#!/usr/bin/perl

# syn analyzes the syntax of HTML pages. It outputs a list of orphan tags
# or an indented list of a single tag type for debugging.
# It accepts file wildcards and outputs a list of all file errors.

# Written by P. Lutus Ashland, Oregon lutusp@arachnoid.com 6/3/96

@orphan = ( # these tags can be unpaired
  "!--",
  "p",
  "br",
  "img",
  "input",
  "bgsound",
  "embed",
  "hr",
  "base",
  "li",
  "!doctype",
  "option",
  "isindex",
  "dt",
  "dd",
  "link",
  "meta",
  "nextid",
);

if($ARGV[0]) {
  $continuous = 0;
  $argstr = 0;
  $argval = "";
  $verbose = 0;
  $indentlist = 0;
  $indent = 0;
  $pline = 0;
  foreach $fn (@ARGV) {
    if($argstr != 0) {
      $argval = $fn;
      $argstr = 0;
    }
    elsif($fn eq "-v") {
      $verbose++;
    }
    elsif($fn eq "-i") {
      $indentlist++;
      $argstr++;
    }
    elsif($fn eq "-c") {
      $continuous++;
    }
    else {
      if(-T $fn) { # if it is a test file
        
        $err = 0;
        undef %ar;
        undef $/; # grab the entire file at once
        open(DATA,$fn);
        $file = <DATA>;
        close DATA;
        $/ = "\n";
        $lines = 0;
        $lines = ($file =~ s/\n/ /g);
        undef %pcs;
        $n = (@pcs = split(/\</,$file));
        if($n > 0) {
          foreach $q (@pcs) {
            if($q ne "") {
              
              $r = $q;
              
# find the token up to but not including ' ','=','>'
              
              $q =~ s/[\s=>].*//;
              
# make it lower case
              
              $q =~ tr/[A-Z]/[a-z]/;
              
# strip and count the '/' delimiter
              
              $decr = ($q =~ s/^\///);
              $m = 0;
              $tagflg = (($q eq $argval) || ($argval eq "all"));
              foreach $z (@orphan) { # is this a tag that can be single?
                if($q eq $z) {
                  $m++;
                }
              }
              if($m == 0) { # not OK to be single
                if($ar{$q}) { # OK exists - do nothing
                }
                else { # create it
                  $ar{$q} = 0;
                }
                $ar{$q} += ($decr)?-1:1;
                if($tagflg) {
                  $indent += ($decr)?-1:1;
                }
              }
            }
            if(($indentlist) && ($tagflg)) {
              $tab = " " x $indent;
              print "$tab$r\n";
              &check_line;
            }
          }
        }
      }
      foreach $k (keys(%ar)) { # now print the errors
        $q = $ar{$k};
        if($q != 0) {
          if($err == 0) {
            print "$fn:\n";
            &check_line;
          }
          $v = ($q < 0)?"</$k>":"<$k>";
          $q = abs($q);
          $err += $q;
          print "$v $q\n";
          &check_line;
        }
      }
      if(($verbose) || ($err > 0)) {
        print "$fn: $lines lines, $err errors.\n";
        &check_line;
      }
    }
  }
}
else {
  print "usage: -c continuous listing (no page stops) -v (verbose)\n";
  print "       -i tagname/\"all\" (indented list of a chosen tag or all)\n";
  print "       names of HTML files inc. wildcards.\n";
}

sub check_line {
  if($continuous == 0) {
    $pline++;
    if($pline > 20) {
      $pline = 0;
      print "(Press Enter for more):";
      <STDIN>;
    }
  }
}