problem with bind flud

[Burke]

hi

i wanted to make my own antiflood script. it will be very useful and easy if i can use the internal flood-**** settings of the bot. them can easy set up. the detection for them is ok too.

Code: Select all

    (30) FLUD (stackable)
         bind flud <flags> <type> <proc>
         procname <nick> <user@host> <handle> <type> <channel>

         Description: any floods detected through the flood control settings
           (like 'flood-ctcp') are sent here before processing. If the proc
           returns 1, no further action is taken on the flood; if the proc
           returns 0, the bot will do its normal "punishment" for the flood.
           The flood types are: pub, msg, join, or ctcp (and can be masked to
           "*" for the bind); flags are ignored.
         Module: server

Ok, there is no info about the size of any floods. If the proc return 0 --> no kick from the internal bot function. this will be good, because i make my own kicks.

Code: Select all

bind flud - "*" floodtest111
proc floodtest111 { nickname hostname handle type channel } { 
putlog "floodscript $nickname $hostname $handle $type $channel" }
return 0

that`s how i start to code. if 1 user flood, it will trigger correctly and show me all needed things on the partyline. just a fine binding and easy variables.

you won`t see the rest of noob coding tcl it`s just getchanhost, mode +b, utimer 10 kick user.

my problem is. the bot will not detect bigger floods. if 1 user flood at the same time the script work just perfect as it should.

but if 2-10 users flood at the same time, the bot won`t do anything.

i tested it without any scripts (no binds flud). the internal eggdrop function works. it will only kick 1 users which is flooding but on big floods the bot do nothing. this last sentence is my question: is this a bug or is my code wrong?

[De Kus]

eggdrop has no built tsunami flood detection, this is provided by sentinel.tcl. only if single users are triggering a flood your bind will be triggered.
but if you mean the bot doesnt trigger the bind if many people individually trigger a flood, this sounds like a major problem with your eggdrop core or you messed up your code totally .

[slennox]

Eggdrop's flood detector seems to be intended for individual flooders--it does not detect flooding where multiple flooders alternate messages. That's why the flud bind is not very useful for flood protection scripts.

[Burke]

your replys sounds good. it`s not the best way to detect them (because it only watch 1 user) but i should work.

for the test i changed flood-chan to 3:20 (this will be changed later, for sure).

- beginn test / demonstration

-- test with 1 flooders

[16:34:05] * g740 (GideonShav@c-24-13-186-77.client.comcast.net) has joined #ah20
[16:34:05] * x6346 (~j8989@lsanca2-ar28-4-47-249-201.lsanca2.dsl-verizon.net) has joined #ah20
[16:34:16] <x6346> test
[16:34:24] <x6346> test
[16:34:25] <x6346> test
[16:34:26] * x6346 was kicked by [-AxF-steve-][o (flood)

- result: all ok.


- test with 3 floodbots
[16:34:39] * x6346 (~j8989@lsanca2-ar28-4-47-249-201.lsanca2.dsl-verizon.net) has joined #ah20
[16:34:46] * n9464 (~w3610@80.81.58.57) has joined #ah20
[16:34:50] <g740> test
[16:34:50] <n9464> test
[16:34:50] <x6346> test
[16:34:52] <g740> test
[16:34:52] <x6346> test
[16:34:52] <n9464> test
[16:34:57] <g740> test
[16:34:57] <n9464> test
[16:34:57] <x6346> test
[16:34:59] <g740> test
[16:34:59] <n9464> test
[16:34:59] <x6346> test
[16:35:01] <g740> test
[16:35:02] <n9464> test
[16:35:02] <x6346> test
[16:35:12] <@Burke^> !tcl binds flud
[16:35:14] <@[-AxF-steve-][o> OK - 0.028 ms
-- result: no one will be kicked

- test with 1 floodbots again
[16:35:32] <@Burke^> ---
[16:35:38] <g740> test
[16:35:40] <g740> test
[16:35:41] <g740> test
[16:35:42] * g740 was kicked by [-AxF-steve-][o (flood)
[16:35:52] <@Burke^> -------
- ok, the flooder will be kicked


- test with 2 floodbots
[16:35:58] <x6346> test
[16:36:02] <n9464> test
[16:36:02] <x6346> test
[16:36:04] <n9464> test
[16:36:04] <x6346> test
[16:36:05] <x6346> test
[16:36:06] <n9464> test
[16:36:07] <n9464> test
[16:36:08] <x6346> test
[16:36:09] <x6346> test
[16:36:10] <n9464> test
[16:36:11] <n9464> test
[16:36:11] <x6346> test
- result: the bot will kick no one

i tested it with my own bot and 2 other bots of friends. for sure it`s eggdrop 1.6.17.

about the hosts of the flooders. - no clones, differnent hosts and diferents ip`s (-> more users).

this seems like a bug in the source of the eggdrop or is this bind just unless and not desinged for that?

[slennox]

Your last test is a exact replica of one I did six or seven years ago. I recall mentioning it to someone on the devteam but didn't make a formal bug report--it was not clear that it was a bug rather than a mere limitation of a very rudimentary detector. In any case, the issue isn't a recent one.

[YooHoo]

Your last test is a exact replica of one I did six or seven years ago. I recall mentioning it to someone on the devteam but didn't make a formal bug report--it was not clear that it was a bug rather than a mere limitation of a very rudimentary detector. In any case, the issue isn't a recent one.

very very interesting, to be sure. May not be a recent issue, but it is a security problem, damn straight...
Excellent time to advertise sentinel.tcl, eh slennox

[Burke]

i maked (ca.) 2 weaks ago a bug report but no one reply me.

that was reason for asking here. i think them ignored me because i am a big noob in tcl and it was my error.

them don`t care about those problem or it will take a while until them reply me to the bug report?