diff options
author | Karl Berry <karl@freefriends.org> | 2015-09-20 22:31:12 +0000 |
---|---|---|
committer | Karl Berry <karl@freefriends.org> | 2015-09-20 22:31:12 +0000 |
commit | 9280f63533af6460c33dbfcf486a5e5368afed53 (patch) | |
tree | 4d9646df783056a13bfdfe201d510c6fe07c5858 /Master/texmf-dist/source/generic/multiexpand | |
parent | bf941dd6fd9a67534bf00d114908c151b5fd1606 (diff) |
multiexpand (20sep15)
git-svn-id: svn://tug.org/texlive/trunk@38416 c570f23f-e606-0410-a88d-b1316a301751
Diffstat (limited to 'Master/texmf-dist/source/generic/multiexpand')
-rw-r--r-- | Master/texmf-dist/source/generic/multiexpand/multiexpand.dtx | 506 | ||||
-rw-r--r-- | Master/texmf-dist/source/generic/multiexpand/multiexpand.ins | 71 |
2 files changed, 577 insertions, 0 deletions
diff --git a/Master/texmf-dist/source/generic/multiexpand/multiexpand.dtx b/Master/texmf-dist/source/generic/multiexpand/multiexpand.dtx new file mode 100644 index 00000000000..e28291565d0 --- /dev/null +++ b/Master/texmf-dist/source/generic/multiexpand/multiexpand.dtx @@ -0,0 +1,506 @@ +% \iffalse meta-comment +%<*internal> +\iffalse +%</internal> +%<*readme> +Variations on the `\expandafter` TeX primitive +============================================== +* E-mail: blflatex@gmail.com +* Released under the LaTeX Project Public License v1.3c or later + See http://www.latex-project.org/lppl.txt + +The package defines `\multiexpand`, `\multiexpandafter`, `\MultiExpand` +and `\MultiExpandAfter`, replacing large chains of `\expandafter`. +These four commands take as an argument the number of expansion to be +done. If eTeX is available, this number is evaluated using `\numexpr`. + +Say we want to expand `\C` five times before `\A` and `\B` in `\A\B\C`. +The traditional approach would be to insert 31 `\expandafter` before +`\A` and the same number before `\B`. With this package one can use +any of + + \expandafter\A\expandafter\B\romannumeral\multiexpand{5}\C + + \expandafter\A\romannumeral\multiexpandafter{5}\B\C + + \MultiExpandAfter{2}\A\MultiExpandAfter{5}\B\C + + \MultiExpandAfter{2}\A\MultiExpandAfter{2}\B\MultiExpand{5}\C + +In one step of expansion (triggered by the `\expandafter`'s), +`\romannumeral\multiexpand{5}` expands the following token 5 times, +whereas `\romannumeral\multiexpandafter{5}` expands the token after +that 5 times. The macros `\MultiExpandAfter` and `\MultiExpand` +take two steps of expansion, but do not require `\romannumeral`. +Another example is that in two steps of expansion, + + \MultiExpandAfter{2}\a\MultiExpandAfter{3}{% + \MultiExpandAfter{9}\b\MultiExpandAfter{10}\c\d} + +expands `\d` 10 times, then `\c` 7 times (9-2), then `\b` once (3-2). + +The package can be built from the file `multiexpand.dtx' by running + + pdflatex multiexpand.dtx + pdflatex multiexpand.dtx + +%</readme> +%<*lvt> + \input regression-test.tex\relax + \START + \AUTHOR{Bruno Le Floch} +%<lvt002> \let\numexpr\relax + \input multiexpand.sty\relax + + %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% + % Introduce commands that expand to each other, to count how many times + % things were expanded. + \OMIT + \def\i{} + \def\v{\i\i\i\i} + \def\x{\i\i\i\i\v} + \def\l{\i\i\i\i\v\x\x\x\x} + \def\c{\i\i\i\i\v\x\x\x\x\l} + \def\d{\i\i\i\i\v\x\x\x\x\l\c\c\c\c} + \def\m{\i\i\i\i\v\x\x\x\x\l\c\c\c\c\d} + \TIMO + + %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% + % We test |\MultiExpand|, which needs two expansions. In particular, we see + % that the implementation is reasonably quick, and we get the $2345$-th + % expansion of |\m\m\m\m|. + % 4000-2345 = 1655 = mdclv + \TEST{MultiExpand many times (vlcdm)}{% + \toks0\expandafter\expandafter\expandafter{\MultiExpand{2345}\m\m\m\m}% + \TYPE{\the\toks0.}% + } + + %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% + % Test |\MultiExpandAfter|, two expansions to get the specified expansion. + % The results indirectly show that expansions happen in the expected order. + % 100 = c + % 100-(27-2) = 75 = lxxv + % 100-(14-2) = 88 = lxxxviii + % 100-38 = 62 = lxii + \TEST{MultiExpandAfter chain (c vxxl iiivxxxl iixl)}{% + \toks0\expandafter\expandafter\expandafter{% + \MultiExpandAfter{27}\c + \MultiExpandAfter{14}\c + \MultiExpandAfter{38}\c\c}% + \TYPE{\the\toks0.}% + } + + %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% + % Test the order of expansion. The example was given in the original + % question that led to writing this package. For the record, this + % package (at some point) used 32 |\expandafter| to do the expansion + % in the right order, which was better than other approaches; however + % this counting ignores the large overheads due to other primitives. + \OMIT + \def\a#1.{#1} + \def\b#1:{#1.} + \def\c#1,{#1:} + \def\d{Hello world!,} + \TIMO + % + \TEST{MultiExpandAfter order}{% + \toks0\MultiExpandAfter{3}{% + \MultiExpandAfter{3}\a + \MultiExpandAfter{3}\b + \MultiExpandAfter{1}\c + \d}% + \TYPE{\the\toks0}% + } + + %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% + % The other example from the same question (expanding seven tokens once + % each in the reverse order) takes 74 \expandafter. + \OMIT + \def\b.{.b} \def\c.{.c} \def\d.{.d} + \def\e.{.e} \def\f.{.f} \def\g.{.g} + \TIMO + % + \TEST{MultiExpandAfter order (2)}{% + \toks0\MultiExpandAfter{2}{% + \MultiExpandAfter3\a \MultiExpandAfter3\b + \MultiExpandAfter3\c \MultiExpandAfter3\d + \MultiExpandAfter3\e \MultiExpandAfter1\f\g.}% + \TYPE{\the\toks0}% + } + + \END +%</lvt> +%<*tlg> +This is a generated file for the l3build validation system. +Don't change this file in any respect. +%% \CharacterTable +%% {Upper-case \A\B\C\D\E\F\G\H\I\J\K\L\M\N\O\P\Q\R\S\T\U\V\W\X\Y\Z +%% Lower-case \a\b\c\d\e\f\g\h\i\j\k\l\m\n\o\p\q\r\s\t\u\v\w\x\y\z +%% Digits \0\1\2\3\4\5\6\7\8\9 +%% Exclamation \! Double quote \" Hash (number) \# +%% Dollar \$ Percent \% Ampersand \& +%% Acute accent \' Left paren \( Right paren \) +%% Asterisk \* Plus \+ Comma \, +%% Minus \- Point \. Solidus \/ +%% Colon \: Semicolon \; Less than \< +%% Equals \= Greater than \> Question mark \? +%% Commercial at \@ Left bracket \[ Backslash \\ +%% Right bracket \] Circumflex \^ Underscore \_ +%% Grave accent \` Left brace \{ Vertical bar \| +%% Right brace \} Tilde \~} +%% +Author: Bruno Le Floch +(multiexpand.sty) +============================================================ +TEST 1: MultiExpand many times (vlcdm) +============================================================ +\v \l \c \d \m . +============================================================ +============================================================ +TEST 2: MultiExpandAfter chain (c vxxl iiivxxxl iixl) +============================================================ +\c \v \x \x \l \i \i \i \v \x \x \x \l \i \i \x \l . +============================================================ +============================================================ +TEST 3: MultiExpandAfter order +============================================================ +Hello world! +============================================================ +============================================================ +TEST 4: MultiExpandAfter order (2) +============================================================ +\a .bcdefg +============================================================ +%</tlg> +%<*internal> +\fi +\def\nameofplainTeX{plain} +\ifx\fmtname\nameofplainTeX\else + \expandafter\begingroup +\fi +%</internal> +%<*install> +\input docstrip.tex +\keepsilent +\askforoverwritefalse +\preamble +---------------------------------------------------------------- +multiexpand --- trigger multiple expansions in one expansion step. +E-mail: blflatex@gmail.com +Released under the LaTeX Project Public License v1.3c or later +See http://www.latex-project.org/lppl.txt +---------------------------------------------------------------- + +\endpreamble +\postamble + +Copyright (C) 2011-2015 by Bruno Le Floch <blflatex@gmail.com> + +This work may be distributed and/or modified under the +conditions of the LaTeX Project Public License (LPPL), either +version 1.3c of this license or (at your option) any later +version. The latest version of this license is in the file: + +http://www.latex-project.org/lppl.txt + +This work is "maintained" (as per LPPL maintenance status) by +Bruno Le Floch. + +This work consists of the file multiexpand.dtx +and the derived files multiexpand.ins, + multiexpand.pdf and + multiexpand.sty. + +\endpostamble +\usedir{tex/latex/multiexpand} +\generate{ + \file{\jobname.sty}{\from{\jobname.dtx}{package}} +} +%</install> +%<install>\endbatchfile +%<*internal> +\usedir{source/latex/multiexpand} +\generate{ + \file{\jobname.ins}{\from{\jobname.dtx}{install}} + \file{\jobname 001.lvt}{\from{\jobname.dtx}{lvt}} + \file{\jobname 002.lvt}{\from{\jobname.dtx}{lvt,lvt002}} +} +\nopreamble\nopostamble +\generate{ + \file{\jobname 001.tlg}{\from{\jobname.dtx}{tlg}} + \file{\jobname 002.tlg}{\from{\jobname.dtx}{tlg}} +} +\usedir{doc/latex/multiexpand} +\generate{ + \file{README.md}{\from{\jobname.dtx}{readme}} +} + +\ifx\fmtname\nameofplainTeX + \expandafter\endbatchfile +\else + \expandafter\endgroup +\fi +%</internal> +%<*driver> +\documentclass[12pt]{ltxdoc} +\usepackage[T1]{fontenc} +\usepackage{hologo} +\usepackage{multiexpand} +\usepackage[margin=4cm]{geometry} +\usepackage{hyperref} +\RecordChanges +\begin{document} + \DocInput{multiexpand.dtx} +\end{document} +%</driver> +% \fi +% +% \CheckSum{126} +% +% \def\fileversion{v1.4} \def\filedate{2015/09/20} +%\title{\texttt{multiexpand}\\ +% Trigger multiple expansions \\ in one expansion step\thanks{This file +% describes version \fileversion, last revised \filedate.}} +%\author{Bruno Le Floch\thanks{E-mail: blflatex@gmail.com} +% \thanks{I have gathered ideas from various posts in the \texttt{\{TeX\}} +% community at \url{http://tex.stackexchange.com}. Thanks to their authors.}} +%\date{Released \filedate} +% +%\maketitle +% +% \tableofcontents +% \changes{v1.0}{2011/02/15}{First version with documentation} +% \changes{v1.1}{2013/01/08}{Version submitted to CTAN} +% \changes{v1.2}{2013/01/09}{Use fewer expandafter for large arguments} +% \changes{v1.2}{2013/01/09}{Change ME prefix to multiexpand} +% \changes{v1.4}{2015/09/20}{Clarify that eTeX is not required} +% +% \section{Two user commands} +% +% \begin{itemize} +% \item For $n>0$, expanding |\MultiExpand{|$n$|}\macro| twice gives +% the $n$-th expansion of |\macro|. +% \item For $n>0$, expanding |\MultiExpandAfter{|$n$|}\macroA\macroB| +% twice expands |\macroB| $n$ times before expanding |\macroA|. +% \end{itemize} +% Note that neither functions work for $n = 0$. +% +% These can typically be combined as +% \begin{verbatim} +% \MultiExpand{7}% +% \MultiExpandAfter{4}\a\MultiExpandAfter{7}\b% +% \MultiExpandAfter{3}\c\d +% \end{verbatim} +% which would expand |\d| $3$~times, then |\c| $5$~times ($2$~of the $7$~times +% were used to expand |\MultiExpandAfter{3}|), then |\b| twice ($4-2$), and +% finally |\a| five times ($7-2$). Note that all this happens in precisely +% \emph{two} steps of expansion. +% +% In some cases, one needs to achieve the same effect in \emph{one} step +% only. For this, we use the first expansion of |\MultiExpand|, which +% is |\romannumeral| |\multiexpand|, or of |\MultiExpandAfter|, which is +% |\romannumeral| |\multiexpandafter|. In detail, expanding |\romannumeral| +% |\multiexpand{|$n$|}| once expands the following token $n$ times, and +% similarly for |\romannumeral| |\multiexpandafter{|$n$|}|. +% +% These are especially useful when we want to expand several times a +% very specific token which is buried behind many others. For instance, +% expanding the following code once +% \begin{verbatim} +% \expandafter\macroA\expandafter\macroB +% \romannumeral\multiexpandafter{4}\macroC\macroD +% \end{verbatim} +% will expand |\macroD| $4$ times before the three other macros. +% +% Note: as we mentionned, this breaks for $n=0$. But in this case, consider +% using |\expandafter\empty|, or a variant thereof. +% +% \section{Implementation} +%\StopEventually{\PrintChanges\PrintIndex} +% \begin{macrocode} +%<*package> +% \end{macrocode} +% +% We work inside a group, to change the catcode of |@|. So we will only +% do |\gdef|s. Note that this code can be read several times with no +% issue; no need to bother to check whether it was already read or not. +% \begin{macrocode} +\begingroup +\catcode `\@=11 +% \end{macrocode} +% +% \subsection{Common to the \hologo{eTeX} and non-\hologo{eTeX} cases} +% +% For the ``lazy'', who do not want to use |\romannumeral|, we provide +% |\MultiExpand| and |\MultiExpandAfter|, simple shorthands. A drawback +% is that they require two steps of expansion rather than only one. +% \begin{macrocode} +\gdef \MultiExpand {\romannumeral \multiexpand } +\gdef \MultiExpandAfter {\romannumeral \multiexpandafter } +% \end{macrocode} +% +% \subsection{Without \hologo{eTeX}'s \cs{numexpr}} +% \changes{v1.3}{2015/03/03}{Support TeX with no numexpr} +% +% No need for the usual |\begingroup\expandafter\endgroup| to prevent +% |\numexpr| from being set to |\relax|, because we are already in a +% group. +% \begin{macrocode} +\expandafter \ifx \csname numexpr\endcsname \relax +% \end{macrocode} +% +% A helper. +% \begin{macrocode} +\long \gdef \multiexpand@gobble #1{} +% \end{macrocode} +% +% The user commands |\multiexpand| and |\multiexpandafter|, to be used +% after |\romannumeral|. They only differ a little bit. +% \begin{macrocode} +\gdef \multiexpand {\multiexpand@aux \multiexpand@ } +\gdef \multiexpandafter {\multiexpand@aux \multiexpand@after } +% \end{macrocode} +% The user commands receives a number, and to accept various forms of +% numbers we hit it with |\number|. If it is non-positive, stop the +% |\romannumeral| expansion with |0| and a space. Otherwise, reverse +% the number, to make it easy to subtract~$1$. +% \begin{macrocode} +\long \gdef \multiexpand@aux #1#2% + {\expandafter \multiexpand@test \number #2;#1} +\gdef \multiexpand@test #1;#2% + {% + \ifnum #1>0 + \multiexpand@reverse #1{?\multiexpand@reverse@end }?;;#2% + \fi + 0 % + } +% \end{macrocode} +% The macro |\multiexpand@reverse| puts characters from the number one +% by one (as~|#1|) after the semicolon, to reverse the number. After +% the last digit, |#1| is |{?\multiexpand@reverse@end}|. The question +% mark is removed by |\multiexpand@gobble|, and the |reverse@end| +% macro cleans up. In particular, one should not forget to close the +% conditional using~|#5|, which is the trailing~|\fi|. At this stage, +% |#4| is the function that distinguishes |\multiexpand| from +% |\multiexpand@after|, and |#3| is the reversed +% number. \begin{macrocode} +\gdef \multiexpand@reverse #1#2;% + {\multiexpand@gobble #1\multiexpand@reverse #2;#1} +\gdef \multiexpand@reverse@end #1;?#2#3;#4#50 + {#5\multiexpand@iterate #41#3;} +% \end{macrocode} +% The macro |\multiexpand@iterate| applies a \meta{function} a certain +% number of times to what follows in the input stream. It expects to +% receive \meta{function} \meta{nines} |1|\meta{reversed number}|;|. +% The argument \meta{nines}, made entirely of the digit~|9|, is used to +% compute carries when subtracting~$1$, and is initally empty. As a +% concrete example, after |\multiexpand{302}| the successive +% calls to |\multiexpand@iterate| would go as follows. +% \begin{verbatim} +% \multiexpand@iterate \multiexpand@ 1203; +% \multiexpand@iterate \multiexpand@ 1103; +% \multiexpand@iterate \multiexpand@ 1003; +% \multiexpand@iterate \multiexpand@ 9 103; +% \multiexpand@iterate \multiexpand@ 99 13; +% \multiexpand@iterate \multiexpand@ 1992; +% \multiexpand@iterate \multiexpand@ 1892; +% \multiexpand@iterate \multiexpand@ 1792; +% \end{verbatim} +% Note in particular how carries are done in several steps. The details +% are left as an exercise to the reader. The most common case is when +% |#2| is empty and |#3| is a non-zero digit. Then |\number| is +% expanded, triggering |\ifcase| which shifts |#3| by one unit, and |#1| +% takes care of expanding the tokens are required by |\multiexpand| or +% |\multiexpandafter|. If |#3| is~$0$, then |\multiexpand@zero| is +% called, closing the conditional with |#1|, and iterating, this time +% with a non-empty \meta{nines}, which are the argument |#2| of a new +% call to |\multiexpand@iterate|. Those \meta{nines} are put back into +% the number by |\multiexpand@iterate|, unless the next significant +% digit is also $0$, in which case |\multiexpand@zero| is called again, +% until finding a non-zero digit; at each step, one more $9$ is added to +% the \meta{nines}. If all digits are zero, we reach |;| this way, and +% end, after cleaning up. +% \begin{macrocode} +\gdef \multiexpand@iterate #1#21#3% + {% + \ifx ;#3\multiexpand@end \fi + \ifx 0#3\multiexpand@zero \fi + \expandafter \multiexpand@iterate + \expandafter #1% + \number 1#2% + \ifcase #3 \or 0\or 1\or 2\or 3\or 4\or 5\or 6\or 7\or 8\fi + #1% + } +\gdef \multiexpand@zero #1#2\number 1#3\ifcase #4\fi #5% + {#1\multiexpand@iterate #59#31} +\gdef \multiexpand@end #1#2\ifcase #3\fi #4{#10 } +% \end{macrocode} +% Finally, the two different expansion commands. +% \begin{macrocode} +\gdef \multiexpand@ #1;{#1\expandafter ;} +\gdef \multiexpand@after #1;{#1\expandafter ;\expandafter } +% \end{macrocode} +% +% \subsection{With \hologo{eTeX}} +% +% \begin{macrocode} +\else +% \end{macrocode} +% +% With \hologo{eTeX}, everything is much easier, since the engine knows +% how to subtract~$1$. +% +% The main looping macros expect their arguments as an integer followed +% by a semicolon. As long as the argument is at least~$2$, decrement +% it, and expand what follows. Once the argument is~$1$ (or less: the +% macros are not meant to handle that case), call |\multiexpand@end| to clean up +% and stop looping. +% \begin{macrocode} +\gdef \multiexpand@ #1;% + {% + \ifnum #1<2 \multiexpand@end \fi + \expandafter \multiexpand@ + \the \numexpr #1-1\expandafter ;% + } +\gdef \multiexpand@after #1;% + {% + \ifnum #1<2 \multiexpand@end \fi + \expandafter \multiexpand@after + \the \numexpr #1-1\expandafter ;\expandafter + } +% \end{macrocode} +% The looping macros are used within an overarching |\romannumeral| +% expansion, which we end with a |0| and a space, as well as the +% appropriate |\expandafter|. Here, |#1| is |\fi| which needs to remain +% to close the conditional, |#2| is |\expandafter|, and there is a +% trailing |\expandafter| in the case of |\multiexpand@after|. +% \begin{macrocode} +\gdef \multiexpand@end #1#2#3;{#10#2 } +% \end{macrocode} +% Finally, user commands, used as |\romannumeral| |\multiexpand(after)|. +% Those evaluate their argument, and pass it to |\multiexpand@(after)|. +% The argument might contain |\par| tokens (who knows) +% \begin{macrocode} +\long \gdef \multiexpand #1% + {\expandafter \multiexpand@ \the \numexpr #1;} +\long \gdef \multiexpandafter #1% + {\expandafter \multiexpand@after \the \numexpr #1;} +% \end{macrocode} +% +% \begin{macrocode} +\fi +% \end{macrocode} +% +% Close the group. +% \begin{macrocode} +\endgroup +% \end{macrocode} +% +% \begin{macrocode} +%</package> +% \end{macrocode} +%\Finale +\endinput + + + diff --git a/Master/texmf-dist/source/generic/multiexpand/multiexpand.ins b/Master/texmf-dist/source/generic/multiexpand/multiexpand.ins new file mode 100644 index 00000000000..956f082a98d --- /dev/null +++ b/Master/texmf-dist/source/generic/multiexpand/multiexpand.ins @@ -0,0 +1,71 @@ +%% +%% This is file `multiexpand.ins', +%% generated with the docstrip utility. +%% +%% The original source files were: +%% +%% multiexpand.dtx (with options: `install') +%% ---------------------------------------------------------------- +%% multiexpand --- trigger multiple expansions in one expansion step. +%% E-mail: blflatex@gmail.com +%% Released under the LaTeX Project Public License v1.3c or later +%% See http://www.latex-project.org/lppl.txt +%% ---------------------------------------------------------------- +%% +\input docstrip.tex +\keepsilent +\askforoverwritefalse +\preamble +---------------------------------------------------------------- +multiexpand --- trigger multiple expansions in one expansion step. +E-mail: blflatex@gmail.com +Released under the LaTeX Project Public License v1.3c or later +See http://www.latex-project.org/lppl.txt +---------------------------------------------------------------- + +\endpreamble +\postamble + +Copyright (C) 2011-2015 by Bruno Le Floch <blflatex@gmail.com> + +This work may be distributed and/or modified under the +conditions of the LaTeX Project Public License (LPPL), either +version 1.3c of this license or (at your option) any later +version. The latest version of this license is in the file: + +http://www.latex-project.org/lppl.txt + +This work is "maintained" (as per LPPL maintenance status) by +Bruno Le Floch. + +This work consists of the file multiexpand.dtx +and the derived files multiexpand.ins, + multiexpand.pdf and + multiexpand.sty. + +\endpostamble +\usedir{tex/latex/multiexpand} +\generate{ + \file{\jobname.sty}{\from{\jobname.dtx}{package}} +} +\endbatchfile +%% +%% Copyright (C) 2011-2015 by Bruno Le Floch <blflatex@gmail.com> +%% +%% This work may be distributed and/or modified under the +%% conditions of the LaTeX Project Public License (LPPL), either +%% version 1.3c of this license or (at your option) any later +%% version. The latest version of this license is in the file: +%% +%% http://www.latex-project.org/lppl.txt +%% +%% This work is "maintained" (as per LPPL maintenance status) by +%% Bruno Le Floch. +%% +%% This work consists of the file multiexpand.dtx +%% and the derived files multiexpand.ins, +%% multiexpand.pdf and +%% multiexpand.sty. +%% +%% +%% End of file `multiexpand.ins'. |