If you've ever tried to back up your Flickr account or download the public Flickr photos of other users, then your options were pretty limited on Linux. After trying and failing to get FlickrDown to work under wine, I decided to write my own... in perl :)

The various Flickr API's on CPAN are incomplete with respect to obtaining the list of public/friends photos. I had to use XML::Parser::Lite::Tree::XPath to extract the photos from each Flickr page.

By default, we obtain the largest photo available (orginal -> large -> medium -> small), populate the EXIF comments field with the Flickr photo title, description and url. We verify the file type with File::MimeInfo::Magic and change/add the file extension as necessary.

During testing I received numerous duplicate files which were mostly resolved by adding a rudimentary duplicate file checker.

It should work on just about any platform that Perl does.

Thoughts?

shell$ zooomr -u pippo@example.com -p pluto foto*.jpg

# Let the program ask you for the password, so that you don't have
# to show it on the screen
shell$ zooomr add -u pippo@example.com photo*.jpg

# Add a tag
shell$ zooomr add -u pippo@example.com --tag awesome photo*.jpg

# Add two tags
shell$ zooomr add -u pippo@example.com -t awesome -t good photo*.jpg
[download]

Update: added more features and checks for good uploads vs fake ones (Zooomr sometimes simply seems to throw an upload in the dust bin).

Update: added support for browser cookie file, managing external authentication via browser (needed for OpenID accounts). Added some missing documentation.

Update: added support for summary file at the end of upload, to report managed files and their URIs (see this post in Zooomr groups).

Update: changed upload URI to conform to current non-flash upload URI.

There's no particular reason to use POE::Component::RSSAggregator rather than XML::RSS::Feed, other than the fact that I heard about the POE version first and was interested in trying something in POE. (Thanks again everyone for helping me get around the problems due to my out-of-date POE in POE::Component::RSSAggregator breaks LWP::Simple::get.)

Also, I heartily recommend this station to everyone!

Downloads all your posted comments to a directory. Generates a yaml file with definitions of all the comments, suitable for other scripts to use for parsing the data.

Can be the first of a number of scripts to e.g. find mod-bombers, etc.

Tab size 4.

Link to Better documentation.
Re: SF_form_secure

More Examples.
SFLEX's scratchpad
I am now working on a 5.0 that will use parts of CGI::Util for the experation time and fix a bug in action 5 that returns the version if the matching code is blank.
Still trying to put together more documentation for this code so one can use it the right way.

Blame the author of JavaScript::XRay. He demoed this at YAPC::NA 2006 but there wasn't a proxy in the distribution. I wrote this so I could use it. Just run this and set your browser to use it as a proxy. The simplest way to use this is just run it and tell your browser to use a proxy at port 8080 on your local host.

bib_canonicalizer.pl - takes a citeulike url or a .bib file and searches it against the Collection of Computer Science Bibliographies, outputing a canonicalized file with all available fields filled. It does not clobber your citeulike URLs or tags.

citeulike_pdf_grabber.pl - downloads and caches all .pdfs it can find from your citeulike account (including private) and spools them to a printer. Will not print duplicates even over multiple runs as long as you don't delete the cache folder. It's good for printing only the things you have most recently added. Outputs a "missing.html" file with links to the citeulike articles it could not find a .pdf for. You will probably have to customize some of the regexs for the databases you use the most.

A simple way to download stuff from KFAI's program archive. I just wrote the thing, it seems to work. I hope so. You'll need mplayer and lame.

KFAI is a public radio station in Minneapolis and it rocketh.

NAME

linkextor — extract particular links from HTML documents

SYNOPSIS

linkextor [ -b baseurl ] [ -f filter ] [ files and urls ]

DESCRIPTION

linkextor prints links found in given HTML documents, filtered by any given expressions. If no files are specified on the command line, the input document is read from STDIN. You can specify URLs on the command line; they will be downloaded and parsed transparently.

By default, no links are filtered, so the output will include references to every external resource found in the file such as stylesheets, external scripts, images, other documents, etc.

After considering criteria, all matching links will be printed to STDOUT. You could pipe them to wget -i - to mass download them, or to while read url ; do ... ; done or xargs to process them further, or do anything else you can do with a list of links.

ARGUMENTS

• -h, --help

  See a synopsis.

• --man

  Browse the manpage.

OPTIONS

• -b, --base

  Sets base URL to use for relative links in the file(s).

  This only applies to local files and input read from STDIN. When parsing documents retrieved from a URL, the source URL is always the base URL assumed for that document.

• -f, --filter

  Defines a filter.

  If no filter has been specified, all external links will be returned. This includes links to CSS and Javascript files specified in <link> elements, image links specified in <img> elements, and so on.

  If one or more filters have been specified, only links conforming to any of the specified criteria will be output.

  Filters take the form tagname:attribute:regex. tagname specifies which tag this filter will allow. attribute specifies which attribute of the allowed tags will be considered. regex specifies a pattern which candidate links must match. You can leave any of the criteria empty. For empty criteria on the right-hand side of the specification, you can even omit the colon.

  Eg, -f 'a:href:\.mp3$' will only extract links from the <href> attributes in the document's <a> tags which end with .mp3. Since <a> tags can only contain links in their <href> attribute you can shorten that to -f 'a::\.mp3$'. If you wanted any and all links from <a> tags, you could use -f a::, where both the attribute and regex component are empty, or just -f a: because both components to the right are empty, you can leave out the colons entirely. Likewise, you could use -f img to extract all images.

BUGS

None currently known.

SEE ALSO

• perlre
• xargs(1)
• wget(1)

COPYRIGHT

(c)2005 Aristotle Pagaltzis

LICENCE

This script is free software; you can redistribute it and/or modify it under the same terms as Perl itself.

Lookup Seagate hard drives by model number.

$ bin/lookup_seagatehd ST118202FC ST336752FC ST150176FC

Desc=ST-118202FC Fibre Channel FC-AL, Dual Port (Cheetah 18)
Capacity=18.2
RPM=10000
Access time=6

Desc=ST-336752FC Fibre Channel FC-AL, Dual Port (Cheetah X15 36LP)
Capacity=36.7
RPM=15000
Access time=3.6

Desc=ST-150176FC Fibre Channel FC-AL, Dual Port (Barracuda 50)
Capacity=50.01
RPM=7200
Access time=7.6
[download]

Three scripts:
• update_awstats.pl Perl control script, and fetches gzipped country database updates
• update_geo-ip_db.sh Shell script to gunzip the compressed update and replace the old database file.
• update_awstats.sh Shell script to update AWStats.

Updated: Fixed bug with retrieving film information by its title.

Tidy is a useful command-line utility that cleans up messy and invalid HTML, producing nice, pristine, XHTML. I wrote this module as a wrapper around tidy to let you clean up html on the fly as it is served by apache. It also uses the Apache::Filter framework so you can use it to clean up the results of any other Filter compliant perl handler (or registry script using Apache::RegistryFilter). This could be very useful if you are trying to get your site to validate but are stuck with an old CMS that produces messy, invalid markup.

You can also download the module from: http://thraxil.org/tidy/Apache-Tidy-0.1.tar.gz

Any and all suggestions are welcome. If no one finds any big problems, I may try to upload it to CPAN.

After rebuilding Apache hundreds of times on just as many boxes, each with their own tweaks and configuration options with mod_php, mod_perl, mod_ssl, MySQL and so on, I was tiring of doing it all manually. Normally you have to do the following:

• Fetching the tarballs
• Verifying the md5sums
• Unpacking the tree
• Configuring the tree
• Building the tree
• Testing the build
• Installing the tree
• ...etc.

• You can't instruct it to use the system-installed MySQL headers and libraries
• It insists on building its own PHP from scratch
• Many other libraries (gd, t1, xpm) all get built by ApacheToolbox and installed in /usr/local
• ...and so on. Not clean, not scalable. Too intrusive to strictly-controlled systems.

Well, now I don't have to!

With this script, all you have to type is 'build', and it will fetch, verify, unpack, configure, build, test, and install Apache, PHP, mm (shared memory), mod_ssl, mod_perl, and so on for you, in your target prefix, all in one shot. You just run 'build' and it does the rest for you as DSO, with just a 14k httpd server daemon when it completes (on my test boxes).

For those wondering what this has to do with Perl, I rely on mod_perl heavily for my work, and rebuilding Apache with each new release of mod_perl, Perl, mod_ssl, and so on gets tedious. This removes that tedium, so I can get back to work, faster, with less babysitting of the code.

There are three files: build (the main script), functions (some helper routines), and source.md5 (md5sums of the target tarballs). You'll need all three in your work directory to properly run this.

Hopefully PerlMonk's wrapping won't kill this too much.

Suggestions and patches are welcome, of course.

This module alters the operation of LWP::Simple so that it keeps track of any cookies presented by the server. Any import options are passed directly to HTTP::Cookies.

When my company's sysadmin emailed me some log files from sysstat (http://perso.wanadoo.fr/sebastien.godard/) with the note "Can you make some graphs out of those?", I started to:

• write a converter from sysstat logs into a gnuplot data file (basically a row to col converter)
• write a form based web-frame to select files, columns and options, generate queries for a gnuplot backend and display the resulting graphs
• finally write a gnuplot backend that outputs corresponding graphs

The gnuplot backend is a module currently called Apache::Gnuplot that can be used for nearly any statistical data files in gnuplot format.

The converter and web-frame are not published yet, but you may contact me if you like to have a look at it. See http://alex.zeitform.de/unsorted/saplot.png for a screenshot.

Fellow monks, I would like to read your comments on possible errors or improvements. And please let me know if this is of any interest to put it on CPAN. Thanks.

Update: code cleanup for readability

Update2: added input validation thanks to Abigail-II

There are two scripts: image upload, image display. With these two files you can have yourself your very own image gallery!

User uploads an image at a time and specifies some image information (such as title and description). The image display prints/displays 20 images per page (5x4) as thumbnails (100x100 with link that opens in new window. Under each image displays: file name, image title, image description, image height, image width.

What I learned from this:

I am still working on my top Seti@home teams script.

This uses: (in no particular order)
grantm's XML::Simple,
samtregar's HTML::Template,
mojotoad's HTML::TableExtract,
gbarr's Net::FTP,
LWP::UserAgent,
HTML::Entities
and DBI.

Many thanks to the authors of these wonderful modules!

Output is available at Top 40 Seti@home Teams.

The URL that Berkeley uses for the XML team stats may change once they "officially" announce it. They still need to fix some things... See my Malformed XML thread at the Seti@home Bulletin Board for more infomation.^{update 4}

Updated the ParseTeamXML subroutine so that it no longer uses a hash as a reference. Thanks tye and thunders for helping me resolve this issue.

Update 2 added color coding for rise/fall in rank.

Update 3 Updated code to work around duplicate teams in the top teams page.

Update 4 Thread is AWOL.

This script is designed to be run from /perl or /cgi-bin on an Apache webserver. Given a root directory, it will create dynamic pages based on the contents of the directory that's referenced. As it navigates the directory hierarchy via user choices, it determines if it's already created a page for that dir, and creates a cache page in that directory, pushing the cached page to the next request unless the directory contents changes, when it will regenerate the cache page.

In short, it gets faster the more it's used.

Enhanced data display for MP3 files via MP3::Info

Examples can be seen here. Project page is here.

Comments, criticism welcome. This is one of my first "real" perl scripts.

-JPJ

Code reviews/critiques are welcomed and requested.

Updated: Jun. 28, 2002

• Squashed 2 bugs
• UserAgent support
• Proxy support
• Progress report

Fetch a random web page from a list of URLs.   Waits a random length of time between fetches.

Written with two purposes in mind:
  * log internet connectivity
  * nail up analog dial connection
Mostly the latter, so my ISP won't drop my connection during looong downloads.

There are a number of commandline options+arguments.   Perhaps the more interesting are

• --verbose=1 which reports only delta of success/fail ^{ala zeno from Internet Connection Uptime Logger}
• --errDelay=1 to cause almost-immediate retry on fetch failure
• --daemonize which backgrounds and disconnects so vty can be closed
• --logging to print progress, options, and versions to file

From a Perlish perspective, this has been an exercise in rand, IO::Tee and Proc::Daemon, plus Getopt::Long bobs that I ended up not using.   I started out "use"ing Proc::Daemon, but had to copy+tweak subs from it to allow logging-to-file while daemonized.

Thanks to tye, fletch, tachyon, lemming, chmrr and others for tips and suggestions.   And as always, constructive criticism is welcome and invited.

This script is derived from one of the examples I developed for the chapter of my book that introduces the SOAP::Lite module.

It posts a journal entry to the journal system at use.perl.org, without using the browser interface, and without faking a POST to the CGI interface. It uses instead the still-alpha SOAP interface that use.perl currently offers.

All elements of the SOAP request may be specified on the command-line, including entry subject, whether to enable comments, uid and passwd, and the option of reading the entry from a file (the default is to read from STDIN, filter-like). It also will attempt to set the authentication credentials from your Netscape cookie file. But as I said, those can be set by the command-line (and the cmd-line switch overrides the Netscape cookie).

In a first-response to this code posting, I will post a XEmacs defun (which should also work under GNU/Emacs) that I use to post from within XEmacs by selecting the text I want to upload into the region. More on that in the follow-up node.

--rjray

The use Perl; site is starting to provide SOAP interface support to the journal system there. Much of the interface is still alpha, and I'm trying to help out pudge by testing what he has, and offering suggestions here and there as I run into things.

This is a simple demonstration/proof-of-concept script that uses SOAP::Lite and XML::RSS to turn the last 15 journal entries from a given user into a RSS 1.0 syndication channel.

Caveats: There is not yet an interface for turning a nickname into a user-ID, so you have to provide the user ID as the first (and only required) parameter to the script. You may provide the nickname as the second parameter. The titles and such are very simple, as this goes for minimal network traffice over feature-bloat. But it's mostly for illustrative purposes. As the interface fleshes out and becomes more stable, I expect to have a more functional version of this. Oh, and it writes to STDOUT, making it suitable as a filter to other XML/RSS processing apps.

A lot of web applications rely on this non-standard behavior in browsers so I think it would be a good idea to integrate this to LWP. See Redirect after POST behavior in LWP::UserAgent differs from Netscape for reference.

Look for the XXX marker in the code to see where this code differs from the one in LWP::UserAgent.

Kudos to dws for his help when I became inevitably entagled. Sadly, I have yet to attain CGI-zen.

Use of this module is very simple with the replacement of the use base qw/CGI::Application/; pragma with use base qw/CGI::Application::Session/;. All data to be stored and retrieved should be placed within the $self->{'__SESSION_OBJ'} hash. Additional new method parameters include SESSIONS_EXPIRY, SESSIONS_NAME and SESSIONS_PATH to set the respective parameters of the client-side ID cookie.

This code can be expanded and improved upon greatly, but it demonstrates a proof-of-concept for session utilisation within a state-based CGI engine.

Replacement for the WWW::Search::Google module. I apologise for the scrappiness of the code, but at least it works.

Thanks crazyinsomniac and hacker.

Update 06/03/2002: Surprisingly, this module still works . After all the changes that Google has gone through since the time I first released it, I would expect it to have broken a long time ago, considering it parses HTML rather than some stable format. There's an interesting story at slashdot about googling via SOAP - maybe this is the future direction this module could take?

The following act can act as a HTML script filter, stripping Javascript, VBScript, JScript, PerlScript, etc. from the HTML code.

This weeds out all the "scriptable" events from the HTML 4.01 specifications and all the <script> elements.

It takes a filename as argument, or if there's no argument, read from STDIN. All the output is done on STDOUT.

This piece of code should be pretty reliable, but I'd be interested to know if there's a flaw in this code.

While working on the re-write I discovered something that turned 75 LWP requests into two requests; one for the indices information, and one for the list of symbols for which I needed data. The discovery was that Yahoo has a publically facing application server that offers up all financial data in CSV rather than html. I could request ALL the symbols I needed at once, and it was returned in orderly, well formatted, easy to parse CSV.

Needless to say, I saw a significant performance increase.

Yesterday I was asked if I could create a small web application that allowed users to get current currency exchange rates, and rather than re-inventing the wheel, I went back to the stock work I had done and found that the same server would do the exchange for you if given the correct URL. I have included an example of a correct URL, but I am leaving getting the correct Country Currency codes as an exercise to the user. They are readily available from a few different sources.

Usage might be somethig like:

use LWP::Simple;
use LWP::UserAgent;
use CGI;

$|++;
print "Content-type: text/html\n\n";
&Header();
&PrintForm();
my $q = new CGI;
if ($q->param("s") && $q->param("s") ne "DUH" && $q->param("t")ne "DUH
+"){ 
    my $sym = $q->param("s").$q->param("t")."=X";
    my ($factor, $date, $time) =  (&GetFactor($sym))[2,3,4];
    $time =~ s/"//g;
    $date =~ s/"//g;
    print '<CENTER><p><b><font face="Arial" size="4">RESULTS:</font></
+b> ';
    printf("<b><font face=\"Arial\" color=\"#0000FF\" size=\"4\">%s %s
+ = %s %s as of %s, %s </font></b></p>",&commify($q->param("a")), $q->
+param("s"), &commify(($factor*$q->param("a"))), $q->param("t"),$date,
+$time);
    print "</CENTER><P>";
}

&PrintFoot(\$q);
[download]

C-.

As strange as it seems, I couldn't find here a code that cleanly modifies HREF attributes in A starting tags in a HTML page.

So here's one that I whipped up quickly to answer a question on fr.comp.lang.perl

It surely could be easily improved to include other links (<script>, <img src="...">, etc.), but you get the idea...

The only (slight) caveats is that the 'a' starting tag is always lowercased and the order of the attributes are lost. But that should not matter at all.
Also, this code won't print 'empty' attributes correctly (though I can't think right now of any empty attributes that are legal with 'a')

To use this script, you have to modify the $new_link variable, and then call the script with the URL of the page to be modified. Every <a href="..."> will have the $new_link added at the start of the href, and the old URL will be properly escaped.

It is probably useless as is, but with a minimum of tweaking, you can easily do what you want.
Actually, it might be a good thing to turn this little script into a module where you would only have to do the URL munging, without worrying about the whole parsing stuff...

Independent subroutines can be specified for different URLs, in the single example provided, the subroutine virus_alert is executed should the Symantec web page have changed since the last execution of the script.

So, I set about writing my own simplified robot code which had one and only one function - return a list of crawled URLs from a start URL address.

#!/usr/bin/perl -w

use Local::SiteRobot;
use strict;

my $robot = Local::SiteRobot->new(
    DEPTH           =>  10,
    FOLLOW_REGEX    =>  '^http://www.cowsnet.com',
    URLS            =>  [ 'http://www.cowsnet.com.au' ]
);
my @pages = $robot->crawl;
print STDOUT $_, "\n" foreach @pages;
[download]

The code I feel is quite self explanatory - /msg me if you have any questions on usage.

This script is a website link checking tool. It extracts and checks *all* links for validity including anchors, http, ftp, mailto and image links.

Script performs a recursive search, width first to a user defined depth. External links are checked for validity but are not followed for obvious reasons - we don't want to check the whole web.

Broken anchors, links are repoted along with the server error. All email addresses harvested are checked for RFC822 compliance and optionally against an MX or A DNS listing.

More details in pod

Don't blame me, ar0n had the idea first.

But nonetheless, it's a fully functional webring.
You too can make your own and impress your friends, by showing them how many people share you love of Cheesy things.

Big thanks to virtualsue for hosting the Cheese Ring! (see node below)

update: Eradicated a couple of bugs that prevented the code to compile. oops.

update - Thu Aug 23 07:47:26 2001 GMT: The truncate filehandle was not right, but virtualsue spotted it!

update - Sun Aug 26 13:32:41 2001 GMT: Fixed the check for http://perlmonks.org, thanks crazyinsomniac!

Sat Sep 22 21:35:45 UTC 2001: Fixed the encoding of characters other than 0-9a-zA-Z.

Fetches HTML files by URL instead of file so html2ps will process images.   Add <!--NewPage--> to HTML as needed for html2ps to process.   Links *not* converted from HTML to PDF   8^(

Requires external libraries html2ps and gs-aladdin, but no perl modules.

From a Perlish perspective, this has been a minor exercise in (open|read)dir, grep, and difference of lists.   As always, comments and critique gladly welcomed.

Latest update(s):     2001-05-30     22:25

• Thanks to jeroenes for noticing funky indents and for suggesting cleaner exclusions syntax.
• Implement and test simpler exclusions syntax.
• Eliminate redundant code with PrintLocs()
• Add explanatory comments for html2ps and ps2pdf syntax.

Update 2001-11-13: Complete rewrite of the adr2xbel script. It follows a bit more closely the python script found in the PyXML library demos.

This program is a web based GUI for your database. It has currently only been tested for MySql, but the idea is for it to work with any DB.

The program give you a list of tables and allows you to Create, Edit and Drop them, as well as viewing their schema and data. It also provides a command line for other functions

Feel free to give me feedback/critisim. If you make any major additions, please let me know.

Update 4/3/2001:
Added privilege granting and revoking funtions.

Simple scripted website generator.
Uses CGI.pm, HTML::Template and param('page') to eliminate duplication of code and markup in toy website.   Also uses CSS to separate style from structure (as much as possible, anyway).   Clearly not suited for anything industrial-strength.

Code and CSS files go in a web-published directory.   Common, content and template files go in a non-webpub directory.   Each web page in the site is defined by its' own content file.

Thanks to ovid, chipmunk, Petruchio, davorg, chromatic and repson for helping me figure this out.

Most recent update:   2001-05-10
Hashamafied a few of the passel o'scalar variables.

Update: 2002-05-11 note to self: look into Storable to potentially reduce response times, or Data::Denter so don't have to unTaint incoming data structure.   Props to crazy for suggesting Denter and for initial benchmark strongly favoring Storable as perf king.

use Benchmark    qw( cmpthese timethese );
use Storable     qw( freeze thaw );
use Data::Denter qw( Indent Undent );

timethese( 2_000, { Data::Denter => \&DENTOR, Storable => \&STORKO });
print "\n\n\n";
cmpthese( 2_000, { Data::Denter => \&DENTOR, Storable => \&STORKO });

sub DENTOR {
  my $in  = Indent \%BLARG;
  my %out = Undent $in;
  return();
}

sub STORKO {
  my $in  = freeze \%BLARG;
  my %out = %{ thaw($in)};
  return();
}
__END__


Benchmark: timing 2000 iterations of Data::Denter, Storable...
Data::Denter: 12 wallclock secs (11.71 usr +  0.01 sys = 11.72 CPU) @ 170.69/s (n=2000)
Storable:     6 wallclock secs  ( 5.60 usr +  0.00 sys =  5.60 CPU) @ 357.27/s (n=2000)


Benchmark: timing 2000 iterations of Data::Denter, Storable...
Data::Denter: 12 wallclock secs (11.80 usr +  0.00 sys = 11.80 CPU) @ 169.53/s (n=2000)
Storable:     6 wallclock secs  ( 5.62 usr +  0.00 sys =  5.62 CPU) @ 356.06/s (n=2000)
              Rate Data::Denter     Storable
Data::Denter 170/s           --         -52%
Storable     356/s         110%           --

cookbook.pl gets & stores daily cookbook recipes from perl.com. Written to run as a cronjob (no output except the file) & includes a sample daily crontab to automate the process. cookbooks.pl lists the saved recipes in alphabetical order. The viewer requires two images to work properly, 1. a small transparent gif named blank.gif, and 2. a Perl Cookbook cover image such as these at fatbrain and amazon, both stored in the same dir as the scripts & saved files.

Update: fixed url to point to the new recipe location at perl.com

A perl module to access the library of congress' book database.

Unlike Amazon or Barnes and Noble, you can't just look up a book in the Library of Congress database with an ISBN; you need to first initialize a session with their web server. That's all this module does, it initializes a session for you and returns either a url of a book's web page or a reference to a hash containing that book's data (author, title, etc.)

This Perl filter fixes bad HTML comments, such as <!----------- really ---------bad ------ comments ---------->. (Such comments are bad because, according to the spec, each -- -- pair within <! > delimits a comment. This means <!-- -- --> is not a complete comment, for example.)

The code reads in the entire file, then finds each occurrence of <!-- ... --> and uses tr/// to squash each run of hyphens to a single hyphen. The assignment to $x is necessary because $1 is read-only.

• Colanderwax@someregistereddomain.com
• JesusRedmond@someregistereddomain.com
• crudbedbug@someregistereddomain.com
• tyingbabies@someregistereddomain.com
• leekchecker@someregistereddomain.com
• hottrousers@someregistereddomain.com

I have posted this code to show an extensible method of serving images in the event that any monks are called upon to write a similar script.

This can easily be modified to allow calls directly from a CGI script in the following format:

<img src="/cgi-bin/images.cgi?image=category1234" height=40 width=40 alt="some text">

Then, you'd add the following to your code:

use CGI;
my $query = new CGI;
[download]

$ENV{'QUERY_STRING'} =~ m!([a-zA-Z]+)(\d+)\.($types)$!;
[download]

$query->param('image') =~ m!([a-zA-Z]+)(\d+)\.($types)$!;
[download]

It's definitely beta quality, so any suggestions would be appreciated.

Gets the Slashdot headlines every 30 minutes.

CustomLog |/path/to/websplit.pl combined

I wrote this mod_perl handler to give easy timeouts to restricted web pages. It is very elementary, but useful. Please give me some comments at my email address above if you wish.

It requires a directory "times" under your /usr/local/apache/conf/ directory, owned by the user:group running the Apache child processes, for your timestamp files.

Usage: See in-code docs.

Update v0.21

• I added better docs and fixed a bug or two.
• I also moved most of the config info into the httpd.conf file and only moved configurable stuff to .htaccess.
• Added concept of Minimum Time Out and Mode.

Update v0.20

• I sped up the routine that checks time since last visit. It now stats a file, grabs the number of seconds since last modification, and uses that for $last_time. Then opens the time file rw to update the modification time.
• I added option to put the DEBUG mode into the .htaccess file.

TO DO:

• Write documentation
• Make into format usable on CPAN

Usage:

% webster_client word
[download]

Updated: as suggested by kudra, added a comment to the code about double-checking the name obtained, and why we don't do it in this case.

Make your own Portal! Amaze your friends! Confuse your enemies!

This quick hack takes a list of RSS feeds and pulls the links into a local database. If you don't know what RSS is, it's the cool XML headline standard used by My Netscape. Lots of sites provide RSS feeds that you can use to make headline links on your own site (like Slashdot.org, ZDNet, etc.). I used this to make a little headline scroller using DHTML on our company Intranet. This script works best with a scheduler (like cron) to update on a periodic basis.

For a comprehensive list of available feeds, take a look at http://www.xmltree.com.

Comments/improvements more than welcome ;)

Since I don't have access to my ISP's server logs but still want to have some idea of who's visiting my website, I developed this little script that can integrate easily with any existing site. All you need is CGI access of some type.

To install, save the script so it is executable. Also, you'll need to set the $home, $logfile, $ips (IP addresses you want to ignore), and %entries (labels and expressions to match) variables. Be sure to "touch" the logfile and make it writable by the web server's user.

Pick an image, preferably a small one, on your page.

<img src="/cgi-bin/showpic/path_to_pic_in_document_root/pic.jpg">
[download]

Each time someone accesses the page with that image, an entry is made in the log with the date, time, and either hostname or IP address. Here's an example of the output. Enjoy.

Wed Apr 05 13:08:26 2000 d208.focal6.interaccess.com resume
Wed Apr 05 13:29:29 2000 audialup197.phnx.uswest.net
Thu Apr 06 01:31:47 2000 adsl-63-194-241-209.dsl.lsan03.pacbell.net

I run it from a cron script every Thursday and Sunday morning.

To use it, try faucet 8080 -io wsproxy.pl

v0.4 and 0.5 added banning of Javascript and HTML pages.
v0.6 had a file locking bug nuked and preliminiary tempcache support
v0.7 now has an embedded gif (here) v0.8 now has mostly functional (self-cleaning) tempcache support.