First i tried this
<cfmodule template="tags/scopecache.cfm" scope="application" clearall="true" cachename="podName">
in Application.cfm and it helped. BUT as result the pages were loading over 2 seconds each time. and of course it was not acceptableSo, i went from Rob Wilkerson's idea of substituting a part of the cached version of the entire page with the needed part. But the problem is that theoretically it's not known what exactly and with what exactly should be changed (because of the dynamic nature of what should be changed and/or because i wanted to make it generalized enough).
So i had to find or have some kind of markers that will allow uniquily identify the needed parts of the source html. And this is not difficult. This way i came to the following process:
1. Inside the pod script, around the code that will be output on the blog, i place a tag that does not really exists. In order to make it simple, i decided to use as so the name given to the pod. So, the pod's script will look something like this:
<podName>
....
here goes pod code
...
</podName>
2. In Applicatiion.cfm, somewhere at the top but after the CFAPPLICATION tag do a CFINCLUDE of a file with the following code (i placed it right after udf.cfm's cfinclude):
<cfset notCachedPodNames = "podName|podName2|podName3">
<cfset notCachedPodFiles = "pod1.cfm|pod2.cfm|pod3.cfm">
<cfif IsDefined("application.SCOPECACHE") and StructKeyExists(application.SCOPECACHE, application.applicationName)>
<cfloop index="podListPoz" from="1" to="#ListLen(notCachedPodNames, '|')#">
<cfset podElement = ListGetAt(notCachedPodNames, podListPoz, "|")>
<cfset podFileElement = ListGetAt(notCachedPodFiles, podListPoz, "|")>
<cfif FindNoCase("<#podElement#>", application.SCOPECACHE[application.applicationName].value, 1)>
<cfset thisAppVar1 = Left(application.SCOPECACHE[application.applicationName].value, FindNoCase("<#podElement#>", application.SCOPECACHE[application.applicationName].value, 1)-1)>
<cfset application.SCOPECACHE[application.applicationName].value = RemoveChars(application.SCOPECACHE[application.applicationName].value, 1, FindNoCase("</#podElement#>", application.SCOPECACHE[application.applicationName].value, 1)+Len(podElement)+2)>
<cfsavecontent variable="thisAppVar2"><cfinclude template="includes/pods/#podFileElement#"></cfsavecontent>
<cfset application.SCOPECACHE[application.applicationName].value = thisAppVar1 & thisAppVar2 & application.SCOPECACHE[application.applicationName].value>
</cfif>
</cfloop>
</cfif>
So.. what do we have here..
First of all, in the notCachedPodNames variable we set all the pod names for the pods that we don't want to be cached and in notCachedPodFiles the actual names of that pods. Then we do a loop for the number of elements in that lists and do the needed substitution. As can be seen, first the script look for the open tag of the current pod name (for example
<podName1>) and save in thisAppVar1 the first part of the source. Then we delete from the top everything including the close tag for the current pod name (for example </podName1>). Now we save in the thisAppVar2 variable the execution of the current pod. And finally we replace the existing cache for the page html source (application.SCOPECACHE[application.applicationName].value) with the concatenated values of thisAppVar1, thisAppVar2 and the rest of application.SCOPECACHE[application.applicationName].value.All DONE. Now our cache contain the modified version of the html source. And it work FAST. For me, loading the blog with that process inside, rarely goes over 70 milliseconds.
Ok, so it looks like I was wrong - the cache surrounds the entire page. I still think your solution is overkill. After relooking at my own code - I only cache the entire page on the home page. Most people read blog entries - not the home page, so typically I'd say to not worry about your pod being cached.
yeah.. maybe it's not deadly needed, but it was more like an interesting problem for some free time in the week end. kind of challenge :)