Revision as of 18:44, 3 September 2015

The $* allows you to iterate over all of the tokens contained within $1-. The way this works is much like a while loop. $* has been omitted from the help file since while loops were introduced.

$* is extremely powerful, because it is much faster than using a while loop on a list of tokens. Understanding how this works is pretty simple, so consider the following:

var %list a b c d,%a 1
while ($gettok(%list,%a,32)) {
echo -a $v1
inc %a
}

This can be written as:

tokenize 32 a b c d | echo -a $*

The above is much smaller, and a lot faster. Pretty cool, isn't it?

Notes & quirks

Why was the $* identifier removed from the help file in the first place? Well, $* was removed because it does not really work the same way that the other identifiers do, and because it is quirky.

Quirky? Quirky, how?

Well, mIRC takes the command $* appears in, and replace all of the occurrences in the line by a special marker: `~$*: You can see that by doing for example:

//tokenize 32 abcd | echo -a $left($*,1) $+ $chr(3) $+ $mid($*,2)

Let's take a look at this and understand why the above return such a value:

//tokenize 32 abcd | echo -a $mid($*,2)

Why isn't this returning bcd? Because of the usage of the special marker `~$*, mIRC has stored that the command is actually "echo -a $mid(`~$*,2)".

Then, for each token, (here only $1 == abcd), mIRC evaluates the line, here $mid(`~$*,2) becomes ~$*, and then mIRC replaces the marker by the token and executes the echo command. However, after an operation, like $mid here, that marker cannot be found;

Basically, it cannot be guaranteed to get the correct value of $* inside an identifier.

There is a workaround for the above issue, and that workaround is by using scid and scon:

//tokenize 32 abcd | scon -r echo -a $!mid( $* ,2)

mIRC replaces $* by the marker, but scon has an extra evaluation system which fits perfectly. The $* mechanism is enabled on scon, and mIRC stores the command of the $* as "scon -r echo -a $!mid( `~$* ,2)"

Then, mIRC evaluates the line for each token, which then becomes: "scon -r echo -a $mid( abcd ,2)", and finally the scon command is executed, resulting in the expected value being echoed.

Another issue is that you cannot call $* more than once in the same scope; the command will simply be skipped. However, this can also be circumvented/worked-around :)

Simply retokenize after using $*, which forces the update of $1-, $* can be used again, but only if the number of tokens passed is higher than the previous number of tokens passed. $* will only start from the previous number of tokens + 1:

//tokenize 32 1 2 3 | echo -a $* | tokenize 32 4 5 6 7 8 | echo -a $* | echo -a here

The first three tokens of the second tokenize, "4 5 6", are dummy tokens which are passed to fill in the gap. $1- can be used in front of the next tokenize command in order to start from those new tokens, much like a normal situation.