hi

or: mikeperry committed revision 21492 (/torflow/trunk/CircuitAnalysis/BuildTimes/CBT-Test):
or: Fix a miscount issue if we desync from the fuzzy threshhold
or: and not the strict one.

I have done a fresh install of the Mac OS X Tor bundle (stable version) but when I launch it for the first time it says that an instance of vidalia is already running. If I tell it to continue I get the same issue over and over and over again. If I tell it to quit I still get Vidalia relaunching each time. I am running OS X 10.6.2 on a Core 2 Duo Aluminum MacBook.
Is this an issue with Snow Leopard's OS? If so I can downgrade the system to OS 10.5.x

Hi. If you run firefox(no tor) at the same time you are running firefox(tor enabled) is this a concern for the security on the firefox(tor enabled)? I mean since the firefox(no tor) has lower security settings.

I think so long as they're running different profiles you should be ok

Why are governments so stupid and create Data Retention Laws

folks, please keep government discussion in #nottor

ioerror_kef: ping

BarkerJr: sir yes sir

or: Juliusz Chroboczek <jch@pps.jussieu.fr>: 2010-01-24 21:02:20 [polipo/master]: Update CHANGES.
or: Juliusz Chroboczek <jch@pps.jussieu.fr>: 2010-01-24 21:02:21 [polipo/master]: Remove dead code in discardObjects. From Honglei Junan.

hello?

ioerror_kef: ping

So the latest Apple Security Update apparently botched the SSL renegotiation? Any fix for this?
What flags do I have to recompile openssl with?

IIUC they rebuilt packages to address that

not yet

Weird. My packages are the latest and the sigs checkout fine but I still have a TLS error with renegotiating in Vidalia, jr_

yes, because apple broke openssl

Guiri: phobos knows best
Guiri: you could recompile it against openssl from macports

Isn't Vidalia bundled with a static OpenSSL though? Thanks phobos, I saw a few comments on the Tor Blog about Apple's security update.

https://bugs.torproject.org/flyspray/index.php?do=details&id=1225
that's the bug
vidalia on osx isn't shipped with static openssl
that's what we're working on

jr__: Recompile tor? Yeah I already have 0.9.8I_0 from macports. I guess my system defaults to the apple version though
(Action) reads bug

Guiri: on FreeBSD they updated the tor port to compile against openssl from ports
will need to do the same with os x

https://bugs.torproject.org/flyspray/index.php?do=details&id=1225&area=comments#3770
that's the current plan

Interesting. So as much as I love Vidalia as a standalone, if I install tor from macports it may work if the macport is updated?
Ah gotcha phobos
Sorry I'm a bit slow
So at this point I'll just recompile tor against my macports ssl and then drop the binary into the Vidalia.app package.. hopefullly

basically

git link for tor? Can't find it on the download page

git://git.torproject.org/

Good evening, can anyone shed light on the seriousness of the recent server compromise? i've read about it, but don't full understand the signifigance

Teal: did you see the blog post?

morning to everyone, i have a warning from libevent that i never see before : janv. 25 04:39:28.180 [Warning] Warning from libevent: Epoll MOD on fd 146 failed. Old events were 6; read change was 0; write change was 2.: Bad file descriptor
it is a bug or only a warn who i don't need to take care ?

SwissTorExit: I'm guessing nickm will be able to answer :)

hello Runa :P
yep, seem like :D

Hello. What would i need to enter into my [proxy] section in irssi's config file to have irssi connect through TOR? I tried using the socks5 address (localhost:9050) but that didn't work.
And I am not able to make sense of the torify howto page, yet.
:)

jazzencat: have you seen http://wiki.noreply.org/noreply/TheOnionRouter/TorifyHOWTO/IrcSilc#Irssi ?

yes, though i'm still in the dark.
socat TCP4-LISTEN:4242,fork SOCKS4A:localhost:foo.onion:6667,socksport=9050
is this the line i need to use to connect through tor? i'd replace foo.onion:6667 with my irc address of choice?

I'm not sure actually, I haven't tried it before
looks like it, though

:)

the documentation says that if you want to connect to an IRC server on, say, barbaz.com, port 6667, you use: socat TCP4-LISTEN:4242,fork SOCKS4A:localhost:barbaz.com:6667,socksport=9050

ok
so now i just have to go reactivate the use_proxy setting
I was kind of hoping someone could point me at a TOR for Dummies page.
because they give me a script as well, but not sure what to do with it
:S

who gave you a script?

the page
# ./socatchk remote-host remote-port
# crudely shutsdown socat (if running) and then restarts it for new host/port
orport=9050
orlisadr=localhost
[ "$(pidof socat >/dev/null 2>&1 && echo $?)" = 0 ] && kill $(pidof socat); [ "$(pidof socat && echo $?)" != 0 ] && socat TCP4-LISTEN:4242,fork SOCKS4A:$orlisadr:$1:$2,socksport=$orport &
that looks like a script to me
it's from the Torify page

next time, please link the url of the page instead of pasting long lines of code here :)

sorry

I'm not sure, though, I might have some time to look at it later today

ok.
i'll check in later
i thought i could just point my IRC client at the socks5 proxy and it would do the rest.. same way that it worked for the browser

I think what tor needs is nightly builds... continuous integration, etc

BarkerJr: what would nightly builds improve? I think the Tor Project is already pushing out new releases quite often :)

the diff between 0.2.2.6 and 0.2.2.7 is 464K, or 5% of the entire application
that's way too many changes to push out at once

BarkerJr: you could send an email to or-talk and ask why there aren't any nightly builds, and also write down why you think there should be :)
I think most people prefer a stable release than a nightly build anyways. If you want to hack on tor, you can always check out the git repo

I also think it's rather bad that most of this diff is changing the way we free memory, and that wasn't even mentioned in the changelog
well, I'm slowely reverting back to 0.2.2.6, cause I 0.2.2.7 crashes too often

but I thought you wanted nightly builds and bleeding edge? :)

right, cause then I'd be able to determine which commit caused it :)

BarkerJr: git bisect !

I also think tor should be able to be upgraded without restarting

BarkerJr: have you reported the bug?

I should. "Tor should run in dynamic load libraries"

:)
BarkerJr: I think that those who want to help and investigate bugs more closely should get the source from git and build it. Maybe this is something you'd like to do?

yeah, probably

great, do you know how to build from source?

yeah, I'd just have to figure out git
can't be much of a learning curve from cvs, though

it is a bit different from cvs and svn, but I don't think you'll run into any big problems :)
check out the documentation on the git website

I suppose I could do that to track down these crashes... I could checkout a snapshot from each day, then runit for 2 days

git bissect ?

BarkerJr: for the record: git-bisect - Find by binary search the change that introduced a bug

ah ok, thx

if there is crash related to memory, maybe valgrind or efence could help ?

BarkerJr: hahaah "can't be much of a learning curve from cvs,"

how hard can it be to checkout a branch? :P

BarkerJr: git clone git://urlhere

BarkerJr: mkdir tmp && cd tmp && git clone git://git.torproject.org/git/tor
:)

omg, there's not git in centos
anyways, gtg, bbl

or: phobos committed revision 21493 (/projects/articles): convert the article into a useful format for me.

hi there!

hai

i'm trying to setup a tor exit node but keep saying "We have no network-status consensus"

what operating system are you on?

gentoo

ok. Can you give me the full log output, please? http://paste.debian.net/ for example.

ok, just a sec
http://paste.debian.net/57615/

ah
your Tor version is old
can you please upgrade to 0.2.1.22?
You're trying to bootstrap from a dirauth that is currently down

ok... I took the stable version (gentoo speaking)

hrm
gentoo should upgrade
please file a gentoo bug

I'll try to update that, should not take long
(Action) installing 0.2.1.22...
bootstrapped 100%
thanks
it seems there is a new tor exit node :)

hah
aweseome
awesome*
let us know if we can help you with more questions

thanks a lot for that. I'll do my homework a bit better next time... ;)

Exit nodes have increased a bit since I started monitoring them: http://noc.nix.is/nix.is/v.nix.is/tor_routers.html

interesting link

avar: neat. You haven't monitored them for very long, though.

nope:)

avar: do you know about archive.torproject.org?
specifically http://archive.torproject.org/tor-directory-authority-archive/

avar: you might be interested in metrics.torproject.org

hah, it's a karsten

ah, neat-o

avar: it has a data page that links to archived directory data.

yeah, like this: http://metrics.torproject.org/consensus-graphs.html

avar: yes. there are only very few graphs so far, though.
i figured exit nodes are most interesting. though there's data on the other flags (as you have graphed them), too.

I see on http://torstatus.kgprog.com that some tor nodes have ORPort set to 443 or even 80...

the whole internet is on 80 ;-)

is it to make it harder to filter tor usage?

yep, mostly

the 8 and the 0 are closeby on my keyboard
so are the 4 and the 3

some ISPs throttle everything but port 80, too

hence 80 and 443

ok. Thus that's a good idea to do that on my node
bye all

is there any fix for this already, like "use latest git" or so? https://bugs.torproject.org/flyspray/index.php?do=details&id=1208

Tas: don't think so
are you experiencing the same issues?
rieo seemed to have some idea

if you still have backuped bridge-stats can you gzip it and send to devs?

yes, on my bridge, on FreeBSD. sdoesn't start anymore. not sure if it's related to a recent OpenSSL update there

no, it is not
please back up your data directory

I think they just added some patch, no change in version number

specifically the bridge stats

yep, I did that already

and send it to tor-assistants@torproject.org

oh, didn't do that yet :-)
without keys, rigfht?

also mention bug 1208 and rieo's comment here, please.

right*

Yes, definitely
no keys, no state file
only the bridge stats

what do we think is wrong with the bridge stats?

err, so only the stats dir? or also cached* stuff?

Tas: not the cached* stuff

memory corruption can breaks something during write it to disk, and during load not all sanitized

karsten: there seem to be parsing errors

bridge stats is only 2 lines or so

like the end of file can be \0\n

Tas: please still send it, or attach it to bug 1208
karsten: there's nothing secret in there, right?

ok

Tas: please gzipped, so we won't corrupt any \0

yep, sure

better do more debug log_warn if bug triggered, without guessing what happens.
I mean print all string.

hm, no zero in there though

rieo: you mean in routerparse.c?

in extrainfo_dump_to_string()