3 # $Cambridge: exim/doc/doc-docbook/Tidytxt,v 1.3 2007/08/29 13:37:28 ph10 Exp $
5 # Script to tidy up the output of w3m when it makes a text file. First we
6 # convert sequences of blank lines into a single blank line, to get everything
7 # uniform. Then we go through and insert blank lines before chapter and
8 # sections, also converting chapter titles to uppercase.
10 # We also have to do some character translation in the first pass. It seems
11 # that xmlto now generates Unicode in its HTML pages. This gives three problems:
12 # (1) It inserts the byte sequence C2 A0 (U+00A0) as a fixed-width space;
13 # (2) It inserts a whole slew of "box drawing" characters round the heading.
14 # (3) It uses U+25CF as its bullet character.
19 foreach $line (@lines)
21 $line =~ s/\x{c2}\x{a0}/ /g;
22 $line =~ s/\x{e2}\x{94}[\x{80}-\x{bf}]/-/g;
23 $line =~ s/\x{e2}\x{97}\x{8f}/*/g;
27 $line = "" if $lastwasblank;
34 # Find start of TOC, uppercasing its title
36 for ($i = 0; $i < scalar @lines; $i++)
38 $lines[$i] = "TABLE OF CONTENTS\n" if $lines[$i] =~ /^Table of Contents/;
39 last if $lines[$i] =~ /^1\. /;
42 # Find start of first chapter
44 for ($i++; $i < scalar @lines; $i++)
45 { last if $lines[$i] =~ /^1\. /; }
47 # Process the body. We can detect the starts of chapters and sections by
48 # looking for preceding and following blank lines, and then matching against
52 for (; $i < scalar @lines; $i++)
54 next if $lines[$i-1] !~ /^$/ || $lines[$i+1] !~ /^$/;
58 if ($lines[$i] =~ /^(\d+)\. / && $1 == $chapter + 1)
62 $lines[$i] = "\n\n" . ("=" x 79) . "\n" . uc($lines[$i]);
65 # Start of next section
67 elsif ($lines[$i] =~ /^(\d+)\.(\d+) / && $1 == $chapter && $2 == $section + 1)
70 $lines[$i] = "\n$lines[$i]" . "-" x (length($lines[$i]) - 1) . "\n";