% Copyright 1990 - 1995 by AT&T Bell Laboratories.

% Permission to use, copy, modify, and distribute this software
% and its documentation for any purpose and without fee is hereby
% granted, provided that the above copyright notice appear in all
% copies and that both that the copyright notice and this
% permission notice and warranty disclaimer appear in supporting
% documentation, and that the names of AT&T Bell Laboratories or
% any of its entities not be used in advertising or publicity
% pertaining to distribution of the software without specific,
% written prior permission.

% AT&T disclaims all warranties with regard to this software,
% including all implied warranties of merchantability and fitness.
% In no event shall AT&T be liable for any special, indirect or
% consequential damages or any damages whatsoever resulting from
% loss of use, data or profits, whether in an action of contract,
% negligence or other tortious action, arising out of or in
% connection with the use or performance of this software.


% Version 0 was prepared (March 1990).
% Version 0.1 implemented virtual fonts. (May 1990)
% Version 0.30 outputs setbounds statments, ignores rules 1sp wide (May 1991)
% Version 0.60 outputs rules as penstrokes with butt ends
% Version 0.62 makes the output more robust when used in a macro definition
% Version 0.63 new version number only (Unix change file has // path searching)
% Version 0.632 outputs a setbounds path even if no 1sp vrule appears (Jan 1997)
% Version 0.64 avoids outputting wrong fonts, improves error handling (Jan 1998)

% Although considerable effort has been expended to make the DVItoMP program
% correct and reliable, no warranty is implied; the author disclaims any
% obligation or liability for damages, including but not limited to
% special, indirect, or consequential damages arising out of or in
% connection with the use or performance of this software.

% This program is loosely based on DVItype Version 3.0
% It converts a DVI file into a sequence of MetaPost picture expressions.

% TeX is a trademark of the American Mathematical Society.

% Here is TeX material that gets inserted after \input webmac
\def\hang{\hangindent 3em\indent\ignorespaces}
\font\ninerm=cmr9
\font\sc=cmcsc10
\let\mc=\ninerm % medium caps for names like SAIL
\def\MP{MetaPost}
\def\LaTeX{{\rm L\kern-.36em\raise.3ex\hbox{\sc a}\kern-.15em
    T\kern-.1667em\lower.7ex\hbox{E}\kern-.125emX}}
\def\PASCAL{Pascal}

\def\(#1){} % this is used to make section names sort themselves better
\def\9#1{} % this is used for sort keys in the index

\def\title{DVI$\,$\lowercase{to}MP}
\def\contentspagenumber{1}
\def\topofcontents{\null
  \def\titlepage{F} % include headline on the contents page
  \def\rheader{\mainfont\hfil \contentspagenumber}
  \vfill
  \centerline{\titlefont The {\ttitlefont DVItoMP} processor}
  \vskip 15pt
  \centerline{(Version 0.64, January 1998)}
  \vfill}
\pageno=\contentspagenumber \advance\pageno by 1

@* Introduction.
The \.{DVItoMP} program reads binary device-independent (``\.{DVI}'')
files that are produced by document compilers such as \TeX, and converts them
into a symbolic form understood by \MP.  It is loosely based on the \.{DVItype}
utility program that produces a more faithful symbolic form of a \.{DVI} file.

The output file is a sequence of \MP\ picture expressions, one for every page
in the \.{DVI} file.  It makes no difference to \.{DVItoMP} where the \.{DVI}
file comes from, but it is intended to process the result of running \TeX\
or \LaTeX\ on the output of \.{MPtoTEX}.  Such a \.{DVI} file will contain
one page for every \.{btex}$\ldots$\.{etex} block in the original input.
Processing in with \.{DVItoMP} creates a corresponding sequence of \MP\ picture
expressions for use as an auxiliary input file.  Since \MP\ expects such files
to have the extension \.{.MPX}, the output of \.{DVItoMP} is sometimes called
an ``\.{MPX}'' file.

The |banner| string defined here should be changed whenever \.{DVItoMP}
gets modified.

@d banner=='% Written by DVItoMP, Version 0.64'
  {the first line of the output file}

@ This program is written in standard \PASCAL, except where it is necessary
to use extensions; for example, \.{DVItoMP} must read files whose names
are dynamically specified, and that would be impossible in pure \PASCAL.
All places where nonstandard constructions are used have been listed in
the index under ``system dependencies.''
@!@^system dependencies@>

Many \.{DVI}-reading programs need the ability to move to a random place
in a binary file.  \.{DVItoMP} does not need to do this, but it does use
a default |case| as in \.{TANGLE}, \.{WEAVE}, etc.

@ The binary input comes from |dvi_file|, and the symbolic output goes to
|mpx_file|.  \PASCAL's standard |output| file is used only to print an error
message if the \.{DVI} file is bad. The term |print| is used instead of
|write| when this program writes on |mpx_file| just in case this helps
some installations deal with system dependencies.
@^system dependencies@>

@d print(#)==write(mpx_file,#)
@d print_ln(#)==write_ln(mpx_file,#)
@d err_print(#)==write(#)
@d err_print_ln(#)==write_ln(#)

@p program DVI_to_MP(@!dvi_file,@!mpx_file,@!output);
label @<Labels in the outer block@>@/
const @<Constants in the outer block@>@/
type @<Types in the outer block@>@/
var @<Globals in the outer block@>@/
procedure initialize; {this procedure gets things started properly}
  var i:integer; {loop index for initializations}
  begin @<Set initial values@>@/
  end;

@ If the program has to stop prematurely, it goes to the
`|final_end|'. Another label, |done|, is used when stopping normally.

@d final_end=9999 {label for the end of it all}
@d done=30 {go here when finished with a subtask}

@<Labels...@>=final_end,done;

@ The following parameters can be changed at compile time to extend or
reduce \.{DVItoMP}'s capacity.

@<Constants...@>=
@!max_fonts=100; {maximum number of distinct fonts per \.{DVI} file}
@!max_fnums=300; {maximum number of fonts plus fonts local to virtual fonts}
@!max_widths=10000; {maximum number of different characters among all fonts}
@!virtual_space=10000;
  {maximum total bytes of typesetting commands for virtual fonts}
@!line_length=79; {maximum output line length (must be at least 60)}
@!stack_size=100; {\.{DVI} files shouldn't |push| beyond this depth}
@!name_size=1000; {total length of all font file names}
@!name_length=50; {a file name shouldn't be longer than this}

@ There is one more parameter that is a little harder to change because
it is of type |real|.

@d font_tolerance==0.00001
  {font sizes should match to within this multiple of $2^{20}$ \.{DVI} units}

@ Here are some macros for common programming idioms.

@d incr(#) == #:=#+1 {increase a variable by unity}
@d decr(#) == #:=#-1 {decrease a variable by unity}
@d do_nothing == {empty statement}

@ A global variable |history| keeps track of what type of errors have
occurred with the hope that that \MP\ can be warned of any problems.

@d spotless=0 {|history| value when no problems have been found}
@d cksum_trouble=1 {|history| value there have been font checksum mismatches}
@d warning_given=2 {|history| value after a recoverable error}
@d fatal_error=3 {|history| value if processing had to be aborted}

@<Glob...@>=
history:spotless..fatal_error;

@ @<Set init...@>=
history:=spotless;

@ If the \.{DVI} file is badly malformed, the whole process must be aborted;
\.{DVItoMP} will give up, after issuing an error message about the symptoms
that were noticed.

Such errors might be discovered inside of subroutines inside of subroutines,
so a procedure called |jump_out| has been introduced. This procedure, which
simply transfers control to the label |final_end| at the end of the program,
contains the only non-local |goto| statement in \.{DVItoMP}.
@^system dependencies@>

@d abort(#)==begin err_print_ln('DVItoMP abort: ',#);
    history:=fatal_error; jump_out;
    end
@d bad_dvi(#)==abort('Bad DVI file: ',#,'!')
@.Bad DVI file@>
@d warn(#)==begin err_print_ln('DVItoMP warning: ',#);
    history:=warning_given;
    end

@p procedure jump_out;
begin goto final_end;
end;

@* The character set.
Like all programs written with the  \.{WEB} system, \.{DVItoMP} can be
used with any character set. But it uses ASCII code internally, because
the programming for portable input-output is easier when a fixed internal
code is used, and because \.{DVI} files use ASCII code for file names.

The next few sections of \.{DVItoMP} have therefore been copied from the
analogous ones in the \.{WEB} system routines. They have been considerably
simplified, since \.{DVItoMP} need not deal with the controversial
ASCII codes less than @'40 or greater than @'176.
If such codes appear in the font names,
they will be printed as question marks.

@<Types...@>=
@!ASCII_code=" ".."~"; {a subrange of the integers}

@ The original \PASCAL\ compiler was designed in the late 60s, when six-bit
character sets were common, so it did not make provision for lower case
letters. Nowadays, of course, we need to deal with both upper and lower case
alphabets in a convenient way, especially in a program like \.{DVItoMP}.
So we shall assume that the \PASCAL\ system being used for \.{DVItoMP}
has a character set containing at least the standard visible characters
of ASCII code (|"!"| through |"~"|).

Some \PASCAL\ compilers use the original name |char| for the data type
associated with the characters in text files, while other \PASCAL s
consider |char| to be a 64-element subrange of a larger data type that has
some other name.  In order to accommodate this difference, we shall use
the name |text_char| to stand for the data type of the characters in the
output file.  We shall also assume that |text_char| consists of
the elements |chr(first_text_char)| through |chr(last_text_char)|,
inclusive. The following definitions should be adjusted if necessary.
@^system dependencies@>

@d text_char == char {the data type of characters in text files}
@d first_text_char=0 {ordinal number of the smallest element of |text_char|}
@d last_text_char=127 {ordinal number of the largest element of |text_char|}

@<Types...@>=
@!text_file=packed file of text_char;

@ The \.{DVItoMP} processor converts from ASCII code to
the user's external character set by means of an array |xchr|
that is analogous to \PASCAL's |chr| function.

@<Globals...@>=
@!mpx_file:text_file; {destination for printed output}
@!xchr: array [0..255] of text_char;
  {specifies conversion of output characters}

@ To prepare the |mpx_file| for output, we |rewrite| it.
@^system dependencies@>

@p procedure open_mpx_file; {prepares to write text on |mpx_file|}
begin rewrite(mpx_file);
end;

@ Under our assumption that the visible characters of standard ASCII are
all present, the following assignment statements initialize the
|xchr| array properly, without needing any system-dependent changes.

@<Set init...@>=
for i:=0 to @'37 do xchr[i]:='?';
xchr[@'40]:=' ';
xchr[@'41]:='!';
xchr[@'42]:='"';
xchr[@'43]:='#';
xchr[@'44]:='$';
xchr[@'45]:='%';
xchr[@'46]:='&';
xchr[@'47]:='''';@/
xchr[@'50]:='(';
xchr[@'51]:=')';
xchr[@'52]:='*';
xchr[@'53]:='+';
xchr[@'54]:=',';
xchr[@'55]:='-';
xchr[@'56]:='.';
xchr[@'57]:='/';@/
xchr[@'60]:='0';
xchr[@'61]:='1';
xchr[@'62]:='2';
xchr[@'63]:='3';
xchr[@'64]:='4';
xchr[@'65]:='5';
xchr[@'66]:='6';
xchr[@'67]:='7';@/
xchr[@'70]:='8';
xchr[@'71]:='9';
xchr[@'72]:=':';
xchr[@'73]:=';';
xchr[@'74]:='<';
xchr[@'75]:='=';
xchr[@'76]:='>';
xchr[@'77]:='?';@/
xchr[@'100]:='@@';
xchr[@'101]:='A';
xchr[@'102]:='B';
xchr[@'103]:='C';
xchr[@'104]:='D';
xchr[@'105]:='E';
xchr[@'106]:='F';
xchr[@'107]:='G';@/
xchr[@'110]:='H';
xchr[@'111]:='I';
xchr[@'112]:='J';
xchr[@'113]:='K';
xchr[@'114]:='L';
xchr[@'115]:='M';
xchr[@'116]:='N';
xchr[@'117]:='O';@/
xchr[@'120]:='P';
xchr[@'121]:='Q';
xchr[@'122]:='R';
xchr[@'123]:='S';
xchr[@'124]:='T';
xchr[@'125]:='U';
xchr[@'126]:='V';
xchr[@'127]:='W';@/
xchr[@'130]:='X';
xchr[@'131]:='Y';
xchr[@'132]:='Z';
xchr[@'133]:='[';
xchr[@'134]:='\';
xchr[@'135]:=']';
xchr[@'136]:='^';
xchr[@'137]:='_';@/
xchr[@'140]:='`';
xchr[@'141]:='a';
xchr[@'142]:='b';
xchr[@'143]:='c';
xchr[@'144]:='d';
xchr[@'145]:='e';
xchr[@'146]:='f';
xchr[@'147]:='g';@/
xchr[@'150]:='h';
xchr[@'151]:='i';
xchr[@'152]:='j';
xchr[@'153]:='k';
xchr[@'154]:='l';
xchr[@'155]:='m';
xchr[@'156]:='n';
xchr[@'157]:='o';@/
xchr[@'160]:='p';
xchr[@'161]:='q';
xchr[@'162]:='r';
xchr[@'163]:='s';
xchr[@'164]:='t';
xchr[@'165]:='u';
xchr[@'166]:='v';
xchr[@'167]:='w';@/
xchr[@'170]:='x';
xchr[@'171]:='y';
xchr[@'172]:='z';
xchr[@'173]:='{';
xchr[@'174]:='|';
xchr[@'175]:='}';
xchr[@'176]:='~';
for i:=@'177 to 255 do xchr[i]:='?';

@* Device-independent file format.
The format of \.{DVI} files is described in many places including
\.{dvitype.web} and Volume~B of D.~E. Knuth's {\sl Computers and Typesetting}.
This program refers to the following command codes.

@d id_byte=2 {identifies the kind of \.{DVI} files described here}
@#
@d set_char_0=0 {typeset character 0 and move right}
@d set1=128 {typeset a character and move right}
@d set_rule=132 {typeset a rule and move right}
@d put1=133 {typeset a character}
@d put_rule=137 {typeset a rule}
@d nop=138 {no operation}
@d bop=139 {beginning of page}
@d eop=140 {ending of page}
@d push=141 {save the current positions}
@d pop=142 {restore previous positions}
@d right1=143 {move right}
@d w0=147 {move right by |w|}
@d w1=148 {move right and set |w|}
@d x0=152 {move right by |x|}
@d x1=153 {move right and set |x|}
@d down1=157 {move down}
@d y0=161 {move down by |y|}
@d y1=162 {move down and set |y|}
@d z0=166 {move down by |z|}
@d z1=167 {move down and set |z|}
@d fnt_num_0=171 {set current font to 0}
@d fnt1=235 {set current font}
@d xxx1=239 {extension to \.{DVI} primitives}
@d xxx4=242 {potentially long extension to \.{DVI} primitives}
@d fnt_def1=243 {define the meaning of a font number}
@d pre=247 {preamble}
@d post=248 {postamble beginning}
@d post_post=249 {postamble ending}
@d undefined_commands==250,251,252,253,254,255

@* Input from binary files.
We have seen that a \.{DVI} file is a sequence of 8-bit bytes. The bytes
appear physically in what is called a `|packed file of 0..255|'
in \PASCAL\ lingo.

Packing is system dependent, and many \PASCAL\ systems fail to implement
such files in a sensible way (at least, from the viewpoint of producing
good production software).  For example, some systems treat all
byte-oriented files as text, looking for end-of-line marks and such
things. Therefore some system-dependent code is often needed to deal with
binary files, even though most of the program in this section of
\.{DVItoMP} is written in standard \PASCAL.
@^system dependencies@>

One common way to solve the problem is to consider files of |integer|
numbers, and to convert an integer in the range $-2^{31}\L x<2^{31}$ to
a sequence of four bytes $(a,b,c,d)$ using the following code, which
avoids the controversial integer division of negative numbers:
$$\vbox{\halign{#\hfil\cr
|if x>=0 then a:=x div @'100000000|\cr
|else begin x:=(x+@'10000000000)+@'10000000000; a:=x div @'100000000+128;|\cr
\quad|end|\cr
|x:=x mod @'100000000;|\cr
|b:=x div @'200000; x:=x mod @'200000;|\cr
|c:=x div @'400; d:=x mod @'400;|\cr}}$$
The four bytes are then kept in a buffer and output one by one. (On 36-bit
computers, an additional division by 16 is necessary at the beginning.
Another way to separate an integer into four bytes is to use/abuse
\PASCAL's variant records, storing an integer and retrieving bytes that are
packed in the same place; {\sl caveat implementor!\/}) It is also desirable
in some cases to read a hundred or so integers at a time, maintaining a
larger buffer.

We shall stick to simple \PASCAL\ in this program, for reasons of clarity,
even if such simplicity is sometimes unrealistic.

@<Types...@>=
@!eight_bits=0..255; {unsigned one-byte quantity}
@!byte_file=packed file of eight_bits; {files that contain binary data}

@ The program deals with two binary file variables: |dvi_file| is the main
input file that we are translating into symbolic form, and |tfm_file| is
the current font metric file from which character-width information is
being read.  It is convenient to have a throw-away variable for function
results when reading parts of the files that are being skipped.

@<Glob...@>=
@!dvi_file:byte_file; {the input file}
@!tfm_file:byte_file; {a font metric file}
@!vf_file:byte_file; {a virtual font file}
@!down_the_drain:integer; {a ``write-only'' variable}

@ To prepare these files for input, we |reset| them. An extension of
\PASCAL\ is needed in the case of |tfm_file|, since we want to associate
it with external files whose names are specified dynamically (i.e., not
known at compile time). The following code assumes that `|reset(f,s)|'
does this, when |f| is a file variable and |s| is a string variable that
specifies the file name. If |eof(f)| is true immediately after |reset(f,s)|
has acted, these routines assume that no file named |s| is accessible.
@^system dependencies@>

@p procedure open_dvi_file; {prepares to read packed bytes in |dvi_file|}
begin reset(dvi_file);
if eof(dvi_file) then abort('DVI file not found');
end;
@#
function open_tfm_file:boolean; {prepares to read packed bytes in |tfm_file|}
begin reset(tfm_file,cur_name);
open_tfm_file:=(not eof(tfm_file));
end;
@#
function open_vf_file:boolean; {prepares to read packed bytes in |vf_file|}
begin reset(vf_file,cur_name);
open_vf_file:=(not eof(vf_file));
end;

@ If you looked carefully at the preceding code, you probably asked,
``What is |cur_name|?'' Good question. It's a global
variable: |cur_name| is a string variable that will be set to the
current font metric file name before |open_tfm_file| is called.

@<Glob...@>=
@!cur_name:packed array[1..name_length] of char; {external name,
  with no lower case letters}

@ It turns out to be convenient to read four bytes at a time, when we are
inputting from \.{TFM} files. The input goes into global variables
|b0|, |b1|, |b2|, and |b3|, with |b0| getting the first byte and |b3|
the fourth.

@<Glob...@>=
@!b0,@!b1,@!b2,@!b3: eight_bits; {four bytes input at once}

@ The |read_tfm_word| procedure sets |b0| through |b3| to the next
four bytes in the current \.{TFM} file.
@^system dependencies@>

@p procedure read_tfm_word;
begin read(tfm_file,b0); read(tfm_file,b1);
read(tfm_file,b2); read(tfm_file,b3);
end;

@ Input can come from from three different sources depending on the settings
of global variables.  When |vf_reading| is true, we read from the \.{VF} file.
Otherwise, input can either come directly from |dvi_file| or from a buffer
|cmd_buf|.  The latter case applies whenever |buf_ptr<virtual_space|.

@<Glob...@>=
@!vf_reading:boolean; {should input come from |vf_file|?}
@!cmd_buf:packed array [0..virtual_space] of quarter_word;
  {commands for virtual characters}
@!buf_ptr:0..virtual_space; {|cmd_buf| index for the next byte}

@ @<Set init...@>=
vf_reading:=false; buf_ptr:=virtual_space;

@ It is probably not critical that |cmd_buf| be packed as efficiently as possible,
but we define a new type just in case it is necessary to for |cmd_buf| entries
to be in the range |-128..127|.
@^system dependencies@>

@d qi(#)==# {convert from |eight_bits| to |quarter_word|}
@d qo(#)==# {convert from |quarter_word| to |eight_bits|}

@<Types...@>=
quarter_word=0..255; {a one byte quantity as stored in |cmd_buf|}

@ We shall use a set of simple functions to read the next byte or bytes from the
current input source. There are seven possibilities, each of which is treated
as a separate function in order to minimize the overhead for subroutine calls.

@p function get_byte:integer; {returns the next byte, unsigned}
var b:eight_bits;
begin @<Read one byte into |b|@>;
get_byte:=b;
end;
@#
function signed_byte:integer; {returns the next byte, signed}
var b:eight_bits;
begin @<Read one byte into |b|@>;
if b<128 then signed_byte:=b @+ else signed_byte:=b-256;
end;
@#
function get_two_bytes:integer; {returns the next two bytes, unsigned}
var a,@!b:eight_bits;
begin @<Read two bytes into |a| and |b|@>;
get_two_bytes:=a*256+b;
end;
@#
function signed_pair:integer; {returns the next two bytes, signed}
var a,@!b:eight_bits;
begin @<Read two bytes into |a| and |b|@>;
if a<128 then signed_pair:=a*256+b
else signed_pair:=(a-256)*256+b;
end;
@#
function get_three_bytes:integer; {returns the next three bytes, unsigned}
var a,@!b,@!c:eight_bits;
begin @<Read three bytes into |a|, |b|, and~|c|@>;
get_three_bytes:=(a*256+b)*256+c;
end;
@#
function signed_trio:integer; {returns the next three bytes, signed}
var a,@!b,@!c:eight_bits;
begin @<Read three bytes into |a|, |b|, and~|c|@>;
if a<128 then signed_trio:=(a*256+b)*256+c
else signed_trio:=((a-256)*256+b)*256+c;
end;
@#
function signed_quad:integer; {returns the next four bytes, signed}
var a,@!b,@!c,@!d:eight_bits;
begin @<Read four bytes into |a|, |b|, |c|, and~|d|@>;
if a<128 then signed_quad:=((a*256+b)*256+c)*256+d
else signed_quad:=(((a-256)*256+b)*256+c)*256+d;
end;

@ @<Read one byte into |b|@>=
if vf_reading then read(vf_file,b)
else if buf_ptr=virtual_space then read(dvi_file,b)
else begin b:=qo(cmd_buf[buf_ptr]);
  incr(buf_ptr);
  end

@ @<Read two bytes into |a| and |b|@>=
if vf_reading then
  begin read(vf_file,a); read(vf_file,b);
  end
else if buf_ptr=virtual_space then
  begin read(dvi_file,a); read(dvi_file,b);
  end
else if buf_ptr+2>n_cmds then
  abort('Error detected while interpreting a virtual font')
@.Error detected while...@>
else begin a:=qo(cmd_buf[buf_ptr]);
  b:=qo(cmd_buf[buf_ptr+1]);
  buf_ptr:=buf_ptr+2;
  end

@ @<Read three bytes into |a|, |b|, and~|c|@>=
if vf_reading then
  begin read(vf_file,a); read(vf_file,b); read(vf_file,c);
  end
else if buf_ptr=virtual_space then
  begin read(dvi_file,a); read(dvi_file,b); read(dvi_file,c);
  end
else if buf_ptr+3>n_cmds then
  abort('Error detected while interpreting a virtual font')
@.Error detected while...@>
else begin a:=qo(cmd_buf[buf_ptr]);
  b:=qo(cmd_buf[buf_ptr+1]);
  c:=qo(cmd_buf[buf_ptr+2]);
  buf_ptr:=buf_ptr+3;
  end

@ @<Read four bytes into |a|, |b|, |c|, and~|d|@>=
if vf_reading then
  begin read(vf_file,a); read(vf_file,b); read(vf_file,c); read(vf_file,d);
  end
else if buf_ptr=virtual_space then
  begin read(dvi_file,a); read(dvi_file,b); read(dvi_file,c); read(dvi_file,d);
  end
else if buf_ptr+4>n_cmds then
  abort('Error detected while interpreting a virtual font')
@.Error detected while...@>
else begin a:=qo(cmd_buf[buf_ptr]);
  b:=qo(cmd_buf[buf_ptr+1]);
  c:=qo(cmd_buf[buf_ptr+2]);
  d:=qo(cmd_buf[buf_ptr+3]);
  buf_ptr:=buf_ptr+4;
  end

@* Data structures for fonts.
\.{DVI} file format does not include information about character widths, since
that would tend to make the files a lot longer. But a program that reads
a \.{DVI} file is supposed to know the widths of the characters that appear
in \\{set\_char} commands. Therefore \.{DVItoMP} looks at the font metric
(\.{TFM}) files for the fonts that are involved.
@.TFM {\rm files}@>

@ For purposes of this program, the only thing we need to know about a
given character |c| in a non-virtual font |f| is the width.  For the font as
a whole, all we need is the symbolic name to use in the \.{MPX} file.

This information appears implicitly in the following data
structures. The current number of fonts defined is |nf|. Each such font has
an internal number |f|, where |0<=f<nf|.  There is also an external number
that identifies the font in the \.{DVI} file.  The correspondence is
maintained in arrays |font_num| and |internal_num| so that |font_num[i]|
is the external number for |f=internal_num[i]|.
The external name of this font is the string that
occupies positions |font_name[f]| through |font_name[f+1]-1| of the array
|names|. The latter array consists of |ASCII_code| characters, and
|font_name[nf]| is its first unoccupied position. The
legal characters run from |font_bc[f]| to |font_ec[f]|, inclusive.
The \.{TFM} file can specify that some of these are invalid, but this doesn't
concern \.{DVItoMP} because it does not do extensive error checking.
The width of character~|c| in font~|f| is given by
|char_width(f)(c)=width[info_base[f]+c]|, and |info_ptr| is the
first unused position of the |width| array.

If font~|f| is a virtual font, there is a list of \.{DVI} commands for each
character.  These occupy consecutive positions in the |cmd_buf| array with
the commands for character~|c| starting at
|start_cmd(f)(c)=cmd_ptr[info_base[f]+c]| and ending just before
|start_cmd(f)(c+1)|.  Font numbers used when interpreting these \.{DVI}
commands occupy positions |fbase[f]| through |ftop[f]-1| in the |font_num|
table and the |internal_num| array gives the corresponding internal font
numbers.  If such an internal font number~|i| does not correspond to
some font occuring in the \.{DVI} file, then |font_num[i]| has not been
assigned a meaningful value; this is indicated by |local_only[i]=true|.

If font~|f| is not virtual, then |fbase[f]=0| and |ftop[f]=0|.  The |start_cmd|
values are ignored in this case.

@d char_width_end(#)==#]
@d char_width(#)==width[info_base[#]+char_width_end
@d start_cmd(#)==cmd_ptr[info_base[#]+start_cmd_end
@d start_cmd_end(#)==#]

@<Glob...@>=
@!font_num:array [0..max_fnums] of integer; {external font numbers}
@!internal_num:array [0..max_fnums] of integer; {internal font numbers}
@!local_only:array [0..max_fonts] of boolean; {|font_num| meaningless?}
@!font_name:array [0..max_fonts] of 0..name_size; {starting positions
  of external font names}
@!names:array [0..name_size] of ASCII_code; {characters of names}
@!area_length:array [0..max_fonts] of integer;
  {length of area part of font name}
@!font_scaled_size:array [0..max_fonts] of real; {scale factors over $2^{20}$}
@!font_design_size:array [0..max_fonts] of real; {design sizes over $2^{20}$}
@!font_check_sum:array [0..max_fonts] of integer;
  {check sum from the |font_def|}
@!font_bc:array [0..max_fonts] of integer; {beginning characters in fonts}
@!font_ec:array [0..max_fonts] of integer; {ending characters in fonts}
@!info_base:array [0..max_fonts] of integer;
  {index into |width| and |cmd_ptr| tables}
@!width:array [0..max_widths] of integer;
  {character widths, in units $2^{-20}$ of design size}
@!fbase:array [0..max_fonts] of integer; {index into |font_num| for local fonts}
@!ftop:array [0..max_fonts] of integer;
  {|font_num| index where local fonts stop}
@!cmd_ptr:array [0..max_widths] of integer; {starting positions in |cmd_buf|}
@!nf:0..max_fonts; {the number of known fonts}
@!vf_ptr:max_fonts..max_fnums;
  {next |font_num| entry for virtual font font tables}
@!info_ptr:0..max_widths; {allocation pointer for |width| and |cmd_ptr| tables}
@!n_cmds:0..virtual_space; {number of occupied cells in |cmd_buf|}
@!cur_fbase,@!cur_ftop:0..max_fnums;
  {currently applicable part of the |font_num| table}

@ @<Set init...@>=
nf:=0; info_ptr:=0; font_name[0]:=0;
vf_ptr:=max_fnums;
cur_fbase:=0; cur_ftop:=0;

@ Printing the name of a given font is easy except that a procedure |print_char|
is needed to actually send an |ASCII_code| to the \.{MPX} file.

@p @<Declare subroutines for printing strings@>@;
procedure print_font(@!f:integer); {|f| is an internal font number}
var k:0..name_size; {index into |names|}
begin if (f<0)or(f>=nf) then bad_dvi('Undefined font')
else  begin for k:=font_name[f] to font_name[f+1]-1 do
    print_char(names[k]);
  end;
end;

@ Sometimes a font name is needed as part of an error message.

@d font_warn(#)==begin err_print('DVItoMP warning: ',#); font_warn_end
@d font_warn_end(#)==err_print_font(#);
    history:=warning_given;
    end
@d font_abort(#)==begin err_print('DVItoMP abort: ',#); font_abort_end
@d font_abort_end(#)==err_print_font(#);
    history:=fatal_error; jump_out;
    end

@p procedure err_print_font(@!f:integer); {|f| is an internal font number}
var k:0..name_size; {index into |names|}
begin for k:=font_name[f] to font_name[f+1]-1 do
  err_print(xchr[names[k]]);
  err_print_ln(' ');
end;

@ When we encounter a font definition, we save the name, checksum, and size
information, but we don't actaully read the \.{TFM} or \.{VF} file until we
are about to use the font.  If a matching font is not already defined, we then
allocate a new internal font number.

The following subroutine does the necessary things when a \\{fnt\_def} command
is encountered in the \.{DVI} file or in a \.{VF} file.  It assumes that the
first argument has already been parsed and is given by the parameter~|e|.

@p @<Declare a function called |match_font|@>@;
procedure define_font(@!e:integer); {|e| is an external font number}
var @!i:integer; {index into |font_num| and |internal_num|}
@!n:integer; {length of the font name and area}
@!k:integer; {general purpose loop counter}
@!x:integer; {a temporary value for scaled size computation}
begin if nf=max_fonts then abort('DVItoMP capacity exceeded (max fonts=',
    max_fonts:1,')!');
@.DVItoMP capacity exceeded...@>
@<Allocate an index |i| into the |font_num| and |internal_num| tables@>;
@<Read the font parameters into position for font |nf|@>;
internal_num[i]:=match_font(nf,true);
if internal_num[i]=nf then
  begin info_base[nf]:=max_widths; {indicate that the info isn't loaded yet}
  local_only[nf]:=vf_reading; incr(nf);
  end;
end;

@ @<Allocate an index |i| into the |font_num| and |internal_num| tables@>=
if vf_ptr=nf then abort('DVItoMP capacity exceeded (max font numbers=',
    max_fnums:1,')');
@.DVItoMP capacity exceeded...@>
if vf_reading then
  begin font_num[nf]:=0; i:=vf_ptr; decr(vf_ptr);
  end
else i:=nf;
font_num[i]:=e

@ @<Read the font parameters into position for font |nf|@>=
font_check_sum[nf]:=signed_quad;
@<Read |font_scaled_size[nf]| and |font_design_size[nf]|@>;
n:=get_byte; area_length[nf]:=n;
n:=n+get_byte;
if font_name[nf]+n>name_size then
  abort('DVItoMP capacity exceeded (name size=',name_size:1,')!');
@.DVItoMP capacity exceeded...@>
font_name[nf+1]:=font_name[nf]+n;
for k:=font_name[nf] to font_name[nf+1]-1 do names[k]:=get_byte

@ The scaled size and design size are stored in \.{DVI} units divided by $2^20$.
The units for scaled size are a little different if we are reading a virtual
font, but this will be corrected when the scaled size is used.  The scaled size
also needs to be truncated to at most 23 significant bits in order to make
the character width calculation match what \TeX\ does.

@<Read |font_scaled_size[nf]| and |font_design_size[nf]|@>=
x:=signed_quad;
k:=1;
while x>@'40000000 do
  begin x:=x div 2; k:=k+k;
  end;
font_scaled_size[nf]:=x*k/1048576.0;
if vf_reading then
  font_design_size[nf]:=signed_quad*dvi_per_fix/1048576.0
else font_design_size[nf]:=signed_quad/1048576.0;

@ @<Glob...@>=
@!dvi_per_fix:real; {converts points scaled $2^{20}$ to \.{DVI} units}

@ The |match_font| function tries to find a match for the font with internal
number~|ff|, returning |nf| or the number of the matching font.  If
|exact=true|, the name and scaled size should match.  Otherwise the scaled
size need not match but the font found must be already loaded, not just
defined.

@<Declare a function called |match_font|@>=
function match_font(ff:integer; exact:boolean):integer;
label done, 99;
var @!f:0..max_fonts; {font number being tested}
@!ss,@!ll:0..name_size; {starting point and length of name of font |ff|}
@!k,@!s:0..name_size; {registers for comparing font names}
begin ss:=font_name[ff]; ll:=font_name[ff+1]-ss;
f:=0;
while f<nf do
  begin if f<>ff then
    begin @<Compare the names of fonts |f| and |ff|; |goto 99| if they differ@>;
    if exact then
      begin if abs(font_scaled_size[f]-font_scaled_size[ff])
          <= font_tolerance then 
        begin if not vf_reading then
          if local_only[f] then
            begin font_num[f]:=font_num[ff]; local_only[f]:=false;
            end
          else if font_num[f]<>font_num[ff] then goto 99;
        goto done;
        end;
      end
    else if info_base[f]<>max_widths then goto done;
    end;
  99:incr(f);
  end;
done:if f<nf then
  @<Make sure fonts |f| and |ff| have matching design sizes and checksums@>;
match_font:=f;
end;

@ @<Compare the names of fonts |f| and |ff|; |goto 99| if they differ@>=
if (area_length[f]<area_length[ff]) or (ll<>font_name[f+1]-font_name[f]) then
  goto 99;
s:=font_name[f];
k:=ll;
while k>0 do
  begin decr(k);
  if names[s+k]<>names[ss+k] then goto 99;
  end

@ @<Make sure fonts |f| and |ff| have matching design sizes and checksums@>=
if abs(font_design_size[f]-font_design_size[ff]) > font_tolerance then
  font_warn('Inconsistent design sizes given for ')(ff)
@.Inconsistent design sizes@>
else if font_check_sum[f]<>font_check_sum[ff] then
  font_warn('Checksum mismatch for ')(ff)
@.Checksum mismatch@>

@* Reading ordinary fonts.
An auxiliary array |in_width| is used to hold the widths as they are
input. The global variable |tfm_check_sum| is set to the check sum that
appears in the current \.{TFM} file.

@<Glob...@>=
@!in_width:array[0..255] of integer; {\.{TFM} width data in \.{DVI} units}
@!tfm_check_sum:integer; {check sum found in |tfm_file|}

@ Here is a procedure that absorbs the necessary information from a
\.{TFM} file, assuming that the file has just been successfully reset
so that we are ready to read its first byte. (A complete description of
\.{TFM} file format appears in the documentation of \.{TFtoPL} and will
not be repeated here.) The procedure does not check the \.{TFM} file
for validity, nor does it give explicit information about what is
wrong with a \.{TFM} file that proves to be invalid. The procedure simply
aborts the program if it detects anything amiss in the \.{TFM} data.

@p procedure in_TFM(@!f:integer);
  {input \.{TFM} data for font |f| or abort}
label 9997, {go here when the format is bad}
  9999;  {go here to exit}
var k:integer; {index for loops}
@!lh:integer; {length of the header data, in four-byte words}
@!nw:integer; {number of words in the width table}
@!wp:0..max_widths; {new value of |info_ptr| after successful input}
begin @<Read past the header data; |goto 9997| if there is a problem@>;
@<Store character-width indices at the end of the |width| table@>;
@<Read the width values into the |in_width| table@>;
@<Move the widths from |in_width| to |width|@>;
fbase[f]:=0; ftop[f]:=0;
info_ptr:=wp; goto 9999;
9997: font_abort('Bad TFM file for ')(f);
@.Bad TFM file@>
9999: end;

@ @<Read past the header...@>=
read_tfm_word; lh:=b2*256+b3;
read_tfm_word; font_bc[f]:=b0*256+b1; font_ec[f]:=b2*256+b3;
if font_ec[f]<font_bc[f] then font_bc[f]:=font_ec[f]+1;
if info_ptr+font_ec[f]-font_bc[f]+1>max_widths then
  abort('DVItoMP capacity exceeded (width table size=',max_widths:1,')!');
@.DVItoMP capacity exceeded...@>
wp:=info_ptr+font_ec[f]-font_bc[f]+1;
read_tfm_word; nw:=b0*256+b1;
if (nw=0)or(nw>256) then goto 9997;
for k:=1 to 3+lh do
  begin if eof(tfm_file) then goto 9997;
  read_tfm_word;
  if k=4 then
    if b0<128 then tfm_check_sum:=((b0*256+b1)*256+b2)*256+b3
    else tfm_check_sum:=(((b0-256)*256+b1)*256+b2)*256+b3;
  end;

@ @<Store character-width indices...@>=
if wp>0 then for k:=info_ptr to wp-1 do
  begin read_tfm_word;
  if b0>nw then goto 9997;
  width[k]:=b0;
  end;

@ No fancy width calculation is needed here because \.{DVItoMP} stores
widths in their raw form as multiples of the design size scaled by $2^{20}$.
The |font_scaled_size| entries have been computed so that the final width
compution can be done in floating point if enough precision is available.

@<Read the width values into the |in_width| table@>=
for k:=0 to nw-1 do
  begin read_tfm_word;
  if b0>127 then b0:=b0-256;
  in_width[k]:=((b0*@'400+b1)*@'400+b2)*@'400+b3;
  end

@ The width compution uses a scale factor |dvi_scale| that will be introduced
later.  It is equal to one when not typesetting a character from a virtual
font.  In that case, the following expressions do the width computation that is
so important in \.{DVItype}.  It is less important here because it is impractical
to guarantee precise character positioning in \MP\ output.  Nevertheless, the
width compution will be precise if reals have at least 46-bit mantissas and
|round(x-.5)| is equivalent to $\lfloor x\rfloor$.  It may be a good idea to
modify this computation if these conditions are not met.
@^system dependencies@>

@<Width of character |c| in font |f|@>=
round(dvi_scale*font_scaled_size[f]*char_width(f)(c)-0.5)

@ @<Width of character |p| in font |cur_font|@>=
round(dvi_scale*font_scaled_size[cur_font]*char_width(cur_font)(p)-0.5)

@ @<Move the widths from |in_width| to |width|@>=
if in_width[0]<>0 then goto 9997; {the first width should be zero}
info_base[f]:=info_ptr-font_bc[f];
if wp>0 then for k:=info_ptr to wp-1 do
  width[k]:=in_width[width[k]]


@* Reading virtual fonts.
The |in_VF| procedure absorbs the necessary information from a \.{VF} file that
has just been reset so that we are ready to read the first byte.  (A complete
description of \.{VF} file format appears in the documention of \.{VFtoVP}).
Like |in_TFM|, this procedure simply aborts the program if it detects anything
wrong with the \.{VF} file.

@p @<Declare a function called |first_par|@>@;
procedure in_VF(f:integer);
  {read \.{VF} data for font |f| or abort}
label 9997, {go here to abort}
  9999; {go here to exit}
var @!p:integer; {a byte from the \.{VF} file}
@!was_vf_reading:boolean; {old value of |vf_reading|}
@!c:integer; {the current character code}
@!limit:integer; {space limitations force character codes to be less than this}
@!w:integer; {a \.{TFM} width being read}
begin was_vf_reading:=vf_reading; vf_reading:=true;
@<Start reading the preamble from a \.{VF} file@>;@/
@<Initialize the data structures for the virtual font@>;@/
p:=get_byte;
while p>=fnt_def1 do
  begin if p>fnt_def1+3 then goto 9997;
  define_font(first_par(p));
  p:=get_byte;
  end;
while p<=242 do
  begin if eof(vf_file) then goto 9997;
  @<Read the packet length, character code, and \.{TFM} width@>;
  @<Store the character packet in |cmd_buf|@>;
  p:=get_byte;
  end;
if p=post then
  begin @<Finish setting up the data structures for the new virtual font@>;
  goto 9999;
  end;
9997:font_abort('Bad VF file for ')(f);
9999: vf_reading:=was_vf_reading;
end;

@ @<Start reading the preamble from a \.{VF} file@>=
p:=get_byte;
if p<>pre then goto 9997;
p:=get_byte; {fetch the identification byte}
if p<>202 then goto 9997;
p:=get_byte; {fetch the length of the introductory comment}
while p>0 do
  begin decr(p); down_the_drain:=get_byte;
  end;
tfm_check_sum:=signed_quad;
down_the_drain:=signed_quad; {skip over the design size}

@ @<Initialize the data structures for the virtual font@>=
ftop[f]:=vf_ptr;
if vf_ptr=nf then abort('DVItoMP capacity exceeded (max font numbers=',
    max_fnums:1,')');
@.DVItoMP capacity exceeded...@>
decr(vf_ptr);
info_base[f]:=info_ptr;
limit:=max_widths-info_base[f];@/
font_bc[f]:=limit; font_ec[f]:=0

@ @<Read the packet length, character code, and \.{TFM} width@>=
if p=242 then
  begin p:=signed_quad; c:=signed_quad; w:=signed_quad;
  if c<0 then goto 9997;
  end
else begin c:=get_byte; w:=get_three_bytes;
  end;
if c>=limit then
  abort('DVItoMP capacity exceeded (max widths=', max_widths:1,')!');
@.DVItoMP capacity exceeded...@>
if c<font_bc[f] then font_bc[f]:=c;
if c>font_ec[f] then font_ec[f]:=c;
char_width(f)(c):=w

@ @<Store the character packet in |cmd_buf|@>=
if n_cmds+p>=virtual_space then
  abort('DVItoMP capacity exceeded (virtual font space=',virtual_space:1,')!');
@.DVItoMP capacity exceeded...@>
start_cmd(f)(c):=n_cmds;
while p>0 do
  begin cmd_buf[n_cmds]:=qi(get_byte);
  incr(n_cmds); decr(p);
  end;
cmd_buf[n_cmds]:=qi(eop); {add the end-of-packet marker}
incr(n_cmds)

@ There are unused |width| and |cmd_ptr| entries if |font_bc[f]>0| but it isn't
worthwhile to slide everything down just to save a little space.

@<Finish setting up the data structures for the new virtual font@>=
fbase[f]:=vf_ptr+1;
info_ptr:=info_base[f]+font_ec[f]


@* Loading fonts.
The character width information for a font is loaded when the font is selected
for the first time.  This information might already be loaded if the font has
already been used at a different scale factor.  Otherwise, we look for a \.{VF}
file, or failing that, a \.{TFM} file.  All this is done by the |select_font|
function that takes an external font number~|e| and returns the corresponding
internal font number with the width information loaded.

@p function select_font(@!e:integer):integer;
var @!f:0..max_fonts; {the internal font number}
@!ff:0..max_fonts; {internal font number for an existing version}
@!k,@!l:integer; {general purpose loop counters}
begin @<Set |f| to the internal font number that corresponds to |e|,
  or |abort| if there is none@>;
if info_base[f]=max_widths then
  begin ff:=match_font(f,false);
  if ff<nf then @<Make font |f| refer to the width information from font |ff|@>
  else begin @<Move the \.{VF} file name into the |cur_name| string@>;
    if open_vf_file then in_VF(f)
    else  begin @<Change \.{.VF} to \.{.TFM} in the |cur_name| string@>;
      if not open_tfm_file then font_abort('No TFM file found for ')(f);
  @.no TFM file found@>
      in_TFM(f);
      end;
    @<Make sure the checksum in the font file matches the one given in the
      |font_def| for font |f|@>;
    end;
  @<Do any other initialization required for the new font |f|@>;
  end;
select_font:=f;
end;

@ @<Set |f| to the internal font number that corresponds to |e|,...@>=
if cur_ftop<=nf then cur_ftop:=nf;
font_num[cur_ftop]:=e;
k:=cur_fbase;
while (font_num[k]<>e)or local_only[k] do incr(k);
if k=cur_ftop then abort('Undefined font selected');
f:=internal_num[k]

@ @<Make font |f| refer to the width information from font |ff|@>=
begin font_bc[f]:=font_bc[ff];
font_ec[f]:=font_ec[ff];
info_base[f]:=info_base[ff];
fbase[f]:=fbase[ff];
ftop[f]:=ftop[ff];
end

@ If |area_length[f]=0|, i.e., if no font directory has been specified,
\.{DVItoMP} is supposed to use the default font directory, which is a
system-dependent place where the standard fonts are kept.
The string variable |default_directory| contains the name of this area.
@^system dependencies@>

@d default_directory_name=='TeXfonts:' {change this to the correct name}
@d default_directory_name_length=9 {change this to the correct length}

@<Glob...@>=
@!default_directory:packed array[1..default_directory_name_length] of char;

@ @<Set init...@>=
default_directory:=default_directory_name;

@ The string |cur_name| is supposed to be set to the external name of the
\.{VF} file for the current font. This usually means that we need to
prepend the name of the default directory, and
to append the suffix `\.{.VF}'. Furthermore, we change lower case letters
to upper case, since |cur_name| is a \PASCAL\ string.
@^system dependencies@>

@<Move the \.{VF} file name into the |cur_name| string@>=
for k:=1 to name_length do cur_name[k]:=' ';
if area_length[f]=0 then
  begin for k:=1 to default_directory_name_length do
    cur_name[k]:=default_directory[k];
  l:=default_directory_name_length;
  end
else l:=0;
for k:=font_name[f] to font_name[f+1]-1 do
  begin incr(l);
  if l+3>name_length then
    abort('DVItoMP capacity exceeded (max font name length=',
      name_length:1,')!');
@.DVItoMP capacity exceeded...@>
  if (names[k]>="a")and(names[k]<="z") then
      cur_name[l]:=xchr[names[k]-@'40]
  else cur_name[l]:=xchr[names[k]];
  end;
cur_name[l+1]:='.'; cur_name[l+2]:='V'; cur_name[l+3]:='F'

@ It is fairly simple to change a \.{VF} file name to a \.{TFM} file name.
@^system dependencies@>

@<Change \.{.VF} to \.{.TFM} in the |cur_name| string@>=
l:=area_length[f];
if l=0 then l:=default_directory_name_length;
l:=l+font_name[f+1]-font_name[f];
if l+4>name_length then
  abort('DVItoMP capacity exceeded (max font name length=',
    name_length:1,')!');
@.DVItoMP capacity exceeded...@>
cur_name[l+2]:='T'; cur_name[l+3]:='F'; cur_name[l+4]:='M'

@ @<Make sure the checksum in the font file matches the one given in the...@>=
begin if (font_check_sum[f]<>0)and(tfm_check_sum<>0)and@|
    (font_check_sum[f]<>tfm_check_sum) then
  begin err_print('DVItoMP warning: Checksum mismatch for ');
@.Checksum mismatch@>
  err_print_font(f);
  if history=spotless then history:=cksum_trouble;
  end;
end

@* Low level output routines.
One of the basic output operations is to write a \MP\ string expression for
a sequence of characters to be typeset.  The main difficulties are that such
strings can contain arbitrary eight-bit bytes and there is no fixed limit on
the length of the string that needs to be produced.  In extreme cases this
can lead to expressions such as
$$\vcenter{
    \hbox{\.{char7\&char15\&char31\&"?FWayzz"}}
    \hbox{\.{\&"zzaF"\&char15\&char3\&char31}}
    \hbox{\.{\&"Nxzzzzzzzwvtsqo"}}}
$$

@ A global variable |state| keeps track of the output process.
When |state=normal| we have begun a quoted string and the next character
should be a printable character or a closing quote.  When |state=special|
the last thing printed was a ``\.{char}'' construction or a closing quote
and an ampersand should come next.  The starting condition |state=initial|
is a lot like |state=special|, except no ampersand is required.

@d special=0 {the |state| after printing a ``\.{char}'' expression}
@d normal=1 {the |state| value in a quoted string}
@d initial=2 {initial |state|}

@<Glob...@>=
state:special..initial; {controls the process of printing a string}
print_col:0..line_length;
  {there are at most this many characters on the current line}

@ To print a string on the \.{MPX} file, initialize |print_col|, ensure that
|state=initial|, and pass the characters one-at-a-time to |print_char|.

@<Declare subroutines for printing strings@>=
procedure print_char(@!c:eight_bits);
var @!printable:boolean; {is it safe to print |xchr[c]|?}
@!l:integer; {number of characters to print |c| or the \.{char} expression}
begin printable:=(c>=" ")and(c<="~")and(c<>"""");
if printable then l:=1
else if c<10 then l:=5
else if c<100 then l:=6
else l:=7;
if print_col+l>line_length-2 then
  begin if state=normal then
    begin print('"'); state:=special;
    end;
  print_ln(' ');
  print_col:=0;
  end;
@<Print |c| and update |state| and |print_col|@>;
end;

@ @<Print |c| and update |state| and |print_col|@>=
if state=normal then
  if printable then print(xchr[c])
  else begin print('"&char',c:1);
    print_col:=print_col+2;
    end
else begin if state=special then
    begin print('&'); incr(print_col);
    end;
  if printable then
    begin print('"',xchr[c]); incr(print_col);
    end
  else print('char',c:1);
  end;
print_col:=print_col+l;
if printable then state:=normal @+else state:=special

@ The |end_char_string| procedure gets the string ended properly and ensures
that there is room for |l| more characters on the output line.

@<Declare subroutines for printing strings@>=
procedure end_char_string(@!l:integer);
begin while state>special do
  begin print('"');
  incr(print_col);
  decr(state);
  end;
if print_col+l>line_length then
  begin print_ln(' '); print_col:=0;
  end;
state:=initial; {get ready to print the next string}
end;

@ Since |end_char_string| resets |state:=initial|, all we have to do is set
|state:=initial| once at the beginning.

@<Set init...@>=
state:=initial;

@ Characters and rules are positioned according to global variables |h| and~|v|
as will be explained later.  We also need scale factors that convert quantities
to the right units when they are printed in the \.{MPX} file.

Even though all variable names in the \MP\ output are made local via \.{save}
commands, it is still desirable to preceed them with underscores.  This makes
the output more likely to work when used in a macro definition, since the
generated variables names must not collide with formal parameters in such
cases.

@<Glob...@>=
@!h,@!v:integer; {the current position in \.{DVI} units}
@!conv:real; {converts \.{DVI} units to \MP\ points}
@!mag:real; {magnification factor times 1000}

@ @p @<Declare a procedure called |finish_last_char|@>@;
procedure do_set_char(@!f,@!c:integer);
begin if (c<font_bc[f])or(c>font_ec[f]) then
  abort('attempt to typeset invalid character ',c:1);
@.attempt to typeset...@>
if (h<>str_h2)or(v<>str_v)or(f<>str_f)or(dvi_scale<>str_scale) then
  begin if str_f>=0 then finish_last_char
  else if not fonts_used then
    @<Prepare to output the first character on a page@>;
  if not font_used[f] then
    @<Prepare to use font |f| for the first time on a page@>;
  print('_s('); print_col:=3;@/
  str_scale:=dvi_scale; str_f:=f; str_v:=v; str_h1:=h;
  end;
print_char(c);
str_h2:=h+@<Width of character |c| in font |f|@>;
end;

@ @<Glob...@>=
@!font_used:array[0..max_fonts] of boolean;
  {has this font been used on this page?}
@!fonts_used:boolean; {has any font been used on this page?}
@!rules_used:boolean; {has any rules been set on this page?}
@!str_h1,str_v:integer; {starting position for current output string}
@!str_h2:integer; {where the current output string ends}
@!str_f:integer; {internal font number for the current output string}
@!str_scale:real; {value of |dvi_scale| for the current output string}

@ The |font_used| array is not initialized until it is actually time to output
a character.

@<Prepare to output the first character on a page@>=
begin k:=0;
while (k<nf) do
  begin font_used[k]:=false; incr(k);
  end;
fonts_used:=true;
print_ln('string _n[];');
print_ln('vardef _s(expr _t,_f,_m,_x,_y)=');
print_ln('  addto _p also _t infont _f scaled _m shifted (_x,_y); enddef;');
end

@ @<Do any other initialization required for the new font |f|@>=
font_used[f]:=false;

@ @<Prepare to use font |f| for the first time on a page@>=
begin font_used[f]:=true;
print('_n',f:1,'=');
print_col:=6;
print_font(f);
end_char_string(1);
print_ln(';');
end

@ We maintain the invariant that |str_f=-1| when there is no output string
under construction.

@<Declare a procedure called |finish_last_char|@>=
procedure finish_last_char;
var @!m,@!x,@!y:real;
  {font scale factor and \MP\ coordinates of reference point}
begin if str_f>=0 then
  begin m:=str_scale*font_scaled_size[str_f]*mag/font_design_size[str_f];@/
  x:=conv*str_h1; y:=conv*(-str_v);
  if (abs(x)>=4096.0)or(abs(y)>=4096.0)or(m>=4096.0)or(m<0) then
    begin warn('text scaled ',m:1:1,@|
        ' at (',x:1:1,',',y:1:1,') is out of range');
    end_char_string(60);
    end
  else end_char_string(40);
  print_ln(',_n',str_f:1,',',m:1:5,',',x:1:4,',',y:1:4,');');
  str_f:=-1;
  end;
end;

@ Setting rules is fairly simple.

@p procedure do_set_rule(@!ht,@!wd:integer);
var @!xx1,@!yy1,@!xx2,@!yy2,@!ww:real;
  {\MP\ coordinates of lower-left and upper-right corners}
begin if wd=1 then @<Handle a special rule that determines the box size@>
else if (ht>0)or(wd>0) then
  begin if str_f>=0 then finish_last_char;
  if not rules_used then
    begin rules_used:=true;@/
    print_ln('interim linecap:=0;');@/
    print_ln('vardef _r(expr _a,_w) =');
    print_ln('  addto _p doublepath _a withpen pencircle scaled _w enddef;');
    end;
  @<Make |(xx1,yy1)| and |(xx2,yy2)| then ends of the desired penstroke
    and |ww| the desired stroke width@>;
  if (abs(xx1)>=4096.0)or(abs(yy1)>=4096.0)or@|
      (abs(xx2)>=4096.0)or(abs(yy2)>=4096.0)or(ww>=4096.0) then
    warn('hrule or vrule near (',xx1:1:1,',',yy1:1:1,') is out of range');
  print_ln('_r((',xx1:1:4,',',yy1:1:4,')..(',xx2:1:4,',',yy2:1:4,
      '), ',ww:1:4,');');
  end;
end;

@ @<Make |(xx1,yy1)| and |(xx2,yy2)| then ends of the desired penstroke...@>=
xx1:=conv*h;
yy1:=conv*(-v);
if wd>ht then
  begin xx2:=xx1+conv*wd;
  ww:=conv*ht;@/
  yy1:=yy1+0.5*ww;
  yy2:=yy1;
  end
else begin yy2:=yy1+conv*ht;
  ww:=conv*wd;@/
  xx1:=xx1+0.5*ww;
  xx2:=xx1;
  end

@ Rules of width one dvi unit are not typeset since \.{MPtoTeX} adds an
extraneous rule of this width in order to allow \.{DVItoMP} to deduce the
dimensions of the boxes it ships out.  The box width is the left edge of the
last such rule; the height and depth are at the top and bottom of the rule.
There should be only one special rule per picture but there could be more if
the user tries to typeset his own one-dvi-unit rules.  In this case the
dimension-determining rule is the last one in the picture.

@<Handle a special rule that determines the box size@>=
begin pic_wd:=h; pic_dp:=v; pic_ht:=ht-v;
end

@ @<Glob...@>=
pic_dp, pic_ht, pic_wd: integer; {picture dimensions from special rule}

@ The following  initialization and clean-up is required.  We do a little more
initialization than is absolutely necessary since some compilers might complain
if the variables are uninitialized when |do_set_char| tests them.

@p procedure start_picture;
begin fonts_used:=false;
rules_used:=false;
str_f:=-1;@/
str_v:=0; str_h2:=0; str_scale:=1.0; {values don't matter}
print_ln('begingroup save _p,_r,_s,_n; picture _p; _p=nullpicture;');
end;
@#
procedure stop_picture;
var @!w,@!h,@!dd:real; {width, height, negative depth in PostScript points}
begin if str_f>=0 then finish_last_char;
@<Print a \&{setbounds} command based on picture dimensions@>;
print_ln('_p endgroup');
end;

@ @<Print a \&{setbounds} command based on picture dimensions@>=
dd:=-pic_dp*conv;
w:=conv*pic_wd; h:=conv*pic_ht;@/
print_ln('setbounds _p to (0,',dd:1:4,')--(',w:1:4,',',dd:1:4,')--');
print_ln(' (',w:1:4,',',h:1:4,')--(0,',h:1:4,')--cycle;')

@* Translation to symbolic form.
The main work of \.{DVItoMP} is accomplished by the |do_dvi_commands|
procedure, which produces the output for an entire page, assuming that the
|bop| command for that page has already been processed. This procedure is
essentially an interpretive routine that reads and acts on the \.{DVI}
commands.  It is also capable of executing the typesetting commands for
a character in a virtual font.

@ The definition of \.{DVI} files refers to six registers,
$(h,v,w,x,y,z)$, which hold integer values in \.{DVI} units.
These units come directly from the input file except they need to be
rescaled when typesetting characters from a virtual font.
The stack of $(h,v,w,x,y,z)$ values is represented by six arrays
called |hstack|, \dots, |zstack|.

@<Glob...@>=
@!w,@!x,@!y,@!z:integer;
  {current state values (|h| and |v| have already been declared)}
@!hstack,@!vstack,@!wstack,@!xstack,@!ystack,@!zstack:
  array [0..stack_size] of integer; {pushed down values in \.{DVI} units}
@!stk_siz:integer; {the current stack size}
@!dvi_scale:real; {converts units of current input source to \.{DVI} units}

@ @<Do initialization required before starting a new page@>=
dvi_scale:=1.0;
stk_siz:=0;
h:=0; v:=0

@ Next, we need procedures to handle |push| and |pop| commands.

@p procedure do_push;
begin if stk_siz=stack_size then
      abort('DVItoMP capacity exceeded (stack size=',stack_size:1,')');
@.DVItoMP capacity exceeded...@>
hstack[stk_siz]:=h; vstack[stk_siz]:=v; wstack[stk_siz]:=w;
xstack[stk_siz]:=x; ystack[stk_siz]:=y; zstack[stk_siz]:=z;
incr(stk_siz);
end;
@#
procedure do_pop;
begin if stk_siz=0 then bad_dvi('attempt to pop empty stack')
else  begin decr(stk_siz);
  h:=hstack[stk_siz]; v:=vstack[stk_siz]; w:=wstack[stk_siz];
  x:=xstack[stk_siz]; y:=ystack[stk_siz]; z:=zstack[stk_siz];
  end;
end;

@ We need to define the |set_virtual_char| procedure now because it is
mutually recursive with |do_dvi_commands|.  This is really a supervisory
@^recursion@>
procedure that calls |do_set_char| or adjusts the input source to read
typesetting commands for a character in a virtual font.

@p procedure do_dvi_commands;forward;@t\2@>
procedure set_virtual_char(@!f,@!c:integer);
var @!old_scale:real; {original value of |dvi_scale|}
@!old_buf_ptr:0..virtual_space; {original value of the input pointer |buf_ptr|}
@!old_fbase,@!old_ftop:0..max_fnums;
  {originally applicable part of the |font_num| table}
begin if fbase[f]=0 then do_set_char(f,c)
else begin old_fbase:=cur_fbase; old_ftop:=cur_ftop;
  cur_fbase:=fbase[f]; cur_ftop:=ftop[f];@/
  old_scale:=dvi_scale;
  dvi_scale:=dvi_scale*font_scaled_size[f];
  old_buf_ptr:=buf_ptr;
  buf_ptr:=start_cmd(f)(c);@/
  do_push;
  do_dvi_commands;
  do_pop;@/
  buf_ptr:=old_buf_ptr;
  dvi_scale:=old_scale;
  cur_fbase:=old_fbase; cur_ftop:=old_ftop;
  end;
end;

@ Before we get into the details of |do_dvi_commands|, it is convenient to
consider a simpler routine that computes the first parameter of each
opcode.

@d four_cases(#)==#,#+1,#+2,#+3
@d eight_cases(#)==four_cases(#),four_cases(#+4)
@d sixteen_cases(#)==eight_cases(#),eight_cases(#+8)
@d thirty_two_cases(#)==sixteen_cases(#),sixteen_cases(#+16)
@d sixty_four_cases(#)==thirty_two_cases(#),thirty_two_cases(#+32)

@<Declare a function called |first_par|@>=
function first_par(o:eight_bits):integer;
begin case o of
sixty_four_cases(set_char_0),sixty_four_cases(set_char_0+64):
  first_par:=o-set_char_0;
set1,put1,fnt1,xxx1,fnt_def1: first_par:=get_byte;
set1+1,put1+1,fnt1+1,xxx1+1,fnt_def1+1: first_par:=get_two_bytes;
set1+2,put1+2,fnt1+2,xxx1+2,fnt_def1+2: first_par:=get_three_bytes;
right1,w1,x1,down1,y1,z1: first_par:=signed_byte;
right1+1,w1+1,x1+1,down1+1,y1+1,z1+1: first_par:=signed_pair;
right1+2,w1+2,x1+2,down1+2,y1+2,z1+2: first_par:=signed_trio;
set1+3,set_rule,put1+3,put_rule,right1+3,w1+3,x1+3,down1+3,y1+3,z1+3,
  fnt1+3,xxx1+3,fnt_def1+3: first_par:=signed_quad;
nop,bop,eop,push,pop,pre,post,post_post,undefined_commands: first_par:=0;
w0: first_par:=w;
x0: first_par:=x;
y0: first_par:=y;
z0: first_par:=z;
sixty_four_cases(fnt_num_0): first_par:=o-fnt_num_0;
end;
end;

@ Here is the |do_dvi_commands| procedure.

@p procedure do_dvi_commands;
label 9999;
var o:eight_bits; {operation code of the current command}
@!p,@!q:integer; {parameters of the current command}
@!cur_font:integer; {current internal font number}
begin if (cur_fbase<cur_ftop) and (buf_ptr<virtual_space) then
  cur_font:=select_font(font_num[cur_ftop-1]) {select first local font}
else cur_font:=max_fnums+1; {current font is undefined}
w:=0; x:=0; y:=0; z:=0; {initialize the state variables}
while true do @<Translate the next command in the \.{DVI} file;
    |goto 9999| if it was |eop|@>;
9999: do_nothing;
end;

@ The multiway switch in |first_par|, above, was organized by the length
of each command; the one in |do_dvi_commands| is organized by the semantics.

@ @<Translate the next command...@>=
begin o:=get_byte; p:=first_par(o);
if eof(dvi_file) then bad_dvi('the DVI file ended prematurely');
@.the DVI file ended prematurely@>
if o<set1+4 then  {|set_char_0| through |set_char_127|, |set1| through |set4|}
  begin if cur_font>max_fnums then
    if vf_reading then
      abort('no font selected for character ',p:1,' in virtual font')
    else bad_dvi('no font selected for character ',p:1);
@.no font selected@>
  set_virtual_char(cur_font,p);
  h:=h+@<Width of character |p| in font |cur_font|@>;
  end
else case o of
  four_cases(put1): set_virtual_char(cur_font,p);
  set_rule: begin q:=trunc(signed_quad*dvi_scale);
    do_set_rule(trunc(p*dvi_scale),q);
    h:=h+q;
    end;
  put_rule: do_set_rule(trunc(p*dvi_scale),trunc(signed_quad*dvi_scale));
  @t\4@>@<Additional cases for translating \.{DVI} command |o| with
      first paramter |p|@>@;
  undefined_commands:bad_dvi('undefined command ',o:1);
@.undefined command@>
  end; {all cases have been enumerated}
end

@ @<Additional cases for translating \.{DVI} command |o|...@>=
four_cases(xxx1): for k:=1 to p do
    down_the_drain:=get_byte;
pre,post,post_post: bad_dvi('preamble or postamble within a page!');
@.preamble or postamble within a page@>

@ @<Additional cases for translating \.{DVI} command |o|...@>=
nop: do_nothing;
bop: bad_dvi('bop occurred before eop');
@.bop occurred before eop@>
eop: goto 9999;
push: do_push;
pop: do_pop;

@ @<Additional cases for translating \.{DVI} command |o|...@>=
four_cases(right1):h:=h+trunc(p*dvi_scale);
w0,four_cases(w1):begin w:=trunc(p*dvi_scale); h:=h+w;
  end;
x0,four_cases(x1):begin x:=trunc(p*dvi_scale); h:=h+x;
  end;
four_cases(down1):v:=v+trunc(p*dvi_scale);
y0,four_cases(y1):begin y:=trunc(p*dvi_scale); v:=v+y;
  end;
z0,four_cases(z1):begin z:=trunc(p*dvi_scale); v:=v+z;
  end;

@ @<Additional cases for translating \.{DVI} command |o|...@>=
sixty_four_cases(fnt_num_0),four_cases(fnt1):
  cur_font:=select_font(p);
four_cases(fnt_def1): define_font(p);

@* The main program.
Now we are ready to put it all together. This is where \.{DVItoMP} starts,
and where it ends.

@p begin initialize; {get all variables initialized}
@<Process the preamble@>;
open_mpx_file;
print_ln(banner);
begin while true do
  begin @<Advance to the next |bop| command@>;
  for k:=0 to 10 do down_the_drain:=signed_quad;
  @<Do initialization required before starting a new page@>;
  start_picture;
  do_dvi_commands;
  if stk_siz<>0 then bad_dvi('stack not empty at end of page');
@.stack not empty...@>
  stop_picture;
  print_ln('mpxbreak');
  end;
done:end;
final_end:end.

@ The main program needs a few global variables in order to do its work.

@<Glob...@>=
@!k,@!p:integer; {general purpose registers}
@!numerator,@!denominator:integer; {stated conversion ratio}

@ @<Process the preamble@>=
open_dvi_file;
p:=get_byte; {fetch the first byte}
if p<>pre then bad_dvi('First byte isn''t start of preamble!');
@.First byte isn't...@>
p:=get_byte; {fetch the identification byte}
if p<>id_byte then
  warn('identification in byte 1 should be ',id_byte:1,'!');
@.identification...should be n@>
@<Compute the conversion factor@>;
p:=get_byte; {fetch the length of the introductory comment}
while p>0 do
  begin decr(p); down_the_drain:=get_byte;
  end

@ The conversion factor |conv| is figured as follows: There are exactly
|n/d| decimicrons per \.{DVI} unit, and 254000 decimicrons per inch,
and |resolution| pixels per inch. Then we have to adjust this
by the stated amount of magnification.  No such adjustment is needed for
|dvi_per_fix| since it is used to convert design sizes.

@<Compute the conversion factor@>=
numerator:=signed_quad; denominator:=signed_quad;
if (numerator<=0)or(denominator<=0) then
  bad_dvi('bad scale ratio in preamble');
@.bad scale ratio@>
mag:=signed_quad/1000.0;
if mag<=0.0 then bad_dvi('magnification isn''t positive');
@.magnification isn't positive@>
conv:=(numerator/254000.0)*(72.0/denominator)*mag;
dvi_per_fix:=(254000.0/numerator)*(denominator/72.27)/1048576.0;

@ @<Advance to the next |bop| command@>=
repeat k:=get_byte;
if (k>=fnt_def1)and(k<fnt_def1+4) then
  begin p:=first_par(k); define_font(p); k:=nop;
  end;
until k<>nop;
if k=post then goto done;
if k<>bop then bad_dvi('missing bop');
@.missing bop@>


@* System-dependent changes.
This section should be replaced, if necessary, by changes to the program
that are necessary to make \.{DVItoMP} work at a particular installation.
It is usually best to design your change file so that all changes to
previous sections preserve the section numbering; then everybody's version
will be consistent with the printed program. More extensive changes,
which introduce new sections, can be inserted here; then only the index
itself will get a new section number.
@^system dependencies@>

@* Index.
Pointers to error messages appear here together with the section numbers
where each ident\-i\-fier is used.