rcs

git-svn-id: svn://tug.org/texlive/trunk@1267 c570f23f-e606-0410-a88d-b1316a301751

1 files changed, 1024 insertions, 0 deletions

diff --git a/Master/texmf-dist/tex/latex/rcs/rcs.sty b/Master/texmf-dist/tex/latex/rcs/rcs.sty
new file mode 100644
index 00000000000..453b947d75c
--- /dev/null
+++ b/Master/texmf-dist/tex/latex/rcs/rcs.sty
@@ -0,0 +1,1024 @@
+% $StyleId: rcs.doc,v 2.10 2003/02/02 20:42:38 schrod Exp $
+%----------------------------------------------------------------------
+% Written by Joachim Schrod <jschrod@acm.org>.
+% Copyright conditions see below.
+
+%
+% LaTeX package rcs
+% to use RCS tag values in one's document, and typeset revision logs
+%
+% [LaTeX in MAKEPROG]
+% (history at end)
+
+% If you have received this style file without the user manual (in the
+% file rcs-user.tex or the respective DVI file), it's incomplete and
+% near to useless. If it was given to you as something that you shall
+% use as an author -- complain bitterly to your provider. You need the
+% documentation and you have a right on it! (Below you can find info
+% where to get the reference version.)
+
+
+
+%%%%
+%%%%
+%%%% These TeX macros were documented with the documentation system
+%%%% MAKEPROG and automatically converted to the current form.
+%%%% If you have MAKEPROG available you may transform it back to
+%%%% the original input: Remove every occurence of three percents
+%%%% and one optional blank from the beginning of a line and remove
+%%%% every line which starts with four percents.  The following Perl
+%%%% command will do this:
+%%%%
+%%%%	perl -pe 's/^%%%%.*\n//; s/^%%%( )?//'
+%%%%
+%%%%	If you just want to print the documentation you may fetch
+%%%% the archive print-makeprog.tar.Z from ftp.th-darmstadt.de (directory
+%%%% pub/tex/latex). It contains *all* used styles -- but beware, they
+%%%% may not be in a documented form...
+%%%%
+%%%%
+%%% \documentclass{progltx}
+
+%%% \usepackage{rcs-doc}
+%%% \usepackage{path}
+%%% \usepackage{alltt}
+%%% \usepackage{fullpage}
+
+
+%%% \RCS $StyleDate: 2003/02/02 20:42:38 $
+
+
+%%% \begin{document}
+
+
+%%% \title{The \texttt{rcs} Package\\
+%%%     {\large (Implementation)}%
+%%%     }
+%%% \author{%                       % LaTeX does not discard unnecessary glue...
+%%%     Joachim Schrod%
+%%%     \thanks{%
+%%%         \protect\raggedright
+%%%         R\"odermark, Germany.
+%%%         Email: \texttt{jschrod@acm.org}%
+%%%         }%
+%%%     }
+%%% \date{%
+%%%     Revision \RCSStyleRevision\\
+%%%     (as of \RCSStyleDate)%
+%%%     }
+
+%%% \maketitle
+
+
+
+%%% % ------------------------------------------------------------
+%%% %
+%%% % subdocument: The user interface of rcs.sty
+%%% %
+
+%%% \input{rcs-user}
+
+%%% %
+%%% % ------------------------------------------------------------
+%%% %
+%%% % subdocument: The internal interface
+%%% %
+
+%%% \input{rcs-conf}
+
+%%% %
+%%% % ------------------------------------------------------------
+
+
+
+%%% \chap Implementation.
+
+%%% This module reserves the namespace |rcs|.
+
+%%% \beginprog
+\ifx \rcs@loaded\undefined
+    \def\rcs@loaded{$StyleRevision: 2.10 $}
+\else
+    \PackageWarningNoLine{rcs}%
+	{Some other package already uses namespace `rcs'}
+\fi
+%%% \endprog
+
+
+%%% \sect Let's identify this package.
+
+%%% Now I'm caught in a catch-22 situation. The code below---part of my
+%%% standard templates for \LaTeX{} packages---uses principles that are
+%%% explained later in this document. In particular, the explanation how
+%%% the information from the following RCS fields are gathered starts at
+%%% chunk~\ref{chunk:value-split}.
+
+%%% There you'll find also the explanation for the funny strings behind
+%%% the RCS fields. Actually, I would not need them any more, they make
+%%% the initial document (before the first check-in) work. But they won't
+%%% be discarded in this package source, as they'll demonstrate another
+%%% utilization of RCS at work.
+
+%%% \beginprog
+\begingroup
+    \def\RCSPackage#1#2 $#3: #4 #5\endRCS $#6: #7 #8\endRCS{%
+	\def\date{#4}\def\id{v#7}%
+	\ProvidesPackage{#1}[\date\space\id\space #2]%
+	}
+
+  \RCSPackage{rcs}{typeset info from RCS fields}
+  $StyleDate: 2003/02/02 20:42:38 $: 9999/00/00 \endRCS
+  $StyleRevision: 2.10 $: 0.0 \endRCS
+\endgroup
+%%% \endprog
+
+
+%%% \sect This module is supported. Send bug reports, comments, and
+%%% repairs.
+
+%%% The reference version may be retrieved via anonymous ftp from CTAN,
+%%% directory \path|/tex-archive/macros/latex/contrib/supported/rcs|.
+
+
+%%% \sect This is freely distributable software; you can redistribute it
+%%% and/or modify it under the terms of the GNU General Public License
+%%% (version~2) as published by the Free Software Foundation.
+
+%%% This software is distributed in the hope that it will be useful, but
+%%% \textbf{without any warranty}; without even the implied warranty of
+%%% \textbf{merchantability} or \textbf{fitness for a particular purpose}.
+%%% See the GNU General Public License for more details.
+
+%%% You should have received a copy of the GNU General Public License in
+%%% the file |License| along with this package; if not, write to the Free
+%%% Software Foundation, Inc., 675~Mass Ave, Cambridge, MA~02139,~USA.
+
+
+%%% \sect Before we start we declare some shorthands for category codes.
+%%% By declaring the underscore~`(|_|)' as letter we can use it in our
+%%% macros. (I agree with {\sc D.~Knuth} that
+%%% |\identifier_several_words_long| is more readable than
+%%% |\IdentifierSeveralWordsLong| and in every case better than
+%%% |\p@@@s|.) As this is a \LaTeX{} style option the at sign is a letter
+%%% anyhow; so we can use the ``private'' Plain and \LaTeX{} macros; and
+%%% with the underscore we can make our own macros more readable.  But as
+%%% we have to restore this category code at the end of this macro file
+%%% we store its former value in the control sequence |\CatUsCode|. This
+%%% method is better than to use a group because not all macros have to
+%%% be defined global this way.
+
+%%% \beginprog
+\ifx \CatEscape\undefined
+    \chardef\CatEscape=0
+    \chardef\CatOpen=1
+    \chardef\CatClose=2
+    \chardef\CatIgnore=9
+    \chardef\CatLetter=11
+    \chardef\CatOther=12
+    \chardef\CatActive=13
+
+    \chardef\CatUsCode=\catcode`\_
+\fi
+
+\catcode`\_=\CatLetter                  % top level macro file
+%%% \endprog
+
+
+
+
+%%% \chap Parsing RCS fields.
+
+%%% Each RCS field consists of a keyword and an optional value. The value
+%%% is delimited by `\verb*|: |' from the keyword, this delimiter is
+%%% missing if no value does exist. At the end of a value there's always
+%%% a space. I.e., the possible input looks like one of the two following
+%%% lines:
+%%%  %
+%%% \begin{quote}
+%%%   \itshape
+%%%   |\RCS $|Keyword\/|$|\\
+%%%   |\RCS $|Keyword\/\verb*|: |value\/\verb*| $|
+%%% \end{quote}
+%%%  %
+%%% where \textit{Keyword} and \textit{value} may be arbitrary strings. In
+%%% fact, we assume that \textit{Keyword} is a sequence of letters---only
+%%% then the creation of a cseq |\RCS|\textit{Keyword} is sensible. Our
+%%% implementation will not produce an error with an obscure keyword, but
+%%% the result will not be of use for the author.
+%%% We must not depend on a fixed set of keywords, this style shall be
+%%% usable with a configurable RCS version. In such a version one can
+%%% bind arbitrary keywords to the information available in RCS\@. This
+%%% means we will parse all RCS fields with the same macro,
+%%% |\rcs_split_field|. This macro will place its results in two macros:
+%%% |\RCS_keyword| and |\RCS_value|. Afterwards we can bind the expansion
+%%% of |\RCS_value| to the target cseq.
+
+
+%%% \sect |\RCS_value| will be set to the expansion of |\RcsEmptyValue|
+%%% if there was no RCS value available, i.e., if the field was not
+%%% processed by RCS yet. This way the values of non-expanded RCS fields
+%%% can be configured. The default value is the empty token list.
+
+%%% \beginprog
+\def\RcsEmptyValue{}
+%%% \endprog
+
+
+%%% \sect For those fields where a special handling is demanded, a cseq
+%%% |\RcsHandle|\textit{Keyword} must exist. This cseq must handle the
+%%% parsing of the value. Usually this means that it will create a macro
+%%% named |\RCS|\textit{Keyword} with a transformed (or filtered) value.
+%%% If you happen to write a handler you might want to look at the
+%%% definition of |\RcsHandleDate| starting at section~\ref{sec:date}
+%%% If a special handler for a field exists, the original value is
+%%% placed automatically in the macro |\RCSRaw|\textit{Keyword}. This is done
+%%% before the handler is called.
+
+
+%%% \sect \label{chunk:value-split} The macro
+%%% |\rcs_split_field| will split the field into |\RCS_keyword| and
+%%% |\RCS_value|. This split is done by the argument parsing capability
+%%% of \TeX{}: We can separate parameters by token lists and \TeX{} will
+%%% use a shortest pattern match to determine the arguments.
+
+%%% Let's look what kind of pattern we can use. As an example we use the
+%%% RCS field `\verb*|$Revision: 2.11 $|', i.e., the unexpanded field is
+%%% `|$Revision: 2.11 $|'. The macro |\RCS| will use dollars to delimit its
+%%% argument, we don't need to worry for them. Eventually we'll have to
+%%% discard the space of the expanded field's value---after all,
+%%% |\RCSRevision| shall expand to `1.1', not to `1.1~'.
+
+%%% First, we can see that the token list $\it (colon, space)$ may be used
+%%% as the seperator of keyword and value:
+%%% %
+%%% \begin{quote}
+%%%   \macroCall{\\rcs_split_field \[Revision\(: \)1.1 \]\\end_value}
+%%% \end{quote}
+%%% %
+%%% ($d_i$ is the separating token list between parameter $i$ and~$i+1$.
+%%% |#1| denotes the argument of |\RCS|. |\end_value| is a stop token,
+%%% used to delimit the last parameter; we have to supply this token at
+%%% the call. In fact, all tokens marked by $s$ and the dotted
+%%% line---i.e., those after |#1|---are supplied by us.)
+
+%%% When we want this token list as a separator, we have to supply it for
+%%% the unexpanded field, since there is no $\it (colon, space)$ in there:
+%%% %
+%%% \begin{quote}
+%%%   \macroCall{\\rcs_split_field \[Revision\]\(: \)\\end_value}
+%%% \end{quote}
+%%% %
+%%% Then the call with the expanded field gets
+%%% %
+%%% \begin{quote}
+%%%   \macroCall{\\rcs_split_field \[Revision\(: \)1.1 \]: \\end_value}
+%%% \end{quote}
+
+%%% Well, we want neither the space nor the colon and the other space
+%%% after the value---we have to get rid of them. We can utilize the fact
+%%% that no RCS field will ever have the token list $\it (space, colon)$
+%%% in it. This means we can use it as a terminator for the RCS value:
+%%% %
+%%% \begin{quote}
+%%%   \macroCall{\\rcs_split_field \[Revision\(: \)1.1\( \]:\) \\end_value}
+%%% \end{quote}
+%%% %
+%%% Please note, one token of the delimiter~$d_2$ comes from the
+%%% |#1|~argument, the other from the tokens we'll add at the call.
+
+%%% Let's check our unexpanded field again---we don't have $d_2$ there. We
+%%% have to add it by the caller:
+%%% %
+%%% \begin{quote}
+%%%   \macroCall{\\rcs_split_field \[Revision\]\(: \)\{\}\( :\)\\end_value}
+%%% \end{quote}
+%%% %
+%%% (Without the empty group the two spaces would be combined into one
+%%% token.) Going back to the expanded field, we notice that these
+%%% additional tokens do not harm us---we'll ignore everything after $d_2$
+%%% anyhow.
+%%% %
+%%% \begin{quote}
+%%%   \macroCall{\\rcs_split_field \[Revision\(: \)1.1\( \]:\) \{\} :\\end_value}
+%%% \end{quote}
+
+%%% We can take another look at the whole affair and point out the
+%%% parameters we get for the split. Remember, that spaces after a control
+%%% word belong to this word. First, the expanded field:
+%%% %
+%%% \begin{quote}
+%%%   \macroCall
+%%%       [p]{\\rcs_split_field \[\(Revision\): \(1.1\) \]:\( \{\} :\)\\end_value}
+%%% \end{quote}
+%%% %
+%%% Then the unexpanded one:
+%%% %
+%%% \begin{quote}
+%%%   \macroCall[p]{\\rcs_split_field \[\(Revision\)\]: \(\{\}\) :\(\|\)\\end_value}
+%%% \end{quote}
+%%% %
+%%% ($p_3$, the third parameter will be empty in this case.)
+
+%%% \medskip
+
+%%% \noindent As explained above: if the value does not exist, a default
+%%% value is supplied.
+
+%%% \beginprog
+\def\rcs_split_field #1: #2 :#3\end_value{%
+    \def\RCS_keyword{#1}%
+    \def\RCS_value{#2}%
+    \ifx \RCS_value\empty
+	\let\RCS_value\RcsEmptyValue
+    \fi
+    }
+%%% \endprog
+
+
+%%% \sect |\RCS| triggers the split. After it has occured, we can bind the
+%%% expansion of |\RCS_value| to a new name. This name is constructed from
+%%% the prefix |RCS|, the string |Raw| iff a handle for this field exists,
+%%% and the keyword itself.
+
+%%% Afterwards we call the handle, on non-existance this will be a no-op.
+
+%%% Gathering the field is a bit more complicated, as we need to pay
+%%% attention to underscores in file names.
+
+%%% \beginprog
+\def\RCS{%
+    \bgroup
+        \catcode`\_ \CatActive
+	\RCS_get_argument
+    }
+\def\RCS_get_argument $#1${%
+        \gdef\RCS_argument{#1}%
+    \egroup
+    \RCS_process
+    }
+
+\def\RCS_process{%
+    \expandafter\rcs_split_field \RCS_argument: {} :\end_value
+    \expandafter\let \csname RCS\rcs_raw_if_handle\RCS_keyword\endcsname
+		     \RCS_value
+    \csname RcsHandle\RCS_keyword\endcsname
+    }
+\def\rcs_raw_if_handle{%
+    \@ifundefined{RcsHandle\RCS_keyword}{}{Raw}%
+    }
+%%% \endprog
+
+
+
+
+%%% \chap Transforming the value of \Date{} fields.
+%%% \label{sec:date}
+
+%%% When we implement the special handling for the \Date{} field, we take
+%%% care that we don't rely on the fact that the keyword is |Date|.
+%%% Certain RCS versions (one of them created at our site) allow the
+%%% configuration of the used keywords. I.e., there might be another
+%%% keyword used for this field, but we still want to use the
+%%% functionality provided here.
+%%% For example, let's assume the keyword of the \Date{} field is
+%%% |StyleDate|. Then the assignment
+%%% %
+%%% \begin{quote}
+%%%   |\let\RcsHandleStyleDate=\RcsHandleDate|
+%%% \end{quote}
+%%% %
+%%% is all that's needed to turn on the `date special handling' for this
+%%% field. (In fact, that's done in this style file you're reading.)
+
+
+%%% \sect The special handler for \Date{} fields sets up
+%%% |\RCS|\textit{Keyword} (|\RCSDate| usually) to produce a textual
+%%% representation of the date part of the \Date{} value in a
+%%% |\today|-like format. Thus the constructed tag can be used in many
+%%% circumstances, e.g., as the argument of the |\date| tag on the
+%%% titlepage.
+
+%%% In addition, |\RCSTime| is defined; it expands to the checkin time.
+
+%%% If the value is empty, i.e., if there's no \Date{} value from RCS
+%%% available, we use the current date. Later we might supply the current
+%%% time, too---currently |\RCSTime| expands to an empty token list in
+%%% such a case.
+%%% Of course, this whole special handling is only done if |\today| is
+%%% defined in some way; otherwise we just handle the \Date{} field like
+%%% every other RCS field: The value is stored unfiltered in
+%%% |\RCS|\textit{Keyword}.
+
+%%% \beginprog
+\def\RcsHandleDate{%
+    \ifx \today\undefined
+	\expandafter\let \csname RCS\RCS_keyword\endcsname \RCS_value
+    \else
+	\ifx \RCS_value\RcsEmptyValue
+	    \@namedef{RCS\RCS_keyword}{\today}%
+	    \let\RCSTime\empty
+	\else
+	    \expandafter\rcs_set_date \RCS_value\end_date
+	\fi
+    \fi
+    }
+%%% \endprog
+
+
+%%% \sect If the value is not empty we split it using the argument parsing
+%%% mechanism of \TeX{}. Then we'll define the value using the current
+%%% definition of |\today|. This way language-specific styles (e.g., from
+%%% the \textsf{babel} system) may supply their own way to present a date.
+
+%%% \beginprog
+\def\rcs_set_date #1/#2/#3 #4\end_date{%
+    \begingroup
+	\day #3			% <-- space!
+	\month #2		% <-- space!
+	\year #1		% <-- space!
+	\expandafter\xdef \csname RCS\RCS_keyword\endcsname {\today}%
+    \endgroup
+    \def\RCSTime{#4}%
+    }
+%%% \endprog
+
+
+
+
+%%% \chap Typesetting revision logs.
+
+%%% If a revision log is to be typeset, we demand a special format.
+%%% Namely, it is to be enclosed in the |rcslog| environment. This
+%%% environment has an optional parameter, the configuration tokens. The
+%%% contents of the environment is the \Log{} field followed by a list of
+%%% revision entries. The \Log{} field must be on a line of its own. Each
+%%% revision entry is introduced by a line consisting of the tag
+%%% |\Revision|, the revision number, the checkin date, the checkin time,
+%%% and the author's user id.
+
+%%% I.e., the input looks like the following example sketch, where italic
+%%% material is substituted by current values.
+%%% %
+%%% \begin{quote}
+%%%   \chardef\\=`\\                        % just for the example
+%%%   \chardef\{=`\{
+%%%   \chardef\}=`\}
+%%%   \begin{alltt}\frenchspacing
+%%% \\begin\{rcslog\}
+%%% $Log: rcs.doc,v $
+%%% Revision 2.11  2003/02/02 22:30:22  schrod
+%%% Removed dependeny on \LaTeX2.09 font selection commands. (Though
+%%% they're still used in the package documentation, as convenience.)
+
+%%% \\Revision {\it revision date time uid}
+%%% \qquad {\it log text\/}
+
+%%% \quad \([\,\ldots\) {\it perhaps more entries\/} \(\ldots\,]\)
+
+%%% \\end\{rcslog\}
+%%%   \end{alltt}
+%%% \end{quote}
+
+%%% We typeset the revision log similar to a |description| environment.
+%%% Each revision entry is an item.
+
+
+%%% \sect The tag for the log intro is implemented as a token register. In
+%%% fact, |\rcsLogIntro| is a hook.
+
+%%% \beginprog
+\newtoks\rcsLogIntro
+	\rcsLogIntro={}
+%%% \endprog
+
+
+%%% \sect The author names are stored in a table named |rcs_author|. We
+%%% supply the lookup macro |\RCS_get_author|, which expands to the table
+%%% value or to the table key itself.
+
+%%% \beginprog
+\def\rcsAuthor#1#2{\@namedef{rcs_author:#1}{#2}}
+\def\RCS_get_author#1{%
+    \@ifundefined{rcs_author:#1}{#1}{\@nameuse{rcs_author:#1}}%
+    }
+%%% \endprog
+
+
+%%% \sect Since this style option shall be usable with many styles, we
+%%% must provide means to configure its behaviour. In particular, it must
+%%% be possible to configure the layout. What remains fixed, is the layout
+%%% of the log like a |description| list. I might add this flexibility if
+%%% there is enough interest (or if somebody sends me tested changes).
+
+%%% The first thing to parametrize is the overall layout and the header.
+%%% This is set up by |\RcsLogStyle|. This cseq \emph{must} define
+%%% |\RcsLogHeading|, which is the macro to set the header. E.g., here
+%%% one could insert a |\newpage| to start a new page for the revision log.
+
+%%% In addition, we reduce the base size of the log. That's the reason why
+%%% these two configurations are combined in one macro---if you change
+%%% one, you also have to consider to change the other.
+
+%%% \beginprog
+\def\RcsLogStyle{%
+    \def\RcsLogHeading{\subsubsection*}%
+    \footnotesize
+    }
+%%% \endprog
+
+
+%%% \sect The header itself is created by |\rcs_log_hdr|, which is called
+%%% with the expanded \Log{} value as its argument, terminated by
+%%% |\end_value|.  We use \TeX{}'s parameter parsing mechanism to chop off
+%%% the tail of the value---if the value exists, it always ends in~`|,v|'.
+%%% Since only an existing value ends in~`|,v|', we have to supply a
+%%% fitting empty value. We will use |\rcs_empty_log_value| for that.
+
+%%% The empty log value gets a bit more complicated: If we have an empty
+%%% \Log{} value, we will not have any revision entries. (Well, except in
+%%% pathological cases.  Then we will produce spurious text.) Missing
+%%% revision entries are missing items; since \LaTeX{} will complain
+%%% about a missing |\item|, we must provide one to stop it from doing
+%%% so. But here is no place for such tokens, the list environment hasn't
+%%% started yet. We will set the flag |@rcs_empty_log@| and check this
+%%% flag after we invoked the list.
+
+%%% \beginprog
+\def\rcs_log_hdr#1,v\end_value{%
+    \@rcs_empty_log@false
+    \RcsLogHeading{\RcsLogHeadingName #1}%
+    \the\rcsLogIntro
+    }
+\def\rcs_empty_log_value{%
+    \global\@rcs_empty_log@true		% that's most probably in a group!
+    \RcsUnknownFile			% text for unknown file name
+    ,v%					% assert calling env
+    }
+\newif\if@rcs_empty_log@
+%%% \endprog
+
+
+%%% \sect The user might want to adapt the text output. He can do so by
+%%% redefining |\RcsLogHeadingName|, |\RcsUnknownFile|, and |\RcsEmptyLog|.
+
+%%% \beginprog
+\def\RcsLogHeadingName{Revision Log for \ttfamily} % file name in monospace
+\def\RcsUnknownFile{\rmfamily\mdseries $\langle\,$Unknown file name$\,\rangle$}
+\def\RcsEmptyLog{{\itshape No log entries available.}}
+%%% \endprog
+
+
+%%% \sect OK, we have fixed the header, let's address the entries. Each
+%%% entry is typeset by |\RcsLogRevision| which is called with four
+%%% arguments: (1)~the revision number, (2)~the checkin date, (3)~the
+%%% checkin time, and (4)~the uid of the author.
+
+%%% Such an entry is inserted as an item in the list. The uid is
+%%% transformed to an author name, if the respective information is
+%%% available. The checkin timestamp is mapped to a neater appearance by
+%%% |\RcsLogDate| and |\RcsLogTime|.
+
+%%% \beginprog
+\def\RcsLogRevision#1#2#3#4{%
+    \item [Revision #1]%
+      (created at \RcsLogDate #2\endDate\space \RcsLogTime{#3} % <-- space!
+      by \RCS_get_author{#4})\\\relax
+    }
+%%% \endprog
+
+
+%%% \sect |\RcsLogDate| sets the date. We can use \TeX{}'s parameter parsing
+%%% mechanism to separate the date value (passed as the argument) in its
+%%% sub-values.  The code is very similar to |\rcs_set_date|.
+
+%%% |\RcsLogTime| sets the time. As explained in the user manual, the time
+%%% is ignored usually. (I don't think that this info is relevant in a
+%%% printout.) Here we assume something about the context it's expanded
+%%% in: Spaces will be before and after the tag. One of them must be
+%%% discarded. The user option |\settime| is implemented here as a
+%%% private macro, too.
+
+%%% \beginprog
+\def\RcsLogDate #1/#2/#3\endDate{%
+    \begingroup
+	\day #3			% <-- space!
+	\month #2		% <-- space!
+	\year #1		% <-- space!
+	\today
+    \endgroup
+    }
+
+\def\RcsLogTime#1{\ignorespaces}
+\def\rcs_settime{%
+    \def\RcsLogTime##1{[##1]}%
+    }
+%%% \endprog
+
+
+%%% \sect Well, that are all possibilities to configure the revision log.
+%%% Let's put together the pieces. Within the |rcslog| environment the
+%%% tags |\Revision| and |\settime| are used, they are implemented in
+%%% the |rcs| namespace.
+
+%%% At the start, we use the optional argument of the environment to
+%%% configure it. |\settime| might be placed here, or font switches, etc.
+
+%%% The next entity shall be the \Log{} field.  In fact, it shall be on
+%%% the next line. If an optional environment argument was there, the
+%%% newline is still in the input and must be discarded before we can use
+%%% |\RCS| to parse the field.
+
+%%% \beginprog
+\def\rcslog{%
+    \@ifnextchar[%			% ] (Emacs)
+	\rcslog_configure
+	{\rcslog_configure[]}%
+    }
+\def\rcslog_configure[#1]{%
+    \let\Revision\rcslog_revision
+    \let\settime\rcs_settime
+    \RcsLogStyle
+    #1%
+    \afterassignment\rcslog_skipcr
+    \let\next
+    }
+%%% \endprog
+
+
+%%% \sect There are users---programmer's life would be so nice without
+%%% them. An empty line (i.e., a |\par|) might have been inserted between
+%%% |\begin{rcslog}| and the \Log{} field. There might be even other
+%%% garbage between them. We have to check these possibilities and provide
+%%% proper responses.
+
+%%% \beginprog
+\def\rcslog_skipcr{%
+    \ifx \next $%
+	\def\next{\rcslog_field $}%
+    \else\ifx \next\par
+	\let\next\rcslog_field
+    \else
+	\PackageError{rcs}{Missing RCS Log field in environment}{%
+The RCS Log field must be the very first text in the rcslog\MessageBreak
+environment. I will discard the rest of your input line and try to\MessageBreak
+recover then. If you think this won't work, exit now.\MessageBreak
+Otherwise press <Return> and cross your fingers!}
+	\let\next\rcslog_discard_line
+    \fi\fi
+    \next
+    }
+%%% \endprog
+
+
+%%% \sect This is the code for the luser. He provided tokens we didn't
+%%% wait for---we discard them and check anew if we've got it right now.
+%%% (Remember, we're waiting for the `|$|' of the \Log{} field.) %$ Emacs
+
+%%% \beginprog
+\def\rcslog_discard_line{%
+    \begingroup
+	\obeylines
+	\rcslog_gobble_line
+    }
+\begingroup
+\obeylines
+\gdef\rcslog_gobble_line #1^^M{%
+    \endgroup%
+    \afterassignment\rcslog_skipcr%
+    \let\next%
+    }
+\endgroup
+%%% \endprog
+
+
+%%% \sect Yep, we've reached a dollar; let's hope that it's our \Log{}
+%%% field. It must be on one line, this line is fetched first and then
+%%% analyzed.
+
+%%% \beginprog
+\def\rcslog_field{%
+    \begingroup
+	\obeylines
+	\catcode`\_ \CatActive
+	\rcslog_get_field
+    }
+
+\begingroup
+\obeylines
+\gdef\rcslog_get_field $#1$^^M{%
+        \gdef\RCS_argument{#1}%
+    \endgroup%
+    \rcslog_set_field%
+    }
+\endgroup
+%%% \endprog
+
+
+%%% \sect When we analyze the \Log{} field, we must supply the default
+%%% value. Within the log list there will be no other RCS fields,
+%%% therefore we can redefine |\RcsEmptyValue| in this environment.
+
+%%% The header gets passed the expanded \Log{} value, as explained above.
+%%% We use |\RCS_value| instead of |\RCSLog| since we don't know which
+%%% RCS keyword is used.
+
+%%% At last, we start the list and check the |@rcs_empty_log@| flag,
+%%% perhaps we have to supply a dummy |\item|.
+
+%%% \beginprog
+\def\rcslog_set_field{%
+    \let\RcsEmptyValue\rcs_empty_log_value
+    \RCS_process
+    \expandafter\rcs_log_hdr \RCS_value\end_value
+    \list{}\RcsLogListStyle		% empty label, layout configurable
+    \if@rcs_empty_log@
+	\item []\RcsEmptyLog		% supply missing item
+    \fi
+    }
+%%% \endprog
+
+
+%%% \sect The setup of the log list is done by a macro from the protected
+%%% interface, the user shall be able to change it. In the default
+%%% supplied, there is no skip between paragraphs, but a quad indentation.
+%%% |rcslog| is here no list environment in the sense like |itemize| or
+%%% |enumerate|, therefore a usage of these environments within an entry
+%%% shall result in top-level item markers.
+
+%%% \textbf{Caveat}:\quad Since we globally set |\@listdepth| at the start
+%%% of the environment, we must not execute |\endlist| that decrements it.
+%%% Instead we use |\endtrivlist| for the termination of the |rcslog|
+%%% environment. It might be, that this must be changed if
+%%% |\RcsLogListStyle| is changed.
+
+%%% \beginprog
+\def\RcsLogListStyle{%
+    \global\@listdepth\z@
+    \labelwidth\z@		    % item label w/o fixed size
+    \itemindent -\leftmargin	    % move label before start of line
+    \advance\itemindent \labelsep   % label was moved too much
+    \def\makelabel##1{\bf ##1}%	    % how to typeset labels
+    \parsep\z@skip		    % no space between pars in items
+    \listparindent 1em		    % <-- space!  indent pars in items
+    }%
+
+\let\endrcslog\endtrivlist
+%%% \endprog
+
+
+%%% \sect The arguments of |\Revision| are separated by spaces and
+%%% terminated by the end of the line. We have to fetch them, then we can
+%%% pass them to |\RcsLogRevision| which does the real work.
+
+%%% \beginprog
+\def\rcslog_revision{%
+    \begingroup
+	\obeylines
+	\rcslog_get_revinfo
+    }
+
+\begingroup
+\obeylines
+\gdef\rcslog_get_revinfo #1 #2 #3 #4^^M{%
+    \endgroup%
+    \RcsLogRevision{#1}{#2}{#3}{#4}%
+    }
+\endgroup
+%%% \endprog
+
+
+
+
+%%% \chap Convenience tags.
+
+%%% As a shortcut we
+%%% support the direct supply of the checkin date for the titlepage
+%%% material. Here we also take care not to depend on the keyword.
+
+%%% \beginprog
+\def\RCSdate $#1${%
+    \RCS $#1$%
+    \expandafter\date \expandafter{\csname RCS\RCS_keyword\endcsname}%
+    }
+%%% \endprog
+
+
+%%% \sect |\RCSID| parses the RCS field and redefines the footline. This
+%%% tag is from Nelson's |rcs| style. In his style he defines that the RCS
+%%% field is passed as an argument to the macro. To be upward compatible,
+%%% we have to support this as well: We look at the next token, if it is
+%%% an opening brace we fetch the param and trigger the `real' invocation
+%%% by |\rcs_id|. If no brace follows, we call |\rcs_id| immediately.
+
+%%% \beginprog
+\def\RCSID{%
+    \@ifnextchar\bgroup
+	\rcsid_get_field
+	\rcs_id
+    }
+\def\rcsid_get_field#1{\rcs_id #1}
+%%% \endprog
+
+
+%%% \sect As a side effect, we'll produce the normal value tags. When we
+%%% define the footline, we must assert that |\RCS_keyword| and
+%%% |\RCS_value| are expanded. (They will get new bindings at the next
+%%% invocation of |\RCS|.) All other macros must not be expanded, they
+%%% shall be used in the context of the footline, not in our definition
+%%% context. We utilize |\protect| for that, |\framebox| is the only macro
+%%% that must be protected.
+
+%%% \beginprog
+\def\rcs_id $#1${%
+    \RCS $#1$%
+    \begingroup
+	\let\protect\noexpand
+	\xdef\@oddfoot{%
+	    \reset@font
+	    \protect\framebox[\textwidth]{\RCS_keyword: \RCS_value}%
+	    }%
+    \endgroup
+    \global\let\@evenfoot\@oddfoot
+    }
+%%% \endprog
+
+
+%%% \sect |\RCSdef| just outputs its parameter before it's passed to
+%%% |\RCS|. This tag is from Tom Verhoeff's |rcs| style. In fact, we're
+%%% not fully compatible: Tom did output `RCS keyword string'. I
+%%% substitute this by `RCS field', this is consistent with the rest of
+%%% this style's description.
+
+%%% \beginprog
+\def\RCSdef $#1${%
+    \typeout{RCS field: $#1$}%
+    \RCS $#1$%
+    }
+%%% \endprog
+
+
+
+
+%%% \chap The end.
+
+%%% Well, after all we're finished with this style. We must not forget to
+%%% restore the underscore catcode.
+
+%%% \beginprog
+\catcode`\_=\CatUsCode
+
+\endinput
+%%% \endprog
+
+
+%%% \sect I would like to thank those who helped me to improve this style.
+%%% % In particular, XXX provided XXXsubstantial parts of the code.
+
+%%% \textsc{Marcus Speh} pointed out that Nelson's style option exists.
+%%% \textsc{Lee Wittenberg} pointed out that Tom style option exists---Tom
+%%% posted the style option on |comp.programming.literate| afterwards. I
+%%% borrowed the idea of appending a second colon to the call of
+%%% |\rcs_split_field| from Tom's style. \textsc{Harald Fuchs} asked for the
+%%% possibility to configure the log list layout.
+
+%%% \textsc{Dave Love} brought the problem with the |-kv| option of |co| to
+%%% my attention. It's a pitty that I haven't found a solution yet. (I
+%%% mean, one that convinces me.)
+
+%%% \textsc{Christine Detig} told me where my documentation was
+%%% uncomprehensible. In particular, you won't have the visual explanation
+%%% of |\rcs_split_field| without her. barbara beeton proofread the user
+%%% manual. In addition, she pointed out where a non-programmer does not
+%%% understand it.
+
+%%% After 7.5 years of quiteness, \textsc{Martin Kneißl} emerged and asked
+%%% for support of underscores in file names. He got it, and I used the
+%%% opportunity to update contact information.
+
+
+
+%%% \vskip \PltxPreSectSkip
+
+%%% \rcsLogIntro{Of course, we shouldn't stop without the RCS log of this
+%%%     style file.}
+
+
+%%% \begin{rcslog}
+%%% $StyleLog: rcs.doc,v $
+%%% \Revision 2.10 2003/02/02 20:42:38 schrod
+%%% Support underscores in file names in RCS field values. E.g., in
+%%% \rcsField{Id} or \Log{}.
+
+%%% Update my contact information.
+
+%%% \Revision 2.9 1995/08/02 12:09:07 schrod
+%%% Transformed this style option into a \LaTeXe{} package.
+
+%%% Updated to \LaTeXe{}.
+
+%%% \Revision 2.8 1993/11/10 12:29:49 schrod
+%%% Added acknowledgement of barbara.
+
+%%% \Revision 2.7 1993/11/08 20:15:15 schrod
+%%% Added possibility to configure the layout of the log list
+%%% (|\RcsLogListStyle|).
+
+%%% Now the log list is on level~0. Thereby item lists or enumerations in
+%%% revision entries get first-level labels. This has consequences for
+%%% the configuration, this implementation strategy must therefore be
+%%% specified in the internal interface description.
+
+%%% \Revision 2.6 1993/11/03 20:04:57 schrod
+%%% Cleaned up for distribution: Added email address to each document,
+%%% added copyright info to |rcs.doc|, added acknowledgements. Checked my
+%%% English and the spacing.
+
+%%% Explained the restriction concerning the |-kv| option of |co|.
+
+%%% Rewrote the explanation to |\rcs_split_field|. Added the visual clues
+%%% to explain the (mis?)usage of macro expansion and pattern matching.
+
+%%% \Revision 2.5 1993/11/02 20:02:51 schrod
+%%% Implemented |\RCSdef|.
+
+%%% Deleted unnecessary macro that did discard the space of an RCS value.
+%%% Moved this in the parameter context of |\rcs_split_field|. Added an
+%%% explanation to |\RCS| where it is called.
+
+%%% \Revision 2.4 1993/11/02 18:44:49 schrod
+%%% Implemented |\RCSID|.\\
+%%% Need |\reset@font| now, for the definition of the footline in |\RCSID|.
+
+%%% Paragraphs in revision logs are not separated by vertical space any
+%%% more. They are indented instead. I use the ``classic'' indentation of
+%%% one quad.
+
+%%% \Revision 2.3 1993/11/01 19:49:49 schrod
+%%% Added subdocument about the internal interface.
+
+%%% Cleaned the cseq names: All public names start with `|RCS|' or
+%%% `|rcs|', protected names start with `|Rcs|', both have only letters.
+%%% Observer names start with `|RCS|', internal names with `|rcs|', both
+%%% feature non-letters in the names.
+
+%%% \Revision 2.2 1993/11/01 19:20:58 schrod
+%%% Cleaned up the documentation. The enhancement of the |rcslog|
+%%% environment was just appended, now it's integrated in the whole
+%%% document. Improved the explanation of the implementation: the valid
+%%% input is now explained more clearly.
+
+%%% Use |rcs-doc.sty| for documentation of |rcs| style. It does not only
+%%% load style options, but does also define |\RCSStyleRevision| for
+%%% access of the style's revision.
+
+%%% Added hint that the user manual is urgently needed.\\
+%%%  The load tag is now in the protected interface.\\
+%%%  Commented out section with definition of |\reset@font|, that cseq is
+%%% not needed in the implementation.\\
+%%%  OALD tells me to use ``heading'' instead of ``header.''
+
+%%% \Revision 2.1 1993/10/29 19:07:38 schrod
+%%% Reorganized the source structure---the style gets too large to keep
+%%% everything in one directory. I use the usual setup for my larger
+%%% styles.
+
+%%% Changed the RCS keywords. All keywords in use now start with
+%%% `|Style|'. Stuff in the text which looks like RCS fields, but where
+%%% the keywords do not start with `|Style|', are examples!
+
+%%% \Revision 1.3 1993/10/29 18:08:19 schrod
+%%% Each field value supplied by RCS ends with a space; this space is
+%%% discarded.
+
+%%% Supports the typesetting of revision logs. The overall documentation
+%%% is not yet fully integrated. Of course, our own revision log is
+%%% typeset at the end, too.\\
+%%%  But we'll have to use other RCS keywords for this document in the
+%%% future. E.g., I can't use original RCS keywords currently in the
+%%% examples without much hassle.
+
+%%% An hook for the configuration of this style option is provided.
+
+%%% \Revision 1.2  1993/09/08  15:49:00  schrod
+%%% Use canonical names for category codes.
+
+%%% \Revision 1.1  1993/09/03  21:01:29  schrod
+%%% Re-implemented |rcs| style option. Made it a documented option.
+
+%%% \end{rcslog}
+
+
+
+
+%%% \end{document}
+
+
+%%% 
+%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
+%%% Local Variables:
+%%% mode: LaTeX
+%%% TeX-brace-indent-level: 4
+%%% indent-tabs-mode: t
+%%% TeX-auto-untabify: nil
+%%% TeX-auto-regexp-list: LaTeX-auto-regexp-list
+%%% End: