Re: GR proposal: code of conduct
On 2014-02-24, Ian Jackson ijack...@chiark.greenend.org.uk wrote: I don't think this is realistic for channels which anyone in the world can join. There are no doubt many people who have private logs and there would be nothing stopping anyone making such a log public without our consent. This is true for any electronic communication. People will probablybe asked to leave the forum if they insist on making logs public. And by asked, I mean forced. Is it really the case that making the logs available as public text files produces too much search engine exposure etc. (which is I guess the real concern) ? Yes. Would you want your chatter in the pub with friends recorded and published? /Sune -- To UNSUBSCRIBE, email to debian-project-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/lehm34$e08$1...@ger.gmane.org
Re: GR proposal: code of conduct
On Mon, Feb 24, 2014 at 05:00:07PM +0800, Paul Wise wrote: For IRC it's a bit more difficult, because we do not long our IRC channels by default (or at least I'm not aware we do), with the exception of meetings run with the help of meetbot. ... i.e. publicly log our IRC channels. That would be nice, the IRC channels are currently a big back-channel that hides a bunch of useful information from the wider public. One could argue that if there is information that is so useful it should be available to the general public then it should be manually polished up and published in designated places (documentation). -- WBR, wRAR -- To UNSUBSCRIBE, email to debian-project-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/20140225093420.ga21...@belkar.wrar.name
Re: State of the debian keyring
On Mon, Feb 24, 2014 at 06:51:37PM -0800, Russ Allbery wrote: Brute-forcing the key just requires compute cycles. There is essentially no chance of discovery and no risky activity at all until you start actually using the key. ...which reminds me of http://www.enricozini.org/2008/tips/audit-uploads/ which was a prototype of creating an audit log of key usage in debian. However, for the audit log to be usable as an audit log without giving a false sense of security, it should be complete, and really cover all instances of key usage in Debian. This means hooking into any place where a signature verification or a decryption actually happens in Debian: I can think of uploads, db.debian.org, voting, keyring requests, RT tickets filed, emails received by lists or the BTS: are there more? I see the job as not so much technically complex[1] as socially complex: since I would not trust auditing an incomplete audit log, I fear that a missing or badly implemented data source could invalidate all the system. So I can't just open vim and write the code: auditing key usage in package uploads requires someone who knows dak inside out, and can commit to maintaining notification triggers in all obscure corners where keys are used, now and in future updates of the ftp-master toolchain. Same goes for any other bit of Debian. The starting point for this work is probably this, then: is it just me, feeling that we have a problem here, or am I actually in the good company of people who can do their bit? Ciao, Enrico [1] For realtime auditing, we now have a rabbitmq server. Or collection could be decoupled in one audit log per team, which are then aggregated by a separate project. Or they can be submitted to a central collection point, like a new ad-hoc bit of contributors.debian.org. I don't see anything technically difficult here. -- GPG key: 4096R/E7AD5568 2009-05-08 Enrico Zini enr...@enricozini.org signature.asc Description: Digital signature
Re: GR proposal: code of conduct
Hi, Andrey Rahmatullin: One could argue that if there is information that is so useful it should be available to the general public then it should be manually polished up and published in designated places (documentation). One could argue that the information / documentation is already readily available at mulitple places, but the person who asks in the channel is too inexperienced / lazy / stupid ^w clueless to actually find it. We all have been guilty of all three of these, at one time or another. -- -- Matthias Urlichs signature.asc Description: Digital signature
Re: GR proposal: code of conduct
On Tue, Feb 25, 2014 at 2:01 AM, Ian Jackson wrote: Is it really the case that making the logs available as public text files produces too much search engine exposure etc. (which is I guess the real concern) ? Several of our derivatives (at least Maemo, Ubuntu) have public logs of their IRC channels. Personally I think it would bring some much needed transparency to what is becoming one of the more essential Debian communication channels to be on. Just like we archive mailing lists and record DebConf talks/BoFs, we should publicly log IRC channels. -- bye, pabs http://wiki.debian.org/PaulWise -- To UNSUBSCRIBE, email to debian-project-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/caktje6forwltdpj7z-wtzfarxqox0ojz7628kbwhqnwuqo3...@mail.gmail.com
Re: State of the debian keyring
On Tue, Feb 25, 2014 at 5:47 PM, Enrico Zini wrote: This means hooking into any place where a signature verification or a decryption actually happens in Debian: I can think of uploads, db.debian.org, voting, keyring requests, RT tickets filed, emails received by lists or the BTS: are there more? I didn't think the BTS cared about OpenPGP keys? mentors.d.n does do signature verification of uploads. -- bye, pabs http://wiki.debian.org/PaulWise -- To UNSUBSCRIBE, email to debian-project-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/CAKTje6HH4ZfhGyUX3BxuXx7f1w=f9qbrx9uo_enhslct5u0...@mail.gmail.com
Re: State of the debian keyring
Ian Jackson dijo [Mon, Feb 24, 2014 at 05:53:58PM +]: Are we now at the stage where it is more important to retire these shortish keys, than to insist on this cross-signatures ? I.e., perhaps it would be better to invite key rollover from a short key to a long one despite the lack of 2 other DD signatures; or perhaps even despite the lack of _any_ other DD signatures. Instead, the keyholder could perhaps present a signed key transition document. A downside is that we would probably have to keep the rolled-over short keys somewhere, at least to maintain the integrity of our records of why a key is in the keyring. Which we do anyway - All retired keys are still in our tree, in the removed-keys-{pgp,gpg} directories (plus the emeritys-keyring-{gpg,pgp}). Of course, they are not installed when you get the generated package (you only get the active keyrings). But they are all there. -- To UNSUBSCRIBE, email to debian-project-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/20140225184458.gh40...@gwolf.org
Re: State of the debian keyring
Ian Jackson dijo [Mon, Feb 24, 2014 at 05:57:57PM +]: I think this is a bug. It can increase security because it can make operations more convenient at the same level of security, and because people trade off convenience for security. For example, it would be possible to have one key for email encryption and a different (more secure) key for package uploads. Debian tools don't care which key you use for email encryption. The extent of actions you interact with debian is easily modeled with a single key; for some time I used to upload with 1024D and sign mails with 4096R because I had not yet pushed my 4096R into the keyring, waiting to get more signatures (yes, also being keyring-maint it took me some time to push it, even if I had all power to do so myself!) -- To UNSUBSCRIBE, email to debian-project-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/20140225184724.gi40...@gwolf.org
Re: State of the debian keyring
Gunnar Wolf gw...@gwolf.org writes: Ian Jackson dijo [Mon, Feb 24, 2014 at 05:57:57PM +]: I think this is a bug. It can increase security because it can make operations more convenient at the same level of security, and because people trade off convenience for security. For example, it would be possible to have one key for email encryption and a different (more secure) key for package uploads. Debian tools don't care which key you use for email encryption. Except for project DPL votes, no? The extent of actions you interact with debian is easily modeled with a single key; for some time I used to upload with 1024D and sign mails with 4096R because I had not yet pushed my 4096R into the keyring, waiting to get more signatures (yes, also being keyring-maint it took me some time to push it, even if I had all power to do so myself!) For email signatures, don't quite a few more things care? All votes, db.debian.org operations, etc. -- Russ Allbery (r...@debian.org) http://www.eyrie.org/~eagle/ -- To UNSUBSCRIBE, email to debian-project-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/87ios3auur@windlord.stanford.edu
Re: GR proposal: code of conduct
On Tue, Feb 25, 2014 at 06:28:39PM +0800, Paul Wise wrote: On Tue, Feb 25, 2014 at 2:01 AM, Ian Jackson wrote: Is it really the case that making the logs available as public text files produces too much search engine exposure etc. (which is I guess the real concern) ? Several of our derivatives (at least Maemo, Ubuntu) have public logs of their IRC channels. Personally I think it would bring some much needed transparency to what is becoming one of the more essential Debian communication channels to be on. Just like we archive mailing lists and record DebConf talks/BoFs, we should publicly log IRC channels. I am generally in favour of more transparency. Logging official project IRC channels would fit well with that. However, I find that it's very difficult to extract useful information from voluminous IRC logs, and official channels are likely to be voluminous. The logs are hard to read, and there's so much irrelevant discussion mixed with the parts that one is looking for that it is much harder to find what one wants. IRC has no threading, so finding the related parts of a discussion is not easy. This is a stark contrast with, say, mailing lists. Thus I suspect that the logs won't be very useful. I would prefer a culture where IRC discussions are ephemeral, and any useful information should end up in debian/changelog, mailing lists, git commit messages, wiki.debian.org, or any of the other places where we already put information. -- http://www.cafepress.com/trunktees -- geeky funny T-shirts http://gtdfh.branchable.com/ -- GTD for hackers -- To UNSUBSCRIBE, email to debian-project-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/20140225190249.GF4722@holywood
Re: State of the debian keyring
On Tue, 25 Feb 2014, Paul Wise wrote: I didn't think the BTS cared about OpenPGP keys? We probably will eventually, but we only use[1] them now to help whitelist mail. 1: By which I mean that if a message seems to have a PGP signature, we think it's probably not spam; we currently don't even bother to check it. -- Don Armstrong http://www.donarmstrong.com Of course, there are cases where only a rare individual will have the vision to perceive a system which governs many people's lives; a system which had never before even been recognized as a system; then such people often devote their lives to convincing other people that the system really is there and that it aught to be exited from. -- Douglas R. Hofstadter _Gödel Escher Bach. Eternal Golden Braid_ -- To UNSUBSCRIBE, email to debian-project-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/20140225190806.gc10...@teltox.donarmstrong.com
Re: GR proposal: code of conduct
Hi, Lars Wirzenius: I would prefer a culture where IRC discussions are ephemeral, and any useful information should end up in debian/changelog, mailing lists, git commit messages, wiki.debian.org, or any of the other places where we already put information. I agree. The second problem I have with IRC logs is that Google is likely to show them if you search for the solution to some problem, partticularly if that problem is asked about often. However, I suspect that most users would extracting a solution from IRC logs, assuming it is buried in there at all, to be very tedious. -- -- Matthias Urlichs -- To UNSUBSCRIBE, email to debian-project-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/20140225200136.ga24...@smurf.noris.de
Re: State of the debian keyring
On Tue, Feb 25, 2014 at 10:51:56AM -0800, Russ Allbery wrote: Gunnar Wolf gw...@gwolf.org writes: Ian Jackson dijo [Mon, Feb 24, 2014 at 05:57:57PM +]: I think this is a bug. It can increase security because it can make operations more convenient at the same level of security, and because people trade off convenience for security. For example, it would be possible to have one key for email encryption and a different (more secure) key for package uploads. Debian tools don't care which key you use for email encryption. Except for project DPL votes, no? I think the keys are used for voting and the email interfance for db.debian.org. I'm not sure if we have any other services checking the gpg signatures of emails. Kurt -- To UNSUBSCRIBE, email to debian-project-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/20140225201000.ga5...@roeckx.be
Re: State of the debian keyring
On Tue, Feb 25, 2014 at 10:51:56AM -0800, Russ Allbery wrote: Gunnar Wolf gw...@gwolf.org writes: Ian Jackson dijo [Mon, Feb 24, 2014 at 05:57:57PM +]: I think this is a bug. It can increase security because it can make operations more convenient at the same level of security, and because people trade off convenience for security. For example, it would be possible to have one key for email encryption and a different (more secure) key for package uploads. ... For email signatures, don't quite a few more things care? All votes, db.debian.org operations, etc. More relevantly an email signature isn't any different to a signature for a package upload, so DDs would have to specify what the use for each key was, keyring-maint would have to maintain appropriate keyrings indicating what the expected use of a key was, and all the project facilities that make use of signatures would have to make decisions about which keyring they were using. (Yes, for encryption that's a different situation but the only example I can think of where the project uses encryption to a key in the keyring at present is the initial account password / a password reset. And for an encryption/signing split subkeys should be able to handle the desired outcome, I think.) J. -- xmpp:nood...@earth.li Time is an illusion. Lunchtime doubly so. -- To UNSUBSCRIBE, email to debian-project-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/20140225201243.go27...@earth.li
Re: State of the debian keyring
On Tue, Feb 25, 2014 at 09:10:00PM +0100, Kurt Roeckx wrote: On Tue, Feb 25, 2014 at 10:51:56AM -0800, Russ Allbery wrote: Gunnar Wolf gw...@gwolf.org writes: Ian Jackson dijo [Mon, Feb 24, 2014 at 05:57:57PM +]: I think this is a bug. It can increase security because it can make operations more convenient at the same level of security, and because people trade off convenience for security. For example, it would be possible to have one key for email encryption and a different (more secure) key for package uploads. Debian tools don't care which key you use for email encryption. Except for project DPL votes, no? I think the keys are used for voting and the email interfance for db.debian.org. I'm not sure if we have any other services checking the gpg signatures of emails. echelon checks the keyring, also. -- Luca Filipozzi http://www.crowdrise.com/SupportDebian signature.asc Description: Digital signature
Re: GR proposal: code of conduct
On Wed, Feb 26, 2014 at 3:02 AM, Lars Wirzenius wrote: Thus I suspect that the logs won't be very useful. Due to Debian being focussed in the European timezones, most of my use of IRC is reading backlog, which is pretty much the same has reading public logs. I still find IRC useful and even essential to be reading. I've also found the public IRC logs of other distributions useful in the past when I wanted to find out what was going on. I've also extracted useful information from public IRC logs I found via search engines. useful information should end up in debian/changelog, mailing lists, git commit messages, wiki.debian.org, or any of the other places where we already put information. I don't think that is happening right now. It might be possible to do this but I expect any effort to do so will end like the debian-private declassification GR; with no-one to doing it. -- bye, pabs http://wiki.debian.org/PaulWise -- To UNSUBSCRIBE, email to debian-project-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/CAKTje6ER8Nvx2s++4DrdT7cRJWH3hNFmTn3sSV=nz0g7dtr...@mail.gmail.com