IRC Logs

Network: freenode Channel: #gentoo-dev

	Enter your search terms Submit search form
	Web www.irclog.org

[03:26]<vsugrd>bye bye dsd_, not unless you have more info about stuff
[03:37]<vsugrd>g2boojum: *poke* how's the financial report coming/dealing with paypal?
[03:48]<zjqqrvq>Flameeyes, pong
[03:48]<ffrdnndns>robbat2, i submitted you a bug, as i thought i was going to sleep :)
[03:49]<ffrdnndns>[i remained only because i had a spree of --as-needed fixes and i wanted to comment on them :P]
[03:51]<vzunmpw>Hi Flameeyes.
[03:51]<ffrdnndns>hi truedfx
[03:51]<ffrdnndns>truedfx, thanks for the pointers on the comment :) they useful
[03:51]<ffrdnndns>*very
[03:51]<zgvzzus>w 5
[03:51]<zgvzzus>gah
[03:51]<ffrdnndns>i'm half asleep if you couldn't tell ;)
[03:51]<vzunmpw>np, and it's about 3AM, so you should be :)
[03:56]<sjfzz>robbat2: you work with rh based distros much?
[03:56]<sjfzz>worked
[03:56]<zjqqrvq>I've done a fair bit in the past yes
[03:56]<zjqqrvq>incl beating their RPM signing stuff to death a few times
[03:57]<sjfzz>from my pov.. here is the simple goal. A user should do the equiv of rpm --import /path/to.master.key 1 time.
[03:57]<sjfzz>and from then on out thats valid for everything. no user interaction after that
[03:57]<ffrdnndns>robbat2, i have to say, the udev rules doesn't work for me at all :/
[03:57]<sjfzz>cept in the case of us getting owned.
[03:57]<ffrdnndns>i launched udevstart at least a couple of times and nothing changed on the devices
[03:57]<sjfzz>which hopefully wont ever happen while we are around
[03:58]<zjqqrvq>solar: there's cases where that falls over, even outside of the owned case
[03:58]<sjfzz>robbat2: is that doable. To not make the end user have to maintain anything ?
[03:58]<zjqqrvq>Flameeyes, could you run udevtest and see what's happening?
[03:58]<ffrdnndns>robbat2, i don't know much about udev so i wouldn't be able to tell even if i'm awake :P
[03:59]<zjqqrvq>solar, yes, it is possible, the question is just doing it with the problems in existing systems
[03:59]<zjqqrvq>the RPM signing is basically the single repo key model
[03:59]<ffrdnndns>can you anyway provide those files if udev is not installed? it's way simpler for me to implement a freebsd-hotplug script rather than a freebsd-udev
[04:00]<zjqqrvq>Flameeyes, submit a different patch to that bug and reopen then
[04:00]<ffrdnndns>will do tomorrow then, i do have to sleep now :)
[04:00]<zjqqrvq>Flameeyes, ok
[04:00]<ffrdnndns>thanks anyway :)
[04:00]<ffrdnndns>night!
[04:01]<zjqqrvq>solar, for key distribution, we'd be closer to the debian model of distributing the master key/keyring as a package
[04:01]<zjqqrvq>that was part of the core system
[04:02]<sjfzz>robbat2: way I se it is to abstract the signing away from the devs also. That guy was spot on with the (either we trust the cvs server or we dont)
[04:03]<zjqqrvq>there's a part of the RH model that's often overlooked, and it is relevant here
[04:03]<sjfzz>robbat2: but really. I'm willing to follow your lead on this. And so is the council. So it's just a matter of when you have some free time
[04:03]<zjqqrvq>their master key is used from repo -> user
[04:04]<zjqqrvq>their developers have keys that they use to submit to the repo
[04:04]<zjqqrvq>debian's model is similar, but just skips the resigning with the master key
[04:04]<zjqqrvq>the debian model is outright scary in some ways
[04:04]<sjfzz>and before my term is up on this run of the council I'd really love to feel like something has been accomplished. And doing just this would be that warn fuzzy.
[04:05]<sjfzz>so if there is anything I can do to help
[04:05]<zjqqrvq>there's an anonymous FTP upload location - and you just upload signed packages to it
[04:05]<rggjgg>fun
[04:06]<rggjgg>sounds similar to the SF model
[04:06]<zjqqrvq>yup
[04:07]<zjqqrvq>solar: thanks, i'll remember that
[04:11]<sjfzz>robbat2: do we need to put a team together for this task?
[04:59]<zjqqrvq>everybody using gpg-agent is using the Assuan stuff already, without realizing it
[05:00]<zjqqrvq>Assuan is the internal language that gpg-agent and gpg use to chat
[05:00]<przzyrr2>robbat2: any bindings folks can use?
[05:00]<przzyrr2>even a c api would do (can just wrap it)
[05:02]<zjqqrvq>the closest thing there is to a usable API right now is gpgme - which has major suckage
[05:03]<zjqqrvq>gpgme talks Assuan on it's outside, but still calls gnupg via execve on the inside
[05:04]<przzyrr2>bleh.
[05:04]<przzyrr2>robbat2: better question- what's required to dodge that hit?
[05:04]<gnzm2jd>i thought it was "metric ass-load"
[05:05]<zjqqrvq>ferringb, bugging mr. koch and the others to get the rest of the API out of SVN
[05:05]<zjqqrvq>incidently, I need to remember to get on that
[05:05]<zjqqrvq>i bugged them 6 months ago about it
[05:06]<zjqqrvq>and haven't been active on that list since then
[05:06]<zjqqrvq>I know it's completed for the S/MIME variant of gnupg
[05:07]<zjqqrvq>just not the OpenPGP variant
[05:07]<zjqqrvq>"gpgsm --server" will get you the S/MIME variant right now
[05:08]<zjqqrvq>digging in some of the current stuff, I see the signing portion has made it to the latest release
[05:11]<zjqqrvq>to see some of the current docs if you have gpg-1.9*, run 'info gpg' and search for "GPGSM SIGN"
[05:11]<rggjgg>robbat2: btw, you still doing mysql?
[05:12]<zjqqrvq>genone: minor stuff on mysql yes, or when vivo is missing and something needs to be touched urgently
[05:12]<rggjgg>was wondering why the current ebuilds unconditionally delete the info files
[05:13]<rggjgg>I find the quite handy
[05:13]<przzyrr2>robbat2: well... api available/required I'd be interested in
[05:13]<przzyrr2>robbat2: need to do a reimplementation of signed tree near future, would be nice to not have to rewrite it ;)
[05:14]<zjqqrvq>look at the gpgsm docs, the OpenPGP variant should be very similar to them
[05:56]<mrrr>average karma: -808.58
[05:57]<mrrr>Got you beat by a lot. :)
[05:57]<vsugrd>wow ouch
[05:58]<mrrr>Yeah. Recently, they've been having ~80 bad urls per comment. Pretty bad.
[05:58]<mrrr>4th most recent caught one had -1782
[05:59]<mrrr>-2188
[05:59]<mrrr>Okay, that takes the cake: -48993
[05:59]<zgzzdgaz>agaffney: you got mail 8-)
[05:59]<vsugrd>dang: thats horrible, I've had 20-30 per mostly
[05:59]<mrrr>303 URLs in that one.
[06:09]<cxzyswxyvn>spin me!
[06:16]<cxzyswxyvn>weee
[06:16]<cxzyswxyvn>vanilla jello!
[06:16]<cxzyswxyvn>so yah, I'm going to Japan on the 28th
[06:17]<przzyrr2>robbat2: need it to be faster single check though.
[06:18]<przzyrr2>robbat2: mass verification, if faster then anything we can come up with for single checks, yes, useful for the tree
[06:18]<cxzyswxyvn>hmm, maybe I can hunt jstubbs down
[06:18]<przzyrr2>robbat2: main interest atm for me is faster repo checks as it's going
[06:18]<cxzyswxyvn>steev-laptop: btw, email me the model of your unit so I can track a battery down for you
[06:19]<svnnv-frlvjl>ChrisWhite: awesome - will do - im not at my place, and won't be till tomorrow so yeah
[06:19]<gnzm2jd>ChrisWhite: did wormo invite you to her sblug meetings?

Back to #gentoo-dev
Or go to some related logs:
#gentoo
#gentoo
#debian
#gentoo
#gentoo