied into the stream contents.

When \Something{object type spec} is given as \type{reserveobjnum},
just a new object number is reserved. The number of the reserved object
is accessible via \type{\pdflastobj}. The object can later be filled
with contents by \Syntax{\Tex{\pdfobj} \Literal{useobjnum}
\Something{number} \Lbrace\Something{balanced text}\Rbrace}.
But the reserved object number can already be used before by other
objects, which provides a forward||referencing mechanism.

The object is kept in memory and will be written to the \PDF\ output only
when its number is referred to by \type{\pdfrefobj} or when \type{\pdfobj}
is preceded by \type{\immediate}. Nothing is appended to the list being
built. The number of the most recently created object is accessible via
\type{\pdflastobj}.

\pdftexprimitive{\Syntax{\Tex{\pdflastobj} \Whatever{read||only integer}}}
\bookmark{\tex{pdflastobj}}

This command returns the object number of the last object created by \type
{\pdfobj}.

\pdftexprimitive{\Syntax{\Tex{\pdfrefobj} \Something{object number}
  \Modelist{h, v, m}}}
\bookmark{\tex{pdfrefobj}}

This command appends a whatsit node to the list being built. When the whatsit
node is searched at shipout time, \PDFTEX\ will write the object
\Something{object number}
to the \PDF\ output if it has not been written yet.

%***********************************************************************

\subsection{Page and pages objects}

\pdftexprimitive{\Syntax{\Tex{\pdfpagesattr} \Whatever{tokens}}}
\bookmark{\tex{pdfpagesattr}}

\PDFTEX\ expands this token list when it finishes the \PDF\ output and
adds the resulting character stream to the root \type{Pages}
object. When defined, these are applied to all pages in the
document. Some examples of attributes are \type{/CropBox}, the rectangle
specifying the region of the page being displayed and printed, and
\type{/Rotate}, the number of degrees (in multiples of 90) the page
should be rotated clockwise when it is displayed or printed.

% /MediaBox is not a good example, since will never take effect

\starttyping
\pdfpagesattr
  { /Rotate 90                % rotate all pages by 90 degrees
    /CropBox [0 0 612 792] }  % the crop size of all pages (in bp)
\stoptyping

\pdftexprimitive{\Syntax{\Tex{\pdfpageattr} \Whatever{tokens}}}
\bookmark{\tex{pdfpageattr}}

This is similar to \type{\pdfpagesattr}, but has priority over it.
It can be used to override any attribute given by \type{\pdfpagesattr}
for an individual page. The token list is expanded when \PDFTEX\ ships out
a page. The contents are added to the attributes of the current page.

If the \type{\pdfpageattr} value contains the string \type{/MediaBox},
then \PDFTEX\ omits outputting its own \type{/MediaBox} value (which is
\type{[0 0 }\Something{\it page\_width} \Something{\it
page\_height}\type{]}).  (This behavior was introduced in version
1.40.18.)

\pdftexprimitive{\Syntax{\Tex{\pdfpageresources} \Whatever{tokens}}}
\bookmark{\tex{pdfpageresources}}

These tokens are added to the resource dictionary for all pages, before
the font, XOBject, and other resources.

\starttyping
\pdfpageresources{/MyPageResourceAttribute /MyValue}
\stoptyping

\pdftexprimitive{\Syntax{\Tex{\pdfpageref} \Something{page number}
  \Whatever{expandable}}}
\bookmark{\tex{pdfpageref}}

This primitive expands to the number of the page object that contains
the dictionary for page \Something{page number}. If the page
\Something{page number} does not exist, a warning will be issued,
a fresh unused \PDF\ object will be generated, and \type{\pdfpageref}
will expand to that object number.

\Eg, if the dictionary for page~5 of the \TEX\ document is contained in
\PDF\ object no.~18, \tex{pdfpageref5} expands to the number 18.

%***********************************************************************

\subsection{Form XObjects}

The next three primitives support a \PDF\ feature called \quote {object
reuse} in \PDFTEX. The idea is first to create a `form XObject' in
\PDF. The content of this object corresponds to the content of a \TEX\
box; it can contain pictures and references to other form XObjects
as well. After creation, the form XObject can be used multiple times
by simply referring to its object number. This feature can be useful
for large documents with many similar elements, as it can reduce the
duplication of identical objects.

These commands behave similarly to \type{\pdfobj}, \type{\pdfrefobj}
and \type{\pdflastobj}, but instead of taking raw \PDF\ code, they handle
text typeset by \TEX.

\pdftexprimitive{\Syntax{\Tex{\pdfxform} \Optional{\Something{attr spec}}
  \Optional{\Something{resources spec}} \Something{box number}
  \Modelist{h, v, m}}}
\bookmark{\tex{pdfxform}}

This command creates a form XObject corresponding to the contents of the
box \Something{box number}. The box can contain other raw objects, form
XObjects, or images as well. It can however {\em not} contain annotations
because they are laid out on a separate layer, are positioned absolutely,
and have dedicated housekeeping. \type{\pdfxform} makes the box void,
as \type{\box} does.

When \Something{attr spec} is given, the text will be written
as additional attribute into the form XObject dictionary. The
\Something{resources spec} is similar, but the text will be added
to the resources dictionary of the form XObject. The text given by
\Something{attr spec} or \Something{resources spec} is written before
other entries of the form dictionary and|/|or the resources dictionary
and takes priority over later ones.

\pdftexprimitive{\Syntax{\Tex{\pdfrefxform} \Something{object number}
  \Modelist{h, v, m}}}
\bookmark{\tex{pdfrefxform}}

The form XObject is kept in memory and will be written to the \PDF\
output only when its object number is referred to by \type{\pdfrefxform}
or when \type{\pdfxform} is preceded by \type{\immediate}. Nothing is
appended to the list being built. The number of the most recently created
form XObject is accessible via \type{\pdflastxform}.

When issued, \type{\pdfrefxform} appends a whatsit node to the list being
built. When the whatsit node is searched at shipout time, \PDFTEX\ will
write the form \Something{object number} to the \PDF\ output if it is
not written yet.

\pdftexprimitive{\Syntax{\Tex{\pdflastxform} \Whatever{read||only integer}}}
\bookmark{\tex{pdflastxform}}

The object number of the most recently created form XObject is accessible
via \type{\pdflastxform}.

As said, this feature can be used for reusing information. This mechanism
also plays a role in typesetting fill||in forms. Such widgets sometimes
depends on visuals that show up on user request, but are hidden otherwise.

\pdftexprimitive{\Syntax{\Tex{\pdfxformname} \Something{object number}
  \Whatever{expandable}}}
\bookmark{\tex{pdfxformname}}

In \PDF\ files produced by \PDFTEX\ one can recognize a form Xobject by
the prefix~\type{/Fm} followed by a number, for instance \type{/Fm2}.
For a given form XObject number, this primitive expands to the number in
the corresponding form XObject name. \Eg, if \type{/Fm2} corresponds to
some form XObject with object number 7, the \type{\pdfxformname7}
expands to the number~\type{2}. \introduced{1.30.0}

%***********************************************************************

\subsection{Graphics inclusion}

\PDF\ provides a mechanism for embedding graphic and textual objects: form
XObjects.  In \PDFTEX\ this mechanism is accessed by means of \type{\pdfxform},
\type{\pdflastxform} and \type{\pdfrefxform}. A special kind of XObjects
are bitmap graphics and for manipulating them similar commands are provided.

\pdftexprimitive{\Syntax{\Tex{\pdfximage}
  \Optional{\Something{rule spec}}
  \Optional{\Something{attr spec}}
  \Optional{\Something{page spec}}
  \Optional{\Something{colorspace spec}}
  \Optional{\Something{pdf box spec}}
  \Something{general text}
  \Modelist{h, v, m}
}}
\bookmark{\tex{pdfximage}}

This command creates an image object. The dimensions of the image can be
controlled via \Something{rule spec}\unkern. The default values are zero
for depth and \quote {running} for height and width. If all of them are
given, the image will be scaled to fit the specified values. If some
(but not all) are given, the rest will be set to a value corresponding
to the remaining ones so as to make the image size to yield the same
proportion of $width : (height+depth)$ as the original image size, where
depth is treated as zero. If none are given then the image will take its
natural size.

An image inserted at its natural size often has a resolution of \type
{\pdfimageresolution} (see below) given in dots per inch in the output
file, but some images may contain data specifying the image resolution,
and in such a case the image will be scaled to the correct
resolution. The dimensions of an image can be accessed by enclosing the
\type{\pdfrefximage} command to a box and checking the dimensions of the
box:

\starttyping
\setbox0=\hbox{\pdfximage{somefile.png}\pdfrefximage\pdflastximage}
\stoptyping

Now we can use \type{\wd0} and \type{\ht0} to question the natural size of
the image as determined by \PDFTEX. When dimensions are specified before the
\type{{somefile.png}}, the graphic is scaled to fit these. Note that, unlike
the \eg\ \type{\input} primitive, the filename is supplied between
braces.

The image type is specified by the extension of the given file name:
\type{.png} stands for \PNG\ image, \type{.jpg} (or \type{.jpeg}) for
\JPEG, \type{.jbig2} (preferred, but \type{.jb2} works also) for \JBIGTWO,
and \type{.pdf} for \PDF\ file. But once \PDFTEX\ has opened the file,
it checks the file type first by looking to the magic number at the
file start, which gets precedence over the file name extension. This
gives a certain degree of fault tolerance, if the file name extension
is stated wrongly.

Similarly to \type{\pdfxform}, the optional text given by \Something{attr
spec} will be written as additional attributes of the image before
other keys of the image dictionary. One should be aware, that slightly
different type of \PDF\ object is created while including \PNG, \JPEG,
or \JBIG2\ bitmaps and \PDF\ images.

While working with \PDF\ or \JBIG2\ images, \Something{page spec}
allows to decide which page of the document is to be included;
the \Something{page spec} is irrelevant for the other two image
formats. Starting with \PDFTEX\ 1.11 one may also decide in the \PDF\
image case, which page box of the image is to be treated as a final
bounding box. If \Something{pdf box spec} is present, it overrides the
default behavior specified by the \type{\pdfpagebox} parameter, and is
overridden by the (obsolete) \type{\pdfforcepagebox} parameter. This
option is irrelevant for non||\PDF\ inclusions.

Starting with \PDFTEX\ 1.21, \type{\pdfximage} command supports
\type{colorspace} keyword followed by an object number (user||defined
colorspace for the image being included). This feature works for \JPEG\
images only. \PNG s are \RGB\ palettes, \JBIG2 s are bitonal, and \PDF\
images have always self||contained color space information.

\pdftexprimitive{\Syntax{\Tex{\pdfrefximage} \Something{object number}}}
\bookmark{\tex{pdfrefximage}}

The image is kept in memory and will be written to the \PDF\ output only when
its number is referred to by \type{\pdfrefximage} or \type{\pdfximage} is
preceded by \type{\immediate}. Nothing is appended to the list being built.

\type{\pdfrefximage} appends a whatsit node to the list being built. When
the whatsit node is searched at shipout time, \PDFTEX\ will write the image
with number \Something{object number} to the \PDF\ output if it has not been
written yet.

\pdftexprimitive{\Syntax{\Tex{\pdflastximage} \Whatever{read||only integer}}}
\bookmark{\tex{pdflastximage}}

The number of the most recently created XObject image is accessible via \type
{\pdflastximage}.

\pdftexprimitive{\Syntax{\Tex{\pdfximagebbox} \Something{integer}
  \Something{integer} \Whatever{expandable}}}
\bookmark{\tex{pdfximagebbox}}

The dimensions of the bounding box of a \PDF\ image loaded with
\type{\pdfximage} are stored in a table. This primitive returns those
dimensions as follows:

\starttyping
\pdfximage{afile.pdf}
\pdfximagebbox\pdflastximage 1 % Returns lower-left x
\pdfximagebbox\pdflastximage 2 % Returns lower-left y
\pdfximagebbox\pdflastximage 3 % Returns upper-right x
\pdfximagebbox\pdflastximage 4 % Returns upper-right y
\stoptyping

\pdftexprimitive{\Syntax{\Tex{\pdflastximagecolordepth}
  \Whatever{read||only integer}}}
\bookmark{\tex{pdflastximagecolordepth}}

The color depth (\type{1} for 1-bit images, \type{2} for 2-bit images,
and so on) of the last image accessed with \type{\pdfximage}.

\pdftexprimitive{\Syntax{\Tex{\pdflastximagepages}
  \Whatever{read||only integer}}}
\bookmark{\tex{pdflastximagepages}}

This read||only register returns the highest page number from a file
previously accessed via the \type{\pdfximage} command.
This is useful only for \PDF\ files; it always returns \type{1}
for \PNG, \JPEG, or \JBIGTWO\ files.

\pdftexprimitive{\Syntax{\Tex{\pdfimageresolution} \Whatever{integer}}}
\bookmark{\tex{pdfimageresolution}}

The integer \type{\pdfimageresolution} parameter (unit: dots per
inch, dpi) is a last resort value, used only for bitmap (\JPEG, \PNG,
\JBIGTWO) images, but not for \PDF{}s. The priorities are as follows:
Often one image dimension (\type{width} or \type{height}) is stated
explicitly in the \TEX\ file. Then the image is properly scaled so
that the aspect ratio is kept. If both image dimensions are given,
the image will be stretched accordingly, whereby the aspect ratio might
get distorted. Only if no image dimension is given in the \TEX\ file,
the image size will be calculated from its width and height in pixels,
using the $x$ and $y$ resolution values normally contained in the image
file. If one of these resolution values is missing or weird (either
$<$~0 or $>$~65535), the \type{\pdfimageresolution} value will be used
for both $x$ and $y$ resolution, when calculating the image size. And
if the \type{\pdfimageresolution} is zero, finally a default resolution
of 72\,dpi would be taken. The \type{\pdfimageresolution} is read when
\PDFTEX\ creates an image via \type{\pdfximage}. The given value is
clipped to the range $0..65535$\,(dpi).

Currently this parameter is used particularly for calculating the
dimensions of \JPEG\ images in \EXIF\ format (unless at least one
dimension is stated explicitly); the resolution values coming with
\EXIF\ files are currently ignored.

\pdftexprimitive{\Syntax{\Tex{\pdfpagebox} \Whatever{integer}}}
\bookmark{\tex{pdfpagebox}}

When \PDF\ files are included, the command \type{\pdfximage} allows the
selection of which \PDF\ page box to use in the optional field
\Something{image attr spec}\unkern. If the option isn't present, the
page box defaults to the value of \type{\pdfpagebox} as follows:
(1)~media box, (2)~crop box, (3)~bleed box, (4)~trim box, and
(5)~artbox.

\pdftexprimitive{\Syntax{\Tex{\pdfforcepagebox} \Whatever{integer}}}
\bookmark{\tex{pdfforcepagebox}}

%- It is now possible to specify the pdf page box to use when including pdfs.
%  The syntax has been extended:
%    \pdfximage [<image attr spec>] <general text>           (h, v, m)
%    <image attr spec> --> [<rule spec>] [<attr spec>] [<page spec>] [<pdf box spec>]
%    <pdf box spec> --> mediabox|cropbox|bleedbox|trimbox|artbox
%  The default is cropbox (which defaults to mediabox), to be compatible with
%  previous versions of pdfTeX.
%    The page box specified at \pdfximage can be globally overridden with the
%  config parameter always_use_pdfpagebox and the command
%  \pdfoptionalwaysusepdfpagebox <integer>, where 0 is the default (use
%  whatever is specified with \pdfximage), 1 is media, 2 is crop, 3 is
%  bleed, 4 is trim and 5 is artbox. This can only be set once per
%  object, i.\,e.\ the value used at \pdfximage is remembered.
%    See the PDF Reference for an explanation of the boxes.

The integer primitive \type{\pdfforcepagebox} allows globally
overriding the choice of the page box used with \type{\pdfximage}. It
takes the same values as \type{\pdfpagebox}. The command is available
starting with \PDFTEX\ 1.30.0, as a shortened synonym of obsolete
\type{\pdfoptionalwaysusepdfpagebox} instruction, but is itself
now considered obsolete --- a mixture of \type{\pdfpagebox} and
\Something{image attr spec} is better.

\pdftexprimitive{\Syntax{\Tex{\pdfinclusionerrorlevel} \Whatever{integer}}}
\bookmark{\tex{pdfinclusionerrorlevel}}

This controls the behavior of \PDFTEX\ when a \PDF\ file is included
that has a newer version than the one specified by
\type{\pdfminorversion}.  If \type{\pdfinclusionerrorlevel} is set to~0
(the default), \PDFTEX\ gives only a warning; if 1, \PDFTEX\ raises an
error; if negative, no diagnostic at all is given.

It was originally a shortened synonym of
\type{\pdfoptionpdfinclusionerrorlevel}, which is now obsolete.
\introduced{1.30.0}

\pdftexprimitive{\Syntax{\Tex{\pdfimagehicolor} \Whatever{integer}}}
\bookmark{\tex{pdfimagehicolor}}

This primitive, when set to~1, enables embedding of \PNG\ images with
16~bit wide color channels at their full color resolution. As such an
embedding mode is defined only from \PDF\ version~1.5 onwards, the
\type{\pdfimagehicolor} functionality is automatically disabled in
\PDFTEX\ if \type{\pdfminorversion}~$<$~5; then each 16~bit color
channel is reduced to a width of 8~bit by stripping the lower 8~bits
before embedding. The same stripping happens when
\type{\pdfimagehicolor} is set to~0. For \type{\pdfminorversion}~$\ge$~5
the default value of \type{\pdfimagehicolor} is~1. If specified, the
parameter must appear before any data is written to the \PDF\ output.
\introduced{1.30.0}

\pdftexprimitive{\Syntax{\Tex{\pdfimageapplygamma} \Whatever{integer}}}
\bookmark{\tex{pdfimageapplygamma}}

This primitive, when set to~1, enables gamma correction while embedding
\PNG\ images, taking the values of the primitives \type{\pdfgamma} as
well as the gamma value embedded in the \PNG\ image into account. When
\type{\pdfimageapplygamma} is set to~0, no gamma correction is
performed.  If specified, the parameter must appear before any data is
written to the \PDF\ output.  \introduced{1.30.0}

\pdftexprimitive{\Syntax{\Tex{\pdfgamma} \Whatever{integer}}}
\bookmark{\tex{pdfgamma}}

This primitive defines the `device gamma' for \PDFTEX. Values are in
promilles (same as \type{\mag}). The default value of this primitive
is~1000, defining a device gamma value of~1.0.

When \type{\pdfimageapplygamma} is set to~1, then whenever a \PNG\ image
is included, \PDFTEX\ applies a gamma correction. This correction is
based on the  value of the \type{\pdfgamma} primitive and the `assumed
device gamma' that is derived from the value embedded in the actual
image. If no embedded value can be found in the \PNG\ image, then the
value of \type{\pdfimagegamma} is used instead.
If specified, the parameter must appear before any data is written to the
\PDF\ output.
\introduced{1.30.0}

\pdftexprimitive{\Syntax{\Tex{\pdfimagegamma} \Whatever{integer}}}
\bookmark{\tex{pdfimagegamma}}

This primitive gives a default `assumed gamma' value for \PNG\ images.
Values are in promilles (same as for \type{\pdfamma}). The default value
of this primitive is~2200, implying an assumed gamma value of~2.2.

When \PDFTEX\ is applying gamma corrections, images that do not have an
embedded `assumed gamma' value are assumed to have been created for
a device with a gamma of 2.2. Experiments show that this default setting
is correct for a large number of images; however, if your images come
out too dark, you probably want to set \type{\pdfimagegamma} to a lower
value, like~1000.
If specified, the parameter must appear before any data is written to the
\PDF\ output.
\introduced{1.30.0}

\pdftexprimitive{\Syntax{\Tex{\pdfpxdimen} \Whatever{dimen}}}
\bookmark{\tex{pdfpxdimen}}

While working with bitmap graphics or typesetting electronic documents,
it might be convenient to base dimensions on pixels rather than \TeX's
standard units like \type{pt} or \type{em}. For this purpose, \PDFTEX\
provides an extra unit called \type{px} that takes the dimension given to
the \type{\pdfpxdimen} primitive. In example, to make the unit \type{px}
corresponding to 96\,dpi pixel density (then 1\,px~=~72/96\,bp), one
can do the following calculation:

\starttyping
\pdfpxdimen=1in % 1 dpi
\divide\pdfpxdimen by 96 % 96 dpi
\hsize=1200px
\stoptyping

Then \type{\hsize} amounts to 1200~pixels in 96\,dpi, which is
exactly 903.375\,pt (but \TeX\ rounds it to 903.36914\,pt). The
default value of \type{\pdfpxdimen} is 1\,bp, corresponding to a pixel
density of 72\,dpi. This primitive is completely independent from the
\type{\pdfimageresolution} and \type{\pdfpkresolution} parameters.
\introduced{1.30.0} It used to be an integer register that gave
the dimension 1\,px as number of scaled points, defaulting to 65536
(1\,px equal to 65536\,sp~$=$~1\,pt). Starting with \PDFTEX\ 1.40.0,
\type{\pdfpxdimen} is now a real dimension parameter.

\pdftexprimitive{\Syntax{\Tex{\pdfinclusioncopyfonts} \Whatever{integer}}}
\bookmark{\tex{pdfinclusioncopyfonts}}

If positive, this parameter forces \PDFTEX\ to include fonts from a \PDF\
file loaded with \type{\pdfximage}, even if those fonts are available on
disk. Bigger files might be created, but included \PDF\ files are sure to
be embedded with the adequate fonts; indeed, the fonts on disk might be
different from the embedded ones, and glyphs might be missing.

\pdftexprimitive{\Syntax{\Tex{\pdfsuppresswarningpagegroup}
  \Whatever{integer}}}
\bookmark{\tex{pdfsuppresswarningpagegroup}}

Ordinarily, \PDFTEX\ gives a warning when more than one included \PDF\
file has a so-called ``page group object'' (\type{/Group}), because only
one can ``win'' --- that is, be propagated to the page level.  Usually
the page groups are identical, but when they are not, the result is
unpredictable.  It would be ideal if \PDFTEX\ in fact detected whether
the page groups were the same and only gave the warning in the
problematic case; unfortunately, this is not easy (a patch would be
welcome).  Nevertheless, often one observes that there is no actual
problem.  Then seeing the warnings on every run is just noise, and can
be suppressed by setting this parameter to a positive number.
\introduced{1.40.15}

%***********************************************************************

\subsection{Annotations}

\PDF\ 1.4 provides four basic kinds of annotations:

\startitemize[packed]
\item hyperlinks, general navigation
\item text clips (notes)
\item movies
\item sound fragments
\stopitemize

The first type differs from the other three in that there is a designated
area involved on which one can click, or when moved over some action occurs.
\PDFTEX\ is able to calculate this area, as we will see later. All
annotations can be supported using the next two general annotation
primitives.

\pdftexprimitive{\Syntax{\Tex{\pdfannot} \Something{annot type spec}
  \Modelist{h, v, m}}}
\bookmark{\tex{pdfannot}}

This command appends a whatsit node corresponding to an annotation to
the list being built. The dimensions of the annotation can be controlled
via the \Something{rule spec}. The default values are running for all
width, height and depth. When an annotation is written out, running
dimensions will take the corresponding values from the box containing
the whatsit node representing the annotation. The \Something{general
text} is inserted as raw \PDF\ code to the contents of annotation. The
annotation is written out only if the corresponding whatsit node is
searched at shipout time.

\pdftexprimitive{\Syntax{\Tex{\pdflastannot} \Whatever{read||only integer}}}
\bookmark{\tex{pdflastannot}}

This primitive returns the object number of the last annotation created by
\type{\pdfannot}. These two primitives allow users to create any annotation
that cannot be created by \type{\pdfstartlink} (see below).

%***********************************************************************

\subsection[linking]{Destinations and links}

The first type of annotation, mentioned above, is implemented by three
primitives. The first one is used to define a specific location as being
referred to. This location is tied to the page, not the exact location on the
page. The main reason for this is that \PDF\ maintains a dedicated list of
these annotations |<|and some more when optimized|>| for the sole purpose of
speed.

\pdftexprimitive{\Syntax{\Tex{\pdfdest} \Something{dest spec}
  \Modelist{h, v, m}}}
\bookmark{\tex{pdfdest}}

This primitive appends a whatsit node which establishes a destination
for links and bookmark outlines; the link is identified by either
a number or a symbolic name, and the way the viewer is to display the
page must be specified in \Something{dest type}\unkern, which must be
one of those mentioned in \in{table}[appearance].

\startbuffer
\starttabulate[|l|l|]
\HL
\NC \bf keyword \NC \bf meaning                                       \NC\NR
\HL
\NC \tt fit    \NC fit the page in the window                         \NC\NR
\NC \tt fith   \NC fit the width of the page                          \NC\NR
\NC \tt fitv   \NC fit the height of the page                         \NC\NR
\NC \tt fitb   \NC fit the \quote{Bounding Box} of the page           \NC\NR
\NC \tt fitbh  \NC fit the width of \quote{Bounding Box} of the page  \NC\NR
\NC \tt fitbv  \NC fit the height of \quote{Bounding Box} of the page \NC\NR
\NC \tt xyz    \NC goto the current position (see below)              \NC\NR
\HL
\stoptabulate
\stopbuffer

\placetable
  [here][appearance]
  {Options for display of outline and destinations.}
  {\getbuffer}

The specification \Literal{xyz} can optionally be followed by \Literal
{zoom} \Something{integer} to provide a fixed zoom||in. The \Something
{integer} is processed like \TEX\ magnification, i.\,e.\ 1000 is the
normal page view. When \Literal{zoom} \Something{integer} is given,
the zoom factor changes to 0.001 of the \Something{integer} value,
otherwise the current zoom factor is kept unchanged.

The destination is written out only if the corresponding whatsit node is
searched at shipout time.

\pdftexprimitive{\Syntax{\Tex{\pdfstartlink}
  \Optional{\Something{rule spec}}
  \Optional{\Something{attr spec}}
  \Something{action spec}
  \Modelist{h, m}
}}
\bookmark{\tex{pdfstartlink}}

This primitive is used along with \type{\pdfendlink} and appends
a whatsit node corresponding to the start of a hyperlink. The whatsit
node representing the end of the hyperlink is created by
\type{\pdfendlink}. The dimensions of the link are handled in the
similar way as in \type{\pdfannot}. Both \type{\pdfstartlink} and
\type{\pdfendlink} must be in the same level of box nesting. A hyperlink
with running width can be multi||line or even multi||page, in which case
all horizontal boxes with the same nesting level as the boxes containing
\type{\pdfstartlink} and \type{\pdfendlink} will be treated as part of
the hyperlink. The hyperlink is written out only if the corresponding
whatsit node is searched at shipout time.

Additional attributes, which are explained in great detail in the
\PDFReference, can be given via \Something{attr spec}\unkern. Typically,
the attributes specify the color and thickness of any border around
the link.  Thus \typ {/C [0.9 0 0] /Border [0 0 2]} specifies a color
(in \RGB) of dark red, and a border thickness of 2~points.

While all graphics and text in a \PDF\ document have relative positions,
annotations have internally hard||coded absolute positions. Again this
is for the sake of speed optimization. The main disadvantage is that these
annotations do {\em not} obey transformations issued by \type
{\pdfliteral}'s.

The \Something{action spec} specifies the action that should be performed
when the hyperlink is activated while the \Something{user-action spec}
performs a user||defined action. A typical use of the latter is to specify
a \URL, like \typ {/S /URI /URI (http://www.tug.org/)}, or a named action like
\typ {/S /Named /N /NextPage}.

A \Something{goto-action spec} performs a GoTo action. Here \Something
{numid} and \Something{nameid} specify the destination identifier (see
below). The \Something{page spec} specifies the page number of the
destination, in this case the zoom factor is given by \Something{general
text}\unkern. A destination can be performed in another \PDF\ file by
specifying \Something{file spec}\unkern, in which case
\Something{newwindow spec} specifies whether the file should be opened
in a new window. A \Something{file spec} can be either a \type{(string)}
or a \type{<<dictionary>>}. The default behavior of the
\Something{newwindow spec} depends on the browser setting.

A \Something{thread-action spec} performs an article thread reading. The
thread identifier is similar to the destination identifier. A thread can be
performed in another \PDF\ file by specifying a \Something{file spec}\unkern.

\pdftexprimitive{\Syntax{\Tex{\pdfendlink} \Modelist{h, m}}}
\bookmark{\tex{pdfendlink}}

This primitive ends a link started with \type{\pdfstartlink}. All text
between \type{\pdfstartlink} and \type{\pdfendlink} will be treated as part
of this link. \PDFTEX\ may break the result across lines (or pages), in which
case it will make several links with the same content.

\pdftexprimitive{\Syntax{\Tex{\pdflastlink} \Whatever{read||only integer}}}
\bookmark{\tex{pdflastlink}}

This primitive returns the object number of the last link created by
\type{\pdfstartlink} (analogous to \type{\pdflastannot}).
\introduced{1.40.0}

\pdftexprimitive{\Syntax{\Tex{\pdflinkmargin} \Whatever{dimen}}}
\bookmark{\tex{pdflinkmargin}}

This dimension parameter specifies the margin of the box representing
a hyperlink and is read when a page containing hyperlinks is shipped out.

\pdftexprimitive{\Syntax{\Tex{\pdfdestmargin} \Whatever{dimen}}}
\bookmark{\tex{pdfdestmargin}}

Margin added to the dimensions of the rectangle around the destinations.

\pdftexprimitive{\Syntax{\Tex{\pdfsuppresswarningdupdest} \Whatever{integer}}}
\bookmark{\tex{pdfsuppresswarningdupdest}}

Ordinarily, \PDFTEX\ gives a warning when the same destination is used
more than once.  However, due to problematic macro packages, sometimes a
document may end up producing the warning through no fault of its own,
and fixing the macros may well not be trivial.  Then seeing the warnings
on every run is just noise, and can be suppressed by setting this
parameter to a positive number.  \introduced{1.40.13}

%***********************************************************************

\subsection{Bookmarks}

\pdftexprimitive{\Syntax{\Tex{\pdfoutline}
  \Optional{\Something{attr spec}}
  \Something{action spec}
  \Optional{\Literal{count} \Something{integer}}
  \Something{general text}
  \Modelist{h, v, m}
}}
\bookmark{\tex{pdfoutline}}

This primitive creates an outline (or bookmark) entry. The first parameter
specifies the action to be taken, and is the same as that allowed for \type
{\pdfstartlink}. The \Something{count} specifies the number of direct
subentries under this entry; specify~0 or omit it if this entry has no
subentries. If the number is negative, then all subentries will be closed and
the absolute value of this number specifies the number of subentries. The
\Something{text} is what will be shown in the outline window. Note that this
is limited to characters in the \PDF\ Document Encoding vector. The outline
is written to the \PDF\ output immediately.

%***********************************************************************

\subsection{Article threads}

\pdftexprimitive{\Syntax{\Tex{\pdfthread} \Optional{\Something{rule spec}}
  \Optional{\Something{attr spec}} \Something{id spec} \Modelist{h, v, m}}}
\bookmark{\tex{pdfthread}}

Defines a bead within an article thread. Thread beads with same
identifiers (spread across the document) will be joined together.

\pdftexprimitive{\Syntax{\Tex{\pdftstartthread}
  \Optional{\Something{rule spec}}
  \Optional{\Something{attr spec}}
  \Something{id spec} \Modelist{v, m}}}
\bookmark{\tex{pdfstartthread}}

This uses the same syntax as \type{\pdfthread}, apart that it must be
followed by a \type{\pdfendthread}.  \type{\pdfstartthread} and the
corresponding \type{\pdfendthread} must end up in vboxes with the same
nesting level; all vboxes between them will be added into the thread.
Note that during output runtime if there are other newly created
boxes which have the same nesting level as the vbox/vboxes containing
\type{\pdfstartthread} and \type{\pdfendthread}, they will be also added
into the thread, which is probably not what you want. To avoid such
unconsidered behavior, it's often enough to wrap boxes that shouldn't
belong to the thread by a box to change their box nesting level.

\pdftexprimitive{\Syntax{\Tex{\pdfendthread} \Modelist{v, m}}}
\bookmark{\tex{pdfendthread}}

This ends an article thread started before by \type{\pdfstartthread}.

\pdftexprimitive{\Syntax{\Tex{\pdfthreadmargin} \Whatever{dimen}}}
\bookmark{\tex{pdfthreadmargin}}

Specifies a margin to be added to the dimensions of a bead within
an article thread.

%***********************************************************************

\subsection{Literals and specials}

\pdftexprimitive{\Syntax{\Tex{\pdfliteral}
  \Optional{\Something{pdfliteral spec}}
  \Something{general text}
  \Modelist{h, v, m}
  }}
\bookmark{\tex{pdfliteral}}

Like \type{\special} in normal \TEX, this command inserts raw
\PDF\ code into the output. This allows support of color and text
transformation. This primitive is heavily used in the \METAPOST\
inclusion macros. Normally \PDFTEX\ ends a text section in the \PDF\
output and sets the transformation matrix to the current location on the
page before inserting \Something{general text}\unkern, however this can be
turned off by giving the optional keyword \Literal{direct}. This command
appends a whatsit node to the list being built. \Something{general text}
is expanded when the whatsit node is created and not when it is shipped
out, as with \type{\special}.

Starting with version 1.30.0, \PDFTEX\ allows to use a new keyword
\type{page} instead of \type{direct}. Both modify the default behavior
of \type{\pdfliteral}, avoiding translation of the coordinates space
before inserting the literal code. The difference is that the \type{page}
keyword instructs \PDFTEX\ to close a \type{BT ET} text block before
inserting anything. It means that the literal code inserted refers to the
origin (lower||left corner of the page) and can be safely enclosed with
\type{q Q}. Note, that in most cases using \type{q Q} operators inside
\type{\pdfliteral} with \type{direct} keyword will produce corrupted
\PDF\ output, as the \PDF\ standard doesn't allow to do anything like
this within a \type{BT ET} block.

% HE: \unkern is a kludge here; wanted to have tight "{pdf:"
\pdftexprimitive{\Syntax{\Tex{\special} \Lbrace\unkern\Literal{pdf:}
  \Something{text} \Rbrace}}
\bookmark{\tex{special}}

This is equivalent to \Syntax{\Tex{\pdfliteral} \Lbrace \Something{text}
\Rbrace}.

\pdftexprimitive{\Syntax{\Tex{\special} \Lbrace\unkern\Literal{pdf:direct:}
  \Something{text} \Rbrace}}
\bookmark{\tex{special\ direct}}

This is equivalent to \Syntax{\Tex{\pdfliteral} \Literal{direct} \Lbrace
\Something{text} \Rbrace}.

\pdftexprimitive{\Syntax{\Tex{\special} \Lbrace\unkern\Literal{pdf:page:}
  \Something{text} \Rbrace}}
\bookmark{\tex{special\ page}}

This is equivalent to \Syntax{\Tex{\pdfliteral} \Literal{page} \Lbrace
\Something{text} \Rbrace}.

%***********************************************************************

\subsection{Strings}

\pdftexprimitive{\Syntax{\Tex{\pdfescapestring} \Something{general text}
  \Whatever{expandable}}}
\bookmark{\tex{pdfescapestring}}

Starting with version 1.30.0, \PDFTEX\ provides a mechanism for converting
a general text into \PDF\ string. Many characters that may be needed inside
such a text (especially parenthesis), have a special meaning inside a \PDF\
string object and thus, can't be used literally. The primitive replaces each
special \PDF\ character by its literal representation by inserting a backslash
before that character. Some characters (\eg\ space) are also converted into
3||digit octal number. In example, \type{\pdfescapestring{Text (1)}} will be
expanded to \type{Text\040\(1\)}. This ensures a literal interpretation of the
text by the \PDF\ viewer.
\introduced{1.30.0}

\pdftexprimitive{\Syntax{\Tex{\pdfescapename} \Something{general text}
  \Whatever{expandable}}}
\bookmark{\tex{pdfescapename}}

In analogy to \type{\pdfescapestring}, \type{\pdfescapename} replaces each
special \PDF\ character inside the general text by its hexadecimal
representation preceded by \type{#} character. This ensures the proper
interpretation of the text if used as a \PDF\ name object. In example,
\type{Text (1)} will be replaced by \type{Text#20#281#29}.
\introduced{1.30.0}

\pdftexprimitive{\Syntax{\Tex{\pdfescapehex} \Something{general text}
  \Whatever{expandable}}}
\bookmark{\tex{pdfescapehex}}

This command converts each character of \Something{general text} into its
hexadecimal representation. Each character of the argument becomes a pair of
hexadecimal digits. \introduced{1.30.0}

\pdftexprimitive{\Syntax{\Tex{\pdfunescapehex} \Something{general text}
  \Whatever{expandable}}}
\bookmark{\tex{pdfunescapehex}}

This command treats each character pair of \Something{general text} as
a hexadecimal number and returns an \ASCII\ character of this code.
\introduced{1.30.0}

\pdftexprimitive{\Syntax{\Tex{\pdfstrcmp} \Something{general text}
  \Something{general text} \Whatever{expandable}}}
\bookmark{\tex{pdfstrcmp}}

This command compares two strings and expands to \type{0} if the strings
are equal, to \type{-1} if the first string ranks before the second, and
to \type{1} otherwise.  \introduced{1.30.0}

\pdftexprimitive{\Syntax{\Tex{\pdfmatch} \Optional{\Literal{icase}}
  \Optional{\Literal{subcount} \Something{integer}} \Something{general text}
  \Something{general text} \Whatever{expandable}}}
\bookmark{\tex{pdfmatch}}

This command implements pattern matching (using the syntax of \POSIX\
extended regular expressions). The first \Something{general text} is a
pattern and the second is a string.  The command expands to \type{-1} if
the pattern is invalid, to \type{0} if no match is found, and to
\type{1} if a match is found. With the \type{icase} option, the matching
is case-insensitive.  The \type{subcount} option sets the size of the
table storing the found (sub)patterns; its default is 10.
\introduced{1.30.0}

\pdftexprimitive{\Syntax{\Tex{\pdflastmatch} \Something{integer}
  \Whatever{expandable}}}
\bookmark{\tex{pdflastmatch}}

The matches found with \type{\pdfmatch} are stored in a table. This
command returns the entry for \Something{integer}, in the format
\Something{position}\type{->}\Something{string}; \Something{position} is
the position of the match (starting at zero) or \type{-1} if no match
was found, and \Something{string} is the matched substring.

Entry~0 contains the match as a whole; the subsequent entries contain
submatches corresponding to the subpatterns, up to \type{subcount-1}.

If \Something{integer} is less than zero, an error is given.

For instance:

\starttyping
\pdfmatch subcount 3 {ab(cd)*ef(gh)(ij)}{abefghij}
\pdflastmatch0 % "0->abefghij"
\pdflastmatch1 % "-1->"
\pdflastmatch2 % "4->gh"
\pdflastmatch3 % "-1->"
\stoptyping

Entry~1 is empty because no match was found for \type{cd}, and entry~3
is empty because it exceeds the table's size as set by \type{subcount}.
\introduced{1.30.0}

%***********************************************************************

\subsection{Numbers}

\pdftexprimitive{\Syntax{\Tex{\ifpdfabsnum} \Whatever{expandable}}}
\bookmark{\tex{ifpdfabsnum}}

This conditional works like the standard \type{\ifnum}, except that it
compares absolute values of numbers. Although it seems to be a trivial
shortcut for a couple of \type{\ifnum x<0} tests, as a primitive it is
considerably more friendly in usage and works a bit faster.
\introduced{1.40.0}

\pdftexprimitive{\Syntax{\Tex{\ifpdfabsdim} \Whatever{expandable}}}
\bookmark{\tex{ifpdfabsdim}}

Analogous to \type{\ifpdfabsnum}, this conditional works like
\type{\ifdim}, except that it compares absolute values of
dimensions. \introduced{1.40.0}

\pdftexprimitive{\Syntax{\Tex{\pdfuniformdeviate} \Something{number}
  \Whatever{expandable}}}
\bookmark{\tex{pdfuniformdeviate}}

Generate a uniformly-distributed random integer value between 0
(inclusive) and \Something{number} (exclusive). This primitive expands
to a list of tokens. \introduced{1.30.0}

\pdftexprimitive{\Syntax{\Tex{\pdfnormaldeviate} \Whatever{expandable}}}
\bookmark{\tex{pdfnormaldeviate}}

Generate a normally-distributed random integer with a mean of~0 and
standard deviation 65\,536.  That is, about 68\% of the time, the result
will be between $-65536$ and $65536$ (one standard deviation away from
the mean).  About 95\% of results will be within two standard
deviations, and 99.7\% within three.  This primitive expands to a list
of tokens. \introduced{1.30.0}

\pdftexprimitive{\Syntax{\Tex{\pdfrandomseed} \Whatever{read||only integer}}}
\bookmark{\tex{pdfrandomseed}}

You can use \type{\the\pdfrandomseed} to query the current seed value,
so you can \eg\ write the value to the log file.  The initial value of
the seed is derived from the system time, and is not more than
1\,000\,999\,999 (this ensures that the value can be used with commands
like \type{\count}). \introduced{1.30.0}

\pdftexprimitive{\Syntax{\Tex{\pdfsetrandomseed} \Something{number}}}
\bookmark{\tex{pdfsetrandomseed}}

Set the random seed (\type{\pdfrandomseed}) to a specific value,
allowing you to replay sequences of semi-randoms at a later moment.
\introduced{1.30.0}

%***********************************************************************

\subsection{Timekeeping}

\pdftexprimitive{\Syntax{\Tex{\pdfelapsedtime} \Whatever{read||only integer}}}
\bookmark{\tex{pdfelapsedtime}}

Return a number that represents the time elapsed from the moment of the
start of the run. The elapsed time is returned in `scaled seconds',
meaning seconds divided by 65536, \eg\ \PDFTEX\ has run for
\the\pdfelapsedtime\ `scaled seconds' when this paragraph was
typeset. Of course, the command will never return a value greater than
the highest number available in \TeX: if the time exceeds 32767 seconds,
the constant value $2^{31}-1$ will be returned.  \introduced{1.30.0}

\pdftexprimitive{\Syntax{\Tex{\pdfresettimer}}}
\bookmark{\tex{pdfresettimer}}

Reset the internal timer so that \type{\pdfelapsedtime}
starts returning micro||time from~0 again. \introduced{1.30.0}

%***********************************************************************

\subsection{Files}

\pdftexprimitive{\Syntax{\Tex{\pdffilemoddate} \Something{general text}
  \Whatever{expandable}}}
\bookmark{\tex{pdffilemoddate}}

Expands to the modification date of file \Something{general text} in the same
format as for \type{\pdfcreationdate}, \eg\ it's {\tt \pdffilemoddate
{./pdftex-t.tex}} for the source of this manual.
\introduced{1.30.0}

\pdftexprimitive{\Syntax{\Tex{\pdffilesize} \Something{general text}
  \Whatever{expandable}}}
\bookmark{\tex{pdffilesize}}

Expands to the size of file \Something{general text}, \eg\ it's {\tt
\pdffilesize {./pdftex-t.tex}} for the source of this manual.
\introduced{1.30.0}

\pdftexprimitive{\Syntax{\Tex{\pdfmdfivesum} file \Something{general text}
  \Whatever{expandable}}}
\bookmark{\tex{pdfmdfivesum\ file}}

Expands to the \MDFIVE\ of file \Something{general text} in uppercase
hexadecimal format (same as \type{\pdfescapehex}), \eg\ it's {\tt
\pdfmdfivesum file {./pdftex-t.tex}} for the source of this manual.
\introduced{1.30.0}

\pdftexprimitive{\Syntax{\Tex{\pdffiledump} \Optional{\Literal{offset}
  \Something{number}} \Optional{\Literal{length}
  \Something{number}} \Something{general text} \Whatever{expandable}}}
\bookmark{\tex{pdffiledump}}

Expands to the dump of the file \Something{general text} in uppercase
hexadecimal format (same as \type{\pdfescapehex}), starting at offset
\Something{number} or 0 with length \Something{number}, if given. The first ten
bytes of the source of this manual are {\tt \pdffiledump length 10
{./pdftex-t.tex}}. \introduced{1.30.0}

%***********************************************************************

\subsection{Color stack}

\PDFTEX\ 1.40.0 comes with color stack support (actually any graphic state
stack).

\pdftexprimitive{\Syntax{\Tex{\pdfcolorstackinit} \Optional{\Literal{page}}
  \Optional{\Literal{direct}} \Something{general text} \Whatever{expandable}}}
\bookmark{\tex{pdfcolorstackinit}}

The primitive initializes a new graphic stack and returns its number. Optional
\Literal{page} keyword instructs \PDFTEX\ to restore the graphic at the
beginning of every new page. Also optional \Literal{direct} has the same effect
as for \Tex{\pdfliteral} primitive. \introduced{1.40.0}

\pdftexprimitive{\Syntax{\Tex{\pdfcolorstack} \Something{stack number}
  \Something{stack action} \Something{general text}}}
\bookmark{\tex{pdfcolorstack}}

The command operates on the stack of a given number. If \Something{stack
action} is \Literal{push} keyword, the new value provided as
\Something{general text} is inserted into the top of the graphic stack
and becomes the current stack value. If followed by \Literal{pop}, the
top value is removed from the stack and the new top value becomes the
current. \Literal{set} keyword replaces the current value with
\Something{general text} without changing the stack size.
\Literal{current} keyword instructs just to use the current stack value
without modifying the stack at all. \introduced{1.40.0}

\subsection{Transformations}

Since the content of \Tex{\pdfliteral} is not interpreted anyhow, any
transformation inserted directly into the content stream, as well as saving
and restoring the current transformation matrix, remains unnoticed by
\PDFTEX\ positioning mechanism. As a consequence, links and other annotations
(that are formed in \PDF\ as different layer then the page content) are not
affected by such user-defined transformations. \PDFTEX\ 1.40.0 solves this
problem with three new primitives.

\pdftexprimitive{\Syntax{\Tex{\pdfsetmatrix}}}
\bookmark{\tex{pdfsetmatrix}}

Afine transformations are normally expressed with six numbers. First
four (no unit) values defining scaling, rotating and skewing, plus two
extra dimensions for shifting. Since the translation is handled by \TeX\
itself, \Tex{\pdfsetmatrix} primitive expects as an argument a string
containing just the first four numbers of the transformation separated
by a space and assumes two position coordinates to be~0. In example,
\type{\pdfsetmatrix{0.87 -0.5 0.5 0.87}} rotates the current space by 30
degrees, inserting \type{0.87 -0.5 0.5 0.87 0 0 cm} into the content
stream. \introduced{1.40.0}

\pdftexprimitive{\Syntax{\Tex{\pdfsave}}}
\bookmark{\tex{pdfsave}}

The command saves the current transformation by inserting \type{q}
operator into the content stream. \introduced{1.40.0}

\pdftexprimitive{\Syntax{\Tex{\pdfrestore}}}
\bookmark{\tex{pdfrestore}}

The command restores previously saved transformation by inserting \type{Q}
operator into the content stream. One should keep in mind that \Tex{\pdfsave}
and \Tex{\pdfrestore} pairs should always be properly nested and should start
and end at the same group level. \introduced{1.40.0}

%***********************************************************************

\subsection{Miscellaneous}

\pdftexprimitive{\Syntax{\tex {ifincsname} \Whatever{expandable}}}
\bookmark{\tex{ifincsname}}

This conditional is true if evaluated inside \type{\csname ... \endcsname},
and false otherwise.


\pdftexprimitive{\Syntax{\tex {ifpdfprimitive} \Something{control sequence}
  \Whatever{expandable}}}
\bookmark{\tex{ifpdfprimitive}}

This condition checks if the following control sequence has its
primitive meaning. If it has, \type{\ifpdfprimitive} returns true. In
any other case (redefined, made \type{\undefined}, has never been
primitive) false is returned.  \introduced{1.40.0}


\pdftexprimitive{\Syntax{\Tex{\pdfcreationdate} \Whatever{expandable}}}
\bookmark{\tex{pdfcreationdate}}

Expands to the date string \PDFTEX\ uses in the info dictionary of the
document, \eg\ for this file {\tt\pdfcreationdate}. \introduced{1.30.0}


\pdftexprimitive{\Syntax{\tex {pdfdraftmode} \Whatever{integer}}}
\bookmark{\tex{pdfdraftmode}}

When set to 1 (or set by the command-line switch \type{-draftmode})
\PDFTEX\ doesn't write the output \PDF\ file and doesn't actually read any
images but does everything else (including writing auxiliary files),
thus speeding up compilations when you know you need an extra run but
don't care about the output, \eg\ just to get the \BIBTEX\ references
right.  \introduced{1.40.0}


\pdftexprimitive{\Syntax{\Tex{\pdfinsertht} \Something{integer}
  \Whatever{expandable}}}
\bookmark{\tex{pdfinsertht}}

If \Something{integer} is the number of an insertion class, this command
returns the height of the corresponding box at the current time.
For instance, the following returns \type{12pt} in plain \TEX:

\starttyping
Abc\footnote*{Whatever.}\par
\pdfinsertht\footins
\stoptyping


\pdftexprimitive{\Syntax{\Tex{\pdflastxpos} \Whatever{read||only integer}}}
\bookmark{\tex{pdflastxpos}}

This primitive returns an integer number representing the absolute $x$
coordinate of the last point marked by \type{\pdfsavepos}. The unit is
`scaled points' (sp).


\pdftexprimitive{\Syntax{\Tex{\pdflastypos} \Whatever{read||only integer}}}
\bookmark{\tex{pdflastypos}}

Completely analogous to \type{\pdflastxpos}, returning the $y$ coordinate.


\pdftexprimitive{\Syntax{\tex {pdfprimitive} \Something{control sequence}}}
\bookmark{\tex{pdfprimitive}}

This command executes the primitive meaning of the following control
sequence, regardless of whether the control sequence has been redefined
or made undefined. If the primitive was expandable, \type{\pdfprimitive}
expands also. On the other hand, if the following control sequence never
was a primitive, nothing happens and no error is raised. (In some
versions of \PDFTEX\ prior to 1.40.19, an error was wrongly given.)
\introduced{1.40.0}


\pdftexprimitive{\Syntax{\Tex{\pdfretval} \Whatever{read||only integer}}}
\bookmark{\tex{pdfretval}}

Set to $-1$ if \type{\pdfobj} ignores an invalid object number.  Perhaps
this will be used to store the error status of other primitives in the
future.


\pdftexprimitive{\Syntax{\Tex{\pdfsavepos} \Modelist{h, v, m}}}
\bookmark{\tex{pdfsavepos}}

This primitive marks the current absolute ($x,y$) position on the
media, with the reference point in the lower left corner. It is active
only during page shipout, when the page is finally assembled. The
position coordinates can then be retrieved by the \type{\pdflastxpos}
and \type{\pdflastypos} primitives, and \eg\ written out to some
auxiliary file. The coordinates can be used only after the current
\type{\shipout} has been finalized, therefore normally two \PDFTEX\
runs are required to utilize these primitives. Starting with
\PDFTEX\ 1.40.0, this mechanism can be used also while running
in \DVI\ mode.


\pdftexprimitive{\Syntax{\Tex{\pdfshellescape} \Whatever{read||only integer}}}
\bookmark{\tex{pdfshellescape}}

This primitive is~1 if \type{\write18} is enabled, 2 if it is
restricted, and 0 otherwise.  (\type{\write18} was
\ifnum\pdfshellescape=0\relax not \fi enabled when this manual was
typeset.)  \introduced{1.30.0}


\pdftexprimitive{\Syntax{\Tex{\pdftexbanner} \Whatever{expandable}}}
\bookmark{\tex{pdftexbanner}}

Returns the \PDFTEX\ banner message, \eg\ for the version used here:
{\tt \pdftexbanner}. \introduced{1.20a}


\pdftexprimitive{\Syntax{\Tex{\pdftexrevision} \Whatever{expandable}}}
\bookmark{\tex{pdftexrevision}}

\def\versplit#1#2#3{#1.#2#3}

Returns the revision number of \PDFTEX, \eg\ for \PDFTEX\ version
\expandafter\versplit\the\pdftexversion.\pdftexrevision\ (used to produce
this document), it returns the number {\tt \pdftexrevision}.


\pdftexprimitive{\Syntax{\Tex{\pdftexversion} \Whatever{read||only integer}}}
\bookmark{\tex{pdftexversion}}

Returns the version of \PDFTEX\ multiplied by 100, \eg\ for \PDFTEX\
version \expandafter\versplit\the\pdftexversion.\pdftexrevision\ used to
produce this document, it returns {\tt \number\pdftexversion}.


\pdftexprimitive{\Syntax{\Tex{\quitvmode}}}
\bookmark{\tex{quitvmode}}

The primitive instructs \PDFTEX\ to quit vertical mode and start
typesetting a paragraph. \type{\quitvmode} has the same effect as
\type{\leavevmode} definition from \type{plain} macro package. Note
however, that \type{\leavevmode} may conflict with \type{\everypar}
tokens list. No such risk while using \type{\quitvmode} instead.
\introduced{1.21a}


\pdftexprimitive{\Syntax{\Tex{\vadjust}
  \Optional{\Something{pre spec}}
  \Something{filler}
  \Lbrace \Something{vertical mode material} \Rbrace
  \Modelist{h, m}
}}
\bookmark{\tex{vadjust}}

The \type{\vadjust} implementation of \PDFTEX\ adds an optional
qualifier \Something{pre spec}, which is simply the string \type{pre}, to
the original \TEX\ primitive with the same name. If
no \type{pre} is given, \type{\vadjust} behaves exactly as the original
(see the \TEX book, p.~281): it appends an adjustment item created
from \Something{vertical mode material} to the current list {\em after}
the line in which \type{\vadjust} appears. However, with the qualifier
\type{pre}, the adjustment item is put {\em before} the line in which
\type{\vadjust pre} appears.

%***********************************************************************

\section{Graphics}

\PDFTEX\ supports inclusion of pictures in \PNG, \JPEG, \JBIGTWO, and
\PDF\ format; a few differences between these are discussed below. The
most common technique with \TEX\ |<|the inclusion of \EPS\ figures|>|
is replaced by \PDF\ inclusion. \EPS\ files can be converted to \PDF\ by
\GHOSTSCRIPT, Adobe Distiller or other \POSTSCRIPT||to||\PDF\ converters.

The \PDF\ format is currently the most versatile source format for
graphics embedding. \PDFTEX\ allows to insert arbitrary pages from
\PDF\ files with their own fonts, graphics, and pixel images into
a document. The cover page of this manual is an example of such an insert,
being a one page document generated by \PDFTEX.

By default \PDFTEX\ takes the BoundingBox of a \PDF\ file from its CropBox
if available, otherwise from its MediaBox. This can be influenced by
the \Something{pdf box spec} option to the \type{\pdfximage} primitive,
or by setting the \type{\pdfpagebox} or \type{\pdfforcepagebox} primitives to
a value corresponding to the wanted box type.

To get the right BoundingBox from a \EPS\ file, before converting to \PDF,
it is necessary to transform the \EPS\ file so that the start point is
at the (0,0)~coordinate and the page size is set exactly corresponding
to the BoundingBox. A \PERL\ script (\EPSTOPDF) for this purpose has been
written. The \TEXUTIL\ utility script and the \PSTOPDF\ program that comes
with \GHOSTSCRIPT\ can so a similar job. (Concerning this conversion,
they can handle complete directories, remove some garbage from files,
takes precautions against duplicate conversion, etc.)

The lossless compressing \PNG\ format is great for embedding crisp pixel
graphics (\eg\ line scans, screen shots). Since \PDFTEX\ 1.30.0 also the
alpha-channel of \PNG\ images is processed if available; this allows
embedding of images with simple transparency. The \PNG\ format does not
support the CMYK color model, which is sometimes required for print media
(this often can be replaced by four component \JPEG\ in high quality or
lossless compression mode). Photos in \PNG\ format have a rather weak
compression; here the \JPEG\ format is preferable.

Embedding \PNG\ images in the general case requires \PDFTEX\ to uncompress
the pixel array and to re||compress it to the \PDF\ requirements; this
process often takes a noticeable amount of time. Since \PDFTEX\ 1.30.0
there is now a fast \PNG\ embedding mode that goes without uncompressing;
the image data are directly copied into the \PDF\ stream, resulting in
a much higher embedding speed. However this mode is only activated, if
the image array structure of the \PNG\ file is compatible with the \PDF\
image structure (\eg\ an interlaced \PNG\ image requires uncompressing to
re||arrange the image lines). Luckily it seems that the most common \PNG\
files also allow fast copying. The use of gamma correction disables fast
copying, as it requires calculations with individual pixels. Whether the
fast copy mode is used for a \PNG\ image can be seen from the log file,
which then shows the string `(PNG copy)' after the \PNG\ file name.

The \JPEG\ format is normally used in lossy mode; then it's ideal for
embedding photos.  It's not recommended for crisp images from synthetic
sources with a limited amount of colors.  Both \JFIF\ and \EXIF\ are
supported for additional information.

The \JBIGTWO\ format works only for bitonal (black and white) pixel
images like scanned line and text documents, but for these it has
typically a much higher compression ratio than the other two pixel
image formats. The \JBIGTWO\ format is part of the \PDF\ standard since
version 1.5; native \JBIGTWO\ image inclusion is available in \PDFTEX\
since version 1.40.0. A \JBIGTWO\ file might contain many images, which
gives an even better compression ratio than with a single image per file,
as \JBIGTWO\ encoders can exploit similarities between bit patterns over
several images. Encoders for \JBIGTWO\ can operate in lossy as well as
lossless modes. Only recently a free \JBIGTWO\ encoder has been written
and made available, see \from[jbig2enc].

Other options for graphics in \PDFTEX\ are:

\description {\LATEX\ picture mode} Since this is implemented simply in terms
of font characters, it works in exactly the same way as usual.

\description {Xy||pic} If the \POSTSCRIPT\ back||end is not requested, Xy||pic
uses its own Type~1 fonts, and needs no special attention.

\description {tpic} The \quote {tpic} \type{\special} commands (used in some
macro packages) can be redefined to produce literal \PDF, using some macros
written by Hans Hagen.

\description {\METAPOST} Although the output of \METAPOST\ is \POSTSCRIPT,
it is in a highly simplified form, and a \METAPOST\ to \PDF\ conversion
(\MPTOPDF, written by Hans Hagen and Tanmoy Bhattacharya) is implemented
as a set of macros which reads \METAPOST\ output and supports all of
its features.

For new work, the \METAPOST\ route is highly recommended. For the future,
Adobe has announced that they will define a specification for \quote
{encapsulated \PDF}.

The inclusion of raw \POSTSCRIPT\ commands |<|a technique utilized
by for instance the \type{pstricks} package|>| cannot directly be
supported. Although \PDF\ is direct a descendant of \POSTSCRIPT, it
lacks any programming language commands, and cannot deal with arbitrary
\POSTSCRIPT.

%***********************************************************************

\section{Character translation}

Characters that are input to \PDFTEX\ are subject to optional
\TEX\ character translation (\TCX) under control of a \TCX\ file.
The \TCX\ maps the input character codes (\eg\ from \type{\input} or
\type{\read}) to the character codes as seen by \PDFTEX.  This mapping
takes place before the characters enter \PDFTEX's `mouth'.  If no \TCX\
file is read, the input characters enter \PDFTEX\ directly; no mapping
is done.

\TCX\ files consist of lines each containing one or two integer numbers
in the range 0..255, either in decimal or hex notation.
% strtol() C-function
A comment sign \type{%} in a \TCX\ line starts a comment until the
end of line.  The first number in each line is for matching the input
character code, the second, optional number is the corresponding \TEX\
character code.  If a line contains only one number, characters with
this code enter \PDFTEX\ unchanged; no mapping is done.

\TCX\ mapping also influences \PDFTEX\ output streams for \type{\message} and
\type{\write}.  Without \TCX\ mapping, only characters that are within
the range 32..126 are flagged as `printable', meaning that these
characters are output directly by \type{\message} and \type{\write}
primitives.  Characters outside the range 32..126 are instead output
in escaped form, \eg\ as \type{^^A} for a character with code
\type{0x01}.  When a character code is mentioned in the 2nd column of the
\TCX\ file, or as the only value in a line, it is flagged as `printable'.
During \type{\message} and \type{\write}, output characters are mapped
in reverse direction: they are looked up in the 2nd column of the \TCX\
file and the corresponding values from the 1st column are output.  Again,
if a \PDFTEX\ character code is found as the only number in a line, no
mapping is done.  Mentioning a character code as the only number on
a line has the sole purpose to flag this code `printable'; remember that
character within the range 32..126 are `printable' anyway.

The characters output into the \PDF\ file, \eg\ by \type{\pdfliteral}
or \type{\special} primitives, are not subject to \TCX\ output remapping.

Beware: Character translation interferes with the \ENCTEX\ primitives; to
avoid surprises, don't use \ENCTEX\ and \TCX\ mapping at the same time.
Further details about \TCX\ file loading can be found in the \TETEX\
manual.

%***********************************************************************

\stopbodymatter

%D We did use some abbreviations. Only those really used will end up in the
%D following list.

\startbackmatter

\writebetweenlist[section]{\blank[line]}

%***********************************************************************

\section{Abbreviations}

In this document we use numerous abbreviations. For convenience we mention
their meaning here.

\placelistofabbreviations

%***********************************************************************

\start \setupalign[nothanging,nohz]

\section{Examples of HZ and protruding}

In the following sections we will demonstrate \PDFTEX's protruding and
\HZ\ features, using a text from E.~Tufte. This sample text has a lot
of punctuation and often needs hyphenation. Former \PDFTEX\ versions
had sometimes problems with combining these features, but from version
1.21a on it should be ok. If you still encounter problems, please try
to prepare a small test file that demonstrates the problem and send it
to one of the maintainers.

\startbuffer
    \input tufte
    \blank[big]
    \startcolumns
        \input tufte
    \stopcolumns
\stopbuffer

\subsection{Normal}     \start                         \getbuffer \stop
\subsection{HZ}         \start \setupalign[hz]         \getbuffer \stop
\subsection{Protruding} \start \setupalign[hanging]    \getbuffer \stop
\subsection{Both}       \start \setupalign[hz,hanging] \getbuffer \stop

\stop

%***********************************************************************

\section[sec.addpdfkeys]{Additional \PDF\ keys}

{\em This section is based on the manual on keys written by Martin
Schr\"oder, one of the maintainers of \PDFTEX.}

A \PDF\ document should contain only the structures and attributes defined
in the \PDF\ specification. However, the specification allows applications
to insert additional keys, provided they follow certain rules.

The most important rule is that developers have to register with Adobe
prefixes for the keys they want to insert.  Hans Hagen has registered the
prefix \type{PTEX} for \PDFTEX.

\PDFTEX\ generates an XObject for every included \PDF. The dictionary of
this object contains these additional keys:

\starttabulate[|lT|l|p|]
\HL
\NC \bf key         \NC \bf type   \NC meaning \NC \NR
\HL
\NC PTEX.FileName   \NC string     \NC The name of the included file as seen by
                                       \PDFTEX. \NC \NR
\NC PTEX.InfoDict   \NC dictionary \NC The document information dictionary of the included
                                       \PDF\ (an indirect object). \NC \NR
\NC PTEX.PageNumber \NC integer    \NC The page number of the included file. \NC \NR
\HL
\stoptabulate

The \PDFReference\ says: \quotation {Although viewer applications can
store custom metadata in the document information dictionary, it is
inappropriate to store private content or structural information there;
such information should be stored in the document catalog instead.}

Although it would seem more natural to put this information in the
document information dictionary, we have to obey the rules laid down in
the \PDFReference. The following key ends up in the document catalog.

\starttabulate[|lT|l|p|]
\HL
\NC \bf key         \NC \bf type   \NC meaning \NC \NR
\HL
\NC PTEX.Fullbanner \NC string     \NC The full version of the \pt binary that
produced the file as displayed by {\tt pdftex \hbox{-{}-version}}, \eg\
{\tt\pdftexbanner}.  This is necessary because the string in the
\type{Producer} key in the info dictionary is rather short,
namely {\tt pdfTeX-\currentpdftex}. \NC \NR
\HL
\stoptabulate

Any or all of these keys can be suppressed with the
\type{\pdfsuppressptexinfo} primitive, described in
\in{section}[sec.docinfocatalog].

%***********************************************************************

\section{Colophon}

This manual is typeset in \CONTEXT. One can generate an A4 version from
the source code by typing:

\starttyping
texexec --result=pdftex-a.pdf pdftex-t
\stoptyping

Or in letter size:

\starttyping
texexec  --mode=letter --result=pdftex-l.pdf pdftex-t
\stoptyping

Given that the A4 version is typeset, one can generate an A5 booklet by
typing:

\starttyping
texexec --pdfarrange --paper=a5a4 --print=up --addempty=1,2
  --result=pdftex-b.pdf pdftex-a
\stoptyping

Odd and even page sets for non-duplex printers can be generated using
\type{-}\type{-pages=odd} and \type{-}\type{-pages=even} options
(which might require some disciplined shuffling of sheet).

This also demonstrates that \PDFTEX\ can be used for page imposition
purposes (given that \PDFTEX\ and the fonts are set up properly).

%***********************************************************************

\page

\definehead[gnusection][subsection]
\definehead[gnusubject][subsubject]

\setuphead[gnusection,gnusubject][style=\bfa,before=\blank,after=\blank,align={right,broad,nothyphenated}]

\setuphead[gnusection][ownnumber=yes]

\section{GNU Free Documentation License}

\startnotmode[screen]
    \switchtobodyfont[4.6pt] % squeeze everything on one page :-}
    \setuplayout[grid=yes]
    \setupcolumns[n=7]
\stopnotmode

\startmode[screen]
    \setupcolumns[n=2]
\stopmode

\startcolumns[tolerance=verytolerant,distance=10pt] \widowpenalty10000 \clubpenalty10000

\startlines
Version 1.2, November 2002
Copyright \copyright\ 2000, 2001, 2002
Free Software Foundation, Inc.
59 Temple Place, Suite 330,
Boston, MA  02111-1307  USA
\stoplines

Everyone is permitted to copy and distribute verbatim copies
of this license document, but changing it is not allowed.

\gnusubject{Preamble}

The purpose of this License is to make a manual, textbook, or other
functional and useful document ``free'' in the sense of freedom: to
assure everyone the effective freedom to copy and redistribute it,
with or without modifying it, either commercially or noncommercially.
Secondarily, this License preserves for the author and publisher a way
to get credit for their work, while not being considered responsible
for modifications made by others.

This License is a kind of ``copyleft'', which means that derivative
works of the document must themselves be free in the same sense.  It
complements the GNU General Public License, which is a copyleft
license designed for free software.

We have designed this License in order to use it for manuals for free
software, because free software needs free documentation: a free
program should come with manuals providing the same freedoms that the
software does.  But this License is not limited to software manuals;
it can be used for any textual work, regardless of subject matter or
whether it is published as a printed book.  We recommend this License
principally for works whose purpose is instruction or reference.

\gnusection{1}{APPLICABILITY AND DEFINITIONS}

This License applies to any manual or other work, in any medium, that
contains a notice placed by the copyright holder saying it can be
distributed under the terms of this License.  Such a notice grants
a world-wide, royalty-free license, unlimited in duration, to use that
work under the conditions stated herein.  The {\bf ``Document''}, below,
refers to any such manual or work.  Any member of the public is
a licensee, and is addressed as {\bf ``you''}.  You accept the license if you
copy, modify or distribute the work in a way requiring permission
under copyright law.

A {\bf ``Modified Version''} of the Document means any work containing the
Document or a portion of it, either copied verbatim, or with
modifications and/or translated into another language.

A {\bf ``Secondary Section''} is a named appendix or a front-matter section of
the Document that deals exclusively with the relationship of the
publishers or authors of the Document to the Document's overall subject
(or to related matters) and contains nothing that could fall directly
within that overall subject.  (Thus, if the Document is in part
a textbook of mathematics, a Secondary Section may not explain any
mathematics.)  The relationship could be a matter of historical
connection with the subject or with related matters, or of legal,
commercial, philosophical, ethical or political position regarding
them.

The {\bf ``Invariant Sections''} are certain Secondary Sections whose titles
are designated, as being those of Invariant Sections, in the notice
that says that the Document is released under this License.  If
a section does not fit the above definition of Secondary then it is not
allowed to be designated as Invariant.  The Document may contain zero
Invariant Sections.  If the Document does not identify any Invariant
Sections then there are none.

The {\bf ``Cover Texts''} are certain short passages of text that are listed,
as Front-Cover Texts or Back-Cover Texts, in the notice that says that
the Document is released under this License. A Front-Cover Text may
be at most 5 words, and a Back-Cover Text may be at most 25 words.

A {\bf ``Transparent''} copy of the Document means a machine-readable copy,
represented in a format whose specification is available to the
general public, that is suitable for revising the document
straightforwardly with generic text editors or (for images composed of
pixels) generic paint programs or (for drawings) some widely available
drawing editor, and that is suitable for input to text formatters or
for automatic translation to a variety of formats suitable for input
to text formatters. A copy made in an otherwise Transparent file
format whose markup, or absence of markup, has been arranged to thwart
or discourage subsequent modification by \PDF\ viewers is not Transparent.
An image format is not Transparent if used for any substantial amount
of text. A copy that is not ``Transparent'' is called {\bf ``Opaque''}.

Examples of suitable formats for Transparent copies include plain
ASCII without markup, Texinfo input format, LaTeX input format, SGML
or XML using a publicly available DTD, and standard-conforming simple
HTML, \POSTSCRIPT\ or \PDF\ designed for human modification.  Examples of
transparent image formats include PNG, XCF and JPG.  Opaque formats
include proprietary formats that can be read and edited only by
proprietary word processors, SGML or XML for which the DTD and/or
processing tools are not generally available, and the
machine-generated HTML, \POSTSCRIPT\ or \PDF\ produced by some word
processors for output purposes only.

The {\bf ''Title Page''} means, for a printed book, the title page itself,
plus such following pages as are needed to hold, legibly, the material
this License requires to appear in the title page.  For works in
formats which do not have any title page as such, ``Title Page'' means
the text near the most prominent appearance of the work's title,
preceding the beginning of the body of the text.

A section {\bf ``Entitled XYZ''} means a named subunit of the Document whose
title either is precisely XYZ or contains XYZ in parentheses following
text that translates XYZ in another language.  (Here XYZ stands for
a specific section name mentioned below, such as {\bf ``Acknowledgements''},
{\bf ``Dedications''}, {\bf ``Endorsements''}, or {\bf ``History''}.)
To {\bf ``Preserve the Title''}
of such a section when you modify the Document means that it remains
a section ``Entitled XYZ'' according to this definition.

The Document may include Warranty Disclaimers next to the notice which
states that this License applies to the Document.  These Warranty
Disclaimers are considered to be included by reference in this
License, but only as regards disclaiming warranties: any other
implication that these Warranty Disclaimers may have is void and has
no effect on the meaning of this License.


\gnusection{2}{VERBATIM COPYING}

You may copy and distribute the Document in any medium, either
commercially or noncommercially, provided that this License, the
copyright notices, and the license notice saying this License applies
to the Document are reproduced in all copies, and that you add no other
conditions whatsoever to those of this License.  You may not use
technical measures to obstruct or control the reading or further
copying of the copies you make or distribute.  However, you may accept
compensation in exchange for copies.  If you distribute a large enough
number of copies you must also follow the conditions in section 3.

You may also lend copies, under the same conditions stated above, and
you may publicly display copies.


\gnusection{3}{COPYING IN QUANTITY}

If you publish printed copies (or copies in media that commonly have
printed covers) of the Document, numbering more than 100, and the
Document's license notice requires Cover Texts, you must enclose the
copies in covers that carry, clearly and legibly, all these Cover
Texts: Front-Cover Texts on the front cover, and Back-Cover Texts on
the back cover.  Both covers must also clearly and legibly identify
you as the publisher of these copies.  The front cover must present
the full title with all words of the title equally prominent and
visible.  You may add other material on the covers in addition.
Copying with changes limited to the covers, as long as they preserve
the title of the Document and satisfy these conditions, can be treated
as verbatim copying in other respects.

If the required texts for either cover are too voluminous to fit
legibly, you should put the first ones listed (as many as fit
reasonably) on the actual cover, and continue the rest onto adjacent
pages.

If you publish or distribute Opaque copies of the Document numbering
more than 100, you must either include a machine-readable Transparent
copy along with each Opaque copy, or state in or with each Opaque copy
a computer-network location from which the general network-using
public has access to download using public-standard network protocols
a complete Transparent copy of the Document, free of added material.
If you use the latter option, you must take reasonably prudent steps,
when you begin distribution of Opaque copies in quantity, to ensure
that this Transparent copy will remain thus accessible at the stated
location until at least one year after the last time you distribute an
Opaque copy (directly or through your agents or retailers) of that
edition to the public.

It is requested, but not required, that you contact the authors of the
Document well before redistributing any large number of copies, to give
them a chance to provide you with an updated version of the Document.


\gnusection{4}{MODIFICATIONS}

You may copy and distribute a Modified Version of the Document under
the conditions of sections 2 and 3 above, provided that you release
the Modified Version under precisely this License, with the Modified
Version filling the role of the Document, thus licensing distribution
and modification of the Modified Version to whoever possesses a copy
of it.  In addition, you must do these things in the Modified Version:

\startitemize[A,packed]
\item
   Use in the Title Page (and on the covers, if any) a title distinct
   from that of the Document, and from those of previous versions
   (which should, if there were any, be listed in the History section
   of the Document).  You may use the same title as a previous version
   if the original publisher of that version gives permission.

\item
   List on the Title Page, as authors, one or more persons or entities
   responsible for authorship of the modifications in the Modified
   Version, together with at least five of the principal authors of the
   Document (all of its principal authors, if it has fewer than five),
   unless they release you from this requirement.

\item
   State on the Title page the name of the publisher of the
   Modified Version, as the publisher.

\item
   Preserve all the copyright notices of the Document.

\item
   Add an appropriate copyright notice for your modifications
   adjacent to the other copyright notices.

\item
   Include, immediately after the copyright notices, a license notice
   giving the public permission to use the Modified Version under the
   terms of this License, in the form shown in the Addendum below.

\item
   Preserve in that license notice the full lists of Invariant Sections
   and required Cover Texts given in the Document's license notice.

\item
   Include an unaltered copy of this License.

\item
   Preserve the section Entitled ``History'', Preserve its Title, and add
   to it an item stating at least the title, year, new authors, and
   publisher of the Modified Version as given on the Title Page.  If
   there is no section Entitled ``History'' in the Document, create one
   stating the title, year, authors, and publisher of the Document as
   given on its Title Page, then add an item describing the Modified
   Version as stated in the previous sentence.

\item
   Preserve the network location, if any, given in the Document for
   public access to a Transparent copy of the Document, and likewise
   the network locations given in the Document for previous versions
   it was based on.  These may be placed in the ``History'' section.
   You may omit a network location for a work that was published at
   least four years before the Document itself, or if the original
   publisher of the version it refers to gives permission.

\item
   For any section Entitled ``Acknowledgements'' or ``Dedications'',
   Preserve the Title of the section, and preserve in the section all
   the substance and tone of each of the contributor acknowledgements
   and/or dedications given therein.

\item
   Preserve all the Invariant Sections of the Document,
   unaltered in their text and in their titles.  Section numbers
   or the equivalent are not considered part of the section titles.

\item
   Delete any section Entitled ``Endorsements''.  Such a section
   may not be included in the Modified Version.

\item
   Do not retitle any existing section to be Entitled ``Endorsements''
   or to conflict in title with any Invariant Section.

\item
   Preserve any Warranty Disclaimers.
\stopitemize

If the Modified Version includes new front-matter sections or
appendices that qualify as Secondary Sections and contain no material
copied from the Document, you may at your option designate some or all
of these sections as invariant.  To do this, add their titles to the
list of Invariant Sections in the Modified Version's license notice.
These titles must be distinct from any other section titles.

You may add a section Entitled ``Endorsements'', provided it contains
nothing but endorsements of your Modified Version by various
parties---for example, statements of peer review or that the text has
been approved by an organization as the authoritative definition of
a standard.

You may add a passage of up to five words as a Front-Cover Text, and
a passage of up to 25 words as a Back-Cover Text, to the end of the list
of Cover Texts in the Modified Version.  Only one passage of
Front-Cover Text and one of Back-Cover Text may be added by (or
through arrangements made by) any one entity.  If the Document already
includes a cover text for the same cover, previously added by you or
by arrangement made by the same entity you are acting on behalf of,
you may not add another; but you may replace the old one, on explicit
permission from the previous publisher that added the old one.

The author(s) and publisher(s) of the Document do not by this License
give permission to use their names for publicity for or to assert or
imply endorsement of any Modified Version.

\gnusection{5}{COMBINING DOCUMENTS}

You may combine the Document with other documents released under this
License, under the terms defined in section 4 above for modified
versions, provided that you include in the combination all of the
Invariant Sections of all of the original documents, unmodified, and
list them all as Invariant Sections of your combined work in its
license notice, and that you preserve all their Warranty Disclaimers.

The combined work need only contain one copy of this License, and
multiple identical Invariant Sections may be replaced with a single
copy.  If there are multiple Invariant Sections with the same name but
different contents, make the title of each such section unique by
adding at the end of it, in parentheses, the name of the original
author or publisher of that section if known, or else a unique number.
Make the same adjustment to the section titles in the list of
Invariant Sections in the license notice of the combined work.

In the combination, you must combine any sections Entitled ``History''
in the various original documents, forming one section Entitled
``History''; likewise combine any sections Entitled ``Acknowledgements'',
and any sections Entitled ``Dedications''.  You must delete all sections
Entitled ``Endorsements''.

\gnusection{6}{COLLECTIONS OF DOCUMENTS}

You may make a collection consisting of the Document and other documents
released under this License, and replace the individual copies of this
License in the various documents with a single copy that is included in
the collection, provided that you follow the rules of this License for
verbatim copying of each of the documents in all other respects.

You may extract a single document from such a collection, and distribute
it individually under this License, provided you insert a copy of this
License into the extracted document, and follow this License in all
other respects regarding verbatim copying of that document.

\gnusection{7}{AGGREGATION WITH INDEPENDENT WORKS}

A compilation of the Document or its derivatives with other separate
and independent documents or works, in or on a volume of a storage or
distribution medium, is called an ``aggregate'' if the copyright
resulting from the compilation is not used to limit the legal rights
of the compilation's users beyond what the individual works permit.
When the Document is included in an aggregate, this License does not
apply to the other works in the aggregate which are not themselves
derivative works of the Document.

If the Cover Text requirement of section 3 is applicable to these
copies of the Document, then if the Document is less than one half of
the entire aggregate, the Document's Cover Texts may be placed on
covers that bracket the Document within the aggregate, or the
electronic equivalent of covers if the Document is in electronic form.
Otherwise they must appear on printed covers that bracket the whole
aggregate.

\gnusection{8}{TRANSLATION}

Translation is considered a kind of modification, so you may
distribute translations of the Document under the terms of section 4.
Replacing Invariant Sections with translations requires special
permission from their copyright holders, but you may include
translations of some or all Invariant Sections in addition to the
original versions of these Invariant Sections.  You may include
a translation of this License, and all the license notices in the
Document, and any Warranty Disclaimers, provided that you also include
the original English version of this License and the original versions
of those notices and disclaimers.  In case of a disagreement between
the translation and the original version of this License or a notice
or disclaimer, the original version will prevail.

If a section in the Document is Entitled ``Acknowledgements'',
``Dedications'', or ``History'', the requirement (section 4) to Preserve
its Title (section~1) will typically require changing the actual
title.

\gnusection{9}{TERMINATION}

You may not copy, modify, sublicense, or distribute the Document except
as expressly provided for under this License.  Any other attempt to
copy, modify, sublicense or distribute the Document is void, and will
automatically terminate your rights under this License.  However,
parties who have received copies, or rights, from you under this
License will not have their licenses terminated so long as such
parties remain in full compliance.

\gnusection{10}{FUTURE REVISIONS OF THIS LICENSE}

The Free Software Foundation may publish new, revised versions
of the GNU Free Documentation License from time to time.  Such new
versions will be similar in spirit to the present version, but may
differ in detail to address new problems or concerns.  See
http:/\!/www.gnu.org/copyleft/.

Each version of the License is given a distinguishing version number.
If the Document specifies that a particular numbered version of this
License ``or any later version'' applies to it, you have the option of
following the terms and conditions either of that specified version or
of any later version that has been published (not as a draft) by the
Free Software Foundation.  If the Document does not specify a version
number of this License, you may choose any version ever published (not
as a draft) by the Free Software Foundation.
\stopcolumns

%***********************************************************************

\stopbackmatter

%D And then we're done.

% Evidently ConTeXt somewhere sets \pdfgentounicode=1.  Or something.
% So insert the mappings.  Seems like a good thing anyway.
\input glyphtounicode
\stoptext