sub size_cmp($$) {
    return (stat($_[0]))[9] <=> (stat($_[1]))[9];
}
sub get_files($) {
    return glob($_[0].'/*.{jpg,jpeg,gif,png}');
}
sub main {
    $col=0;
    print "<table>\n<tr>\n";
    @imgs = get_files($dir);
    for (@imgs) {
        print "\t<td width=\"200\">\n<img src="$Path/$_"></td>\n";
        print "\n</tr>\n<tr>" if 0 == (3 % ++$col);
    }
    print "\n</tr>" if 0 != (3 % $col);
    print "</table>\nthere are ".@imgs." thumbs<br>\n";
}
[download]

broquaint

I have been burnt by glob in the past. I can no longer recall what I did (wrong)?, but I wound up with filenames containing spaces split across seperate elements of the returned array.

Older versions of person require spawning a child process to an external command (tcsh) in order to perform the glob. This may be a cost too high to bear.

Truly, opendir/readdir/closedir are your friends ;).

--

g r i n d e r

Thanks for your suggestions.

I switched to file globbing. It seems simpler and is built into perls greater than 5.6.0

I eliminated one of the map lines as the glob now returns the full path. I think this sped things up a bit

I switched to mod ing by three as you suggested. % is probably 100 times more expensive than addition and eliminated but 1/100000 of a second isn't much more than a 1/1000 and the modulo is easier to read.

I left the error checking in. --A few times it was handy to check to see that stat actually succeeded

I left the sort in, I do want the most recent images first. I also suspect that the $a $b variables supplied by sort are quicker than $_[0] and $_1

I left the printed comments and the print "content-type: text/html\n\n"; in as the former help me figure out problems when looking at HTML source and the later <large>:-></large> is needed for the HTML to display.

#!/usr/local/bin/perl 

require 5.6.0;
use warnings;
use strict;
use File::Spec::Functions;
use File::Basename;



my $gHTMLPath='/go/pics/thumbs/';

my $gthumb_dir=catdir('..','go','pics','thumbs');    


##################################
#                                #
# no user servicable parts below # 
#                                #
##################################
if ($gthumb_dir=~m|/|){
    $gthumb_dir.='/';
}elsif ($gthumb_dir=~m|\\|){
    $gthumb_dir.='\\';
}elsif ($gthumb_dir=~m|:|){
    $gthumb_dir.=':'; # we're on a Mac ??? not tested

}else{ 
    die "on unknown OS or couldn't get path at start \n"
}




# FILE: listthumbs.pl  REVISION DATE: 09-13-2001
#
# CGI to list display jpeg, gif & png files in a directory

#####################
#  Declarations     #
#####################

sub get_files($);    # get a list of jpeg files;

sub size_cmp();        # used in sort function compares two values    


sub main();        # 

######################
# Run our program    #
######################

main();

######################
#    Definitions     #
######################

sub size_cmp(){
    my @a_stat=stat($a) or die "couldn't stat\n $a \n$!\n";
    my @b_stat=stat($b) or die "couldn't stat \n $b\n$!\n";

    #7 is size  so
    # sorting by size puts the duplicates next to each other

    #9 is modification time
    # sort reverse order
    return $b_stat[9] <=> $a_stat[9];
}


sub get_files($) {
    my @result=glob("$_[0]*.{jpg,jpeg,png}");

    if (0==@result) {
        die "couldn't glob in get_Files()";
    }
    return @result;
    
}


sub main(){
    my @pics=get_files($gthumb_dir);
    @pics=sort size_cmp @pics;

    for (my $i=0;$i<@pics;$i++){
        $pics[$i]=basename($pics[$i]);
    }

    print "content-type: text/html\n\n";
    print "<!-- start code by $0 -->\n" ;

    my $count=0;

    print '<TABLE>';
    print '<TR>';
    foreach (@pics){
        $count++;
    
        print "<TD width=200 align=left>\n";
        print '<IMG SRC='. $gHTMLPath . $_ .'><BR>'."\n";
        print ;
        print "</TD>\n";
        if (0==$count% 3){
            print "</TR>\n";
            print '<TR>';

        }

    } # end loop
    print "\n</tr>" if 0 != (3 % $count);
    print '</TABLE>';
    print "there are $count thumbs<BR>\n";
    print "<!-- end code by $0 -->\n" ;
[download]



--mandog

Perhaps even -

sub get_files {
    return sort {
        (stat($a))[9] <=> (stat($b))[9]
    } glob($_[0]."/*.{jpg,jpeg,png}") or
         die($_[0]." contained no images\n");
}
[download]

And if you're worried about speed, then the modulo operator is probably the least of your worries, as pretty much all operations are going to be costly (but if you were really looking for speed, then you'd be using C ;o) This should eliminate a lot of unncessary operations ie saving an array into memory but only using one element, and having to enter the size_cmp function for every comparison.
HTH

broquaint