#! /usr/bin/perl
-# $Cambridge: exim/doc/doc-docbook/Tidytxt,v 1.1 2005/06/16 10:32:31 ph10 Exp $
+# $Cambridge: exim/doc/doc-docbook/Tidytxt,v 1.2 2006/02/01 11:01:01 ph10 Exp $
-# Script to tidy up the output of w3m when it makes a text file. We convert
-# sequences of blank lines into a single blank line.
+# Script to tidy up the output of w3m when it makes a text file. First we
+# convert sequences of blank lines into a single blank line, to get everything
+# uniform. Then we go through and insert blank lines before chapter and
+# sections, also converting chapter titles to uppercase.
-$blanks = 0;
-while (<>)
+@lines = <>;
+
+$lastwasblank = 0;
+foreach $line (@lines)
{
- if (/^\s*$/)
+ if ($line =~ /^\s*$/)
{
- $blanks++;
+ $line = "" if $lastwasblank;
+ $lastwasblank = 1;
next;
}
- print "\n" if $blanks > 0;
- $blanks = 0;
- print;
+ $lastwasblank = 0;
+ }
+
+# Find start of TOC, uppercasing its title
+
+for ($i = 0; $i < scalar @lines; $i++)
+ {
+ $lines[$i] = "TABLE OF CONTENTS\n" if $lines[$i] =~ /^Table of Contents/;
+ last if $lines[$i] =~ /^1. /;
+ }
+
+# Find start of first chapter
+
+for ($i++; $i < scalar @lines; $i++)
+ { last if $lines[$i] =~ /^1. /; }
+
+# Process the body. We can detect the starts of chapters and sections by
+# looking for preceding and following blank lines, and then matching against
+# the numbers.
+
+$chapter = 0;
+for (; $i < scalar @lines; $i++)
+ {
+ next if $lines[$i-1] !~ /^$/ || $lines[$i+1] !~ /^$/;
+
+ # Start of chapter
+
+ if ($lines[$i] =~ /^(\d+)\. / && $1 == $chapter + 1)
+ {
+ $chapter++;
+ $section = 0;
+ $lines[$i] = "\n\n" . ("=" x 79) . "\n" . uc($lines[$i]);
+ }
+
+ # Start of next section
+
+ elsif ($lines[$i] =~ /^(\d+)\.(\d+) / && $1 == $chapter && $2 == $section + 1)
+ {
+ $section++;
+ $lines[$i] = "\n$lines[$i]" . "-" x (length($lines[$i]) - 1) . "\n";
+ }
}
+print @lines;
+
# End