But there’s no need to have the bloat of Thickbox and jQuery loading on all pages when most of them don’t even use it. That’s why Joost de Valk detailed a nice tip to only load thickbox when needed by checking the page content for references to it. All it takes is a few lines of code in your functions.php file.

Unfortunately, you end up with some broken images because the built-in Thickbox defines their locations in thickbox.js as relative paths:

var tb_pathToImage = "../wp-includes/js/thickbox/loadingAnimation.gif";
var tb_closeImage = "../wp-includes/js/thickbox/tb-close.png";

Those paths get tacked onto the URL of whatever page you’re on, which results in an invalid URL and a couple 404 errors for each use. Now that’s not unique to Joost’s solution. It happens whenever Thickbox is loaded with wp_enqueue_script( ).

It can easily be fixed by adding a couple lines of javascript in the footer to redefine the image path variables. That’s as simple as a few more lines of code in your functions.php file. We’ll use Joost’s technique of only loading it when the page content requires it:

function thickbox_image_paths() {
	global $post;
	wp_reset_query();
	if (is_singular() && strpos($post->post_content,'class="thickbox"') !== false) {
		$thickbox_path = get_option('siteurl') . '/wp-includes/js/thickbox/';
		echo "<script type=\"text/javascript\">\n";
		echo "	var tb_pathToImage = \"${thickbox_path}loadingAnimation.gif\";\n";
		echo "	var tb_closeImage = \"${thickbox_path}tb-close.png\";\n";
		echo "</script>\n";
	}
}
add_action('wp_footer', 'thickbox_image_paths');

My theme uses multiple loops so I had to add the wp_reset_query() to be able to examine the original content. Otherwise it would be looking at the last post in my recent posts box and not see that Thickbox was used in the current content.

There you have it. A nice technique from Joost and a simple addition to complement it.

We’re on our way to building an AppleScript to update all of the contacts in your Mac Address Book with pictures from the Gravatar service. This is part 3 of the series.

Part 1, Gravatar, AppleScript, and Address Book Pictures, introduced the topic and ended with a short AppleScript to construct a Gravatar URL given an e-mail address.

Then in part 2, Setting an Address Book Picture with AppleScript, we built a script showing how to successfully set a contact photo in Address Book.

Now we need to detect when the Gravatar service does not have an image on file for a given e-mail address. The problem is that Gravatar does one of several things when it doesn’t have a matching image: return a default image, return an image you specify, or generate an image in one of several styles. But in every case it returns an image.

This is all well and good if you’re displaying Gravatars on a blog comments page. The Gravatar URL ends up inside an img tag and this keeps it dead simple. It’s also nice for everyone to get some kind of avatar even if they haven’t created one.

But for our purposes we don’t need a default image that ends up being attached to a contact. Not only does it clutter our contacts, but it would make it more difficult if we later wanted our script to only update contacts with no existing picture and leave the others alone.

Detecting a no match condition would also be useful in a WordPress plugin even if a default image was used. For example, I’ve always wanted to display some sort of indicator and a link to Gravatar for those who don’t have a Gravatar yet. If we can detect that a default image was used then we can set a CSS class on the image or do other creative things.

So how do we tell if there is no image match? It’s simple with a little undocumented trick in the Gravatar URL.

There is a d= parameter where we can specify a default image to return when there is no match. If we set that to null by using two empty URL encoded quotes then instead of a default image we get nothing. For example:

For our Address Book script we need to save the image to a file anyway so we just check for the existence of the file or a zero size. We can skip that entry or take other action if there is nothing there.

In the case of a WordPress plugin it might be less convenient because we may have to hit the Gravatar servers twice; once to check for a no match condition, then again to get the image with whatever default we actually want. Of course, if we use a local default file then it would still only be one call to the Gravatar service but that rules out using one of the cool generated defaults.

(If any of the Gravatar folks are listening, it would be great to have an option in the URL that means I’m using this for further processing and not in an img tag then also pass back some sort of status telling if the image is a default.)

Here’s a demonstration AppleScript that makes use of the above technique to download a Gravatar for a given e-mail address and take action only if there was a matched image: (Open in Script Editor)

(Updated with a small bug fix on 03/25/2009.)
(Updated on 04/23/2009 to remove my testing e-mail address and display the downloaded file path.)

-- Download a Gravatar image and see if it's a match or default.
-- Released under GPL.
-- by Doug Smith, http://smithsrus.com
 
-- The e-mail address to look up.
set email to "someone@somewhere.com"
 
-- Calculate an MD5 for the e-mail address.
set md5_email to do shell script "md5 -q -s `echo " & email & " | tr '[:upper:]' '[:lower:]'`"
 
-- Construct the Gravatar URL.
set grav_url to quoted form of ("http://gravatar.com/avatar/" & md5_email & "?d=" & "%22%22" as text)
 
-- Make a file name in which to temporarily save the Gravatar.
set grav_file to (path to temporary items) & "gravatar" & md5_email & ".jpg" as text
set grav_POSIX_file to quoted form of POSIX path of grav_file
 
-- Download the Gravatar image to the temporary file with a timeout of 10 seconds.
do shell script "curl " & grav_url & " -m 10 -o " & grav_POSIX_file
 
tell application "System Events"
	if exists file grav_file then
		tell me to display dialog "Yes, we have a matching Gravatar downloaded to " & grav_POSIX_file & "." buttons {"Okay"} default button 1
		delete file grav_file
	else
		tell me to display dialog "Sorry, there was no matching Gravatar." buttons {"Okay"} default button 1
	end if
end tell

Well, we’re almost to our goal of a script to automatically update all of our Address Book contacts with Gravatars. I’ll put it all together next time in part 4.

It connected to my blog effortlessly and allowed me to easily add categories, tags, and a photo.

Great job WordPress team!

I only have a couple minor gripes that I hope will be addressed later:

1. It would be easier to type if the application supported landscape mode and the wider keyboard that provides.

2. It’s too bad WordPress does not extract and store the GPS information along with the other metadata it grabs from photos.

I would love to hear comments from others who have tried it. How has it worked for you?

Fortunately, WordPress 2.5 has a way for plugins to add additional color schemes. So I whipped up a quick plugin to add the style necessary to mark those visited links in a different color. You can grab a copy from my downloads page for admin color scheme: Fresh Plus Visited.

I believe they are still taking registrations, so you may not be too late to sign up if you haven’t already.

Then the spam flood came that most of us have seen by now. These spam comments are harder to detect because they seem to be written by a human and customized to somewhat fit the post content. You can often tell only by the URL they promote and the awkward wording of the comment that shows they don’t truly grasp the topic. More of them are slipping through the trusty spam-fighting superheroes: Akismet, Bad Behavior, and Spam Karma.

A little research showed that my URLs were ending up on lists of dofollow blogs on sites made just for the purpose of gaming Google’s Pagerank. (I won’t link to them and give them publicity.) The lists are all neatly sorted into groups with the highest Pagerank sites at the top.

And then today, I found comments with a referrer from a site that sells a script to help you find blogs without nofollow. I was tempted to buy it to see how it worked so I could block it, but I just couldn’t stand the idea of giving them any money. (It’s dofollowblogs.net, which I also will not link to.) What really irks me is that they refer to this as white-hat link building to make it sound palatable to the average Joe! (Oh joy, here’s another one: http://www.commenthut.com/)

I still believe that using nofollow fails to stop spam, however, I am now convinced that:

1. Turning off nofollow can be a spammer magnet.
2. Some sites are promoting dofollow to increase the number of sites available to abuse.
3. Some sites are promoting displaying a dofollow graphic on your site only to make it easier for them to find you.
4. Not using nofollow may effect Google search results ranking. It may be coincidence, but I had a site that by every statistic should have ranked #1 for certain keywords, yet it remained at the #3 position for many months. The site leapt past the competition to #1 within a month of turning nofollow back on. I can’t prove a correlation, but further study is certainly warranted.

Well, I’m still not ready to give up. So what can we do about? Here are a few possibilities:

1. When you run across sites like this, report to Google through their webmaster tools.
2. WordPress users should flag these comments as spam in Akismet. If we all work together, we can make a dent in the sources.
3. Use a dofollow plugin with advanced features, like the one I mentioned above.
4. It may be possible to block some of these with some .htaccess rules looking for referrers including “nofollow” or “dofollow”. Be careful, though, as it’s easy to block legitimate traffic or yourself with this, e.g., I had to quit blocking those so Google searches for this post would get through.
5. Don’t display the little graphics announcing that you dofollow unless you rename the files so they’re not so easy to find with a Google search.

Anyone have any more ideas or know anything about the scripts being used to abuse our blogs?

• Additional checks for plugins knowing about the new tag system so they don’t get flagged as incompatible. i.e., Sitemap Generator 3.0.
• Internationalization of text. The .po file is included. If you do a translation send the files over and I’ll be happy to post them on the download page.
• Miscellaneous clean up.
• Added GPL license text.

Get it on the Upgrade Preflight Plugin page. (My apologies for this not being on the WordPress Extend plugins directory yet. I’ve applied for access but they apparently have quite a backlog in approving new plugins.)

You can download the translated version from his site.

I plan on doing proper internationalization in the next version of the plugin. That should make translation easier for anyone who wants to make one.

I decided to release it in case it’s of use to others. More information and a download are on my WordPress Upgrade Preflight Check plugin page.

This is my first plugin that I’ve released publicly. I’m certainly not an expert PHP coder, so please go easy on me. Suggestions for improvements are welcome.

Hopefully, this will entice me to blog about a few things as I think of them because it’s now so easy to do. Some of my old articles are still getting a decent amount of views so I’ll soon be moving those into posts and redirecting traffic there.

I still haven’t settled on a theme. I’d like to customize one for this site but I just don’t have the time right now. Thanks to Jim Mitchell for his bogart theme. It’s clean and easy on the eye so it will definitely do the job for the time being.