[00:15] <\u03b5> is there a way to mark releases as testing?
[00:17] <wgrant> \u03b5: No :(
[00:18] <NakedNous> Hi
[00:18] <NakedNous> I have never used freenode
[00:18] <\u03b5> any way to mark one as current stable then?
[00:19] <NakedNous> can anyobe see me?
[00:19] <\u03b5> NakedNous, yes, Hi
[00:19] <NakedNous> say great! Hi all!
[00:19] <NakedNous> say sorry for my English, it is not my first language
[00:20] <\u03b5> you don't need to write "say" in front of everything :)
[00:20] <NakedNous> I'm a packager newbie and it seems I lost my private key when upgrading to karmic
[00:20] <NakedNous> my packages are located at: https://launchpad.net/~nakednous/+archive/ppa
[00:21] <NakedNous> and I want to upgrade them to karmic
[00:21] <\u03b5> you need to make a new key then
[00:21] <\u03b5> then set it up in launchpad
[00:21] <\u03b5> then sign the repo with your new key
[00:22] <NakedNous> is it the only way?
[00:22] <\u03b5> hopefully it is :)
[00:22] <NakedNous> I don't want the public part of the kay to be changed
[00:23] <\u03b5> then you have to do the math yourself
[00:23] <\u03b5> why is that important?
[00:23] <NakedNous> I thought the private key was also store somewhere at the ppa server
[00:24] <NakedNous> I don't want people using my software to sign it again
[00:24] <wgrant> So, there are two keys involved in PPA.
[00:24] <wgrant> *PPAs
[00:24] <wgrant> One is the one where *only* Launchpad has the private key. that's the one that your users see.
[00:24] <wgrant> And one where only you have the private key.
[00:25] <wgrant> That's the one you use to uploda.
[00:25] <NakedNous> I think so
[00:25] <wgrant> Nobody cares about that once you've uplodaed.
[00:25] <wgrant> Wow, I cannot type today.
[00:26] <NakedNous> do mean that the public part will remain unchanged?
[00:26] <NakedNous> the one my users see?
[00:26] <wgrant> Right.
[00:26] <wgrant> You never had the private key for the key that users see.
[00:26] <NakedNous> thats great!
[00:27] <NakedNous> so I need to regenerate it, but what if I need it at another computer in the future?
[00:27] <NakedNous> I will check how to retrieve it
[00:28] <wgrant> You can copy your keys (they live in ~/.gnupg) between machines. But guard them well.
[00:28] <NakedNous> I see, need to study
[00:29] <NakedNous> finally, could you please point me the url where to generate the kyes, I forgot it
[00:30] <wgrant> NakedNous: See https://launchpad.net/people/+me/+editpgpkeys
[00:31] <NakedNous> say thank you all!
[00:31] <NakedNous> have a nice day
[00:31] <wgrant> You too, NakedNous.
[00:39] <DBO> bzr is the slow, I assume this is known so just popping in to ensure this is the case
[00:40] <wgrant> DBO: What operation is slow?
[00:40] <dtchen> bzr itself, or pushing to somewhere hosted by Canonical?
[00:40] <DBO> someone pushed a new branch, and nobody can access it yet
[00:40] <DBO> been about half an hour
[00:40] <DBO> dtchen!
[00:40] <DBO> wanna know something amusing?
[00:40] <dtchen> not really.
[00:40] <DBO> my speakers came back to life
[00:40] <DBO> for no friggin reason
[00:41] <DBO> about 2 months ago, out of the blue, been working ever sense
[00:41] <dtchen> congrats.
[00:41] <DBO> also I work for Canonical now :)
[00:41] <dtchen> excellent.
[00:42] <wgrant> The branch scanner appears to be doing bad things.
[00:42] <wgrant> (ie. nothing)
[00:42] <DBO> awesome
[00:42] <wgrant> mthaddon: ^^?
[00:43] <wgrant> Hm. Odd that no yellow boxes are showing up.
[00:46] <wgrant> Hmm. I wonder if it's related to the large number of new branches that have been created in the past 12 or so hours.
[00:46] <wgrant> (due to Lucid)
[00:55] <wgrant> rockstar: You're not still around?
[00:56] <rockstar> wgrant, I am.
[00:57] <wgrant> rockstar: What is going on with the branch scanner?
[00:57] <rockstar> wgrant, I didn't know anything was going on with it.  One sec.
[00:57] <wgrant> Just busy with the 16000 new branches?
[00:57] <rockstar> wgrant, oh yes, that's VERY likely.
[00:58] <wgrant> Might want to dent and topic it, unless it's finishing soon (I presume you have pretty graphs)
[01:03] <rockstar> wgrant, investigating it.
[01:46] <mwhudson> yes, branch updates seem clogged by the 16000 new branches
[01:46] <mwhudson> probably requesting them all be mirrored in one hit isn't so smart
[01:46] <wgrant> Probably not. How's it going?
[01:47] <mwhudson> wgrant: probably 25% done
[01:47] <wgrant> mwhudson: Eeep.
[01:47] <mwhudson> yeah :(
[01:47] <mwhudson> after 3 hours of progress
[01:47] <mwhudson> not so great
[01:47] <wgrant> Still, better than the old days when LP had to be taken down entirely for 12 hours for a new release's *translations*.
[01:52] <wgrant> Adding it to the topic and denting it seems like a very good idea at this point.
[01:52]  * mwhudson dents
[01:54] <wgrant> mwhudson: Thanks.
[01:55] <exarkun> 16000 is a lot of branches isn't it
[01:55] <wgrant> mwhudson: Now, remember to log back in as your normal user, or we'll have a repeat of mrevell-as-Launchpad.net.
[01:56] <mwhudson> wgrant: heh, i dented from a client that supports multiple accounts (on my phone, ffs)
[01:57] <mwhudson> i guess i should finally set up gwibber on my machine
[01:57] <lifeless> exarkun: its a measurable fraction of our total
[01:57] <wgrant> mwhudson: Gwibber 2.0 really sucks.
[01:57] <exarkun> I guess the lucid developers are really prolific?
[01:57] <lifeless> exarkun: its one branch per package
[01:57] <wgrant> exarkun: They're copies of all the old Karmic branches.
[01:57] <lifeless> more or less
[01:57] <wgrant> One for each source package.
[01:58] <exarkun> Oh
[01:58] <lifeless> exarkun: Ubuntu as a whole doesn't have 'trunk', it has one 'branch' per release, and each 'branch' is > 16000 real branches
[01:58] <exarkun> You're not talking about an editor for gnome, you're talking about karmic+1
[01:58] <wgrant> exarkun: Right.
[01:58] <exarkun> (lucid is an editor for gnome)
[01:58] <wgrant> Ahh
[01:59] <exarkun> that makes sense then I guess
[02:34] <RenatoSilva> any problem with the branch ui updates in LP?
[02:34] <RenatoSilva> pushed 2 revisions 10 minutes ago and nothing yet
[02:45] <wgrant> RenatoSilva: There are long delays - see the end of the topic.
[02:46] <RenatoSilva>  /topic does not return anything
[02:46] <RenatoSilva> but ok
[02:56] <lifeless> abentley: hi, do you remember the error you had when you tried --parallel?
[02:57] <abentley> lifeless: IIRC, subunit wasn't packaged in ubuntu and easy_install didn't work.
[02:58] <lifeless> ok; thats been covered in the thread - subunit uses autoconf (because its many-languages)
[02:58] <lifeless> I'm not sure how to many something that uses autoconf easy_install-able
[03:04] <lifeless> abentley: would more documentation in bzr have helped (to provide the expectation that you'd need to configure && make?
[03:06] <abentley> lifeless: Yes, I definitely felt at a loss as to how I was supposed to get subunit.  And I was surprised that it wasn't in Ubuntu because I'd been hearing about it from Canonical folk for years.
[03:08] <abentley> lifeless: I can't see any description of the argument to --parallel in the command help.
[03:09] <lifeless> how do you get help on a registry?
[03:12] <abentley> lifeless: You can set value_switches=True, like I did with branch formats, or you can write up a help topic, like we did with transports.  I don't think there's support for autogenerated help otherwise.
[03:13] <lifeless> what does value_switches do - would it stop the registry being lazy?
[03:14] <abentley> lifeless: value_switches causes each value to be a separate option, e.g. --fork, --subprocess.
[03:14] <abentley> lifeless: It would not stop the registry from being lazy.
[03:15] <lifeless> ok
[03:15] <lifeless> I think we should make that change, and file a bug about not getting help otherwise
[03:22] <abentley> lifeless: Anyhow, parallel=fork and parallel=subprocess both work now.
[03:24] <lifeless> col
[03:24] <lifeless> cool
[04:55] <_habnabit> Is there any sort of ETA on the branch update backlog?
[04:57] <wgrant> _habnabit: I estimate another 4 or 5ish hours, but that's working from this morning's prediction.
[04:57] <_habnabit> Dang it. :(
[04:58] <wgrant> Yes. This issue was not forseen, since the process has not been run on production before today.
[05:11] <abentley> _habnabit: It's going through ~2.5k/h, with 14k remaining.
[05:11] <_habnabit> Mmkay.
[05:14] <abentley> _habnabit: actually, closer to 3.0k than 2.5k.
[05:14] <lifeless> the lines have crossed ;)
[05:15] <lifeless> its 22000 seconds behind at the moment
[05:15] <lifeless> 6 hours to get where it is, 6 hours to go
[05:15] <wgrant> Aha.
[05:16] <lifeless> so it will be about 44K seconds behind when it catches up completely
[05:26] <Peng> This is wildly off-topic and very unimportant, but could someone near Canonical network-wise do me a favor? How accurate is ntp.ubuntu.com? ISTM it's about 2 ms slow, but I'm far away and it could just be an Internet issue.
[05:28] <wgrant> I've a host a couple of hops away.
[05:28]  * wgrant checks latency.
[05:36] <Peng> Thanks. :)
[05:38] <wgrant> Peng: I can't find another NTP server equally close (<1ms), but most others nearby seem to reliably be about 1-2ms ahead.
[05:44] <Peng> wgrant: So it really is a little inaccurate? Interesting, thanks. :)
[05:45] <wgrant> Peng: I'd say so.
[05:46] <Peng> wgrant: Thanks a lot. :)
[05:55] <lifeless> Peng: ask in #canonical-sysadmins
[05:55] <lifeless> Peng: or a question on launchpad
[05:55] <lifeless> ntp is meant to do latency adjustments
[05:59] <Peng> lifeless: Yeah, but it doesn't handle asymmetric latency.
[06:01] <lifeless> Peng: it doesn't? damn - I haven't read the spec to see
[06:02] <wgrant> How could it?
[06:03] <Peng> I think I've heard that Chrony tries to handle it better than the reference implementation.
[06:08] <lifeless> wgrant: because traffic gets sent in both directions
[06:09] <Peng> Usually it's close enough to symmetric that it's not a problem.
[06:18] <lifeless> when ntp was designed and built asymmetry was very common
[06:18] <lifeless> so  its surprising to me that it wouldn't be catered or
[06:18] <lifeless> *for*
[07:20] <mneptok> i'd like Ubuntu to use an NTP pool instead of Canonical instances
[07:21] <lifeless> mneptok: do you mean regional ones/
[07:21] <lifeless> mneptok: I recall some concerns about swamping unprepared time sources
[07:22] <mneptok> server 0.pool.ntp.org
[07:22] <mneptok> server 1.pool.ntp.org
[07:22] <mneptok> server 2.pool.ntp.org
[07:22] <mneptok> like so
[07:22] <mneptok> grabs servers from a round-robin
[07:23] <wgrant> And possibly increases load on that round-robin by a couple of orders of magnitude.
[07:23] <wgrant> lifeless: How's the scanner graph looking?
[07:23] <lifeless> 8K to go
[07:23] <mneptok> if you don't want that load, don't make your stuff public and encourage its use
[07:24] <mneptok> http://www.pool.ntp.org/en/
[07:24] <lifeless> mneptok: folk base load estimates on  current data; dropping several million machines onto an unprepard pool would be pretty harsh
[07:25] <mneptok> lifeless: that's why a Blueprint and UDS discussion would be in order ;)
[07:25] <lifeless> true
[07:25] <mneptok> and add Canonical servers to the pool
[07:25] <mneptok> "We're bringing the horde, but we have the salad and dessert."
[07:37] <thumper> wgrant: we seem to be processing just over 2k an hour, so probably another 4 hours (ish) till complete
[07:37] <wgrant> thumper: Ah. Not too bad.
[07:38]  * thumper gets ready to watch the rugby
[07:40] <lifeless> thumper: 3K
[07:41] <lifeless> 2hr to go
[07:42] <lifeless> mwhudson: if the puller had nothing to do, did you consider just lying about last-mirror?
[08:05] <vasi> hey, i can't figure out how to link my launchpad bug report to the equivalent report on SourceForge
[08:05] <vasi> should i be using the "Also affects project" link? and what if the project isn't there?
[08:06] <wgrant> vasi: You need to click 'Also affects project', select the upstream project, and enter the SourceForge.net URL
[08:06] <wgrant> If the project isn't there, you'll need to create it.
[08:06] <vasi> oh, random users can do that? cool
[08:06] <wgrant> Yep.
[08:06] <vasi> i kinda assumed that was for only admins
[08:06] <wgrant> It happens all the time.
[08:07] <vasi> ok, that was surprisingly easy....sorry for the newbie question :-)
[08:07] <vasi> (i did look at the docs, i swear!)
[08:08] <wgrant> Great.
[08:43] <goodnight> what's a "triaged" bug?
[08:45] <mwhudson> lifeless: yeah, it's a little more complicated than just lying about last mirror
[08:45] <mwhudson> but it's probably (a) doable and (b) worth it
[09:06] <lifeless> wgrant: 3K
[09:06] <wgrant> lifeless: Not bad.
[09:06] <wgrant> Unlike the lack of tests for error cases in archiveuploader.
[09:07] <lifeless> there are no error cases
[09:07] <lifeless> :)
[09:07] <lifeless> we know this, cause there aren't any tests....
[09:09] <wgrant> Looks like lp:launchpad/devel just got scanned.
[09:14] <lifeless> 1.5K now
[09:16] <wgrant> Odd.
[09:17] <lifeless> latency in the graph data I suspect
[09:17] <wgrant> Ah.
[09:22] <lifeless> but there is also a bend in the curve
[09:22] <lifeless> may be something associated with the size of the data set
[09:22] <_habnabit> Yaay! My branch finally updated.
[09:22] <lifeless> I now see ~500
[09:23] <lifeless> we're down to 9000secs latency
[09:23] <wgrant> Excellent.
[10:20] <gioele> hello, can SSH pubkeys be fully removed from a Launchpad account or will Launchpad just set them as "old"?
[12:05] <alefteris> I have checked out a mirrored branch and pushed it into a new lp hosted branch, but when doing bzr info,  parent branch still mentions the mirrored one
[12:05] <alefteris> How can I change it? thanks
[12:09] <lifeless> change what
[12:13] <maxb> bzr pull --remember parent-you-want ?
[12:17] <alefteris> thanks maxb
[12:19] <alefteris> Now I have another problem, when trying to change the development focus, the new branch is not there for selection, only the mirrored on is
[12:19] <alefteris> any idea what I'm doing wrong? thanks
[12:39] <alefteris> nevermind I found it
[12:39] <alefteris> the dev forus is a release series not a branch :/
[12:42] <alefteris> when someone aproves a merge request from the web, is the merge peroformed automaticaly in the branch or I have to do it also manualy from the terminal?
[12:44] <lifeless> you have to do it
[12:46] <alefteris> lifeless, lifeless I notised in the web merge request it has a commit message field? what is that used for?
[12:47] <lifeless> communication
[15:10] <Peng> mneptok: Adding a couple servers to the pool would not make up for the load Ubuntu users would bring...
[15:50] <carresmd> currently launchpad is _very_ slow.. having problems?
[15:52] <carresmd> ah, it's getting better
[16:13] <dtchen> filing a bug using edge appears to OOPS
[16:13] <dtchen> three in a row, the most recent being OOPS-1400EB434
[18:33] <mneptok> Peng: all Ubuntu use those servers now, so ...
[19:46] <hyc> hey, I added an upstream bug to launchpad #291760 but it used the wrong Project name
[19:46] <hyc> now when I try to change the Project it times out and says Bad Gateway
[19:47] <hyc> the upstream is gnome 580185 filed against NetworkManager
[19:48] <hyc> but the LP project says Linux, I wanted to change it to NetworkManager
[19:48] <hyc> anyone know why trying to change this fails?
[21:23] <lifeless> hyc: it already has a network manager task
[22:16] <hyc> yes, but it's all messed up. The NetworkManager task points to a kernel bug report
[22:16] <hyc> and the Linux task points to a NetworkManager bug report
[22:19] <wgrant> hyc: Click the little arrow at the left of each task, and select the correct bug watch for each.
[22:56] <RenatoSilva> is it common to report development bugs for documentation or whatever?
[22:57] <RenatoSilva> that is, report bugs that affect your development version, but no released version at all
[23:11] <hyc> wgrant: yes, I tried that but as I said at the beginning of this conversation, I was getting timeouts / Bad Gateway responses from that
[23:15] <wgrant> hyc: Reproducibly?
[23:17] <wgrant> hyc: It's incredibly slow, but I just swapped them