The sad part about this little tidbit is that it would have saved me and the team I’m on a ton of time. You see, we work quite a bit with secure web sites using https, SSL protocol. But it’s a fickle thing, because web browsers hate it if anything on the page is linked without SSL if you’re already in SSL. In our development environments, we don’t have certificates installed, so sometimes we forget to put that elusive little “s” at the end of the http. What happens then for those who don’t know is your customer’s browser is experience on your site is abruptly halted to prompt you that there’s unsecured content on the page! GASP!

Needless to say that’s mud in your face when you present your awesome secured site to your customers. What ensues then is a mad search and replace for that lone link sitting in some obscure corner of your html that has that hard coded “http”!

The epiphany is you can use what’s called a Protocol-Relative Hyperlink, so you can just create a image link for example like this:

<img src="//yoursite.com/yourimage.gif" />

instead of coding in the full URL:

<img src="https://yoursite.com/yourimage.gif" />

So the // in front acts just like a relative link for your protocol, picking up whatever you were at, SSL or not! It’s so simple it’s hard to believe I missed it, heck, that’s probably why we missed it! Unreal…

So, now I know, and I’m writing it in here so I never forget, and who knows, maybe it’ll help someone else too.

It takes a URL as input and does a pre-flight check to ensure that the endpoint is available before launching the user to the destination.

Most web developers probably never need to take this extra step, since it’s another layer in the link chain, but this becomes important if you are dealing with a large multi-environment site or external links that are a critical portion of your web site. The trend (at least in the company I work for) seems to be a move move toward outsourced web services, where whole portions of your website functionality is literally outsourced to a vendor. The outsource component is more complex than it seems, because this isn’t just sending the user blindly over to the vendor; we must perform authentication in some fashion, a single sign on, if you will, to allow for a more seamless experience. In the case of a large internal multi-environment web site, you have multiple pieces, which are expected to behave as a single entity, so any failure in one part ruins the cohesive single experience.

The code below also provides a simple mechanism to disable parts of my web site, based upon trigger files which I can set to determine if I should even send the user over, if the destination is available.

Wait – did you say TRIGGER FILES? Yes I did.

There are those that will probably argue with me how poor this concept is and how mundane, but honestly, it works, and it’s pretty quick to put together.

I had about little over an hour to whip the code up to do this, or suffer from manually fudging the site to take links down during an upcoming maintenance window (I won’t go into how we got to this state.. That’s for another story). Of course the more PRO way of doing this would have been to either create a cool Coldfusion application that read a database table or read an XML file to get all sorts of nifty data to show and a cool kick ass admin interface to handle it all! But honestly, I’m not that good at Coldfusion, and I don’t have a guru handy at the moment to help me figure out the CFML to parse XML, nor a DBA available at my beckon call to whip up an awesome Schema for me. So, in the time I’m given, trigger files were easy.

Hopefully someone might find this code handy, or at least parts of it, since I’ve not found anything this specific online. I paraphrased some of the details, like the exact messages, but you should get the idea. Enjoy

<!--- dest is the parameter passed in --->
<cfif isDefined('url.dest')>
	<cfhttp url='#url.dest#' method="head" resolveurl="no" throwonerror="no" />
	<cfif NOT IsDefined("cfhttp.responseheader.status_code") OR cfhttp.responseheader.status_code EQ "404">
			I just checked the header response for a 404.
<pre style="font: normal normal normal 12px/18px Consolas, Monaco, 'Courier New', Courier, monospace;">			Sorry the destination is down</pre>
<cfelse>
 <!--- ok the destination is there, can I go to it? --->
 <!--- here, I check to see if I'm calling this proxy script from my member section --->
 <cfif #CGI.HTTP_REFERER# contains "member">
 <cfif fileexists(expandpath("/outage/member.html"))>
 <cfset outage = 1>
 <cfelse>
 <cfset outage = 0>
 </cfif>
 <!--- I do this for other sections of my site --->
 <cfelse>
 <cfset outage = 0>
 </cfif>
 <!--- Now I check to see if there was an outage, if not, we send the user to the destination --->
 <cfif outage eq 1>
 <p>We're sorry, but due to maintenance, this part of the site is down. </p>
 <cfelse>
 <cflocation url='#url.dest#' />
 </cfif>
 </cfif>
<cfelse>
 <!--- Do Nothing, since no Parameter came in --->
</cfif>

I got back from Voices That Matter, and I was so inspired by Erica Sudan’s Dell Mini 9 Hacintosh project I had to try it. I’m happy to say that I’m a proud owner of a HP Mini 110 Mac! I’m going to write about my experiences doing it, and using it soon! I’ve also made my Dell 17″ E1705 laptop into a nice 17″ Macbook – Bwahahaha!

I’ve also been checking out Windows 7, now that I have it. I’ve installed it a couple of times now, and the short of it is… I like it! It is so much better than Vista… If Microsoft brought this out as Vista, Apple wouldn’t have so much trash to talk about. I’ve not delved too deep into it, but from what I’ve experienced so far from installation to initial usage it’s been awesome. Kudos to Microsoft!

In gaming, my buddy James got me Borderlands… Damn him! I’ve been playing that more than I should, and in reality, not as much as I want to… Damn real life job!

So, between those three big things, and the impending launch of Motorola’s Droid on Verizon tomorrow, I’ve got no shortage of things to type about… I just need more hours in the day…

Once dinner was done, I headed to my gate. To my dismay, the airport had barely any outlets to plug anything in, and my laptop was down to 30% and my iPhone wasn’t far behind, what outlets I did find were far from the gate, or otherwise occupied. I noticed alot of people sitting around, and my phone kept getting notifications from Orbitz about my flight being delayed. It didn’t take long for the cancellation to come. I really wanted to head home… But the weather had turned bad; ice formed on the planes, and the earlier Rochester flight was still at the gate awaiting departure. The next direct flight would be tomorrow morning at 1:30pm – Ouch!

This is the first time I’ve been stranded (it always happens to those other people on TV!). But I remained calm, since I’m of the philosophy that getting angry will probably not solve anything, and just aggravate the situation. I asked at the counter nicely about what my overnight options were, and the woman behind the counter indicated that no-one is allowed to remain at the terminals when the airport closes, I would need to wait in the area outside security (which is REALLY sparse). But she said that the lady at the info booth near baggage claim could help me out with discounted lodging at hotels near the airport (see it pays to be nice). I went down to see this woman who pointed me to the hotel phones where I could enter in a special code which got me to someone right away. The gentleman on the phone gave me rates for 2 hotels close by. The hotel at the airport was still expensive, even with the discount, but the Wyndham Chelsea, about 2 miles from the airport, had a reasonable rate of $109, and had a free shuttle.

So, I’m here now at my room in the Wyndham Chelsea (which is quite nice), finishing off my weekend blog and taking advantage of the extra quiet time.

Time for bed, and hopefully my flight won’t be further delayed tomorrow.