Built in debug mode

benkillin

This isn't really a critique thred, but I didn't know where to put it.

When I write code I always build in debug modes into the code to make development time decrease. The way I do it is very simple:

define('DEBUG', true); // if this var is true, numerous debug statements will be printed out in the 
						   // course of the execution of this script
function dbg($string="DEBUG STATEMENT")
{/*******************************************\
  * This function prints a string $string if 
  * the constant DEBUG is set to true
 \*******************************************/
	if(DEBUG)
	{
		print("<div id=\"".rand(0,2147483648)."\"><p>" . $string . "</p></div>\r\n");
	}
}

Every time I query a databse, or do some loops, I put a dbg call afterwards to know exactly when my code breaks, and whenyou are done developing you can just set one variable to false and all the debug statements go away.

You could go without the div and paragraph tag, but I put it in there to be compliant with my site's requirements.

I was also wondering if anybody else did something like this or if I was just wierd.

bubblenut

No, you're not wierd. In fact, I think people who don't have some kind of debugging output are wierd 😉

Something which can be very usefull for debugging (and logging etc.) is AOP (aspect oriented programming). I recently found an implementation in PHP, I haven't checked out the source code yet but I've looked through the examples and it seems to have what you need. Check it out here

The great thing about AOP is that you don't have to place the debugging/logging in your code, you just attach it to the functions/methods you want it to affect. This means less typing for you, and also the ability to not only turn it off but remove it's presence all together. Pretty neat huh?

benkillin

Woulden't that create alot more overhead when declaring a class, instantiating it, accessing its features, and the such instead of just a funtion?

bubblenut

Well, it depends on how many layers you're building into it. It's one wrapping function per aspect per function. But then pretty much all high level programming constructs are less effecient than their lower level counterparts, what you get in the tradeoff is convenience. You can drop in/out the code as and when you please. Also, remember that debugging is not a day to day occurrence (or at least it shouldn't be) and it happens in a controlled environment, so, in this case, on the whole, it's more effecient because when you take the debugging out it's out of the code alltogether. ie it doesn't even get parsed.