Test version of sentinel.tcl v4.09.1

[slennox]

I've finally made some time to work on sentinel.tcl and prepare an updated version. While preparing it, I had a look at the overhauled version that I'd been working on a few years ago, and although it appears mostly complete, I concluded that it would take too much time to sufficiently refamiliarise myself with it and retest everything. So good old v4.09 with an oft-requested feature or two added will have to do.

v4.09.1-test1 for netbots.tcl users (I don't know yet whether I'll be releasing a standalone version) fixes a small bug in the nick flood detection and adds that most requested feature: custom lock modes. There is one caveat: the +i and +m modes remain. This was to reduce the complexity of the modifications and prevent an end-run attack on weaker modes.

You can download test1 here. To install it, just overwrite the existing sentinel.tcl and .rehash. You'll find the new sl_locktimes setting at the top of the script, including a description of it. I wouldn't bother moving this setting to netset.tcl for now; just adjust it in place. Note that the sl_ilocktime and sl_mlocktime settings in your netset.tcl are no longer used and will be ignored. After loading the new version, type .sentinel to check that your lock modes/times are set as expected.

Problems should be reported in this thread. Logfiles demonstrating problems can be e-mailed to me directly.

[slennox]

Test2 is now available. This adds some text and notice flood detection in a way that was relatively simple to implement--as part of an amalgamated text/notice/tsunami/avalanche mechanism. Please refer to the section at the top of the test script for more information.

[awyeah]

Oh a new final version of sentinel is finally out! after waiting a long interval of 3 years

[Alchera]

I've been using Test2 version of sentinel and am pleased with its' performance. So far, with a number of attempts to text flood a number of channels I have bots in, it's performed as expected. In one channel (which I am actually sitting in to watch) last night 6 flood bots entered and only managed to get a total of 12 lines in before the channel locking kicked in; bearing in mind that I have yet to still "play" with the sl_txlength setting.                                            

[Vixs]

I am running eggdrop v1.6.18+morecopyrights and have loaded the test2 file and this is wat the error it shows

Code: Select all

<RockBabe> [16:38] can't read "sl_bxsimul": no such variable
<RockBabe>     while executing
<RockBabe> "if {$sl_bxsimul} {
<RockBabe>   bind raw - 001 sl_bxserverjoin
<RockBabe>   if {![info exists sl_bxonestack]} {
<RockBabe>     set sl_bxonestack 0
<RockBabe>   }
<RockBabe>   if {![info exists sl_bxversion..."
<RockBabe>     (file "scripts/sentinel.tcl" line 1115)
<RockBabe>     invoked from within
<RockBabe> "source scripts/sentinel.tcl"
<RockBabe>     (file "eggdrop.conf" line 1371)

Edit: I realised , Does this require netbots to work?

[Alchera]

Edit: I realised , Does this require netbots to work?

Oh, very much so.

[mm]

Hi, how the setting will work for undernet, since undernet has mode +r and may be doesn't need +i
so should i set
set sl_locktimes {r:120 m:60}?

thanks

[Alchera]

Yes, that should work fine.

[Sir_Fz]

There are two rules. First, the minimum locktime for any mode is 30 seconds. Second, modes +i and +m must be included.

I guess you can't remove +i and +m.

[mm]

Yeah, well i've upload the test2 with the following setting

set sl_locktimes {i:30 m:60 r:60}
set sl_txflood 6:15
set sl_txlength 180
set sl_nclength 1
set sl_linecap 50:15

then i have used few clone ip's to test this using control char, bold etc, but it didn't trigger at all..
when i do .sentinel i can see the setting is there

[De Kus]

with a bit teaking I got this test version running as standalone. you just need the settings from netset.tcl and the procs nb_sendcmd and nb_killutimer (i have bitchx simul disabled ) from netbot.tcl.

PS: beat me, if I missed any change I made .

PPS: I changed the locktime setting to this

Code: Select all

if { $network == "QuakeNet" } {
	set sl_locktimes {i:30 m:30 r:300}
} elseif { $network == "euIRCnet" } {
	set sl_locktimes {i:30 m:30 R:300}
} else {
	set sl_locktimes {i:120 m:30}
}

probably only intresting when 2 bots running on diffrent networks are running on same directory. But maybe you consider using the "only regististred" Mode in flood cases, this mode will block most multi-host spamer and leave the channel free for kind registred chatters .

[mm]

Sorry test2 version is working great. thank you to slennox

[arvent]

Hii. I use netbots.tcl v4.10. When sentinel get someone banned it get double * in front of the inent.
I have set it with set sl_masktype 2 in order to get this ban mask: *!*ident@*.domain.com, but instead get the: *!**ident@*.domain.com.
Am I doing something wrong and the most important what can I do to fix it?
Thanks in advance.
P.S.
Great work with the scripts. Brings much joy and safety.
Chears.

[slennox]

I have set it with set sl_masktype 2 in order to get this ban mask: *!*ident@*.domain.com, but instead get the: *!**ident@*.domain.com.
Am I doing something wrong and the most important what can I do to fix it?

Looks as though an argument was added to the masktype command in eggdrop 1.6.20 allowing selection of a hostmask type, but the new default type is different from what the command would've returned prior to 1.6.20, which can break compatibility with scripts that expect it to behave as documented in 1.6.19 and earlier.

A small change to line 1155 (for the sentinel.tcl included with netbots.tcl v4.10) will probably fix it:

Replace:

Code: Select all

    2 {return *!*[lindex [split [maskhost $uhost] "!"] 1]}

With:

Code: Select all

    2 {return *!*[lindex [split [maskhost $uhost 0] "!"] 1]}

[arvent]

You are correct about the eggdrop version. I am using 1.6.21.
The ban mask change with the new line to: *!*ident@machine.domain.com as it is in sl_masktype 1.
Thanks for your replay and efforts.
Best