JQuery to the rescue!  The previous web guy had used Superfish, which is a fine system on its own, but he’d only used the bits and pieces he really needed.  Thus, when two-level nested menus became three, the script was not able to handle it.

First step: Use what they already had without reinventing the wheel:

<?php wp_list_pages('title_li=&depth=2'); ?>

Became:

<?php wp_list_pages('title_li=&depth=3'); ?>

And Christian at Kresi.at provided a nice starting point for me to jump from using JQuery in WordPress menus.  He really should be applauded for such a simple solution.  I just had to style the 3rd tier menus and I was set!  (Text blurred to protect client identity).

Nice.

I also managed to wrangle in Image Uploading issues – caused by two separate problems on two different sites.  Freeing up a bit of file permissions did the trick on one, and discovering that there was an entire duplicate WordPress installation elsewhere and re-directing the uploads folder fixed the second.

The whole line of fixes and menu additions took a little over 3 hours.  A total success for myself and the client!

I’m crazy like Gary Busey when it comes to tracking bugs – I know exactly what changes I made, when I made them, and which files were affected.  Since WordPress doesn’t really have a client version control system (particularly for plugins) you have to track for yourself.

The first thing I did was ask for help.  I know this might seem like I started in reverse, but I like to have as many eyes as possible looking with me for a possible problem.  I got some very good recommendations on possible problems – including taking a closer look at the headers, making the site XHTML 1.0 – strict compliant, and so on.

After testing out the site in a few different browsers (all DOM Level 2 browsers were fine, IE8 was fine) I determined that there were multiple problems.

1. Facebook Connect needs the widget to be present on the page to work.  You can’t just install it and not add the widget, because it sends an asynchronous request to Facebook.  The request object comes back… and FBC looks for the widget to display the information.  If the widget isn’t there, FBC thinks that something was wrong with the data, so it asks Facebook for the info again.  This creates an infinite request loop (in IE7).  This might be because I was still using the beta version of FBC… so, who knows.  This fix allowed the website to show up in IE8.
2. I downloaded HttpWatch to load my header information in IE7… just to see where the page stopped loading.  Note: the full retail version of this program costs $400… but I wasn’t that desperate.  The program was giving me inconsistent results – twice it would load two parts of the page and then stop, then five more tests revealed that it would load 5-8 items and then stop.  This told me that a plug-in was the likely culprit, but I needed to eliminate other sources of errors before I could narrow it down for sure.
3. XHTML Validation – I used the W3C XHTML validator and discovered that I had incorrectly coded a small portion of the “print” plug-in, causing a mismatch in opening and closing elements.  This ended up being a minor problem – and in some areas where I need centered graphics and advertisements, I’m not sure how to get around the <center> tag.  So, I got it down to 5 errors rather than the 33 that I started with.
4. iTouchTheme plugin (to make your site easy to use for iPhones and iTouches) was using conditional formatting and was operating incorrectly.  I have a sinking suspicion that this was the straw that broke the camel’s back.  As soon as I patched this plugin to the next version, the site began to show up in IE7.
5. As a side project, I had a hell of a time trying to get the menu to appear in the same place in IE browsers and DOM Level 2 browsers… but as it turns out, the Margin property of a Div is handled differently when it is contained within a parent Div element… (in Freshy – the header div contains the menu div).  I eliminated the offending tag and now the menu displays properly (and in the right spot!) in all of my test browsers.

It was definitely aggrivating to keep on hearing, “Hey dude, your link is dead,” but the debugging experience was pretty cool, and I solved some issues on my website that I wouldn’t have otherwise thought about.  The total time invested was about 5 hours.

The culmination of ONE amp energy drink, 2 hours of continuous work, a 30 minute break and then another 8 hours of continuous work… I’m pretty proud of this one.  I think it turned out very well.  We’ll see how Deloris likes it tomorrow, but I have a feeling she’ll like this a lot better than the old version… which I would have taken a picture of, but since I designed THAT too, I’ll let it die in obscurity.  Bwahaha.

It’s definitely been a fun project and it’s only about 90% completed, but the heavy lifting in the design work is almost completely done.  Barring any bugs that may crop up, I’m satisfied that this design will last for a good long while.