#* 
   translate.perl.pss 

   This is a parse-script which translates parse-scripts into perl
   code, using the 'pep' tool. The script creates a standalone 
   perl program
   
   The virtual machine and engine is implemented in plain c at
   http://bumble.sf.net/books/pars/pep.c. This implements a script
   language with a syntax reminiscent of sed and awk (much simpler than
   awk, but more complex than sed).
   
   This version tries to simplify the grammar and uses a "parse()" method

STATUS

  Aug 2022: very simple scripts working.

NOTES
   
   This script could form the basis of a interp.perl.pss script
   which would translate a script into perl and execute it immediately
   using an "exec" or "eval" command. This requires that the script is
   read from a file (with an -f switch for example) because stdin is 
   required for the input stream.

   The script is very close to being an interpreter, but it is hard 
   to think about. It only works if we run the script on itself
   and then use as an interpreter.

   I will "reform" this grammer, which was originally written for 
   calculating absolute jumps in the "assembler" output code in
   compile.pss. This is not necessary here, and complicates matters.
   * So: some BNF rules
   --------
     !*test* -> nottest*
     !*class* -> nottest*
     !*quote* -> nottest*
     E*quote* -> test*
     B*quote* -> test*
     "(eof)" -> test*
     "(==)" -> test*

     test*,*test* -> orlist*
     orlist*,*test* -> orlist*
     test*.*test* -> andlist*
     andlist*.*test* -> andlist*
     (*andlist*)* -> exp*
     (*orlist*)* -> exp*

     exp*,exp* -> orlist*
     exp*,test* -> orlist*
     exp*,nottest* -> orlist*
     exp*,class* -> orlist*
     exp*,quote* -> orlist*

     exp*.exp* -> andlist*
     exp*.test* -> andlist*
     exp*.nottest* -> andlist*
     exp*.class* -> andlist*
     exp*.quote* -> andlist*
   
   ,,,,

   some strange bug with quote parsing in the compiler compile.pss
   seen in the parse(""); 

   man perlfunc is a handy reference for perl functions
   perl has a "goto"

   In other translation scripts, we use labelled loops and
   break/continue to implement the parse> label and .reparse .restart
   commands. Breaks could be used to implement the quit command but aren't.

   Does perl support labelled loops? No, one option is to set a flag
   when .restart .reparse is called to break the outer loop

   We can use "run once" loops eg " while true do ... break; end " an 
   example is in the translate.tcl.pss script. 

TODO

   convert the generated code to use a "parse" method with 
   some kind of a stream reader. 

SEE ALSO
   
   At http://bumble.sf.net/books/pars/

   tr/translate.java.pss, tr/translate.py.pss
     similar scripts for compiling scripts into java and python
     (also go, ruby, c, tcl, javascript).

   compile.pss
     compiles a script into an "assembly" format that can be loaded
     and run on the parse-machine with the -a  switch. This performs
     the same function as "asm.pp" 

PERL SYNTAX


 * open and read from file
 -----
  open F, "<file.txt" or die $!;
  while (<F>) { print $_; }
  close F;
 ,,,

 * read from a string as if it were a file
 -----
  my $foo = "abc\ndef\n";
  open my $fh, "<", \$foo;
  while (<$fh>) {
    print "line $.: $_";
  }
 ,,,

TEST SCRIPTS

  * test escaping
  >> pep.rb 'r;escape "c";t;d;' "abcxcabc"
  >> pep.rb 'r;unescape "c";t;d;' 'ab\cx\\cabc'

  so unescaping should be more intelligent.
  
TESTING

   * test simple scripts (using helper function), 1st and 2nd generation
   >> pep.tt perl

   * use helper to test 2nd gen script translation
   >> pep.plss "r;a':';t;d;" "abc"

   * use helper function in peprc to see palindromes
   >> pep.plff eg/palindrome.pss /usr/share/dict/words

   * use the bash helper functions to test (from helpers.pars.sh)
   >> pep.plf eg/json.check.pss '{"here":2}'

   The line above compiles the script to perl in the folder
   pars/eg/perl/json.check.pss and runs it with the input.

   * one comprehensive test is to run the script on itself
   >> pep -f translate.rb.pss translate.rb.pss > eg/perl/translate.rb.rb
   >> chmod a+x eg/perl/translate.perl.pl
   >> echo "r;t;d;" | eg/perl/translate.perl.pl
   
   I call this "2nd generation" script compilation.

GOTCHAS

  treatment of regexes is different (for while whilenot etc). Eg in
  ruby [[:space:]] is unicode aware but \s is not

  until needs to actually count how many trailing '\' or
  escape chars there are !...

  Make sure escaping and multiline arguments work.

  "until" may not read at least one character.

BUGS
   
  many.

  2nd gen is not working  yet

SOLVED BUGS TO WATCH FOR 

  * the line below was throwing an error (compile.pss)
  >> add '", "\\'; get; add '")'; --; put; clear;

  This needed to be fixed in compile.pss and also in eg translation
  script.

  Multiline add should not add extra spaces,

  escape needs to use the machine escape char. 
  found and fixed a bug in java whilenot/while. The code exits if the 
  character is not found, which is not correct.

  Found and fixed a bug in the (==) code ie in java (stringa == stringb)
  doesnt work. 

  "until" bug where the code did not read 
  at least one character.

  Read must exit if at end of stream, but while/whilenot/until, no.

TASKS 

HISTORY

  22 aug 2022
    Wrote escapeChar and unescapeChar properly. These are "expensive"
    operations when the workspace is big. Wrote parse methods.
    Can read and write to strings now.
    Will try to change the grammar to be more logical. Eg, will 
    use tokens test* nottest* exp* quote*. So begintext* and endtext*
    should get ellided. Changed to parse method. Also
    will write "struct.syntax.txt" which will have all syntax structures.
    This will be useful for seeing how translators work.
    Saved an original version in tr/translate.perl.orig.pss
  20 aug 2022
    Debugging and testing. A few tests in tr/tr.test.txt now work 
  13 july 2022
    Converting from a normal perl hash, to an object oriented 
    perl idiom. (bless etc). Perl oo seems quite odd. Simplest script
    'r;t;d' is now working.
  25 june 2022
    A lot of conversion work.
  24 June 2022
    Began to adapt from the ruby code.

*#

  read;
  #--------------
  [:space:] {
    clear; .reparse
  }

  #---------------
  # We can ellide all these single character tests, because
  # the stack token is just the character itself with a *
  # Braces {} are used for blocks of commands, ',' and '.' for concatenating
  # tests with OR or AND logic. 'B' and 'E' for begin and end
  # tests, '!' is used for negation, ';' is used to terminate a 
  # command.
  "{", "}", ";", ",", ".", "!", "B", "E" {
    put; add "*"; push; .reparse 
  }

  #---------------
  # format: "text"
  "\"" {
    # save the start line number (for error messages) in case 
    # there is no terminating quote character.
    clear; add "line "; lines; add " (character "; chars; add ") ";
    put; clear; add '"';
    until '"'; 
    !E'"' { 
      clear; add 'Unterminated quote character (") starting at ';
      get; add ' !\n'; 
      print; quit;
    }
    # check for empty quotes as arguments for escape etc 
    replace "#{" "\\#{"; 
    put; clear;
    add "quote*"; push;
    .reparse 
  }

 #---------------
 # format: 'text', single quotes are converted to double quotes
 # but we must escape embedded double quotes.
  "'" {
    # save the start line number (for error messages) in case 
    # there is no terminating quote character.
    clear; add "line "; lines; add " (character "; chars; add ") ";
    put; clear;
    until "'"; 
    !E"'" { 
      clear; add "Unterminated quote (') starting at ";
      get; add '!\n'; 
      print; quit;
    }
    clip; escape '"'; 
    # #{ does string interpolation in ruby which is not 
    # what we want. Also unescape \'
    replace "#{" "\\#{"; 
    unescape "'";
    put; clear;
    add "\""; get; add "\"";
    put; clear; add "quote*";
    push; .reparse 
  }

  #---------------
  # formats: [:space:] [a-z] [abcd] [:alpha:] etc 
  # should class tests really be multiline??!
  "[" {
    # save the start line number (for error messages) in case 
    # there is no terminating bracket character.
    clear; add "line "; lines; add " (character "; chars; add ") ";
    put; clear; add "[";
    until "]"; 
    "[]" {
      clear; add "pep script error at line "; lines;
      add " (character "; chars; add "): \n";
      add "  empty character class [] \n";
      print; quit;
    }
    !E"]" { 
      clear; add "Unterminated class text ([...]) starting at "; get; 
      add "
      class text can be used in tests or with the 'while' and 
      'whilenot' commands. For example: 
        [:alpha:] { while [:alpha:]; print; clear; }
      ";
      print; quit;
    }

    # need to escape quotes? ruby uses /.../ to match
    escape '"';
    # the caret is not a negation operator in pep scripts
    replace "^" "\\^";

    # save the class on the tape
    put;
    clop; clop;
    !B"-" {
      # not a range class, eg [a-z] so need to escape '-' chars
      clear; get; replace '-' '\\-'; put;
    }
    B"-" {
      # a range class, eg [a-z], check if it is correct
      clip; clip; 
      !"-" {
        clear;
        add "Error in pep script at line "; lines;
        add " (character "; chars; add "): \n";
        add " Incorrect character range class "; get;
        add "
   For example:
     [a-g]  # correct
     [f-gh] # error! \n";
        print; clear; quit;

      }
    }
    clear; get;  # restore class text
    B"[:".!E":]" { 
      clear; add "malformed character class starting at ";
      get; add '!\n'; 
      print; quit;
    }
    # class in the form [:digit:]
    B"[:".!"[:]" {
      clip; clip; clop; clop;
      # unicode posix character classes 
      # Also, abbreviations (not implemented in gh.c yet.)
      # this should not be tricky in ruby because posix is supported
      "alnum","N" { clear; add "[[:alnum:]]"; }
      "alpha","A" { clear; add "[[:alpha:]]"; }
      # ? can use s.ascii_only?()
      "ascii","I" { clear; add "[[:ascii:]]"; }
      # non-standard ruby posix class 'word'
      "word","W" { clear; add "[[:word:]]"; }
      "blank","B" { clear; add "[[:blank:]]"; }
      "cntrl","C" { clear; add '[[:cntrl:]]'; }
      "digit","D" { clear; add "[[:digit:]]"; }
      "graph","G" { clear; add '[[:graph:]]'; }
      "lower","L" { clear; add '[[:lower:]]'; }
      "print","P" { clear; add "[[:print:]]"; }
      "punct","T" { clear; add '[[:punct:]]'; }
      "space","S" { clear; add "[[:space:]]"; }
      "upper","U" { clear; add '[[:upper:]]'; }
      "xdigit","X" { clear; add "[[:xdigit:]]"; }

      !B"[[" {
        put; clear;
        add "pep script error at line "; lines;
        add " (character "; chars; add "): \n";
        add "Unknown character class '"; get; add "'\n";
        print; clear; quit;
      }

    }

    put; 
    clear;
    add '/^'; get; add '+$/'; put;
    clear;
    # add quotes around the class and limits around the 
    # class so it can be used with the string.matches() method
    # (must match the whole string, not just one character)
    add "class*"; push;
    .reparse 
  }

 #---------------
 # to implement and expression grammar need to free up ( and )
 # formats: (eof) (EOF) (==) etc. 
  "(" {
    clear; until ")"; clip;
    put; 
    "eof","EOF" { clear; add "eof*"; push; .reparse } 
    "==" { clear; add "tapetest*"; push; .reparse } 
    add " << unknown test near line "; lines;
    add " of script.\n";
    add " bracket () tests are \n";
    add "   (eof) test if end of stream reached. \n";
    add "   (==)  test if workspace is same as current tape cell \n";
    print; clear;
    quit;
  }

  #---------------
  # multiline and single line comments, eg #... and #* ... *#
  "#" {
    clear; read;
    "\n" { clear; .reparse }

    # checking for multiline comments of the form "#* \n\n\n *#"
    # these are just ignored at the moment (deleted) 
    "*" { 
      # save the line number for possible error message later
      clear; lines; put; clear;
      until "*#"; 
      E"*#" {
        # convert to python comments (#), python doesnt have multiline
        # comments, as far as I know
        clip; clip; replace "\n" "\n#"; 
        put; clear; 
        # create a "comment" parse token
        # comment-out this line to remove multiline comments from the 
        # compiled python
        # add "comment*"; push; 
        .reparse  
      }
      # make an unterminated multiline comment an error
      # to ease debugging of scripts.
      clear; 
      add "unterminated multiline comment #* ... *# \n";
      add "stating at line number "; get; add "\n";
      print; clear;
      quit;
    }

    # single line comments. some will get lost.
    put; clear; add "#"; get; until "\n"; clip;
    put; clear; 
    # comment out this below to remove single line comments
    # from the output
    add "comment*"; push; 
    .reparse 
  }

 #----------------------------------
 # parse command words (and abbreviations)

 # legal characters for keywords (commands)
 ![abcdefghijklmnopqrstuvwxyzBEKGPRUWS+-<>0^] {
   # error message about a misplaced character
   put; clear;
   add "!! Misplaced character '";
   get;
   add "' in script near line "; lines;
   add " (character "; chars; add ") \n";
   print; clear; quit;
 }

   # my testclass implementation cannot handle complex lists
   # eg [a-z+-] this is why I have to write out the whole alphabet

   while [abcdefghijklmnopqrstuvwxyzBEOFKGPRUWS+-<>0^];
   #----------------------------------
   # KEYWORDS 
   # here we can test for all the keywords (command words) and their
   # abbreviated one letter versions (eg: clip k, clop K etc). Then
   # we can print an error message and abort if the word is not a 
   # legal keyword for the parse-edit language

   # make ll an alias for "lines" and cc an alias for chars
   "ll" { clear; add "lines"; }
   "cc" { clear; add "chars"; }
   # one letter command abbreviations
   "a" { clear; add "add"; }
   "k" { clear; add "clip"; }
   "K" { clear; add "clop"; }
   "D" { clear; add "replace"; }
   "d" { clear; add "clear"; }
   "t" { clear; add "print"; }
   "p" { clear; add "pop"; }
   "P" { clear; add "push"; }
   "u" { clear; add "unstack"; }
   "U" { clear; add "stack"; }
   "G" { clear; add "put"; }
   "g" { clear; add "get"; }
   "x" { clear; add "swap"; }
   ">" { clear; add "++"; }
   "<" { clear; add "--"; }
   "m" { clear; add "mark"; }
   "M" { clear; add "go"; }
   "r" { clear; add "read"; }
   "R" { clear; add "until"; }
   "w" { clear; add "while"; }
   "W" { clear; add "whilenot"; }
   "n" { clear; add "count"; }
   "+" { clear; add "a+"; }
   "-" { clear; add "a-"; }
   "0" { clear; add "zero"; }
   "c" { clear; add "chars"; }
   "l" { clear; add "lines"; }
   "^" { clear; add "escape"; }
   "v" { clear; add "unescape"; }
   "z" { clear; add "delim"; }
   "S" { clear; add "state"; }
   "q" { clear; add "quit"; }
   "s" { clear; add "write"; }
   "o" { clear; add "nop"; }
   "rs" { clear; add "restart"; }
   "rp" { clear; add "reparse"; }

   # some extra syntax for testeof and testtape
   "<eof>","<EOF>" { put; clear; add "eof*"; push; .reparse }
   "<==>" { put; clear; add "tapetest*"; push; .reparse }

   "jump","jumptrue","jumpfalse",
   "testis","testclass","testbegins","testends",
   "testeof","testtape" {
     put; clear;
     add "The instruction '"; get; add "' near line "; lines; 
     add " (character "; chars; add ")\n";
     add "can be used in pep assembly code but not scripts. \n";
     print; clear; quit;
   }
   
   # show information if these "deprecated" commands are used
   "Q","bail" {
     put; clear;
     add "The instruction '"; get; add "' near line "; lines; 
     add " (character "; chars; add ")\n";
     add "is no longer part of the pep language (july 2020). \n";
     add "use 'quit' instead of 'bail', and use 'unstack; print;' \n";
     add "instead of 'state'. \n";
     print; clear; quit;
   }
   
   "add","clip","clop","replace","upper","lower","cap","clear","print","state",
   "pop","push","unstack","stack","put","get","swap",
   "++","--","mark","go","read","until","while","whilenot",
   "count","a+","a-","zero","chars","lines","nochars","nolines",
   "escape","unescape","delim","quit",
   "write","nop","reparse","restart" {
     put; clear;
     add "word*";
     push; .reparse
   }
   
   #------------ 
   # the .reparse command and "parse label" is a simple way to 
   # make sure that all shift-reductions occur. It should be used inside
   # a block test, so as not to create an infinite loop. There is
   # no "goto" in java so we need to use labelled loops to 
   # implement .reparse/parse>

   "parse>" {
     clear; count;
     !"0" {
       clear; 
       add "script error:\n";
       add "  extra parse> label at line "; lines; add ".\n";
       print;
       quit;
     }
     clear; add "# parse> parse label"; put;
     clear; add "parse>*"; push;
     # use accumulator to indicate after parse> label
     a+; .reparse 
   }

   # --------------------
   # implement "begin-blocks", which are only executed
   # once, at the beginning of the script (similar to awk's BEGIN {} rules)
   "begin" {
     put; add "*"; push; .reparse 
   }

   add " << unknown command on line "; lines; 
   add " (char "; chars; add ")"; 
   add " of source file. \n"; 
   print; clear; quit;

# ----------------------------------
# PARSING PHASE:

# Below is the parse/compile phase of the script. Here we pop tokens off the
# stack and check for sequences of tokens eg "word*semicolon*". If we find a
# valid series of tokens, we "shift-reduce" or "resolve" the token series eg
# word*semicolon* --> command*
#
# At the same time, we manipulate (transform) the attributes on the tape, as
# required. 
#

parse>

#-------------------------------------
# 2 tokens
#-------------------------------------
  pop; pop;

  # All of the patterns below are currently errors, but may not
  # be in the future if we expand the syntax of the parse
  # language. Also consider:
  #    begintext* endtext* quoteset* notclass*, !* ,* ;* B* E*
  # It is nice to trap the errors here because we can emit some
  # (hopefully not very cryptic) error messages with a line number.
  # Otherwise the script writer has to debug with
  #   pep -a asm.pp -I scriptfile 
  #

  "word*word*","word*}*","word*begintext*","word*endtext*", "word*!*",
  "word*,*","quote*word*", "quote*class*", "quote*state*", "quote*}*",
  "quote*begintext*", "quote*endtext*", "class*word*", "class*quote*",
  "class*class*", "class*state*", "class*}*", "class*begintext*",
  "class*endtext*", "class*!*", "notclass*word*", "notclass*quote*",
  "notclass*class*", "notclass*state*", "notclass*}*" {
    add " (Token stack) \nValue: \n"; get; 
    add "\nValue: \n"; ++; get; --; add "\n";
    add "Error near line "; lines; add " (char "; chars; add ")"; 
    add " of pep script (missing semicolon?) \n";
    print; clear; 
    quit;
  }  

  "{*;*", ";*;*", "}*;*" {
    push; push;
    add "Error near line "; lines; add " (char "; chars; add ")"; 
    add " of pep script: misplaced semi-colon? ; \n";
    print; clear; quit;
  }

  ",*{*" {
    push; push;
    add "Error near line "; lines; add " (char "; chars; add ")"; 
    add " of script: extra comma in list? \n";
    print; clear; quit;
  }

  "command*;*","commandset*;*" {
    push; push;
    add "Error near line "; lines; add " (char "; chars; add ")"; 
    add " of script: extra semi-colon? \n";
    print; clear; quit;
  }

  "!*!*" {
    push; push;
    add "error near line "; lines; add " (char "; chars; add ")"; 
    add " of script: \n double negation '!!' is not implemented \n";
    add " and probably won't be, because what would be the point? \n";
    print; clear; quit;
  }

  "!*{*","!*;*" {
    push; push;
    add "error near line "; lines;
    add " (char "; chars; add ")"; 
    add " of script: misplaced negation operator (!)? \n";
    add " The negation operator precedes tests, for example: \n";
    add "   !B'abc'{ ... } or !(eof),!'abc'{ ... } \n";
    print; clear; quit;
  }

  ",*command*" {
    push; push;
    add "error near line "; lines;
    add " (char "; chars; add ")"; 
    add " of script: misplaced comma? \n";
    print; clear; quit;
  }

  "!*command*" {
    push; push;
    add "error near line "; lines;
    add " (at char "; chars; add ") \n"; 
    add " The negation operator (!) cannot precede a command \n";
    print; clear; quit;
  }

  ";*{*", "command*{*", "commandset*{*" {
    push; push;
    add "error near line "; lines;
    add " (char "; chars; add ")"; 
    add " of script: no test for brace block? \n";
    print; clear; quit;
  }

  "{*}*" {
    push; push;
    add "error near line "; lines;
    add " of script: empty braces {}. \n";
    print; clear; quit;
  }

  "B*class*","E*class*" {
    push; push;
    add "error near line "; lines;
    add " of script:\n  classes ([a-z], [:space:] etc). \n";
    add "  cannot use the 'begin' or 'end' modifiers (B/E) \n";
    print; clear; quit;
  }

  "comment*{*" {
    push; push;
    add "error near line "; lines;
    add " of script: comments cannot occur between \n";
    add " a test and a brace ({). \n";
    print; clear; quit;
  }

  "}*command*" {
    push; push;
    add "error near line "; lines;
    add " of script: extra closing brace '}' ?. \n";
    print; clear; quit;
  }

  #*
  E"begin*".!"begin*" {
    push; push;
    add "error near line "; lines;
    add " of script: Begin blocks must precede code \n";
    print; clear; quit;
  }
  *#

  #------------ 
  # The .restart command jumps to the first instruction after the
  # begin block (if there is a begin block), or the first instruction
  # of the script.
  ".*word*" {
    clear; ++; get; --;
    "restart" {
      clear; count;
      # this is the opposite of .reparse, using run-once loops
      # cant do next before label, infinite loop.
      # need to set flag variable. There is a subtlety: .restart can
      # exist without parse> (although it would be unusual). 

      # before (or without) the parse> label
      "0" { 
        clear; 
        # use the comment '# restart' so we can replace
        # this with 'break' if the parse> label appears later
        add "restart = $true; next; # restart"; 
      }   
      "1" { clear; add "break;"; }   # after the parse> label
      put; clear;
      add "command*";
      push; .reparse 
    }
    "reparse" {
      clear; count; 
      # check accumulator to see if we are in the "lex" block
      # or the "parse" block and adjust the .reparse compilation
      # accordingly.
      "0" { clear; add "break;"; }
      "1" { clear; add "next;"; }
      put; clear;
      add "command*";
      push; .reparse 
    }
    push; push;
    add "error near line "; lines;
    add " (char "; chars; add ")"; add " of script:  \n";
    add " misplaced dot '.' (use for AND logic or in .reparse/.restart \n";
    print; clear; quit;
  }

  #---------------------------------
  # Compiling comments so as to transfer them to the java 
  "comment*command*","command*comment*","commandset*comment*" {
    clear; get; add "\n"; ++; get; --; put; clear;
    add "command*"; push; .reparse
  }

  "comment*comment*" {
    clear; get; add "\n"; ++; get; --; put; clear;
    add "comment*"; push; .reparse
  }

  # -----------------------
  # negated tokens.
  #
  # This is a new more elegant way to negate a whole set of 
  # tests (tokens) where the negation logic is stored on the 
  # stack, not in the current tape cell. We just add "not" to 
  # the stack token.

  # eg: ![:alpha:] ![a-z] ![abcd] !"abc" !B"abc" !E"xyz"
  #  This format is used to indicate a negative test for 
  #  a brace block. eg: ![aeiou] { add "< not a vowel"; print; clear; }

  "!*quote*","!*class*","!*begintext*", "!*endtext*",
  "!*eof*","!*tapetest*" {
    # a simplification: store the token name "quote*/class*/..."
    # in the tape cell corresponding to the "!*" token. 
    replace "!*" "not"; push;
    # this was a bug?? a missing ++; ??
    # now get the token-value
    get; --; put; ++; clear;
    .reparse
  }

  #-----------------------------------------
  # format: E"text" or E'text'
  #  This format is used to indicate a "workspace-ends-with" text before
  #  a brace block.
  "E*quote*" {
     clear; add "endtext*"; push; get; 
     '""' {
       # empty argument is an error
       clear;
       add "pep script error near line "; lines;
       add " (character "; chars; add "): \n";
       add '  empty argument for end-test (E"") \n';
       print; quit;
     }
     --; put; ++;
     clear; .reparse
  } 

  #-----------------------------------------
  # format: B"sometext" or B'sometext' 
  #   A 'B' preceding some quoted text is used to indicate a 
  #   'workspace-begins-with' test, before a brace block.
  "B*quote*" {
     clear; add "begintext*"; push; get; 
     '""' {
       # empty argument is an error
       clear;
       add "pep script error near line "; lines;
       add " (character "; chars; add "): \n";
       add '  empty argument for begin-test (B"") \n';
       print; quit;
     }
     --; put; ++;
     clear; .reparse
  } 

  #--------------------------------------------
  # ebnf: command := word, ';' ;
  # formats: "pop; push; clear; print; " etc
  # all commands need to end with a semi-colon except for 
  # .reparse and .restart
  #
  "word*;*" {
     clear;
     # check if command requires parameter
     get;
     "add","while","whilenot","mark","go",
     "escape","unescape","delim","replace" {
       put; clear; add "'"; get; add "'";
       add " << command needs an argument, on line "; lines; 
       add " of script.\n";
       print; clear; quit;
     }

     # the new until tape command
     "until" { 
       clear; 
       add "$self->until(${$self->{'tape')}[$self->{'cell'}]);  # until (tape)  ";
       put; 
     }
     "clip" { 
       clear; 
       add "chop($self->{'work'});       # clip";
       put; 
     }
     # check indices
     "clop" { 
       clear; 
       add "$self->{'work'} =~ s/^.//s;       # clop \n";
       put; 
     }
     "clear" { clear; add "$self->{'work'} = '';        # clear"; put; }
     "upper" { clear; add "$self->{'work'} = uc($self->{'work'});  # upper"; put; }
     "lower" { clear; add "$self->{'work'} = lc($self->{'work'});  # lower"; put; }
     "cap" { clear;   add "$self->{'work'} .capitalize! # capital"; put; }
     "print" { 
       clear; 
       add '$self->{"output"}->print($self->{"work"});   # print'; 
       put;
     }
     "state" { clear; add '$self->printState();         # state'; put; }
     "pop" { clear; add "$self->popToken();"; put; }
     "push" { clear; add "$self->pushToken();"; put; }
     "unstack" { 
       clear; add "while ($self->popToken()) { next; }  # unstack "; put; }
     "stack" { 
       clear; add "while ($self->pushToken()) { next; } # stack "; put; }
     "put" { 
       clear; 
       add "${$self->{'tape'}}[$self->{'cell'}] = $self->{'work'};   # put "; 
       put; 
     }
     "get" { 
       clear; 
       add "$self->{'work'} .= ${$self->{'tape'}}[$self->{'cell'}];  # get";
       put;
     }
     "swap" { 
       clear; 
       add "$self->{'work'}, @{$self->{'tape'}}[$self->{'cell'}] = \n";
       add "  @{$self->{'tape'}}[$self->{'cell'}], $self->{'work'};    # swap ";
       put;
     }
     "++" { clear; add "$self->increment();      # ++"; put; }
     "--" {
       clear; add "if ($self->{'cell'} > 0) { $self->{'cell'} -= 1; } # --"; put;
     }
     "read" { clear;  add "$self->readChar();         # read"; put; }
     "count" { 
       clear; add "$self->{'work'} .= $self->{'counter'};     # count "; put; }
     "a+" { clear;    add "$self->{'counter'} += 1;  # a+ "; put; }
     "a-" { clear;    add "$self->{'counter'} -= 1;  # a- "; put; }
     "zero" { clear;  add "$self->{'counter'} = 0;   # zero "; put; }
     "chars" { clear; add "$self->{'work'} .= $self->{'charsRead'}; # chars "; put; }
     "lines" { clear; add "$self->{'work'} .= $self->{'linesRead'}; # lines "; put; }
     "nochars" { clear; add "$self->{'charsRead'} = 0;           # nochars "; put; }
     "nolines" { clear; add "$self->{'linesRead'} = 0;           # nolines "; put; }
     # use a labelled loop to quit script.
     "quit" { clear; add "exit();"; put; }
     # inline this?
     "write" { clear; add "File->write('sav.pp', $self->{'work'} )"; put; }
     # convert to "pass" which does nothing. 
     "nop" { clear; add "# nop: no-operation"; put; }

     clear; add "command*";
     push; .reparse
   }

  #-----------------------------------------
  # ebnf: commandset := command , command ;
  "command*command*", "commandset*command*" {
    clear;
    add "commandset*"; push;
    # format the tape attributes. Add the next command on a newline 
    --; get; add "\n"; ++; get; --; put; ++; 
    clear; .reparse
  } 

  #-------------------
  # here we begin to parse "test*" and "ortestset*" and "andtestset*"
  # 

  # This will become "test*","class*","quote*","nottest*"
  #-------------------
  # eg: B"abc" {} or E"xyz" {}
  # transform and markup the different test types
  "begintext*,*","endtext*,*","quote*,*","class*,*",
  "eof*,*","tapetest*,*",
  "begintext*.*","endtext*.*","quote*.*","class*.*",
  "eof*.*","tapetest*.*",
  "begintext*{*","endtext*{*","quote*{*","class*{*",
  "eof*{*","tapetest*{*" {

    B"begin" { clear; add "$self->{'work'} =~ /^"; }
    B"end" { clear; add "$self->{'work'} .end_with?("; }
    B"quote" { clear; add "$self->{'work'} eq "; }
    B"class" { 
      clear; add "$self->{'work'} =~ /";
      # unicode categories are also regexs 
    } 
    # clear the tapecell for testeof and testtape because
    # they take no arguments. 
    B"eof" { clear; put; add "$self->{'eof'}"; }
    B"tapetest" { 
      clear; put; 
      add "$self->{'work'} eq ${$self->{'tape'}}[$self->{'cell'}]"; 
    }
    get; 
    # a hack
    #B"mm.work.match?" { add ')'; }
    !B"$self->{'eof'}".!B"$self->{'work'} eq" { add ")"; }
    put; 
    #*
    #  maybe we could ellide the not tests by doing here
    B"not" { clear; add "!"; get; put; }
    *#
    clear; add "test*"; push;
    # the trick below pushes the right token back on the stack.
    get; add "*"; push; .reparse
  }

  #-------------------
  # negated tests
  # eg: !B"xyz {} !(eof) {} !(==) {}
  #     !E"xyz" {} 
  #     !"abc" {}
  #     ![a-z] {}
  "notbegintext*,*","notendtext*,*","notquote*,*","notclass*,*",
  "noteof*,*","nottapetest*,*",
  "notbegintext*.*","notendtext*.*","notquote*.*","notclass*.*",
  "noteof*.*","nottapetest*.*",
  "notbegintext*{*","notendtext*{*","notquote*{*","notclass*{*",
  "noteof*{*","nottapetest*{*"
  {

    B"notbegin" { clear; add "$self->{'work'} !~ /^"; }
    B"notend" { clear; add "!$self->{'work'} =~ /"; }
    B"notquote" { clear; add "$self->{'work'} != "; }
    B"notclass" { 
      clear; add "!$self->{'work'} =~ /";
      # perl unicode categories are regexs ?
    }
    # clear the tapecell for testeof and testtape because
    # they take no arguments. 
    B"noteof" { clear; put; add "!$self->{'eof')"; }
    B"nottapetest" { 
      clear; put; 
      add "$self->{'work'} != @{$self->{'tape'}}[$self->{'cell'}]"; 
    }
    get; 
    !B"!$self->{'eof'}".!B"$self->{'work'}  !=" { add ")"; }
    put; clear; add "test*"; push; 
    # the trick below pushes the right token back on the stack.
    get; add "*"; push; .reparse
  }

  #-------------------
  # 3 tokens
  #-------------------

  pop;

  #-----------------------------
  # some 3 token errors!!!
 
  # not a comprehensive list of 3 token errors
  "{*quote*;*","{*begintext*;*","{*endtext*;*","{*class*;*",
  "commandset*quote*;*", "command*quote*;*" {
    push; push; push;
    add "[pep error]\n invalid syntax near line "; lines;
    add " (char "; chars; add ")"; 
    add " of script (misplaced semicolon?) \n";
    print; clear; quit;
  }  

  # to simplify subsequent tests, transmogrify a single command
  # to a commandset (multiple commands).
  "{*command*}*" {
    clear; add "{*commandset*}*"; push; push; push;
    .reparse
  }

  # errors! mixing AND and OR concatenation
  ",*andtestset*{*",
  ".*ortestset*{*" {
    # push the tokens back to make debugging easier
    push; push; push; 
    add " error: mixing AND (.) and OR (,) concatenation in \n";
    add " in pep script near line "; lines;
    add " (character "; chars; add ") \n";
    add ' 
  For example:
     B".".!E"/".[abcd./] { print; }  # Correct!
     B".".!E"/",[abcd./] { print; }  # Error! \n';
    print; clear; quit;
  }

  #--------------------------------------------
  # ebnf: command := keyword , quoted-text , ";" ;
  # format: add "text";

  "word*quote*;*" {
    clear; get;
    "replace" {
       # error 
       add "< command requires 2 parameters, not 1 \n";
       add "near line "; lines;
       add " of script. \n";
       print; clear; quit;
    }

    # check whether argument is single character, otherwise
    # throw an error. Also, check that argument is not empty
    # eg "". Its probably silly to allow while/while not to
    # have a quote argument, since the same can be achieved with
    # >> while [a]; etc
    "escape", "unescape", "while", "whilenot" {
      # This is trickier than I thought it would be.
      clear; ++; get; --; 

      # check that arg not empty, (but an empty quote is ok 
      # for the second arg of 'replace'
      '""' {
        clear; 
        add "[pep error] near line "; lines;
        add " (or char "; chars; add "): \n"; 
        add "  command '"; get; 
        add '\' cannot have an empty argument ("") \n';
        print; quit;
      }

      # quoted text has the quotes still around it.
      # also handle escape characters like \n \r etc
      # this needs to be better
      clip; clop; clop; clop;
      # B "\\" { clip; } 
      clip; 
      !"" {
        clear; 
        add "Pep script error near line "; lines;
        add " (character "; chars; add "): \n"; 
        add "  command '"; get; 
        add "' takes only a single character argument. \n";
        print; quit;
      }
      clear; get;
    }

    "mark" {
      clear;
      add "@{$self->{'marks'}}[$self->{'cell] = "; ++; get; --; add " # mark";
      put; clear; add "command*"; push; .reparse
    }

    "go" {
      clear;
      add "goToMark($mm, "; ++; get; --; add ")  # go";
      put; clear; add "command*"; push; .reparse
    }

    "delim" {
      clear;
      # the delimiter should be a single character, no?
      add "$self->{'delimiter'} = "; ++; get; --; add " # delim ";
      put; clear; add "command*"; push; .reparse
    }

    "add" {
      clear;
      add "$self->{'work'} .= "; ++; get; --; add ";";
      # handle multiline text
      # check this! \\n or \n
      replace "\n" '"\n$self->{"work"} .= "\\n';
      put; clear; add "command*"; push; .reparse
    }

    "while" {
      clear;
      add "while ($self->{'peep'} eq "; ++; get; --;
      add ");   # while \n"; 
      add "  if ($self->{'eof'}) { break; }\n  $self->readChar();\n"; 
      add "}"; 
      put; clear; add "command*"; push; .reparse
    }

    "whilenot" {
      clear;
      add "while $self->{'peep'} != "; ++; get; --;
      add ";   # whilenot \n"; 
      add "  if ($self->{'eof'}) { break; }\n  read(%mm);\n}"; 
      put; clear; add "command*"; push; .reparse
    }

    "until" {
       clear; add "$self->until("; 
       ++; get; --; 
       # error until cannot have empty argument
       '$self->until(""' { 
         clear; 
         add "Pep script error near line "; lines;
         add " (character "; chars; add "): \n";
         add " empty argument for 'until' \n";
         add " 
   For example:
     until '.txt'; until \">\";    # correct   
     until '';  until \"\";        # errors! \n";
         print; quit;
       }
       # handle multiline argument
       replace "\n" "\\n";
       add ');'; put; clear;
       add "command*"; push; .reparse
     }

    "escape" {
       clear; ++;
       # argument still has quotes around it
       # it should be a single character since this has been previously
       # checked.
       # Actually this should be handled by a function, since it is 
       # a bit complex. (need to check that the char is not already
       # escaped, which means counting the number of \\ or other escape
       # chars before it.
       get; clip; clop; escape "/"; 
       # should be 'unescape' but this isnt well implemented in many
       # translators.
       replace '\\"' '"'; put; clear;
       add '$self->{"work"} =~ s/\\Q'; get; add '/\\E'; 
       # will this work for mm.escape != '\\'
       # E'"\'", "' { add "$self->{'escape'}"; }
       add '$self->{"escape"}\\Q'; get; add '/g;   # escape'; --; put; clear;
       add "command*"; push; .reparse
     }

    # replace \n with n for example (only 1 character)
    "unescape" {
       clear; ++;
       # unescape is not trivial, need to walk the string
       # hence the method rather than one-liner
       #add 'mm.unescapeChar('; get; add ')  # unescape'; 
       add '$self->{"work"} =~ s/$self->{"escape"}+'; get; 
       add ', '; get; add ')  # escape'; --; put; clear;
       add "command*"; push; .reparse
     }

     # error, superfluous argument
     add ": command does not take an argument \n";
     add "near line "; lines;
     add " of script. \n";
     print; clear;
     #state
     quit;
   }

   #----------------------------------
   # format: "while [:alpha:] ;" or whilenot [a-z] ;

   "word*class*;*" {
     clear; get;

     "while" {
       clear;
       add "# while  \n";
       # the ruby pat.match? method should be faster than others
       add "while "; ++; get; --; add ".match?($self->{'peep)\n";
       add "  if $self->{'eof { break }\n  mm->read()\n}"; 
       put; clear; add "command*"; push; .reparse
     }

     "whilenot" {
       clear;
       add "# whilenot  \n";
       add "while !"; ++; get; --; add ".match?($self->{'peep)\n";
       add "  if $self->{'eof { break }\n";
       add "  mm.read()\n}"; 
       put; clear; add "command*"; push; .reparse
     }

     # error 
     add " < command cannot have a class argument \n";
     add "line "; lines; add ": error in script \n";
     print; clear; quit;
   }


  # arrange the parse> label loops
  # also, change .restart code before the parse label
  (eof) {
    "commandset*parse>*commandset*","command*parse>*commandset*",
    "commandset*parse>*command*","command*parse>*command*" {
      clear; 
      # indent both code blocks
      add "  "; get; replace "\n" "\n  "; 
      # change .restart code before parse> label
      replace "next # restart" "break; # restart"; 
      put; clear; 
      ++; ++; add "  "; get; replace "\n" "\n  "; put; clear; --; --;
      # add a block so that .reparse works before the parse> label.
      add "\n# lex block \n";
      add "while 1 {\n";
      get; add "\n  break;\n}\n"; ++; ++;
      add "if (restart) { restart = 0; next; }\n";
      # indent code block
      # add "  "; get; replace "\n" "\n  "; put; clear;
      # ruby doesnt support labelled loops (but swift does, and go?)
      # add "parse: \n";
      add "\n# parse block \n";
      add "while 1 {\n"; get;
      add "\n  break \n} # parse\n"; 
      --; --; put; clear;
      add "commandset*"; push; .reparse
    }
  }

  # -------------------------------
  # 4 tokens
  # -------------------------------

  pop;

  #-------------------------------------
  # bnf:     command := replace , quote , quote , ";" ;
  # example:  replace "and" "AND" ; 

  "word*quote*quote*;*" {
    clear; get;
    "replace" {
      #---------------------------
      # a command plus 2 arguments, eg replace "this" "that"
      clear; 
      # remove quotes around replace args
      # everything is done with regex in perl, no replace string
      ++; get; clip; clop; 
      escape "/"; replace '\\"' '"'; put; clear;
      ++; get; clip; clop; 
      escape "/"; replace '\\"' '"'; put; clear;
      --; --;
      # the \\Q stops special chars being interpreted.
      add "$self->{'work'} =~ s/\\Q";
      ++; get; add "/\\Q";
      ++; get; add "/;  # replace"; 
      --; --; put;
      clear; add "command*"; push; .reparse
    }

    add "Pep script error on line "; lines; 
    add " (character "; chars; add "): \n";
    add "  command does not take 2 quoted arguments. \n";
    print; quit;
  }

  #-------------------------------------
  # format: begin { #* commands *# }
  # "begin" blocks which are only executed once (they
  # will are assembled before the "start:" label. They must come before
  # all other commands.

  # "begin*{*command*}*",
  "begin*{*commandset*}*" {
     clear; 
     ++; ++; get; --; --; put; clear;
     add "beginblock*";
     push; .reparse
   }

   # -------------
   # parses and compiles concatenated tests
   # eg: 'a',B'b',E'c',[def],[:space:],[g-k] { ...

   # these 2 tests should be all that is necessary
   "test*,*ortestset*{*",
   "test*,*test*{*" {
     clear; get; add " || ";
     ++; ++; get; --; --; put; clear; 
     add "ortestset*{*";
     push; push;
     .reparse
   }

   # dont mix AND and OR concatenations 

   # -------------
   # AND logic 
   # parses and compiles concatenated AND tests
   # eg: 'a',B'b',E'c',[def],[:space:],[g-k] { ...
   # it is possible to elide this block with the negated block
   # for compactness but maybe readability is not as good.

   # negated tests can be chained with non negated tests.
   # eg: B'http' . !E'.txt' { ... }

   "test*.*andtestset*{*",
   "test*.*test*{*" {
     clear; get; add " && ";
     ++; ++; get; --; --; put; clear; 
     add "andtestset*{*";
     push; push; .reparse
   }

  #-------------------------------------
  # we should not have to check for the {*command*}* pattern
  # because that has already been transformed to {*commandset*}*

  "test*{*commandset*}*",
  "andtestset*{*commandset*}*",
  "ortestset*{*commandset*}*" { 
     clear; 
     # indent the code for readability
     ++; ++; add "  "; get; replace "\n" "\n  "; put; --; --; 
     clear; add "if ("; get; add ") {\n";
     ++; ++; get;
     # block end required
     add "\n}"; 
     --; --; put; clear;
     add "command*";
     push;
     # always reparse/compile
     .reparse
   }

  # -------------
  # multi-token end-of-stream errors
  # not a comprehensive list of errors...
  (eof) {
    E"begintext*",E"endtext*",E"test*",E"ortestset*",E"andtestset*" {
      add "  Error near end of script at line "; lines;
      add ". Test with no brace block? \n";
      print; clear; quit;
    }

    E"quote*",E"class*",E"word*"{
      put; clear; 
      add "Error at end of pep script near line "; lines; 
      add ": missing semi-colon? \n";
      add "Parse stack: "; get; add "\n";
      print; clear; quit;
    }

    E"{*", E"}*", E";*", E",*", E".*", E"!*", E"B*", E"E*" {
      put; clear; 
      add "Error: misplaced terminal character at end of script! (line "; 
      lines; add "). \n";
      add "Parse stack: "; get; add "\n";
      print; clear; quit;
    }
  }

  # put the 4 (or less) tokens back on the stack
  push; push; push; push;

  (eof) {
    print; clear;

    add '#!/usr/bin/perl
use strict;
use autodie;
#use warnings;

  # create the virtual machine object code and save it
  # somewhere on the tape.

  # code generated by "translate.perl.pss" a pep script
  # see http://bumble.sf.net/books/pars/tr/

  # mm is the machine object. 
package Machine;
  my $true = 1;
  my $false = 0;

  sub new {
    my $class = shift;  # 

    my $self = {
      # output =>    # A reference to the output and input streams
      # input =>     # if <stdin> and <stdout> are not used.
      input => *STDIN,
      output => *STDOUT,
      size => 300,      # how many initial elements in tape/marks array
      eof => 0,         # end of stream reached?
      charsRead => 0,   # how many chars already read
      linesRead => 1,   # how many lines already read
      escape => "\\\\",
      delimiter => "*", # push/pop delimiter (default "*")
      counter => 0,     # a counter for anything
      work => "",       # the workspace
      stack => (),      # stack for parse tokens 
      cell => 0,                # current tape cell
      tape => (),       # a list of attribute for tokens 
      marks => (),      # marked tape cells 
      peep => "" 
    };


    #$self->{"tape"} = 100;
    #$self->{"marks"} = 100;
    #@{$self->{"tape"}} = ("") x 100;
    #@{$self->{"marks"}} = ("") x 100;
    #$self->{"size"} = 100;
    # or use this, which may not duplicate the references
    # $my @arr = map { [] } 1..100;

    bless $self, $class;
    return $self;
  }

  # if (@tape < 5) check if length of tape is 4 or less
  sub setInput {
    my $self = shift;   # pointer to the machine
    my $newInput = shift; 
    print "to be implemented";
  }

  # read one character from the input stream and 
  #    update the machine.
  sub readChar {
    #$self->{"peep"} = getc(STDIN);

    my $self = shift;   # the pep/nom virtual-machine object
    if ($self->{"eof"}) { exit; }
    $self->{"charsRead"} += 1;
    # increment auto line counter
    if ($self->{"peep"} eq "\\n") { $self->{"linesRead"} += 1; }
    $self->{"work"} .= $self->{"peep"};
    $self->{"peep"} = getc($self->{"input"});
    #check
    if (eof($self->{"input"})) { $self->{"eof"} = 1; }
  }

  # test if all chars in workspace are in unicode category
  sub isInCategory {
    my $self = shift;   # pointer to the machine
    my $cat = shift; 
    #for ch in $self->{"work"}
    #  if not category(ch).start_with?(cat) then return false end
    #return True
  }

  # this walks the string and determines if the given char 
  # is already escaped or not # eg "ab\cab\\cab\c"
  sub unescapeChar {
    my $self = shift;   # the machine
    my $c = shift;
    my $cc = "";
    my $result = "";
    my $isEscaped = $false;

    foreach $cc (split(//,$self->{"work"})) {
      if (($isEscaped == $false) && ($cc eq $self->{"escape"})) {
        $isEscaped = $true;
      } else { $isEscaped = $false; }
      # remove the last escape character (usually backslash)
      if (($isEscaped == $true) && ($cc eq $c)) {
        $result =~ s/.$//s; 
      }
      $result .= $cc; 
    }
    $self->{"work"} = $result;
  }

  # this needs to walk the string and determine if the given char 
  # is already escaped or not # eg "ab\cab\\cab\c"
  sub escapeChar {
    my $self = shift;   # the machine
    my $c = shift;
    my $cc = "";
    my $result = "";
    my $isEscaped = $false;

    foreach $cc (split(//,$self->{"work"})) {
      if (($isEscaped == $false) && ($cc eq $self->{"escape"})) {
        $isEscaped = $true;
      } else { $isEscaped = $false; }
      if (($isEscaped == $true) && ($cc eq $c)) {
        $result .= $self->{"escape"};
      }
      $result .= $cc; 
    }
    $self->{"work"} = $result;
  }

  # a helper for the multiescape until bug
  sub countEscaped {
    my $self = shift;   # the machine
    my $suffix = shift; 
    my $count = 0;
    # no check
    my $s = $self->{"work"};
    $s =~ s/$suffix$//;
    while ($s =~ /$self->{"escape"}$/) {
      $count += 1;
      $s =~ s/$self->{"escape"}$//;
    }
    return $count;
  }

  # reads the input stream until the workspace end with text 
  sub until {
    my $self = shift;   # the machine
    my $suffix = shift; 
    # read at least one character
    if ($self->{"eof"}) { return; }
    # pass a reference to the machine hash with \% not %
    $self->readChar();
    while (1) { 
      if ($self->{"eof"}) { return; }
      # need to count the @escape chars preceding suffix
      # if odd, keep reading, if even, stop
      if ($self->{"work"} =~ /\Q$suffix$/) { 
        if ($self->countEscaped($suffix) % 2 == 0) { return; }
      }
      $self->readChar()
    }
  }  

  # this implements the ++ command incrementing the tape pointer
  # and growing the tape and marks arrays if required
  sub increment {
    my $self = shift;   # the machine
    $self->{"cell"} += 1;
    if ($self->{"cell"} >= $self->{"tape"}) { 
      # lengthen the tape and marks arrays by assigning to
      # length var
      $self->{"tape"} = $self->{"tape"} + 40;
      $self->{"marks"} = $self->{"marks"} + 40;
      $self->{"size"} = $self->{"tape"};
    }
  }

  # pop the first token from the stack into the workspace */
  sub popToken {
    my $self = shift;   # the machine, not local
    if (!$self->{"stack"}) { return 0; }
    $self->{"work"} = pop(@{$self->{"stack"}}) + $self->{"work"};
    if ($self->{"cell"} > 0) { $self->{"cell"} -= 1; }
    return 1;
  }

  # push the first token from the workspace to the stack 
  sub pushToken {
    my $self = shift;   # a pointer to the machine
    # dont increment the tape pointer on an empty push
    if ($self->{"work"} eq "") { return 0; }
    # need to get this from the delimiter.
    my $iFirst = index($self->{"work"}, $self->{"delimiter"});
    if ($iFirst == -1 ) {
      push(@{$self->{"stack"}}, $self->{"work"}); 
      $self->{"work"} = ""; return 1;
    }
    # s[i..j] means all chars from i to j
    # s[i,n] means n chars from i
    push(@{$self->{"stack"}}, substr($self->{"work"}, 0, $iFirst));
    $self->{"work"} = substr($self->{"work"}, $iFirst+1, -1);
    $self->increment();
    return 1;
  }

  sub printState {
    # print "Stack[${@self->{\"stack\"}.join(, )}] Work[#{$self->{\"work\"}}] Peep[#{$self->{\"peep\"}}]"
   # print "Acc:#{$self->{"counter"}} Esc:#{$self->{"escape"}} Delim:#{$self->{"delimiter"}} Chars:#{$self->{"charsRead"}}" +
   #      " Lines:#{$self->{"linesRead"}} Cell:#{$self->{"cell"}}"
  }

  sub goToMark {
    my $self = shift;   # a pointer to the machine
    my $mark = shift; 
    # search the marks 
    for my $ii (0..$self->{"marks"}) {
      if (@{$self->{"marks"}}[$ii] eq $mark) {
        $self->{"cell"} = $ii; 
        return;
      }
    }
    # mark was not found- fatal error!
    print("bad mark \'$mark!\'");
    exit();
  }

  # parses from a string and returns the result as a string.
  # useful for programatic use. Chopping last char??
  # does nothing if no input given.
  sub parseString {
    my $out = "";
    my $self = shift;       # the parse machine 
    my $in = shift;         # input string
    if (!defined $in) { return ""; }
    open my $input, "<", \\$in or return "[no input!]";
    open my $output, ">", \\$out or return "[no output!]";
    $self->parse($input, $output);
    return $out;
  }

  # parses from a file and writes to a string
  # useful for creating a pep/nom interpreter.
  sub parseFile {
    my $out = "";
    my $self = shift;          # the parse machine 
    my $infile = shift;         # input filename
    if (!defined $infile) { return ""; }
    open my $input, "<", $infile or return "[no input file!]";
    open my $output, ">", \\$out or return "[no output!]";
    $self->parse($input, $output);
    return $out;
  }

';  

    # save preamble (class) code in the current tape cell
    put; clear;

    #---------------------
    # check if the script correctly parsed (there should only
    # be one token on the stack, namely "commandset*" or "command*").
    pop; pop;

    "commandset*","command*",
    "beginblock*commandset*", "beginblock*command*" {
      "commandset*","command*" {
        clear;
        # indenting generated code  for readability.
        add "\n"; get; replace "\n" "\n  "; put; clear;
        add '\nwhile (!$self->{"eof"}) {\n'; get;
        add "\n}\n";
        replace "\n" "\n    "; put; clear;
        # restore the perl preamble (machine definition) from the tape
        ++; get; --;
        #add 'script: \n';
      }
      "beginblock*commandset*", "beginblock*command*" {
        clear; 
        # indent begin-block and main code for readability.
        add "\n"; ++; get; replace "\n" "\n  "; put; --; clear;
        add "\n"; get; ++; 
        add '\nwhile (!$self->{"eof"}) {\n'; get; add "\n}\n";
        replace "\n" "\n    "; --; put; clear;
        # get perl preamble from tape
        ++; ++; get; --; --;
        add "\n";
      }

      add '

  # This could have an input/output stream argument, or make the 
  # streams an attribute of the object.
  sub parse {
    my $self = shift;  # a machine reference
    my $in = shift;   # the input stream
    my $out = shift;  # output stream
    if (!defined $in) { $in = *STDIN; }
    if (!defined $out) { $out = *STDOUT; }
    #$self->{"output"} = *STDOUT;
    $self->{"output"} = $out;
    $self->{"input"} = $in;
    $self->{"peep"} = getc($self->{"input"});
    # the restart flag allows .restart to work before the 
    # parse label, in languages that dont have labelled loops
    my $restart = $false; \n';
    # the while loop has already been added around the 
    # script commands.
    get;

    add '
  }
  # end of Machine methods definition

  my $mm = Machine->new();

  # reads from <stdin> and writes to <stdout>
  $mm->parse();

  # --------
  # Other ways to call parse(), parseString() and parseFile()
  # Use an input file and print to stdout.
  # open IN, "<pars-book.txt" or die $!;
  # $mm->parse(*IN);

  # write results to an output file, reading from stdin
  # open OUT, ">test.txt" or die $!;
  # $mm->parse(*STDIN, *OUT);

  # parse/translate a string and return the result as string
  # my $result = $mm->parseString("new.new");
  # print $result;

  # parse/translate a file and return the result as a string
  # my $result = $mm->parseFile("test.txt");
  # print $result;

  # end of generated perl code\n';

      # put a copy of the final compilation into the tapecell
      # so it can be inspected interactively.
      put; print; clear; quit;
    }

    push; push;
    # try to explain some more errors
    unstack;
    B"parse>" {
      put; 
      clear; 
      add "[error] pep syntax error:\n";
      add "  The parse> label cannot be the 1st item \n"; 
      add "  of a script \n"; 
      print; quit;
    }
    put; clear;

    clear;
    add "After compiling with 'translate.perl.pss' (at EOF): \n ";
    add "  parse error in input script. \n ";
    print; clear; 
    unstack; put; clear;
    add "Parse stack: "; get; add "\n";
    add "   * debug script ";
    add "   >> pep -If script -i 'some input' \n ";
    add "   *  debug compilation. \n ";
    add "   >> pep -Ia asm.pp script' \n ";
    print; clear; 
    quit;

  } # not eof

  # there is an implicit .restart command here (jump start)