| |
| |
| |
|
Page: 1 2
<Gangi> Hi, anyone know where I can find some documentation for setting up Ratbox services please?
<jilles> INSTALL / README in the source distribution
<jilles> it's not very complete
<Gangi> Ya, I discovered that.
<anfl> src/client.c:#ifdef HIDE_SPOOF_IPS
<anfl> src/client.c:#ifndef HIDE_SPOOF_IPS
<anfl> theyre the only three checks left now :)
<AndroSyn> nice
<anfl> theyre about to dissapear into a conf option
<anfl> [19:14 leeh@fortune:~/ircd-memcleanup/include]$ wc -l ~/ircd-ratbox/include/config.h config.h
<anfl> 162 /home/leeh/ircd-ratbox/include/config.h
<anfl> 134 config.h
<anfl> \o/ :)
<anfl> another 6 lines gone for an unused define :)
<anfl> im tempted to move everything else barring ts6 and file paths into defaults.h
<anfl> btw, you really need to review everything ive done for spoof ips/server ips
<anfl> do make sure I havent introduced a leak
<anfl> as its quite important I havent :)
<anjuta> will you make an effort to simplify the making of a debian package ? :)
<AndroSyn> yeah i'll take a look at it at some point
<AndroSyn> :P
<anfl> did we not get DESTDIR in 2.x?
<AndroSyn> i thought we had..
<anfl> its in the Makefile..
<anfl> though if by simplifying you mean altering the directory structure, the answer is no.
<anfl> :)
<anjuta> why not ?
<AndroSyn> one of the things a package like that really needs to do though is..
<anjuta> it's really bad to dump cores in /usr
<AndroSyn> when running, it needs to really chdir the ircd into a directory it can write..
<AndroSyn> yes..make it chdir to /var/run/ircd or something
<AndroSyn> then it'll dump cores into /var/run/ircd instead of /usr
<anfl> we chdir() to DPATH
<anfl> and changing the directory structure will do little more than piss people off
<AndroSyn> that is why dpkg lets you add patches :P
<anfl> im not doing it for the sake of "the unix way"
<anjuta> it will not piss off people that prefer a debian style install
<AndroSyn> it will piss people off if they get useless cores
<anfl> itll piss off everyone who currently has an ircd installed from source
<anjuta> the others can still go through ./configure;make;..
<AndroSyn> anjuta: the solution is simple..
<AndroSyn> just add a patch to your package
<AndroSyn> that twiddles the minor things you need twiddled
<AndroSyn> mostly you just need to fix the chdir() part
<anfl> personally, I prefer ircd being inside a single directory :)
<AndroSyn> i agree
<anjuta> can you think of anything else than the chdir() ?
<AndroSyn> what else do you need to twiddle?
<AndroSyn> everything else change be tweaked as it is
<AndroSyn> the chdir() thing fixes the core dump part
<anjuta> anfl, i agree too, but I can't fight people that are pissed of ircd's being installed not the right way on their debian mashine
<anfl> there is no "right" way.
<AndroSyn> s/right way/debian way/
<anjuta> especially when this mashine happens to be a primary debian mirror uh oh
<anfl> given debian will install the files itself
<anfl> you can set the paths fine to configure
<anfl> as you wont be running make install
<AndroSyn> anjuta: everything you need to tweak things is already there
<anfl> ill admit putting the pid in etc/ is kinda wrong
<anjuta> there was a problem with the help files.. /quote help
<anfl> but theres a lot of **** that depends on it
<anjuta> and the core dump.. that's all I know of..
<AndroSyn> ./configure --prefix=/usr --sysconfdir=/etc/ratbox --localstatedir=/var/run/ircd --libdir=/usr/lib/ratbox
<AndroSyn> or so..
<AndroSyn> --with-logdir=/var/log/ratbox --with-helpdir=/usr/share/ratbox/help --with-moduledir=/usr/lib/ratbox/modules
<anjuta> can the chdir() take the value of --localstatedir
<anjuta> ?
<AndroSyn> wouldn't be hard to make it so
<anjuta> i mean.. isn't it logical to make it default for you too ?
<anjuta> probably not..
<AndroSyn> if we force the localstatedir to be the same as prefix, sure
<anjuta> ok, i'll take a look at it later tonight
<AndroSyn> i'm not opposed to adding some autoconf-fu to make things a little easier
<AndroSyn> but the defaults are going to stay as they are
<AndroSyn> as that is one of the points of using autoconf is letting the sysadmin do all the weird **** he likes
<anfl> ok
<anfl> im gunna add /chantrace
<anfl> so you get an etrace of a channel, providing youre a member :)
<sjk> :)
<sjk> \o/
<anfl> I will add operspy for it too
<anfl> [20:58] --- User default testy4 root 192.168.0.6 192.168.0.6 :Da Cl0ne!
<anfl> [20:58] --- User default testy5 root 192.168.0.6 192.168.0.6 :Da Cl0ne!
<anfl> [20:58] --- Oper default anfl leeh 127.0.0.1 127.0.0.1 :who knows
<anfl> :)
<anfl> I might as well make the cl*** field the channel name actually
<anfl> all done :)
<anfl> you can have it for pte if you want :p
<sjk> Neato!
<anfl> http://svn.ircd-ratbox.org/svnroot/ircd-ratbox/branches/RATBOX_2_2/modules/m_etrace.c
<anfl> :)
<anfl> the new m_testmask.c is quite good too
<anfl> lets you do testmaskgecos :p
<anfl> and testmask n!u@h
<anfl> :)
<sjk> Awesome!
<anjuta> :>
<anfl> its nice being able to give urls for stuff thats in svn that work over the web ;-)
<anfl> I may remove the oper only restriction on it
<anfl> havent decided yet :)
<anfl> its a lot easier to grasp than who extensions though
<anfl> :p
<sjk> /chantrace could be useful to normal users as well yeah
<jilles> kinda like /who, except showing IP address instead of channel membership type
<sjk> yes
<anjuta> spoofs will be 255.255.255.255 ?
<sjk> rocks when sending abusemails, etc
<jilles> will be 0
<anjuta> great..
<sjk> too bad you cannot do /chantrace for foreign clients though :|
<jilles> you can
<anjuta> now this would be the easiest way to spot spoofs
<jilles> the 0 is inconsistent with other commands that hide IP addresses which show 255.255.255.255
<jilles> even with /etrace <nick>
<AndroSyn> with TS6, you'd get IP addresses for users as well..
<AndroSyn> foreign users that is
<sjk> AndroSyn: but not all servers use ts6 (I honestly cannot understand why this is not the case)
<AndroSyn> well the ones that don't are running ircd's that don't support it
<AndroSyn> meaning ircd-ratbox 1.x, hybrid 6.x & 7.0.x, and csircd
<sjk> Okay, the csircd ones I can understand
<sjk> but for crying out loud, ratbox 1.x, hybrid 6&7? Upgrade!
<AndroSyn> irc admins are lazy :P
<jilles> admittedly, code quality for 2.x hasn't always been that great
<anjuta> they sure are.. my rate is 1 migration every 5 days :) i need 3 more months to get everything to ratbox :)
<sjk> I mean, jebus. Takes less than an hour to set everything up from scratch
<AndroSyn> sjk: well for some servers its a case of, if it ain't broke, don't fix it
<anjuta> sjk, takes much days to set the things in your head
<anjuta> AndroSyn, make more coding mistakes :)
<AndroSyn> i'm not claiming that what they are running doesn't have bugs in it..
<AndroSyn> its just that they haven't had problems to date with what they are running
<sjk> AndroSyn: yeah - but still, it's not really a lot of work to upgrade.
<AndroSyn> thus feel no need to run the latest and greatest
<anjuta> cant the advantages of ts6 be a factor for the whole network ?
<sjk> They should see the benefits of using an up-to-date ircd, though.
<AndroSyn> like take for example, verio has been running ratbox 1.2-3 and has an uptime of 650+ days..
<jilles> with several of the servers I admin I just don't want to restart them if it's not necessary
<jilles> even though I know most users reconnect just fine without noticing much
<AndroSyn> sjk: well thats up to the admin to decide, unless there is a vote called to mandate upgrades..
<AndroSyn> and good luck with that ;)
<AndroSyn> all considering the size of the TS6 cluster on efnet has grown faster than i expected
<sjk> AndroSyn: hehe
<sjk> AndroSyn: I wonder how many admins would vote yes
<jilles> you probably need a more pressing reason to get rid of csircd
<AndroSyn> probably not many..
<jilles> and the other stuff
<sjk> Csircd isn't being developed any longer?
<jilles> comstud is gone
Return to ratbox
or
Go to some related
logs:
exlack for kids
ibm x41 gentoo
|
|