Parametric RepTags

let me put an example more abstract

Lets define $sometag(arg1,arg2) = Hi. My name is %arg1 and my age is %arg2

The = sign i put was for default values we sometimes put in functions if arguments are omitted.
so for example if i define the tag as

$sometag(arg1="tom", arg2="19") = Hi. My name is %arg1 and my age is %arg2
if i write
$sometag(nick,25) it will output-> Hi. My name is nick and my age is 25
if i just write
$sometag it will output -> Hi. My name is tom and my age is 19
which are the default values.

I hope this example is simpler.

Why would it introduce a security hole. ?

It does not matter if its a sitewide tag or a special module tag (if i really understood your question)

Now when the reptag finds the $sometag$ it replaces it with the defined text.
What i was suggesting is making it a step ahead. replace the text and if i find a place holder %arg for example replace it with the first argument, if i find a second %arg replace it with the second argument and so on.

For example

sometag(arg1,arg2,arg3) = sometext %arg and sometext %arg and even more text.

Now if i type $sometag$ it will output--> sometext %arg and sometext %arg and even more text.
if the module makes a second pass and replaced the first %arg with arg1 the second %arg with arg2 and the third %arg with arg3.

it would really be helpfull.

Correction

"Now if i type $sometag$ it will output--> sometext %arg and sometext %arg and even more text.

BUT IF i type $sometage(arg1,arg2,arg3)$ then if the module makes a second pass and replaced the first %arg with arg1 the second %arg with arg2 and the third %arg with arg3.

it would really be helpfull.

Case 1 is highly problematic. The reason it deprives the essense of the parametric approach, the dynamic nature

Case 2 would be nice (if it is possible) I think ( beeing reserved is not so important) . You can always use { or [ or [[ or $$ or what ever you like to signal start and end of place holding.

If it cannot be done .. no prob :)

if the problem its UI you dont need something new.
In the existing UI

LEFT you define tag right the reptag

sometag hi i am %arg and my age is %arg

TILL HERE nothing new.

The only thing that needs to change is parsing . Let me place it in althorithmic steps.
LET the calling reptag be

$sometag(nick,34)$

1) If inside the $ $ there is no ( and ) then use the normal procedure and just output the text
2) if the are () then create an array of the values as args[] = {nick,34} in our case (its just splitting the string when comma occurs)
3) then to the output text (which is currently "hi i am %arg and my age is %arg")
do the following :replace the first instance of %arg with args[0] the second instance of %args with args[1]
4) output the newly create string.

Very simple string manipulation i think and you dont need to change a bit of the UI.

Bullseye.

Although defaults are not a must. Just an extra feature of arguments.
So if you skip default values you dont need ANYTHING new in the definition

$SOMETAG$ will just do for definition

and then $SOMETAGE(a1,a2,a3,a4,a5)$ {it does not matter how many arguments} because the procedure is just the same

1) split the string a1,a2,a3,a4,a5 with delimiter , and create array args
2) currentargument=0;position=0;
2) while (position = findnext(outputtext,"%arg",position))
{ replace(outputtext,"%arg",args[currentargument++],position) //just replace in the string outputtext the string %arg with the argument stored in the array args and its the currentargument , and the starting position of the %arg string is position
}//endwhile

ofcourse functions are schematic but the procedure is as simple as that (unfortunatelly i dont code in php :)

So if you leave defaults out, the procedure is really easy i think.
Dont you agree?

actually keyed or not args is the same. I thought it would be a little bit easier to program non keyed args and much FASTER because replacement requires only ONE pass (as each time you encounter an %arg you trigger a replacement). The keyed replacement is slower as you require to pass the output string as many times as the number of arguments to find %argN.

Just that. Actually there is not difference in the result. In the unkeyed solution order is defined by occurence of %arg. In the keyed order is defined by naming convention.

The final result is the same but i think unkeyed is way much faster to parse and replace (did i convince you? xexe.)

It seems that keyed are faster but in fact they are not.
Let me tell you why.

What a find_replace(sourcestring,findstring,replacestring) function does is search through an array (string = array of chars) to find an instrance of findstring
This is LINEAR procedure. Let me discribe a worst case senario.

Let as have a string of N characters where N is big. And we have to replace K strings that are ALL at the end of the source string ( i mean "blablabla.... %arg1 %arg2...)
The search function will find the arg1 after searching all characters before arg1 . The same applies to arg2 and so on.
So the search function will have to traverse the array of chars N*K times appox (where K is the number of arguments)

In the unkeyed way you the search will just be N times (because the last search provides the position the %arg placeholder was found)

If N is big and K increases we have a huge difference in implementation times.

In the keyed version the overload is just hidden under the search function.

I dont know if i explained it it well

May i provide some simple example.

the image tag (which exist in a module ofcourse but its an easy example)
You would not need to write a module to post an image. Just define a tag as
$IMAGE(path,size,title) for example.. and you have your image

I for example use it to embed a chess game viewer in a post. How?
To embedd it i just need to copy paste a certain code + the body of the chess game. So my solution now is to create TWO TAGS
a $STARTTAG$ and and $ENDTAG$. They simple hold the code that must be place on the body of the post and are SPLITTED at the place the gamescore (moves) must be inserted. So the user writes $STARTTAG$chessmoves....$ENDTAG$ and its working perfectly.

So i am simulating an ONE parameter TAG. TAG(chessmoves). If i want to simulate 2 parameters i have to manually create 3 tags and so on.

I can list so many examples.

I let my users use the tags but ofcourse i never allow them to create tags (for security reasons)

Just download the latest dev install and then go to your admin section at reptags settings page and look under Expert settings.
You have to enable parametric tags.
It works fine by the way (i have not tested for default arguments) but parametric tags are GREAT