path: root/Master/texmf-dist/doc/fonts/fontinst/test/comparemetrics.sty

% We start by making some default catcode assignments, in case we are 
% using ini\TeX{}.
%    \begin{macrocode}
%<*pkg>
\catcode`\{=1
\catcode`\}=2
\catcode`\#=6
\catcode`\^=7
%</pkg>
%    \end{macrocode}
%
% \package{fontinst} uses some unusual, but convenient, settings of 
% |\catcode|. |@| and |_| are made letters, |~| is made a space, and 
% space and newline are ignored. Before setting those however, we save 
% the current values of the catcodes, so that they can be restored at 
% the end of \texttt{fontinst.sty}.
%
%    \begin{macrocode}
%<*pkg>
\edef\spacecatcode{\the\catcode`\ }
\edef\nlcatcode{\the\catcode`\^^M}
\edef\atcatcode{\the\catcode`\@}
\edef\underscorecatcode{\the\catcode`\_}
\edef\tildecatcode{\the\catcode`\~}
%    \end{macrocode}
%
%    \begin{macrocode}
\catcode`\ =9
\catcode`\^^M=9
\catcode`\@=11
\catcode`\_=11
\catcode`\~=10
%</pkg>
%    \end{macrocode}
%


\newlinechar=`\^^J

% \begin{macro}{\a_read}
% \begin{macro}{\b_read}
% \begin{macro}{\a_filelineno}
% \begin{macro}{\b_filelineno}
%   These control sequences are used for keeping track of the ``physical'' 
%   side of the two input files. |\a_read| and |\b_read| are the stream 
%   numbers. |\a_filelineno| and |\b_filelineno| are the respective 
%   line numbers of the lines most recently read from these files.
%    \begin{macrocode}
\newread\a_read
\newread\b_read 
\newcount\a_filelineno
\newcount\b_filelineno
%    \end{macrocode}
% \end{macro}\end{macro}\end{macro}\end{macro}
% 
% \begin{macro}{\a_lineno}
% \begin{macro}{\b_lineno}
% \begin{macro}{\a_bufbase}
% \begin{macro}{\b_bufbase}
% \begin{macro}{\a_bufmid}
% \begin{macro}{\b_bufmid}
% \begin{macro}{\a_buftop}
% \begin{macro}{\b_buftop}
%   These count registers are used for managing the buffert of lines 
%   that are read from the input files when a match between lines is 
%   being looked for. This buffert consists of lines which are saved in 
%   the macros
%   \begin{quote}
%     |\|\meta{letter}|_saved-|\meta{ofs}
%   \end{quote}
%   where \meta{letter} is |a| or |b| and \meta{ofs} is an integer, 
%   being the index into the buffert. |\a_bufbase| and |\b_bufbase| 
%   contain the \meta{ofs} of the first line currently in the buffert, 
%   whereas |\a_buftop| and |\b_buftop| contain the \meta{ofs} of the 
%   last such line. |\a_bufmid| and |\b_bufmid| hold the buffert 
%   positions of the lines which are currently considered the next by 
%   |\get_next_line|. If \(\mathtt{buftop}<\mathtt{bufmid}\) then the 
%   buffert is empty. |\a_lineno| and |\b_lineno| are the respective 
%   line numbers of the |\a_bufbase| and |\b_bufbase| lines.
%   
%   All these registers should be set globally.
%    \begin{macrocode}
\newcount\a_lineno   \newcount\b_lineno
\newcount\a_bufbase  \newcount\b_bufbase
\newcount\a_bufmid   \newcount\b_bufmid
\newcount\a_buftop   \newcount\b_buftop
%    \end{macrocode}
% \end{macro}\end{macro}\end{macro}\end{macro}\end{macro}\end{macro}
% \end{macro}\end{macro}


% \begin{macro}{\get_next_line}
%   The |\get_next_line| macro is called as
%   \begin{quote}
%     |\get_next_line|\marg{letter}
%   \end{quote}
%   where \meta{letter} is |a| or |b|. This sets the 
%   |\|\meta{letter}|_line| macro to the next line of that file and 
%   updates the value of |\|\meta{letter}|_lineno|.
%   
%   The ``next line'' is normally read from input stream 
%   |\|\meta{letter}|_read|, but if a line mismatch recently occured 
%   then it might happen that the next line has already been read from 
%   the file, and in that case it is taken from one of the 
%   |\|\meta{letter}|_saved-|\meta{ofs} macros.
%    \begin{macrocode}
\def\get_next_line#1{
   \ifnum \csname#1_buftop\endcsname<\csname#1_bufmid\endcsname
      \ifeof \csname#1_read\endcsname
         \x_cs\let{#1_line}\end_of_file_line
      \else
         \read\csname#1_read\endcsname t\x_cs{o}{#1_line}
         \global\advance \csname#1_filelineno\endcsname \@ne
      \fi
   \else
      \expandafter\let \csname#1_line\expandafter\endcsname
         \csname#1_saved-\the\csname#1_bufmid\endcsname\endcsname
   \fi
}
%    \end{macrocode}
% \end{macro}

% \begin{macro}{\end_of_file_line}
%   This macro expands to the text \verb*|(ENDOFFILE) | with the 
%   catcodes it would have if read from an input file. It is used as an 
%   end-of-file marker.
%    \begin{macrocode}
\begingroup
   \catcode`\(=13
   \catcode`\)=13
   \gdef\end_of_file_line{(ENDOFFILE)~}
\endgroup
%    \end{macrocode}
% \end{macro}

% \begin{macro}{\the_line}
%   The |\the_line| macro is called as
%   \begin{quote}
%     |\the_line|\marg{letter}\marg{ofs}
%   \end{quote}
%   where \meta{letter} is |a| or |b|, and \meta{ofs} is a \TeX\ number. 
%   It expands to the control sequence |\|\meta{letter}|_saved-|^^A
%   \meta{ofs}.
%    \begin{macrocode}
\def\the_line#1#2{\csname#1_saved-\number#2\endcsname}
%    \end{macrocode}
% \end{macro}

% \begin{macro}{\initialise_search}
%   The |\initialise_search| macro sets everything up for a search for 
%   a match between lines.
%   
%    \begin{macrocode}
\def\initialise_search{
   \global\a_bufbase=\a_bufmid
   \global\b_bufbase=\b_bufmid
   \add_to_buffert{a}
   \add_to_buffert{b}
   \global\advance \a_bufmid \@ne
   \global\advance \b_bufmid \@ne
}
%    \end{macrocode}
% \end{macro}

% \begin{macro}{\add_to_buffert}
%   The |\add_to_buffert| macro is called as
%   \begin{quote}
%     |\add_to_buffert|\marg{letter}
%   \end{quote}
%   It adds the line returned by |\get_next_line| to the corresponding 
%   buffert (if it wasn't already there) and updates the counters for 
%   that buffert. After this, the line following the line being put into 
%   the buffert will be the line that |\get_next_line| returns.
%    \begin{macrocode}
\def\add_to_buffert#1{
   \ifnum \csname#1_buftop\endcsname<\csname#1_bufmid\endcsname
      \expandafter\let 
         \csname#1_saved-\the\csname#1_bufmid\endcsname
            \expandafter\endcsname
         \csname#1_line\endcsname
      \global\advance \csname#1_buftop\endcsname \@ne
   \fi
}
%    \end{macrocode}
% \end{macro}



\newcount\first_integer
\newcount\second_integer

% \begin{macro}{\match_lines}
%   The |\match_lines| macro tests if two lines can be considered the 
%   same (up to metrics changes, which need not be small). It is called as
%   \begin{quote}
%     |\match_lines|\marg{line a}\marg{line b}
%   \end{quote}
%   \meta{line a} and \meta{line b} may be any amount of text which 
%   eventually expands to the lines in question. |\match_lines| returns 
%   its answer by setting the |_a_| switch to |false| (if the lines 
%   match) or |true| (if the lines don't match).
%   
%   |\match_lines| also causes the |\first_dimen| and |\second_dimen| 
%   registers, as well as the |_histogram_dimen_| switch to be updated, 
%   so it can be immediately followed by a test of the metrics.
%    \begin{macrocode}
\def\match_lines#1#2{
   \first_integer=\z@
   \second_integer=\z@
   \first_dimen=\z@
   \second_dimen=\z@
   \x_cs\let{string-+}\empty_command
   \x_cs\let{string--}\empty_command
%    \end{macrocode}
%    \begin{macrocode}
   \def\sign{+}
   #1\line_stop
   \def\sign{-}
   #2\line_stop
%    \end{macrocode}
%    \begin{macrocode}
   \_a_false
   \expandafter\ifx \csname cmd-+ \expandafter\endcsname
      \csname cmd--\endcsname
   \else  \_a_true  \fi
   \ifnum \first_integer=\z@ \else  \_a_true  \fi
   \ifnum \second_integer=\z@ \else  \_a_true  \fi
   \expandafter\ifx \csname string-+ \expandafter\endcsname
      \csname string--\endcsname
   \else  \_a_true  \fi
}
%    \end{macrocode}
% \end{macro}

% \begin{macro}{\try_matching}
% \begin{macro}{\a_bufP}
% \begin{macro}{\b_bufP}
%   The |\try_matching| descends through the buffert until the |_a_| 
%   switch gets |false| or it has tried the bottommost lines in the 
%   buffert. Since |\body| is already defined by the |\loop| in 
%   |\process_lines|, |\try_matching| uses an explicit end recursion to 
%   make the loop.
%    \begin{macrocode}
\newcount\a_bufP
\newcount\b_bufP
\def\try_matching{
   \global\advance \a_bufP \m@ne
   \global\advance \b_bufP \m@ne
%    % Begin debug
%    \wlog{a_bufbase:~\the\a_bufbase\space\space\space
%          b_bufbase:~\the\b_bufbase}
%    \wlog{a_bufmid:~\the\a_bufmid\space\space\space
%          b_bufmid:~\the\b_bufmid}
%    \wlog{a_buftop:~\the\a_buftop\space\space\space
%          b_buftop:~\the\b_buftop}
%    \wlog{a_bufP:~\the\a_bufP\space\space\space
%          b_bufP:~\the\b_bufP}
%    \begingroup
%       \def\do_left_paren{(}\def\do_right_paren{)}
%       \wlog{a_line:~\a_line}
%       \wlog{b_buf_line:~\the_line{b}{\b_bufP}}
%       \wlog{b_line:~\b_line}
%       \wlog{a_buf_line:~\the_line{a}{\a_bufP}^^J}
%    \endgroup
%    % End debug
   \match_lines{\a_line}{\the_line{b}{\b_bufP}}
   \if_a_
      \match_lines{\the_line{a}{\a_bufP}}{\b_line}
      \if_a_ \else
         \global\a_bufmid=\a_bufP
      \fi
   \else
      \global\b_bufmid=\b_bufP
   \fi
   \ifnum \a_bufP>\a_bufbase \if_a_
      \expandafter\expandafter\expandafter\try_matching
   \fi \fi
}
%    \end{macrocode}
% \end{macro}

% \begin{macro}{\report_mismatch}
%   The |\report_mismatch| macro does precisely what it says. It also increases 
%   the \texttt{bufbase} and \texttt{lineno} registers to point at the 
%   matching lines.
%    \begin{macrocode}
\def\report_mismatch{
   \begingroup
      \def\do_left_paren{(}
      \def\do_right_paren{)}
      \wlog{Mismatch~in~first~file~(at~\the\a_lineno):}
      \loop
         \wlog{\the_line{a}{\a_bufbase}}
      \ifnum \a_bufbase<\a_bufmid
         \global\advance \a_bufbase \@ne
         \global\advance \a_lineno \@ne
      \repeat
      \wlog{Match~found~at~line~\the\a_lineno.}
      \wlog{Mismatch~in~second~file~(at~\the\b_lineno):}
      \loop
         \wlog{\the_line{b}{\b_bufbase}}
      \ifnum \b_bufbase<\b_bufmid
         \global\advance \b_bufbase \@ne
         \global\advance \b_lineno \@ne
      \repeat
      \wlog{Match~found~at~line~\the\b_lineno.}
   \endgroup
   \glyphified_location
   \ifnum 4>\compared_status
      \global\chardef\compared_status=4
   \fi
}
%    \end{macrocode}
% \end{macro}

\def\glyphified_location{
   \ifx \encoding_name\empty_command
      \ifnum \m@ne<\last_character
         \wlog{This~is~related~to~slot~\the\last_character.^^J}
      \else
         \wlog{This~is~in~the~header.^^J}
      \fi
   \else
      \if_undefined{name-\encoding_name-\the\last_character}\then
         \ifnum \last_character=\m@ne
            \wlog{This~is~in~the~header.^^J}
         \else
            \wlog{This~is~related~to~slot~\the\last_character\space
               (unencoded).^^J}
         \fi
      \else
         \wlog{This~is~related~to~the~glyph~
            \csname name-\encoding_name-\the\last_character\endcsname
            .^^J}
      \fi
   \fi
}
% \begin{macro}{\compared_status}
%   |\compared_status| is a \meta{chardef token} which stores the 
%   current status of the comparision, according to the following table
%   \begin{enumerate}
%     \item[0] The property lists are equal---there may be checksum 
%       differences, but nothing else.
%     \item[1] The property lists are almost equal---there are 
%       differences in the metrics, but these are within the |\tolerance| 
%       level.
%     \item[2] The property lists are essentially equal---there are 
%       differences in the metrics, but these are within the |\tolerance| 
%       level, with the exception of some \texttt{CHARHT}, 
%       \texttt{CHARDP}, and \texttt{CHARIC} values.
%     \item[3] The property lists have metric differences---there are 
%       metric differences not with category 2.
%     \item[4] The property lists are not completely matched.
%   \end{enumerate}
% \end{macro}

% \begin{macro}{\step_linenos}
%   The |\step_linenos| macro increments the |_lineno| line numbers. If 
%   the current line was fetched from the buffert, then it also 
%   increments the |_bufmid| and |_bufbase| registers, and checks 
%   whether the buffert ended (in which case it is normalized to its 
%   starting position, to conserve hash table space).
%    \begin{macrocode}
\def\step_linenos{
   \global\advance \a_lineno \@ne
   \ifnum \a_bufmid=\a_buftop
      \global\a_bufbase=\z@
      \global\a_bufmid=\z@
      \global\a_buftop=\m@ne
   \else\ifnum \a_bufmid<\a_buftop
      \global\advance \a_bufmid \@ne
      \global\advance \a_bufbase \@ne
   \fi\fi
   \global\advance \b_lineno \@ne
   \ifnum \b_bufmid=\b_buftop
      \global\b_bufbase=\z@
      \global\b_bufmid=\z@
      \global\b_buftop=\m@ne
   \else\ifnum \b_bufmid<\b_buftop
      \global\advance \b_bufmid \@ne
      \global\advance \b_bufbase \@ne
   \fi\fi
}
%    \end{macrocode}
% \end{macro}


\newdimen\first_dimen
\newdimen\second_dimen
\newcount\ht_hist_i
\newcount\ht_hist_ii
\newcount\ht_hist_iii
\newcount\ht_hist_iv
\newcount\dp_hist_i
\newcount\dp_hist_ii
\newcount\dp_hist_iii
\newcount\dp_hist_iv
\newcount\ic_hist_i
\newcount\ic_hist_ii
\newcount\ic_hist_iii
\newcount\ic_hist_iv

% \begin{macro}{\clear_histograms}
%   This macro clears all the histograms.
%    \begin{macrocode}
\def\clear_histograms{
   \global\ht_hist_i=\z@
   \global\ht_hist_ii=\z@
   \global\ht_hist_iii=\z@
   \global\ht_hist_iv=\z@
   \global\dp_hist_i=\z@
   \global\dp_hist_ii=\z@
   \global\dp_hist_iii=\z@
   \global\dp_hist_iv=\z@
   \global\ic_hist_i=\z@
   \global\ic_hist_ii=\z@
   \global\ic_hist_iii=\z@
   \global\ic_hist_iv=\z@
}
%    \enc{macrocode}
% \end{macro}

% \begin{macro}{\present_histograms}
%   This macro writes the histogram data to the log file.
%    \begin{macrocode}
\def\present_histograms{
   \wlog{\string\tolerance*~\space CHARHT~\space CHARDP~\space CHARIC}
   \wlog{~\space\space 1--~\space\space 10\hist_to_eight{ht}{i}
      \hist_to_eight{dp}{i}\hist_to_eight{ic}{i}}
   \wlog{~\space 10--~\space 100\hist_to_eight{ht}{ii}
      \hist_to_eight{dp}{ii}\hist_to_eight{ic}{ii}}
   \wlog{~100--~1000\hist_to_eight{ht}{iii}\hist_to_eight{dp}{iii}
      \hist_to_eight{ic}{iii}}
   \wlog{1000--10000\hist_to_eight{ht}{iv}\hist_to_eight{dp}{iv}
      \hist_to_eight{ic}{iv}}
}
\def\hist_to_eight#1#2{\x_cs\hist_to_eight_i{#1_hist_#2}}
\def\hist_to_eight_i#1{
   \ifnum #1<\@M
      \four_spaces
      \ifnum #1<\one_hundred
         \space\space
         \ifnum 10>#1 ~\fi
      \else \ifnum #1<\one_thousand \space\fi \fi
   \else
      \ifnum 1000000>#1
         \space\space
         \ifnum 100000>#1 ~\fi
      \else \ifnum 1000000>#1 ~\fi \fi
   \fi
   \the#1
}
%    \end{macrocode}
% \end{macro}


% \begin{macro}{\metric_statistics}
% \begin{macro}{\metric_origin}
%   The |\metric_statistics| macro takes care of comparing the metrics 
%   by inspecting the values of |\first_dimen| and |\second_dimen|. It 
%   has four classes of differences, which are represented by 
%   |\compared_status| values 0--3: no difference, tolerable (almost no)
%   difference, histogrammed (essentially no) difference, and 
%   non-tolerable difference. If a non-tolerable difference is found, 
%   then the current lines are written to the log file. Histogrammed 
%   differences causes the corresponding 
%   |\|\meta{type}|_hist_|\meta{group} register to be incremented, and 
%   tolerable differences only affect the value of |\compared_status|.
%   
%   The |\metric_origin| macro contains the \meta{type} of the metric 
%   data. It is usually empty, but PL properties with metrics that have 
%   histograms set it so that the correct histogram is updated.
%   
%    \begin{macrocode}
\def\metric_statistics{
   \_a_false
   %
   \ifdim \first_dimen<\z@ \first_dimen=-\first_dimen \fi
   \ifdim \first_dimen=\z@
   \else\ifdim \first_dimen<\tolerance
      \ifnum 1>\compared_status
         \global\chardef\compared_status=1
      \fi
   \else\ifx \metric_origin\empty_command
      \_a_true
   \else
      \a_count=\@ne
      \def\a_macro{10}
      \loop\ifdim \a_macro\tolerance<\first_dimen
         \edef\a_macro{\a_macro 0}
         \advance \a_count \@ne
      \repeat
      \if_undefined{\metric_origin _hist_ \romannumeral\a_count}\then
         \_a_true
      \else
         \global\advance \csname \metric_origin _hist_ 
            \romannumeral\a_count \endcsname \@ne
         \ifnum 2>\compared_status
            \global\chardef\compared_status=2
         \fi
      \fi
   \fi\fi\fi
   %
   \ifdim \second_dimen<\z@ \second_dimen=-\second_dimen \fi
   \ifdim \second_dimen=\z@
   \else\ifdim \second_dimen<\tolerance
      \ifnum 1>\compared_status
         \global\chardef\compared_status=1
      \fi
   \else\ifx \metric_origin\empty_command
      \_a_true
   \else
      \a_count=\@ne
      \def\a_macro{10}
      \loop\ifdim \a_macro\tolerance<\second_dimen
         \edef\a_macro{\a_macro 0}
         \advance \a_count \@ne
      \repeat
      \if_undefined{\metric_origin _hist_ \romannumeral\a_count}\then
         \_a_true
      \else
         \global\advance \csname \metric_origin _hist_ 
            \romannumeral\a_count \endcsname \@ne
         \ifnum 2>\compared_status
            \global\chardef\compared_status=2
         \fi
      \fi
   \fi\fi\fi
   %
   \if_a_
      \ifnum 3>\compared_status
         \global\chardef\compared_status=3
      \fi
      \begingroup
         \def\do_left_paren{(}
         \def\do_right_paren{)}
         \wlog{Metric~differences~at~lines~\the\a_lineno\space
            and~\the\b_lineno\space respectively:}
         \wlog{\a_line}
         \wlog{\b_line}
      \endgroup
      \glyphified_location
   \fi
}
%    \end{macrocode}
% \end{macro}\end{macro}

  


\newif\if_b_

% \begin{macro}{\process_lines}
%   The |\process_lines| macro is the cheif executive in the comparision. 
%   Its normal mode of operation is to retrive the next two lines from 
%   the input files, match them, and update the metric differences 
%   statistics. If the lines do not match, however, it starts building a 
%   buffert of lines in which it searches for two matching lines. When 
%   these two lines are found the mismatching portions of the files are 
%   written to the log file, and the metric difference statistics are 
%   updated for the two lines that matched.
%   
%    \begin{macrocode}
\def\process_lines{
   \get_next_line{a}
   \get_next_line{b}
   \match_lines{\a_line}{\b_line}
   \if_a_
%    \end{macrocode}
%   Now the process of trying to find a match begins. At each step of 
%   the search for a match, it tries all possible matches with the 
%   lines read so far that haven't been tried at an earlier step.
%    \begin{macrocode}
%       \tracingmacros=\@ne\tracingcommands=\tw@ % DEBUG
      \initialise_search
      \loop
         \get_next_line{a}
         \get_next_line{b}
         \add_to_buffert{a}
         \add_to_buffert{b}
         \match_lines{\a_line}{\b_line}
         \if_a_
            \global\a_bufP=\a_bufmid
            \global\b_bufP=\b_bufmid
            \try_matching
         \fi
      \if_a_
         \global\advance \a_bufmid \@ne
         \global\advance \b_bufmid \@ne
         \ifnum \a_bufmid>\sixt@@n 
            \errmessage{Big~buffert.~\the\a_lineno[\the\a_filelineno]:
              \the\b_lineno[\the\b_filelineno]}
         \fi % DEBUG
      \repeat
      \report_mismatch
%       \tracingcommands=\z@\tracingmacros=\z@ % DEBUG
%    \end{macrocode}
%    \begin{macrocode}
   \fi
   \metric_statistics
   \step_linenos
   \par % DEBUG
   \if \ifx \a_line\end_of_file_line 
      \ifx \b_line\end_of_file_line 1\else 0\fi\else 0\fi 0
   % then
      \expandafter\process_lines
   \fi
}
%    \end{macrocode}
% \end{macro}

\def\loadencoding#1{{
   \def\do_slot{\x_cs\xdef{name-#1-\the\slot_number}{\slot_name}}
   \inputetx{#1}
}}

% \begin{macro}{\comparePLs}
%   The |\comparePLs| command compares two property lists files. It is 
%   geared against verifying that two different such files are as good 
%   as equal, and it might take a very long time before it gives up on 
%   trying to find any resemblance between them. It is furthermore only 
%   meant to be used on (V)PL files generated by \package{TFtoPL} or 
%   \package{VFtoVP}, because it assumes the property lists follow a 
%   stricter syntax than \package{PLtoTF} and \package{VPtoVF} requires 
%   them to. Finally it needs to be said that the comparision is rather 
%   stupid in many respects, as it is completely line-oriented and 
%   takes no notice of what kind of property list it is reading. This 
%   may in some cases cause it to find ``matches'' between lines which 
%   for syntactical reasons alone cannot match.
%   
%   |\comparePLs| is called as
%   \begin{quote}
%     |\comparePLs|\marg{first PL}\marg{second PL}\marg{encoding}
%   \end{quote}
%   where \meta{first PL} and \meta{second PL} are the complete names 
%   (including extensions) of the files to compare. \meta{encoding} is 
%   the name of the encoding to use when translating slots to glyph 
%   names; leave it empty if you don't want to specify one. (One could 
%   use the \texttt{CODINGSCHEME} property to determine the encoding in 
%   that case, but this is not currently done). The slot to glyph 
%   translation is only done to make it easier to locate the position of 
%   some mismatch; it has not effect on the actual comparision.
%    \begin{macrocode}
\newcount\last_character
\def\comparePLs#1#2#3{{
   \def\encoding_name{#3}
   \last_character=\m@ne
   \openin\a_read=#1\x_relax
   \openin\b_read=#2\x_relax
   \immediate\write\sixt@@n{#1~and^^J #2}
   \pl_catcodes
   \let\do_left_paren=\do_property
   \let\do_right_paren=\empty_command
   \global\chardef\compared_status=\z@
   \clear_histograms
   \global\a_filelineno=\@ne
   \global\b_filelineno=\@ne
   \global\a_lineno=\@ne
   \global\b_lineno=\@ne
   \global\a_bufbase=\z@
   \global\b_bufbase=\z@
   \global\a_bufmid=\z@
   \global\b_bufmid=\z@
   \global\a_buftop=\m@ne
   \global\b_buftop=\m@ne
   \process_lines
   \closein\a_read
   \closein\b_read
   \ifcase \compared_status
      \immediate\write\sixt@@n{are~equal.^^J}
   \or
      \immediate\write\sixt@@n{are~almost~equal.^^J}
   \or
      \immediate\write\sixt@@n{are~essentially~equal.}
   \or
      \immediate\write\sixt@@n{have~non-tolerated~metric~differences.}
   \or
      \immediate\write\sixt@@n{do~not~completely~match.}
   \else
      \immediate\write\sixt@@n{could~not~be~matched!}
   \fi
   \ifnum 1<\compared_status
      \immediate\write\sixt@@n{See~log~file~for~specification.}
      \present_histograms
      \immediate\write\sixt@@n{}
   \fi
}}
%    \end{macrocode}
% \end{macro}   


\begingroup
   \catcode`\(=13 \catcode`\)=13
   \gdef\pl_catcodes{
      \catcode`\(=13 \def({\do_left_paren}
      \catcode`\)=13 \def){\do_right_paren}
   }
\endgroup

\def\do_property#1~{
   \x_cs\edef{cmd-\sign}{#1}
   \let\metric_origin=\empty_command
   \x_cs\ifx{pl-#1}\x_relax
      \expandafter\ignore_parens
   \else
      \csname pl-#1 \expandafter\endcsname
   \fi
}
\def\pl_def_s#1(#2){\x_cs\def{pl-#1}#2~}
\begingroup
   \catcode`\!=13\lccode`\!=`\)
\lowercase{\endgroup
   \def\pl_def_p#1(#2){\x_cs\def{pl-#1}#2!}
}
\def\pl_let#1#2{
   \expandafter\let 
      \csname pl-#1 \expandafter\endcsname \csname pl-#2 \endcsname
}



\newdimen\tolerance  \tolerance=110sp

% The following is for gobbling to the next unmatched right parenthesis. 
% The |\catcode| trick used by the PL-to-MTX converter's 
% |\ignore_parens| cannot be used since the characters are already 
% tokenized, but one can do something similar with an |\edef| since one 
% does not have to worry about the PL files containing material which 
% expands in an undesirable way.
% 
% The |\line_stop| is so that linebreaks in the parenthesis being 
% gobbled will not cause total havok (although the parenthesis won't be 
% gobbled completely). It inserts end-of-group tokens until the |\edef| 
% is terminated, and then the |\afterassignment| token will close the 
% group and restore the normal definition.

\def\ignore_parens{
   \begingroup
   \def\do_left_paren{{\iffalse}\fi}
   \def\do_right_paren{\iffalse{\fi}}
   \def\line_stop{\do_right_paren\line_stop}
   \afterassignment\endgroup
   \edef\a_macro{\iffalse}\fi
}
\def\line_stop{\x_relax}
%    \end{macrocode}
%
\begingroup
   \catcode`\(=1
   \catcode`\)=2
   \gdef\pl_string{
      \begingroup
      \def\do_left_paren{(\iffalse)\fi}
      \def\do_right_paren{\iffalse(\fi)}
      \def\line_stop{\iffalse{\fi}\line_stop}
      \afterassignment\pl_string_i
      \edef\a_macro{\iffalse}\fi
   }
\endgroup
\def\pl_string_i{
   \expandafter\endgroup \expandafter\def 
      \csname string-\sign \expandafter\endcsname 
      \expandafter{\a_macro}
}

\def\pl_real#1{\sign#1\hundred_pt}
\newdimen\hundred_pt  \hundred_pt=100pt

% Convert a PL int to a \TeX{} int, assuming it's prefixed
% by |C|, |D|, |O|, or |H|.
%
%    \begin{macrocode}
\def\pl_int#1#2{
   \sign \ifx#1C `#2
   \else\ifx#1D #2
   \else\ifx#1O '#2
   \else\ifx#1H "#2
   \else -1\errmessage{Unknown~PL~number~prefix~`#1'}
   \fi\fi\fi\fi
}
%    \end{macrocode}


%    \begin{macrocode}
\pl_def_p{NEXTLARGER}(#1~#2){\advance \first_integer \pl_int{#1}{#2}}
\pl_let{BOUNDARYCHAR}{NEXTLARGER}
%    \end{macrocode}
% 
% 
%    \begin{macrocode}
\x_cs\let{pl-VTITLE}=\pl_string
\pl_let{FONTNAME}{VTITLE}
\pl_let{FONTAREA}{VTITLE}
% \pl_let{CHECKSUM}{VTITLE} 
% % Not really a string, but unsigned numbers can become too large for TeX.
% \pl_let{FONTCHECKSUM}{CHECKSUM}
% Commented out since checksums vary too much.
\pl_def_p{FONTAT}(R~#1){\advance \first_dimen \pl_real{#1}}
\pl_let{FONTDSIZE}{FONTAT}
\pl_let{SELECTFONT}{NEXTLARGER}
\pl_let{SETCHAR}{NEXTLARGER}
\pl_def_p{SETRULE}(R~#1~R~#2){
   \advance \first_dimen \pl_real{#1}
   \advance \second_dimen \pl_real{#2}
}
\x_cs\let{pl-POP}\empty_command
\pl_let{PUSH}{POP}
\pl_let{MOVERIGHT}{FONTAT}
\pl_let{MOVELEFT}{FONTAT}
\pl_let{MOVEUP}{FONTAT}
\pl_let{MOVEDOWN}{FONTAT}
\pl_let{SPECIAL}{VTITLE}
\pl_let{SPECIALHEX}{VTITLE}
%    \end{macrocode}
% 
%

\pl_let{CODINGSCHEME}{VTITLE}
\pl_let{DESIGNSIZE}{FONTAT}

\def\pl_parameter#1#2#3{
   \advance \first_integer \pl_int{#1}{#2}
   \advance \first_dimen \pl_real{#3}
}

\pl_def_p{PARAMETER}(#1~#2~R~#3){\pl_parameter{#1}{#2}{#3}}

\pl_def_p{SLANT}(R~#1){\pl_parameter{D}{1}{#1}}
\pl_def_p{SPACE}(R~#1){\pl_parameter{D}{2}{#1}}
\pl_def_p{STRETCH}(R~#1){\pl_parameter{D}{3}{#1}}
\pl_def_p{SHRINK}(R~#1){\pl_parameter{D}{4}{#1}}
\pl_def_p{XHEIGHT}(R~#1){\pl_parameter{D}{5}{#1}}
\pl_def_p{QUAD}(R~#1){\pl_parameter{D}{6}{#1}}
\pl_def_p{EXTRASPACE}(R~#1){\pl_parameter{D}{7}{#1}}

\pl_def_p{NUM1}(R~#1){\pl_parameter{D}{8}{#1}}
\pl_def_p{NUM2}(R~#1){\pl_parameter{D}{9}{#1}}
\pl_def_p{NUM3}(R~#1){\pl_parameter{D}{10}{#1}}
\pl_def_p{DENOM1}(R~#1){\pl_parameter{D}{11}{#1}}
\pl_def_p{DENOM2}(R~#1){\pl_parameter{D}{12}{#1}}
\pl_def_p{SUP1}(R~#1){\pl_parameter{D}{13}{#1}}
\pl_def_p{SUP2}(R~#1){\pl_parameter{D}{14}{#1}}
\pl_def_p{SUP3}(R~#1){\pl_parameter{D}{15}{#1}}
\pl_def_p{SUB1}(R~#1){\pl_parameter{D}{16}{#1}}
\pl_def_p{SUB2}(R~#1){\pl_parameter{D}{17}{#1}}
\pl_def_p{SUPDROP}(R~#1){\pl_parameter{D}{18}{#1}}
\pl_def_p{SUBDROP}(R~#1){\pl_parameter{D}{19}{#1}}
\pl_def_p{DELIM1}(R~#1){\pl_parameter{D}{20}{#1}}
\pl_def_p{DELIM2}(R~#1){\pl_parameter{D}{21}{#1}}
\pl_def_p{AXISHEIGHT}(R~#1){\pl_parameter{D}{22}{#1}}

\pl_let{DEFAULTRULETHICKNESS}{NUM1}
\pl_let{BIGOPSPACING1}{NUM2}
\pl_let{BIGOPSPACING2}{NUM3}
\pl_let{BIGOPSPACING3}{DENOM1}
\pl_let{BIGOPSPACING4}{DENOM2}
\pl_let{BIGOPSPACING5}{SUP1}

\def\boundarychar_name{BOUNDARYCHAR}
\pl_def_s{LABEL}(#1){
   \def\a_macro{#1}
   \ifx \a_macro\boundarychar_name
      \x_cs\let{string-\sign}\a_macro
   \else
      \def\a_macro{\csname pl-LABEL(i\endcsname #1~}
      \expandafter\a_macro
   \fi
}
\pl_def_p{LABEL(i}(#1~#2){
   \last_character=\first_integer
   \advance \first_integer \pl_int{#1}{#2}
}
\pl_def_p{LIG}(#1~#2~#3~#4){
   \advance \first_integer \pl_int{#1}{#2}
   \advance \second_integer \pl_int{#3}{#4}
}
\pl_let{/LIG}{LIG}
\pl_let{/LIG>}{LIG}
\pl_let{LIG/}{LIG}
\pl_let{LIG/>}{LIG}
\pl_let{/LIG/}{LIG}
\pl_let{/LIG/>}{LIG}
\pl_let{/LIG/>>}{LIG}

\pl_let{HEADER}{LIG}

% \def\pl_lig#1~#2~#3~#4~#5~{
%    \x_cs\def{string-\sign}{#1}
% \def\LIG#1~{
%    \if\ifx #1C 0
%       \else\ifx #1D 0
%       \else\ifx #1H 0
%       \else\ifx #1O 0 \else 1\fi\fi\fi\fi 0
%    % then
%       \expandafter\both_of_two
%    \else
%       \expandafter\first_of_two
%    \fi{\pl_lig LIG}{~}#1~
% }
% \def\/{\pl_lig/}

\pl_let{KRN}{PARAMETER}
\pl_let{SKIP}{NEXTLARGER}

\pl_def_s{CHARACTER}(#1~#2){
   \last_character=\first_integer
   \advance \first_integer \pl_int{#1}{#2}
}
\pl_let{CHARWD}{FONTAT}
\pl_def_p{CHARHT}(R~#1){
   \def\metric_origin{ht}
   \advance \first_dimen \pl_real{#1}
}
\pl_def_p{CHARDP}(R~#1){
   \def\metric_origin{dp}
   \advance \first_dimen \pl_real{#1}
}
\pl_def_p{CHARIC}(R~#1){
   \def\metric_origin{ic}
   \advance \first_dimen \pl_real{#1}
}

\pl_let{TOP}{NEXTLARGER}
\pl_let{MID}{NEXTLARGER}
\pl_let{BOT}{NEXTLARGER}
\pl_let{REP}{NEXTLARGER}

\pl_let{FONTDIMEN}{POP}
\pl_let{LIGTABLE}{POP}
\pl_let{VARCHAR}{POP}
\pl_def_s{MAPFONT}(#1~#2){\advance \first_integer \pl_int{#1}{#2}}
\pl_let{MAP}{POP}


%    \begin{macrocode}
\catcode`\@=\atcatcode
\catcode`\^^M=\nlcatcode
\catcode`\ =\spacecatcode
\catcode`\~=\tildecatcode
\catcode`\_=\underscorecatcode
%    \end{macrocode}