[01:03] <ndowens> I am packaging newlisp to be put into a PPA, and it is version 10.4.4, when I try ucan --report-status it gives me newlisp-wiki and newlisp-ide is newer. I am only packaging newlisp itself
[01:04] <ndowens> The download url is http://newlisp.org/downloads/newlisp-(.*)\.tgz
[01:04] <ndowens> in the watch file
[01:05] <ndowens> i tried to mangle wiki and then it gives that ide is newer, I can't figure out how to use multiple mangle
[01:25] <ndowens> never mind figured it out
[01:25] <infinity> ndowens: newlisp-([\d\.]*).tgz?
[01:26] <infinity> Or nevermind, if you got there just now. :P
[01:26] <ndowens> heh. yea the \d\ thing worked :)
[01:27] <infinity> \d and \. actually.
[01:27] <infinity> As in "only strings that match a sequence of \d (digits) and/or \. (dot).
[01:27] <infinity> "
[01:27] <cjwatson> Except \. doesn't mean that within [].
[01:30] <ndowens> Can you point me to where if arch=386 it would use a dependency than x86_64. It needs ffi.h so i386 would use libffi-dev and then x86_64 would use lib64ffi-dev
[01:30] <infinity> ndowens: Erm.  What?  It should always be libffi-dev.
[01:30] <cjwatson> [\d] isn't documented in perlrecharclass(1).  I think uscan(1) is being naughty in documenting this.
[01:30] <ndowens> ah
[01:31] <cjwatson> I'd write it either [[:digit:].]* or (?:\d|\.)*
[01:31] <ndowens> figured there would be a different dep, because there is a libffi-dev package and then lib64ffi-dev
[01:31] <infinity> cjwatson: It's documented in pcrepattern(3)
[01:31] <cjwatson> ndowens: lib64ffi-dev is only for when building 64-bit programs on i386, which is a rare case.
[01:32] <infinity> cjwatson: Generic character types.
[01:32] <ndowens> k. i package newlisp for Fedora, so I have to use conditonals for the diff deps
[01:32] <cjwatson> infinity: Yes.  I think that's cheating since uscan uses actual Perl regexps, though :)
[01:33] <infinity> cjwatson: Oh, I didn't realise it was actually perl.
[01:33] <infinity> Though, I suppose it makes sense that it is.
[01:33] <infinity> What sort of loon would write that in C?
[01:33] <cjwatson> ndowens: There is a syntax for that (libffi-dev [i386] | lib64ffi-dev [amd64]) but you almost certainly do not want to use it here.
[01:34] <infinity> cjwatson: So, this is a case of the tail wagging the dog, and Perl has become PCRE-compatible? :P
[01:34] <ndowens> heh. Seems Perl is used alot for code replacement annd such
[01:34] <cjwatson> ndowens: As a matter of design we try to avoid using different library paths or package names for the default form of a library on different architectures.
[01:34] <infinity> (Or, knowing Perl, it probably always had the undocumented shorthand forms, and PCRE just dutifully replicated it)
[01:34] <cjwatson> Yeah, more likely that the docs are incomplete.
[01:34] <cjwatson> Or that I'm failing to find the documentation of this.
[01:35] <sarnold> cjwatson: perlreref(1) documents  \d and \D as "work within or without a character class."
[01:35] <infinity> cjwatson: The strange bit here being that my brain saw the perlre(1) reference at the end of uscan(1) and just went "finger, yo, type 'man pcre'".
[01:35] <ndowens> I use to dislike Ubuntu, so I used Debian. Main reason was because I like the rolling-distro idea, and I used Debian unstable for the latest apps
[01:36] <ndowens> Ihave a package in Debian currently. So I figure I would just play around with packaging in PPA
[01:37] <cjwatson> sarnold: Ah, yes.  Thanks.
[01:37] <cjwatson> perlre has expanded, er, rather a lot since I last read it start-to-finish.
[01:38] <infinity> That's because you started cheating on Perl with Python.  You know, that language where you fork sed when you want to do things.
[01:38] <cjwatson> Which was in about 1999.
[01:38] <cjwatson> infinity: You're going to make me fix that just as an anti-shaming device, aren't you
[01:39] <infinity> Nonsense.  If you fix it, I'll have to find something else to pick on you for.
[01:39] <infinity> And when I run out of code examples, it's going to get nasty and personal instead.
[01:39] <infinity> So, best to leave it there.
[01:42] <cjwatson> # XXX: do not fix this bug or else infinity's going to mention that time with the watermelon
[01:43] <infinity> I thought it was a cantaloupe?
[01:45] <cjwatson> Shhh.
[01:50] <jetsaredim> is there a way to get a given package updated from upstream?
[01:52] <RAOF> jetsaredim: If the package has an active maintainer, they know about the new upstream. If the package does not have an active maintainer, you can become one :)
[01:54] <jetsaredim> RAOF: i was afraid you'd say that
[01:54] <RAOF> You were asking for it :)
[01:55] <RAOF> We're also after feature-freeze, so the default answer to ‘should we update this package to a new upstream’ is going to be ‘no’.
[01:55] <jetsaredim> its broken
[01:55] <cjwatson> You can also file a bug, of course, but there's no particular guarantee of when anyone will look at it.  Most of the time the best way to get a package updated is to follow its downstreams step by step: that is, if it's packaged in Debian, file a Debian bug to get the Debian maintainer to package a new version, and then we can sync/merge that into Ubuntu.
[01:55] <RAOF> If there are particularly good reasons for updating then that answer can become ‘yes’ :)
[01:55] <cjwatson> (As a general rule.  There are always exceptions.)
[01:56] <jetsaredim> how would i figure out if there is a downstream?
[01:56] <cjwatson> Look at the package changelog and see if it's all Ubuntu or if there are Debian entries there too
[01:57] <cjwatson> Or http://packages.qa.debian.org/<package name>
[01:57]  * cjwatson -> second attempt to sleep
[01:58] <jetsaredim> interestingly enough there is a downstream but its older than the 12.10 package
[04:34] <pitti> Good morning
[04:43] <ion> rning
[07:04] <dholbach> good morning
[07:05] <didrocks> hey Mr Holbach
[07:05] <dholbach> Salut Monsieur Rouche - comment ça va aujourd'hui?
[07:07] <didrocks> dholbach: ça va bien, et toi?
[07:08] <didrocks> (toujours en train de tousser par contre)
[07:08] <dholbach> oui, ça va - encore un peu malade, mais ça va
[08:43] <panchiniak> Hi. This page http://people.canonical.com/~ogra/mobile/ links to a not found one. Is there an update on "Ubuntu UMPC Edition"? Where could I find the cool .img used by ogra?
[08:45] <xnox> panchiniak: maybe #ubuntu-arm people will know.
[08:46] <xnox> panchiniak: looks very old, e.g. old branding. Currently for quantal regular Ubuntu Desktop is used on panda boards.
[08:50] <ev> those of you who use https://errors.ubuntu.com/, please be aware that the latest version of the website drops the launchpad=false option. The information fetched from Launchpad is now loaded asynchronously, so it will not slow down the loading of the most common problems table or cause it to hit a timeout.
[08:52] <xnox> nice & slick =)
[08:53] <ev> thanks
[08:53] <ev> that said, it's still way too slow
[08:53] <ev> I'm continuously looking into this
[08:54] <xnox> well I opened the page. Loaded a graph & table; 1-2 seconds and the launchpad settings flashed in for the whole page.
[08:54] <xnox> ... switching to "most common problems in the past year" failed with "An error occured while trying to load the most common problems" =(
[08:56] <ev> xnox: what does your javascript console say?
[08:56] <ev> any errors?
[08:57] <xnox> hmm... Failed to load resource: the server responded with a status of 504 (Gateway Time-out)
[08:57] <ev> right
[08:57] <ev> the request is taking too long
[08:57] <xnox> ev.... did you create a big website to make fun of my internet connection? =)))))) *giggle*
[08:58] <ev> I'll see what I can do as I rework the month view into a 30 day window and the day view into a 24 hour one
[08:58] <ev> xnox: oh no, it's entirely on our side
[08:58] <xnox> Well past month & day work =)
[08:58] <ev> xnox: your standard British broadband connection of bits of string and blue tack is not to blame here
[08:59] <ev> yeah, but at the start of every day the totals drop to 0 and the same goes for the start of the month
[09:00] <ev> https://bugs.launchpad.net/errors/+bug/1033813 https://bugs.launchpad.net/errors/+bug/1008730
[09:00] <xnox> ev: well actually I have Virgin Media cheese strings: 100 Mbps, which actually is 30 Mbps.....
[09:00] <Laney> it's very location dependent. My VM 100mbit actually is :-)
[09:01] <Laney> cdimage is good for maxing me out
[09:01] <ev> :)
[09:01] <panchiniak> xnox: thank you by the advice. I've seen a regular Ubuntu Desktop running on a Galaxy Tab, but it needs keyboard and mouse attachments. So I thought that that "Ubuntu UMPC Edition" would be enough to take the risks of hacking a non panda board (closed) machine. Otherwise I think I'm going just withdraw from the game.
[10:43] <tsdgeos> if i have a patch against ubuntu gtk patches, which bzr branch do i have to branch and submit a MR against?
[10:43] <tsdgeos> https://code.launchpad.net/~ubuntu-branches/ubuntu/quantal/gtk+3.0/quantal ?
[10:46] <cjwatson> $ apt-cache showsrc gtk+3.0 | grep -m1 Vcs-Bzr
[10:46] <cjwatson> Vcs-Bzr: https://code.launchpad.net/~ubuntu-desktop/gtk/ubuntugtk3
[10:46] <doko> lucas__, do you have any quick ideas about the ruby-* ftbfs in universe (see http://qa.ubuntuwire.com/ftbfs/)
[10:57] <tsdgeos> cjwatson: that makes much mroe sense, thanks
[11:06] <pitti> tsdgeos: FYI, 'debcheckout gtk+3.0' DTRT
[11:07] <tsdgeos> pitti: neat!
[11:15]  * tsdgeos shakes fist against quilt
[12:53] <smoser> anyone around familiar with libvirt?  kind of specific quesiton though.
[12:53] <smoser> i'm interested in getting it to not run dns server on  network.
[13:07] <ev> mpt: as it turns out, top-k is complex for arbitrary ranges
[13:07] <ev> who could've guessed
[13:07] <mpt> ev, I don't know what that means. What is top-k?
[13:07] <ev> so for a set of dates, get the 100 items with the greatest count
[13:09] <SpamapS> oh yeah
[13:09] <SpamapS> thats been perplexing database folks for years
[13:10] <SpamapS> ev: materialized views of common ranges are the simplest answer
[13:10] <SpamapS> but, not the most flexible
[13:13] <lifeless> ev: distributed map reduce.
[13:13] <lifeless> ev: clearly the answer.
[13:13] <ev> lifeless: storm?
[13:13] <ev> have you played with it yet, or any alternatives for that matter?
[13:13] <lifeless> its on my when-I-get-time-to-scratch-myself list.
[13:14] <ev> SpamapS: indeed - we can do that for the month view (get the past 30 days once a day, calculate, stuff somewhere), but for anything where realtime counts (like the past 24 hours making up a day) we're boned
[13:14] <ev> this is for the most common problems table on errors.ubuntu.com, for what it's worth
[13:14] <ev> lifeless: looks like I've got an itch :)
[13:15] <lifeless> ev: I can suggest an iterative algorithm that will stop short of a full scan
[13:15] <ev> lifeless: please do
[13:15] <ev> SpamapS: or indeed when people select arbitrary date ranges for the table
[13:16] <lifeless> ev: take the topk from each bucket, sum common values together and order; take lowest value V, then retrieve all items from each bucket where the count in that bucket is >= V/buckets.
[13:17] <lifeless> ev: sum and order again, cut at k, and say fuckit, thats good enough.
[13:17] <ev> ah excellent
[13:17] <ev> so that will require sorted buckets
[13:17] <lifeless> this is off the cuff at 1am.
[13:17] <lifeless> It may be terrible.
[13:17] <ev> so intermediary column families as cassandra doesn't support counters as part of a composite column name
[13:17] <ev> so no real time
[13:18] <ev> hm
[13:18] <lifeless> alternatively, maintain a sliding window topK family
[13:18] <lifeless> pick the ranges you want to support, and walk them forward
[13:18] <ev> still trying to wrap my head around this: http://www.acunu.com/2/post/2012/07/approximate-analytics.html
[13:18] <ev> and this http://www.cs.ucsb.edu/research/tech_reports/reports/2005-14.pdf
[13:19] <lifeless> in my tabs now, will read tomorrow
[13:19] <ev> lifeless: can you elaborate? (whenever you have time, that is)
[13:19] <ev> I don't want to keep you up
[13:19] <ev> but I'm not sure I follow what you mean by a sliding window top k family
[13:19] <ev> I'm assuming this involves timeuuids, I'm just not sure how :)
[13:22] <lifeless> ev: so, I'm thinking about the following two questions
[13:22] <lifeless> if you have a time range T, and some topK data in it - e.g. the top(K*2) or something
[13:23] <lifeless> and you wanted to know for range T+1, ending one bucket width later
[13:23] <lifeless> what the topK would be
[13:23] <lifeless> under what conditions would something that isn't topK in T become topK in T+1
[13:23] <lifeless> under what conditions would something that is topK in T stop being topK in T+1
[13:24] <lifeless> then repeat that for the case where you have T but want to know for T where you reduce it by one day
[13:24] <lifeless> bucket
[13:24] <lifeless> thing
[13:24] <lifeless> if you could do that fairly efficiently
[13:24] <lifeless> you could get T shifted by - say- 2 buckets, in realtime
[13:25] <lifeless> and just keep some number of T's prepped ready to roll, and a 'head' T that you update every day to keep things moving forward and answer 'lastT' queries
[13:25]  * lifeless stops handwaving
[13:26] <ev> yeah, I'm just not sure the data makes that possible. Of course these papers I'm skimming seem to argue in your direction :)
[13:28] <lifeless> ev: things like, if in the +1 bucket you have some count for value V of N, it can only become topK, if the count of V in T +N > topk[k]
[13:30] <xnox> for what it's worth postgres 9.2 supports ranges natively, I don't know how good the performance is as I have not played with it yet.
[13:30] <xnox> http://wiki.postgresql.org/wiki/What%27s_new_in_PostgreSQL_9.2#Range_Types
[13:30] <xnox> and it can give you intersections....
[13:30] <xnox> but you are not using posgres at all are you, ev?
[13:31] <xnox> use postgres 9.2. db as a "cache" for the top-k view..... lol
[13:32] <ev> lifeless: right, but if we looked at 24 hours of data, broken down in to hour buckets, and in all the hours except the most recent, a key has a value of 0, but has a value greater than the highest in top-k, it breaks down surely
[13:32] <ev> which is entirely possible when we're talking about crash signatures mapping to their frequency
[13:32] <ev> an issue lands and it suddenly spikes
[13:32] <ev> xnox: well you joke, but one option (before we got into this arbitrary range thing) was to use redis for the top-k
[13:33] <lifeless> ev: well, say your T range is 24 hours
[13:33] <lifeless> ev: so you want to merge in one additional hour
[13:33] <lifeless> ev: if its greater than any of the top-k in that hour, that key will be added in, no prob
[13:34] <ev> right
[13:34] <ev> ah
[13:35] <lifeless> ev: you can get to two hours by induction
[13:36] <lifeless> ev: so if it takes 10 seconds to load the nearest T range
[13:36] <lifeless> ev: and 20 seconds per day
[13:36] <lifeless> ev: you could do 3 days out in 70 seconds
[13:36] <lifeless> ev: and run T's overlapping by day, with <=3 days to the nearest T
[13:37] <lifeless> ev: night
[13:37] <ev> lifeless: thanks and goodnight
[13:38] <ev> I'll try to make sense of all that now :)
[13:42] <SpamapS> ev: to the map/reduce point, this is pretty much the approach the twitters of the world use to get realtime stats
[13:43] <ev> SpamapS: ja. Nathan Marz of Twitter wrote Storm
[13:44] <ev> though I think they had something in place before that came on the scene
[13:55] <SpamapS> probably
[14:28] <stokachu> xnox: whats the imap sync package you maintain?
[14:29] <xnox> stokachu: offlineimap ?
[14:29] <stokachu> xnox: yea, does that fetch all sub-folders too?
[14:29] <xnox> stokachu: yeah
[14:29] <stokachu> ok cool thanks
[14:30] <xnox> stokachu: you can specify python filters if you need to e.g. include only those folders that contain interesting names/topics/ and e.g. to exclude big and boring.
[14:31] <xnox> stokachu: I then sync all of that to a local imap server and access it with thunderbird or any other sensible client (emacs)
[14:32] <stokachu> xnox: cool yea this is exactly what i was looking for
[14:32] <stokachu> fetchmail doesn't do all folders without specifying each one
[14:33] <xnox> stokachu: offlineimap can sync all the read/unread flags two-way =) which fetchmail can't
[14:33] <xnox> stokachu: and offlineimap is two-way sync
[14:33] <xnox> stokachu: you can set read-only if you need / want
[14:34] <stokachu> xnox: do i have to run it on my shell for canonical email?
[14:34] <stokachu> for the 2-way sync
[14:34] <xnox> stokachu: no. you run it anywhere. It works via IMAP(S) protocol.
[14:34] <stokachu> ah ok
[14:35] <xnox> stokachu: e.g. I run offlineimap on my laptop.
[14:35] <stokachu> folderfilter = lambda folder: folder.startswith('MyLabel')  <- thats awesome
[14:40] <stokachu> xnox: this app is so awesome
[14:41] <xnox> stokachu: mind the gap.... that is mind the memory leaks, you may want to do `offlineimap -o` (run-once) if you are affected by the memory leaks
[14:42] <xnox> patches welcome ;-)
[14:42] <stokachu> haha ok
[14:43] <xnox> stokachu: also if you have a lot of mail use: status_backend = sqlite
[14:43] <xnox> to speed things up.
[14:44] <stokachu> xnox: cool yea i set that up with max connections 2
[15:00] <mpt> ev, what time of day does the graph update?
[15:00] <ev> mpt: I suspect that's broken.
[15:00] <ev> I've been manually updating it despite my branch to do it in cron landing
[15:00] <ev> will look at that in a bit
[15:00] <ev> still trying to brain dump the previous conversation
[15:01] <mpt> k
[15:01] <smoser> anyone have ideas?
[15:01] <smoser> if i run ubuntu-bug on a quantal instance today i see
[15:01] <smoser> Cannot connect to crash database, please check your Internet connection.
[15:01] <smoser> 'int' object has no attribute 'isspace'
[15:06] <mitya57> smoser: bug 1052754
[15:09] <smoser> mitya57, thanks
[15:15] <bgamari> Does anyone know anything about Xorg crashing in quantal right after logging in?
[15:15] <bgamari> This is on an Intel i915 box
[15:16] <bgamari> It might not be a crash since from the backtrace it seems that Xorg is falling gracefully out of its event loop
[15:17] <bgamari> unfortunately I'm not really sure what else to blame it on
[15:18] <bgamari> On attempting to login lightdm disappears, things look for a second like I'll be dropped into a session, then suddenly I'm back at lightdm
[15:19] <bgamari> Ahh, the guest session works so it seems the user's session is crashing
[15:19] <xnox> bgamari: does it crash with a brand-spanking new user account?
[15:19] <bgamari> about to find out
[15:20] <bgamari> looking through the failing user's .xsession-errors turns up quite a number of errors
[15:20] <xnox> bgamari: $ ubuntu-bug xorg ?
[15:21] <bgamari> although it's unclear whether this is the failing session or simply the end of the last working session
[15:22] <bgamari> xnox, Brand new user succeeds
[15:23] <xnox> bgamari: check your user settings or start-up apps? did you tinker with compiz or .profile?
[15:24] <bgamari> xnox, .xession-errors isn't touched during login of the failing account
[15:24] <xnox> horum
[15:24] <bgamari> So it's unlikely that compiz or start-up apps are the culprit
[15:24] <cjwatson> bryceh: the package-team-mapping spreadsheet I have maps coq to universe, but https://bazaar.launchpad.net/~bryce/arsenal/2.x/view/head:/reports/package-team-mapping.csv maps coq to foundations.  Do you know (a) what reports.qa actually looks at directly and (b) what needs to be done to fix this and get coq bugs off our list?
[15:52] <jamespage> james_w, would you be able to take a look at the packaging branch for oprofile? it appears to be quite broken and arosales is trying to ressurect it for quantal
[15:55] <james_w> jamespage, I'll take a look
[15:55] <arosales> james_w: specifically when I do a branch of oprofile for oneiric I get http://paste.ubuntu.com/1215045/
[15:56] <cjwatson> mvo: does bug 1016040 ring any bells?  I hope I was correct in reassigning to apt; the cache code breaks my head
[15:56] <james_w> arosales, ah
[15:57] <mvo> cjwatson: that sounds right to reassign
[15:57] <james_w> arosales, sounds like https://bugs.launchpad.net/bzr/+bug/888615 see the last two comments
[15:58] <arosales> james_w: taking a look . . .
[16:00] <arosales> james_w: doing the suggestion in comment 4 looks to have worked around the issue http://paste.ubuntu.com/1215054/
[16:00] <james_w> great
[16:01] <arosales> james_w: thanks for the pointer. Looks like the bug is still under investigation, but the work around does indeed work, thanks!
[16:02] <arosales> jamespage: thanks for getting sync'ed up with james_w
[16:02] <arosales> jamespage: I mean thanks for helping me get sync'ed up with james_w
[16:02] <arosales> :-)
[16:03] <jamespage> np
[16:08] <plars> didrocks: hi, quick question for you.  I'm working on upgrade testing and upgrading from precise-quantal seems to fail to preserve keyboard shortcuts right now.  I think this is pretty much the same as https://bugs.launchpad.net/ubuntu/+source/compiz/+bug/1041169.  Is that something already on your radar, or do you want a different bug for this, since that one was against the +bzr package?
[16:08] <cjwatson> psivaa: is there any way to get even manual logs for bug 1052605?
[16:09] <cjwatson> psivaa: a tarball of /var/log/, perhaps
[16:09] <cjwatson> psivaa: I have some suspicions but I need to know the order in which things have happened here
[16:10] <psivaa> cjwatson, just destroyed the machine and doing some fresh install, with released version of precise, with the intention of reproducing it
[16:10] <didrocks> plars: it's on our radar, unfortunately, we don't have a good story to fix the bug
[16:10] <didrocks> plars: it will get better, but not ideal
[16:10] <didrocks> with tomorrow's update
[16:10] <plars> didrocks: what do you mean by a good story? is that something I can help with?
[16:10] <didrocks> plars: if you know about debugging glib/gsettings and the dconf writer at its root, yeah :)
[16:11] <Riddell> stgraber: can you give me admin on http://packages.qa.dev.stgraber.org/ ?
[16:11] <Riddell> me and baloons are looking at testcases
[16:11] <plars> didrocks: no, sorry, but I'd be happy to provide more details or work with someone to help debug if needed
[16:11] <cjwatson> psivaa: thanks; I'll mark incomplete for now
[16:11] <didrocks> plars: we know exactly what is happening, just don't know how to fix it
[16:13] <psivaa> cjwatson, ack
[16:13] <plars> didrocks: on an unrelated note, I get a lot of compiz crashes trying to run tests under kvm, and not sure how to capture more on it.  apport fails to actually open a bug on it, and after rebooting I get kicked out on login.  Any idea if this is something well known already? I haven't seen it on real hardware so far, but most of my testing is on kvm at the moment
[16:14] <cjwatson> psivaa: also, bug 1050436 - is it reasonably straightforward to find out if this happens on precise as well?
[16:15] <didrocks> plars: it's when you close a window right?
[16:15] <plars> didrocks: near as I can tell, at least the xorg crash is the same as https://bugs.launchpad.net/ubuntu/+source/xorg-server/+bug/1043513
[16:15] <didrocks> plars: tomorrow, there is a compiz upload fixing most of them
[16:15] <stgraber> Riddell: you should be admin as a member of ~ubuntu-release. Maybe try to logout and login again, making sure the ubuntu-release checkbox is ticked on the SSO login page?
[16:15] <didrocks> plars: can you please ping me back after this upload?
[16:15] <plars> didrocks: no, in this case, it was when trying to change the theme
[16:15] <plars> didrocks: will do
[16:15] <plars> thanks!
[16:15] <didrocks> plars: yeah, same
[16:16] <psivaa> cjwatson, i saw that's happening in precise as well
[16:16] <didrocks> plars: juts keep me posted then, thanks :)
[16:16] <plars> didrocks: will do, thanks
[16:16] <didrocks> thanks to you :)
[16:17] <cjwatson> psivaa: ok, thanks
[16:21] <cjwatson> psivaa: if I threw up a random test package somewhere would you be able to give it a go?
[16:22] <bgamari> xnox, Yeah, I'm not really sure how to approach this short of wiping ~/.*
[16:22] <cjwatson> there's one post-2.00 patch upstream that *might* be relevant
[16:22] <bgamari> It seems to throw up so early that there's no evidence to help identify the issue
[16:22] <xnox> bgamari: ? i am sorry, context?
[16:22] <bgamari> Anyone know how to debug issues early in X session setup?
[16:23] <bgamari> xnox, Regarding the X session crash right after login on quantal
[16:23] <psivaa> cjwatson, i also reported another bug re: usb keyboard on amd64+mac, bug 1050855. and yes i could try the fix later today
[16:24] <bgamari> It seems that XFCE starts as expected
[16:24] <bgamari> So it's just a Unity/Gnome issue
[16:25] <cjwatson> psivaa: OK - not related to bug 1017879 FWIW, as in that case it worked in the desktop CD
[16:28] <psivaa> cjwatson, yes hence a new bug :)
[16:30] <cjwatson> right, but you said it was very similar to that other bug, and I'm just pointing out the important distinction that means it would require a different (and typically harder) kind of fix
[16:37] <psivaa> cjwatson, understand
[16:46] <bryceh> cjwatson, I've made the change in bzr and will coordinate with bdmurry on getting reports.qa updated.
[16:47] <bryceh> cjwatson, I believe the data originates from kate so we may need to make sure her coq mapping is correct as well.
[16:48] <bdmurray> cjwatson: I updated the spreadsheet during or after the meeting and the reports will be updated shortly
[16:53] <cjwatson> bryceh,bdmurray: great, thanks
[16:53] <cjwatson> bryceh: I did check the coq mapping in what I believe is Kate's spreadsheet
[16:58] <brendand> before i go wasting my time, does anyone know of a very simple script to file bugs in launchpad against a specified project? i'm not talking about ubuntu-bug style, just want to give a project name, a title and description and be done
[17:04] <bryceh> brendand, I don't know if such a script already exists but would be trivial to make using lpltk
[17:04] <cjwatson> or plain launchpadlib come to that
[17:05] <bryceh> even launchpadlib, although you'd need to muss with credentials and such, but probably could cargo cult that
[17:06] <cjwatson> It's hardly difficult
[17:06] <brendand> nope, hardly difficult
[17:06] <cjwatson> launchpad = Launchpad.login_with("lp-file-bug", "production")
[17:06] <cjwatson> done
[17:09] <smoser> hey. maybe someone can help me.
[17:09] <smoser> i'm really out of ideas.
[17:09] <smoser> bug 1052664
[17:10] <smoser> it seems to me, that when invoked in boot (and only intermittently even then), 'hostname -b -F /etc/hostname' does not ignore '#' comments.
[17:10] <smoser> i know you're going to tell me that doesnt make sense.
[17:10] <smoser> it doesnt
[17:13] <smoser> come on.
[17:13] <smoser> someone has to be able to spot what i'm doing wrong there, or at least tell me i'm compeltely wrong.
[17:14] <sarnold> smoser: is /bin/hostname at early boot provided by e.g. busybox in an initrd/initramfs and _doesn't_ ignore the # ?
[17:15] <SpamapS> cjwatson: would you expect boot-time fscks on Ubuntu server to display the same level of detail on the plymouth details plugin as they do on the graphical plymouth bootup?
[17:16]  * SpamapS apologizes for the weird parsing required to udnerstand that
[17:22] <lamont> in the new world of quantal, how do I tell nautilus to burn an iso to the CDR?  "write to disk" seems to have gone *poof*
[17:26] <smoser> sarnold, i had that thought, possibly that the initramfs was opening it. but my wrapped script did get called by the upstart job (which starts on 'starting').
[17:27] <smoser> so really outside of the initramfs setting it, i dont know what would do it.
[17:27] <smoser> and then, even if the initramfs set it wrong,the real /bin/hostname was told to set it, so it should have done that.
[17:35] <bryceh> brendand, in any case let me know if you write the script, I'd like to include such a script in lpltk
[17:35] <obounaim> Can anybody review my merge proposal.
[17:37] <bryceh> infinity, if you're still around, alberto uploaded nvidia-graphics-drivers-experimental-304 to replace nvidia-graphics-drivers-experimental.  Same bits as before, just the corrected name.  Mind waving it through NEW?
[17:39] <bryceh> infinity, there is also a nvidia-settings-experimental-304 to replace nvidia-settings-experimental, so the latter should be dropped from the archive
[17:43] <sarnold> smoser: the /bin/hostname source code even looks like it should match the description given in the manpage. I'm as confused as you. Is something calling hostname(2) directly without going through the /bin/hostname program?
[17:44] <infinity> bryceh: Looks like someone else NEWed them, but I'll make sure everything's sane WRT old versions being removed, etc.
[17:47] <smoser> sarnold, oh. good you're still paying attention.
[17:47] <smoser> i thoguht that too
[17:47] <smoser> and that is possible
[17:48] <smoser> however, i updated the wrapper program
[17:48] <smoser> http://paste.ubuntu.com/1215249/
[17:48] <smoser> so *if* that is the case, then that other hostname setter ran in between my read, invoke, read
[17:49] <bryceh> infinity, thanks
[17:51] <sarnold> smoser: that's a _good_ one. Wow.
[17:54] <smoser> the other hint is that this seems a recent regression.
[17:55] <smoser> i've only ever seen this in builds from today.
[17:55] <smoser> i've also grepped through /etc/ for occurence of '/etc/hostname' in the hope that whatever was doing it would showup there.
[17:56] <smoser> but the only time that string occurs is in the upstart job
[18:28] <smoser> sarnold, you want to poke at this system and see if you cant make heads or tails of this ?
[18:28] <smoser> i'm just completely baffled.
[18:28] <smoser> i can't waste more time on it and i'm just going to remove the comment from the file
[18:28] <smoser> but i'd love to knwo what is going on
[18:28] <sarnold> smoser: I'm not sure what i'd do next. aiming the kernel tracing at it seems most likely, but I don't know that off the top of my head, sorry
[18:28] <smoser> i straced
[18:28] <smoser> and i see hostname calling without the '#'
[18:29] <smoser> so it really seems like some other thing is reading that file
[18:32] <mdeslaur> smoser: what's echo $HOSTNAME give you?
[18:33] <smoser> the garbage
[18:33] <smoser> its really annoying.
[18:34] <smoser> its in my prompt
[18:36] <mdeslaur> smoser: what about uname -a?
[18:36] <smoser> garbage
[18:36] <smoser> it *did* set the hostname to that.
[18:41] <mdeslaur> smoser: dhcp or fixed ip?
[18:41] <smoser> dhcp.
[18:41] <smoser> i considered that too
[18:41] <smoser> but the race just seems so odd
[18:41] <mdeslaur> could it be racing with dhclient-script?
[18:41] <smoser> you're thinking ddns ? and dhcp setting it.
[18:41] <smoser> yeah.
[18:41] <smoser> but *every time* i hit this race.
[18:41] <mdeslaur> maybe try commenting it out in dhclient-script just for kicks?
[18:47] <smoser> mdeslaur, it sure doesn't look like it is doing it.
[18:47] <smoser> i put a print right there, and its not calling it
[18:48] <mdeslaur> hrm
[18:51] <smoser> fixed
[18:51] <smoser> sudo apt-get --purge remove network-manager
[18:51] <mdeslaur> smoser: oh! what was it?
[18:51] <smoser> GAH!
[18:51] <mdeslaur> on server??
[18:52] <smoser> this was on desktop cloud image
[18:52] <smoser> but i sweare i saw it in server
[18:52] <smoser> but i must hvae been wrong
[18:57] <sarnold> smoser: excellent! :) well. Sortof. Stupid NM.
[18:57] <smoser> yeah, i'm gonna hope a bug on it.
[18:58] <smoser> and then stop cloud-init from writing a comment there.
[19:24] <cjwatson> SpamapS: I don't quite recall, TBH; I think I would expect them to provide whatever you'd get by calling the appropriate fsck command from a console
[19:25] <cjwatson> it's possible the -C (?) progress interface gives the smarter splashes more than that
[19:31] <SpamapS> indeed, we have 'console output' on mountall, which runs the fsck, and I don't see where mountall closes stdout/stderr so it should arrive on console
[19:38] <brendand> bryceh - this is the basics. i used launchpadlib directly: https://code.launchpad.net/~brendan-donegan/+junk/lp-file-bug
[19:42] <bryceh> brendand, thanks
[20:24] <brendand> bryceh - where are you going to put it? i can propose a merge, no problem
[20:27] <bryceh> brendand, the scripts directory of python-launchpadlib-toolkit, once I finish a couple MIR+SRUs
[20:28] <bryceh> merge proposals to https://code.launchpad.net/~arsenal-devel/lpltk/2.x
[20:36] <brendand> bryceh, LaunchpadService.create_bug insists on filing into a package?
[20:39] <bryceh> brendand, it expects that yes, but that can certainly be changed to be optional.
[20:40] <bryceh> bbiab (AC repairman coming)
[21:13] <cjwatson> psivaa: deb http://people.canonical.com/~cjwatson/tmp/grub2/for-psivaa/ ./
[21:13] <cjwatson> (unsigned)
[21:19] <brendand> bryceh, hopefully all in order: https://code.launchpad.net/~brendan-donegan/lpltk/lp-file-bug-lpltk/+merge/125346
[21:24] <psivaa> cjwatson, thanks, ill give it a try in a little while
[21:25] <cjwatson> may not make much difference, but based on https://lists.gnu.org/archive/html/grub-devel/2012-07/msg00053.html and upthread
[21:25] <cjwatson> "broken BIOSes" is often a decent characterisation of Macs, so it's possible :-)
[21:54] <bryceh> brendand, thanks, merged
[21:57] <brendand> bryceh, was it added to scripts so that if i install lpltk it will be in the path?
[22:03] <bryceh> brendand, yep
[22:06] <brendand> bryce - heh. tiny little typo in there :/
[22:07] <bryceh> brendand, oh?
[22:07] <brendand> yeah, i'll push something real quick
[22:07] <bryceh> ok cool
[22:08] <bryceh> $ ./scripts/lp-file-bug foobar
[22:08] <bryceh> ... NameError: global name 'CalledProcessError' is not defined
[22:08] <brendand> oh, sure i'll take care of that too
[22:09] <brendand> but that's weird. CalledProcess is part of python subprocess. is it removed from your system somehow?
[22:11] <jtaylor> its subprocess.CalledProcessError
[22:11] <jtaylor> no global name
[22:12] <bryceh> brendand, fwiw I ran it on a precise box, if that matters
[22:12] <brendand> ok, i'm on it
[22:13] <brendand> oh wait, do you have nano installed?
[22:13] <brendand> and EDITOR is not set?
[22:15] <brendand> it wants to use either EDITOR or nano if EDITOR isn't specified
[22:17] <bryceh> $ echo $EDITOR
[22:17] <bryceh> emacs -nw
[22:18] <bryceh> maybe the space screwed it up?
[22:19] <bryceh> yep
[22:19] <bryceh> $ EDITOR=emacs ./scripts/lp-file-bug foobar
[22:19] <bryceh> Unable to file bug in:
[22:19] <bryceh> 	project: ubuntu
[22:19] <jtaylor> the argumet brakes it use shlex.split
[22:33] <psivaa> cjwatson, that did not make it work, i'm afraid.
[22:34] <psivaa> cjwatson, so i had to make the grub timeout to be -1 inorder for the grub menu to be displayed, that made the grub menu displayed, but none of the keystrokes work on the grub menu
[22:36] <psivaa> cjwatson, that has now left the system stay on grub menu forever and needs reinstalling :) ( which i will have to do anyway for further testing)
[23:00] <cjwatson> psivaa: ok, worth a try, thanks.  (you could modify the grub configuration without reinstalling with the aid of a live CD)
[23:01] <psivaa> cjwatson, i do not have an external cd drive though
[23:01] <cjwatson> or live USB image
[23:02] <psivaa> cjwatson, ahh ok, that i could do then, thanks that helps
[23:03] <psivaa> cjwatson, editing the grub during live session persists?
[23:06] <cjwatson> psivaa: mounting the target filesystem and saving changes to /boot/grub/grub.cfg within it does
[23:06] <cjwatson> psivaa: you can also edit /etc/default/grub and run update-grub, which is technically better because you aren't editing generated code, but that's more effort as you have to bind-mount all of /dev, /proc, and /sys into the target system and chroot in
[23:06] <cjwatson> which you may or may not be familiar with doing
[23:07] <psivaa> cjwatson, thanks, ill try doing the first, the second option has more newer stuff :)