diff options
| -rw-r--r-- | Master/texmf-dist/doc/dvipdfmx/dvipdfmx.pdf | bin | 183084 -> 259383 bytes | |||
| -rw-r--r-- | Master/texmf-dist/doc/dvipdfmx/dvipdfmx.tex | 1739 |
2 files changed, 1336 insertions, 403 deletions
diff --git a/Master/texmf-dist/doc/dvipdfmx/dvipdfmx.pdf b/Master/texmf-dist/doc/dvipdfmx/dvipdfmx.pdf Binary files differindex 3a9ac317fa3..b3e282a32c6 100644 --- a/Master/texmf-dist/doc/dvipdfmx/dvipdfmx.pdf +++ b/Master/texmf-dist/doc/dvipdfmx/dvipdfmx.pdf diff --git a/Master/texmf-dist/doc/dvipdfmx/dvipdfmx.tex b/Master/texmf-dist/doc/dvipdfmx/dvipdfmx.tex index df6c7cabe9e..e20b81cff63 100644 --- a/Master/texmf-dist/doc/dvipdfmx/dvipdfmx.tex +++ b/Master/texmf-dist/doc/dvipdfmx/dvipdfmx.tex @@ -1,13 +1,16 @@ -\documentclass[a4paper,xetex]{article} +\documentclass[a4paper,xetex,oneside]{book} \usepackage{xltxtra} \usepackage{fontspec} \usepackage{microtype} +\usepackage{xunicode} +\usepackage{unicode-math} %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% PLATFORM DEPENDENT INSTRUCTION HERE %% Font setup for body text. \setmainfont{Constantia} \setsansfont{Cambria} \setmonofont[Scale=0.92]{Consolas}% Adjust xheight difference +\setmathfont{Cambria Math} %% The following section is for showing some fancy examples. %% Some fonts used here may not be availabel on your system. %% Please modify this. Just replacing with empty macros is OK. @@ -16,20 +19,20 @@ \newcommand{\jpninezeroexamples}{{% \setmainfont[Scale=5,RawFeature=+jp90]{SourceHanSerifJP-Light.otf}葛祇逢}} %% We use PostScript raw code here to test dvipdfmx's capability. -%% TFM files `rml' and `rmlv' are distributed with ASCII pTeX. +%% TFM files `uprml' and `uprmlv' are distributed with upTeX. %% TrueType font `yumindb.ttf' is bundled with Windows 10. \newcommand{\jphoritext}{% \makebox[112bp][l]{% \raisebox{88bp}[112bp][0bp]{% -\special{pdf:mapline rml UniJIS-UTF8-H yumindb.ttf} -\special{ps: rml findfont 16 scalefont setfont +\special{pdf:mapline uprml UniJIS-UTF8-H yumindb.ttf} +\special{ps: uprml findfont 16 scalefont setfont currentpoint moveto (「こんにちは」) show}% }}} \newcommand{\jpverttext}{% \makebox[16bp][l]{% \raisebox{112bp}[112bp][0bp]{% -\special{pdf:mapline rmlv UniJIS-UTF8-V yumindb.ttf} -\special{ps: rmlv findfont 16 scalefont setfont +\special{pdf:mapline uprmlv UniJIS-UTF8-V yumindb.ttf} +\special{ps: uprmlv findfont 16 scalefont setfont currentpoint moveto (「こんにちは」) show}% }}} %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% @@ -39,6 +42,7 @@ colorlinks=true,% urlcolor=[rgb]{0.2,0.2,0.4}, linkcolor=[rgb]{0.2,0.2,0.4}, + citecolor=[rgb]{0.2,0.2,0.4}, pdftitle={The Dvipdfmx User's Manual},% pdfauthor={The dvipdfmx project team},% pdfsubject={A User's Manual for Dvipdfmx and Xdvipdfmx.},% @@ -67,34 +71,49 @@ belowskip=2em } \usepackage{mflogo} +\usepackage{lipsum} +\usepackage{array} \usepackage{marginnote} \renewcommand*{\marginfont}{\footnotesize\itshape} \usepackage[noorphans,font=itshape]{quoting} -\newcommand{\package}[1]{{\itshape #1}} -\newcommand{\code}[1]{\texttt{#1}} -\newcommand{\option}[1]{`\texttt{#1}'} +\newcommand{\package}[1]{\textit{#1}} +\newcommand{\code}[1]{\mbox{\texttt{#1}}} +\newcommand{\keyword}[1]{\textit{#1}} +\newcommand{\option}[1]{\mbox{`\texttt{#1}'}} \newcommand{\dvipdfm}{\texttt{dvipdfm}} \newcommand{\dvipdfmx}{\texttt{dvipdfmx}} \newcommand{\xdvipdfmx}{\texttt{xdvipdfmx}} \newcommand{\deprecated}[1]{\marginnote{\addfontfeatures{Color=CC3333}#1}} +\newcommand{\newfeature}[1]{\marginnote{\addfontfeatures{Color=3366CC}#1}} \newcommand{\lnum}[1]{{\addfontfeatures{RawFeature=+lnum}#1}} +% For placeing drawings via \special +\newcommand{\specialbox}[3]{% +\makebox[#1][l]{\raisebox{-#2}[0bp][#2]{\special{#3}}}} \usepackage{fancyhdr} \pagestyle{fancy} -%\lhead{} -\title{The \dvipdfmx\ User's Manual} -\author{The \dvipdfmx\ project team\footnote{Jin-Hwan Cho, Matthias Franz, -and \href{mailto:shunsaku.hirata74@gmail.com}{Shunsaku Hirata}}} -\date{May 6, 2017} +%%DVIPDFMX SETTINGS +\AtBeginDocument{% + \special{dvipdfmx:config O 1}% + \special{dvipdfmx:config V 7} +} \begin{document} -\maketitle -%\tableofcontents +\begin{titlepage} + \begin{raggedleft} + {\Huge\bfseries The Dvipdfmx User's Manual}\\[\baselineskip] + \Large The Dvipdfmx Project Team\\ + August 27, 2018\par + \end{raggedleft} +\end{titlepage} + +\tableofcontents + +\chapter{Getting Started} \section{Introduction} -The \dvipdfmx\ (formerly \dvipdfm-cjk) project -provides an extended version of the \dvipdfm, a DVI to PDF translator developed -by Mark~A.~Wicks. +The \dvipdfmx\ (formerly \dvipdfm-cjk) project provides an extended version of +the \dvipdfm, a DVI to PDF translator developed by Mark~A.~Wicks. The primary goal of this project is to support multi-byte character encodings and large character sets such as for East Asian languages. @@ -104,7 +123,7 @@ Shunsaku Hirata and its modified one, \dvipdfm-kor, by Jin-Hwan Cho. Extensions to \dvipdfm\ include, \begin{itemize} \item Support for OpenType and TrueType font, including partial support - for OpenType Layout for finding glyphs and vertical writing. + for the OpenType Layout for finding glyph variants and vertical writing. \item Support for CJK-\LaTeX\ and H\LaTeX\ with Subfont Definition Files. \item Support for various legacy multi-byte encodings via PostScript CMap Resources. @@ -130,17 +149,17 @@ Some important features are still missing: \end{itemize} \dvipdfmx\ is now maintained as part of \TeX\ Live. Latest source code can -be found at \TeX\ Live SVN repository. For an instruction on accessing the -development sources for TeX Live, see,\medskip +be found at the \TeX\ Live SVN repository. For an instruction on accessing the +development sources for \TeX\ Live, see,\medskip \url{http://www.tug.org/texlive/svn/} \medskip -This manual, ``The \dvipdfmx\ User's Manual'', is originally prepared for -\TeX\ Live 2017. Current maintainer of this manual is Shunsaku Hirata. +This document, ``The \dvipdfmx\ User's Manual'', was originally prepared for +\TeX\ Live 2017. Current maintainer of this document is Shunsaku Hirata. Latest version and contact information can be found at:\medskip -\url{http://github.com/shirat74/dvipdfm-x-doc/} +\url{http://github.com/shirat74/dvipdfm-x-doc} \medskip \noindent{}Please send quenstions or suggestions. @@ -151,7 +170,7 @@ Latest version and contact information can be found at:\medskip \dvipdfmx. The \xdvipdfmx\ extensions provides support for the Extended DVI (.xdv) format -generated by \XeTeX which includes support for platform-native fonts and the +generated by \XeTeX\ which includes support for platform-native fonts and the \XeTeX\ graphics primitives, as well as Unicode text and OpenType font. \XeTeX\ originally used a Mac-specific program called \code{xdv2pdf} as a @@ -183,17 +202,19 @@ available from its CTAN site:\medskip \url{http://www.ctan.org/tex-archive/dviware/dvipdfm} \medskip -The minimal requirements for building \dvipdfmx\ is the \code{kpathsea} library. -\code{zlib} for compression and \code{libpng} for PNG inclusion are highly -recommended. Optionally, the libpaper library may be used to handle paper sizes. +The minimal requirements for building \dvipdfmx\ is the \keyword{kpathsea} library. +the \keyword{zlib} library for compression and the \keyword{libpng} library for PNG +inclusion are highly recommended. +Optionally, the \keyword{libpaper} library might be used to handle paper size. -This document only describes additions and modifications to \dvipdfm. +This document mainly focuses on the additions and modifications to \dvipdfm. Please refer the ``\href{http://mirrors.ctan.org/dviware/dvipdfm/dvipdfm.pdf}{Dvipdfm User's Manual}'' available from the CTAN site mentioned above for basic usage. Some additional command line options recognized by \dvipdfmx\ are listed in -Table~\ref{TABLE:options}. Try +Table~\ref{TABLE:options}. In addition to this, the \code{-V} option for specifying +the output PDF version now accepts the verion specification of a form \code{2.0}. Try \begin{lstlisting} dvipdfmx --help \end{lstlisting} @@ -221,6 +242,7 @@ for the list of command line options and thier explanations. \code{-M} & Process \MP\ generated PostScript file.\\ \code{-E} & Always try to embed fonts \emph{regardless of liscensing}.\\ + \code{-O} \textit{number} & Set maximum depth of open bookmark item.\\ \hline \end{tabular} \caption{Additional command line options recognized by \dvipdfmx.}% @@ -236,19 +258,22 @@ p\TeX, and up\TeX. This section is intended to be a quick guide for each users. \subsection{\texorpdfstring{\XeTeX}{XeTeX}} -For \XeTeX\ users, most part of this document is irrelevant except section of -``\hyperref[SEC:graphics]{Graphics and Image Formats}'' and -``\hyperref[SEC:specials]{DVI Specials}''. + +\XeTeX\ users +normally do not invoke the \dvipdfmx\ command directly. To contorol the +behavior of \dvipdfmx, please consider using the \code{dvipdfmx:config} +special explained in the section of ``\hyperref[SEC:specials]{Specials}.'' +Some part of this document is irrelevant for \XeTeX\ users. \subsection{p\TeX} p\TeX\ users are at least required to install several auxilially files mentioned in the section of ``\hyperref[SEC:auxfiles]{Auxilially Files}'' -and to setup fontmappings. Just install the \package{adobemappings} and +and to setup font-mappings. Just install the \package{adobemappings} and \package{glyphlist} for auxilially files. (As \TeX\ Live basic intallation requires them, they are probably already installed for \TeX\ Live users.) Setting up fontmaps can be done easily with help of the \package{ptex-fontmaps} -package. For examples, to use with IPA fonts (contained +package. For example, to use with the IPA fonts (contained in the \package{ipaex} package), run, \begin{lstlisting} updmap ipaex @@ -270,7 +295,7 @@ the following \code{special} command, \AtBeginDocument{\special{pdf:tounicode 90ms-RKSJ-UCS2}} \end{lstlisting} in the preamble. -The above \code{special} command instructs \dvipdfmx\ to convert text encoded +The above \code{special} command instruct let \dvipdfmx\ to convert text encoded in Shift-JIS to Unicode. For EUC-JP, replace 90ms-RKJK-UCS2 with EUC-UCS2. \subsection{up\TeX} @@ -278,9 +303,9 @@ up\TeX\ users are basically the same as p\TeX\ users but there are two choices for setting fontmaps. Setup fontmaps as mentioned in the above for p\TeX, or use keyword \code{unicode} in the encoding field of fontmap files. -The former case might be easier as auto-creation of fontmap files can be done +The former might be easier as the auto-creation of fontmap files can be done with the \code{updmap} program and the \package{ptex-fontmaps} package. But in -this method there are some difficulties when using fonts which employ character +this method there are some difficulties when using a font which employs a character collection (glyph repertoire) other than Adobe-Japan1 in the case of PostScript flvavored OpenType fonts. In the later case, the \package{adobemappings} package is not reuiqred @@ -290,7 +315,8 @@ can be used too. Using \code{unicode} is more simpler and intuitive thus it is recommended to use this method.\footnote{For \TeX\ Live 2017. Earlier versions have buggy support.} -Typical example fontmap entries for using Adobe's SouceHan fonts look like: +A typical example fontmap entries for using Adobe's SouceHan series +might look like: \begin{lstlisting} urml unicode SourceHanSerifJP-Light.otf urmlv unicode SourceHanSerifJP-Light.otf -w 1 @@ -298,8 +324,8 @@ ugbm unicode SourceHanSansJP-Medium.otf ugbmv unicode SourceHanSansJP-Medium.otf -w 1 \end{lstlisting} -As in p\TeX, the following \code{special} instruction is necessary to PDF -document information and annotations to be shown correctly: +As in p\TeX, the following \code{special} instruction might be necessary for PDF +document information and annotations to be shown correctly: \begin{lstlisting} \AtBeginDocument{\special{pdf:tounicode UTF8-UCS2}} \end{lstlisting} @@ -307,8 +333,8 @@ Here, input encodig is assumed to be UTF-8. \subsection{CJK-\LaTeX} -CJK-\LaTeX users are required to have Subfont Definition Files to be installed. -They are available as part of the \package{ttfutils} package. +CJK-\LaTeX\ users are required to have \keyword{Subfont Definition Files} +to be installed. They are available as part of the \package{ttfutils} package. To use TrueType Ariphic fonts provided by the \package{arphic-ttf} package: \begin{lstlisting} @@ -325,22 +351,24 @@ To use TrueType Ariphic fonts provided by the \package{arphic-ttf} package: \end{CJK} \end{document} \end{lstlisting} -Here, \code{pdf:mapline} special is used to setup fontmappings. +Here, \code{pdf:mapline} special is used to setup a font-mapping. \section{Auxiliary Files}\label{SEC:auxfiles} This section is mostly for supporting legacy encodings and legacy font format such as PostScript Type1 font. \XeTeX\ users may skip this section. -\dvipdfmx\ has a capability to handle various input encodings from 7-bit +\dvipdfmx\ has a capacity to handle various input encodings from 7-bit encodings to variable-width multi-byte encodings. It also has some sort of Unicode support. Several auxiliary files which are not common to \TeX\ users are needed to enable those features. This section shortly describes about them. \subsection{PostScript CMap Resources} -PostScript CMap Resources\footnote{See, ``\href{http://www.adobe.com/content/dam/Adobe/en/devnet/font/pdfs/5014.CIDFont\_Spec.pdf}{Adobe CMap and CIDFont Files Specification}''} -are required for supporting lagacy encodings such as Shift-JIS, EUC-JP, Big5, +\keyword{PostScript CMap Resources}\footnote{See, +``\href{http://www.adobe.com/content/dam/Adobe/en/devnet/font/pdfs/5014.CIDFont\_Spec.pdf}% +{Adobe CMap and CIDFont Files Specification}''} +are required for supporting lagacy encodings such as Shift-JIS, EUC-JP, \lnum{Big5}, and other East Asian encodings. \dvipdfmx\ internally identifies glyphs with identifiers (CIDs\footnote{PostScript terminology ``Character IDentifier''.}) represented as an integer ranging from 0 to 65535 in the CID-based glyph access. @@ -363,9 +391,9 @@ want to install them but they are not required. CJK fonts usually contain several thousands of glyphs. For using such fonts with (original) \TeX, which can only handle 8-bit encodings, it is necessary to -split fonts into several ``subfonts''. Subfont Definition File (SFD) +split a font into several ``subfonts''. Subfont Definition File (SFD) specify the way how those fonts are split into subfonts. \dvipdfmx\ uses SFD -files to convert subfonts back to a single font. +files to convert a set of subfonts back to a single font. SFD files are not required for use with \TeX\ variants which can handle multi-byte character encodings and large character sets such as p\TeX, @@ -376,8 +404,9 @@ for \TeX\ Live users. \subsection{The Adobe Glyph List and ToUnicode Mappings} -The Adobe Glyph List\footnote{See, ``\href{http://github.com/adobe-type-tools/agl-specification}{Adobe Glyph List Specification}''} (AGL) -describes correspondence between PostScript glyph names (e.g., \code{AE}, +The Adobe Glyph List\footnote{See, +``\href{http://github.com/adobe-type-tools/agl-specification}{Adobe Glyph List Specification}''} +(AGL) describes correspondence between PostScript glyph names (e.g., \code{AE}, \code{Aacute},...) and Unicode character sequences representing them. Some features described in the section ``Unicode Support'' requires AGL file. @@ -389,7 +418,7 @@ fonts when the font itself does not provide a mapping from PostScript glyph names to glyph indices (version 2.0 ``post'' table), and when the encoding \code{unicode} is specified for Type1 font. -AGL file is included in the \package{glyphlist} package. The latest version +The AGL file is included in the \package{glyphlist} package. The latest version can be found at Adobe's GitHub site:\medskip \url{http://github.com/adobe-type-tools/agl-aglfn} @@ -398,10 +427,10 @@ can be found at Adobe's GitHub site:\medskip ToUnicode Mappings are similar to AGL but they describe correspondence between CID numbers (instead of glyph names) and Unicode values. The content of those files are the same as CMap Resources. -They are required when using TrueType fonts emulated as CID-keyed fonts. +They are required when using TrueType fonts emulated as a CID-keyed font. They should be found in the same directory as ordinary CMap files. -ToUnicode Mapping files are included in \package{adobemapping} package. +ToUnicode Mapping files are included in the \package{adobemapping} package. Those files are not required for \XeTeX\ users. \section{Overview of Extensions} @@ -420,12 +449,12 @@ users who are not involved in CJK languages. \dvipdfmx\ has an extensible support for encodings by means of PostScript CMap Resources. Just like \code{enc} files are written for 8-bit encodings, one can write their own CMap files to support custom encodings. -See, Adobe's technical notes for details on PostScript CMap Resources. +See, Adobe's technical notes\cite{ADOBE} for details on PostScript CMap Resources. \subsubsection{Vertical Writing} -\dvipdfmx\ supports vertical writing extension used by p\TeX\ and up\TeX. -A DVI instruction to set writing mode is supported. OpenType Layout +\dvipdfmx\ supports the vertical writing extension used by p\TeX\ and up\TeX. +A DVI instruction to set the writing mode is supported. The OpenType Layout GSUB Feature is supported for selecting vertical version of glyphs. \begin{figure} @@ -438,25 +467,26 @@ left and right corner brackets are replaced with their vertical counterparts.}% \subsection{Unicode Support} -Unicode support here consists of two parts: Supporting Unicode as input -encodings and making output PDF files ``Unicode aware''. +Unicode support here consists of two parts: Supporting Unicode as an input +encoding and making output PDF files ``Unicode aware''. \subsubsection{Unicode as Input Encoding} \dvipdfmx\ recognizes an additional keyword \code{unicode} in fontmap files -to declare that Unicode values are used in input DVI files. Unicode support is +to declare that Unicode values are used in the input DVI file. Unicode support is basically limited to the Basic Multilingual Plane (BMP) since there are no support for code ranges that requires more than three bytes in TFM and extended TFM formats. \subsubsection{ToUnicode CMap Support} -In PDF,it is often the case that text is not encoded in Unicode. +In PDF, it is often the case that text is not encoded in Unicode. However, modern applications usually want them represented in Unicode to make -it usable. ToUnicode CMaps are a bridge between PDF text string encodings -and Unicode encodings, and they makes it possible to extract text in PDF as +it usable as text information. +The ToUnicode CMap is a bridge between PDF text string encodings and Unicode +encodings, and makes it possible to extract text in PDF files as Unicode encoded strings. It is important to make resulting PDF search-able and -copy-and-past-able. Dvipdfmx supports the auto-creation of ToUnicode CMaps. +copy-and-past-able. Dvipdfmx supports auto-creation of ToUnicode CMaps. It will not work properly for multiply encoded glyphs due to fundamental limitations of Unicode conversion mechanism with ToUnicode CMaps. @@ -466,19 +496,28 @@ limitations of Unicode conversion mechanism with ToUnicode CMaps. \dvipdfmx\ can generate encrypted PDF documents to protect its contents from unauthorized access. It is limited to passorwd-based authentication, and public-key based authentication is not supported. The 256-bit AES encryption is -also supported for PDF version 1.7 setting although it may not be supported +also supported for PDF version 1.7 adn 2.0 setting although it may not be supported by PDF viewers. There are various other improvements over \dvipdfm. The most notable one is more improved PDF input and output support: The cross-reference stream and -object stream introduced in PDF-1.5 are also supported. +object stream introduced in \lnum{PDF-1.5} are also supported. + +\chapter{Graphics}\label{SEC:graphics} -\section{Graphics and Image Format}\label{SEC:graphics} +\section{Image Inclusion} -Graphics support was mostly rewritten. Support for BMP and JPEG\lnum{2000} is -added. An effort to preserve more information originally found in included -images, e.g., embedded ICC Profiles and XMP Metadata, was made. +The basics of incorporating images into output PDF is the same as in \dvipdfm. +To do this, \LaTeX\ users can simply use the \package{graphicx} package. +(possibly with the driver option \code{dvipdfmx}) +This section is \emph{not} for providing a how-to guide to include images but +just for supported graphics and image formats along with supported features. + +Graphics support was mostly rewritten in \dvipdfmx. +Support for BMP and JPEG\lnum{2000} was added. An effort to preserve more +information originally found in included images, e.g., embedded ICC Profiles +and XMP Metadata, was made. However, \dvipdfmx\ does not support various features common to graphics manipulation programs such as resampling, color conversion, and selection of @@ -489,53 +528,150 @@ in image manipulation for preparation of images. Supported formats are, PNG, JPEG, \lnum{JPEG2000}, BMP, PDF, and \MP\ generated EPS. All other format images, such as SVG and PostScript, must be converted to -other format supported by \dvipdfmx\ before inclusion. -The \option{-D} option, as in \dvipdfm, can be used for filtering images. +PDF before inclusion. The \option{-D} option, as in \dvipdfm, can be used for +filtering images. + +\subsubsection{Notes on PNG Support} + +PNG is supported as in \dvipdfm\ with many improvements. + +PNG support includes most of important features of PNG format such as color +palette, transparency, 16-bit bit-depth color, embedded ICC Profiles, +calibrated color, and embedded XMP Metadata. + +In including PNG images, \dvipdfmx\ first decompresses image data and then +compresses (if requested) it again. +For better compression ratio, a preprocessing filter (Predictor filter) might +be applied before compression. +\dvipdfmx\ supports the TIFF Predictor 2 and the PNG optimum filter. +However, there is yet no way to specify which predictor function is to be used +and currently PNG optimum filter is always employed. + +Predictor filter is a preprocessing filter to image data for improving compression. +Using a predictor filter usually gives better compression +but in many cases compression speed becomes significantly slower. +Try \option{-C 0x20} command line option to disable predictor filters and to +check if slowness is due to filtering. + +For the PNG optimum filter, a heuristic approach, ``minimum sum of absolute +differences'', is employed for finding the most optimal filter to be used. +See, discussion in the PNG Specification ''Filter selection'':\medskip + +\url{http://www.w3.org/TR/2003/REC-PNG-20031110/\#12Filter-selection} +\medskip + +As built-in support for the sRGB color space is absent in PDF, +the sRGB color can only be represented precisely by means of the sRGB ICC Profile. +However, for sRGB color PNG images, \dvipdfmx\ uses an approximate calibrated +RGB color space instead. +For approximating the sRGB color, the gamma and CIE \lnum{1931} chromaticity values +mentioned in the PNG Specification is used. +See, the following page for more information:\medskip -\subsection{PNG and JPEG} +\url{http://www.w3.org/TR/2003/REC-PNG-20031110/\#11sRGB} +\medskip + +\dvipdfmx\ also supports calibrated color with the \code{gAMA} and the \code{cHRM} chunk. +These chunks carry information on more accurate color representation. +Some software programs, however, write only \code{cHRM} but do not record the gamma value +although the PNG specification recommends to do so. Gamma value 2.2 is assumed if only +\code{cHRM} is present but \code{gAMA} is not. + +Some PNG features are unavailable depending on output PDF version setting. Please refer +Table~\ref{TABLE:PNGfeat} for more details. + +\begin{table} + \centering + \begin{tabular}{lp{8cm}}\hline + Feature & PDF Version Required \\ \hline\hline + 16-bit Color Depth & Version 1.5 \\ + Transparency & Full support for alpha channel requires PDF version 1.4. + Color key masking (a specific color is treated as fully transparent) + requires 1.3.\\ + XMP Metadata & Version 1.4 \\ + ICC Profile & Version 1.3 \\ + \hline + \end{tabular} + \caption{PNG features and corresponding PDF versions required.}% + \label{TABLE:PNGfeat} +\end{table} -PNG and JPEG are supported as in \dvipdfm\ but it is mostly rewritten. +\subsubsection{JPEG and \lnum{JPEG2000}} -PNG support includes most of important features of PNG such as color palette, -transparency, 16bit bit-depth color, embedded ICC Profiles, and calibrated -colors specified by \code{gAMA} and \code{cHRM} chunks. XMP Metadata is also -supported. Predictor filter may be applied for Flate compression which result in -better compression for larger images. Application of the predictor filter -sometimes makes compression speed very slow. Please try \option{-C 0x20} -(disable predictor filters) command line option to check if slowness is due to -the predictor filter. +JPEG format is supported as in \dvipdfm. In addition to this, \lnum{JPEG2000} +is also supported. + +JPEG and \lnum{JPEG2000} image inclusion is basically done as ''pass-through'', +that is, image data is not decompressed before inclusion. So, although these +formats are basically lossy, there should be no quality loss when including +images. JPEG is relatively well supported. \dvipdfmx\ supports embedded ICC Profiles -and CMYK color. Embedded XMP metadata is also supported. -\dvipdfmx\ uses JFIF or Exif data to determin image's physical size. +and CMYK color. Embedded XMP metadata is also preserved in the output PDF. +JFIF or Exif data might be used to determin image's physical size. + +As the PDF specification does not require information irrelevant to +displaying images to be embedded, \dvipdfmx\ does not embedd whole data. +Especially, not all application specific data is retained. Application +specific data such as JFIF, Exif, and \code{APP14} Adobe marker are +preserved. +Please note that XMP and Exif data which may contain location information +where the photo was taken is always preserved in the output PDF file. -\subsection{PDF and ``MPS''} +\lnum{JPEG2000} is also supported. It is restricted to JP2 and JPX baseline +subset as required by the PDF specification. It is not well supported and still in +an experimental stage. J2C format and transparency are not supported. + +\subsubsection{PDF Support} PDF inclusion is supported as in \dvipdfm, with various important enhancement -over \dvipdfm\ for robust inclusion. \dvipdfmx\ can handle cross-reference +over \dvipdfm\ for more robust inclusion. \dvipdfmx\ can handle cross-reference streams and object streams introduced in \lnum{PDF-1.5}. \dvipdfmx\ also supports inclusion of PDF pages other than the first page. However, tagged PDF may cause problems and annotations are not kept. -For \MP\ generated EPS files, multi-byte encoding support is added. -\dvipdfmx\ also supports ``\MP\ mode'': When \dvipdfmx\ is invoked with -\option{-M} option, it enters in \MP\ mode and processes a \MP\ -generated EPS file as input.% -\footnote{\code{prologue} must be set to \code{2}.} +As there are no clear way to determin the natural extent of the graphics content +to be clipped, \dvipdfmx\ preferably try to find the \emph{crop box} to decide +image size. If there are no crop box \emph{explicitly} specified,% +\footnote{There are some accusations by Japanese \TeX\ users as +''violating the PDF spec.'' regarding this point. However, what we are +talking about it how to guess the \emph{natural} or \emph{intended} size of +images but not the default value of the PDF crop box itself.} then it +tries to refer other boundary boxes such as the \emph{art box}. If there are no +possible boundaries of the graphics content explicitly specified, the \emph{media box}, +which is the boundaries of the physical medium on which the page is to be printed, +is used as the last resort. + +The \code{pdf:image} special supports additional keys, ``\code{page}'' and +``\code{pagebox}''. The \code{page} key takes an integer value representing +the page number of PDF page to be included, and the \code{pagebox} takes one +of the keywords \code{mediabox}, \code{cropbox}, \code{artbox}, \code{bleedbox}, +or \code{trimbox} for selecting page's boundary box to be used. For example, +\begin{lstlisting} +\special{pdf:image pagebox artbox page 3 (foo.pdf)} +\end{lstlisting} +includes 3rd page of `foo.pdf' with the boundary box set to art box. -\subsection{Additional Supported Formats} +\subsubsection{Other Image Formats} -BMP is supported but limited to uncompressed or RLE-compressed raster -images. Extensions are not (won't be) supported. +For \MP\ generated Encapsulated PostScript (EPS) files, multi-byte encoding +support is added. +\dvipdfmx\ also supports ``\MP\ mode''. When \dvipdfmx\ is invoked with +\option{-M} option, it enters in \MP\ mode and processes a \MP\ +generated EPS file as an input.% +\footnote{\code{prologue} should be set to \code{2}.} -\lnum{JPEG2000} is also supported. It is restricted to JP2 and JPX baseline -subset as required by PDF spec. It is not well supported and still in -an experimental stage. J2C format and transparency are not supported. +BMP support is also added. It is limited to uncompressed or RLE-compressed +raster images. Extensions are not (won't be) supported. +For image formats not natively supported, the \code{-D} option can be +used to convert images to PDF format before inclusion, as in \dvipdfm. +In \dvipdfmx, the letter \code{v} in the \code{-D} option argument is expanded +to the output PDF version. \subsection{Image Cache} -Chaching of images generated via filtering command specified with \option{-D} +Caching of images generated via filtering command specified with \option{-D} option is supported. It solves the problems that image inclusion becomes very slow when external filtering program such as GhostScript is invoked each time images are included. @@ -545,137 +681,336 @@ Use \option{-I} option to enable this feature: -I 24 \end{lstlisting} where the integer represents the life of cache files, 24 hours in this example. +Zero and negative values have a special meaning. +Value 0 for ``erase old cached images while leaving newly created one'', +-1 for ``erase all cached images'', and -2 for ``ignore image cache''. +Default value is set to -2. + +\section{Graphics Drawing} +\dvipdfmx\ does not offer a high level interface to draw graphics objects. +A possible way to draw grahpics is to write raw PDF graphics drawing codes and +then to insert them into the output via \code{special} commands. + +To show an example, the following code: +\begin{lstlisting} +\special{pdf:content + 1 0 0 1 0 0 cm + 0 100 m + 80 100 120 80 120 0 c + S +} +\end{lstlisting} +draws a Bézier curve, +\begin{center} +\specialbox{120bp}{100bp}{pdf:content 1 0 0 1 0 0 cm 0 100 m 80 100 120 80 120 0 c S} +\end{center} +The \code{pdf:content} special is used here which is useful for inserting an +isolated graphics object. + +The above example illustrates a typical example of PDF graphics drawing. +It consists of three parts; setting graphics state, constructing a path, and painting +a path. A Graphic object are specified as a sequence of operators and thier operands +using \emph{postfix notation}. \keyword{Graphics state operators} comes first, +\code{cm} in this example sets the current transformation matrix (CTM). Then, +\keyword{path construction} operators follow; move to position $(0, 100)$, +append a Bézier curve from current point to $(120, 0)$ with control points +$(80, 100)$ and $(120, 80)$. Finally, a \keyword{path painting} operator +comes to draw the constructed path. +In this example the stroking operator \code{S} +is used. + +\subsection{The \code{pdf:content} Special} + +The \code{pdf:content} special can be used for drawing an \emph{isolated} +graphics object. It sets the origin of graphics drawing operators supplied to +this command to the position where it is inserted. +The whole content is enclosed by a pair of graphics state save-restore +operators. So for example, a color change made within a \code{pdf:content} +command takes an effect only within the content of this special. + +\subsection{Guide to PDF Graphics Operators} + +PDF employs essentially the same imaging model as PostScript. +So, it is easy to learn about PDF graphics drawing for people who are +well accustomed to PostScript. This section is intended to be a short guide +for PDF graphics operators. + +\subsubsection{Graphics State Operators} + +The \code{cm} operator modifies CTM by concatinating the specified matrix. +Operands given to this operators are six numbers each representing +transformation matrix elements: +translation represented as $[1, 0, 0, 1, t_x, t_y]$, +scaling $[s_x, 0, 0, s_y, 0, 0]$, +rotation $[\cos\theta, \sin\theta, -\sin\theta, \cos\theta, 0, 0]$. + +To uniformly scale the object, just use +\begin{lstlisting} +2.0 0 0 2.0 0 0 cm +\end{lstlisting} -\section{DVI Specials}\label{SEC:specials} +The \code{w} operator sets the line width, e.g., `\code{2 w}' sets the line +width to 2. Here is an example of drawing a rotated rectangle: +\begin{lstlisting} +0.866 0.5 -0.5 0.866 30 2 cm 5 w 0 0 100 50 re S +\end{lstlisting} +\begin{center} +\specialbox{120bp}{96bp}{pdf:content 0.866 0.5 -0.5 0.866 30 2 cm 5 w 0 0 100 50 re S}% +\end{center} -\dvipdfmx\ is mostly compatible to \dvipdfm. -Several DVI \code{special} commnads are added for more flexible PDF generation: -Creation of arbitrary stream object, controlling \dvipdfmx\ behavior, and some -specials which might be useful for graphics drawing packages. +Transformations can be sequentially applied; for the above example, +\begin{lstlisting} +1 0 0 1 30 2 cm 0.866 0.5 -0.5 0.866 0 0 cm +5 w 0 0 100 50 re S +\end{lstlisting} +gives the same result. +To specify colors, use \code{RG}, \code{rg}, \code{K}, and \code{k} operators, +for RGB or CMYK color for stroking (upper-case) and nonstroking (lower-case). +\begin{lstlisting} +0.866 0.5 -0.5 0.866 30 2 cm 5 w +1 0.4 0 0 K 1 0 0 0 k +0 0 100 50 re B +\end{lstlisting} +where the \code{B} operator fill and then stroke the path. +\begin{center} +\specialbox{120bp}{96bp}{pdf:content 0.866 0.5 -0.5 0.866 30 2 cm 5 w +1 0.4 0 0 K 1 0 0 0 k 0 0 100 50 re B} +\end{center} -\subsection{Additions to \dvipdfm's \code{pdf:} Special} +A dash pattern can be specified with the \code{d} operator. +Operands for this operator are the \keyword{dash array} and +the \keyword{dash phase}. +For example, to specify a dash pattern with 6-on 4-off starting with phase 0: +\begin{lstlisting} +[6 4] 0 d 2 w 0 0 m 320 0 l S +\end{lstlisting} +draws the following dashed line: +\begin{center} +\specialbox{320bp}{30bp}{pdf:content +1 0 0 1 0 20 cm [6 4] 0 d 2 w 0 0 m 320 0 l S} +\end{center} -The \code{pdf:fstream} special is added, which enables creation of PDF stream -object from an existing file. +Other important operators are \code{q} and \code{Q}, which saves and restores +the graphics state. +\begin{lstlisting} +1 0 0 1 30 2 cm +q +0.866 0.5 -0.5 0.866 0 0 cm +[6 4] 0 d 2 w 0 0 100 50 re S +Q +-30 0 m 90 0 l S +0 -2 m 0 96 l S +\end{lstlisting} +In the above example, \code{d}, \code{w}, and rotation only take effect within +the \code{q}-\code{Q} block. The portion drawing two straight lines is unaffected +by these graphics state change. +\begin{center} +\specialbox{120bp}{96bp}{pdf:content +1 0 0 1 30 2 cm +q +0.866 0.5 -0.5 0.866 0 0 cm +[6 4] 0 d 2 w 0 0 100 50 re S +Q +-30 0 m 90 0 l S +0 -2 m 0 96 l S +} +\end{center} + +For a (incomplete) list of graphics state operators, see +Talbe~\ref{TAB:operatorsGS}. + +\begin{table} + \centering + \begin{tabular}{llp{7.6cm}}\hline + Operands & Operator & Description\\ \hline\hline + --- & \code{q} & Save the current graphics state.\\ + --- & \code{Q} & Restore the previously saved graphics state.\\ + $a$ $b$ $c$ $d$ $e$ $f$ & \code{cm} & Modify the current transformation matrix + by concatinating the specified matrix.\\ + \textit{width} & \code{w} & Set the line width.\\ + \textit{array} \textit{phase} & \code{d} & Set the line dash pattern.\\ + $r$ $g$ $b$ & \code{RG} & Set the stroking color space to RGB and + set the stroking color as specified.\\ + $r$ $g$ $b$ & \code{rg} & Set the nonstroking color space to RGB and + set the nonstroking color as specified.\\ + $c$ $m$ $y$ $k$ & \code{K} & Set the stroking color space to CMYK and + set the stroking color as specified.\\ + $c$ $m$ $y$ $k$ & \code{k} & Set the nonstroking color space to CMYK and + set the nonstroking color as specified.\\ + \hline + \end{tabular} + \caption{A few examples of graphics state operators and color operators.}% + \label{TAB:operatorsGS} +\end{table} + +\subsubsection{Path Construction Operators} + +A path construction normally starts with a \code{m} operator which moves the +current point to the specified position and then sequences of other path +construction operators follow. The path currently under construction is +called the \keyword{current path}. A sequence of path construction operators +appends segments of path to the current path and then move the +\keyword{current point} to the end point of appended path. +A typical sequence of path construction looks like, +\begin{lstlisting} +100 50 m +100 78 78 100 50 100 c +22 100 0 78 0 50 c +0 22 22 0 50 0 c +78 0 100 22 100 50 c +S +\end{lstlisting} +\begin{center} +\specialbox{120bp}{80bp}{pdf:content +100 50 m +100 78 78 100 50 100 c +22 100 0 78 0 50 c +0 22 22 0 50 0 c +78 0 100 22 100 50 c +S} +\end{center} +This example is an approximated circle drawn by four Bézier curves. + +Table~\ref{TAB:operators} shows a list of path construction operators. +Those who are accustomed to the PostScript language should note that in PDF +the current path is not a part of the graphics state, +and hece is \emph{not} saved and restored along with the other graphics state +parameters. + +\begin{table} + \centering + \begin{tabular}{llp{6.9cm}}\hline + Operands & Operator & Description \\ \hline\hline + $x$ $y$ & \code{m} & Begin a new path by moving the current point specified by given operands.\\ + $x$ $y$ & \code{l} & Append a line segment from the current point to the point specified.\\ + $x_1$ $y_1$ $x_2$ $y_2$ $x_3$ $y_3$ & \code{c} & Append a Bézier curve to the current path. + Two Control points and the end point given as operands.\\ + $x_2$ $y_2$ $x_3$ $y_3$ & \code{v} & Append a Bézier curve to the current path. Using the current + point and first two operand as the Bézier control points.\\ + $x_1$ $y_1$ $x_3$ $y_3$ & \code{y} & Append a Bézier curve to the current path. The second + control point coincides with the end point.\\ + --- & \code{h} & Close the current path by appending a straight line segment from the current point + to the starting point of the path.\\ + $x$ $y$ \textit{width} \textit{height} & \code{re} & Append a rectangle. First two operands for the + position of lower-left corner, third and forth operand representing width and + height.\\ + \hline + \end{tabular} + \caption{List of path construction operators. All operators move the current point to the end + point of appended path.}\label{TAB:operators} +\end{table} + +\subsubsection{Path Painting Operators} + +There are basically four kind of path painting operators: \code{S}, \code{f}, +\code{B}, and \code{n}. The first three for ``stroke'', ``fill'', and +``fill then stroke'' operators respectively, and the last one \code{n} +paints nothing but end the path object. For filling operators, there are +two kind of operators depending on how ``insideness'' of points are determined: +the \keyword{non-zero winding number rule} and the \keyword{even-odd rule}. +The even-odd rule operators are \code{f*} and \code{B*}. + +The following example illustrates the difference: +\begin{lstlisting} +0 0 100 100 re 20 20 60 60 re f +1 0 0 1 120 0 cm +0 0 100 100 re 20 20 60 60 re f* +\end{lstlisting} +\begin{center} +\specialbox{220bp}{100bp}{pdf:content +0 0 100 100 re 20 20 60 60 re f +1 0 0 1 120 0 cm +0 0 100 100 re 20 20 60 60 re f*} +\end{center} +The ``interior'' of the ``inner'' square has a non-zero even winding number. +(In this example, counter-clockwise direction is assumed for both of +two \code{re} operators.) + +\chapter{Specials} + +\section{PDF Specials}\label{SEC:specials} + +\dvipdfmx\ recognizes various special commands originally introduced in +\dvipdfm. Please refer to the ``Dvipdfm User's Manual''\cite{DVIPDFM} for detailed +information on PDF specials. + +\subsection{Additions to PDF Specials} + +Several \code{special} commnads are added for more flexible PDF generation: +creation of arbitrary stream objects, controlling \dvipdfmx\ behavior, and some +specials which might be useful for graphics drawing. + +\subsubsection{PDF Object Manipulation} + +PDF object manipulation is a key feature of PDF specials. +The \code{pdf:fstream} special is added in \dvipdfmx\ which enables creation of +PDF stream object from an existing file. The syntax of this special is, \begin{lstlisting} pdf:fstream @identifier (filename) <<dictionary>> \end{lstlisting} where identifier and filename (specified as a PDF string object) are -mandatory and a dictionary object which is to be added to the stream -dictionary following the filename is optional. +mandatory and a dictionary object, following the filename, which is to be added +to the stream dictionary is optional. -For examples, to incorporate a XMP Metadata, +For example, to incorporate XMP Metadata stored from the file \code{test.xmp}, \begin{lstlisting} \special{pdf:fstream @xmp (test.xmp) << - /Type /Metadata + /Type /Metadata /Subtype /XML >>} \special{pdf:put @catalog << /Metadata @xmp >>} \end{lstlisting} Similary, \code{pdf:stream} special can be used to create a PDF stream -object from a PDF string instead of a file: +object from a PDF string instead of a file. \begin{lstlisting} -\special{pdf:stream @MyPattern - (0.16 0 0 0.16 0 0 cm 4 w - 50 0 m 50 28 28 50 0 50 c S 100 50 - m 72 50 50 28 50 0 c S - 50 100 m 50 72 72 50 100 50 c S - 0 50 m 28 50 50 72 50 100 c S - 100 50 m 100 78 78 100 50 100 c 22 100 0 78 0 50 c - 0 22 22 0 50 0 c 78 0 100 22 100 50 c S - 0 0 m 20 10 25 5 25 0 c f 0 0 m 10 20 5 25 0 25 c f - 100 0 m 80 10 75 5 75 0 c f - 100 0 m 90 20 95 25 100 25 c f - 100 100 m 80 90 75 95 75 100 c f - 100 100 m 90 80 95 75 100 75 c f - 0 100 m 20 90 25 95 25 100 c f - 0 100 m 10 80 5 75 0 75 c f - 50 50 m 70 60 75 55 75 50 c 75 45 70 40 50 50 c f - 50 50 m 60 70 55 75 50 75 c 45 75 40 70 50 50 c f - 50 50 m 30 60 25 55 25 50 c - 25 45 30 40 50 50 c f - 50 50 m 60 30 55 25 50 25 c 45 25 40 30 50 50 c f) - << - /BBox [0 0 16 16] - /PaintType 2 - /PatternType 1 - /Resources << - /ProcSet [/PDF] - >> - /TilingType 3 - /Type /Pattern - /XStep 16 - /YStep 16 - >> -} +pdf:stream @identifier (stream contents) <<dictionary>> \end{lstlisting} -The above example defines a tiling pattern. With the following code: + +This special might be useful for creating a simple image inline. \begin{lstlisting} -\special{pdf:put @resources - << - /ColorSpace << /CS01 [/Pattern /DeviceRGB] >> - /Pattern << /PT01 @MyPattern >> - >> -} -\special{pdf:content - q 0.8 0.3 0.3 RG /CS01 cs 0.8 0.3 0.3 /PT01 scn - 0 0 240 80 re f +\special{pdf:stream @myim01 + <5500AAC05500AAC05500AAC05500AAC05500> + << + /Type /XObject + /Subtype /Image + /BitsPerComponent 1 + /ColorSpace /DeviceGray + /Width 9 + /Height 9 + >> } +\special{pdf:put @resources << + /XObject << /MyIM01 @myim01 >> +>>} +\special{pdf:content 81 0 0 81 0 0 cm /MyIM01 Do} \end{lstlisting} -it draws a box filled with a tiling pattern as shown in -Figure~\ref{FIG:pattern}. - -\special{pdf:stream @MyPattern - (0.16 0 0 0.16 0 0 cm 4 w - 50 0 m 50 28 28 50 0 50 c S 100 50 - m 72 50 50 28 50 0 c S - 50 100 m 50 72 72 50 100 50 c S - 0 50 m 28 50 50 72 50 100 c S - 100 50 m 100 78 78 100 50 100 c 22 100 0 78 0 50 c - 0 22 22 0 50 0 c 78 0 100 22 100 50 c S - 0 0 m 20 10 25 5 25 0 c f 0 0 m 10 20 5 25 0 25 c f - 100 0 m 80 10 75 5 75 0 c f - 100 0 m 90 20 95 25 100 25 c f - 100 100 m 80 90 75 95 75 100 c f - 100 100 m 90 80 95 75 100 75 c f - 0 100 m 20 90 25 95 25 100 c f - 0 100 m 10 80 5 75 0 75 c f - 50 50 m 70 60 75 55 75 50 c 75 45 70 40 50 50 c f - 50 50 m 60 70 55 75 50 75 c 45 75 40 70 50 50 c f - 50 50 m 30 60 25 55 25 50 c - 25 45 30 40 50 50 c f - 50 50 m 60 30 55 25 50 25 c 45 25 40 30 50 50 c f) - << - /BBox [0 0 16 16] - /PaintType 2 - /PatternType 1 - /Resources << - /ProcSet [/PDF] - >> - /TilingType 3 - /Type /Pattern - /XStep 16 - /YStep 16 - >> -}% +The above example draws an image like Figure~\ref{FIG:stream}. \begin{figure} -\centering -\raisebox{-80bp}[0bp][80bp]{\makebox[240bp][l]{\special{pdf:put @resources - << - /ColorSpace << /CS01 [/Pattern /DeviceRGB] >> - /Pattern << /PT01 @MyPattern >> - >> -}% -\special{pdf:content - 0.8 0.3 0.3 RG /CS01 cs 0.8 0.3 0.3 /PT01 scn - 0 0 240 80 re f -}}}% -\caption{An example of tiling pattern.}\label{FIG:pattern} + \centering + \makebox[81bp][l]{\raisebox{-81bp}[0bp][81bp]{% + \special{pdf:stream @myim01 <5500AAC05500AAC05500AAC05500AAC05500> + << + /Type /XObject + /Subtype /Image + /BitsPerComponent 1 + /ColorSpace /DeviceGray + /Width 9 + /Height 9 + >>} + \special{pdf:put @resources << + /XObject << /MyIM01 @myim01 >> + >>} + \special{pdf:content 81 0 0 81 0 0 cm /MyIM01 Do} +}} +\caption{An image created by \code{pdf:stream} special.}% +\label{FIG:stream} \end{figure} +\subsubsection{Controlling Font Mappings} + \code{pdf:mapline} and \code{pdf:mapfile} specials can be used to append a fontmap entry or to load a fontmap file: \begin{lstlisting} @@ -683,34 +1018,52 @@ pdf:mapline foo unicode bar pdf:mapfile foo.map \end{lstlisting} +\subsubsection{Specifying The Output PDF Version} + \code{pdf:majorversion} and \code{pdf:minorversion} specials can be used to specify major and minor version of output PDF. \begin{lstlisting} pdf:minorversion 3 \end{lstlisting} +\subsubsection{Encryption} + To protect output PDF with encryption, use \code{pdf:encrypt} special \begin{lstlisting} pdf:encrypt userpw (foo) ownerpw (bar) length 128 perm 20 \end{lstlisting} -where user-password and owner-password must be specified as PDF string objects. -(which can be empty) Numbers specifying key-length and permission flags here -must be decimal numbers. See, section of ``\hyperref[SEC:encryption]{Encryption -Support}'' for a brief description of permission flags. +where user-password (\code{userpw}) and owner-password (\code{ownerpw}) must be +specified as PDF string objects. (which can be empty) +Numbers specifying key-length and permission flags here are decimal numbers. +See, section ``\hyperref[SEC:encryption]{Encryption Support}'' for a brief +description of permission flags. + +\subsubsection{PDF Document Creation} + +\newfeature{Addition in \TeX\ Live 2019} + +As a convenience, the \code{pageresources} special is added, which puts given page +resources into subsequent page's \keyword{Resource Dictionary}. For example, +\begin{lstlisting} +\special{pdf:pageresources << + /ExtGState << /MyGS01 << /ca 0.5 /CA 0.5 >> >> +>>} +\end{lstlisting} +puts an ExtGState resource named \code{MyGS01} into the current page's and all +subsequent pages' resource dictionary. Other notable extensions are \code{code}, \code{bcontent}, and \code{econtent}. The \code{code} special can be used to insert raw PDF graphics -instructions into output page content stream. It is different from \dvipdfm's +instructions into the output. It is different from \dvipdfm's \code{content} special in that it does not enclose contents with a \code{q} and \code{Q} (save-restore of graphics state) pair. A typical usage of this special is: \begin{lstlisting} \special{pdf:code q 1 Tr} -text goes here... +...some text goes here... \special{pdf:code Q} \end{lstlisting} -which changes text rendering mode to 1, as shown in Figure~\ref{FIG:trmode} for -example. +which changes text rendering mode to 1, as shown in Figure~\ref{FIG:trmode}. \begin{figure} \centering @@ -722,28 +1075,173 @@ example. \end{figure} Be careful on using this special as it is very easy to generate -broken PDF files. The \code{bcontent} and \code{econtent} pair is fragile and -often incompatible to other groups of special commands. It is not -always guaranteed to work as `expected'. +broken PDF files. The \code{bcontent} and \code{econtent} pair is somewhat +fragile and might be incompatible to other groups of special commands. +It may not always be guaranteed to work as `expected'. + +\subsection{ToUnicode Special} + +PDF allows users to attach various additional information such as document information, +annotation, and navigation information (like bookmarks) to their document. +All human-readable text, \keyword{text string}, contained in such information must +be encoded either in \keyword{PDFDocEncoding} or UTF-16BE with Unicode byte order +marker. However, as many users can't prepare their document with text strings +properly encoded, there is a special kind of special command, \code{pdf:tounicode}, +which can be used to convert text strings into the appropriate Unicode form. +Note that this feature is provided just as a remedy for users incapable of encoding +text strings properly. + +For example, +\begin{lstlisting} +\special{pdf:tounicode 90ms-RKSJ-UCS2} +\end{lstlisting} +declares that \emph{some} of text strings should be re-encoded according to specified +conversion CMap \code{90ms-RKSJ-UCS2}. + +Conversion is done only for arguments to several PDF special commands such as +\code{docinfo}, \code{ann}, and \code{out} but not for that of general object creation +specials. Note that not all PDF string objects are subject to this conversion. +By default, only dictionary entries listed below are converted. +\begin{lstlisting} +Title Author Subject Keywords Creator Producer Contents Subj +TU T TM +\end{lstlisting} -\subsection{Dvipdfmx Extensions} +\newfeature{Addition in \TeX\ Live 2019} -A new special \code{dvipdfmx:config} was introduced in \TeX Live 2016 which -makes it possible to invoke a command line option. Most single letter command -line options except \option{D} are supported. For examples, +The list of dictionary entries subect to conversion can be extended by supplying +additional dictionary keys as an array object: \begin{lstlisting} -dvipdfmx:config C 0x10 + \special{pdf:tounicode 90ms-RKSJ-UCS2 [/RC /DS]} +\end{lstlisting} + +If the name of conversion CMap contains one of the keywords RKSJ, \lnum{B5}, GBK, and KSC, +PDF string objects are treated specially when they are parsed. A two byte sequence +starting with the first byte's hight bit set is treated ''as is'' so that +the \code{0x5c} byte appears in the second byte is not treated as an escape sequence. +This behavior is not complaiant to the PDF specification. + +\subsection{PDF Special Examples} + +This section shows several examples of special command usage. +It is intended to be a hint for advanced users, so uninterested users can safely skip +this section. In many cases, using \dvipdfmx\ PDF specials requires knowledge on PDF. +Please refer to Adobe's ''PDF Reference''\cite{ADOBE}. + +\subsubsection{Annotations} + +In this section, some useful special commands for creating \keyword{annotations} +are explained. Note that viewer support is required for annotations to be +displayed correctly. + +First start with a very simple \keyword{Text Annotation} for attaching a comment. +This feature is supported by many PDF viewer applications. +\marginnote{\special{pdf:ann width 20bp height 20bp + << + /Type /Annot + /Subtype /Text + /Name /Comment + /T (Text Annotation Example) + /Subj (An Example of Text Annotations) + /Contents (A Quick Brown Fox Jumped Over The Lazy Dog.) + >> +}} + +\begin{lstlisting} +\special{pdf:ann width 20bp height 20bp + << + /Type /Annot + /Subtype /Text + /Name /Comment + /T (Text Annotation Example) + /Subj (An Example of Text Annotations) + /Contents (A Quick Brown Fox Jumped Over The Lazy Dog.) + >> +} +\end{lstlisting} +\code{pdf:ann} special is used to create an annotation. A small icon shall +be shown on the side margin. Here, dictionary entry \code{T} is for the tilte, +\code{Subj} for the subject of this annotation, and \code{Contents} for +the text string to be shown in the body of this annotation. + +Likewise, \keyword{Rubber Stamp Annotation}, which places a rubber stamp +like figure, +\begin{center} +\specialbox{150bp}{50bp}{pdf:ann width 150bp height 50bp + << + /Type /Annot + /Subtype /Stamp + /Name /Approved + >> +} +\end{center} + +\begin{lstlisting} +\special{pdf:ann width 150bp height 50bp + << + /Type /Annot + /Subtype /Stamp + /Name /Approved + >> +} \end{lstlisting} -sets the \dvipdfmx's compatibility flag. +Other keywords such as \code{Expired}, \code{Final}, \code{Draft}, and so on, +can be used in place of \code{Approved}. -\subsection{DVI Special Examples} +One can create stamps of thier own style. For this purpose, other specials +\code{pdf:bxobj} and \code{pdf:exobj} can be used for designing stamps. +Those specials ``capture'' all typset material enclosed by them into a PDF +\keyword{Form XObject}, which is a reusable graphics object like included images. + +For a simple example, +\begin{lstlisting} +\special{pdf:bxobj @MyStamp + width 280pt height 0pt depth 40pt} +\addfontfeature{Scale=4,Color=FF9933}My Own Stamp +\special{pdf:exobj} +\end{lstlisting} +It captures typeset material ``My Own Stamp'' (this example uses \code{fontspec} +package's command for changing font size and text color) into the object labeled as +\code{MyStamp} for later reuse. +Then, \code{AP} (\keyword{appearance dicrionary}) entry for controling the appearance +of annotations is used as, +\begin{lstlisting} +\special{pdf:ann width 280pt height 40pt + << + /Type /Annot + /Subtype /Stamp + /AP << /N @MyStamp >> + >> +} +\end{lstlisting} +The image captured into the object \code{MyStamp} is used as ``Normal'' +(\code{AP} dictionary entry \code{N}) appearance. +(\code{R} for ``Rollover'' and \code{D} for ``Down'' can be used.) + +\parbox[t][0pt][c]{280pt}{% +\special{pdf:bxobj @MyStamp width 280pt height 0pt depth 40pt}% +\addfontfeature{Scale=4,Color=FF9933}My Own Stamp% +\special{pdf:exobj}% +}% -The following code reads \TeX\ source file and creates a stream object named -as \code{@SourceFile}, then creates a file attachment annotation. +The result is: +\begin{center} +\specialbox{280bp}{40bp}{pdf:ann width 280pt height 40pt + << + /Type /Annot + /Subtype /Stamp + /AP << /N @MyStamp >> + >> +} +\end{center} + +With the following code, \dvipdfmx\ reads the source file and creates a stream +object named \code{SourceFile}, and then creates file attachment annotation. \marginnote{% \special{pdf:fstream @SourceFile (\jobname.tex)}% \special{pdf:ann width 10bp height 20bp - << /Type /Annot + << + /Type /Annot /Subtype /FileAttachment /FS << /Type /Filespec @@ -752,7 +1250,7 @@ as \code{@SourceFile}, then creates a file attachment annotation. >> /Name /PushPin /C [0.8 0.2 0.2] - /T (The dvipdfmx project team) + /T (The Dvipdfmx User's Manual) /Subj (The Dvipdmfx User's Manual) /Contents (XeLaTeX source file of this manual.) >> @@ -760,7 +1258,8 @@ as \code{@SourceFile}, then creates a file attachment annotation. \begin{lstlisting} \special{pdf:fstream @SourceFile (\jobname.tex)}% \special{pdf:ann width 10bp height 20bp - << /Type /Annot + << + /Type /Annot /Subtype /FileAttachment /FS << /Type /Filespec @@ -769,25 +1268,52 @@ as \code{@SourceFile}, then creates a file attachment annotation. >> /Name /PushPin /C [0.8 0.2 0.2] - /T (The dvipdfmx project team) + /T (The Dvipdfmx User's Manual) /Subj (The Dvipdfmx User's Manual) /Contents (XeLaTeX source file of the manual.) >> } \end{lstlisting} -An Example of separation color space: +A push-pin image must be shown on the margin if viewer supports this kind of +annotation. +PDF viewer applications are required to provide predefined icon appearances +at least for the following standard icons: \code{Graph}, \code{PushPin}, +\code{PaperClip}, and \code{Tag}. + +\subsubsection{Special Color Space} + +This section shows various examples of using \keyword{Special color spaces}. +Examples in this section have a common structure. They consist of essentially +three parts. The first part is for definining color space itself. +PDF object creation commands like \code{pdf:obj} and \code{pdf:stream} are used +for this purpose. Next is for registering color space resources in the page's +\keyword{Resource Dicrionary}. It can be done via \code{pdf:put} command as, +\begin{lstlisting} +\special{pdf:put @resource << + /Category << ...key-value pairs... >> +>>} +\end{lstlisting} +where \code{@resource} is a special keyword representing current page's +Resource Dictionary and \code{Category} (to be replaced by actual category +name) is a category name such as \code{ColorSpace}. +Finally, graphics objects are placed, with or with a combination of +text and, PDF drawing operators inserted by the \code{pdf:code} or the +\code{pdf:contents} special. + +%%EXAMPLE: Separation Color Space +The first example is the \keyword{Separation} color space: \special{pdf:stream @TintTransform1 ({0 exch dup 0.62 mul exch 0}) << /FunctionType 4 - /Domain [ 0.0 1.0 ] - /Range [ 0.0 1.0 0.0 1.0 0.0 1.0 0.0 1.0 ] + /Domain [0.0 1.0] + /Range [0.0 1.0 0.0 1.0 0.0 1.0 0.0 1.0] >> }% \special{pdf:stream @TintTransform2 ({dup 0.78 mul exch dup 0.05 mul exch 0.71 mul 0}) << /FunctionType 4 - /Domain [ 0.0 1.0 ] - /Range [ 0.0 1.0 0.0 1.0 0.0 1.0 0.0 1.0 ] + /Domain [0.0 1.0] + /Range [0.0 1.0 0.0 1.0 0.0 1.0 0.0 1.0] >> }% \special{pdf:obj @Orange [ @@ -846,78 +1372,355 @@ An Example of separation color space: \special{pdf:code Q} } \end{lstlisting} -Transparency example: + +\code{TintTransform}'s defined here are functions for transforming +\keyword{tint} values into approximate colors in the +\keyword{alternate color space} +(\code{DeviceCMYK} in this example). PostScript calculator functions are used +for converting a single component value representing ``Orange'' or ``Green'' +into four component CMYK values to approximate those colors. The ``Orange'' +color $v$ is approximated as $(0, 0.62v, v, 0)$ in CMYK color space for +alternate display here. + +The \code{cs} operator for selecting color space and the \code{scn} operator +for color values are used in the \code{pdf:code} special. +Be sure that the \code{pdf:put} command, which puts color space resources +into the current page's Resource Dictionary, goes into the same page as +subsequent drawing commands. + +\dvipdfmx\ currently does not have an easy interface for using various color +space families such as CIE-Based color spaces (e.g., calibrated colors and +color space with an ICC profile) and Special color spaces (e.g., indexed, +separation, shading and patterns). + +%%EXAMPLE: Shading +Another example is a \keyword{shading pattern}: +\begin{lstlisting} +\special{pdf:put @resources << + /Shading << + /SH01 << + /ShadingType 2 + /ColorSpace @Orange + /Coords [0 0 320 20] + /Extend [true true] + /Function << /FunctionType 2 /Domain [0 1] /N 1.0 >> + >> + >> +>>} +\special{pdf:content 0 0 320 20 re W n /SH01 sh} +\end{lstlisting} +where the ``Orange'' separation color space defined before is used again. +This example shows an axial shading (\code{ShadingType} 2) pattern. +\begin{center} +\makebox[320pt][l]{ + \special{pdf:put @resources << + /Shading << + /SH01 << + /ShadingType 2 + /ColorSpace @Orange + /Coords [0 0 320 20] + /Extend [true true] + /Function << /FunctionType 2 /Domain [0 1] /N 1.0 >> + >> + >> + >>} + \raisebox{-20pt}[0pt][20pt]{\special{pdf:content 0 0 320 20 re W n /SH01 sh}} +} +\end{center} + +The shading pattern requires coordinate values to be mapped into color values. +Type 2 (Exponential Interpolation) \keyword{Function} is used for defining how +this mapping should occur here. +The above example, with the exponent $N=1$, is just a simple linear-gradient. + +The final examples is a \keyword{tiling pattern}. +\begin{lstlisting} +\special{pdf:stream @MyPattern + (0.16 0 0 0.16 0 0 cm 4 w + 50 0 m 50 28 28 50 0 50 c S 100 50 + m 72 50 50 28 50 0 c S + 50 100 m 50 72 72 50 100 50 c S + 0 50 m 28 50 50 72 50 100 c S + 100 50 m 100 78 78 100 50 100 c 22 100 0 78 0 50 c + 0 22 22 0 50 0 c 78 0 100 22 100 50 c S + 0 0 m 20 10 25 5 25 0 c f 0 0 m 10 20 5 25 0 25 c f + 100 0 m 80 10 75 5 75 0 c f + 100 0 m 90 20 95 25 100 25 c f + 100 100 m 80 90 75 95 75 100 c f + 100 100 m 90 80 95 75 100 75 c f + 0 100 m 20 90 25 95 25 100 c f + 0 100 m 10 80 5 75 0 75 c f + 50 50 m 70 60 75 55 75 50 c 75 45 70 40 50 50 c f + 50 50 m 60 70 55 75 50 75 c 45 75 40 70 50 50 c f + 50 50 m 30 60 25 55 25 50 c + 25 45 30 40 50 50 c f + 50 50 m 60 30 55 25 50 25 c 45 25 40 30 50 50 c f) + << + /BBox [0 0 16 16] + /PaintType 2 + /PatternType 1 + /Resources << + /ProcSet [/PDF] + >> + /TilingType 3 + /Type /Pattern + /XStep 16 + /YStep 16 + >> +} +\end{lstlisting} + +The above example defines a tiling pattern. The content stream containing +painting operators, \code{m} for ``move-to'', \code{c} for ``curve-to'', +\code{f} for ``fill'', and \code{S} for ``stroke'', defines the appearance of +the \keyword{pattern cell} for this tiling pattern. With the following code, +\begin{lstlisting} +\special{pdf:put @resources + << + /ColorSpace << /CS01 [/Pattern /DeviceRGB] >> + /Pattern << /PT01 @MyPattern >> + >> +} +\special{pdf:content + q 0.8 0.3 0.3 RG /CS01 cs 0.8 0.3 0.3 /PT01 scn + 0 0 320 100 re f +} +\end{lstlisting} +a box filled with the tiling pattern defined above is drawn. +\special{pdf:stream @MyPattern + (0.16 0 0 0.16 0 0 cm 4 w + 50 0 m 50 28 28 50 0 50 c S 100 50 + m 72 50 50 28 50 0 c S + 50 100 m 50 72 72 50 100 50 c S + 0 50 m 28 50 50 72 50 100 c S + 100 50 m 100 78 78 100 50 100 c 22 100 0 78 0 50 c + 0 22 22 0 50 0 c 78 0 100 22 100 50 c S + 0 0 m 20 10 25 5 25 0 c f 0 0 m 10 20 5 25 0 25 c f + 100 0 m 80 10 75 5 75 0 c f + 100 0 m 90 20 95 25 100 25 c f + 100 100 m 80 90 75 95 75 100 c f + 100 100 m 90 80 95 75 100 75 c f + 0 100 m 20 90 25 95 25 100 c f + 0 100 m 10 80 5 75 0 75 c f + 50 50 m 70 60 75 55 75 50 c 75 45 70 40 50 50 c f + 50 50 m 60 70 55 75 50 75 c 45 75 40 70 50 50 c f + 50 50 m 30 60 25 55 25 50 c + 25 45 30 40 50 50 c f + 50 50 m 60 30 55 25 50 25 c 45 25 40 30 50 50 c f) + << + /BBox [0 0 16 16] + /PaintType 2 + /PatternType 1 + /Resources << + /ProcSet [/PDF] + >> + /TilingType 3 + /Type /Pattern + /XStep 16 + /YStep 16 + >> +}% +\begin{center} +\raisebox{-100bp}[0bp][100bp]{\makebox[320bp][l]{\special{pdf:put @resources + << + /ColorSpace << /CS01 [/Pattern /DeviceRGB] >> + /Pattern << /PT01 @MyPattern >> + >> +}% +\special{pdf:content + 0.8 0.3 0.3 RG /CS01 cs 0.8 0.3 0.3 /PT01 scn + 0 0 320 100 re f +}}}% +\end{center} + +\subsubsection{Transparency} + +%%EXAMPLE: Transparency +\XeTeX's transparency feature is currently lost in \xdvipdfmx, but the same +effect can be achieved by setting graphics state parameters with +\code{ExtGState} resources and \code{gs} operator. Here is a simple +transparency example: \special{pdf:obj @gs01 << - /Type /ExtGState /CA 0.6 /ca 0.6 /AIS false + /Type /ExtGState /CA 0.5 /ca 0.5 /AIS false >>}% -\begin{center} +\begin{figure}[ht] +\centering \mbox{% + %\raisebox{120pt}{\parbox[t]{0.8\textwidth}{% + % \addfontfeature{Color=444444}\lipsum[1]}% + %}% + %\hspace{-0.8\textwidth}% \special{pdf:put @resources << /ExtGState << /GS01 @gs01 >> >>}% - \fontsize{200pt}{200pt}\selectfont - \special{pdf:code q 1.0 0.8 0.2 rg} - \textalpha% - \special{pdf:code /GS01 gs 0.4 0.8 0.4 rg}% - \hspace{-0.3em}\textbeta% + \fontsize{220pt}{220pt}\selectfont + \special{pdf:code q /GS01 gs 1.0 0.8 0.2 rg}% + α% + \special{pdf:code 0.4 0.8 0.4 rg}% + \hspace{-0.3em}% + β% \hspace{-0.3em}\raisebox{0.5ex}{% \special{pdf:code 0.4 0.4 0.8 rg}% - \textpi% - } + π% + }% + \special{pdf:code 0.6 0.2 0.8 rg}% + \hspace{-0.2em}% + γ% \special{pdf:code Q}% } -\end{center} +\end{figure} \begin{lstlisting} \special{pdf:obj @GS01 << - /Type /ExtGState /CA 0.6 /ca 0.6 /AIS false + /Type /ExtGState /CA 0.5 /ca 0.5 /AIS false >>}% \mbox{% \special{pdf:put @resources << /ExtGState << /GS01 @GS01 >> >>}% - \special{pdf:code q 1.0 0.8 0.2 rg}% - \textalpha% - \special{pdf:code /GS01 gs 0.4 0.8 0.4 rg}% - \hspace{-0.3em}\textbeta% + \special{pdf:code q /GS01 gs 1.0 0.8 0.2 rg}% + α% + \special{pdf:code 0.4 0.8 0.4 rg}% + \hspace{-0.3em}% + β% \hspace{-0.3em}\raisebox{0.5ex}{% \special{pdf:code 0.4 0.4 0.8 rg}% - \textpi% + π% }% + \special{pdf:code 1.0 0.2 0.4 rg}% + \hspace{-0.2em}% + γ% \special{pdf:code Q}% } \end{lstlisting} -where \code{CA} and \code{ca} represent opacity of stroke and fill color -individually. +where values for \code{CA} and \code{ca} represent opacity of stroke and +fill color respectively. Again, \code{pdf:put} command must go into the same +page as subsequent grahics and text drawing operators. -\section{Font Mapping} +\section{Dvipdfmx Extensions} -Syntax of fontmap file is basically the same as \dvipdfm. There are few -extensions in \dvipdfmx. In addition to 8-bit \code{enc} files and -keyword \code{builtin} and \code{none}, \dvipdfmx\ accepts CMap name and -the keyword \code{unicode} in the encoding field. +\newfeature{Addition in \TeX\ Live 2016} -This section is completely irrevant to \XeTeX\ users. +\noindent{}A new special \code{dvipdfmx:config} was introduced in +\TeX Live 2016 which makes it possible to invoke a command line option. +Several single letter command line options except \option{D} are supported. +For example, +\begin{lstlisting} +dvipdfmx:config C 0x10 +\end{lstlisting} +sets the \dvipdfmx's compatibility flags. See, the section +``\hyperref[SEC:compatibility]{Incompatible Changes}'' for an explanation of +compatibility flags. + +\section{PS Specials} + +PS (PostScript) specials can be used to insert a raw PostScript code for drawing +graphics objects and transforming subsequent text and graphics. +Please note that support for PostScript operators in \dvipdfmx\ is very +limited. It is just enough for interpreting PostScript figures output by \MP. +Only a basic set of operators for arithmetic and math, stack operation and manupilation, +graphics state, path construction and painting, glyph and font, are supported. +See, Table~\ref{TABLE:PS} for the list of +recognized PostScript operators. + +\begin{table} + \centering + \begin{tabular}{p{3cm}>{\raggedright\arraybackslash}p{8cm}}\hline + Classification & Operators \\ \hline\hline + Arithmetic \& Math & \code{add} \code{sub} \code{mul} \code{div} \code{neg} \code{truncate}\\ + Stack Operation & \code{clear} \code{pop} \code{exch}\\ + Graphis State & \code{gsave} \code{grestore} + \code{setlinewidth} \code{setdash} \code{setlinecap} \code{setlinejoin} \code{setmiterlimit} \code{setgray} \code{setrgbcolor} \code{setcmykcolor} \\ + Coordinate System & \code{concat} \code{scale} \code{translate} \code{rotate} \code{idtransform} \code{dtransform}\\ + Path Construction & \code{currentpoint} \code{newpath} \code{closepath} \code{moveto} \code{rmoveto} \code{lineto} \code{rlineto} \code{curveto} \code{rcurveto} \code{arc} \code{arcn} \code{clip} \code{eoclip} \\ + Painting & \code{stroke} \code{fill} \\ + Glyph \& Font & \code{show} \code{findfont} \code{scalefont} \code{setfont} \code{currentfont} \code{stringwidth}\\ + \hline + \end{tabular} + \caption{List of PostScript operators recognized by \dvipdfmx.}% + \label{TABLE:PS} +\end{table} + +It might be enough for the purpose of basic graphics drawings but as there are +no support for conditionals and controls it is not enough for complicated tasks, +espacially, the PSTricks package is not supported. + +In \dvipdfmx, text handling is extended to support CJK text. +The following code draws Japanese text like shown in Figure~\ref{FIG:verttext}: +\begin{lstlisting} +\special{pdf:mapline uprml UniJIS-UTF8-H yumindb.ttf} +\special{ps: uprml findfont 16 scalefont setfont + currentpoint moveto + (...some Japanese text goes here...) show +} +\end{lstlisting} + +\chapter{Fonts and Encodings} + +\section{Fonts and Encodings Support} + +In \dvipdfmx, all font formats supported by \dvipdfm\ are also supported with +many improvements: The CFF conversion for PostScript Type1 fonts\footnote{PostScript +Type1 font support is restricted to the binary format as in \dvipdfm.} is +implemented which greatly reduces the output file size. Embedded TrueType fonts are +now subsetted. The OpenType font format is also supported.\footnote{Its +implementation is based on the OpenType specification version 1.4. +Newly added features such as color fonts and variable fonts are not supported yet.} + +There are various enhancements made for supporting Unicode and legacy multi-byte +character encodings for East Asian languages. + + +\section{Font Mappings} + +The Syntax of font-mapping (fontmap) files is basically the same as in \dvipdfm. +There are few extensions available in \dvipdfmx. In addition to the 8-bit \code{enc} +file and keywords \code{builtin} and \code{none}, \dvipdfmx\ accepts a PostScript CMap +Resource name and the keyword \code{unicode} in the encoding field. + +When the keyword \code{unicode} is specified in the encoding field of fontmap +files, it is assumed that Unicode values are used in the input DVI file. + +\begin{lstlisting} + urml unicode SourceHanSerifJP-Light.otf +\end{lstlisting} + +Although the DVI format allows 3-byte and 4-byte character codes to be used, +\dvipdfmx\ only supports up to 2-byte range since there is no TFM format supporting +3-byte or 4-byte codes. + +For PostScript Type1 fonts which do not support Unicode natively, an auxilially file, +the Adobe Glyph List, is required to make it possible to use fonts with Unicode access. + +As a general framework for supporting legacy multi-byte encodings, \dvipdfmx\ employs +PostScript CMap Resources for handling input strings encoded in various +character encodings. A CMap name can be specified in the encoding field just like +the encoding name for 8-bit encodings. For example, to specify the CMap ``UniJIS-UCS2-H'', + +\begin{lstlisting} + urml UniJIS-UCS2-H HiraMinPro-W3.otf +\end{lstlisting} + +For information on the Adobe Glyph List and PostScript CMap Resources, see, +the section \ref{SEC:auxfiles}, ``Auxilially Files.'' \subsection{Extended Syntax and Options} -Few options are available in \dvipdfmx\ in addition to the original -dvipdfm's one. All options that makes \dvipdfmx\ to use unembedded -fonts are deprecated as by using them makes \dvipdfmx\ to create PDF -files which can be not compliant to the ISO standards. +Few options are available in \dvipdfmx\ in addition to the original dvipdfm's +one. Please note that all features which makes \dvipdfmx\ to use +non-embedded fonts are deprecated, as by doing so it makes \dvipdfmx\ to create +PDF files which can be non-compliant to the ISO standards. \subsubsection{SFD Specification} -For bundling up fonts split into multiple subfonts via SFD back into -a single font, dvipdfmx supports an extended sytax of the form +For bundling up a font split into multiple subfonts via SFD back into +a single font, dvipdfmx supports extended sytax of the form \begin{lstlisting} tfm_name@SFD@ encoding filename options \end{lstlisting} - A typical example looks like: \begin{lstlisting} gbsn@EUC@ GB-EUC-H gbsn00lp \end{lstlisting} - where TFMs \code{gbsn00}, \code{gbsn01}, \code{gbsn02}... are mapped into a single font named \code{gbsn00lp} via the rule described in the SFD file \code{EUC}. @@ -925,11 +1728,10 @@ single font named \code{gbsn00lp} via the rule described in the SFD file \subsubsection{TrueType Collection Index} TrueType Collection index number can be specified with \code{:n:} -in front of TrueType font name: +in front of the TrueType font name: \begin{lstlisting} min10 H :1:mincho \end{lstlisting} - In this example, the option \code{:1:} tells \dvipdfmx\ to select first TrueType font from the TTC font \code{mincho.ttc}. Alternatively, the \option{-i} option can be used in the option field to specify TTC index: @@ -940,10 +1742,10 @@ min10 H mincho -i 1 \subsubsection{Non-embedding Switch} \deprecated{Use of this option is deprecated.} -\noindent{}The character \code{!} in front of the font name can be used to +\noindent{}The character \option{!} in front of the font name can be used to indicate that the font shall not be embedded. This feature greatly reduces the -size of the final PDF output, but the PDF file may not be viewed exactly in -other systems on which appropriate fonts are not installed. +size of the final PDF output, but the PDF file may not be viewed exactly the same +in other systems on which appropriate fonts are not installed. \bigskip \noindent{}NOTE: \dvipdfmx\ always converts input encodings to CIDs and then @@ -966,7 +1768,7 @@ described here since it might still be useful for some situations. \dvipdfmx\ recognizes several `Standard' CJK fonts although there are no such notion in PDF. In older days where there were not so many freely available CJK -fonts, it was sometimes useful to create PDF files without embedded fonts and +fonts, it was sometimes useful to create PDF files without embedding fonts and let PDF viewers or printers to use substitute fonts (tend to be higher quality) installed in their systems. \dvipdfmx\ `knows' several fonts which might be available in PostScript printers and PDF applications such as Acrobat Reader, @@ -1015,7 +1817,7 @@ supported for those fonts. used to create synthetic bold, italic, and bolditalic style variants from other font using PDF viewer's (or OS's) function. \begin{lstlisting} -jbtmo@UKS@ UniKSCms-UCS2-H :0:!batang,Italic +jbtmo@UKS@ UniKSCms-UCS2-H :0:!batang,Italic \end{lstlisting} Availability of this feature highly depends on the implementation of PDF @@ -1025,14 +1827,15 @@ use of it is deprecated. \subsection{Specifying Unicode Plane} -As there are no existing 3-bytes or 4-bytes TFM formats, the only way to use -Unicode characters other than the BMP is to map code range 0-65535 to different -planes via (e.g., to plane 1) \option{-p 1} fontmap option. This option is +As there are no existing TFM formats supporting 3-byte or 4-byte character +encodings, the only way to use Unicode characters other than the BMP is to +map the code range 0-65535 to different planes via (e.g., to plane 1) +the \option{-p 1} fontmap option. This option is available only when \code{unicode} is specified in the encoding field. \subsection{OpentType Layout Feature} -OpenType Layout Feature fontmap options mentioned below are only meaningful +The OpenType Layout Feature fontmap options mentioned below are only meaningful when \code{unicode} is specified in the encoding field. With the \option{-w} option, writing mode can be specified. @@ -1040,12 +1843,14 @@ With the \option{-w} option, writing mode can be specified. enables an OpentType Layout Feature related to vertical writing, namely, \code{vert} or \code{vrt2}, to choose proper glyphs for vertical text. -The \option{-l} option can be used to enable various OpenType Layout GSUB -Features. For examples, \option{-l jp04} -enables \code{jp04} feature to select \lnum{JIS2004} forms for Kanjis. +\newfeature{Addition in \TeX\ Live 2017.} + +The \option{-l} (lower case el) option can be used to enable various +OpenType Layout GSUB Features. For example, \option{-l jp04} enables +\code{jp04} feature to select \lnum{JIS2004} forms for Kanjis. Features can be specified as a ``:'' separated list of OpenType Layout -Feature tags like \option{-l vkna:jp04}. -Script and language may be additionally specified as +Feature tags like \option{-l vkna:jp04}. Script and language may be +additionally specified as \option{-l kana.JAN.ruby}. An example can be @@ -1053,7 +1858,7 @@ An example can be uprml-v unicode SourceHanSerifJP-Light.otf -w 1 -l jp90 \end{lstlisting} which declares that font should be treated as for vertical writing and -use \lnum{JIS1990} form for Kanjis. (See, Figure~\ref{FIG:jp90} for an example) +use \lnum{JIS1990} form for Kanjis. \begin{figure} \centering @@ -1062,94 +1867,251 @@ use \lnum{JIS1990} form for Kanjis. (See, Figure~\ref{FIG:jp90} for an example) \end{figure} This feature is limited to the single substitution, there are no way to select -a glyph from multiple candidates, such as in \code{aalt}, and specifying +a glyph from multiple candidates, such as in the \code{aalt} feature, and specifying general many-to-many glyph substitutions does not take effect. +\section{Other Improvements} + +This section briefly describes other improvements made for \dvipdfmx. +There is an extension to glyph name handling in the \code{enc} file for +seamless support of both PostScript Type1 and TrueType fonts. +PostScript Type1 font support is enhanced although this format might be +considered obsolete. + +\subsection{Extended Glyph Name Syntax} + +\dvipdfmx\ accepts the following syntax for glyph names in the \code{enc} file: +\code{uni0130}, \code{zero.onum} and \code{T\_h.liga}. +Each represents a glyph accessed with Unicode value \code{U+0130}, +oldstyle number for zero and ``Th'' ligature accessed via the OpenType +Layout GSUB Feature \code{onum} and \code{liga}, respectively. +Note that \dvipdfmx\ does not understand glyph names which directly +use a glyph index such as \code{index0102} or \code{gid2104}. + +When \dvipdfmx\ encounters a glyph name, e.g., \code{T\_h.liga}, it first looks +for OpenType \code{post} table if such glyph name exists; if it exists, then +\dvipdfmx\ simply uses \code{post} table and maps the glyph name to the glyph index; +if not, \dvipdfmx\ tries to convert \code{T\_h} to a Unicode sequence (U+0054 +U+0068 in this example) via the AGL mapping; then, OpenType \code{cmap} table is used +to further convert the resulting Unicode sequence to the sequence of glyph indices; +finally, the OpenType Layout Feature \code{liga} is applied to get the desired glyph. + +A glyph name of a form \code{a.swsh2} can be specified to denote the 2nd swash +variant form of the letter `a'. + +\subsection{CFF Conversion} + +\dvipdfmx\ supports on-the-fly PostScript Type1 to CFF (Type1C) conversion +which greatly reduces size of the resulting PDF file when using Type1 fonts. +Conversion is essentially `lossless' and there should not be any quality loss. +However, due to differences in the ability of rasterizers, there might be +noticeable differences on rendering result. + +When an older Type1 font is used, \dvipdfmx\ may give the following warning +message: +\begin{lstlisting} +Obsolete four arguments of "endchar" will be used for Type1 +"seac" operator. +\end{lstlisting} +It happens when there is an accented character made as a composite glyph using +the ``seac'' operator. +This warning is issued as conversion can't be done without relying on the +\emph{deprecated} usage of the \code{endchar} operator. However, as mentioned in +``Appendix C Compatibility and Deprecated Operators'' of Adobe Technical +Note \#5177, +``\href{http://wwwimages.adobe.com/content/dam/Adobe/en/devnet/font/pdfs/5177.Type2.pdf}{Type 2 Charstring Format}'', +PDF applications should support this operator and hence this warning +message can be ignored. + +Use of Type1 font should be avoided as much as possible. +Please consider using OpenType version instead. + +\section{Font Licensing} + +In OpenType font format, information regarding how a font should be treated +when creating a document can be recorded.% +\footnote{See, +``\href{http://www.microsoft.com/typography/otspec/os2.htm}{OpenType Sepcification: +OS/2 -- OS/2 and Windows Metrics Table}''.} +\dvipdfmx\ uses this information to decide whether font embedding is permitted. + +This font licensing information is indicated by the flag called \code{fsType} +recorded in OpenType font files; each bits representing different restrictions on +font embedding. If multiple flag bits are set in \code{fsType}, the least restrictive +license granted takes precedence in \dvipdfmx. +The \code{fsType} flag bit recognized by \dvipdfmx\ is as follows: +\begin{itemize} + \item Installable embedding + \item Editable embedding + \item Embedding for Preview \& Print only +\end{itemize} +\dvipdfmx\ issues the following warning message for fonts with `Preview \& Print only' +setting: +\begin{verbatim} + This document contains 'Preview & Print' only licensed font +\end{verbatim} + +For a font with this type of licensing, font embedding is allowed solely for the +purpose of (on-screen) viewing and/or printing; further editing of the document +or extracting embedded font data for other purposes are not allowed. +One way to ensure this condition is to protect your document with a non-empty +password. + +All other flags are treated as more restrictive license than any of +the above flags and treated as ``No embedding allowed''; e.g., if both +of the editable-embedding flag and unrecognized license flag is set, +the font is treated as editable-embedding allowed, however, if only +unrecognized flags are set, the font is not embedded. + +Font Embedding flags are preserved in the embedded font if they are embedded +as a TrueType font or a CIDFontType2 CID-keyed font. +For all fonts embedded as a PostScript font (Type1C and CIDFontType0 +CID-keyed font), they are not preserved. +Only \code{Copyright} and \code{Notice} in the \keyword{FontInfo} dictionary +are preserved in this case. + +Some font vendors put different embedding restrictions for different +condition; e.g., font embedding might not be permitted for the commercial +use unless you acquire the ``commercial license'' separately. +Please read EULA carefully before making decision on the font usage. + +See, for example, +\href{http://www.adobe.com/products/type/font-licensing/font-embedding-permissions.html}{Adobe's site on font embedding permissions} +for the font in the Adobe Type Library. +Microsoft also has a +\href{http://www.microsoft.com/typography/RedistributionFAQ.mspx}{FAQ page on Font Redistribution}. + +For Japanese font in general, embedding permission tend to be somewhat +restrictive. Japanese users should read the statement regarding font +embedding from Japan Typography Association (in Japanese):\medskip + +\url{http://www.typography.or.jp/act/morals/moral4.html} +\medskip + +\dvipdfmx\ does not support full embedding. Only subset embedding is supported. + +\chapter{Encryption} + \section{Encryption Support}\label{SEC:encryption} -\dvipdfmx\ offers basic PDF password security support including -256-bits AES encryption. Only ``Standard'' security handler is supported and -public-key security handlers are currently not supported. +\dvipdfmx\ offers basic PDF password security support including the 256-bit AES encryption. +Only the ``Standard'' security handler is supported and the Public-key security handler is not +supported. + Encryption is enabled by \option{-S} command line option. +\begin{lstlisting} + dvipdfmx -S -K 128 -P 0x14 +\end{lstlisting} +where \code{-K} and \code{-P} options are used to specify encryption key length and +permission flags respectively, and are briefly explained in Table~\ref{TABLE:enc-options}. + +\begin{table} + \centering + \begin{tabular}{lp{8cm}}\hline + Option & Description \\ \hline\hline + \code{-S} & Enable PDF encryption. \\ + \code{-K} \textit{number} & Set encryption key length. The default value + is 40.\\ + \code{-P} \textit{number} & Set permission flags for PDF encryption. + The \textit{number} is a 32-bit unsigned integer representing permission + flags. + See, Table~\ref{TABLE:flags}. The default value is \code{0x003C}.\\ + \hline + \end{tabular} + \caption{Command line options for encryption.}% + \label{TABLE:enc-options} +\end{table} + +When \dvipdfmx\ is invoked with encryption via the \code{-S} option, +passwords will be asked. +However, in some circumstances, it might be desirable to avoid being prompted for +passwords. In that case, use the \code{pdf:encrypt} special to supply passwords in +the \TeX\ file, as, +\begin{lstlisting} + \special{pdf:encrypt userpw (foo) ownerpw (bar) length 128 perm 20} +\end{lstlisting} +Here, user and owner passwords are supplied as PDF string objects (\code{foo} and \code{bar} +in the example above) which can be empty. -When encrypting the document, up to two passwords can be specified: -an owner password and a user password. If a user attempts to open an encrypted -document that has a user password, PDF applications prompt for a password. -Correctly supplying either password enables the user to open the document to -display or to access to the contents. +Up to two passwords can be specified for a document -- an owner password and a user password. +If a user attempts to open an encrypted document with user password being set, PDF application +should prompt for a password. Users are allowed to access the contents of the document only when +either password is correctly supplied. Depending on which password (user or owner) was supplied, additional operations -allowed for an opened document is determined; full access for users who opened -the document with the correct owner password or additional operations -controlled by permission flags supplied by the owner of the document for users -who opened the document with the correct user password. - -Access permision flags can be specified via the \option{-P} option. Each bits -of the (unsigned) integer number given to this option represents user access -permissions; -e.g., bit position 3 for allowing ``print'', 4 for ``modify the contents'', -5 for ``copy or extract text'', and so on. -\begin{table}[b] +allowed for the opened document is determined; full access for users who opened +with the correct owner password or additional operations +controlled by permission flags for users who opened with the correct user password. + +Although PDF specification allows various character encodings other than \code{US-ASCII} +for entering password, \dvipdfmx\ is unable to handle it properly. +Thus it must be assumed that \code{US-ASCII} is used for password strings. + +Access permision flags can be specified via \option{-P} command-line option. +Each bits of (32-bit unsigned) integer number given to this option represents user +access permissions; e.g., bit position 3 for allowing ``print'', 4 for +``modify'', 5 for ``copy or extract'', and so on. See, Table~\ref{TABLE:flags}. +For example, \code{-P 0x34} allows printing, copying and extraction of text, and adding and +modifying text annotation and filling in interactive form fields (but disallows +modification of the contents of the document). +\begin{table} \centering \begin{tabular}{lp{8.3cm}}\hline Bit Position & Meaning \\ \hline\hline - 3 & Print the document. \\ - 4 & Modify the contents of the document. \\ + 3 & (Revision 2) Print the document. \\ + & (Revision 3 or greater) Print the document. Print quality depending on bit 12.\\ + 4 & Modify the contents of the documentby operations other than those controlled + by bits 6, 9, and 11. \\ 5 & Copy or extract text and graphics from the document. \\ 6 & Add or modify text annotations, fill in interactive form fields. Creation and modification of interactive form field is also allowed if bit 4 is set.\\ + 9 & (Revision 3 or greater) Fill in existing interactive form fields + (including signature fields), even if bit 6 is clear.\\ + 10 & \em{Deprecated in PDF 2.0} \\ + & (Revision 3 or greater) Extract text and graphics (in support of accessibility to + users with disabilities or for other purposes).\\ + 11 & (Revision 3 or greater) Assemble the document + (insert, rotate, or delete pages and create document outline items or thumbnail + images), even if bit 4 is clear.\\ + 12 & (Revision 3 or greater) High-quality printing. + When this bit is clear (and bit 3 is set), printing shall be limited to a low-level, + possibly of degraded quality.\\ \hline \end{tabular} - \caption{Flag bits and their short explaination for the Revision 2 - Standard Security Handler.} + \caption{Flag bits and their short explaination. + Revision 2 is used when encryption key length is 40 bits or when PDF output + version is less than 1.5. Otherwise, Rivision 3 or greater is used.}\label{TABLE:flags} \end{table} -For examples, -\begin{lstlisting} -dvipdfmx -S -P 0x34 foo.dvi -\end{lstlisting} -allows printing, copying and extraction of text, and adding and modifying -text annotation and filling in interactive form fields (but disallows -modification of the contents of the document). The \option{-K} option can be used to specify the encryption key length. The key length must be multiple of 8 in the range 40 to 128, or 256 (for PDF -version 1.7 plus Adobe Extension or forthcoming PDF version 2.0). - -Password will be asked when encryption is enabled. It may not work well on -Windows platforms. Use the \code{pdf:encrypt} special instead of command line -options in this case. - -The default values for \option{K} is 40 and for \option{-P} is \code{0x003C0}. - -\section{Other Improvements} - -This section briefly describes other improvements made for \dvipdfmx. - -\subsection{Extended Glyph Name Syntax} - -\dvipdfmx\ accepts the following syntax for glyph names in \code{enc} files: -\code{uni0130}, \code{zero.onum} and \code{T\_h.liga}. -Each represents a glyph accessed with Unicode value \code{U+0130}, -oldstyle number for zero and ``Th'' ligature accessed via OpenType -Layout GSUB Feature \code{onum} and \code{liga} respectively. -Note that \dvipdfmx\ does not understand glyph names which directly -use glyph indices such as \code{index0102} or \code{gid2104}, since those -indices are private to each font. +version 1.7 plus Adobe Extension or PDF version 2.0). Please note +that when key length 256 is specified for PDF version 1.7 output, it requires +Adobe's Extension to \lnum{PDF-1.7} and hence PDF applications may not support it. +PDF version 1.4 is required for key length more than 40 bits. AES encryption algorithm requires +PDF version 1.6. + +To show some examples:\\ +128-bit AES encryption with print-only (high-quality) setting, +\begin{lstlisting} + dvipdfmx -V 5 -S -K 128 -P 0x804 input.dvi +\end{lstlisting} +256-bit AES encryption with print (low-quality), adding and modifying text annotations +allowed, +\begin{lstlisting} + dvipdfmx -V 2.0 -S -K 256 -P 0x24 input.dvi +\end{lstlisting} -When \dvipdfmx\ encounters a glyph name, e.g., \code{T\_h.liga}, it first looks -for OpenType \code{post} table if such glyph exists; if it exists, then -\dvipdfmx\ simply uses \code{post} talbe for mapping glyph name to glyph index; -if not, \dvipdfmx\ tries to convert \code{T\_h} to Unicode sequence (U+0054 -U+0068 in this example) via the AGL mapping; then, OpenType cmap table is used -to further converting resulting Unicode sequence to a sequence of glyph indices; -finally, OpenType Layout Feature \code{liga} is applied to get desired glyph. +The default value for \option{K} is 40 and for \option{-P} is \code{0x003C0} +(all bits from bit-position 3 to 6 set). +\chapter{Compatibility} \section{Incompatible Changes}\label{SEC:compatibility} There are various minor incompatible changes to \dvipdfm. -The \option{-C} command line option may be used for compatibility to +The \option{-C} command line option may be used for compatibility to \dvipdfm\ or older versions of \dvipdfmx. The \option{-C} option takes flags meaning \begin{itemize} @@ -1169,71 +2131,38 @@ The remap option \option{-r} in fontmaps is no longer supported and is silently ignored. The command line option \option{-e} to disable partial (subset) font embedding is not supported. -\section{Font Licensing and Embedding} +\section{Important Changes} -In OpenType format, information regarding how a font should be treated -when creating documents can be recorded.% -\footnote{See, `` \href{http://www.microsoft.com/typography/otspec/os2.htm}{OpenType Sepcification: OS/2 -- OS/2 and Windows Metrics Table}''.} -\dvipdfmx\ uses this information to decide whether embedding font is permitted. - -This font embedding information is indicated by a flag called -\code{fsType}; each bit representing different restrictions on font -embedding. -If multiple flag bits are set in \code{fsType}, the least restrictive -license granted takes precedence in \dvipdfmx. -The \code{fsType} flag bits recognized by \dvipdfmx\ is as follows: +Here is a list of important changes since the \TeX\ Live 2016 release: \begin{itemize} - \item Installable embedding - \item Editable embedding - \item Embedding for Preview \& Print only +\item Changes to make PDF/A creation easier: Always write CIDSet and CharSet +for embedded fonts. Do not compress XMP metadata. +\item Merge from libdpx for p\TeX-ng by Clerk Ma. +\item Addition of \code{STHeiti-Regular-Acro} for CJK `Standard' fonts. +\item Command line option \option{-p} takes precedence over \code{papersize} +and \code{pagesize} specials. +\item Fixed serious bugs in supporting `\code{unicode}' encoding: +OpenType Layout Feature \code{vert} and \code{vrt2} was not enabled. +Support for format 2 CFF charsets was broken. +\item Added simplified version of OpenType Layout support: The `\option{-l}' +option in fontmaps. \end{itemize} -\dvipdfmx\ give the following warning message for fonts with `Preview \& -Print only' setting: -\begin{verbatim} - This document contains 'Preview & Print' only licensed font -\end{verbatim} - -For fonts with this type of licensing, font embedding is allowed solely for the -purpose of (on-screen) viewing and/or printing; further editing of the document -or extracting embedded font data for other purposes are not allowed. -One way to ensure this condition is to protect your document with a non-empty -password. - -All other flags are treated as more restrictive license than any of -the above flags and treated as ``No embedding allowed''; e.g., if both -of the editable-embedding flag and unrecognized license flag is set, -the font is treated as editable-embedding allowed, however, if only -unrecognized flags are set, the font is not embedded. - -Embedding flags are preserved in embedded font if the font is embedded -as a TrueType font or a CIDFontType2 CID-keyed font. -For all font embedded as a PostScript font (Type1C and CIDFontType0 -CID-keyed font), they are not preserved. -Only \code{Copyright} and \code{Notice} in the FontInfo dictionary are -preserved in this case. - -Some font vendors put different embedding restrictions for different -condition; e.g., font embedding might be not permitted for commercial -materials unless you acquire ``commercial license'' separately. -Please read EULA carefully before making decision on font usage. - -See, for examples, -\href{http://www.adobe.com/products/type/font-licensing/font-embedding-permissions.html}{Adobe's site on font embedding permissions} -for fonts in the Adobe Type Library. -Microsoft also has a -\href{http://www.microsoft.com/typography/RedistributionFAQ.mspx}{FAQ page on Font Redistribution}. - -For Japanese font in general, embedding permission tend to be somewhat -restrictive. Japanese users should read the statement regarding font -embedding from Japan Typography Association (in Japanese):\medskip +The full \code{ChangeLog} entries can be viewed via the web interface of the +\TeX\ Live SVN repository: +\medskip -\url{http://www.typography.or.jp/act/morals/moral4.html} +\url{http://www.tug.org/svn/texlive/trunk/Build/source/texk/dvipdfm-x} \medskip -\dvipdfmx\ does not support full embedding. Only subset embedding is supported. +There was an undocumented feature for supporting OpenType Layout but it was +dropped. Simplified support for the OpenType Layout was intorduced instead. + -\renewcommand{\refname}{Further Reading} +%\renewcommand{\refname}{Further Reading} +\renewcommand{\bibname}{Further Reading} \begin{thebibliography}{99} +\bibitem{DVIPDFM} ''\href{http://mirrors.ctan.org/dviware/dvipdfm/dvipdfm.pdf}% +{Dvipdfm User's Manual}'' written by Mark~A.~Wicks. \bibitem{ADOBE} Adobe's PDF References and a free copy of \lnum{ISO 32000-1:2008} standard are available from ``\href{http://www.adobe.com/devnet/pdf.html}{PDF Technology Center}'' @@ -1242,12 +2171,16 @@ on \href{http://www.adobe.com/devnet.html}{Adobe Developer Connection}. site: ``\href{http://www.microsoft.com/en-us/Typography/OpenTypeSpecification.aspx}% {OpenType Specification}''. +\bibitem{PNGSPEC} ''\href{https://www.w3.org/TR/2003/REC-PNG-20031110/}% +{Portable Network Graphics (PNG) Specification (Second Edition)}''. \bibitem{CHOF} An article regarding DVI specials: Jin-Hwan Cho, ``\href{http://www.tug.org/TUGboat/tb30-1/tb94cho.pdf}{DVI specials for PDF generation}'', TUGboat, 30(1):6-11, 2009. \end{thebibliography} -\section*{GNU Free Documentation License}\label{SEC:FDL} +\appendix + +\chapter{GNU Free Documentation License}\label{SEC:FDL} This document is distributed under the term of the GNU Free Documentation License. See, the attached file for copying conditions.% @@ -1263,8 +2196,8 @@ License. See, the attached file for copying conditions.% >> /Name /PushPin /C [0.8 0.2 0.2] - /T (Free Software Foundation, Inc.) - /Subj (GNU Free Documentation License) + /T (GNU Free Documentation License) + /Subj (GNU FDL) /Contents (Plain text version of the GNU Free Documentation License.) >> }}% |