| |
| |
| |
|
Page: 1 2
<spoogle> csircd == ***
<anfl> andro: I emailed efnet-drama about testmaskgecos/chantrace to preclude any *****ing
<anfl> jilles: I made spoofed sockhost 255.255.255.255
<jilles> ok
<anfl> andro: ive been debating moving testmask* over to match_esc()
<anfl> which reminds me actually
<anfl> still need to fix that kline *@* bug
<anfl> ah
<anfl> I see now
<anfl> id done 0 for a reason I wasnt sure of
<anfl> now I know what it was -- the sockhost of a remote spoofed user is "0"
<jilles> that's a TS6 peculiarity really :)
<anfl> yup, but it looks odd in chantrace without it
<anfl> 22:26] --- Oper #ratbox ShadowDog esast delusion.shadow.no 255.255.255.255 :E. Alexander Stromme
<anfl> [22:26] --- Oper #ratbox SalBlsd cking starfire.blessed.net 0 :Steven King
<anfl> im just debating whether servername is important..
<jilles> why do you store a "0" sockhost as such anyway
<anfl> but I dont think its worth the changes itd involve
<jilles> wouldn't it be better to keep it the empty string then?
<anfl> it cant be empty
<anfl> without breaking the protocol :)
<anfl> and you should add exceptions for stuff you send, not what you receive :)
<anfl> if I wanna add servername to it itd be another numeric
<anfl> hum
<anfl> ill think about it :p
<jilles> this gets rather messy
<anfl> yup
<anfl> its that or output servername instead of channel
<anfl> ;-)
<anfl> (.. instead of cl***)
<jilles> chantrace lacks an end numeric
<anfl> whoops
<anfl> ill fix that tomorrow :)
<anfl> and replace channel name with servername
<anfl> and probably go back to sending "0"
<anfl> dunno bout that yet tho :)
<anfl> I may even decide to differentiate between spoofs and TS5
<anfl> make spoofs "0" and TS5 "255.255.255.255"
<anfl> bed now tho :)
<anjuta> is it possible to check a hostmask against the placed glines ?
<anjuta> similar to testmask, but for active glines
<jilles> /quote testline doesn't work?
<anjuta> nice.. i knew there should be something :)
<Cdogg> can services change a users nick?
<jilles> yes, using ENCAP RSFNC
<Cdogg> is there docs on that?
<Cdogg> heh googled it and it returns
<Cdogg> a (crude) shot at an ENCAP implementation, by jilles
<Cdogg> famous now :/
<jilles> doc/services.txt, modules/m_services.c
<Quai> but how many clients supports nickchanging from the server?
<jilles> I don't think many clients will have a problem with it
<jilles> dalnet/ircnet already do it
<AndroSyn> KLINE 1440 foo@10.99.1.* foo
<AndroSyn> :jupiter.ratbox.org NOTICE synandro :Added temporary 1440 min. K-Line [foo@10.99.1.0/24]
<wfl> :)
<wfl> what about 10.99.*? :p
<AndroSyn> i'm intentionally not gonna support that ;)
<AndroSyn> i could i suppose, but i don't like it
<AndroSyn> it'll do 10.99.*.*
<AndroSyn> i guess it wouldn't be hard to do 10.99.*
<wfl> as you will :)
<AndroSyn> heh
<AndroSyn> KLINE 1440 *@12.* :moo
<AndroSyn> :jupiter.ratbox.org NOTICE * :*** Notice -- androsyn!androsyn@127.0.0.1{AndroSyn} added temporary 1440 min. K-Line for [*@12.0.0.0/8] [moo]
<AndroSyn> KLINE 1440 *@12.0.* moo
<AndroSyn> :jupiter.ratbox.org NOTICE androsyn :[*@12.0.0.0/16] already K-Lined by [*@12.0.0.0/8] - Temporary K-line 1440 min. - moo (2005/11/23 12.16)
<AndroSyn> its a really ugly hack too..
<SiD3WiNDR> nice :p
<AndroSyn> basically i just mangle the text string into a cidr mask and then feed that into parse_netmask() :P
<AndroSyn> the mangling part itself gets called from in find_user_host()
<AndroSyn> committed to devel
<AndroSyn> and 2.2 as well
<anjuta> uhh, Too many global connections .. it's reporting the ip instead of the host, but the ip has never been seen on the network..
<anjuta> isn't it more logical to report the actual host used ?
<AndroSyn> if its TS6..
<AndroSyn> it sees the IP
<anjuta> it's not ts6
<anjuta> it's not used on any server atm
<AndroSyn> well it could be too many global connections from either the IP or the host
<AndroSyn> i think returning the IP is more useful
<anjuta> it doesn't matter for the server, isn't it ?
<AndroSyn> hmm?
<anjuta> oops, in the context of placing a gline.. sorry
<anjuta> uff, whatever
<anjuta> the cold outside isn't healthy for my brain
<AndroSyn> if you place a gline on the ip, it'll match it as well
<anjuta> any approximation for the time of the next stable release? just to make my own schedule whether or not to wait for it in order to push some additions to my patchset ?
<anfl> I wont be releasing 2.2.0 until after the new year
<anjuta> so if 2.1.8 comes out, it would be a bugfix release ?
<anfl> yes.
<anfl> though itd have to be a bloody serious bug :)
<anjuta> the trace bug was fun :)
<dougk_ff7> like "half the network goes boom" serious? :D
<anfl> half the network goes boom and you cant fix it with a module :)
<dougk_ff7> lol
<dougk_ff7> well, at least unlike the older ircds, it's not like there are no modules to fix with :D
<AndroSyn> there is very little that can't be fixed with a module if you're brave ;)
<dougk_ff7> this is true... and I love it for that reason :D
<dougk_ff7> the only exception might be if the bug is in modload/modunload itself....
<AndroSyn> well..the stuff in src generally cant be..
<AndroSyn> unless you're willing to run one of my hack jobs of scribbling over the function in question with shell code...
<dougk_ff7> lol
<anfl> andro: ive been debating moving bits of serv_estab() into m_server.c
<AndroSyn> hmm..
<dougk_ff7> I still liked it when I saw someone with a 700 day uptime or something on an ancient version of Linux.... and after a pretty nasty security exploit was released, he fixed it with a module :D
<AndroSyn> i'm not sure its worth it for 2.2
<anfl> aye
<AndroSyn> though you should be able to move server_estab() completely into m_server.c
<AndroSyn> without breaking anything
<anfl> you know
<anfl> I really dislike this unused stuff
<AndroSyn> heh
<anfl> I think -Wno-unused was nicer
<AndroSyn> well i think i'd rather just point out to it what we intentionally have unused..
<anfl> it looks so goddamned ugly tho
<AndroSyn> we can change the macro a bit if you like
<anfl> and I really really dont like setting UNUSED() on parc/parv
<AndroSyn> well it just tells gcc its possibly not used
<anfl> if its sole purpose is to be able to compile with -Wno-unused
<anfl> err, without
<anfl> id rather not bother :p
<anfl> hell, if anything gcc should optimise them itself anyway.
<AndroSyn> well having -Wunused in there, is a good thing though..
<AndroSyn> as it catchs crap that happens to get defined but then nobody uses it because they forgot about it
<anfl> id be happy periodically compiling with -Wunused to check
<AndroSyn> well my original version of the macro did #define UNUSED(x) UNUSED_ ## x __attribute__((unused))
<AndroSyn> which ended up mangling the name ;)
<anfl> I still think it achieves little except making the code horribly ugly
<anfl> its not often we'll get something itll pickup anyway
<anfl> I really dont like it ;-)
<AndroSyn> heh
<anfl> can I remove it? :)
<AndroSyn> sure :P
<anfl> besides
<anfl> you actually wrapped some unused stuff in UNUSED() when it should be removed :p
<anfl> void
<anfl> init_dlink_nodes(size_t UNUSED(dh_size))
<anfl> {
<anfl> dnode_heap = BlockHeapCreate(sizeof(dlink_node), dh_size);
<anfl> I dont even wanna ask about that one
<anfl> :p
<AndroSyn> -DNOBALLOC
<AndroSyn> :P
<AndroSyn> BlockHeapCreate() becomes dnode_heap = sizeof(dlink_node)
<AndroSyn> :P
<AndroSyn> i do actually use -DNOBALLOC combined with efence once in a while ;)
<anfl> surely its better to just define BlockHeapCreate() to nothing, and BlockHeapAlloc() to malloc() ? :)
Return to ratbox
or
Go to some related
logs:
"'static' should not be used on member functions d
|
|