Todd Sharp recently pinged me about an interesting problem. He was working with Images generated in ColdFusion 8 and was trying to use this imaga via the HTML IMG tag, like so:

<img src="foo.cfm">

The code in foo.cfm generated the graphic and he couldn't quite figure out how to display it. One thing he tried was the WriteToBrowser action:

<cfimage action="writeToBrowser" ...>

This didn't work - and if you view source on a file using this action you will see why. The WriteToBrowser action actually generates HTML that points to a 'magic' URL to serve the image.

He also tried cfcontent:

<cfimage action="read" source="paris.jpg" name="paris"> <cfcontent type="image/jpg" variable="#paris#">

This also failed as well. Why? The cfcontent tag doesn't like the Image variable. ColdFusion will throw the following error:

coldfusion.image.Image is not a supported variable type. The variable is expected to contain binary data.

If you think about it - this makes sense. Image variables aren't binary data in ColdFusion. While they may store binary data, ColdFusion wraps up the data in a new "Image" variable type. Now personally I think cfcontent should be smart enough to recognize a image variable and deal with it - but what do you in the meantime?

ColdFusion provides a ImageGetBlob function. This returns the binary data of the image and can be safely used with cfcontent like so:

<cfimage action="read" source="paris.jpg" name="paris"> <cfcontent type="image/jpg" variable="#imageGetBlob(paris)#">

But wait - there is a catch! This works ok as long as you begin your image with a real image. Notice above I have a image file I'm beginning with. If I use a 100% virtual image then it doesn't work. Consider:

<cfset canvas = imageNew("", 100, 100, "rgb", "white")>

If you run imageGetBlob on this, you get:

The source file should contain an extension,so that ColdFusion can determine the image format.

So I'm not 100% sure - but this is how I read this. I had created an image of a certain type of image type. But there was no file type for the image. Obviously I could write the variable out to a file, but there is no way to go directly from an "pure" Image variable to a blob. I'm going to file an enhancement request to add support for this. In the meantime if you needed to do something like this, I'd recommend creating a "canvas" graphic in your desired format and seed your dynamic image with that.

Ok - so now for the contest part. I whipped up a quick demo to show the code described above in action. I call it the Paris-Talkometer. Let me show you the code and then I'll link to the application.

<cfparam name="url.caption" default="">

<cfajaxproxy bind="javascript:updImage({caption})">

<script> function updImage(str) { if(str != '') document.getElementById('myimg').src = 'imgs3.cfm?caption='+escape(str); } </script>

<h2>Paris-Talkometer</h2>

<cfform name="form"> <cfinput type="text" name="caption" value="#url.caption#" > <cfinput name="mybutton" type="button" value="Update"><br /> <cfinput type="image" name="myimage" src="imgs3.cfm?caption=#url.caption#" id="myimg"> </cfform>

What we have is is a basic form. There is a text box for a caption, and a cfinput image type that points to imgs3.cfm. Notice though that it passes a URL variable. Go back up to the top of the file and see how I use cfajaxproxy to bind to the caption. Whenever the caption changes, I use a bit of JavaScript to the change the SRC of the image. Here is the code behind imgs3.cfm:

<cfparam name="url.caption" default="">

<cfimage action="read" source="paris.jpg" name="paris">

<cfif len(trim(url.caption))>

<cfset imageSetDrawingColor(paris, "black")> <cfset imageDrawRect(paris, 1, 1, paris.width, 40, true)>

<cfset imageSetAntialiasing(paris, "on")> <cfset imageSetDrawingColor(paris, "white")> <cfset text = { style='bold', size=12, font='verdana' }> <cfset imageDrawText(paris, url.caption, 10, 25, text)>

</cfif>

<cfcontent type="image/jpg" variable="#imageGetBlob(paris)#">

All I'm really doing here is looking for a URL.caption value. If it exists, I draw some text over the picture. The last thing I do is serve up the image using cfcontent and imageGetBlob.

You can see this in action here:
http://www.coldfusionjedi.com/demos/imagecfcontent/test3.cfm

Now for the contest. If you look at the code for test3.cfm, you will notice that you can seed the caption value via the URL. Here is an example.

Your content is to make Paris intelligent - or at least witty. While this is a somewhat herculean task - I'm sure some of you can do it. Just post your URL as a comment (and of course, you can comment on the main part of this blog post as well). Please keep your captions work safe. Work safe doesn't mean boring - just keep it safe please.