Skip to main content
Ben Nadel at NCDevCon 2016 (Raleigh, NC) with: Dan Wilson
Ben Nadel at NCDevCon 2016 (Raleigh, NC) with: Dan Wilson

Need Help Making A Viable ExpandServerPath() ColdFusion UDF

By
Published in Comments (17)

ColdFusion provides a built-in function, ExpandPath(), which gives you a fully expanded server path based on the currently executing script. The problem is, a lot of our application templates are NOT on the same level as the currently executing script. Furthermore, if you are using some sort of URL re-writing, the real executing script and the web-visible executing script are not the same values.

What we need is a new function: ExpandServerPath(). This function would act in exactly the same way, only the passed in path (function argument) would be relative to the calling template and would have nothing to do with the executing script.

Back in November, I talked about using a CFTry / CFCatch block with a manually thrown exception to find out the calling template of the given function. I have taken that methodology and wrapped it up in the new ExpandServerPath() ColdFusion user defined function:

<cffunction
	name="ExpandServerPath"
	access="public"
	returntype="string"
	output="false"
	hint="Expands the given path based on the callind template, not the executing script name.">

	<!--- Define arguments. --->
	<cfargument
		name="Path"
		type="string"
		required="true"
		hint="The path we are expanded."
		/>

	<!--- Define the local scope. --->
	<cfset var LOCAL = StructNew() />


	<!---
		Throw an error so that we can get the stack
		trace from the manually caught exception object.
	--->
	<cftry>
		<cfthrow
			type="ManualException"
			message="This is a planned exception."
			/>

		<!--- Catch the exception. --->
		<cfcatch>

			<!---
				Store the tag context array from the
				exception object. Since the first tag context
				item will be this template (the one containing
				the UDF), we will need to grab the second one.
			--->
			<cfset LOCAL.Context = CFCATCH.TagContext[ 2 ] />

		</cfcatch>
	</cftry>


	<!---
		Now that we have our tag context item, we can
		grab the directory of the template on which this
		method was called.
	--->
	<cfset LOCAL.Template = GetDirectoryFromPath(
		LOCAL.Context.Template
		) />


	<!--- Get the proper slash for this system. --->
	<cfset LOCAL.Slash = Right(
		LOCAL.Template,
		1
		) />


	<!---
		Clean the path that was passed into the document.
		We want to switch all slashes to be the proper slash
		for the expanded path.
	--->
	<cfset ARGUMENTS.Path = ARGUMENTS.Path.ReplaceAll(
		"[\\\/]",
		"\#LOCAL.Slash#"
		) />

	<!---
		Now that we have the proper directory, we are
		going to need to apply the standard path actions
		(such as ./ and ../). In order to do so, we are
		going to loop over the path to expand as if it
		were a list and apply each element.
	--->
	<cfloop
		index="LOCAL.PathElement"
		list="#ARGUMENTS.Path#"
		delimiters="#LOCAL.Slash#">

		<!--- Check to see the current path element type. --->
		<cfif (LOCAL.PathElement EQ ".")>

			<!---
				This is a "same directory" construct. These
				don't really need to be applied and can just
				be dropped from the path.
			--->

		<cfelseif (LOCAL.PathElement EQ "..")>

			<!---
				This is a "move up one directory" directive.
				For this, we need to remove the final slash
				and directory name from the exsisting template
				path.
			--->
			<cfset LOCAL.Template = LOCAL.Template.ReplaceFirst(
				"[^\\\/]+[\\\/]$",
				""
				) />

		<cfelse>

			<!---
				We are dealing with an actual path element. Add
				it to the existing path followed by a slash.
			--->
			<cfset LOCAL.Template = (
				LOCAL.Template &
				LOCAL.PathElement &
				LOCAL.Slash
				) />

		</cfif>

	</cfloop>


	<!---
		At this point, we have an expanded path that has a
		trailing slash. However, we only want the trailing
		slash if the path passed into this method had a
		trailing slash. Other wise, we just want to leave
		it as it came in (assuming that it was a file).
	--->
	<cfif (Right( ARGUMENTS.Path, 1 ) NEQ LOCAL.Slash)>

		<!--- Remove last slash. --->
		<cfset LOCAL.Template = LOCAL.Template.ReplaceFirst(
			"[\\\/]$",
			""
			) />

	</cfif>


	<!--- Return the expanded path. --->
	<cfreturn LOCAL.Template />
</cffunction>

This function would then just be called like ExpandPath():

#ExpandServerPath( "./data/2007_07_02.dat" )#

This would be looking for a DAT file within the directory "data" which is a sub directory of the directory containing the currently executing TEMPLATE (not the root script).

It works, so what's the problem? The problem is, it's not a good methodology. It's one thing to have something like this in a development environment where speed / processing drain is not that critical. But, putting something like this up in a production server is a huge No No. Creating exceptions is a very processing heavy activity. This is certainly not something that we want to be doing all the time.

Does anyone have any suggestions on where to improve this? I tried using the ColdFusion debugging information, but information about a processing template only gets put into it once the template has finished processing (which would not work since we need information about the currently processing template).

Help??

Want to use code from this post? Check out the license.

Reader Comments

10 Comments

Hey Ben! It was nice meeting you at the CFUNITED conference. Again, thanks for all of the posts.

I'm wondering why you wouldn't use GetBaseTemplatePath(). Doesn't that return the same information that grabbing the TagContext provides? Or am I missing something...

15,848 Comments

@Toby,

Really good to meet you to. I don't know why, but when you were talking, you sound exactly like Rob Lowe to me. I don't know if anyone has ever said that or if I have just watched Tommy Boy like a couple hundred too many times :)

As far as GetBaseTemplatePath(), you can actually use something similar to get a relative path, except with GetCurrentTemplatePath(). So, for the example above:

GetDirectoryFromPath( GetCurrentTemplatePath() ) & "data/2007_07_02.dat"

... would give us the same thing. But, to me,

ExpandServerPath( "./data/2007_07_02.dat" )

... just seems more elegant. Of course, I tend to do things that people later will tell me are bad practice, so it's possible that there is no need for this type of a template.

Additionally, my UDF will integrate the path constructs like "./" and "../" right into the returned path (rather than just appending it).

Of course, I might be totally missing something, so if you have an example to look at that you think might help, I would as always, be very greatful.

15,848 Comments

But, doesn't GetBaseTemplatePath() only return the top-level ColdFusion page. This won't work for any CFIncluded pages that need to get a relative path?

198 Comments

@Ben:

It's never good to throw an error when you can avoid it. (Throw/catching errors is slow.) Instead, use the code you posted as your starting point:

GetDirectoryFromPath( GetCurrentTemplatePath() ) & arguments.Path

You can then always clean up the path if you want so that the directory pathing ("." and "..") is cleaned up.

15,848 Comments

@Dan,

Certainly, I do not want to throw and catch errors on the production site. This is a new function, I don't actually use this anywhere. But, I like the idea of it. I wouldn't use it, as I know it is slow. I am posting it up just so people can see what I want to try and do and then offer suggestions. Of course, it might just not be possible.

My current methodology is using the GetCurrentTemplatePath() and all that jazz, but it just never feels elegant.

15,848 Comments

@Dan,

Because the GetCurrentTemplatePath() will give me the path of the template in which the UDF is defined. I need to go up one level to get to the template that invoked the UDF. That's why when I grab out of the Exception tag context, I grab index 2.

39 Comments

I can't think of any other way to get the calling template. It would be interesting to see *how* slow your UDF is; is it really that bad performance-wise?

15,848 Comments

@Aaron,

Good point. I have never tested this for speed. I have only been told that it is fast. I assume that it is nothing that would scale, but I'd be willing to run a loop :)

26 Comments

Ben,

I'm not sure I understand your desired function completely... you want the full path to any file where you give the relative path to the calling template? Well, that's exactly what ExpandPath() does...

ExpandPath( "./data/2007_07_02.dat" ) will give you the full path, so what's the difference to what you need?

Chris

15,848 Comments

@Chris,

ExpandPath() goes relative to the base template. For instance, imagine we had this code in the base template:

<!--- Called from top level. --->
#ExpandPath( "./" )#
#ExpandServerPath( "./" )#

<!--- Include sub directory template. --->
<cfinclude template="expand_server_path/test.cfm" />

... And then we had this in the include:

<!--- Called from sub-directory level. --->
#ExpandPath( "./" )#
#ExpandServerPath( "./" )#

... This would give us the following output:

D:\....\site_v1\testing\
D:\....\site_v1\testing\

D:\....\site_v1\testing\
D:\....\site_v1\testing\expand_server_path\

Notice that at the top level template, both in fact do the same thing. However, once you go into an include, ExpandPath() is still relative to the base template, where as ExpandServerPath() is relative to the currently executing template.

Does that clear up the difference in functionality?

15,848 Comments

@Pete,

Thanks. If for no other reason, it's always good for people to see more recursive function examples so they can wrap their heads around it.

1 Comments

If I know that the path to my root is 5 slashes into the hard drive, I can use this:

<cfloop index="i" from="5" to="#ListLen(GetBaseTemplatePath(), Chr(92))#">../</cfloop>

to get the proper relationship.

I believe in love. I believe in compassion. I believe in human rights. I believe that we can afford to give more of these gifts to the world around us because it costs us nothing to be decent and kind and understanding. And, I want you to know that when you land on this site, you are accepted for who you are, no matter how you identify, what truths you live, or whatever kind of goofy shit makes you feel alive! Rock on with your bad self!
Ben Nadel