Wow, Ray. I thought *I* was OCD when it came to code organization. Yeah...alphabetical? I wouldn't be able to find ANYTHING in my cfc's if I did that. Here's how one of mine looks, collapsed, mind you (and I trimmed it to shorten the line lengths...)

<cfcomponent name="cmsAdmin" hint="Admin CMS relate
<!--- ------------------------------------------
<!--- GENERAL Related functions --->
<!--- ------------------------------------------
<cffunction name="init" returntype="cmsAdmin" ou
<cffunction name="broadcastCMSChange" access="pu
<!--- ------------------------------------------
<!--- CONTENT Related functions --->
<!--- ------------------------------------------
<cffunction name="cmsContentSearchFormParam" ret
<cffunction name="generateCMSContentSearchWhereC
<cffunction name="getCMSContentPagination" acces
<!--- ------------------------------------------
<!--- PAGE (Layout) Related functions --->
<!--- ------------------------------------------
<cffunction name="cmsPageSearchFormParam" return
<cffunction name="generateCMSPageSearchWhereClau
<cffunction name="getCMSPagePagination" access="

[more functions removed to slim it down]

<!--- ------------------------------------------
<!--- ------------------------------------------
<cfinclude template="/com/model/mixinsSimple.cfm
<cfinclude template="/com/model/mixinsCMS.cfm" /
<!--- ------------------------------------------
</cfcomponent>

You're just plain nuts, dude... :D

If I'm going to participate in something this pointless it better be because there's an iPhone on the line or something. hehe

I concur with Jared... hehe

insane man, but I would like to see the finalized code...
;-)

so you want code, to open code up, snip and tuck that code, reordering the methods by alphabet, then save it, and have a cfc that basically cleans up your mis-ordered methods?

woah.

Should be easy actually with reMatch. :)

I am sure I can make it super complicated somehow... it's that part of the fun :D

Oh - if you want complex - I can make it complex.

The function should take, as an optional second parameter, ANOTHER udf that will work as the sort function. The sort function takes 2 items and returns -1 if the item A should be sorted before, 1 if after, and 0 if the same.

In other words - allow for custom sorts.

If you want to cheat - there is a version of this up on CFLib already.

What is passed to the sorting function? The name of the CFFunction?

That would be telling. ;)

Look to the lib....

Do it "quick"ly...

Well, luckily, that date I had with this super model just fell through, so I should have some time to take a look at this over the weekend :) ... at least that sounds like a better excuse.

Guys... just use CFEclipse and forget about it. ;-)

Heh, I _do_ use CFEclipse. Thing is - I'm anal about keeping my code view big - and I'm too lazy to turn the view back on. ;)

Oh, this is going to be fun ;~)

Like Will B, I prefer to keep semi-related functions close together (or ones that call each other, as all the functions in a class should be related).

I'll just open up the side bar in cfeclipse and sort them alphabetically there if I need to find by name (or hit ctrl-f and search for it that way).

This place ROCKS!!!

like running a 4 d Array of these topics Yea, I would love to see the finish; good luck on this Ray!

Here's my crack at it. No custom sorting, but sorts alphabetically and puts any init functions first.

<cffunction name="sortMethods" output="false" returntype="void">
<cfargument name="filePath" type="string" required="true">
<cfset var funcRegex = "<cffunction [^>]*>.*?<\/cffunction>">
<cfset var fileContent = "">
<cfset var allFuncs = "">
<cfset var nameMatch = "">
<cfset var qFunctions = queryNew("name,theFunction,sortOrder")>
<cfset var i = 0>

<!--- Verify the file exists --->
<cfif fileExists(arguments.filePath)>
<!--- Read file --->
<cffile action="read" file="#arguments.filePath#" variable="fileContent">

<!--- Find all functions --->
<cfset allFuncs = reMatch(funcRegex, fileContent)>

<cfloop from="1" to="#arrayLen(allFuncs)#" index="i">
<cfset queryAddRow(qFunctions)>

<!--- Store function name and contents in query --->
<cfset nameMatch = reMatch("name=[""'][^""']*[""']", allFuncs[i])>
<cfset qFunctions.name[qFunctions.recordCount] = replace(nameMatch[1], "'", '#chr(34)#', "ALL")>
<cfset qFunctions.theFunction[qFunctions.recordCount] = allFuncs[i]>

<!--- If the function name is "init", sort it first --->
<cfif qFunctions.name[qFunctions.recordCount] IS NOT 'name="init"'>
<cfset qFunctions.sortOrder[qFunctions.recordCount] = 2>
<cfelse>
<cfset qFunctions.sortOrder[qFunctions.recordCount] = 1>
</cfif>
</cfloop>

<!--- Sort the functions by name (with init on top) --->
<cfquery name="qFunctions" dbtype="query">
SELECT name, theFunction
FROM qFunctions
ORDER BY sortOrder, name
</cfquery>

<!--- Prepare file to be written --->
<cfset fileContent = reReplace(fileContent, "\n*" & funcRegex & "\n*", "", "ALL")>
<cfset fileContent = reReplace(fileContent, "<\/cfcomponent>( |\n)*", "", "ALL")>

<cfsavecontent variable="fileContent"><cfoutput>#left(fileContent, len(fileContent)-14)#</cfoutput>
<cfoutput query="qFunctions">
#theFunction#
</cfoutput>
<cfoutput>#chr(60)#/cfcomponent#chr(62)#</cfoutput>
</cfsavecontent>

<!--- Write the file back to the source --->
<cffile action="write"
file="#arguments.filePath#"
output="#fileContent#">
</cfif>

</cffunction>

Tyson, thanks for sharing the code. Let me make one style comment. I don't like code where 99% of the functionality is all within one CFIF. Your code does nothing if the cfif condition is false, so I would have put this up top instead:

<cfif not fileExists(arguments.filepath)><cfreturn></cfif>

I didn't really like the idea of having the code in one CFIF, but that was the first idea that came to me, so I just went with it. I'll be interested to see how others implement the custom sort function. Also, this was the first time I'd used reMatch and I see myself using a lot from now on.

Unfortunately this is much more complicated than you think it is.

Tyson's solution works, right until it runs into a cffunction with the char > anywhere in an attribute. It'll also choke if the function contains the text </cffunction> anywhere, for instance <cfset myvar = "<cffunction>">.

If you really want to blow your mind think about:

<
cffunction
name
=
"foo"
hint =
"<cffunction name='bar'></cffunction>"
>
<
/
cffunction
>

That's syntactically valid CFML.

What you want can really only be reliably done with a full CFML parser. Implementing it in CFEclipse would probably be the easiest solution because the thing already has a parser, but I'm curious if CFEclipse even parses like the Macromedia/Adobe ColdFusion parser does.

Regular expressions are nifty, but alone they're not really useful for parsing. There's no finite length regular expression that expresses what a cffunction declaration could look like.

Good point Elliot. Although I'd still say that I'd find this code useful. I think I'd know ahead of time if there was a </cffunction> anywhere in the code (I mean as a string). I'd assume that would be rare (except in the case of the code for this blog entry ;)

Yeah, true. I didn't have time to look at this this weekend (unexpected events). However, I cannot imagine that:

<
cffunction

is considered valid? I am pretty sure that need to have the < and the cf right next to each other.

Also, yeah, its true - people really put some crack-head information into the attributes. Just the other day I saw someone do like a whole documentation schema inside of the Hint attribute.

@Ben

Nope, you don't need the < and the cf next to each other. Try the code snippet I pasted. ;)

CF has a very very lenient syntax. You can add new lines and whitespace between all kinds of things and it won't even blink.

I started working on breaking CF's parsing rules down into tokens for documentation purposes, I should really post that somewhere. I have a feeling that many of the CF engines don't really follow the same rules as the ColdFusion Server, but rather the conventional CFML syntax. I definitely know that Smith's parser is wrong/too strict in a few places.

Oh my god! You are totally right. I just ran the craziest code and it ran just fine. That is just insane. Not sure how I feel about that :) I think it's a little frustrating that they allow syntax like:

<
cfset
x = 4
/>

... and at the same time, they do NOT allow you do this:

<cfset GetFunctionPointer()() />

... or :

<cfset GetArray()[ 4 ] />

Those seem like infinitely more important chainnings and yet the will not parse propertly :( Silly ColdFusion parser.