Broken things:
* Duplicated anchors are Ok, and should be made unique by PoD::HTML
* 'Contents' from deeply nested links is broken 
* You have to keep style.css in each dir with pods
* non-pod files, like pdfs and html are not linked to the index.html

The broken things are mostly due to the multi-levelness of the files,
which wasn't the intention at the guide level. So hardcoded links in the
template files are broken. Will have to make these dynamically (i.e.
replace by place holder and make styles.css a special file.

======================================================================
Need to find a way to put the files:

  sample-config/html2ps-global.conf
  sample-config/html2ps.html
  sample-config/hyphen.tex

Into a global directory using the Makefile.PL and make the code find
them. I don't know how to do it, since I've no idea where these can
fit on the filesystem.

====================================================================== 

I've included the hyphen.tex in the bundled html2ps info, now I should find out
how to make the config file to point to it. Will it accept a relative path?
I'll try

====================================================================== 

if you get a moment speedup a bit the s/// code:

eric:

New version!

Benchmark: timing 5000 iterations of HTML pattern, Non Greedy...
HTML pattern:  6 wallclock secs ( 5.40 usr +  0.00 sys =  5.40 CPU) @ 926.19/s (n=5000)
Non Greedy:  7 wallclock secs ( 6.96 usr +  0.00 sys =  6.96 CPU) @ 718.29/s (n=5000)

I felt it was unfair to include the $rexx compilation inside the
benchmark loop. Also note that I've taken out the (), instead I
initialize $pre. This assumes that the <pre> tags are correctly
balanced and not nested.

#!/usr/bin/perl

use Benchmark;

$text = join "\n", map {"<pre>" . "aaaaaaa"x5 . "$_</pre>"} 'aa'..'cc';
$|=1;
my $rex = qr!</?pre>!;

sub non_greedy {
        my $count = 0;
        $count++ while $text =~ m|<pre>(.*?)</pre>|g;
}
sub html_pat {
        my $count = 0;
        my $pre = ($text =~ m/^$rex/);
        for (split $rex, $text) {
                ++$count if $pre;
                $pre = !$pre;
        }
}
timethese(5000, {
        'Non Greedy' => \&non_greedy,
        'HTML pattern' => \&html_pat,

});



====================================================================== 

Still broken: links to the code in external files!!!

########################## code references:
#sub process_code {
#    my($filename, $comment) = @_;
#    $OUT .= qq{
#	       <p><a href="code/$filename"><code>$filename</code></a> -- $comment <i>(Note: See the html version for the source code)</i></p>
#	      };
#}

====================================================================== 

I don't think this is important...

########################## 
#
# check this difference you made in the two files:
# 

## pdf
#    process_text(\$linktext, 0);
#    if ($link) {
#	$link =~ s|(\./)+||;
#	$link =~ s|/+||;
#	$link =~ s|(\.html)?#|/|;
#	$link =~ s|^/|$curr_base/|; # normalize internal links
#	$s1 = "<A HREF=\"$link\">$linktext</A>";
#    } else {

## html
#    process_text(\$linktext, 0);
#    if ($link) {
#	$s1 = "<A HREF=\"$link\">$linktext</A>";
#	$link =~ s|(\./)+||;
#	$link =~ s|/+||;
#	$link =~ s|(\.html)?#|/|;
#	$link =~ s|^/|$curr_base/|; # normalize internal links
#	push @{$r_links_to_check->{$curr_base}}, $link ;
#    } else {


* Guide<->Book difference:
PS2html:

====================================================================== 

### ps2html TODO ###

code to fix the () problem in the ToC.  I did end up having to figure
it out afterall since the codewalkthrough part of the WebDB docs has a
bunch of chapter headings going over each method (eg The new() Method,
The addRecord() Method, etc).


Sample quickie build.pl code:

# Generate PS files
if ($generate_ps) {
   make_indexps_file();
   print "Generating a PostScript\n";
   my $command = "html2ps -f html2psrc -o $ps_root/temp.ps ";
   $command .= join " ", map {"$ps_root/$_.html"} "index", @ordered_srcs;
   print "Doing $command\n";
   system $command;
   open PSFILE, "<$ps_root/temp.ps";
   open PSFILE2, ">$ps_root/temp2.ps";
   while (<PSFILE>) {
     s/\\201\\202/\(\)/;
     print PSFILE2 $_;
   }
   close PSFILE2;
   close PSFILE;
   unlink "$ps_root/temp.ps";
   rename "$ps_root/temp2.ps", "$ps_root/extropia_app_guide.ps";
}


Unlike the spaces problem, I could not solve the () problem in
html2ps.  Whatever it is doing to replace the () with \240 and \241 is
not consistent and I end up corrupting the resulting postscript
code. So what I did was made it open the ps file and filter the ()
manually (rather than modifying the html2ps code) after it is
generated but before the PDF generation happens.

> I guess that if we are going to bundle the source of html2ps, it's
> wiser to fix the innards and not the resulting ps. What do you
> think? I quess it'd be easy to find the code that generates the TOC
> and correct it.

It wasn't easy for me. I found the several places where the () were
being replaced and tried them one by one (and turned off previous
changes) and what resulted were corrupted files for some reason--
could be that I dont know enough about postscript. But I wasn't too
keen on spending a day to solve a problem that is ancillary to getting
something out.

====================================================================== 
