#!/bin/sh
# Compiles all .rl files for Ohcount.
# Written by Mitchell Foral. mitchell<att>caladbolg<dott>net.

for file in *.rl
do
  echo "Compiling $file"
  if [ -z "`grep '#EMBED' $file`" ]
  then
    # no embedded language
    ragel $file -o "`cut -d '.' -f1 <<< "$file"`.h"
  else
    # embedded language(s)
    elangs=`grep -o "#EMBED([a-z_]\+)" $file | 
      sed -e 's/#EMBED(\([a-z_]\+\))/\1/'`

    # embed all languages into this temporary file to be compiled
    parser=${file}.tmp

    # grab all lines before #EMBED statements
    e_line=`grep -n -m 1 "#EMBED([a-z_]\+)" $file | cut -d ':' -f1`
    sed -n -e "1,$e_line p" $file > $parser

    # embed each language in the temporary file
    for lang in $elangs
    do
      echo -e "}%%\n%%{\n" >> $parser

      e_parser=${lang}.rl
      # extract the embedded language contents for embedding
      s_line=`grep -n "^%%{" $e_parser | cut -d ':' -f1`
      e_line=`grep -n "^}%%" $e_parser | cut -d ':' -f1`
      s_line=`expr $s_line + 1`
      e_line=`expr $e_line - 1`
      sed -n "$s_line,$e_line p;" $e_parser |
        sed -e "s/^[ \t]*machine[^;]\+;//;
                s/^[ \t]*write[^;]\+;//;
                s/^[ \t]*include[^;]\+;//;" >> $parser
    done

    # grab all lines after #EMBED statements
    s_line=`grep -n "#EMBED([a-z_]\+)" $file | cut -d ':' -f1 | sort -r |
      grep -m 1 "^[0-9]\+"`
    sed -n -e "$s_line,$ p" $file >> $parser

    ragel $parser -o "`cut -d '.' -f1 <<< "$file"`.h"
    rm $parser
  fi
done