--- trunk/spider/progspider 2004/04/06 15:06:58 72 +++ trunk/spider/progspider 2004/08/28 22:15:59 81 @@ -1,4 +1,4 @@ -#!/usr/local/bin/perl -w +#!/usr/bin/perl -w use strict; use File::Find; use Getopt::Long; @@ -50,7 +50,7 @@ Path-Name: $path Content-Length: $size Last-Mtime: $mtime -Document-Type: HTML +Document-Type: html* EOF print $contents; @@ -76,6 +76,8 @@ } close(F); + return if (! $html); + my ($pre_html,$pages,$post_html) = ('
',$html,''); ($pre_html,$pages,$post_html) = ($1,$2,$3) if ($html =~ m/^(.+?
)(.+)(<\/pre>.+?)$/si); @@ -97,7 +99,7 @@ } else { - return if (! -f $path || ! m/\.html*$/i); + return if (! -f $path || ! m/\.(html*|php|pl|txt|info|log|text)$/i); # skip index files return if (m/index_[a-z]\.html*/i || m/index_symbol\.html*/i);