% $Id: genops.tex,v 1.6 2020/01/01 23:12:25 karl Exp $ % This is part of the book TeX for the Impatient. % Copyright (C) 2003-2020 Paul W. Abrahams, Kathryn A. Hargreaves, Karl Berry. % See file fdl.tex for copying conditions. \input macros \chapter {Commands for \linebreak general operations} \chapterdef{general} This section covers \TeX's ^{programming features} and everything else that doesn't fit into the categories of commands in the previous chapters. For an explanation of the conventions used in this section, see \headcit{Descriptions of the commands}{cmddesc}. \begindescriptions %========================================================================== \section {Naming and modifying fonts} \begindesc \bix^^{fonts//naming and modifying} \cts font {} \aux\cts font {\ = \} \aux\cts font {\ = \ {\bt scaled} \} \aux\cts font {\ = \ {\bt at} \} \explain Used alone, the |\font| control sequence designates the current font. |\font| isn't a true command when it's used alone, since it then can appear only as an argument to another command. For the other three forms of |\font|, \ names a set of files that define a font. These forms of |\font| are commands. Each of these forms has two effects: {\tighten \olist \li It defines \ as a name that selects the font \, possibly magnified (see below). \li It causes \TeX\ to load the font ^{metrics file} (^{\tfmfile}) for \. \endolist }% end tighten \noindent The name of a font file usually indicates its design size. For example, |cmr10| indicates Computer Modern roman with a design size of $10$ points. The design size of a font is recorded in its metrics file. If neither |scaled| \ nor |at| \ is present, the font is used at its design size---the size at which it usually looks best. Otherwise, a magnified version of the font is loaded: \ulist \li If |scaled| \ is present, the font is magnified by a factor of $\hbox{\}/1000$. \li If |at| \ is present, the font is scaled to \ by magnifying it by $\hbox{\}/ds$, where $ds$ is the design size of \. \ and $ds$ are nearly always given in points. \endulist \noindent Magnifications of less than $1$ are possible; they reduce the size. You usually need to provide a shape file (\xref{shape}) for each magnification of a font that you load. However, some ^{device drivers} can utilize fonts that are resident in a printer. ^^{resident fonts} Such fonts don't need shape files. See \conceptcit{font} and \conceptcit{magnification} for further information. \example \font\tentt = cmtt10 \font\bigttfont = cmtt10 scaled \magstep2 \font\eleventtfont = cmtt10 at 11pt First we use {\tentt regular CM typewriter}. Then we use {\eleventtfont eleven-point CM typewriter}. Finally we use {\bigttfont big CM typewriter}. | \produces \font\regttfont = cmtt10 \font\bigttfont = cmtt10 scaled \magstep 2 \font\eleventtfont = cmtt10 at 11pt First we use {\regttfont regular CM typewriter}. Then we use {\eleventtfont eleven-point CM typewriter}. Finally we use {\bigttfont big CM typewriter}. \endexample \enddesc \begindesc \cts fontdimen {\ \\param{dimen}} \explain ^^{fonts//parameters of} These parameters specify various dimensions associated with the font named by the control sequence \ (as distinguished from the \ that names the font files). Values of these parameters are specified in the metrics file for \, but you can retrieve or change their values during a \TeX\ run. The numbers and meanings of the parameters are: \display{\halign{\hfil#\hfil\quad&#\hfil\cr \it Number&\it Meaning\cr \noalign{\vskip 1\jot}% 1&slant per point\cr 2&interword space\cr 3&interword stretch\cr 4&interword shrink\cr 5&x-height (size of |1ex|)\cr 6&quad width (size of |1em|)\cr 7&extra space\cr}} \noindent \TeX\ uses the slant per point for positioning accents. It uses the interword parameters for producing interword spaces (see |\spaceskip|, \xref\spaceskip) and the extra space parameter for the additional space after a period (see |\xspaceskip|, \xref\xspaceskip). The values of these parameters for the \plainTeX\ fonts are enumerated on \knuth{page~433}. Math symbol fonts have $15$ additional parameters, which we won't discuss here. Beware: assignments to these parameters are \emph{not} undone at the end of a group. If you want to change these parameters locally, you'll need to save and restore their original settings explicitly. \example Here's a line printed normally.\par \dimen0=\fontdimen2\font \fontdimen2\font=3\fontdimen2\font % triple word spacing \noindent Here's a really spaced-out line. \fontdimen2\font=\dimen0 | \produces Here's a line printed normally.\par \dimen0=\fontdimen2\font \fontdimen2\font=3\fontdimen2\font % triple word spacing \noindent Here's a really spaced-out line. \fontdimen2\font=\dimen0 \endexample \enddesc \begindesc \cts magnification {{\bt =} \} \cts mag {\param{number}} \explain \margin{{\tt\\mag} and {\tt\\magnification} combined.} An assignment to |\magnification| establishes the ``^{scale factor}'' $f$ that determines the \minref{magnification} ratio of your document \seeconcept{magnification}. The assignment to |\magni!-fication| must occur before the first page of your document has been shipped out. The assignment sets $f$ to \ and also sets |\hsize| and |\vsize| ^^|\hsize//set by {\tt\\magnification}| ^^|\vsize//set by {\tt\\magnification}| respectively to |6.5true in| and |8.9true in|, the values appropriate for an $8 \frac1/2$-% by-$11$-inch page. $f$ must be between $0$ and $32768$. The \minref{magnification} ratio of the document is $f/1000$. A scale factor of $1000$ provides unit magnification, i.e., it leaves the size of your document unchanged. It's customary to use powers of $1.2$ as scale factors, and most libraries of fonts are based on such factors. You can use the ^|\magstep| and ^|\magstephalf| commands to specify magnifications by these factors. |\magnification| is not a parameter. You can't use it to \emph{retrieve} the scale factor. If you write something like |\dimen0 = \mag!-nifi!-cation|, \TeX\ will complain about it. The |\mag| parameter contains the scale factor. Changing the value of |\mag| rescales the page dimensions, which is not usually what you want. Therefore it's usually better to change the magnification by assigning to |\magnification| rather than to |\mag|. \example \magnification = \magstep2 % magnify fonts by 1.44 (=1.2x1.2) | \endexample \enddesc \begindesc \cts magstep {\} \explain This command expands to the \minref{magnification} ratio needed to magnify everything in your document (other than |true| dimensions) by $1.2^r$, where $r$ is the value of \. \ must be between $0$ and $5$. \example \magnification = \magstep1 % Magnify by ratio of 1.2. | \endexample \enddesc \begindesc \cts magstephalf {} \explain This command expands to the \minref{magnification} ratio needed to magnify everything in your document (other than |true| dimensions) by $\sqrt{1.2}$, i.e., halfway between $1$ and $1.2$. \example \magnification = \magstephalf | \endexample \eix^^{fonts//naming and modifying} \enddesc %========================================================================== \section {Converting information to tokens} \subsection {Numbers} \begindesc \xrdef{convert} \bix^^{numbers//converting to characters} % \cts number {\} \explain This command produces the representation of a \minref{number} as a sequence of character \minref{token}s. The number can be either an explicit integer, a \ parameter, or a \ register. \example \number 24 \quad \count13 = -10000 \number\count13 | \produces \number 24 \quad \count13 = -10000 \number\count13 \endexample \enddesc \begindesc ^^{Roman numerals} \easy\cts romannumeral {\} \explain This command produces the roman numeral representation of a \minref{number} as a sequence of character \minref{token}s. The number can be either an explicit integer, a \ parameter, or a \ register. If the number is zero or negative, |\romannumeral| produces no tokens. \example \romannumeral 24 \quad (\romannumeral -16)\quad \count13 = 6000 \romannumeral\count13 | \produces \romannumeral 24 \quad (\romannumeral -16)\quad \count13 = 6000 \romannumeral\count13 \endexample \eix^^{numbers//converting to characters} \enddesc %========================================================================== \subsection {Environmental information} \begindesc ^^{time of day} \cts time {\param{number}} \explain \TeX\ sets this parameter to the number of minutes that have elapsed since midnight (of the current day). At noon, for instance, |\time| is $720$. This command and the next three make use of the time and date as recorded in your computer. \TeX\ retrieves them just once, at the beginning of your run, so |\time| at the end of the run always has the same value as |\time| at the beginning of the run (unless you've explicitly changed it). \enddesc \bix^^{date} \begindesc \cts day {\param{number}} \explain \TeX\ sets this parameter to the current day of the month. It is a number between $1$ and $31$. |\day| is set at the beginning of your run (see the comments on |\time| above). \enddesc \begindesc \cts month {\param{number}} \explain \TeX\ sets this parameter to the current month. It is a number between $1$ and $12$. |\month| is set at the beginning of your run (see the comments on |\time| above). \enddesc \begindesc \cts year {\param{number}} \explain \TeX\ sets this parameter to the current year ({\sc A.D.}). It is a number such as $1991$. |\year| is set at the beginning of your run (see the comments on |\time| above). \eix^^{date} \enddesc \begindesc ^^{version number} \cts fmtname {} \cts fmtversion {} \explain These commands produce the name and version number of the \TeX\ format, e.g., \minref{\plainTeX} or ^{\LaTeX}, that you're using. The |\fmtversion| string contains a long list of supported languages, so is omitted here. \example This book was produced with the \fmtname\ format. | \produces This book was produced with the \fmtname\ format. \endexample \enddesc \begindesc \cts jobname {} \explain This command produces the base name of the file with which \TeX\ was invoked. For example, if your main input file is |hatter.tex|, |\jobname| {\parfillskip=0pt\par\eject\noindent} will expand to |hatter|. |\jobname| is most useful when you're creating an auxiliary file to be associated with a document. ^^{auxiliary files} \example \newwrite\indexfile \openout\indexfile = \jobname.idx % For input file `hatter.tex', open index file `hatter.idx'. | \endexample\enddesc %========================================================================== \subsection {Values of variables} \begindesc \cts meaning {\} \explain ^^{tokens//showing the meaning of} This command produces the meaning of \. It is useful for diagnostic output. You can use the ^|\the| command (\xref\the) in a similar way to get information about the values of \minref{register}s and other \TeX\ entities. \example [{\tt \meaning\eject}] [\meaning\tenrm] [\meaning Y] | \produces [{\tt \meaning\eject}] [\meaning\tenrm] [\meaning Y] \endexample\enddesc \begindesc \cts string {\} \explain ^^{control sequences//converting to strings} This command produces the characters that form the name of \, including the \minref{escape character}. The escape character is represented by the current value of ^|\escapechar|. ^^{escape character//represented by \b\tt\\escapechar\e} \TeX\ gives the characters in the list a category code of $12$ (other). You can perform the reverse operation with the ^|\csname| command (\xref \csname), which turns a string into a control sequence. \example the control sequence {\tt \string\bigbreak} | \produces the control sequence {\tt \string\bigbreak} \endexample\enddesc \begindesc \cts escapechar {\param{number}} \explain This parameter specifies the \ascii\ code \minrefs{\ascii} of the character that \TeX\ uses to represent the \minref{escape character} ^^{escape character//represented by \b\tt\\escapechar\e} when it's converting a control sequence name to a sequence of character tokens. This conversion occurs when you use the |\string| command and also when \TeX\ is producing diagnostic messages. The default value of the escape character is $92$, the {\ascii} character code for a ^{backslash}. If |\escapechar| is not in the range $0$--$255$, \TeX\ does not include an escape character in the result of the conversion. \example \escapechar = `!! the control sequence {\tt \string\bigbreak} | \produces \escapechar = `! the control sequence {\tt \string\bigbreak} \endexample \enddesc \begindesc \cts fontname {\} \explain ^^{fonts//names of} This command produces the filename for \. The filename is the \ that was used to define \. \example \font\myfive=cmr5 [\fontname\myfive] | \produces \font\myfive=cmr5 [\fontname\myfive] \endexample \enddesc %========================================================================== \section {Grouping} \begindesc \bix^^{groups} % \cts begingroup {} \cts endgroup {} \explain These two commands begin and end a \minref{group}. A |\begingroup| does not match up with a right brace, nor an |\endgroup| with a left brace. \TeX\ treats |\begingroup| and |\endgroup| like any other \minref{control sequence} when it's scanning its input. In particular, you can define a \minref{macro} that contains a |\begingroup| but not an |\endgroup|, and conversely. ^^{macros//using \b\tt\\begingroup\e\ and \b\tt\\endgroup\e\ in} This technique is often useful when you're defining paired macros, one of which establishes an environment and the other of which terminates that environment. You can't, however, use |\begingroup| and |\endgroup| as substitutes for braces other than the ones that surround a group. \example \def\a{One \begingroup \it two } \def\enda{\endgroup four} \a three \enda | \produces \def\a{One \begingroup \it two } \def\enda{\endgroup four} \a three \enda \endexample \enddesc \begindesc \makecolumns 4/2: \easy% \ctsact { \xrdef{@lbrace} \cts bgroup {} \ctsact } \xrdef{@rbrace} \cts egroup {} \explain The left and right braces are commands that begin and end a \minref{group}. The |\bgroup| and |\egroup| \minref{control sequence}s are equivalent to `|{|' and `|}|', except that \TeX\ treats |\bgroup| and |\egroup| like any other \minref{control sequence} when it's scanning its input. |\bgroup| and |\egroup| can be useful when you're defining paired macros, one of which starts a brace-delimited construct (not necessarily a group) and the other one of which ends that construct. ^^{macros//using \b\tt\\bgroup\e\ and \b\tt\\egroup\e\ in} You can't define such macros using ordinary braces---if you try, your macro definitions will contain unmatched braces and will therefore be unacceptable to \TeX. Usually you should use these commands only when you can't use ordinary braces. \example Braces define the {\it boundaries\/} of a group. | \produces Braces define the {\it boundaries\/} of a group. \nextexample \def\a{One \vbox\bgroup} % You couldn't use { instead of \bgroup here because % TeX would not recognize the end of the macro \def\enda#1{{#1\egroup} two} % This one is a little tricky, since the \egroup actually % matches a left brace and the following right brace % matches the \bgroup. But it works!! \a \enda{\hrule width 1in} | \produces \def\a{One \vbox\bgroup} % You couldn't use { instead of \bgroup here because % TeX would not recognize the end of the macro \def\enda#1{{#1\egroup} two} % This one is a little tricky, since the \egroup actually % matches a left brace and the following right brace % matches the \bgroup. But it works! \a \enda{\hrule width 1in} \endexample \enddesc \begindesc \cts global {} \explain This command makes the following definition or \minref{assignment} \minref{global} \seeconcept{global} so that it becomes effective independent of \minref{group} boundaries. You can apply a |\global| prefix to any kind of definition or \minref{assignment}, including a \minref{macro} definition or a \minref{register} assignment. \example {\global\let\la = \leftarrow} $a \la b$ | \produces % for safety's sake we fake this one! \let\la = \leftarrow $a \la b$ \endexample \enddesc \begindesc \cts globaldefs {\param{number}} \explain This parameter controls whether or not \TeX\ takes definitions and other assignments to be \minref{global}: \ulist \li If |\globaldefs| is zero (as it is by default), a definition is global if and only if it is preceded by |\global| either explicitly or implicitly. (The ^|\gdef| and ^|\xdef| commands (\xref\gdef) have an implicit |\global| prefix). \li If |\globaldefs| is greater than zero, all assignments and definitions are implicitly prefixed by ^|\global|. \li If |\globaldefs| is less than zero, all ^|\global| prefixes are ignored. \endulist \enddesc \begindesc \margin{Order of {\tt\\aftergroup} and {\tt\\afterassignment} changed.} \cts aftergroup {\} \explain When \TeX\ encounters this command during input, it saves \. After the end of the current \minref{group}, it inserts \ back into the input and expands it. If a group contains several |\aftergroup|s, the corresponding tokens are \emph{all} inserted following the end of the group, in the order in which they originally appeared. The example that follows shows how you can use |\aftergroup| to postpone processing a token that you generate within a \minref{conditional test}. \example \def\neg{negative} \def\pos{positive} % These definitions are needed because \aftergroup applies % to a single token, not to a sequence of tokens or even % to a brace-delimited text. \def\arith#1{Is $#1>0$? \begingroup \ifnum #1>-1 Yes\aftergroup\pos \else No\aftergroup\neg\fi , it's \endgroup. } \arith 2 \arith {-1} | \produces \def\neg{negative} \def\pos{positive} % These definitions are needed because \aftergroup applies % to a single token, not a sequence of tokens or even % a group. \def\arith#1{Is $#1>0$? \begingroup \ifnum #1>-1 Yes\aftergroup\pos \else No\aftergroup\neg\fi , it's \endgroup. } \arith 2 \arith {-1} \endexample \eix^^{groups} \enddesc \begindesc \cts afterassignment {\} \explain When \TeX\ encounters this command it saves \ in a special place. After it next performs an \minref{assignment}, it inserts \ into the input and expands it. If you call |\afterassignment| more than once before an assignment, only the last call has any effect. One use of |\afterassignment| is in writing \minref{macro}s for commands intended to be written in the form of assignments, as in the example below. See \knuth{page~279} for a precise description of the behavior of |\afterassignment|. \example \def\setme{\afterassignment\setmeA\count255} \def\setmeA{$\number\count255\advance\count255 by 10 +10=\number\count255$} Some arithmetic: \setme = 27 % After expanding \setme, TeX sets \count255 to 27 and % then calls \setmeA. | \produces \def\setme{\afterassignment\setmeA\count255} \def\setmeA{$\number\count255\advance\count255 by 10 +10=\number\count255$} Some arithmetic: \setme = 27 % After expanding \setme, TeX sets \count255 to 27 and % then calls \setmeA. \endexample \enddesc %========================================================================== \section {Macros} %========================================================================== \subsection {Defining macros} \begindesc \bix^^{macros} \bix^^{macros//defining} \xrdef{mac1}% begin the section on macros % \cts def {\ \ \rqbraces{\}} \explain This command defines \ as a \minref{macro} with the specified \ and \. See \xrefpg{macro} for a full explanation of how to write a macro definition. \example \def\add#1+#2=?{#1+#2&= \count255=#1 \advance\count255 by #2 \number\count255\cr} $$\eqalign{ \add 27+9=? \add -5+-8=?}$$ | \dproduces \def\add#1+#2=?{#1+#2&= \count255=#1 \advance\count255 by #2 \number\count255\cr} $$\eqalign{ \add 27+9=? \add -5+-8=?}$$ \endexample \enddesc \begindesc \cts edef {\ \ \rqbraces{\}} \explain This command defines a macro in the same general way as |\def|. The difference is that \TeX\ expands the \ of an |\edef| immediately (but still without executing anything). Thus any definitions within the \ are expanded, but assignments and commands that produce things such as boxes and glue are left as is. For example, an |\hbox| command within the \ of an |\edef| remains as a command and is not turned into a box as \TeX\ is processing the definition. It isn't always obvious what's expanded and what isn't, but you'll find a complete list of expandable control sequences on \knuth{pages~212--215}. You can inhibit the expansion of a control sequence that would otherwise be expanded by using |\no!-expand| (\xref\noexpand). ^^|\noexpand| You can postpone the expansion of a control sequence by using ^|\expandafter| (\xref\expandafter). The |\write|, |\message|, |\errmessage|, |\wlog|, and |\csname| commands expand their token lists using the same rules that |\edef| uses to expand its replacement text. ^^|\write//expanded by {\tt\\edef} rules| ^^|\message//expanded by {\tt\\edef} rules| ^^|\errmessage//expanded by {\tt\\edef} rules| ^^|\wlog//expanded by {\tt\\edef} rules| ^^|\csname//expanded by {\tt\\edef} rules| \example \def\aa{xy} \count255 = 1 \edef\bb{w\ifnum \count255 > 0\aa\fi z} % equivalent to \def\bb{wxyz} \def\aa{} \count255 = 0 % leaves \bb unaffected \bb | \produces \def\aa{xy} \count255 = 1 \edef\bb{w\ifnum \count255 > 0\aa\fi z} % equivalent to \def\bb{wxyz} \def\aa{} \count255 = 0 % leaves \bb unaffected \bb \endexample \enddesc \begindesc \cts gdef {\ \ \rqbraces{\}} \explain This command is equivalent to |\global\def|. \enddesc \begindesc \cts xdef {\ \ \rqbraces{\}} \explain This command is equivalent to |\global\edef|. \enddesc \begindesc \cts long {} \explain This command is used as a prefix to a \minref{macro} definition. It tells \TeX\ that the arguments to the macro are permitted to include |\par| tokens (\xref{\@par}), which normally indicate the end of a paragraph. ^^|\par//in macro arguments| If \TeX\ tries to expand a macro defined without |\long| and any of the macro's arguments include a |\par| token, \TeX\ will complain about a runaway argument. The purpose of this behavior is to provide you with some protection against unterminated macro arguments. |\long| gives you a way of bypassing the protection. \example \long\def\aa#1{\par\hrule\smallskip#1\par\smallskip\hrule} \aa{This is the first line.\par This is the second line.} % without \long, TeX would complain | \produces \medskip \long\def\aa#1{\par\hrule\smallskip#1\par\smallskip\hrule} \aa{This is the first line.\par This is the second line.} % without \long, TeX would complain \endexample \enddesc \begindesc \cts outer {} \explain \null ^^{outer} This command is used as a prefix to a \minref{macro} definition. It tells \TeX\ that the macro is outer (\xref{outer}) and cannot be used in certain contexts. If the macro is used in a forbidden context, \TeX\ will complain. \example \outer\def\chapterhead#1{% \eject\topglue 2in \centerline{\bf #1}\bigskip} % Using \chapterhead in a forbidden context causes an % error message. | \endexample \enddesc \begindesc \cts chardef {\=\} \explain ^^{characters//defined by \b\tt\\chardef\e} This command defines \ to be \. Although |\chardef| is most often used to define characters, you can also use it to give a name to a number in the range $0$--$255$ even when you aren't using that number as a character code. \example \chardef\percent = `\% 21\percent, {\it 19\percent} % Get the percent character in roman and in italic | \produces \chardef\percent = `\% 21\percent, {\it 19\percent} % You'll get the percent character in roman and in italic \endexample \enddesc \begindesc ^^{math characters} ^^{mathcodes} \cts mathchardef {\=\} \explain This command defines \ as a math character with the given \minrefs{mathcode}\. The control sequence will only be legal in math mode. \example \mathchardef\alphachar = "010B % As in plain TeX. $\alphachar$ | \produces \mathchardef\alphachar = "010B % As in plain TeX. $\alphachar$ \endexample \eix^^{macros//defining} \enddesc %========================================================================== \subsection {Other definitions} \begindesc \cts let {\ = \} \explain ^^{control sequences//defining with \b\tt\\let\e} \minrefs{token} This command causes \ to acquire the current meaning of \. Even if you redefine \ later, the meaning of \ will not change. Although \ is most commonly a control sequence, it can also be a \minref{character} token. \enddesc \begindesc \cts futurelet {\ \ \} \explain This command tells \TeX\ to make \ the meaning of \ (as would be done with |\let|), and then to process \ and \ normally. |\futurelet| is useful at the end of macro definitions because it gives you a way of looking beyond the token that \TeX\ is about to process before it processes it. \example \def\predict#1{\toks0={#1}\futurelet\next\printer} % \next will acquire the punctuation mark after the % argument to \predict \def\printer#1{A \punc\ lies ahead for \the\toks0. } \def\punc{% \ifx\next;semicolon\else \ifx\next,comma\else ``\next''\fi\fi} \predict{March}; \predict{April}, \predict{July}/ | \produces \def\predict#1{\toks0={#1}\futurelet\next\printer} \def\printer#1{A \punc\ lies ahead for \the\toks0. } \def\punc{% \ifx\next;semicolon\else \ifx\next,comma\else ``\next''\fi\fi } \predict{March}; \predict{April}, \predict{July}/ \endexample \enddesc \begindesc \cts csname {\ {\bt \\endcsname}} \xrdef{\endcsname} \explain This command produces a control sequence from \. It provides a way of synthesizing control sequences, including ones that you can't normally write. \ can itself include control sequences; it is expanded in the same way as the replacement text of an |\edef| definition (\xref\edef). If the final expansion yields anything that isn't a character, \TeX\ will complain. |\csname| goes from a list of tokens to a control sequence; you can go the other way with ^|\string| \ctsref\string. \example \def\capTe{Te} This book purports to be about \csname\capTe X\endcsname. | \produces \def\capTe{Te} This book purports to be about \csname\capTe X\endcsname. \endexample \enddesc %========================================================================== \subsection {Controlling expansion} \begindesc \bix^^{macros//controlling expansion of} \cts expandafter {\ \} \explain This command tells \TeX\ to expand \ according to its rules for \minref{macro} expansion \emph{after} it has expanded \ by one level. It's useful when \ is something like `|{|' ^^|{//with {\tt\\expandafter}| or ^|\string| that inhibits expansion of \, but you want to expand \ nevertheless. \example \def\aa{xyz} \tt % Use this font so `\' prints that way. [\string\aa] [\expandafter\string\aa] [\expandafter\string\csname TeX\endcsname] | \produces \def\aa{xyz} \tt [\string\aa] [\expandafter\string\aa] [\expandafter\string\csname TeX\endcsname] \endexample \enddesc \begindesc \cts noexpand {\} \explain This command tells \TeX\ to suppress expansion of \ if \ is a \minref{control sequence} that can be expanded. If \ can't be expanded, e.g., it's a letter, \TeX\ acts as though the |\noexpand| wasn't there and processes \ normally. In other words the expansion of `|\noexpand|\' is simply \ no matter what \ happens to be. \example \def\bunny{rabbit} \edef\magic{Pull the \noexpand\bunny\ out of the hat!! } % Without \noexpand, \bunny would always be replaced % by `rabbit' \let\oldbunny=\bunny \def\bunny{lagomorph} \magic \let\bunny=\oldbunny \magic | \produces \def\bunny{rabbit} \edef\magic{Pull the \noexpand\bunny\ out of the hat! } % Without \noexpand, \bunny would always be replaced % by `rabbit' \let\oldbunny=\bunny \def\bunny{lagomorph} \magic \let\bunny=\oldbunny \magic \endexample \enddesc \begindesc \cts the {\} \explain This command generally expands to a list of \minref{character} tokens that represents \. \ can be any of the following: \ulist\compact \li a \TeX\ \minref{parameter}, e.g., |\parindent| or |\deadcycles| ^^{parameters//using \b\tt\\the\e\ with} \li a \minref{register}, e.g., |\count0| ^^{registers//with \b\tt\\the\e} \margin{Item for special registers removed} \li a code associated with an input character, e.g., |\catcode`(| \li a font parameter, e.g., |\fontdimen3\sevenbf| \li the ^|\hyphenchar| or ^|\skewchar| of a font, e.g., |\skewchar\teni| \li ^|\lastpenalty|, ^|\lastskip|, or ^|\lastkern| (values derived from the last item on the current horizontal \minrefs{horizontal list} or \minref{vertical list}) \li a control sequence defined by ^|\chardef| or ^|\mathchardef| \endulist \noindent In addition, |\the| can expand to noncharacter tokens in the following two cases: \ulist\compact \li |\the| \, which expands to the most recently defined control sequence that selects the same font as the control sequence \ \li |\the| \, which expands to a copy of the value of the variable, e.g., |\the\everypar| \endulist See \knuth{pages~214--215} for a more detailed description of what |\the| does in various cases. \example The vertical size is currently \the\vsize. The category code of `(' is \the\catcode `(. | \produces The vertical size is currently \the\vsize. The category code of `(' is \the\catcode `(. \endexample \enddesc {\tighten \see \headcit{Converting information to tokens}{convert}, |\showthe| (\xref\showthe). \par} \eix^^{macros//controlling expansion of} %========================================================================== \subsection {Conditional tests} \begindesc \xrdef{conds} \bix^^{conditional tests} % \ctspecial if {\ \}\ctsxrdef{@if} \explain {\emergencystretch=1em This command tests if \ and \ have the same \minref{character} code, independent of their \minref{category code}s. Before performing the test, \TeX\ expands tokens following the |\if| until it obtains two tokens that can't be expanded further. These two tokens become \ and \. The expansion includes replacing a control sequence |\let| equal to a character token by that character token. A \minref{control sequence} that can't be further expanded is considered to have character code $256$.\par} \example \def\first{abc} \if\first true\else false\fi; % ``c'' is left over from the expansion of \first. % It lands in the unexecuted ``true'' part. \if a\first\ true\else false\fi; % Here ``bc'' is left over from the expansion of \first \if \hbox\relax true\else false\fi % Unexpandable control sequences test equal with ``if'' | \produces \def\first{abc} \if\first true\else false\fi; % ``c'' is left over from the expansion of \first. % It lands in the unexecuted ``true'' part. \if a\first\ true\else false\fi; % Here ``bc'' is left over from the expansion of \first \if \hbox\relax true\else false\fi % Unexpandable control sequences test equal with ``if'' \endexample \enddesc \begindesc \ctspecial ifcat {\ \}\ctsxrdef{@ifcat} \explain ^^{category codes//testing} This command tests if \ and \ have the same \minref{category code}. Before performing the test, \TeX\ expands tokens following the |\ifcat| until it obtains two tokens that can't be expanded further. These two tokens become \ and \. The expansion includes replacing a control sequence |\let| equal to a character token by that character token. A \minref{control sequence} that can't be further expanded is considered to have category code $16$. \example \ifcat axtrue\else false\fi; \ifcat ]}true\else false\fi; \ifcat \hbox\day true\else false\fi; \def\first{12345} \ifcat (\first true\else false\fi % ``2345'' lands in the true branch of the test | \produces \ifcat axtrue\else false\fi; \ifcat ]}true\else false\fi; \ifcat \hbox\day true\else false\fi; \def\first{12345} \ifcat (\first true\else false\fi % ``2345'' lands in the true branch of the test \endexample \enddesc \begindesc \ctspecial ifx {\ \}\ctsxrdef{@ifx} \explain This command tests if \ and \ agree. Unlike |\if| and |\ifcat|, |\ifx| does \emph{not} expand the tokens following |\ifx|, so \ and \ are the two tokens immediately after |\ifx|. There are three cases: \olist \li If one token is a \minref{macro} and the other one isn't, the tokens don't agree. \li If neither token is a macro, the tokens agree if: \olist \li both tokens are characters (or control sequences denoting characters) and their \minref{character} codes and \minref{category code}s agree, or \li both tokens refer to the same \TeX\ command, font, etc. \endolist \li If both tokens are macros, the tokens agree if: \olist\compact \li their ``first level'' expansions, i.e., their replacement texts, are identical, and \li they have the same status with respect to ^|\long| (\xref\long) and ^|\outer| (\xref\outer). \endolist Note in particular that \emph{any two undefined control sequences agree}. \endolist \noindent This test is generally more useful than |\if|. \example \ifx\alice\rabbit true\else false\fi; % true since neither \rabbit nor \alice is defined \def\a{a}% \ifx a\a true\else false\fi; % false since one token is a macro and the other isn't \def\first{\a}\def\second{\aa}\def\aa{a}% \ifx \first\second true\else false\fi; % false since top level expansions aren't the same \def\third#1:{(#1)}\def\fourth#1?{(#1)}% \ifx\third\fourth true\else false\fi % false since parameter texts differ | \produces \ifx\alice\rabbit true\else false\fi; % true since neither \rabbit nor \alice is defined \def\a{a}% \ifx a\a true\else false\fi; % false since one token is a macro and the other isn't \def\first{\a}\def\second{\aa}\def\aa{a}% \ifx \first\second true\else false\fi; % false since top level expansions aren't the same \def\third#1:{(#1)}\def\fourth#1?{(#1)}% \ifx\third\fourth true\else false\fi % false since parameter texts differ \endexample \enddesc \begindesc \ctspecial ifnum {\ \ \}\ctsxrdef{@ifnum} \explain ^^{numbers//comparing} This command tests if \ and \ satisfy \, which must be either `|<|', `|=|', or `|>|'. The numbers can be constants such as |127|, count registers such as |\pageno| or |\count22|, or numerical parameters such as |\hbadness|. Before performing the test, \TeX\ expands tokens following the |\ifnum| until it obtains a sequence of tokens having the form \ \ \, followed by a token that can't be part of \. \example \count255 = 19 \ifnum \count255 > 12 true\else false\fi | \produces \count255 = 19 \ifnum \count255 > 12 true\else false\fi \endexample \enddesc \begindesc \ctspecial ifodd {\}\ctsxrdef{@ifodd} \explain ^^{numbers//testing for odd/even} This command tests if \ is odd. Before performing the test, \TeX\ expands tokens following the |\ifodd| until it obtains a sequence of tokens having the form \, followed by a token that can't be part of \. \example \count255 = 19 \ifodd 5 true\else false\fi | \produces \ifodd 5 true\else false\fi \endexample \enddesc \begindesc \ctspecial ifdim {\ \ \}\ctsxrdef{@ifdim} \explain ^^{dimensions//comparing} This command tests if \ and \ satisfy \, which must be either `|<|', `|=|', or `|>|'. The dimensions can be constants such as |1in|, dimension registers such as |\dimen6|, or dimension parameters such as |\parindent|. Before performing the test, \TeX\ expands tokens following the |\ifdim| until it obtains a sequence of tokens having the form \ \ \, followed by a token that can't be part of \. \example \dimen0 = 1000pt \ifdim \dimen0 > 3in true\else false\fi | \produces \dimen0 = 1000pt \ifdim \dimen0 > 3in true\else false\fi \endexample \enddesc \begindesc \ctspecial ifhmode {}\ctsxrdef{@ifhmode} \ctspecial ifvmode {}\ctsxrdef{@ifvmode} \ctspecial ifmmode {}\ctsxrdef{@ifmmode} \ctspecial ifinner {}\ctsxrdef{@ifinner} \explain ^^{horizontal mode//testing for} ^^{vertical mode//testing for} ^^{math mode//testing for} ^^{internal mode//testing for} These commands test what \minref{mode} \TeX\ is in: \ulist \li |\ifhmode| is true if \TeX\ is in ordinary or restricted horizontal mode. \li |\ifvmode| is true if \TeX\ is in ordinary or internal vertical mode. \li |\ifmmode| is true if \TeX\ is in text math or display math mode. \li |\ifinner| is true if \TeX\ is in an ``internal'' mode: restricted horizontal, internal vertical, or text math. \endulist \example \def\modes{{\bf \ifhmode \ifinner IH\else H\fi \else\ifvmode \ifinner \hbox{IV}\else \hbox{V}\fi \else\ifmmode \hbox{M}\else error\fi\fi\fi}} Formula $\modes$; then \modes, \hbox{next \modes\ and \vbox{\modes}}. \par\modes | \produces \def\modes{{\bf \ifhmode \ifinner IH\else H\fi \else\ifvmode \ifinner \hbox{IV}\fi \else\ifmmode \hbox{M}\else error\fi\fi\fi}} Formula $\modes$; then \modes, \hbox{next \modes\ and \vbox{\modes}}. \par\noindent{\bf V} % sorry folks, we have to fake this one \endexample \enddesc \begindesc \ctspecial ifhbox {\}\ctsxrdef{@ifhbox} \ctspecial ifvbox {\}\ctsxrdef{@ifvbox} \ctspecial ifvoid {\}\ctsxrdef{@ifvoid} \explain ^^{hboxes//testing for} ^^{vboxes//testing for} ^^{boxes//testing if void} These commands test the contents of the box register numbered \. Let \ be $n$. Then: \ulist \li |\ifhbox| is true if |\box|$\,n$ is an \minref{hbox}. \li |\ifvbox| is true if |\box|$\,n$ is an \minref{vbox}. \li |\ifvoid| is true if |\box|$\,n$ is void, i.e, doesn't have a box in it. \endulist \example \setbox0 = \vbox{} % empty but not void \setbox1 = \hbox{a} \setbox2 = \box1 % makes box1 void \ifvbox0 true\else false\fi; \ifhbox2 true\else false\fi; \ifvoid1 true\else false\fi | \produces \setbox0 = \vbox{} \setbox1 = \hbox{a} \setbox2 = \box1 % empties box1 \ifvbox0 true\else false\fi; \ifhbox2 true\else false\fi; \ifvoid1 true\else false\fi \endexample \enddesc \begindesc \ctspecial ifeof {\}\ctsxrdef{@ifeof} \explain ^^{end of file, testing for} \minrefs{file} This command tests an input stream for end of file. It is true if input stream \ has not been opened, or has been opened and the associated file has been entirely read in (or doesn't exist). \enddesc \begindesc \ctspecial ifcase {\\ {\bt \\or }\ {\bt \\or} $\ldots$ {\bt \\or} \\hfil\break \hglue 3pc{\bt \\else} \ {\bt \\fi}} \ctsxrdef{@ifcase} \ctsxrdef{@or} \explain ^^{case testing} This command introduces a test with numbered multiple cases. If \ has the value $k$, \TeX\ will expand \ if it exists, and \ if it doesn't. You can omit the |\else|---% in this case, \TeX\ won't expand anything if none of the cases are satisfied. \example \def\whichday#1{\ifcase #1\or Sunday\or Monday% \or Tuesday\or Wednesday\or Thursday\or Friday% \or Saturday\else Nonday\fi \ is day \##1. } \whichday2 \whichday3 \whichday9 | \produces \def\whichday#1{\ifcase #1\or Sunday\or Monday% \or Tuesday\or Wednesday\or Thursday\or Friday% \or Saturday\else Nonday\fi \ is day \##1. } \whichday2 \whichday3 \whichday9 \endexample \enddesc \begindesc \ctspecial iftrue {}\ctsxrdef{@iftrue} \ctspecial iffalse {}\ctsxrdef{@iffalse} \explain These commands are equivalent to tests that are always true or always false. The main use of these commands is in defining macros that keep track of the result of a test. \example \def\isbigger{\let\bigger=\iftrue} \def\isnotbigger{\let\bigger=\iffalse} % These \let's MUST be buried in macros!! If they aren't, % TeX erroneously tries to match them with \fi. \def\test#1#2{\ifnum #1>#2 \isbigger\else\isnotbigger\fi} \test{3}{6} \bigger$3>6$\else$3\le6$\fi | \produces \def\isbigger{\let\bigger=\iftrue} \def\isnotbigger{\let\bigger=\iffalse} % These \let's MUST be buried in macros! % If they aren't, TeX erroneously tries to match them with \fi \def\test#1#2{\ifnum #1>#2 \isbigger\else\isnotbigger\fi} \test{3}{6} \bigger$3>6$\else$3\le6$\fi \endexample \enddesc \begindesc \ctspecial else {} \ctsxrdef{@else} \explain This command introduces the ``false'' alternative of a conditional test. \enddesc \begindesc \ctspecial fi {} \ctsxrdef{@fi} \explain This command ends the text of a conditional test. \enddesc \begindesc \ctspecial newif {{\bt \\if}\}\ctsxrdef{@newif} \explain This command names a trio of control sequences with names |\alpha!-true|, |\alphafalse|, and |\ifalpha|, where |alpha| is \. You can use them to define your own tests by creating a logical variable that records true\slash false information: \ulist\compact \li |\alphatrue| sets the logical variable |alpha| true. \li |\alphafalse| sets the logical variable |alpha| false \li |\ifalpha| is a conditional test that is true if the logical variable |alpha| is true and false otherwise. \endulist The logical variable |alpha| doesn't really exist, but \TeX\ behaves as though it did. After |\newif\ifalpha|, the logical variable is initially false. |\newif| is an outer command, so you can't use it inside a macro definition. \example \newif\iflong \longtrue \iflong Rabbits have long ears. \else Rabbits don't have long ears.\fi | \produces \newif\iflong \longtrue \iflong Rabbits have long ears.\else Rabbits don't have long ears.\fi \endexample \eix^^{conditional tests} \enddesc %========================================================================== \subsection {Repeated actions} {\def\test{{\bt \\if}$\Omega$}% \begindesc \bix^^{repeated actions} \bix^^{loops} \cts loop {$\alpha$ {\test} $\beta$ {\bt \\repeat}} \ctspecial repeat {}\ctsxrdef{@repeat} \explain These commands provide a looping construct for \TeX. Here $\alpha$ and $\beta$ are arbitrary sequences of commands and \test\ is any of the conditional tests described in \headcit{Conditional tests}{conds}. The |\repeat| replaces the |\fi| corresponding to the test, so you must not write an explicit |\fi| to terminate the test. Nor, unfortunately, can you associate an |\else| with the test. If you want to use the test in the opposite sense, you need to rearrange the test or define an auxiliary test with |\newif| (see above) and use that test in the sense you want (see the second example below). \TeX\ expands |\loop| as follows: \olist \li $\alpha$ is expanded. \li {\test} is performed. If the result is false, the loop is terminated. \li $\beta$ is expanded. \li The cycle is repeated. \endolist \example \count255 = 6 \loop \number\count255\ \ifnum\count255 > 0 \advance\count255 by -1 \repeat | \produces \count255 = 6 \loop \number\count255\ \ifnum\count255 > 0 \advance\count255 by -1 \repeat \nextexample \newif\ifnotdone % \newif uses \count255 in its definition \count255=6 \loop \number\count255\ \ifnum\count255 < 1 \notdonefalse\else\notdonetrue\fi \ifnotdone \advance\count255 by -1 \repeat | \produces \newif\ifnotdone \count255=6 \loop \number\count255\ \ifnum\count255 < 1 \notdonefalse\else\notdonetrue\fi \ifnotdone \advance\count255 by -1 \repeat % \eix^^{repeated actions} \eix^^{loops} % \endexample \enddesc } % end scope of definition of \test %========================================================================== \subsection {Doing nothing} \begindesc \cts relax {} \explain This command tells \TeX\ to do nothing. It's useful in a context where you need to provide a command but there's nothing that you want \TeX\ to do. \example \def\medspace{\hskip 12pt\relax} % The \relax guards against the possibility that % The next tokens are `plus' or `minus'. | \endexample \enddesc \begindesc \cts empty {} \explain This command expands to no tokens at all. It differs from |\relax| in that it disappears after macro expansion. % \xrdef{mac2}% end the section on macros \eix^^{macros} \enddesc %========================================================================== \section {Registers} %========================================================================== \subsection {Using registers} \begindesc \bix^^{registers} \makecolumns 11/2: \cts count {\ {\bt =} \} \cts dimen {\ {\bt =} \} \cts skip {\ {\bt =} \} \cts muskip {\ {\bt =} \} \cts toks {\ {\bt =} \} \aux\cts toks {\ {\bt =} \rqbraces{\}} \aux\cts count {\} \aux\cts dimen {\} \aux\cts skip {\} \aux\cts muskip {\} \aux\cts toks {\} \explain ^^{assignments//of registers} The first six commands listed here assign something to a register. The |=|'s in the assignments are optional. The remaining five control sequences are not true commands because they can only appear as part of an argument. They yield the contents of the specified register. Although you can't use these control sequences by themselves as commands in text, you can use ^|\the| to convert them to text so that you can typeset their values. You can name and reserve registers with the ^|\newcount| command and its relatives (\xref{\@newcount}). Using these commands is a safe way to obtain registers that are known not to have any conflicting usage. ^^{count registers} A |\count| register contains an integer, which can be either positive or negative. Integers can be as large as you're ever likely to need them to be.\footnote {Here's the only exercise in this book: find out what's the largest integer that \TeX\ will accept.} \TeX\ uses count registers $0$--$9$ to keep track of the page number (see \knuth{page~119}). |\count255| is the only count register available for use without a reservation. \example \count255 = 17 \number\count255 | \produces \count255 = 17 \number\count255 \endexample \medskip\noindent ^^{dimension registers} A |\dimen| register contains a dimension. Registers |\dimen0| through |\dimen9| and |\dimen255| are available for scratch use. \example \dimen0 = 2.5in \hbox to \dimen0{$\Leftarrow$\hfil$\Rightarrow$} | \produces \dimen0 = 2.5in \hbox to \dimen0{$\Leftarrow$\hfil$\Rightarrow$} \doruler{\8\8\8}3{in} \endexample \medskip\noindent ^^{skip registers} A |\skip| register contains the dimensions of glue. Unlike a |\dimen| register, it records an amount of shrink and stretch as well as a natural size. Registers |\skip0| through |\skip9| and |\skip255| are available for use without a reservation. \example \skip2 = 2in $\Rightarrow$\hskip \skip2 $\Leftarrow$ | \produces \skip2 = 2in $\Rightarrow$\hskip \skip2 $\Leftarrow$\par \noindent\hphantom{$\Rightarrow$}\ruler{\8\8}2{in} \endexample \medskip\noindent ^^{muskip registers} A |\muskip| register is like a |\skip| register, but the glue in it is always measured in ^|mu| \seeconcept{mathematical unit}. The size of a |mu| depends on the current font. For example, it's usually a little smaller in a subscript than in ordinary text. Registers |\muskip0| through |\muskip9| and |\muskip255| are available for use without a reservation. \example \muskip0 = 24mu % An em and a half, no stretch or shrink. $\mathop{a \mskip\muskip0 b}\limits^{a \mskip\muskip0 b}$ % Note the difference in spacing. | \produces \muskip0 = 24mu % an em and a half $\mathop{a \mskip\muskip0 b}\limits^{a \mskip\muskip0 b}$ % Note the difference in spacing \endexample \medskip\noindent ^^{token registers} You can assign either a token variable (a register or a parameter) or a token list to a |\toks| register. When you assign a token list to a token register, the tokens in the token list are \emph{not} expanded. Once the tokens in a token list have been inserted into text using ^|\the|, they are expanded just like tokens that were read in directly. They have the category codes that they received when \TeX\ first saw them in the~\hbox{input}. \example \toks0 = {the \oystereaters\ were at the seashore} % This assignment doesn't expand \oystereaters. \def\oystereaters{Walrus and Carpenter} \toks1 = \toks0 % the same tokens are now in \toks0 and \toks1 Alice inquired as to whether \the\toks1. | \produces \toks0 = {the \oystereaters\ were at the seashore} % This assignment doesn't expand \oystereaters \def\oystereaters{Walrus and Carpenter} \toks1 = \toks0 % the same tokens are now in \toks0 and \toks1 Alice inquired as to whether \the\toks1. \endexample \enddesc \begindesc \cts maxdimen {} \explain ^^{dimensions//maximum} This control sequence yields a \ that is the largest dimension acceptable to \TeX\ (nearly 18 feet). It is not a true command because it can only appear as part of an argument to another command. \example \maxdepth = \maxdimen % Remove restrictions on \maxdepth. | \endexample \enddesc \see |\advance| (\xref\advance), |\multiply|, |\divide| (\xref\divide), |\set!-box|, |\box| (\xref\box). %========================================================================== \subsection {Naming and reserving registers, etc.} \begindesc \bix^^{registers//reserving} \makecolumns 11/2: \ctspecial newcount \ctsxrdef{@newcount} \ctspecial newdimen \ctsxrdef{@newdimen} \ctspecial newskip \ctsxrdef{@newskip} \ctspecial newmuskip \ctsxrdef{@newmuskip} \ctspecial newtoks \ctsxrdef{@newtoks} \ctspecial newbox \ctsxrdef{@newbox} \ctspecial newread \ctsxrdef{@newread} \ctspecial newwrite \ctsxrdef{@newwrite} \ctspecial newfam \ctsxrdef{@newfam} \ctspecial newinsert \ctsxrdef{@newinsert} \ctspecial newlanguage \ctsxrdef{@newlanguage} \explain These commands reserve and name an entity of the indicated type: \ulist \li |\newcount|, |\newdimen|, |\newskip|, |\newmuskip|, |\newtoks|, |\newbox| each reserve a \minref{register} of the indicated type. ^^{count registers//reserved by \b\tt\\newcount\e} ^^{dimension registers//reserved by \b\tt\\newdimen\e} ^^{skip registers//reserved by \b\tt\\newskip\e} ^^{muskip registers//reserved by \b\tt\\newmuskip\e} ^^{token registers//reserved by \b\tt\\newtoks\e} ^^{box registers//reserved by \b\tt\\newbox\e} \li |\newread| and |\newwrite| reserve an input stream and an output stream \minrefs{input stream}\minrefs{output stream} respectively. ^^{input streams//reserved by \b\tt\\newread\e} ^^{output streams//reserved by \b\tt\\newwrite\e} \li |\newfam| reserves a \minref{family} of math fonts. ^^{family//reserved by \b\tt\\newfam\e} \li |\newinsert| reserves an insertion type. (Reserving an insertion type involves reserving several different registers.) ^^{insertions//numbers reserved by \b\tt\\newinsert\e} \li |\newlanguage| reserves a set of hyphenation patterns. \endulist You should use these commands whenever you need one of these entities, other than in a very local region, in order to avoid numbering conflicts. There's an important difference among these commands: \ulist \li The control sequences defined by |\newcount|, |\newdimen|, |\newskip|, |\newmuskip|, and |\newtoks| each designate an entity of the appropriate type. For instance, after the command: \csdisplay \newdimen\listdimen | the control sequence |\listdimen| can be used as a dimension. \li The control sequences defined by |\newbox|, |\newread|, |\newwrite|, |\newfam|, |\newinsert|, and |\newlanguage| each evaluate to the \emph{number} of an entity of the appropriate type. For instance, after the command: \csdisplay \newbox\figbox | the control sequence |\figbox| must be used in conjunction with a |\box|-like command, e.g.: \csdisplay \setbox\figbox = \vbox{!dots} | \endulist \enddesc \begindesc \cts countdef {\ {\bt =} \} \cts dimendef {\ {\bt =} \} \cts skipdef {\ {\bt =} \} \cts muskipdef {\ {\bt =} \} \cts toksdef {\ {\bt =} \} \explain These commands define \ to refer to the \minref{register} of the indicated category whose number is \. Normally you should use the commands in the previous group (|\newcount|, etc.) in preference to these commands in order to avoid numbering conflicts. The commands in the previous group are defined in terms of the commands \hbox{in this group}. \example \countdef\hatters = 19 % \hatters now refers to \count19 \toksdef\hares = 200 % \hares now refers to \toks200 | \endexample \enddesc \see |\newif| (\xref{\@newif}), |\newhelp| (\xref{\@newhelp}). \eix^^{registers//reserving} %========================================================================== \subsection {Doing arithmetic in registers} \begindesc \bix^^{arithmetic} \bix^^{registers//arithmetic in} % \cts advance {\ {\bt by} \} \aux\cts advance {\ {\bt by} \} \aux\cts advance {\ {\bt by} \} \aux\cts advance {\ {\bt by} \} \explain This command adds a compatible quantity to a register. For \ or \ all three components (natural value, stretch, and shrink) \minrefs{glue} are added. Any of the quantities can be negative. For purposes of these calculations (and other assignments as well), \ can be converted to a \ by dropping the stretch and shrink, and a \ can be converted to a \ by taking its value in scaled points \seeconcept{dimension}. You can omit the word |by| in these commands---\TeX\ will understand them anyway. \example \count0 = 18 \advance\count0 by -1 \number\count0\par \skip0 = .5in \advance\skip0 by 0in plus 1in % add stretch \hbox to 2in{a\hskip\skip0 b} | \produces \count0 = 18 \advance\count0 by -1 \number\count0\par \skip0 = .5in \advance\skip0 by 0in plus 1in % add stretch \hbox to 2in{a\hskip\skip0 b} \doruler{\8\8}2{in} \endexample \enddesc \begindesc \cts multiply {\ {\bt by} \} \cts divide {\ {\bt by} \} \explain These commands multiply and divide the value in \ by \ (which can be negative). The register can be a ^|\count|, ^|\dimen|, ^|\skip|, or ^|\muskip| register. For a ^|\skip| or ^|\muskip| register (\xref\skip), all three components of the \minref{glue} in the register are modified. You can omit the word |by| in these commands---\TeX\ will understand them anyway. You can also obtain a multiple of a \ by preceding it by a \ \minrefs{number} or decimal constant, e.g., |-2.5\dimen2|. You can also use this notation for \, but watch out---the result is a \, not \. Thus |2\baselineskip| yields a \ that is twice the natural size of |\baselineskip|, with no stretch or shrink. \example \count0 = 9\multiply \count0 by 8 \number\count0 ; \divide \count0 by 12 \number\count0 \par \skip0 = 20pt plus 2pt minus 3pt \multiply \skip0 by 3 Multiplied value of skip0 is \the\skip0.\par \dimen0 = .5in \multiply\dimen0 by 6 \hbox to \dimen0{a\hfil b} | \produces \count0 = 9\multiply \count0 by 8 \number\count0 ; \divide \count0 by 12 \number\count0 \par \skip0 = 20pt plus 2pt minus 3pt \multiply \skip0 by 3 Multiplied value of skip0 is \the\skip0.\par \dimen0 = .5in \multiply\dimen0 by 6 \hbox to \dimen0{a\hfil b} \doruler{\8\8\8}3{in} \endexample \eix^^{arithmetic} \eix^^{registers//arithmetic in} \eix^^{registers} \enddesc %========================================================================== \section {Ending the job} \begindesc ^^{ending the job} \easy\ctspecial bye \ctsxrdef{@bye} \explain This command tells \TeX\ to fill out and produce the last page, print any held-over \minref{insertion}s, and end the job. It is the usual way to end your input file. \enddesc \begindesc \cts end {} \explain This command tells \TeX\ to produce the last page and end the job. It does not fill out the page, however, so it's usually better to use |\bye| rather than |\end|. \enddesc %========================================================================== \section {Input and output} %========================================================================== \subsection {Operations on input files} \begindesc \bix^^{files} \bix^^{input files} \easy\cts input {\} \explain \minrefs{file}\minrefs{file name} This command tells \TeX\ to read its input from file \. When that file is exhausted, \TeX\ returns to reading from its previous input source. You can nest input files to any level you like (within reason). When you're typesetting a large document, it's usually a good idea to structure your main file as a sequence of |\input| commands that refer to the subsidiary parts of the document. That way you can process the individual parts easily as you're working on drafts. It's also a good practice to put all of your \minref{macro} definitions into a separate file and summon that file with an |\input| command as the first action in your main file. \TeX\ uses different rules for scanning file names than it does for scanning \minref{token}s in general (see \xref{file name}). If your implementation expects file names to have extensions (usually indicated by a preceding dot), then \TeX\ provides a default extension of |.tex|. \example \input macros.tex \input chap1 % equivalent to chap1.tex | \endexample \enddesc \begindesc \cts endinput {} \explain This command tells \TeX\ to stop reading input from the current file when it next reaches the end of a line. \enddesc \begindesc \cts inputlineno {} \explain This command yields a number (not a string) giving the line number of the current line, defined to be the number that would appear in an error message if an error occurred at this point. \enddesc \begindesc \cts openin {\ {\bt =} \} \explain This command tells \TeX\ to open the file named \ and make it available for reading via the input stream designated by \. ^^{input streams//opening} \ must be between $0$ and $15$. Once you've opened a file and connected it to an input stream, you can read from the file using the |\read| command with the input stream's number. You can associate more than one input stream with the same file. You can then read from several different positions within the file, one for each input stream. You should allocate |\openin| stream numbers with |\newread| (\xref{\@newread}). \example \newread\auxfile \openin\auxfile = addenda.aux % \auxfile now denotes the number of this opening % of addenda.aux. | \endexample \enddesc \begindesc\secondprinting{\vglue-.5\baselineskip\vskip0pt} \cts closein {\} \explain This command tells \TeX\ to close the \minref{input stream} numbered \, i.e., end the association between the input stream and its file. The input stream with this number then becomes available for use with a different file. You should close an input stream once you're finished using its file. \example \closein\auxfile | \endexample \enddesc \begindesc\secondprinting{\vglue-.5\baselineskip\vskip0pt} \cts read {\ {\bt to} \} \explain ^^{input streams//reading with \b\tt\\read\e} ^^{reading a file} This command tells \TeX\ to read a line from the file associated with the \minref{input stream} designated by \ and assign the tokens on that line to \. The \minref{control sequence} then becomes a parameterless \minref{macro}. No macro expansion takes place during the reading operation. If the line contains any unmatched left braces, \TeX\ will read additional lines until the braces are all matched. If \TeX\ reaches the end of the file without matching all the braces, it will complain. If \ is greater than $15$ or hasn't been associated with a file using ^|\openin|, \TeX\ prompts you with `\ |=|' on your terminal and waits for you to type a line of input. It then assigns the input line to \. If \ is less than zero, it reads a line of input from your terminal but omits the prompt. \example \read\auxfile to \holder % Expanding \holder will produce the line just read. | \endexample \eix^^{input files} \enddesc \secondprinting{\vfill\eject} %========================================================================== \subsection {Operations on output files} \bix^^{output files} \begindesc \cts openout {\ {\bt =} \} \explain ^^{output streams//opening} This command tells \TeX\ to open the file named \ and make it available for writing via the \minref{output stream} designated by \. \ must be between $0$ and $15$. Once you've opened a file and connected it to an output stream, you can write to the file using the |\write| command with the output stream's number. An |\openout| generates a whatsit that becomes part of a box. The |\openout| does not take effect until \TeX\ ships out that box to the \dvifile, unless you've preceded the |\openout| with ^|\immediate|. \TeX\ won't complain if you associate more than one output stream with the same file, but you'll get garbage in the file if you try it! You should allocate stream numbers for |\openout| using |\newwrite| (\xref{\@newwrite}). \example \newwrite\auxfile \openout\auxfile = addenda.aux % \auxfile now denotes the number of this opening % of addenda.aux. | \endexample \enddesc \begindesc \cts closeout {\} \explain ^^{output streams//closing} This command tells \TeX\ to close the \minref{output stream} numbered \. i.e., end the association between the output stream and its file. The output stream with this number then becomes available for use with a different file. You should close an output stream once you're finished using its file. A |\closeout| generates a whatsit that becomes part of a box. The |\closeout| does not take effect until \TeX\ ships out that box to the \dvifile, unless you've preceded the |\closeout| with ^|\immediate|. \example \closeout\auxfile | \endexample \enddesc \begindesc \cts write {\ \rqbraces{\}} \explain ^^{output streams//writing} ^^{writing a file} This command tells \TeX\ to write \ to the file associated with the \minref{output stream} designated by \. It generates a whatsit that becomes part of a box. The actual writing does not take place until \TeX\ ships out that box to the \dvifile, unless you've preceded the |\write| with ^|\immediate|. For a |\write| that is not immediate, \TeX\ does not expand macros in \ until the token list is actually written to the file. The macro expansions follow the same rules as |\edef| (\xref\edef). In particular, any control sequence that is not the name of a macro is written as ^|\escapechar| followed by the control sequence name and a space. Any `|#|' tokens in \ are doubled, i.e., written as `|##|'. If \ is not in the range from $0$ to $15$, \TeX\ writes \ to the log file. ^^{log file//written by \b\tt\\write\e} If \ is greater than $15$ or isn't associated with an output stream, \TeX\ also writes \ to the terminal. \example \def\aa{a a} \write\auxfile{\hbox{$x#y$} \aa} % Writes the string `\hbox {$x##y$} a a' to \auxfile. | \endexample \enddesc \begindesc \cts immediate {} \explain This command should precede an |\openout|, |\closeout|, or |\write|. ^^|\write//with {\tt\\immediate}| ^^|\openout//with {\tt\\immediate}| ^^|\closeout//with {\tt\\immediate}| It tells \TeX\ to perform the specified file operation without delay. \example \immediate\write 16{I'm stuck!!} % has the same effect as \message | \endexample\enddesc \begindesc \cts special {\rqbraces{\}} \explain This command tells \TeX\ to write \ directly to the \dvifile\ when it next ships out a page. A typical use of |\special| would be to tell the device driver to incorporate the contents of a named graphics file into the output page. ^^{device drivers//instructions from \b\tt\\special\e} The |\special| command produces a whatsit that associates \ with a particular position on the page, namely, the position that a zero-size box would have had if such a box had appeared instead of the |\special| command. Any use you might make of |\special| depends strictly on the ^{device drivers} that you have available. \example \special{graphic expic} % Display the graphics file `expic' here. | \endexample \enddesc \begindesc \cts newlinechar {\param{number}} \explain This parameter contains a character that indicates a new line on output. When \TeX\ encounters this character while reading the argument of a |\write|, |\message|, or |\errmessage| command, it starts a new line. If |\newlinechar| is not in the range $0$--$255$, there is no character that indicates a new line on output. \PlainTeX\ sets |\newlinechar| to $-1$. \example \newlinechar = `\^^J \message{This message appears^^Jon two lines.} | \logproduces This message appears on two lines. | \endexample \enddesc \see |\newread|, |\newwrite| (\xref{\@newwrite}). \eix^^{files} \eix^^{output files} %========================================================================== \subsection {Interpreting input characters} \begindesc \cts catcode {\ \tblentry{number}} \explain ^^{category codes//in \b\tt\\catcode\e\ table} This table entry contains the \minref{category code} of the character whose \ascii\ code is \. The category codes are listed on \xrefpg{catcodes}. By changing the category code of a character you can get \TeX\ to treat that character differently. \example \catcode `\[ = 1 \catcode `\] = 2 % Make [ and ] act like left and right braces. | \endexample \enddesc \begindesc \cts active {} \explain This command contains the category code for an active character, namely, the number $13$. \example \catcode `\@ = \active % Make @ an active character. | \endexample \enddesc \begindesc \cts mathcode {\ \tblentry{number}} \explain This table entry contains the \minref{mathcode} of the character whose \ascii\ code is \ \seeconcept{mathcode}. The mathcode specifies that character's interpretation in math mode. \example \mathcode\> = "313E % as in plain TeX % The > character has class 3 (relation), family 1 (math % italic), and character code "3E | \endexample \enddesc \begindesc \margin{{\tt\\delcode} was explained in two places. This explanation combines them. (The other place was in the math section.)} \cts delcode {\\tblentry{number}} \explain ^^{delimiter codes} This table entry specifies the \minref{delimiter} code for the input character whose \ascii\ code is \. The delimiter code tells \TeX\ how to find the best output character to use for typesetting the indicated input character as a delimiter. \ is normally written in hexadecimal notation. Suppose that \ is the hexadecimal number $s_1s_2s_3\, l_1l_2l_3$. Then when the character is used as a delimiter, \TeX\ takes the character to have small variant $s_1s_2s_3$ and large variant $l_1l_2l_3$. Here $s_1s_2s_3$ indicates the math character found in position $s_2s_3$ of family $s_1$, and similarly for $l_1l_2l_3$. This is the same convention as the one used for ^|\mathcode| (\xref \mathcode), except that |\mathcode| also specifies a class. \example \delcode `( = "028300 % As in plain TeX. | \endexample \enddesc \begindesc \cts endlinechar {\param{number}} \explain This parameter contains the character code for the character that \TeX\ appends to the end of each input line. ^^{input lines} A value not in the range $0$--$255$ indicates that no character should be appended. \PlainTeX\ leaves |\end!-line!-char| at |`\^^M| (the {\ascii} code for \asciichar{return}). \enddesc \begindesc \cts ignorespaces {} \explain This command tells \TeX\ to read and expand tokens until it finds one that is not a space \minref{token}, ignoring any space tokens that it finds on the way. |\ignorespaces| is often useful at the end of a \minref{macro} as a way of making the macro insensitive to any spaces or ends of line that might follow calls on it. (An empty line after |\ignorespaces| still produces a |\par| token, however.) \example \def\aa#1{yes #1\ignorespaces} \aa{may} be | \produces \def\aa#1{yes #1\ignorespaces} \aa{may} be \endexample \enddesc %========================================================================== \section {Controlling interaction with \TeX} \begindesc \bix^^{controlling \TeX} \bix^^{running \TeX} \cts errorstopmode {} \explain This command tells \TeX\ to stop for interaction whenever it finds an error. This is the normal mode of operation. \enddesc \begindesc \cts scrollmode {} \explain This command tells \TeX\ not to stop for most errors, but to continue displaying the error messages on your terminal. Typing `|S|' or `|s|' in response to an error message puts you into scroll mode. \enddesc \begindesc \cts nonstopmode {} \explain This command tells \TeX\ not to stop for errors, even those pertaining to files that it can't find, but to continue displaying the error messages on your terminal. Typing `|R|' or `|r|' in response to an error message puts you into nonstop mode. \enddesc \begindesc \cts batchmode {} \explain This command tells \TeX\ not to stop for errors and to suppress all further output to your terminal. Typing `|Q|' or `|q|' in response to an error message puts you into batch mode. \enddesc \begindesc \cts pausing {\param{number}} \explain If this parameter is greater than zero, \TeX\ will pause at each line of input to give you an opportunity to replace it with a different line. If you type in a replacement, \TeX\ will use that line instead of the original one; if you respond with \asciichar{return}, \TeX\ will use the original line. Setting |\pausing| to $1$ can be useful as a way of patching a document as \TeX\ is processing it. For example, you can use this facility to insert ^|\show| commands (see below). \eix^^{running \TeX} \eix^^{controlling \TeX} \enddesc %========================================================================== \section {Diagnostic aids} \subsection{Displaying internal data} \begindesc \bix^^{tracing} \bix^^{debugging} \bix^^{diagnostic aids} \cts show {\} \cts showthe {\} \cts showbox {\} \cts showlists {} \explain These commands record information in the log of your \TeX\ run: \ulist \li |\show| records the meaning of \.\minrefs{token} ^^{tokens//displayed by \b\tt\\show\e} \li |\showthe| records whatever tokens would be produced by |\the| \ (see \xref \the). \li |\showbox| records the contents of the \minref{box} \minref{register} numbered \. The number of leading dots in the log indicates the number of levels of nesting of inner boxes. \li |\showlists| records the contents of each of the lists that \TeX\ is currently constructing. (These lists are nested one within another.) See \knuth{pages~88--89} for further information about interpreting the output of |\showlists|. \endulist For |\show| and |\showthe|, \TeX\ also displays the information at your ^{terminal}. For |\showbox| and |\showlists|, \TeX\ displays the information at your terminal only if ^|\tracingonline| (\xref \tracingonline) is greater than zero; if ^|\tracingonline| is zero or less (the default case), the information is not displayed. Whenever \TeX\ encounters a |\show|-type command it stops for interaction. The request for interaction does \emph{not} indicate an error, but it does give you an opportunity to ask \TeX\ to show you something else. If you don't want to see anything else, just press \asciichar{return}. You can control the amount of output produced by |\showbox| by setting |\show!-box!-breadth| and |\show!-box!-depth| (\xref\showboxbreadth). ^^|\showboxbreadth| ^^|\showboxdepth| These parameters respectively have default values of $5$ and $3$, which is why just five items appear for each box described in the log output below. (The `|..etc.|' indicates additional items within the boxes that aren't displayed.) \example \show a \show \hbox \show \medskip \show & | \logproduces > the letter a. > \hbox=\hbox. > \medskip=macro: ->\vskip \medskipamount . > alignment tab character &. | \nextexample \showthe\medskipamount \toks27={\hbox{Joe's\quad\ Diner}} \showthe\toks27 | \logproduces > 6.0pt plus 2.0pt minus 2.0pt. > \hbox {Joe's\quad \ Diner}. | \nextexample \setbox 3=\vbox{\hbox{A red dog.}\hrule A black cat.} \showbox 3 | \logproduces > \box3= \vbox(16.23332+0.0)x53.05565 .\hbox(6.94444+1.94444)x46.41675 ..\tenrm A ..\glue 3.33333 plus 1.66498 minus 1.11221 ..\tenrm r ..\tenrm e ..\tenrm d ..etc. .\rule(0.4+0.0)x* .\hbox(6.94444+0.0)x53.05565 ..\tenrm A ..\glue 3.33333 plus 1.66498 minus 1.11221 ..\tenrm b ..\tenrm l ..\tenrm a ..etc. | \endexample \vfil\eject \example \vbox{A \hbox {formula $x \over y\showlists$}} | \logproduces ### math mode entered at line 3 \mathord .\fam1 y this will be denominator of: \fraction, thickness = default \\mathord \.\fam1 x ### restricted horizontal mode entered at line 2 \tenrm f \tenrm o \tenrm r \tenrm m \kern-0.27779 \tenrm u \tenrm l \tenrm a \glue 3.33333 plus 1.66666 minus 1.11111 spacefactor 1000 ### horizontal mode entered at line 1 \hbox(0.0+0.0)x20.0 \tenrm A \glue 3.33333 plus 1.66498 minus 1.11221 spacefactor 999 ### internal vertical mode entered at line 1 prevdepth ignored ### vertical mode entered at line 0 prevdepth ignored | \endexample \enddesc \see |\showboxbreadth|, |\showboxdepth| \ctsref\showboxbreadth. \subsection{Specifying what is traced} \begindesc \cts tracingonline {\param{number}} \explain If this parameter is greater than zero, \TeX\ will display the results of tracing (including ^|\showbox| and ^|\showlists|) at your terminal in addition to recording them in the log file. \enddesc \begindesc \cts tracingcommands {\param{number}} \explain If this parameter is $1$ or greater, \TeX\ will record in the log file most commands that it executes. If ^|\tracingonline| is greater than zero, this information will also appear at your terminal. Typesetting the first character of a word counts as a command, but (for the purposes of the trace only) the actions of typesetting the subsequent characters and any punctuation following them do not count as commands. If |\tracingcommands| is $2$ or greater, \TeX\ will also record commands that are expanded rather than executed, e.g., conditional tests and their outcomes. \example \tracingcommands = 1 If $x+y>0$ we quit.\par On the other hand, \tracingcommands = 0 | \logproduces {vertical mode: the letter I} {horizontal mode: the letter I} {blank space } {math shift character $} {math mode: the letter x} {the character +} {the letter y} {the character >} {the character 0} {math shift character $} {horizontal mode: blank space } {the letter w} {blank space } {the letter q} {blank space } {\par} {vertical mode: the letter O} {horizontal mode: the letter O} {blank space } {the letter t} {blank space } {the letter o} {blank space } {the letter h} {blank space } {\tracingcommands} | \endexample \enddesc \begindesc \cts tracinglostchars {\param{number}} \explain If this parameter is greater than zero, \TeX\ will record an indication in the log file of each time that it drops an output character because that character does not exist in the current font. If ^|\tracingonline| is greater than zero, this information will also appear at your terminal. \PlainTeX\ defaults it to $1$ (unlike the others). \example \tracinglostchars = 1 A {\nullfont few} characters. | \logproduces Missing character: There is no f in font nullfont!! Missing character: There is no e in font nullfont!! Missing character: There is no w in font nullfont!! | \endexample \enddesc \begindesc \cts tracingmacros {\param{number}} \explain If this parameter is $1$ or greater, \TeX\ will record in the log file the expansion and arguments of every macro that it executes. ^^{macros//tracing} If |\tracingmacros| is $2$ or greater, \TeX\ will record, in addition, every expansion of a \minref{token} list such as |\output| or |\everycr|. If ^|\tracingonline| is greater than zero, this information will also appear at your terminal. \example \def\a{first \b, then \c} \def\b{b} \def\c{c} \tracingmacros = 2 Call \a once. | \logproduces \a ->first \b , then \c \b ->b \c ->c | \endexample \enddesc \begindesc \cts tracingoutput {\param{number}} \explain If this parameter is greater than zero, \TeX\ will record in the log file the contents of every box that it sends to the \dvifile. ^^{\dvifile//boxes recorded in log file} If ^|\tracingonline| is greater than zero, this information will also appear at your terminal. The number of leading dots in each line of the trace output indicates the nesting level of the box at that line. You can control the amount of tracing by setting ^|\showboxbreadth| and ^|\showboxdepth| (\xref\showboxbreadth). Setting |\tracingoutput| to $1$ can be particularly helpful when you're trying to determine why you've gotten ^{extra space} on a page. \example % This is the entire file. \tracingoutput = 1 \nopagenumbers One-line page. \bye | \logproduces Completed box being shipped out [1] \vbox(667.20255+0.0)x469.75499 .\vbox(0.0+0.0)x469.75499, glue set 13.99998fil ..\glue -22.5 ..\hbox(8.5+0.0)x469.75499, glue set 469.75499fil ...\vbox(8.5+0.0)x0.0 ...\glue 0.0 plus 1.0fil ..\glue 0.0 plus 1.0fil minus 1.0fil .\vbox(643.20255+0.0)x469.75499, glue set 631.2581fill ..\glue(\topskip) 3.05556 ..\hbox(6.94444+1.94444)x469.75499, glue set 386.9771fil ...\hbox(0.0+0.0)x20.0 ...\tenrm O ...\tenrm n ...\tenrm e ...\tenrm - ...etc. ..\glue 0.0 plus 1.0fil ..\glue 0.0 plus 1.0fill .\glue(\baselineskip) 24.0 .\hbox(0.0+0.0)x469.75499, glue set 469.75499fil ..\glue 0.0 plus 1.0fil | \endexample \enddesc \begindesc \cts tracingpages {\param{number}} \explain If this parameter is greater than zero, \TeX\ will record in the log file its calculations of the cost of various page breaks that it tries. ^^{page breaks//tracing} If |\tracing!-online| ^^|\tracingonline| is greater than zero, this information will also appear at your terminal. \TeX\ produces a line of this output whenever it first places a box or \minref{insertion} on the current page list, and also whenever it processes a potential break point for the page. Examining this output can be helpful when you're trying to determine the cause of a bad page break. See \knuth{pages~112--114} for an illustration and explanation of this output. Some production forms of \TeX\ ignore the value of |\tracingpages| so that they can run faster. If you need to use this parameter, be sure to use a form that responds to it. \enddesc \begindesc \cts tracingparagraphs {\param{number}} \explain If this parameter is greater than zero, \TeX\ will record in the log file its calculations of the cost of various line breaks that it tries. ^^{line breaking//tracing} If ^|\tracingonline| is greater than zero, this information will also appear at your terminal. \TeX\ produces this output when it reaches the end of each paragraph. See \knuth{pages~98--99} for an illustration and explanation of this output. Some production forms of \TeX\ ignore the value of |\tracing!-para!-graphs| so that they can run faster. If you need to use this parameter, be sure to use a form that responds to it. \enddesc \begindesc \cts tracingrestores {\param{number}} \explain If this parameter is greater than zero, \TeX\ will record in the log file the values that it restores when it encounters the end of a \minref{group}. If ^|\tracingonline| is greater than zero, this information will also appear at your terminal. Some production forms of \TeX\ ignore the value of |\tracing!-restores| so that they can run faster. If you need to use this parameter, be sure to use a form that responds to it. \enddesc \begindesc \cts tracingstats {\param{number}} \explain If this parameter is $1$ or greater, \TeX\ will include a report on the resources that it used to run your job (see \knuth{page~300} for a list and explanation of these resources). Moreover, if |\tracingstats| is $2$ or greater, \TeX\ will report on its memory usage whenever it does a ^|\shipout| (\xref \shipout) for a page. The report appears at the end of the log file. ^^{log file//tracing statistics in} If ^|\tracingonline| is greater than zero, the information will also appear at your terminal. If you're having trouble with \TeX\ exceeding one of its capacities, the information provided by |\tracingstats| may help you pinpoint the cause of the difficulty. Some production forms of \TeX\ ignore the value of |\tracingstats| so that they can run faster. If you need to use this parameter, be sure to use a form that responds to it. The following example shows a sample of the tracing output you'd get on one implementation of \TeX. It may be different on other implementations. {\codefuzz = 1in \example \tracingstats=1 | \logproduces Here is how much of TeX's memory you used: 4 strings out of 5540 60 string characters out of 72328 5956 words of memory out of 262141 921 multiletter control sequences out of 9500 14794 words of font info for 50 fonts, out of 72000 for 255 14 hyphenation exceptions out of 607 7i,4n,1p,68b,22s stack positions out of 300i,40n,60p,3000b,4000s | \endexample }% end scope of codefuzz \enddesc \begindesc \cts tracingall {} \explain This command tells \TeX\ to turn on every available form of tracing. It also sets ^|\tracingonline| to $1$ so that the trace output will appear at your terminal. \enddesc \begindesc \cts showboxbreadth {\param{number}} \explain This parameter specifies the maximum number of list items that \TeX\ displays for one level of one box when it is producing the output for ^|\showbox| or ^|\tracingoutput|. \PlainTeX\ sets |\showboxbreadth| to $5$. \enddesc \begindesc \cts showboxdepth {\param{number}} \explain This parameter specifies the level of the deepest list that \TeX\ displays when it is producing the output for ^|\showbox| or ^|\showlists|. \PlainTeX\ sets |\showboxdepth| is $3$. \eix^^{tracing} \eix^^{debugging} \eix^^{diagnostic aids} \enddesc %========================================================================== \subsection {Sending messages} \begindesc \bix^^{messages, sending} \bix^^{error messages} \cts message {\rqbraces{\}} \cts errmessage {\rqbraces{\}} \explain These commands display the message given by \ on your terminal and also enter it into the log. Any \minref{macro}s in the message are expanded, but no commands are executed. This is the same rule that \TeX\ uses for |\edef| (\xref \edef). For |\errmessage|, \TeX\ pauses in the same way that it does for one of its own error messages and displays the |\errhelp| tokens if you ask for help. You can generate multiline messages by using the ^|\newlinechar| character (\xref \newlinechar). \example \message{Starting a new section.} | \endexample \enddesc \begindesc \cts wlog {\rqbraces{\}} \explain This command writes \ on the log file. ^^{log file//written by \b\tt\\wlog\e} \minrefs{log file} \TeX\ expands \ according to the same rules that it uses for |\edef| (\xref\edef). \example \wlog{Take two aspirins and call me in the morning.} | \logproduces Take two aspirins and call me in the morning. | \endexample \enddesc \begindesc \cts errhelp {\param{token list}} \explain This parameter contains the token list that \TeX\ displays when you ask for help in response to an |\errmessage| command. We recommend that when you're generating an error message with |\errmessage|, you set |\errhelp| to a string that describes the nature of the error and use |\newhelp| to produce that string. You can use the ^|\newlinechar| character to produce multiline messages. \enddesc \begindesc \ctspecial newhelp \ctsxrdef{@newhelp} {\ \rqbraces{\}} \explain This command assigns the ^{help message} given by \ to \. It provides an efficient way of defining the ^{help text} that further explains an error message. Before issuing the error message with the |\errmessage| command, you should assign \ to ^|\errhelp|. The help text will then appear if the user types `|H|' or `|h|' in response to the error message. \example \newhelp\pain{Your input includes a token that I find^^J to be offensive. Don't bother me again with this^^J document until you've removed it.} \errhelp = \pain \newlinechar = `\^^J % ^^J will start a new line \errmessage{I do not appreciate receiving this token} | \logproduces !! I do not appreciate receiving this token. l.8 ...t appreciate receiving this token.} ? H \Your input includes a token that I find to be offensive. Don't bother me again with this document until you've removed it. | \endexample \enddesc \begindesc \cts errorcontextlines {\param{number}} \explain This parameter determines the number of pairs of context lines, not counting the top and bottom pairs, that \TeX\ prints when it encounters an error. By setting it to $0$ you can get rid of long error messages. You can still force out the full context by typing something like: \csdisplay I\errorcontextlines=100\oops | in response to an error, since the undefined control sequence |\oops| will cause another error. \PlainTeX\ sets |\error!-context!-lines| to $5$. \enddesc \see |\write| (\xref \write), |\escapechar| (\xref \escapechar). \eix^^{messages, sending} \eix^^{error messages} %========================================================================== \section {Initializing \TeX} \begindesc \cts dump {} \explain This command, which must not appear inside a group, dumps the contents of \TeX's memory to a ^{format file} (\xref{format file}). By using ^|virtex|, a special ``virgin'' form of \TeX, you can then reload the format file at high speed and continue in the same state that \TeX\ was in at the time of the dump. |\dump| also ends the run. Since |\dump| can only be used in ^|initex|, not in production forms of \TeX, it is only useful to people who are installing \TeX. \enddesc \begindesc \cts everyjob {\param{token list}} \explain This parameter contains a \minref{token} list that \TeX\ expands at the start of every job. Because an assignment to |\everyjob| cannot affect the current run (by the time you've done the assignment it's already too late), it is only useful to people who are preparing format files. \enddesc \enddescriptions \endchapter \byebye