1 |
dpavlin |
46 |
#!/usr/local/bin/perl -w |
2 |
|
|
use strict; |
3 |
|
|
use File::Find; |
4 |
dpavlin |
56 |
use Getopt::Long; |
5 |
dpavlin |
46 |
|
6 |
dpavlin |
56 |
my $collection; # name which will be inserted |
7 |
|
|
my $path_add; # add additional info in path |
8 |
|
|
my $verbose; |
9 |
dpavlin |
46 |
|
10 |
dpavlin |
56 |
my $result = GetOptions( |
11 |
|
|
"collection=s" => \$collection, |
12 |
|
|
"path=s" => \$path_add, |
13 |
|
|
"verbose!" => \$verbose, |
14 |
|
|
"debug!" => \$verbose, |
15 |
|
|
); |
16 |
|
|
|
17 |
dpavlin |
46 |
my $dir = shift @ARGV || die "usage: $0 [dir]"; |
18 |
|
|
|
19 |
dpavlin |
56 |
|
20 |
dpavlin |
46 |
my $basedir = $0; |
21 |
|
|
$basedir =~ s,/[^/]+$,/,; |
22 |
|
|
require "$basedir/filter.pm"; |
23 |
|
|
|
24 |
dpavlin |
56 |
|
25 |
dpavlin |
46 |
find({ wanted => \&file, |
26 |
|
|
follow => 1, |
27 |
|
|
no_chdir => 1 |
28 |
|
|
}, $dir); |
29 |
|
|
|
30 |
|
|
sub file { |
31 |
|
|
|
32 |
|
|
return if (! -f || ! m/\.html*/i); |
33 |
|
|
|
34 |
dpavlin |
56 |
# skip index files |
35 |
|
|
return if (m/index_[a-z]\.html*/i || m/index_symbol\.html*/i); |
36 |
|
|
|
37 |
dpavlin |
46 |
my $path = $_; |
38 |
|
|
|
39 |
|
|
open(F,"$path") || die "can't open file: $path"; |
40 |
dpavlin |
56 |
print STDERR "$path" if ($verbose); |
41 |
dpavlin |
46 |
my $contents; |
42 |
|
|
while(<F>) { |
43 |
dpavlin |
56 |
$contents .= "$_"; |
44 |
dpavlin |
46 |
} |
45 |
dpavlin |
56 |
$contents .= "\n\n"; |
46 |
dpavlin |
46 |
|
47 |
dpavlin |
56 |
$contents = filter($contents,$collection); |
48 |
dpavlin |
46 |
|
49 |
dpavlin |
50 |
# die "zero size content in '$path'" if (! $contents); |
50 |
|
|
return if (! $contents); # don't die on empty files |
51 |
|
|
|
52 |
dpavlin |
46 |
my $mtime = time; |
53 |
dpavlin |
56 |
use bytes; |
54 |
dpavlin |
46 |
my $size = length $contents; |
55 |
|
|
|
56 |
dpavlin |
56 |
print STDERR " [$size]\n" if ($verbose); |
57 |
dpavlin |
46 |
|
58 |
dpavlin |
56 |
# add optional components to path |
59 |
|
|
$path .= " $path_add" if ($path_add); |
60 |
|
|
|
61 |
dpavlin |
46 |
# Output the document (to swish) |
62 |
|
|
print <<EOF; |
63 |
|
|
Path-Name: $path |
64 |
|
|
Content-Length: $size |
65 |
|
|
Last-Mtime: $mtime |
66 |
|
|
Document-Type: HTML |
67 |
|
|
|
68 |
|
|
EOF |
69 |
|
|
print $contents; |
70 |
|
|
|
71 |
|
|
} |