detect channel advertising

[ProXy]

Hello, I just use Nosense to protect my chan against channel advertising. Theres only a little error in nosense: It doesn`t match Channel like ## oder ###. these are valid qnet chans, so i wanted to rewrite the regexp.

I want to check if one or more char is after # and save the channel in a var called $match.

I tried to match using this source:

Code: Select all

[regexp -nocase {#.+} $text match]

But this does not work, because my regexp doesn`t stop before the next space. If the string is "i am on #htp and you are not", $match only should have the value "#htp". Could anyone please tell me how to use regexp to match the right part of the string?

ProXy

[caesar]

foreach word [split $text] {
if {![string match "*#*" $word] || [string match "#" $word] || [string compare -nocase $word $chan] == 0} { continue }
rest of the code
}

This way will see if someones uses a #something and will compare the #somehting with the channel that #something was said in and if it's the same will continue, if not will do your proc. Hope this helps you..

[Papillon]

Code: Select all

regexp {((#)+([a-zA-Z0-9])+)} $text match

$match will then contain "#blafblsfbljd" or "##afjhfef34" or whatever =)

[ProXy]

But i guess this will not match ##, right?

[ppslim]

Few test examples, by using "tclsh"

proc tm {in} {
regexp {((#)+([a-zA-Z0-9])+)} $in match
return $match
}

tm "hello to #all out there"
#all

tm "Funny old ##world"
##world

tm "##"
can't read "match": no such variable

tm "hello ## all"
can't read "match": no such variable

As such, it causes error when ## is on it's own

proc tm {in} {
regexp {((#)+([\#a-zA-Z0-9]{0,})+)} $in match
return $match
}

tm "##"
##

tm "hello ## all"
##

The regexpr works by looking for a #, followed by and uper/lower case letter or numbers, until the next non-matched char.

Because the # wasn't matched within it, it coudln't pick it up.

At the same time, it didn't allow you to check for single #'s. The {0,} means 0 or more matches.

As a # doesn't mean a spam (usualy), you can change the {0,} to {1,} requiring at leave one character after the #.

Aditionaly, you will find that you need more characters in the name matching, as channel names can contain other charcters.

Objects to add may be -, _, !, colour character, bold, underline, reverse, normal, ctcp, and many many more (best to see the IRC RFC).

[ProXy]

Well I think it`s better to catch EVERYTHINK against the # and just split the result string to get the channel. This will also work for every speacial char...

[Papillon]

just a little sidenot... instead of using {0,} you can just use *, or + instead of {1,} ... this will do exactly the same thing

[user]

How about matching anything but space?

eg: regexp {#[^ ]*} $the_string chan

[Papillon]

How about matching anything but space?

eg: regexp {#[^ ]*} $the_string chan

even better =)

[ProXy]

Well if i use this, also bla#bla ist detected as channel advertising...

[TsT]

Code: Select all

# returns channel name if found,
# else return nothing
proc getchannelname_ifexists {text} {
    if {[regexp {(#[^, ]+)} $arg _ chan]} {
        return $chan
    }
}

[user]

regexp {(?:^|[ ])(#[^, ]*)} $text chan chan

Notice the *, not +, because "#" is a valid channel name.

The first part of the rule is to make shure there's a space before the # (or that # is the first char in the text)

[ProXy]

Sorry, use rbut your source even detects abc#abc as a channel ad...that`s not want i`m looking for.

Allowed:
bla#bla
bla##
bla#

NOT allowed:
#
##
#bla
#bla#

[ProXy]

This is a nearly Solution: regexp -all -nocase -- {^(#.*)} $text match

The last problem. If someone say "#abc abc abc", then not only "#abc" is saved to $match, but the whole string

[user]

Solution: regexp -all -nocase -- {^(#.*)} $text match

^ matches the start of a line, so for your rule to match the channel name must be the first word in the text, the * quantifier is greedy, so you'll get the complete line try my previous suggestion...