One example that I have is when I have a free text field, (ie: in a form
builder) and the person building the form is allowed to use coldfusion
variables to pull out values at run time. I parse he string for any #
symbols and then build an appropiate string to evaluate. I find evaluate
works very well in this situation as I have now idea what will be in the
text field.
I do agree, however that it should not be used unless alternatives have been
considered. I also think that there is not a noticeable performance hit
when using it in cf7.
For those that are interested my 2 functions which will parse any text are
below. The one to call is evaluateString("your text with possible cf
variables in it").
Just in case someone asks , I looked at tokensing the string but found the
method I use to be fast and also made it easier to deal with complex
variables.
//hint=Evaluate a string looking for any variable names then process and
return completed string
// Example 1 "The value for email is #headEmail#"
// Example 2 "The http referer is #http_referer# or #cgi.http_referer#
// NB: I have decided to use the find function with mid
// The other options would have been to tokenise the string or to
strip the original
// string down as I am processing to make it shorter an remove the mid
function
function evaluateString(str)
{
var newStr = "";
var finished = false;
var itterations = 0;
var marker = 1;
var spos = 1;
var epos=1;
var strLen = len(str);
var evaled_var = "";
var midStr = "";
var startBPos = 0; // End bracket position
var endBPos = 0; // End bracket position
var startVar = "";
var endVar = "";
var variable = "";
var tempTxt = "";
//writeOutput("<br>evaluateString(#htmlEditFormat(str)#) of length
#len(str)#<br>");
if (find("##", str, 0) gt 0)
{
do
{
spos = 0;
if (marker lte strLen)
spos = find("##", str, marker);
if (spos eq 0 or spos eq strLen)
{
// Then we have hit the end of any variables. Finish string
//writeOutput("Hit end of string<br>");
finished = true;
if (marker lte strLen)
newStr = newStr & mid(str, marker, strLen - marker + 1);
}
else
{
midStr = mid(str, marker, spos - marker);
//writeOutput("midStr = [#htmlEditFormat(midStr)#]<br>");
newStr = newStr & midStr;
marker = spos; // Move marker along
// Look for 2 pounds together
if (mid(str, marker +1 , 1) eq "##")
{
// Then we found a double pound
//writeOutput("Found [####]<br>");
newStr = newStr & "##";
marker = marker + 2;
}
else
{
// Look for the closing pound
epos = find("##", str, marker + 1);
if (epos gt 0)
{
variable = mid(str, spos + 1, epos - spos -1);
//writeOutput("variable =
[#htmlEditFormat(variable)#]<br>");
if (find("(",variable) gt 0)
{
// Only need to make recursive call if there is
an array
// cause of the limitations of cfmx
if (find("[", variable, 1))
{
// Pull out what is between the ()
startBPos = find("(", variable, 1);
endBPos = find(")", variable, 1);
tempTxt = mid(variable, startBPos + 1,
endBPos - startBPos - 1); // No braces please
// Processes the middle text
tempTxt = evaluateString("###tempTxt###");
tempTxt = replace(tempTxt, """", """""",
"All");
startVar = mid(variable, 1, startBPos);
endVar = mid(variable, endBPos, strLen);
//writeOutput("Want to evaluate
#htmlEditFormat("[#startVar#][#tempTxt#][#endVar#]")#<br>");
tempTxt = "#startVar#""#tempTxt#""#endVar#";
evaled_var = evaluate("#tempTxt#");
}
else
evaled_var = evaluate(variable);
}
else if (find("[",variable) gt 0)
evaled_var = findStructArrayValue("#variable#");
else if (isDefined("caller.#variable#"))
evaled_var = evaluate("caller.#variable#");
else if (isDefined(variable))
evaled_var = evaluate(variable);
else if (isDefined("session.eForms.values.#variable#
"))
evaled_var = evaluate("
session.eForms.values.#variable#");
else
evaled_var = "{null}";
newStr = newStr & evaled_var;
marker = epos + 1;
evaled_var = "";
// If at the end if the string then finish
if (epos eq strLen)
finished = true;
}
}
}
itterations = itterations + 1;
//writeOutput("itteration [#itterations#] marker=[#marker#]
spos=[#spos#] epos=[#epos#] finished=[#finished#] :
[#htmlEditFormat(newStr)#]<br>");
}
while (finished eq false and itterations lte 1000);
}
else
newStr = str;
//writeOutput("return [#htmlEditFormat(newStr)#]<br>");
return newStr;
}
//hint=If given a string that is a pointer to a structure with an array in
it then find the associated value
// Example 1 session.eForms.keys.pages[7].trigger.prePage.email.to
function findStructArrayValue(str)
{
var sStr = ""; // Start string
var eStr = ""; // End string
var startBPos = 0; // End bracket position
var endBPos = 0; // End bracket position
var newStr = "";
var element = 0;
var x = 1;
var lLen = listLen(str, "."); // Length of the struct element list
var finished = false;
var itterations = 0;
var curStruct = "";
var arrayPos = 1; // The array position for scalar variable when found
var arrayTxt = ""; // The stuff in the brackets
//writeOutput("<br>findStructArrayValue(#htmlEditFormat(str)#) of length
#len(str)#<br>");
if (isDefined(listGetAt(str, x, ".")))
{
sStr = listGetAt(str, x, ".");
curStruct = evaluate(sStr); // Get first item as struct. Just a
pointer according to Macromedia
do
{
x = x + 1;
itterations = itterations + 1;
if (x gt lLen)
{
newStr = "<strong>Ran out of elements processing
""#str#""</strong>";
finished = true;
}
else
{
element = listGetAt(str, x, ".");
// Process array if required
startBPos = find("[", element, 1);
if (startBPos gt 0)
{
arrayTxt = mid(element, startBPos + 1, len(element) -
startBPos - 1); // No braces please
element = mid(element, 1, startBPos - 1);
//writeOutput("Finding array [#element#] with var
[#arrayTxt#]<br>");
if (isNumeric(arrayTxt))
arrayPos = arrayTxt;
else
arrayPos = evaluateString(arrrayStr);
//writeOutput("Array Pos [#arrayPos#]<br>");
}
//writeOutput("Processing element [#element#]<br>");
if (structKeyExists(curStruct, element))
{
if (x eq lLen)
{
// A the end and we have a value
newStr = structFind(curStruct, element);
if (isArray(newStr))
newStr = newStr[arrayPos]; // Worked out
previously and if not then it defaults to 1
finished = true;
}
else
{
// Move to next struct element
curStruct = structFind(curStruct, element);
if (isArray(curStruct))
curStruct = curStruct[arrayPos]; // Worked out
previously and if not then it defaults to 1
}
}
else
{
newStr = "<strong>Element ""#listGetAt(str, x, ".")#""
missing from ""#str#""</strong>";
finished = true;
}
}
}
while (finished eq false and itterations lte 1000);
}
else
newStr = "<strong>Element ""#listGetAt(str, 1, ".")#"" missing from
""#str#""</strong>";
//writeOutput("return [#htmlEditFormat(newStr)#]<br>");
return newStr;
}
On 08/12/06, Brett Payne-Rhodes <[EMAIL PROTECTED]> wrote:
>
>
> Good points all...
>
> What is needed are some good examples so that we can understand what the
> alternatives are and why they are better than evaluate(). Examples of what
> evaluate *should* be used for would also be useful.
>
> For me, dynamic forms were where I learnt to NOT use evaluate.
>
> a form field like <input name="name#userId#">
> I would access via something like <cfset theName = evaluate("
> form.name#userId#")>
>
> When in reality I can use <cfset theName = form["name#userId#"]>
>
> which might not be much different in execution time but is a much better
> coding practice simply because it reduces the complexity of the code for the
> trainee programmer who comes along to alter it later.
>
> Can anyone supply a simple example of an 'acceptable' use of evaluate()?
>
> Cheers,
>
> Brett
> B)
>
>
> Adam Cameron wrote:
> > G'day
> > My testing recently suggested that terrible the performance hit one
> > used to get with evaluate() has all but gone now. So that's good.
> >
> > The only gripe I have with it is that people tend to over-use it: it's
> > seldom the correct answer to a question about dynamic variables or any
> > other situation where one's inclination might be to use it.
> >
> > You touch on one instance when it seems to be necessary to use it, and
> > I can think of a couple of others, but a lot of people seem to use it
> > every time things get slightly confusing to them. These are perhaps
> > the same people #who# #put# #pound-signs# #around# #everything#, "just
> > in case".
> >
> > My position is this:
> > 1) if one can reasonably NOT use evaluate(), then don't.
> > 2) if one thinks "I need to use evaluate() for that", have another
> > small think about it before deciding it's the right answer.
> > 3) and, yes, sometimes it's unavoidable. In these cases, there's no
> > need to feel like one is doing something wrong.
> >
> > I think it's a three-step process before settling on using it.
> >
>
>
> >
>
--
Cheers
Simon Haddon
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"cfaussie" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to [EMAIL PROTECTED]
For more options, visit this group at
http://groups.google.com/group/cfaussie?hl=en
-~----------~----~----~----~------~----~------~--~---
