#*

 This script attempts to convert filenames and urls into html links using some
 simple word/line parsing. This could be expanded into recognising multiword
 patterns. 

NOTES
  
 "name.ext" [code|script|file|image|document]
 name.ext [code|script|file|image|document]
 [code|script|file|image|document] "name.ext"
 [code|script|file|image|document] name.ext

HISTORY

 19 june 2022
  just begun, not complete. As usual this becomes trickier 
  than one at first thinks. How to handle quotes, quotes with 
  multiple words, quotes containing links etc.
   
*#

  read;
  # ignore \r
  [\r] { clear; (eof) {.reparse} .restart }
  
  # dont tokenize non-leading space. One space will be printed
  # between each word.
  [ \t\f] { while [ \t\f]; clear; (eof) {.reparse} .restart }
  
  [\n] { 
    # make character count relative to line.
    nochars;
    # save the leading space in the nl* token 
    while [:space:];
    put; clear; add "nl*"; push; .reparse
  }

  # how to parse single line quoted text.
  '"' { 
    whilenot ["\n]; 
    (eof) { 
      put; clear; add "word*"; push; .reparse
    }
    read; 
    E'\n' {
      nochars; put; clear; add "word*"; push; .reparse
    }
    '""' {
      put; clear; add "word*"; push; .reparse
    }
    E'"' {
      put; clear; add "word*"; push; .reparse 
    }
  }

  # everything else is a word
  !"" { 
    whilenot [:space:]; 
    put; clear;
    add "word*"; push; .reparse 
  }

parse>
  # to visualise parse token reductions
   add "line "; lines; add " char "; chars; add ": "; print; clear; 
   unstack; add "\n"; print; clip; stack; 

  #-------
  # 1 token
  pop; 

  "word*" {
    # a trick!
    clear; get;
    # replace some html entities
    replace ">" ">";
    replace "<" "&lt;";
    put;
    B"http://",B"https://",B"www." {
      !"http://".!"https://".!"www." { 
        clear; add "<a href='"; get; add ">"; get; add "</a>";
      }
    }
    put; clear; add "word*"; 
  }

  # the last line of the file with no final newline char
  (eof) {
    "word*","text*" {
      clear; add "line*"; push; .reparse
    }
  }

  #-------
  # 2 tokens
  pop; 
  # I want to recognise 2 word structures, so need to separate
  # the text*word* reduction from the word*word* rule. 

  # is there any need for a file* token, link token etc? 
  "word*word*","text*word*" {
     clear; get; add " "; ++; get; --; put; 
     clear; add "text*"; push; .reparse
  }

  "word*nl*","text*nl*" {
    clear; get; ++; get; --; put; clear;
    add "line*"; push; .reparse
  }

  "line*line*","lineset*line*" {
    clear; get; ++; get; --; put; clear;
    add "lineset*"; push; .reparse
  }

  push; push;

  (eof) {
    pop;
    "word*","text*","line*","lineset*" { 
      clear; get; add "\n"; print; clear; quit;
    }
  }