%%%%% Active8MacCM.tex
  %%% November 1992, Mar 1994

%%%% NB: 7bit ascii file

%%% Author:
  %% Laurent Siebenmann, Orsay, France
  %% [email protected] <= bug reports here please!

%%% Master Posting:
  %% ftp matups.matups.fr in directory TypingTeX.dir

%%% AIM:
  %% Exploit TeX's active character mechanism
  %% in conjunction with CaesarCM
  %% to allow direct typing of prose for TeX
  %% using all accented characters of Times or other
  %% Macintosh fonts (Helvetica, Palatino etc.)
  %% A fortiori, one can use the less rich but more readable
  %% bitmapped fonts Monaco, Geneva etc.

%%% INSTRUCTIONS:
 %% You have merely  to \input
 %% this file on top of any format in which the
 %%  7-bit typing system for accented characters
 %% CaesarCM has been incorporated.
 %% format-dumper-CM  is a tool to create such
 %% formats.  Then direct typing
 %% of accented characters is acceptable to TeX.
 %% TeX will use a ready-made accented character
 %% if these are available in the current font
 %% and create a composite character otherwise.
 %% LaTeX users must \input the file
 %%   --- AFTER LaTeX.tex and
 %%   --- BEFORE   \documentstyle

 %%%  CASE CHANGE RESTORED!
 %% Such macros as these necessarily fail to
 %% extend the action  of the primitives
 %% \uppercase and \lowercase
 %% to directly typed accented characters.
 %% To replace (or indeed redefine) these, consider
 %% using \Uppercase and \Lowercase
 %% Or \eUppercase and \eLowercase from caesarcm.tex.
 %% See documentation following  \endinput

%%%% \Uppercase and \Lowercase to extend action of
 %%% \uppercase and \lowercase to 128-255.  Robust.
 %%% Misbehave on \ss, \ae, \oe, \o and caps (leave them fixed)
 %%% \eUppercase and \eLowercase of caesarcm.tex
 %%% have perfect behavior --- but are fragile.

\begingroup
 \catcode`\@=11
 \catcode`\/=13

\global\let\primuppercase\uppercase
\global\let\primlowercase\lowercase

%%\count@ is Plain scratch counter \count255
%%% These counts exist globally global despite appearances
\ifx\undefined\Ct@\csname newcount\endcsname\Ct@\fi
\ifx\undefined\Ct@@\csname newcount\endcsname\Ct@@\fi
\ifx\undefined\Ct@@@\csname newcount\endcsname\Ct@@@\fi

  \count@=12\Ct@=128\Ct@@=255
  \loop
    \catcode\Ct@=\count@
  \ifnum\Ct@<\Ct@@\relax
    \advance\Ct@ by \@ne
  \repeat

%%%%% Casing alphabetic Mac Characters
  %%
\def\Alphaholder@{%
  /^^e7^^87./^^83^^8e./^^ea^^92./^^ee^^97./^^f2^^9c.%
  /^^cb^^88./^^e9^^8f./^^ed^^93./^^f1^^98./^^f4^^9d.%
  /^^e5^^89./^^e6^^90./^^eb^^94./^^ef^^99./^^f3^^9e.%
  /^^80^^8a./^^e8^^91./^^ec^^95./^^85^^9a./^^86^^9f.%
  /^^d9^^d8./^^cc^^8b./^^84^^96./^^cd^^9b./^^81^^8c.%
  /^^82^^8d./^^ae^^be./^^ce^^cf./^^af^^bf.%
  }
 %\def\CheckPeriod#1{\if#1.\else %
 %  \errmessage{ !!! Error in character data}\fi}
\def\space{ }

\def/#1#2.{\lccode\number`#1=\number`#2\lccode\number`#2=\number`#2}

  \xdef\LAlphaPrep@{\Alphaholder@}%
    %\show\LAlphaPrep@


\def/#1#2.{\uccode\number`#1=\number`#1\uccode\number`#2=\number`#1}
  \xdef\UAlphaPrep@{\Alphaholder@}%
  %\show\UAlphaPrep@

%%%%% Casing nonalphabetic Mac Characters
\def\Otherholder@{%
 /^^a0./^^a1./^^a2./^^a3./^^a4./^^a5./^^a6./^^a7.%
 /^^a8./^^a9./^^aa./^^ab./^^ac./^^ad.%
 /^^b0./^^b1./^^b2./^^b3./^^b4./^^b5./^^b6./^^b7.%
 /^^b8./^^b9./^^ba./^^bb./^^bc./^^bd.%
 /^^c0./^^c1./^^c2./^^c3./^^c4./^^c5./^^c6./^^c7.%
 /^^c8./^^c9./^^ca./^^d0./^^d1./^^d2./^^d3./^^d4.%
 /^^d5./^^d6./^^d7./^^da./^^db./^^dc./^^dd./^^de.%
 /^^df./^^e0./^^e1./^^e2./^^e3./^^e4./^^f0./^^f5.%
 /^^f6./^^f7./^^f8./^^f9./^^fa./^^fb./^^fc./^^fd.%
 /^^fe./^^ff.}

\def/#1.{\uccode\number`#1\z@}
\xdef\UOtherPrep@{\Otherholder@}
%\show\UOtherPrep@

\def/#1.{\lccode\number`#1\z@}
\xdef\LOtherPrep@{\Otherholder@}
%\show\LOtherPrep@

\gdef\SetCorkULCodeSeg{%
 \loop
  \Ct@=\count@\advance\Ct@ by 32
  \uccode\count@=\count@
  \lccode\count@=\Ct@
  \uccode\Ct@=\count@
  \lccode\Ct@=\Ct@
 \ifnum\count@<\Ct@@
  \relax\advance\count@ by 1
 \repeat
 }

\gdef\SetCorkULCodes{%
 %%% Impose Cork lc uc and sf codes
 \count@=128 \Ct@@=158 \SetCorkULCodeSeg %% segment "80 to "BF
 %%%  Cork "9F is \S and "BF  is pound sign lc/uc codes 0
 \count@=192 \Ct@@=223 \SetCorkULCodeSeg %% segment "C0 to "FF
 }

\gdef\Uppercase#1{\UAlphaPrep@\UOtherPrep@
     \primuppercase{#1}%
     \SetCorkULCodes
     }

\gdef\Lowercase#1{\LAlphaPrep@\LOtherPrep@
     \primlowercase{#1}%
     \SetCorkULCodes
     }

\endgroup

 \begingroup
 \catcode`\@=11
 \global\let\EX@\expandafter

  %\show\CaesarCMLoaded
  {\catcode`!=12
   \ifx\CaesarCMLoaded\undefined
     \immediate\write16{}%
     \immediate\write16{ !!! CaesarCM not loaded}
     \gdef\temp{}\else  \gdef\temp{}
     \fi}
   \temp %\endgroup\end

 { \catcode`!=11
   \ifx\ActiveEight\undefined
     \gdef\temp{}
   \else
     \W@{}%
     \W@{ !!! Stop! An Active8MacCM typing setup has already been
loaded.
     \gdef\temp{\endgroup\end}\else  \gdef\temp{}
     \global\let\ActiveEight\relax
   \fi}
   \temp

  %%% Globally set high ASCII to 13
   %%
  \count@="80\Ct@@="FF
  \loop
    \global\catcode\count@=\active
  \ifnum\count@<\Ct@@\relax
    \advance\count@ by \@ne
  \repeat

\def\CheckPeriod#1{\if#1.\else
       \errmessage{ !!! Error in character data}\fi}

 %%% \lcucActAccDef ^^83^^8eEe\'. %%is example!
  %% Presumably uc and lc codes already set
\def\lcucActAccDef#1#2#3#4#5#6{%
  \CheckPeriod{#6}%
  \global\catcode`#1=\active
  \global\catcode`#2=\active
    \gdef#1{#5#3}%% #3 and #4 ascii catcode 11; #5 is \' say
    \gdef#2{#5#4}%
  }

 %%% \CharAct@ ^^d4^^60. %% is example!
    %% leaves uc lc unresolved
\def\CharAct@#1#2#3{%
 \global\catcode`#1=\active
 \CheckPeriod{#3}%
  \global\let#1#2%
  }

 %%% \QuoteAct@ ^^d4^^60. %% is example!
    %%
\def\QuoteAct@#1.#2#3{%
 \CheckPeriod{#3}%
  \global\catcode`#1=\active
  \xdef#1{\string#2}%
  \global\uccode`#2=0\global\lccode`#2=0
  }

 %%% \Forbid@
    %%
\def\Forbid@ #1{%
   \global\catcode`#1=15 }

 \catcode`/=13
 \let/\lcucActAccDef

 \catcode`\*=13
 \let*=\QuoteAct@

 \catcode`\|=13
 \let|=\CharAct@

%%% / is \lcucActAccDef
 /^^e7^^87Aa\'.
 /^^83^^8eEe\'.
 /^^ea^^92Ii\'.
 /^^ee^^97Oo\'.
 /^^f2^^9cUu\'.
 /^^cb^^88Aa\`.
 /^^e9^^8fEe\`.
 /^^ed^^93Ii\`.
 /^^f1^^98Oo\`.
 /^^f4^^9dUu\`.
 /^^e5^^89Aa\^.
 /^^e6^^90Ee\^.
 /^^eb^^94Ii\^.
 /^^ef^^99Oo\^.
 /^^f3^^9eUu\^.
 /^^80^^8aAa\".
 /^^e8^^91Ee\".
 /^^ec^^95Ii\".
 /^^85^^9aOo\".
 /^^86^^9fUu\".
 /^^d9^^d8Yy\".
 /^^cc^^8bAa\~.
 /^^84^^96Nn\~.
 /^^cd^^9bOo\~.
 /^^81^^8cAa\*.%\show^^81
 /^^82^^8dCc\c.

 %%% | is \CharAct@
 |^^ae\AE.
 |^^be\ae.
 |^^ce\OE.
 |^^cf\oe.
 |^^af\O.
 |^^bf\o.
 |^^a7\ss.

 %%% * is \QuoteAct@
   {\catcode"5C=12 *^^d2.^^5c.}%% backslash code!!
   *^^d3.^^22.%%  ''
   *^^d4.^^60.%%  `
   *^^d5.^^27.%%  '

 \gdef\Symb@@#1{%
   \EX@ \ifx \csname #1@\endcsname\relax
            ^^58%% Print X if nothing better
      \else
            \csname #1@\endcsname
      \fi}

 \xdef\questiondown@{\string^^3e}%
 \xdef\exclamdown@{\string^^3c}%
 %\def\guillemotleft@{\DLG}%% in gllmts.tex
 %\def\guillemotright@{\DRG}%% in gllmts.tex
 \def\dagger@{$\m@th\dag$}%
 \gdef\daggerdbl@{$\m@th\ddag$}%
 \gdef\section@{$\m@th\S$}
 \gdef\paragraph@{$\m@th\P$}
 \gdef\pound@{{\ifdim\fontdimen\@ne\font >\z@
     \it\else \font\ft=cmu10\ft\fi\char`\$}}%% cheat
 %\gdef\paragraph@{X}%% unavailable

    \gdef^^c0{\questiondown}
    \gdef^^c1{\exclamdown}
    \gdef^^c7{\guillemotleft}
    \gdef^^c8{\guillemotright}
    \gdef^^a4{\section}
    \gdef^^a3{\pound}
    \gdef^^a0{\dagger}
    \gdef^^e0{\daggerdbl}
    \gdef^^a6{\paragraph}
    \gdef^^e4{\perthousand}

 %%% If Caesar not present \Symb@@=\Symb@
    \gdef\questiondown{\Symb@@{questiondown}}
    \gdef\exclamdown{\Symb@@{exclamdown}}
    \gdef\guillemotleft{\Symb@@{guillemotleft}}
    \gdef\guillemotright{\Symb@@{guillemotright}}
    \gdef\section{\Symb@@{section}}
    \gdef\pound{\Symb@@{pound}}
    \gdef\dagger{\Symb@@{dagger}}
    \gdef\daggerdbl{\Symb@@{daggerdbl}}
    \gdef\paragraph{\Symb@@{paragraph}}
    \gdef\perthousand{\Symb@@{perthousand}}

 \ifx\CaesarCMLoaded\relax
    \global\let\Symb@@=\Symb@ %% of Caesar
 \fi

  \endgroup

%%%%%% LaTeX material

 { \catcode`!=11 \catcode`\@=11
  \ifx\undefined\@@line %% \@@line unique to LPlain
    \gdef\temp{\endinput}
  \else
    \gdef\temp{}
  \fi}
   \temp  %% \endinput unless LaTeX

 { \catcode`!=11\catcode`\@=11
     \ifx\@@end\undefined %% \@@end  unique to LaTeX
       \W@{}%
       \W@{ !!! LaTeX must be fully loaded }%
       \W@{ !!! before Active8MacCM.tex}%
       \gdef\temp{\end}%
     \else
      \gdef\temp{}
     \fi}
   \temp

 { \catcode`!=11\catcode`\@=11
     \ifx\subsubsection\undefined %% yes before \documentstyle
       \gdef\temp{}
     \else
       \W@{}%
       \W@{ !!! Active8MacCM.tex must be input}%
       \W@{ !!! before \noexpand\documentstyle occurs.}%
       \gdef\temp{\end{document}}%
     \fi}
   \temp

%%%%%% end of LaTeX material

\endinput

%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%
Active8MacCM.doc


       Documentation for Active8MacCM.tex

                  Alpha version, November 1992
                  Beta version, Mar 1994

********************************************************
NOTA BENE In this documentation please read CM for CX and cm for
cx. (The following serves also for the CK (Cork) font encoding
--- reading CK for CX etc.)
********************************************************

   Although Active8MacCX.tex can be used with most TeX formats,
it complements the accent administration package caesarcx.tex in a
well-planned way.  When the two are to be used together, be sure
to obtain versions of Active8MacCX.tex and caesarcx.tex that are
in step. The master postings are currently on ftp
matups.matups.fr. Then on typing a given eight bit Macintosh
character, you will get the best rendition of that character that
is available in the current font.

   There is alarming misbehavior of

 ---  \uppercase, \lowercase

for characters in the range 128-255. Not only does
\uppercase{<eacute>} give <eacute> rather than <Eacute>, but
\uppercase{<Ugrave>} gives <Udieresis>.  Good hyphenation with all
languages served by the Cork norm more or less dictates Cork
standard uc and lc code assignments.  And the misbehavior just
noted is consequence of using them. The one reasonable way to
systematically eliminate the misbehavior is to use "I/O filters"
for TeX to align the code of characters typed to the Cork code.
But few TeX implementations have this desirable flexibility; on
the Macintosh only Direct TeX has it.

   A rather straightforeward remedy is provided by the substitute
macros \Uppercase and \Lowercase defined here. They coincide with
\uppercase and \lowercase  on 7-bit (ascii) TeX input, but are
designed to behave correctly on directly typed Macintosh
characters in the range: 128-255.

   \Uppercase and \Lowercase are robust in the sense that they
are unlikely to cause their argument to blow up no matter how
complex it may be. In particular no \edef-type expansion is used.
Additionally they themselves can hopefully be subjected to
\edef-type expansion. These macros work by temporarily adjusting
many catcodes for the application of \uppercase and \lowercase.
Hence, like, \verbtim macros, they are somewhat slow. Further in
spite of their robustness, these macros may misbehave where
\edef-type expansion is used indescriminately (notably in LaTeX)
the reason is that after expansion the Cork uc and lc codes are
more appropriate, whereas these macros insist on using Macintosh
uc and lc codes.

   A second solution and more ambitions solution is provided
by macros

         \eUppercase, \eLowercase

defined already in the accent administration package caesarcx.tex;
the performance edge is double:

(i)  these macros deal correctly with \ae,\oe,\o,\ss,\i and
capitalized versions, whereas all competeting macros leave them
invariant.
(ii) inlike \Uppercase, \Lowercase often give correct results
when applied to characters in the 128-255 range that have been
subjected to \edef-type expansion.

These macros work by expanding (via the \edef mechanism) until
\uppercase and \lowercase produce the right effect; hence they are
fragile. The "e" in \eUppercase should recall "expansion" or
"\edef". No massive use uc and lc code changes is involved; these
macros are fairly fast.

    \eUppercase, \eLowercase should not be applied to any
material that is not known to be *safe* under \edef-type
expansion; all category 11 and 12 characters are safe; beginners
should regard *all* other material with suspicion, notably font
changes.

    It is tempting to hide the misbehavior of \lowercase and
\uppercase from the  eye of the user by \let'ing the primitive
commands \lowercase and \uppercase equal to \relax.   But this is
in fact dangerous, especially if macros are loaded thereafter,
because programmers use \lowercase and  \uppercase as low level
tools for macro loading.  THOU SHALT NOT GLOBALLY REDEFINE TEX
PRIMITIVES is a commandment deserving respect.

      It is tempting and often very effective to sort out all
case change problems at a blow by redefining via

    \let \uppercase\Uppercase \let \lowercase\Lowercase

or

    \let \uppercase\eUppercase \let \lowercase\eLowercase

although (as explained above) this is antisocial. Use it privately
so long as it gives good results, but  never pass on such a
dangerous system to an unsuspecting third party. Everything within
a TeX group of moderate size can legitimately be considered
private! So this efficient device has  reasonable scope in
practice.  Incidentally, caesarcx.tex Active8MacCX.tex and
preserve intact versions \primuppercase, \primlowercase for
convenience and security.


***** Occasional problems and their current remedies *****

    If TeX issues mysterious complaints in response to the
(local!) redefinition above, try to pacify it by using standard
7-bit  ascii typing there.  Consider also using \ufcr
(=\UpdateFontClassRegister) command of Caesar before the
troublesome part, and selectively replacing the complex \uppercase
by the simple \primuppercase. As a last resort use \noCaesar to
locally to revert to composite characters.

           -----------------------------

    Some special problems occur for LaTeX.  By default a command

   \CaesarProtectLaTeX

is in effect to avoid them.  This makes Caesar use composite
accented characters in its headings. It does so by (re)defining
parameters \LHdShield,  \RHdShield. Other remedies for other
problems are included.

    To undo this at a later point (to perfect the LaTeX
typography) use

   \CaesarDeprotectLaTeX

Then LaTeX may get stuck or lose accented characters in headings
--- especially those involving special font changes and accented
characters. In this case try 7-bit ascii typing for the
troublesome headings.  If that does not work or is too much
bother, revert locally to \CaesarProtectLaTeX.

  Beware that \uppercase is used by LaTeX behind the user's back
and mixed with \edef type expansions.

NB. These difficulties with LaTeX are due to  unnecessary use by
LaTeX of expansion of the sort done by \edef, (and \write, \mark,
etc.).  Hopefully they will all dissappear without trace in
version 3 of LaTeX.  Present versions af LaTeX are fragile, hence
provisional.


***** Related tools *****

(A) Bernd Raichle <[email protected]>, has
a patch for the TeX program offering improved versions of
\uppercase, \lowercase that do not incur any problems, since they
operate after macro expansion, at the horizontal list level.  It
is not yet available.  Raichle's patch and Active8MacCX.tex both
solve case change problems by *delaying* the case changes, but of
course to different degrees and in different ways.

(B) I/O filters available in DirectTeX of W. Ricken let one adopt
the easy solution that uses the Cork code only within TeX. To
enjoy the flexibility of Caesar (use of ready-made accented
characters if they happen to be present), input characters do have
to be active.

(C) The same is accomplished for any TeX on the Macintosh through
use of Cork-encoded screen fonts allied with a special Macintosh
KCHR resource. These have been provided by N. Jungers and are
available on ftp matups.matups.fr.


 ***** Thanks and programmer notes *****

       Comments of Daniel Taupin  and Bernd Raichle in the GUT
TeX forum October 1992 brought into focus the multiple case change
problems stirred up by use of active characters. Taupin noticed
that automatically capitalised headlines in the LaTeX book style
caused difficulties for all then existing 8-bit typing systems
which used active characters to align input to norms for output
(see my Prague proceedings article 1992, where the case change
problem is mentioned). This stirred me to introduce the two
replacement pairs \Uppercase,\Lowercase and \eUppercase,
\eLowercase for \uppercase, \lowercase.  From the programmer's
viewpoint these replacements are the net contribution of these
macros.

       The author's sincere thanks go to them.


***** References *****

---   L. Siebenmann,   *La mise en application d\'efinitive de
la norme de Cork*,   pages 107 to 118 in EuroTeX92, Proceedings of
the 7th European TeX Conference, Prague, Sept 1992, Ed. J. Zlatu\v
ska <[email protected]>, ISBN 80-210-0480-0, Published in
book form, by Czechoslovak TeX Users Group,  Matematick\'y \'ustav
UK, Sokolovsk\'a 83, CS 18600, Praha, Mailorders: 30 DM hardcover
and 25 DM soft cover, preprint possibly also posted on ftp
matups.matups.fr.

---  B. Jackowski and M. Ry\'cko, *Polishing \TeX*, ibid. pages
119-134.

You are viewing proxied material from ftp.icm.edu.pl. The copyright of proxied material belongs to its original authors. Any comments or complaints in relation to proxied material should be directed to the original authors of the content concerned. Please see the disclaimer for more details.