precedent over any other scope (since they were introduced in CF5).
Because they have no prefix, this MUST be so. Otherwise you couldn't
access them in a loop over a query. Best practices is to always scope
your vars. You will never run into trouble!
For queries, you don't have to do it longhand if you are in a <cfloop
query=""> or <cfoutput query="">. Just write #queryname.columnname#.
This has always worked. It is just if you want to use array notation
that you need the currentrow:
<cfset col = "columnname">
<cfoutput>#queryname[col][currentRow]#</cfoutput>
And I think that for most developers, the output of your test is just as
it should be. I think you should never rely on the order of evaluation
for unscoped vars. ALWAYS scope or you will run into trouble sometime.
Pascal Peters
Certified ColdFusion MX Advanced Developer
Macromedia Certified Instructor
LR Technologies
Av. E. De Mot, 19
1000 BRUSSELS, BELGIUM
Tel: +32 2 639 68 70
Fax: +32 2 639 68 99
Email: [EMAIL PROTECTED]
Web: www.lrt.be
> -----Original Message-----
> From: Barney Boisvert [mailto:[EMAIL PROTECTED]
> Sent: donderdag 1 april 2004 23:50
> To: CF-Talk
> Subject: More local variables scoping weirdness
>
> So I happened to loop over a recordset in a function today
> and one of the column names had the same name as an argument
> to the function. I thought nothing of it, and assumed that
> inside my CFLOOP tag, the column's value in the current
> record would have the highest priority in the scope search
> queue. But that's NOT the case.
>
> Local variables have a higher priority that columns when
> looping a recordset. Thus, if you want to reference a
> recordset column within a function that has either an
> argument or a local variable with the same name, you have to
> do it longhand: #recordsetname.columnname[currentrow]#,
> rather that the usual #columnname#
>
> Anyone have any clever insights about this, and/or know if it
> is/should be a bug? I think it should, but we all know my
> opinions on scoping are "interesting".
>
> To prove to myself I wasn't crazy, I wrote a simple test case
> and included it below. While it will happen rarely (this is
> the first time it's happened to me), it seems
> counterintuitive that the local variables scope would be
> searched before the recordset, since the recordset loop is a
> narrower scope than the local variables scope. Both CFC
> methods loop the recordset and create a list of the 'id'
> field's values for the records. The output should, IMHO, be
> "1,2,3" for all three, but it is in fact "42,42,42",
> "37,37,37", and "1,2,3".
>
> test.cfm:
> ---------
> <cfset q = queryNew("id") />
> <cfset queryAddRow(q) /><cfset querySetCell(q, "id", 1) />
> <cfset queryAddRow(q) /><cfset querySetCell(q, "id", 2) />
> <cfset queryAddRow(q) /><cfset querySetCell(q, "id", 3) />
> <cfoutput> #createObject("component", "test").testArg(q,
> 42)#<br /> #createObject("component",
> "test").testLocal(q)#<br /> #createObject("component",
> "test").testFullScope(q)# </cfoutput>
>
> test.cfc:
> ---------
> <cfcomponent>
>
> <cffunction name="testArg" output="true">
> <cfargument name="rs" />
> <cfargument name="id" />
> <cfset var result = "" />
> <cfloop query="rs">
> <cfset result = listAppend(result, id) />
> </cfloop>
> <cfreturn result />
> </cffunction>
>
> <cffunction name="testLocal" output="true">
> <cfargument name="rs" />
> <cfset var id = 37 />
> <cfset var result = "" />
> <cfloop query="rs">
> <cfset result = listAppend(result, id) />
> </cfloop>
> <cfreturn result />
> </cffunction>
>
> <cffunction name="testFullScope" output="true">
> <cfargument name="rs" />
> <cfset var id = 37 />
> <cfset var result = "" />
> <cfloop query="rs">
> <cfset result = listAppend(result, rs.id[currentRow]) />
> </cfloop>
> <cfreturn result />
> </cffunction>
>
> </cfcomponent>
> ---
> Barney Boisvert, Senior Development Engineer AudienceCentral
> [EMAIL PROTECTED] voice : 360.756.8080 x32
> fax : 360.647.5351
>
> www.audiencecentral.com
>
>
>
[Todays Threads] [This Message] [Subscription] [Fast Unsubscribe] [User Settings]
