Forums Scripts & Popups $regsub

1. how can i know the position of \1 in the text ?
so in $regsub(abcd,(b),\1,%x) \1 should replaced with something that will return 2 (cuz that's the position of b)
2. $regsub(abcd,(a),$len(\0),%x) why the \0 doesn't return its value to identifiers? it returns 2 (the length of "\0") except 4
3. and another question, what is the maximum times that regex can match with //g or a lot of ( )'s ?

1. $regml(1).pos

2. It's just the way it works. mirc first evaluates the content of <subtext>, seeing nothing special in \1 (just like it does with every other identifier), then passes the (evaluated) string to the appropriate PCRE function. What you want might be a bit of a pain to implement but is not impossible (and, if I remember what I read correctly, the new version of PCRE would make things easier towards this direction). I would personally like to see <subtext> working in a manner similar to the <command> parameter in $findfile(), ie filling $1- with each captured match and allowing you to pass it to identifiers.

3. I am pretty sure it's no less than 999. What is important to know is that it definitely works with up to ~930 matches, which is about the max number of characters you can have in a mirc string. So any limitation in the number of captures wouldn't show up in mirc anyway.

oh, thanks
about the first question, i wanted to return the position of the \1 (in $regsub, not $regex and then $regml) like $regsub(abcd,(b),\1.pos,%x) - %x = a2cd

$regml() works with $regsub() too. However, what you want to do here is impossible. There is no way to retrieve the position of \N dynamically from within <subtext>.

You can use $replace as a work around. You just need a character or string of characters you know will not exist in the original string.
Not exactly what you want but it should work.

nah.. that's not a good idea. i want to make it with one $regsub.
thanks anyway

If you are simply looking to do exactly what your example shows, replace 'b' or any character with its relative position in the string, this can easily be done using $bfind and /bset with your string in a &binvar.

As far as getting $regsub to perform miracles, it simply wont happen. $regsub is incapable of dynamic replace strings. Eg, $upper(\1) will not give you an uppercase 'B'. This may be possible to do in PERL, but not with PCRE, and is completely out of Khaled's hands.

- Raccoon

As far as getting $regsub to perform miracles, it simply wont happen.

It's not a miracle he's asking for... Getting $regsub to work like this:

$regsub(abcd,/(.)/g,$upper($1),%a)

is very much possible. If something like this can work in $findfile() why couldn't it in $regsub()? Hell, I could even script this using only $regml() and $regml().pos, if only there was a way to avoid evaluation of the 3rd parameter in scripts (ie in $myregsub(), $3 would have to return "$upper($1)", not the evaluated content of it) without having to escape anything with ! or $+ or $eval(...,0) etc.

I originally thought that $regsub() calls a ready-made PCRE function, so that Khaled had no actual control over it (unless he attempted to mess with the PCRE source himself). However, I was browsing through the PCRE source today and found no function that does any $regsub-related job, so I can only assume that $regsub() is hand-made. If this is the case, it's good news; implementing this feature would be WAY easier (although I insist that it would still be possible even if this wasn't the case).

yeah, i have suggested that once that $neval($N) will return the $Nth parameter from the alias without evaluating it