-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Hi s7r!
On 27/03/16 11:30, s7r wrote:
Hi Karsten,
Does the bridge with hashed fingerprint 678912ABD7398DF8EFC8FA2BC7DEF610710360C4 fulfill the requirements you are looking for? It appears to me it handles more clients than antirio bridge, yet I ma unsure about the IPv4 and IPv6 distribution (don't know how you count them) - if this is ok please let me know exactly what you want me to do and how to check the statistics you mention, and i'll revert with a complete report.
It looks like that bridge doesn't see any IPv6 clients:
bridge-stats-end 2016-03-24 17:02:15 (86400 s) bridge-ips cn=120,us=88,ir=56,ru=48,gb=32,tw=32,[...] bridge-ip-versions v4=528,v6=0 bridge-ip-transports <OR>=16,obfs3=120,obfs4=400
See the bridge-ip-versions line with v6=0.
Or are you seeing actual IPv6 clients on that bridge that are not reported there? In theory, that statistic should be unaffected by this bug.
Note that even if there are a few dozen IPv6 clients, I'd rather want to test this patch on a bridge that has at least as many IPv6 clients as IPv4 clients, because that will tell us whether the patch works or not. That's why antirio would be the perfect guinea pig.
All the best, Karsten
Is it just building Tor from your `task-18460-2` and starting Tor with the same bridge identity?
P.S. this bridge is already running 0.2.8.1-alpha git 1f679d4ae11cd976+26ab2e0 - I assume your patch isn't merged already?
-s7r
On 3/27/2016 10:11 AM, Karsten Loesing wrote:
Hi everyone,
does anybody here know who operates the bridge with nickname antirio?
https://atlas.torproject.org/#details/16609212922F6F1077A1BBA299709E19F9A3FB...
I'm asking, because that bridge has a nice distribution of IPv4 and
IPv6 clients:
bridge-ip-versions v4=48,v6=64 bridge-ip-versions v4=56,v6=72
We have a bug where IPv6 addresses are included in bridge-ip-versions statistics but where consensus downloads via IPv6 addresses are not counted. It would be very valuable to test the patch for this bug on the antirio bridge or on another bridge with at least 1/2 of clients connecting via IPv6. (When looking yesterday I didn't find another bridge with that property.)
The patch is commit b79d859 in my task-18460-2 branch:
https://gitweb.torproject.org/karsten/tor.git/log/?h=task-18460-2
In theory it should be sufficient to cherry-pick that commit from any
other recent tor branch. It just changes three lines of code.
More details on the ticket:
https://trac.torproject.org/projects/tor/ticket/18460
Thanks!
All the best, Karsten