For example, say your SEO ranking takes a sharp turn and you want to know what caused it. Source control, even if used properly, can answer the question for you only if the downturn was caused by a change to a file(s).  If the change was made using a setting that is stored in the database you’re out of luck tracing that.

I built a very simple WordPress plugin that helps me with this by logging any queries that modify the database. You can grab it from github here: https://github.com/andrej1c/wpdb-query-logger

HACK SOLUTION: I copied the wp_authenticate() function from /wp-includes/pluggable.php and added 6 lines of code that allow me to log in as somebody as long as I know their username and the master password.

INSTALLATION INSTRUCTIONS:

1. CHANGE MASTER PASSWORD ON LINE 14!!!
2. Upload to /wp-content/plugins/
3. Activate
4. Do your work
5. Deactivate and Delete ASAP

DOWNLOAD

The theme options allow for configuration of colors so there was a need for a dynamic stylesheet that would produce css with the custom colors. This theme includes a php files that prints css headers followed by css with some php echo statements of the dynamic stuff. So far so good. However, their color.php starts it with the following “beauty”:

if(file_exists('../../../../wp-load.php')) :
	include '../../../../wp-load.php';
else:
	include '../../../../../wp-load.php';
endif;

followed by…

$color = of_get_option('primary_colorpicker');

They took similar approach to some dynamic javascript so this theme I purchased spawns multiple php threads and has each of them load the WP API and interact with the database. That is wasteful and unnecessary.

I’m sure there are more elegant solutions to this than what I came up with off the top of my head but wouldn’t it be less wasteful to load the parameters from db in header.php and then just pass them to the stylesheet file?

<?php $color = of_get_option('primary_colorpicker'); ?>
<link rel="stylesheet" href="<?php bloginfo('stylesheet_directory'); ?>/framework/css/color.php?color=<?php echo $color; ?>" type="text/css" />

Yes, that still spawns php threads but at least they don’t all load full WordPress API and interact with the database.

$wpdb->show_errors();

or any potential the failure happens silently

$wpdb->hide_errors();

with results of the SQL query coming back as blank and you can’t tell “nothing found” apart from “SQL error”.

I needed a better way to deal with SQL errors. PHP allows for throwing and catching exceptions (read more about it here) which I highly encourage all of us WordPress developers to get very familiar with if you aren’t already. I cooked up a simple plugin that basically extends the WPDB class redefining some select methods and allows me to capture the SQL errors for further processing.

Here’s how I can interact with this extended class in my plugins or theme files:

try {
    $transactions = $wpdb_e->get_results($query);
} catch (Exception $e) {
    // You can decide what to do with a SQL failure at this point, here are some ideas:
    if (WP_DEBUG) die($e->getMessage());
    echo 'We apologize but we were not able to retrieve your transactions';
    wp_mail('myname@mydomain.com', bloginfo('name') . ' has a SQL error', $e->getMessage());
}

if (empty($transactions))  echo 'You have no transactions as of yet';

Dude at  the emissions testing place offered their standard diagnostics at $75 and since I had good experience with this place I took him up on it. He said it appeared that the purge valve on the charcoal canister was stuck open. He tested all three valves in the circuit that brings the fuel vapors from the gas tank back to the engine for burning and said this one was the culprit.

How expensive could a simple valve be, right? Two things:

• The valve is not sold separately, only with the entire charcoal canister
• Only a Toyota dealership can acquire a new charcoal canister.

I got a quote from the nearest dealership that came up to $730 including labor and taxes.

No way.

I called a number of junk yards in the area and found a totaled 2002 Sienna that still had the canister on it. Bought the canister for $85, replaced it myself following this blog post and this one.

I’m not a mechanic by any means and this was a very doable repair. Just pay attention to what you remove and put it back the same way.

But what do you do when you get a frantic client phone call reporting an issue on the live site? You can’t just enable WP_DEBUG for everybody and very likely break the site’s design as various error messages of sloppy plugin or WordPress’ core code might render.

Solution is simple: Enable WP_DEBUG just for you.

First, you’ll need to know your outgoing IP address. When I’m lazy I go to my friend Topher’s beautiful website which tells me my IP address.

Then edit your wp-config.php (CAREFUL) by replacing:

define('WP_DEBUG', false);

with

if ($_SERVER['REMOTE_ADDR'] == '123.123.123.123') {
	define('WP_DEBUG', true);	
} else {
	define('WP_DEBUG', false);
}

(don’t forget to replace 123.123.123.123 with your outgoing IP address)

By the way this is a great solution if you’re working from home or are tethering using your phone so that you are the only one coming from this IP address. If you’re in the same building with your client and try to debug the site without freaking out the CEO this might not be the best because odds are you and her are coming from the same IP address and she’ll see the same debug junk you do.

Please let me know if this was any helpful to you.

After spending some time on setting things up on my local box I uploaded it to a staging environment on our server and noticed a “mysql_real_escape_string() [function.mysql-real-escape-string]: Access denied for user: ‘nobody@localhost’” error.

If you look up mysql-real-escape-string in the php manual you’ll notice that the second and optional parameter is link_identifier (the db resource). The manual describes it as:

The MySQL connection. If the link identifier is not specified, the last link opened by mysql_connect() is assumed. If no such link is found, it will try to create one as if mysql_connect() was called with no arguments. If no connection is found or established, an E_WARNING level error is generated.

There might be another solution to the problem, for example changing a setting on the server, but if you don’t know how to change it or don’t have permissions to do so, you can modify code of Pods CMS so that anytime it calls mysql_real_escape_string it passes the link_identifier. Here is a patch I created.

It is quite simple. I created a folder in my Bookmarks Toolbar in Firefox and placed the bookmarks I use the most in it, as shown below:

The way to make this URL/blog independent is link to relative page, for example “Add New Post” points to post-new.php (which needs to point to www.andrejciho.com/wp-admin/post-new.php on this blog). The “relativeness” is done via javascript so the “Location” of the bookmark is:

javascript:window.location=’post-new.php’

So you can really roll your own bookmarks following the pattern above. The links in the

javascript:window.location=’post-new.php’
javascript:window.location=’edit.php’
javascript:window.location=’page-new.php’
javascript:window.location=’edit-pages.php’
javascript:window.location=’link-add.php’
javascript:window.location=’link-manager.php’
javascript:window.location=’plugins.php’

Note: These bookmarks only work once you are already logged in and within WordPress admin.

DENIAL.

I drive to a coffee shop because I actually have billable work to get done and would rather do that than waste time on messing with the ISP.

ANGER.

I come home hours later to find that the Internet connection is still spotty. Time to stop billable work and tinker with the issue at hand.  I power down the cable model and the wifi router, disable and enable wireless card,… the whole nine yards – I did work tech support 5+ years, I do know the drill.

Surprise surprise: It’s not me, it’s them.

BARGAINING.

I call Comcast to report the problem. I sit on hold for a while. Phone tech apologizes for the inconvenience and assures me they will fix this free of charge. Ha! I bite my tongue and not mention the cost of lost time and income, cost of gas driving somewhere with reliable internet connection, etc.

The tech looks up the history of connection between their servers and my cable model and confirms that there are issues, sometimes they can’t see the cable model online at all. Tech offers to reset the cable model remotely. This kicks me off the phone since I’m on the Comcast home phone (rather than burning cell minutes). Comcast tech calls back and says that the issue is still occurring.

We schedule a time for a tech to come out.

DEPRESSION.

Tech comes out a couple days later. Surprisingly he knows the exact date/time when the issue started occuring. How? Because that’s when Comcast issued the regular IP refresh signal to my cable model. Because of the tech’s lack of knowledge/experience, I end up training him on how the Internet works behind the scenes (DNS, IP addresses, packets, cable modems, routers). I provide this to Comcast free of charge, every time. Then I walk him step by step through the indisputable proof that my computer is NOT the issue – also providing training on how one can best diagnose whether it is the computer, the website(s) being pinged, or the ISP’s equipment/service.

Tech tries everything and anything to resolve the issue and in the end “adjusts the signal stength something-or-another on the router which should improve the problem over time.” He admits that the model of wifi router issued by Comcast is giving a lot of people trouble but Comcast will not issue a different one to me – I’d have to buy one myself.

After 2 hours of my time, the tech says it shouldn’t be an issue any more and leaves. The internet connection seems to be working OK at this point.

ACCEPTANCE.

I cross my fingers in hopes that the “solution” works for as long as possible before another ice age hits.