some regexp problems

[entrapmen]

Hi,

i was working on spammers a while ago and im back again. but now spammers get developed. they really got weird spam methods.

here are some of the examples:

ºwºwºwº
I¤R¤C

(using with color and it seem like w w w and I R C)

lots of thing like that. i think they all can be catch by using "regexp" but i couldnt figure it out. i dont wanna add all spam words which i was doing to that time. speacial charecters allways changing (, . ? ' ( ) and words). i want to catch them if there is "w w w" in the sentence. it doesnt matter where they are. (ex: where is will wonder) <<-- that person must be caught.

got an another problem, the channel im protecting is at about 300 person and its very active i m using greet message to catch some kind spammers which are using the way "away message". so i have to message all the users but when i give that work to 1 bot it get lagged or "eflood"ed. so i want to cut it to half. if nick starts with "a-k" message else dont if nick starts with "k-z && "special charecters??"" message else dont.

sorry for bad english, i tried my best. thanks

[greenbear]

You could remove everything thats not a normal char or a number from the string before doing any regexp matches on it.

Code: Select all

regsub -nocase -all {[^a-z0-9]} $line {} line

[Alchera]

| stripcodes <strip-flags> <string>
| Description: strips specified control characters from the string given.
| strip-flags can be any combination of the following:
| b - remove all boldface codes
| c - remove all color codes
| r - remove all reverse video codes
| u - remove all underline codes
| a - remove all ANSI codes
| g - remove all ctrl-g (bell) codes
| Returns: the stripped string.
| Module: core

NB: eggdrop v1.6.17.0

[demond]

/me makes a note to himself to check that (source code for stripping) out

[De Kus]

regsub -nocase -all {[^a-z0-9]} $line {} line

Just think what will happen to "vist us on http://www.mystupid.ad and our #stupid-channel". yes, it will become "vistusonhttpwwwmystupidadandourstupidchannel". Good luck .
you won't be able to find advirtising spam that way, you should keep at least the characters " ", "#", "/", ":" and ".". So I would rather prefer: [^a-z0-9\./:# ]


However, for general purpose stripcodes is the best and probably the fastest solution.

[demond]

this will strip mIRC color and control codes:

Code: Select all

regsub -all {([\002\017\026\037]|[\003]{1}[0-9]{0,2}[\,]{0,1}[0-9]{0,2})} $str {} str

and this will match a hotlink (clickable URL) or chan ad:

Code: Select all

regexp {(?i)(http://|www\.|irc\.|\s#)} $str

[entrapmen]

this will strip mIRC color and control codes:

Code: Select all

regsub -all {([\002\017\026\037]|[\003]{1}[0-9]{0,2}[\,]{0,1}[0-9]{0,2})} $str {} str

and this will match a hotlink (clickable URL) or chan ad:

Code: Select all

regexp {(?i)(http://|www\.|irc\.|\s#)} $str

thats the problem, i know those codes and was using them. but unfortunately they are no more working. coz bots doesnt advertise clickable links.

[entrapmen]

about the second problem can someone help me?

i think one if and else should be enough, like if nick starts with the letter a-q do something else do nothing...

[demond]

this will strip mIRC color and control codes:

Code: Select all

regsub -all {([\002\017\026\037]|[\003]{1}[0-9]{0,2}[\,]{0,1}[0-9]{0,2})} $str {} str

and this will match a hotlink (clickable URL) or chan ad:

Code: Select all

regexp {(?i)(http://|www\.|irc\.|\s#)} $str

thats the problem, i know those codes and was using them. but unfortunately they are no more working. coz bots doesnt advertise clickable links.

really? give me just ONE clickable link that these don't match

[entrapmen]

this will strip mIRC color and control codes:

Code: Select all

regsub -all {([\002\017\026\037]|[\003]{1}[0-9]{0,2}[\,]{0,1}[0-9]{0,2})} $str {} str

and this will match a hotlink (clickable URL) or chan ad:

Code: Select all

regexp {(?i)(http://|www\.|irc\.|\s#)} $str

thats the problem, i know those codes and was using them. but unfortunately they are no more working. coz bots doesnt advertise clickable links.

really? give me just ONE clickable link that these don't match

i say it really works for clickable links but bots(spammers) doesnt use that method anymore. thats what i am saying. sorry for bad understandings and bad english

[demond]

so what do you care if they use another method with no clickable links? nobody will bother to manually strip the spam and paste it in their browser - people are lazy, spammers know that; spam without clickable links is harmless (annoying yes, but that's all about it)

[entrapmen]

so what do you care if they use another method with no clickable links? nobody will bother to manually strip the spam and paste it in their browser - people are lazy, spammers know that; spam without clickable links is harmless (annoying yes, but that's all about it)

as you say they are annoying. i m trying to stop them before they message to users. users are lazy if they would //mode their self +R they would not see the spammers...

anyway demond can you help me about the second problem

[demond]

if you've managed to write some script satisfying some of your needs, you should be able to grasp the [regexp] concept and eliminate nick patterns that are of no interest to you, it's not that hard