#! /usr/bin/perl

# $Cambridge: exim/doc/doc-docbook/TidyHTML-filter,v 1.2 2005/11/10 12:30:13 ph10 Exp $

# Script to tidy up the filter HTML file that is generated by xmlto. The
# following changes are made:
#
# 1. Split very long lines.
# 2. Create reverse links from chapter and section titles back to the TOC.
# 3. Turn <div class="literallayout"><p> into <div class="literallayout"> and
#    a matching </p></div> into </div> to get rid of unwanted vertical white
#    space.


$tocref = 1;
$thisdiv = 0;

# Read in the filter.html file.

open(IN, "filter.html") || die "Failed to open filter.html for reading: $!\n";
@text = <IN>;
close(IN);

# Insert a newline after every > because the whole toc is generated as one
# humungous line that is hard to check. Then split the lines so that each one
# is a separate element in the vector.

foreach $line (@text) { $line =~ s/>\s*/>\n/g; }
for ($i = 0; $i < scalar(@text); $i++)
  { splice @text, $i, 1, (split /(?<=\n)/, $text[$i]); }

# We want to create reverse links from each chapter and section title back to
# the relevant place in the TOC. Scan the TOC for the relevant entries. Add
# an id to each entry, and create tables that remember the new link ids. We
# detect the start of the TOC by <div class="toc" and the end of the TOC by
# <div class="chapter".

# Skip to start of TOC

for ($i = 0; $i < scalar(@text); $i++)
  {
  last if $text[$i] =~ /^<div class="toc"/;
  }

# Scan the TOC

for (; $i < scalar(@text); $i++)
  {
  last if $text[$i] =~ /^<div class="chapter"/;
  if ($text[$i] =~ /^<a href="(#[^"]+)">/)
    {
    my($ss) = $1;
    my($id) = sprintf "%04d", $tocref++;
    $text[$i] =~ s/<a/<a id="toc$id"/;
    $backref{"$ss"} = "toc$id";
    }
  }

# Scan remainder of the document

for (; $i < scalar(@text); $i++)
  {
  if ($text[$i] eq "<div class=\"literallayout\">\n" && $text[$i+1] eq "<p>\n")
    {
    $text[++$i] = "";
    $thisdiv = 1;
    }
  elsif ($thisdiv && $text[$i] eq "</p>\n" && $text[$i+1] eq "</div>\n")
    {
    $text[$i] = "";
    $thisdiv = 0;
    }
  elsif ($text[$i] =~ /^<h[23] /)
    {
    $i++;
    if ($text[$i] =~ /^<a( xmlns="[^"]+")? id="([^"]+)">$/)
      {
      my($ref) = $backref{"#$2"};
      $text[$i++] = "<a$1 href=\"#$ref\" id=\"$2\">\n";
      my($temp) = $text[$i];
      $text[$i] = $text[$i+1];
      $text[++$i] = $temp;
      }
    }
  }

# Write out the revised file

open(OUT, ">filter.html") || die "Failed to open filter.html for writing: $!\n";
print OUT @text;
close(OUT);

# End