[12:04] <SteveA> i'm trying the .deb of integration now
[12:04] <SteveA> i'll see how that goes
[12:04] <Kinnison> okies
[12:04] <Kinnison> mdz: Updated, can you look at the Packages file again?
[12:05] <mdz> Kinnison: loads better
[12:07] <Kinnison> mdz: those two words just made my day
[12:07] <mdz> Kinnison: most of the noise I see is now #3535
[12:08] <mdz> Kinnison: is the fact that Release.gpg is missing expected?  does that happen at the very end or something?
[12:09] <mdz> I'll report it as a bug and trust that you'll DTRT if it's expected
[12:12] <Kinnison> mdz: It's a case of "I don't have a key, so there's no final signing of the Release files" :-)
[12:14] <mdz> Kinnison: the diff is still too long to review due to #3535, but I did find one other issue (#3541) which could be gina-related
[12:16] <Kinnison> bug 3535 is possibly related to our lack of extra-overrides support
[12:16] <Ubugtu> Malone bug #3535: publisher (upstream) - Missing Bugs: and Origin: headers (overrides?) Fix req. for: launchpad-publisher (upstream), Severity: Normal, Assigned to: Nobody, Status: New http://launchpad.net/malone/bugs/3535
[12:16] <Kinnison> as for bug 3541...
[12:16] <Ubugtu> Malone bug #3541: publisher (upstream) - Missing linux-meta Fix req. for: launchpad-publisher (upstream), Severity: Normal, Assigned to: Nobody, Status: New http://launchpad.net/malone/bugs/3541
[12:17] <Kinnison> oddly the source appears to be imported
[12:17] <Kinnison> but the binaries are borked
[12:18] <mdz> perhaps related to the fact that bash source and the 'bash' binary are in main, while these are in universe?
[12:18] <mdz> the bash binary itself is OK
[12:19] <Kinnison> http://librarian.staging.launchpad.net/1130018/1130051/qICk9NNKg9TgcW97hom6cTbqes3.txt
[12:20] <Kinnison> I think it's to do with the component, certainly
[12:20] <Kinnison> kiko had a fix for it I think
[12:32] <SteveA> Kinnison: still have the problem.  Can I get a tarball please?
[12:33] <Kinnison> SteveA: https://chinstrap.ubuntu.com/~dsilvers/integration.tar.bz2
[12:33] <Kinnison> urgh
[12:34] <Kinnison> there's distrorelease specific stuff in denise for the overrides
[12:36] <Kinnison> gahgahgahgahgahgahgah
[12:36] <Kinnison> sodding sodding gah
[12:39] <SteveA> Kinnison: thanks
[12:42] <Kinnison> s'okay
[12:46] <elmo> Kinnison: I forgot about that myself, until I was opening dapper, FWIW - sorry
[12:46] <dilys> Merge to rocketfuel@canonical.com/launchpad--devel--0: [r=stevea]  Add dependency-aware scoring algorithm and [trivial]  tweaking buildd-scoring values again, heavy weight for package component. (patch-2716: celso.providelo@canonical.com)
[12:58] <Kinnison> elmo: s'okay, you've had a lot on your mind
[01:01] <Kinnison> Be want is croissant trammel
[01:01] <Kinnison> what a fantastic spam subject
[01:03] <Nafallo> lol
[01:45] <Kinnison> ciao all
[01:51] <sabdfl> night Kinnison
[02:08] <sivang> night all
[02:28] <SteveA> hi SteveA 
[02:28] <SteveA> hi stub 
[02:28] <stub> Morning
[02:28] <SteveA> i'm having some problems getting postgreql 8.0 set up for lp
[02:28] <SteveA> root@einheit:~# sudo -u postgres createuser -a -d steve
[02:28] <SteveA> createuser: could not connect to database template1: could not connect to server: No such file or directory
[02:28] <SteveA>         Is the server running locally and accepting
[02:28] <SteveA>         connections on Unix domain socket "/var/run/postgresql/.s.PGSQL.5432"?
[02:28] <SteveA> 
[02:29] <SteveA> root@einheit:~# ls /var/run/postgresql/.s.PGSQL.*
[02:29] <SteveA> /var/run/postgresql/.s.PGSQL.5433  /var/run/postgresql/.s.PGSQL.5433.lock
[02:29] <SteveA> 
[02:30] <SteveA> where does postgresql find what socket it should be using?
[02:30] <stub> Your psycopg is build for 7.4. You will need to set unix_socket_directory in /etc/postgresql/postgresql.conf to '/tmp' I think (and restart postgres)
[02:30] <SteveA> what has psycopg to do with it?
[02:31] <stub> oh... erm. Sorry. Answering a differnet question.
[02:32] <SteveA> changing the unix_socket_directory didn't help
[02:32] <stub> I don't know why createdb is trying to connect to a non-existant socket. Is PostgreSQL running? Is the version of createdb you are running the 7.4 binary and not the 8.0 binary?
[02:32] <stub> SteveA: You would need to change it to /var/run/postgresql I'm sorry
[02:33] <SteveA> yeah, i have various things from 7.4
[02:33] <SteveA> i'll remove packages and try again
[02:34] <sabdfl> stub: i just landed some bits on rocketfuel, will you be updating the staging code soon?
[02:34] <stub> I will be now
[02:35] <sabdfl> also, i think it may be a good idea to bring all this gina testing up on drescher and let staging do its normal daily update thing
[02:35] <sabdfl> when do you head for montreal?
[02:38] <quannum> Hi, any ideas who I should contact to have two launchpad accounts consolidated? I seem to have ended up with two :-o
[02:42] <jamesh> quannum: you can request the merge yourself if you have access to the registered emails of the two accounts
[02:42] <jamesh> quannum: https://launchpad.net/people/+requestmerge
[02:43] <jamesh> SteveA: the Postgres 8.0 Ubuntu package runs on a non-standard port
[02:44] <quannum> jamesh: Thanks, that did the trick
[02:44] <jamesh> SteveA: if you set it to 5432, things seem to work as expected
[02:45] <SteveA> jamesh: okay, that seems to fix it
[02:45] <Nafallo> jamesh: doesn't it use that by default?
[02:45] <Nafallo> atleast it does here.
[02:47] <jamesh> Nafallo: when I tried Postgres 8.0, the package configured itself to run on 5433
[02:48] <jamesh> SteveA: one thing I noticed is that recreating the sameple data with 8.0 ended up with more differences than expected
[02:48] <Nafallo> odd
[02:48] <jamesh> which can cause problems during merges
[02:56] <stub> sabdfl: Whats a dresher? Elmo made a cryptic comment about this box but I don't know what it is earmarked for.
[02:57] <stub> PG 8 won't run on the standard PG port because it is designed to run simultaneously with PG 7.4 and PG 8.1. And each postmaster needs a unique port.
[02:57] <stub> sabdfl: I think I fly out on Monday
[02:57] <sabdfl> drescher is earmarked as a parallel server for "this sort of stuff", i too it to mean archive publishing from elmo's comment
[02:58] <stub> Sounds like a replacement asuka
[02:58] <sabdfl> so, asuka would continue as "staging, running yesterday's data against today's code"
[02:58] <sabdfl> and drescher would be more flexible
[02:58] <sabdfl> i think
[02:58] <sabdfl> elmo: ?
[03:16] <elmo> drescher is the new ftp-master
[03:17] <elmo> IMNSHO the box that runs apt-ftparchive needs it's own box; I don't  much care what else runs on there, but what kinnison and I discussed and what makes sense for me is for the uploader, queue processor, CAP and apt-ftparchive to run on there
[03:17] <elmo> it has nothing to do with staging (in the long term) - it's meant for production, as soon as LP is ready to do the archive in production
[03:18] <elmo> oh, it's also obviously, ftp-master in the sense that mirrors rsync the final, live, production etc. copy of the archive from
[03:19] <elmo> as jackass is to dak, drescher is to LP - for anyone who knows what jackass is of course..
[03:21] <stub> ok. Would Gina run on there or stay on macquarie?
[03:22] <elmo> *shrug* either is fine by me
[03:23] <elmo> since gina AIUI, doesn't require much in the way of resources except for the initial run
[03:23] <elmo> (i.e. once dapper is on LP, it'll mostly be monitoring pocket updates)
[03:27] <stub> ok. I think we should move Gina to dresher at some point, as it keeps all the stuff together that messes with the distro in one controlled environment, which will be good if I can tighten up the database permissions.
[03:27] <stub> Might even make the archive mirror on macquarie redundant
[03:28] <elmo> yeah, I guess so, I don't think anything else uses it
[03:30] <mdz> I get occasional RequestQueryTimedOut errors from the spec tracker; a retry generally works.  is there some horrifically slow query in there?
[03:30] <stub> which particular URL?
[03:31] <mdz> whatever https://launchpad.net/distros/ubuntu/+addspec posts to
[03:31] <mdz> looks like it posts back to itself
[03:31] <stub> Sometimes it is caused by page rendering times, other times by delays waiting for database locks where the ugly whery is from another process
[03:32] <mdz> I got one just a moment ago
[03:39] <stub> mdz: I'm pretty sure that is being triggered because something else is locking resources in the db
[03:51] <Seveas> :S is the launchpad calendar buggy as hell ar am i doing something wrong?
[03:52] <Seveas> date/duration of events cannot be set correctly
[03:54] <Seveas> it always adds 40 minutes to the time
[03:55] <Seveas> and duration handling is beyond recognition 
[03:57] <mdz> I think jamesh is the person to ask
[04:05] <sabdfl> night all
[04:44] <stub> Anyone know how to specify a URL to a local directory in a config-manager config file?
[04:49] <jblack> Try file:///this/dir/here
[04:50] <jblack> Though I'm surprised that /this/dir/here doesn't work.
[04:52] <jblack> stub: that work for you?
[04:53] <stub> That was the second thing I tried. ValueError: unknown url type 'file:///home/warthogs/archives/rocketfuel/hct/1/devel'
[04:53] <stub> I guess it needs to know what sort of archive to find there so might need some magic prefix
[04:57] <jblack> heh. the cm code is in __init_.py
[04:59] <jblack> It should work.
[04:59] <jblack> For bzr, that is.
[04:59] <jblack> if from_location.startswith("file://"):
[04:59] <jblack>    from_location = from_location[7:] 
[05:00] <jamesh> the code for detecting branch types looks like it needs some work
[05:00] <jblack> Yeah.
[05:00] <jblack> Lifeless has a style about him...
[05:01] <jamesh> you can't just check the URL scheme to differentiate between sftp://chinstrap/home/warthogs/archives/rocketfuel@canonical.com/launchpad--devel--0 and sftp://chinstrap/home/warthogs/archives/rocketfuel/launchpad/devel
[05:02] <jblack> stub: Does it give a trace for which line is failing? Its all wrapped in a big eception. 
[05:04] <stub> Traceback (most recent call last):
[05:04] <stub>   File "/home/warthogs/source/config-manager/cm.py", line 24, in ?
[05:04] <stub>     main(sys.argv)
[05:04] <stub>   File "/home/warthogs/source/config-manager/lib/config_manager/__init__.py", line 205, in main
[05:04] <stub>     config.update(os.path.abspath(os.curdir))
[05:04] <stub>   File "/home/warthogs/source/config-manager/lib/config_manager/__init__.py", line 74, in update
[05:04] <stub>     entry.update(dir)
[05:04] <stub>   File "/home/warthogs/source/config-manager/lib/config_manager/__init__.py", line 156, in update
[05:04] <stub>     return self.build(path)
[05:04] <stub>   File "/home/warthogs/source/config-manager/lib/config_manager/__init__.py", line 144, in build
[05:04] <stub>     raise ValueError("unknown url type '%s'" % self.url)
[05:04] <stub> ValueError: unknown url type 'bzr:///home/warthogs/archives/rocketfuel/hct/1/devel'
[05:04] <stub> I'm attempting to build a launchpad tree on chinstrap without using sftp
[05:07] <jblack> Would you mind walking it through pdb? 
[05:09] <mdz> str could use "strip this prefix" and "strip this suffix" methods
[05:09] <mdz> like ${#} and ${%} in shell
[05:11] <stub> >>> 'foo'.strip('oo')
[05:11] <stub> 'f'
[05:11] <stub> >>> 'foo'.rstrip('oo')
[05:11] <stub> 'f'
[05:11] <stub> >>> 'foo'.lstrip('f')
[05:11] <stub> 'oo'
[05:13] <stub> mdz: ^^
[05:15] <mdz> stub: >>> 'foo'.rstrip('o')
[05:15] <mdz> 'f'
[05:15] <stub> close but no cigar?
[05:15] <mdz> it's interpreted as a list of characters, and all occurrences are stripped
[05:15] <mdz> rather than a string suffix/prefix
[05:16] <mdz> I was referring to this use case:
 if from_location.startswith("file://"):
    from_location = from_location[7:] 
[05:16] <mdz> which I see all the time
[05:17] <jblack> Yeah. strip takes a list of characters, and keeps stripping until the first failure.
[05:17] <jblack> >>> 'abcdabcd'.strip('abd')
[05:17] <jblack> 'cdabc'
[05:23] <stub> 121                         from bzrlib.merge import merge
[05:23] <stub> ImportError: 'No module named bzrlib.merge'
[05:24] <stub> I need to set my PYTHONPATH better
[05:30] <stub> w00t
[06:45] <brenden> it's be nice if it were possible to search the irc logs...
[06:45] <brenden> s/it's/it'd/
[07:08] <fabbione> so guys
[07:08] <fabbione> what's new in bzr-integration ?
[07:11] <fabbione> specially... which one should be using?
[07:39] <dilys> Merge to rocketfuel@canonical.com/launchpad--devel--0: [trivial]  make ILibrarianClient.remoteAddFile() send the content type.  Fixes bug #3347 (patch-2717: james.henstridge@canonical.com)
[08:52] <dilys> Merge to rocketfuel@canonical.com/launchpad--devel--0: [trivial]  production and staging config updates and topPeople performance regression fix (patch-2718: stuart.bishop@canonical.com)
[10:12] <sivang> Good morning all
[10:35] <carlos> hi
[10:35] <ajmitch> hello
[10:36] <sivang> hey carlos
[10:57] <Keybuk> http://people.ubuntu.com/~scott/launchpad-770.jpg
[10:58] <fabbione> Keybuk: nice :))
[11:01] <sivang> Keybuk: lol
[11:01] <sivang> Keybuk: what's that machine you're holding there?
[11:02] <Keybuk> sivang: Nokia 770
[11:02] <sivang> Keybuk: is it a combined Cell+PDA ?
[11:02] <Keybuk> sivang: no, it's neither ... it's a pocket-sized web browser/email client
[11:02] <sivang> Keybuk: ah, nice indeed
[11:03] <Keybuk> the "thumb-nail sized" portlets scare me
[11:03] <Keybuk> the screen on it is utterly gorgeous, you really can read them
[11:17] <dilys> Merge to rocketfuel@canonical.com/launchpad--production--1.38: [trivial]  Production 1.38 cherry picks (patch-1: stuart.bishop@canonical.com, rocketfuel@canonical.com, christian.reis@canonical.com)
[11:22] <Kinnison> Morning
[11:23] <Kinnison> Keybuk: aye, the 770's screen is nice
[11:24] <sivang> Kinnison++
[11:24] <Keybuk> it does have an unusual killer feature I wasn't expecting; when it can see my phone via bluetooth, the phone just appears in the file manager
[11:25] <Keybuk> and I can browse it and grab things off it without thinking
[11:25] <Keybuk> makes it the handiest thing for taking photos on the phone, and sticking them on the web
[11:27] <Keybuk> and I must admit, it also does satisfy that "must quickly look something up on the web" need without having to drag the laptop out and boot it
[11:28] <Kinnison> I guess so
[11:28] <Kinnison> but without ubiqutous networking, or cheap gprs, it's a bit useless when sat in the pub with no wireless
[11:29] <Kinnison> erm, is staging's librarian down?
[11:34] <Kinnison> Is there a way to aggregate specs across different products?
[11:45] <Kinnison> stub: ping?
[11:45] <Kinnison> hey carlos
[11:45] <carlos> Kinnison, hi
[11:51] <stub> Kinnison: pong
[11:52] <stub> Kinnison: I'm currently attempting to rebuild staging with bzr
[11:52] <stub> You need it urgently? Or can I keep it down for a while longer?
[11:53] <Kinnison> stub: I can wait
[11:53] <Kinnison> stub: I just figured I'd try and get as full an archive as we have published
[11:53] <Kinnison> stub: before mdz wakes up
[11:54] <stub> ok. I was just going to try another rebuild of the tree, which takes 20 minutes interestingly enough.
[11:54] <Kinnison> okay well let me know when it's back up
[12:08] <Ubugtu> Malone bug #3552: no way to cross-link specs across products Fix req. for: launchpad (upstream), Severity: Normal, Assigned to: Nobody, Status: New http://launchpad.net/malone/bugs/3552
[12:08] <Kinnison> good, then I'm not alone :-)
[12:08] <ajmitch> except that my specs are on the ubuntu distro, rather than products
[12:19] <stub> Kinnison: staging is back up
[12:20] <Kinnison> stub: thanks, I'm gonna run a CAP run
[12:20] <Kinnison> stub: then can we re-run gina to see what's left for her to import?
[12:20] <stub> Sure
[12:20] <stub> What is CAP btw?
[12:21] <Kinnison> Canonical Archive Publisher
[12:21] <Kinnison> launchpad.net/products/launchpad-publisher
[12:22] <Kinnison> hey sabdfl 
[12:22] <Kinnison> sabdfl: any chance of bug 3552 being fixed before UBZ?
[12:22] <Ubugtu> Malone bug #3552: no way to cross-link specs across products Fix req. for: launchpad (upstream), Severity: Normal, Assigned to: Nobody, Status: New http://launchpad.net/malone/bugs/3552
[12:27] <Kinnison> erm stub: the librarian on staging appears to be naffed
[12:27] <stub> What are you seeing?
[12:27] <Kinnison> https://chinstrap.ubuntu.com/~dsilvers/paste/fileCE8fvJ.html
[12:28] <stub> 2005/10/25 11:26 BST [-]  Starting factory <twisted.web.proxy.ProxyClientFactory
[12:28] <stub> instance at 0x2a9af71b48>
[12:28] <stub> 2005/10/25 11:26 BST [ProxyClient,client]  Stopping factory <twisted.web.proxy.ProxyClientFactory instance at 0x2a9af71b48>
[12:28] <stub> That is useful...
[12:28] <Kinnison> Oi, spiv!
[12:29] <stub> Oohh! The files arn't there because I rebuilt staging. Hang on and I'll put them back
[12:29] <stub> You added any new files?
[12:29] <stub> since I just brought it up?
[12:29] <Kinnison> nope
[12:29] <Kinnison> just trying to publish what should have been there from before
[12:31] <stub> Kinnison: ok. they should be there now
[12:31] <sabdfl> stub: while you are working on production, can you change the owner of the ubz sprint do 'ubuntu-drivers' please?
[12:33] <stub> sabdfl: done
[12:44] <Keybuk> random Q ... on lp/people/techboard there are two mdz's, one of whom is "exmdz-merged" who when clicked gives you a 404
[12:44] <Keybuk> known bug or not?
[12:48] <Kinnison> known peoplemerge bug
[12:53] <sabdfl> stub: thanks muchly
[12:58] <matsubara> good morning!
[12:59] <carlos> SteveA, the login form from https://launchpad.net/distros/ubuntu/+spec/langpacks-desktopfiles/+subscribe/+login does not redirects you to the +subscribe page, instead it says "You are already logged in"
[01:04] <mpt> Gooooooooooooooooood morning
[01:04] <cprov> mpt: morning
[01:07] <salgado> cprov, have you seen https://launchpad.net/errors/showEntry.html?id=1130236133.50.0141519736646 (not sure if it's your department, but it seems to be)
[01:07] <cprov> salgado: let's have a look
[01:09] <cprov> salgado: yes, it is, I've catched it sometime ago too, hope it's already solved, I  will rollout my last merge 
[01:12] <cprov> salgado: it's not merged yet, will do it ASAP, thank you 
[01:15] <salgado> cprov, cool. ta
[01:18] <stub> database is going down in 11 minutes. The wikis will be in read only mode and launchpad and shipit unavailable. Estimated downtime is 30 mins.
[01:20] <mdke> oof
[01:26] <mpt> SteveA: I think "To _____, you need to log in" shouldn't look like an error message, it should just be a normal paragraph
[01:29] <ajmitch> any idea when lifeless will be back on the ground & on irc?
[01:33] <jblack> You probably wont see much of him until you're in person.
[01:33] <jblack> probably a few pockets, with luck. Email is a better shot.
[01:34] <ajmitch> right
[01:34] <ajmitch> was just wanting to ask about pqm, packaging it now
[01:35] <stub> ajmitch: He won't land until tomorrow sometime. Should be on the wiki.
[01:35] <jblack> anything I might be able to asnwer?
[01:35] <ajmitch> pqm is the preferred name?\
[01:35] <jblack> Yeah. pqm. _not_ bzrpqm. :)
[01:35] <ajmitch> copyright info has canonical as (c) 2004
[01:35] <ajmitch> not arch-pqm? ;)
[01:35] <jblack> hmmm.
[01:36] <jblack> this new pqm does arch as well.
[01:36] <ajmitch> yes
[01:36] <ajmitch> it was called arch-pqm upstream
[01:36] <jblack> Is this a permanant decision that will haunt us for all eternity?
[01:36] <ajmitch> but it does bzr now as well
[01:36] <ajmitch> it's *possible* to rename
[01:37] <jblack> WHen I called it bzrpqm, he was clear to tell me that its called strictly "pqm"
[01:37] <ajmitch> ok
[01:37] <jblack> However, that wasn't in the context of package names.
[01:37] <ajmitch> pqm is what I've called it
[01:37] <mpt> The Launchpad product is called pqm :-)
[01:38] <jblack> I don't see how it'll be a big issue, as thats' whre it is everywhere.
[01:38] <ajmitch> jblack: consistency :)
[01:38] <ajmitch> I retitled the debian RFP from arch-pqm to pqm
[01:39] <jblack> Ok. Yeah. Call it pqm. If he seems miffed, tell him I took charge.
[01:39] <ajmitch> he'll understand
[01:39] <jblack> He'll be so thrilled that I took a risk that he'll only ream me half way. (chuckle. :)
[01:39] <jblack> Sure. gimme a good twenty seconds..
[01:39] <ajmitch> thanks
[01:40] <jblack> btw, did you depend upon gpg? 
[01:41] <ajmitch> not yet, I'm just sorting out these parts before I read through the code in depth
[01:41] <jblack> Gpg isn't strictly a requirement, but if you're not using it, then you're also likely the sort that drives around on motorcycles. without a helmet. Backwards. While eating tacos.
[01:41] <ajmitch> I'll put it as a depends
[01:41] <ajmitch> everyone should have it anyway
[01:42] <ajmitch> for the apt repository signing
[01:42] <fabbione> ajmitch: force it as Depends: becaue it is not mandatory to have repo signing
[01:43] <ajmitch> fabbione: I saw :)
[01:47] <jblack> How's this for a shot:
[01:47] <jblack> Patch Queue manager is an tool that provides an email interface for the
[01:47] <jblack> management of Bazaar and Bazaar-NG branches. A variety of capabilities are
[01:47] <jblack> supported including merging with or without checking test cases prior to merge. This tool is particularly useful to multi-developer projects that would like
[01:47] <jblack> to serialize write access to Bazaar and Bazaar-NG branches.
[01:48] <ajmitch> jblack: very useful, thanks
[01:49] <jblack> ajmitch: Anything else I can do for you before I hit the sack?
[01:49] <jblack> Or anyone else, for that matter?
[01:50] <ajmitch> jblack: no, you've been a good help
[01:51] <jblack> I hope so, because you're doing me a heck of a favor.
[01:51] <Kinnison> stub: out of interest, has staging's tables been analysed since gina ran?
[01:51] <jblack> Oh, my quick hack for the day is to flip the lcd brightness each time i get a mail.
[01:52] <stub> Kinnison: In theory yes, autovacuum is running. But I don't trust it that much in 7.4
[01:52] <stub> Running an analyze now
[01:54] <SteveA> hi
[01:54] <ajmitch> ok, deb built, now to clean it up
[01:55] <Kinnison> stub: okay, s'just that the publishing run is taking ages to extract the overrides
[01:56] <Kinnison> and this is plausibly a lack of analysis
[01:56] <stub> Could be. Anyway - it is running now.
[01:56] <stub> Which will slow things down further for a few more minutes ;)
[02:07] <Kinnison> well, CAP is off into apt-ftparchive now anyway
[02:09] <\sh> jblack: btw...you come to ubz? if yes, when? and would u give me a small sprint over all those magic LP support tools like baz/bzr/tla/hct?
[02:10] <ajmitch> judging from the state of what make dist & friends do
[02:12] <sivang> What's the maintainance launchpad is undergoing ?
[02:17] <stub> Weekly code update, which includes database patches and data migration stuff
[02:25] <sivang> stub: ah, cool
[02:44] <carlos> later
[02:48] <mpt> argh, the "Register Specification" link is so hard to find
[02:49] <sabdfl> stub: is production up and running with the new code?
[02:50] <sabdfl> it looks rather like the old code... staging seems quite different
[03:14] <stub> sabdfl: Production is running new code. patch-2711 with cherry picks of  2713, 2714 and 2718
[03:17] <\sh> hmm...the weekly view of the LP calendar is somehow broken...
[03:17] <\sh> you see only the start and the end of an appointment..while looking on daily basis, you see the appointed in this interval..
[03:30] <dilys> Merge to rocketfuel@canonical.com/launchpad--devel--0: r=salgado. first go at pre-defined bug listings. this merge also includes mpt's bug-listings-love, r=bjornt for that one. (patch-2719: bjorn.tillenius@canonical.com, mpt@canonical.com)
[03:32] <sabdfl> stub: can we plan for another rollout on Sunday?
[03:32] <sabdfl> i suspect we'll be tweaking various bits throughout the week and want to push those out before the conf
[03:33] <mpt> woohoo!
[03:33] <stub> ok.
[03:33] <mpt> thanks BjornT
[03:34] <stub> sabdfl: Although Sat would be better. Too many people will be in the air on Sunday
[03:35] <stub> Might be able to do it with cherry picks anyway. We can wait and see what lands.
[03:36] <Kinnison>  1800 launchpa  15   0 3076m 1.8g 2140 D  0.7 91.3   9:18.45 python
[03:36] <Kinnison> go python
[03:36] <Kinnison> it's your birthday
[03:38] <Kinnison> sqlobject and its caching really do make for big python processes
[03:38] <Kinnison> Is there a "flush the sqlobject cache" type call?
[03:38] <Kinnison> And a subsequent "now sodding garbage collect" call?
[03:39] <stub> Kinnison: Mind if I kick off another Gina run now?
[03:39] <Lathiat> you have to fight with the python garbage collector 
[03:39] <Kinnison> stub: staging is kinda busy still
[03:39] <SteveA> Kinnison: if the size of the process is an issue, there are knobs to tweak about that
[03:39] <stub> Kinnison: You want SteveA on that I think - he was last seen playing with the caches
[03:39] <SteveA> Kinnison: so, add a bof for it or something
[03:39] <stub> Kinnison: Yer, but is it ok to make it busier?
[03:40] <Kinnison> stub: it currently has a mere 30M of free ram
[03:40] <stub> Oh.. you are running that on asuka ;)
[03:40] <stub> asuka only has 1GB btw
[03:40] <Kinnison> 2
[03:41] <Kinnison> and I'm using both of 'em
[03:41] <stub> Can you or SteveA kick off /srv/launchpad.ubuntu.com/gina.sh when you have finished?
[03:41] <Kinnison> one sec
[03:41] <stub> As the launchpad user
[03:41] <mpt> SteveA: mpt@canonical.com/launchpad--design-fascism--0 is now conflict-free
[03:41] <Kinnison> screw death row
[03:41] <Kinnison> you can start gina
[03:42] <sabdfl> stub: will you not be in montreal by sunday?
[03:43] <stub> Nope
[03:43] <Kinnison> sabdfl: I now land saturday night in montreal
[03:43] <sabdfl> i will also want to do a production rollout the following sunday (day before lp week)
[03:44] <sabdfl> so that everything that got coded during the distro week lands for them the day they get home
[03:44] <sabdfl> and lastly, i'd like to bring gina / cap etc into production next week or over the weekend
[03:45] <sabdfl> if you want to do it saturday, let all of us (and me specifically) know which version of RF you are tagging off so we can have stuff that's easily cherry-pickable
[03:45] <stub> I thought we would turn on Gina this week?
[03:45] <sabdfl> stub: we will run gina in parallel this week, if the archives are perfect and builds look good too then i'm happy to go this week
[03:46] <sabdfl> next week will really have everyone in one place so i think kinnison will make very quick progress on any outstanding items
[03:49] <sabdfl> salgado: ping
[03:50] <berend_> Hi all!
[03:50] <salgado> sabdfl, pong
[03:51] <sabdfl> salgado: a note on style
[03:51] <sabdfl>     def getAllOptions(self):
[03:51] <sabdfl>         """See IPoll."""
[03:51] <sabdfl>         return getUtility(IPollOptionSet).selectByPoll(self)
[03:51] <sabdfl> don't overuse FooSet's
[03:51] <sabdfl> this would be far better as:
[03:51] <sabdfl> class Poll:
[03:51] <sabdfl>     def getAllOptions(self):
[03:52] <sabdfl>         return PollOption.selectBy(pollID=self.id)
[03:52] <sabdfl> make sense?
[03:52] <mpt> yay, baz corrupted my tree again
[03:52] <sabdfl> mpt: it hardly ever corrupts mine, are you using your branches too long?
[03:53] <sabdfl> i find if i make more, shorter-lived branches, i get better justice
[03:53] <matsubara> it seems i'm not the only lucky one! :)
[03:53] <sabdfl> anyhow, soon to be replaced with a whole new load of bugs
[03:53] <berend_> the launchpad bug system is behaving weird, it'll add a new comment whenever I try to edit the description.
[03:53] <salgado> sabdfl, yes, it does
[03:53] <berend_> not expected behaviour i'd say..?
[03:53] <sabdfl> salgado: if you have a poll, and you want to get things related to it, just get them directly
[03:54] <sabdfl> salgado: ok. i'm adding a new portlet for the poll options, too
[03:55] <salgado> sabdfl, I'm trying to remember why the hell I wrote that. in general, having a method in a FooSet that takes a Foo object as argument is an indication that something might be wrong
[03:55] <mpt> sabdfl: Both the branch I was switching from and the one I was switching to have only been around as long as it's taken me to finish them
[03:55] <mpt> a couple of weeks for both
[03:55] <sabdfl> salgado: also, please move IPollOption into interfaces/polloption.py, and similarly for database, zcml, browser etc
[03:56] <BjornT> berend_: i assume this only happens when you edit the description of a bug that has no comments? this is intentional, in order to preserve the original bug report. if you edit the description once again, no new comment should be added.
[03:56] <sabdfl> mpt: that means many merges from rocketfuel
[03:56] <sabdfl> mpt: in general, with bzr, i would encourage you to have smaller chunks that you put up for review and landing
[03:56] <berend_> BjornT: ok, I was about to file a bug report. Still a bit strange.
[03:56] <berend_> but thanx
[03:56] <salgado> sabdfl, sure. will do it
[03:57] <sabdfl> salgado: you have a FooSet method that takes a Bar as an option
[03:57] <sabdfl> but since every Foo has a Bar, it makes more sense to make it a method on Bar
[03:57] <sabdfl> salgado: and please do that for any other classes you are responsible for
[03:57] <sabdfl> each table gets a class in a file
[03:57] <uws> BjornT, berend_: What's the rationale behind this behaviour?
[03:57] <sabdfl> it has to be that way, so everyone else knows where to poke at your stuff
[03:57] <BjornT> berend_: the comment is actually already there, it's just not shown if it's the same as the description. i agree that it can be somewhat confusing, though.
[03:58] <sabdfl> berend_: ah. i added that feature
[03:58] <sabdfl> we keep the original description as the first comment
[03:58] <sabdfl> if we displayed it all the time, many bugs would look like comment...comment.
[03:58] <berend_> makes some sense, unless you're fixing typos
[03:58] <mpt> berend_: That should be fixed as part of KeepingBugsConcise
[03:58] <mpt> hmm, I should register that
[03:59] <sabdfl> salgado: why do we still have IPollOption.shortname?
[03:59] <sabdfl> wtf is a shortname?
[03:59] <sabdfl> have you seen a shortname on any other LP class?
[04:00] <sabdfl> we have: name, displayname, title, summary, description... you pick only the ones that make sense
[04:00] <salgado> sabdfl, that's something kiko suggested to make sure we can fit an option in a select box or something like that
[04:00] <sabdfl> in this case, name and title i think are all that are required
[04:00] <sabdfl> salgado: we have plenty of select boxes with titles displayed
[04:00] <salgado> he argued that we could have too long options, and that would be a problem
[04:00] <sabdfl> please eliminate shortname, we only need name and title
[04:00] <SteveA> jbailey: ping
[04:01] <sabdfl> in that case, add a IPollOption.summary
[04:01] <sabdfl> then you could display it as radio buttons, with the title in bold, and the summary in class="lesser" text below
[04:01] <sabdfl> that would be very neat
[04:01] <sabdfl> we can pair program it in Montreal, remind me
[04:02] <salgado> sure. that will be great. /me makes a note
[04:02] <jbailey> SteveA: pong
[04:02] <SteveA> mpt: i'd like to see a page, on the debug layer of launchpad, that gives an example of page layouts we use, but with the text used named after the CSS styles used.  So the title might say <h1>This is an h1 title, for the description of what you're looking at</h1>
[04:03] <SteveA> jbailey: i'm getting a 404 error when trying to install the latest bzrtools package from the 'small red circle' update manager in breezy
[04:03] <SteveA> jbailey: i think maybe the packages file is not consistent with the available .debs
[04:04] <jbailey> SteveA: Update the listings.  When I checked it this morning, something had gone wrong and there were no debs there.
[04:04] <jbailey> SteveA: I'm still not sure why, as when I ran the scripts by hand all the debs landed correctly.
[04:04] <jbailey> SteveA: But in any event, they're there now.  I'll watch for it tomorrow.
[04:04] <jbailey> I rewrote all the scripts yesterday to be consistant with one another (since I'm now doing nightlies for bzr, bzr-integration, bzrtools and bzrk)
[04:05] <SteveA> jbailey: yep, working now.  thanks.
[04:07] <mpt> SteveA: ok, on the to-do list
[04:07] <SteveA> https://launchpad.net/products/launchpad/+bug/3565
[04:07] <Ubugtu> Malone bug #3565: Demonstration pages of launchpad CSS. Fix req. for: launchpad (upstream), Severity: Normal, Assigned to: Matthew Paul Thomas, Status: New http://launchpad.net/malone/bugs/3565
[04:10] <sabdfl> salgado
[04:10] <sabdfl>               <tal:block condition="python: x != y">
[04:10] <sabdfl>                 <td tal:condition="python: x != 0 and y != 0"
[04:10] <sabdfl>                     style="text-align: right">
[04:10] <sabdfl>                   <span tal:replace="python: pairwise_matrix[x] [y] " />
[04:10] <sabdfl>                 </td>
[04:10] <sabdfl>                 <td tal:condition="python: x == 0 or y == 0">
[04:10] <sabdfl>                   <span tal:replace="python: pairwise_matrix[x] [y] " />
[04:10] <sabdfl>                 </td>
[04:10] <sabdfl> please don't put python in page templates
[04:12] <uws> sabdfl: The bug aliasing stuff sounds cool
[04:12] <SteveA> salgado: if you want to produce a matrix, make your view class generate a list of tuples
[04:13] <SteveA> and then have the page template simply go through that list of tuples, presenting the data
[04:13] <SteveA> there are special tal:repeat features to let you do special things in the first or the last column or row
[04:14] <SteveA> so you never need to compare to say 'is the row number 0' or anything like that
[04:17] <SteveA> salgado: here is some documentation for SimpleTAL, which is a little bit different from what we are using, but not significantly so.
[04:17] <SteveA>   http://www.owlfish.com/software/simpleTAL/tal-guide.html
[04:18] <Kinnison> SteveA: urgh, gina got stuck, know how to fix it?
[04:18] <SteveA> note in the example how you can say  <li tal:repeat="fruit context/getTheFruit"><span tal:replace="repeat/fruit/number">: <span tal:replace="fruit"/></li>
[04:18] <salgado> SteveA, I had two tal:repeat blocks, and had to compare the index in both (like <tal:block condition="python: x != y">, where x is the index I'm in on the first repeat block and y is the index on the second block), and I couldn't find out how to write that without python:
[04:20] <SteveA> salgado: so, return a list of tuples.  Inside each tuple you have dicts.  The dict looks like {'cssclass': 'whatever', 'value': 'valueofcell'}
[04:20] <SteveA> then you can precompute what you need in the view class
[04:20] <SteveA> and simply render it using simple tal:repeat and tal:condition
[04:20] <SteveA> Kinnison: i need a bit more information than that.   gina on staging?
[04:20] <mpt> ffs
[04:21] <Kinnison> SteveA: yes
[04:21] <Kinnison> SteveA: she seems blocked on writing her log
[04:21] <SteveA> mpt: pleasantly enough, that's what i'm working on now
[04:22] <SteveA> Kinnison: what makes you think that?
[04:22] <mpt> SteveA: rock
[04:22] <Kinnison> SteveA: strace
[04:24] <SteveA> Kinnison: can you pastebin the last part of the trace?
[04:24] <Kinnison> SteveA: I've just resumed the screen session
[04:24] <Kinnison> that appears to have unblocked her
[04:24] <SteveA> oh
[04:24] <SteveA> does screen have a max buffer size or something?
[04:24] <Kinnison> not a clue
[04:24] <Kinnison> sux if it does
[04:25] <Kinnison> I'll leave the screen session open
[04:26] <cprov> SteveA: hi, do you have time today to review my buildd-loading-zcml branch ? 
[04:28] <SteveA> cprov: i expect so
[04:28] <SteveA> is it on the reviews page?
[04:29] <cprov> SteveA: yes, it is
[04:30] <Kinnison> Do we have any idea when kiko will get to montreal?
[04:30] <cprov> SteveA: it's not that huge, but includes many files, I hope you can do it 
[04:31] <cprov> Kinnison: me ? he is leaving br tonight 
[04:31] <Kinnison> I thought he left yesterday
[04:33] <dilys> Merge to rocketfuel@canonical.com/launchpad--devel--0: [trivial]  Speed up fti.py (patch-2720: stuart.bishop@canonical.com)
[04:34] <cprov> Kinnison: me too, but just recieved a call from our travel agent looking for him .. he probably had troubles with his flight
[04:35] <Kinnison> SteveA: also you have the queue processor in your review bundle
[04:35] <Kinnison> SteveA: that's kinda important to get done soon
[04:36] <SteveA> ok
[04:36] <bradb> BjornT: around? I was wondering if you had a moment for a question related to the +filebug refactor review.
[04:39] <SteveA> jblack: around?
[04:44] <SteveA> cprov: celso.providelo@c.c/launchpad--buildd-loading-zcml--0 ?
[04:44] <SteveA> Kinnison: daniel.silverstone@c.c--desktop/launchpad--upload-and-queue--2--patch-38 ?
[04:44] <cprov> SteveA: yup
[04:44] <SteveA> cprov: there is a conflict
[04:45] <cprov> SteveA: will sort it out now .. go to dsilvers's one before
[04:45] <Kinnison> SteveA: aye
[04:47] <sabdfl> lifeless: in baz, how do i tell the last revision of mainline launchpad that was merged into my branch?
[04:51] <sabdfl> Kinnison: hope you recover quickly!
[04:56] <SteveA> ddaa: ping?
[05:24] <doko> hmm, launchpad doesn't let me add a new spec ...
[05:25] <doko> https://launchpad.net/distros/ubuntu/+addspec
[05:25] <doko> any hints?
[05:26] <sivang> doko: works for me
[05:26] <sivang> doko: maybe try again :)
[05:31] <doko> sivang: same error
[05:31] <sivang> doko: are you logged etc?
[05:31] <sivang> (stupid question, but still)
[05:32] <doko> sivang: I am logged in
[05:32] <Lathiat> mm interested bug
[05:32] <Lathiat> open a bug report, click on the status
[05:32] <Lathiat> then click the login loink
[05:33] <Lathiat> once you login you get two notices
[05:33] <Lathiat>  * The requested page is protected. You will need to login.
[05:33] <Lathiat>  * You have been logged in
[05:33] <Lathiat> and the page appears properly
[05:33] <Lathiat> i guess the check/login check is just out of order
[05:33] <mdke> do all specs have to be in launchpad?
[05:33] <Lathiat> mdke: i beleive so
[05:33] <Lathiat> for ubz anyway
[05:33] <mdke> what is the LP/wiki relationship?
[05:34] <Lathiat> you link to the wiki page with the spec
[05:34] <Lathiat> but pu tthe name/summary/people doing it in LP
[05:34] <mdke> hmm
[05:35] <mdke> where do I file the spec if it is related to ubuntu but not to a package (e.g. the website)?
[05:36] <Lathiat>  /distros/ubuntu/+specs
[05:37] <mdke> and also, can somebody adjust https://wiki.ubuntu.com/SpecSpec to reflect the fact that the spec needs to be linked to the spec and shouldn't have the name/summary/people doing it on the page?
[05:37] <Lathiat> it probably should
[05:37] <mdke> ok, but if I understood you correctly, you said that that stuff should be in LP
[05:38] <mdke> perhaps just show me a model spec? ;)
[05:38] <sivang> mdke: did you get the -devel emails me and sabdfl sent? that's the workflow, and you can choose when creating a new spec on the wiki to use "SpecTempalte"
[05:38] <mdke> loads of the ones I see don't have LP links
[05:38] <sivang> mdke: that should make things cleaer
[05:39] <sivang> mdke: well, if you create a new one, make it follow the current SpecTemplate
[05:39] <mdke> sivang, no, and the existing specs all say "See SpecSpec" for a template
[05:39] <sivang> mdke: SpecSpec is not a template, it contains guidelines
[05:39] <doko> mdke, sivang: so you can create new specs?
[05:39] <sivang> mdke: SpecTemplate is a template in which you can fill in data
[05:39] <sivang> doko: yes
[05:39] <mdke> sivang, dude, look at it
[05:40] <mdke> "This is an example Ubuntu specification"
[05:40] <sivang> mdke: At spec template?
[05:40] <sivang> :)
[05:40] <mdke> anyhow, i see the other template page now
[05:41] <mdke> having 2 is a bit confusing
[05:41] <jbailey> Looking at this bug that doko's seeing, it says to report the bug but doesn't give any information to add to the report.
[05:41] <jbailey> Do we just give the date and time and assume that it'll get dug out of the log files/
[06:02] <dilys> Merge to rocketfuel@canonical.com/launchpad--devel--0: [trivial]  Various small publisher fixes (patch-2721: daniel.silverstone@canonical.com)
[06:02] <SteveA> mpt: approved your winnowing branch.  but see email.
[06:03] <mpt> ok, thanks
[06:07] <SteveA> Keybuk: question about the librarian find-by-digest API in your email.
[06:09] <Keybuk> SteveA: huh-whuh? -ECONTEXT
[06:10] <Keybuk> ahh, you mean in my INBOX :p  not [the email you sent] 
[06:13] <bradb> BjornT: You've got mail. :)
[06:15] <SteveA> Keybuk: yes, 'your email' not 'your email'
[06:16] <mpt> that's what happens when you engage in synecdoching
[06:16] <SteveA> thanks scott
[06:20] <bradb> Hm, didn't we used to have a soyuz product, like, yesterday?
[06:20] <bradb> I could have swore that's where I saw the PackageSubscriptions spec reg'd
[06:32] <ddaa> sabdfl: ping
[06:32] <sabdfl> ddaa: pong. on calls, so will be slow
[06:33] <ddaa> sabdfl: I'd like to have a small meeting with your and niemeyer in #canonical-meeting
[06:33] <sabdfl> ddaa: i'm on spec calls for distro team will 23:30 UTC
[06:33] <sabdfl> s/will/till/
[06:35] <mpt> salgado, you beat me to that duplicate by about 30 seconds
[06:37] <ddaa> sabdfl: okay, then we'll have that meeting tomorrow.
[06:46] <mpt> bradb: Yes, it caused a 404 on https://launchpad.net/specs
[06:48] <bradb> mpt: What caused a 404 on that page?
[06:48] <mpt> bradb: The list of most recent specs included a link to https://launchpad.net/products/soyuz/+spec/something that no longer existed because https://launchpad.net/products/soyuz no longer exists
[06:49] <bradb> ah
[06:49] <bradb> It looks like we didn't migrate the data over either (ISTR there being several specs reg'd on soyuz), but maybe I'm missing something.
[06:56] <SteveA> mpt: design-fascism reviewed
[06:57] <SteveA> cprov: how's that updated branch doing?
[06:57] <SteveA> ddaa: hi
[06:57] <cprov> SteveA: almost there, archive-mirror takes ages 
[06:57] <SteveA> ddaa: i want to get my mirrored archive from chinstrap, and make it my working archive on my laptop 
[06:57] <mpt> SteveA: THANKYOU
[06:57] <SteveA> ddaa: other than the mechanics of copying the archive to my laptop, and registering it there, anything else i need to do?
[06:58] <ddaa> SteveA: you need to remove the =meta-info/mirror there
[06:58] <ddaa> that turns a non-committable mirror into a committable master
[06:58] <SteveA> ok
[06:59] <ddaa> and be sure that you do not have any cache data that refers to stuff not in branch
[06:59] <ddaa> (variant of "do not remove comitted revisions)
[07:01] <SteveA> um
[07:01] <SteveA> you mean, nuke my revlib?
[07:01] <mpt> Kinnison: ping
[07:22] <Kinnison> mpt: pong
[07:22] <Kinnison> sorry I took so long
[07:22] <Kinnison> I fell asleep
[07:23] <mpt> Kinnison: "No Commits have occurred, cannot perform incremental tobaz"
[07:23] <mpt> Kinnison: What's a tobaz?
[07:23] <mpt> (apart from being the sibilant version of a topaz)
[07:25] <mpt> I'm not getting the error, I'm copyediting it :-)
[07:26] <Kinnison> looks like my headache affected me more than I thought
[07:26] <Kinnison> Wanna try giving me some context
[07:26] <Kinnison> ?
[07:26] <mpt> Kinnison: So will the "tobaz" term already be familiar to those encountering this error?
[07:27] <Kinnison> how the heck should I know?
[07:27] <mpt> oh, this isn't your code?
[07:27] <mpt> apologies
[07:28] <Kinnison> tell me where the error is and I may be able to help
[07:29] <mpt> lib/importd/JobStrategy.py
[07:30] <Kinnison> ddaa or lifeless code then
[07:30] <Kinnison> Right, it's a totla() which converts to baz instead of tla
[07:31] <Kinnison> it's part of the process where we import all svn, cvs, etc repositories into the supermirror
[07:31] <Kinnison> an "incremental tobaz" is where we already converted some of the branch and we're converting new commits
[07:31] <Kinnison> the error is clearly because there haven't been any
[07:31] <mpt> right
[07:32] <mpt> yeah, that's what I was thinking
[07:32] <mpt> I suppose it's a bit unnecessary to error it at all anyway -- it should be more "yay! nothing to do!"
[07:32] <mpt> but anyway, I was just copyediting
[07:32] <mpt> it came up on a search for "occured"
[07:33] <mpt> thanks Kinnison
[07:35] <Kinnison> s'okay
[07:35] <Kinnison> glad to be of help
[07:35] <Kinnison> anyone know if kiko's around?
[07:38] <mpt> I think he's on a plane somewhere
[07:42] <Kinnison> arse :-)
[07:54] <Kinnison> SteveA: thanks for the review, I'll get to it when I get back from the cinema
[07:54] <SteveA> ok
[07:54] <SteveA> what ar eyou seeing?
[07:55] <Kinnison> curse of the were rabbit
[07:55] <SteveA> cool
[07:55] <SteveA> i want to see that
[07:56] <Kinnison> and I wanted to see something but didn't want to drive, so cunningly choosing something a friend wants to see means I can cadge a lift :-)
[07:56] <dilys> Merge to rocketfuel@canonical.com/launchpad--devel--0: [trivial]  Revert fix for https://launchpad.net/products/launchpad/+bug/2662 (PersonAccountToMergeVocabulary() contains the person that is logged in) as that caused people merge to always time out. (patch-2722: guilherme.salgado@canonical.com)
[07:57] <mpt> whoa, there's a Brad Shuttleworth
[07:57] <Kinnison> mpt: pardon?
[07:57] <SteveA> sure there is
[07:57] <mpt> that's going to be confusing
[07:57] <Kinnison> The gestalt entity brabdflb
[07:57] <SteveA> one of mark's brothers
[07:59] <SteveA> he did some nice things with the schooltool logo / website some time ago
[07:59] <SteveA> although i think the way the site looks has been changed since then
[08:03] <salgado> is production down?
[08:03] <salgado> or is it just our crap link here in .br?
[08:04] <cprov> salgado: second option ;)
[08:04] <SteveA> it isn't happy
[08:06] <cprov> SteveA: what shall I do, cry instead ? 
[08:06] <cprov> the current link we have is the most blocker I've ever had to work
[08:09] <SteveA> production will be back up shortly
[08:10] <SteveA> back
[08:10] <Kinnison> good dinner
[08:11] <SteveA> the box had run out of space because of a huge logfile of some sort
[08:11] <SteveA> elmo was on the case already
[08:30] <SteveA> cprov: did you make any changes on that branch i need to review, other than to remove the conflict?
[08:33] <SteveA> on a vespa?
[08:33] <SteveA> ciao
[08:52] <mdz> mpt: please subscribe to https://launchpad.net/distros/ubuntu/+spec/ubuntu-express
[08:55] <bradb> BjornT: around?
[08:55] <BjornT> bradb: yeah, just got back. i'm reading your mail now.
[08:56] <bradb> ok, cool
[08:57] <mpt> mdz: done
[08:57] <SteveA> cprov: reviewed
[08:57] <cprov> SteveA: great !
[09:02] <mdz> mpt: thanks
[09:24] <SteveA> mpt: reviewed stylesheet-sanity
[09:26] <mpt> SteveA: thanks :-)
[09:26] <dilys> Merge to thelove@canonical.com/bazaar--devo--1.4: bugfix from 1.4.3 (patch-72: Matthieu.Moy@imag.fr, thelove@canonical.com)
[09:29] <SteveA> mpt: i'll actually attach the review...
[09:36] <BjornT> bradb: you've got mail
[09:36] <bradb> thanks
[10:17] <mpt> hi sabdfl, SteveA has reviewed mpt@canonical.com/launchpad--stylesheet-sanity--0510
[10:17] <mpt> it reduces our style sheet bandwidth from 56 KB to 18.5 KB
[10:17] <mpt> (I could have reduced it further, but I was getting sick of resolving conflicts)
[10:17] <salgado> jbailey, what's the best place to ask for help on initramfs-tools?
[10:26] <jbailey> salgado: #ubuntu-devel, me or infinity.
[10:32] <mpt> SteveA: bug 1588 is fixed now, right?
[10:32] <Ubugtu> Malone bug #1588: Plain text exceptions in failing pagetests Fix req. for: launchpad (upstream), Severity: Normal, Assigned to: Steve Alexander, Status: New http://launchpad.net/malone/bugs/1588
[10:37] <SteveA> mpt: you now get plain text exceptions in page tests
[10:37] <mpt> yes
[10:40] <SteveA> bug 1588
[10:40] <Ubugtu> Malone bug #1588: Plain text exceptions in failing pagetests Fix req. for: launchpad (upstream), Severity: Normal, Assigned to: Steve Alexander, Status: Fixed http://launchpad.net/malone/bugs/1588
[10:42] <dilys> Merge to rocketfuel@canonical.com/launchpad--devel--0: [r=SteveA]  Abolishes the /foaf URL-space, and the wonders therein. (patch-2723: mpt@canonical.com)
[10:42] <mpt> w00t!
[10:48] <sabdfl> mpt: good landing. tell me about stylesheet sanity..?
[10:49] <sabdfl> i had to RE fix anchors in portals yesterday, and the spacing of LI items in portals too
[10:49] <sabdfl> i don't really want to fix those again
[10:49] <sabdfl> portlet links should not be underlined. portlet lists should not be widely spaced
[10:51] <mpt> sabdfl: In rocketfuel, about half of launchpad.css is fixing bugs or poor design decisions in plone.css
[10:51] <mpt> Then there's a lot of plone.css that we'll never use because Launchpad isn't a magazine site
[10:51] <SteveA> mpt: i think the size of .css files should not be all that important.  we should add chacheing headers to them so that they are generally loaded only once.
[10:52] <mpt> And it's hard to fix bugs because of all the odd stuff plone.css is doing
[10:52] <mpt> e.g. making the tab borders break
[10:53] <mpt> so launchpad--stylesheet-sanity--0510 merges the style sheets, merges the conflicting rules, and removes the unused rules
[10:54] <mpt> sabdfl: w.r.t. the links in portlets, I think that was a cherrypick gone wrong -- for several weeks they were without underlines on production, but with them in rocketfuel
[10:55] <SteveA> hard to write a test that portlets are rendered without underlines
[10:56] <mpt> Yes, it's hard to test anything related to CSS
[10:56] <mpt> unless Firefox's DOM Inspector grew an API :-)
[10:58] <SteveA> there is something to be said for keeping up with improvements that the plone guys make to their basic stylesheets.  we'd want to have a regular system of checking their SVN tree, or looking at a bzr import of that, for new additions.
[10:59] <SteveA> people aren't doing a whole lot new with it at the moment, i think.
[10:59] <SteveA> the time of fixing / detecting issues in different browsers has passed, mostly.  although, that might change with the new IE release.
[11:11] <mpt> SteveA: w.r.t. your review question about style="100%", once the style sheet is in a maintainable state, it'll be easier to work on reducing the use of style= in templates
[11:11] <mpt> (style="width: 100%", I mean)
[11:11] <SteveA> this is why i want those debug pages that demonstrate the different styles
[11:11] <SteveA> as a launchpad developer, i don't know when to use what styles where
[11:12] <SteveA> or which styles are 'supported' in launchpad
[11:12] <SteveA> i can copy other templates
[11:12] <SteveA> but they might well be wrong
[11:12] <mpt> yeah
[11:12] <SteveA> so, if we have a few pages, registered on the root object
[11:12] <mpt> post-fascism, they should be much less wrong
[11:12] <SteveA> these will be the canonical reference
[11:12] <SteveA> and also, it will be easy to choose a style based on what it looks like
[11:13] <mpt> yep
[11:17] <SteveA> i've found the correct places in the infrastructure to hook in the Vary header stuff
[11:22] <mpt> SteveA: What is it set to currently?
[11:23] <SteveA> set to?
[11:23] <SteveA> right now, there is no Vary header, and the cookie value never varies
[11:23] <SteveA> also, becuase of brad's changes i think, we're maintaining sessions even for not-logged-in users
[11:24] <bradb> BjornT: reply sent
[11:24] <mpt> bah, the online cacheability checkers don't do HTTPS
[11:24] <SteveA> which makes the problem worse
[11:24] <SteveA> in fact...
[11:24] <SteveA> i think brad's adding that navigation box has made this problem appear now
[11:24] <SteveA> and an instant fix would be to turn it off
[11:24] <bradb> SteveA: Interesting, because it hasn't landed. :)
[11:24] <SteveA> the long term fix would be to do the Vary stuff properly
[11:24] <SteveA> bradb: not in RF yet?
[11:24] <bradb> no, and not for a while
[11:25] <SteveA> there's something in RF which is causing cookies to be issued for anonymous
[11:25] <mpt> That's just what I was going to ask
[11:25] <SteveA> something is using the session stuff
[11:25] <mpt> Why is Launchpad emitting a cookie for every image and every everything else
[11:26] <SteveA> i don't know
[11:26] <SteveA> i'll grep to see what is using sessions
[11:27] <mpt> All our images are being sent with "Pragma: no-cache   Cache-Control: no-cache"
[11:27] <mpt> that's unnecessary
[11:27] <mpt> All images except hackergotchis can be allowed to cache for a week, if not longer
[11:28] <SteveA> notification stuff maybe
[11:28] <mpt> notification stuff?
[11:28] <SteveA> bradb: ideally, i don't want anonymous users to have sessions created for them
[11:28] <SteveA> bradb: we need to see if it is possible to do this
[11:28] <mpt> oh, for the session stuff
[11:29] <mpt> The style sheets have "Pragma: no-cache   Cache-Control: no-cache" too
[11:29] <mpt> that's 56 KB with every page load
[11:29] <SteveA> bradb: so that might be a case of noting if we need notifications for anon users, and not offering the 'navigation breadcrumbs' to anon users, perhaps
[11:29] <SteveA> mpt: yeah.  we can fix that easily enough.  but i don't want to look into it right now.
[11:29] <SteveA> i want to get the login bug fixed first
[11:30] <mpt> fair enough
[11:30] <bradb> hm
[11:30] <bradb> it seems to me that we need notifications for anon users
[11:31] <SteveA> what do they do that is worthy of notification?
[11:31] <bradb> register
[11:31] <SteveA> they're not anon then
[11:31] <SteveA> so that's okay
[11:31] <SteveA> they're registering
[11:32] <bradb> aren't they?
[11:32] <bradb> until they confirm their email?
[11:32] <mpt> sabdfl: So, okay to land?
[11:32] <SteveA> i don't mind giving such users a session
[11:34] <SteveA> bradb: i'm more concerned about making the site work well, even for anonymous users -- like with your breadcrumb innovation
[11:34] <SteveA> bradb: but allowing us to keep it scalable
[11:35] <bradb> what does session data get written to right now?
[11:35] <SteveA> in-memory zodb
[11:35] <SteveA> we'll be moving it to its own DB connection onto the main database eventually i think
[11:35] <SteveA> and that will support entirely persistent sessions, across restarts
[11:36] <SteveA> we might even use zodb and DB together
[11:36] <SteveA> because some session data is important -- like login stuff
[11:36] <SteveA> but some we can lose without consqeuence, like breadcrumb stuff
[11:37] <SteveA> and it is nice to be able to make innovations using the flexibility of the zodb
[11:39] <SteveA> oh poo
[11:39] <SteveA> i can't fix the vary header and have logins keep working
[11:39] <SteveA> because of what pound is doing
[11:39] <SteveA> i need to talk with stub
[11:41] <SteveA> as a workaround, i can use a separate cookie
[11:41] <SteveA> that is used simply to tell the browser that things have varied in the launchpad cookie
[11:44] <bradb> btw, re: the nav portlet, i was going to revisit it in front of users at UBZ, because mpt told me package subscriptions are more important (which I'm working on now.) and, though kiko liked the nav portlet right away, mpt's first words were to say that he didn't understand it. i'm hoping that people who have little familiarity with LP can shed some light on the idea one way or the other.
[11:46] <bradb> meanwhile, my UI feedback messages branch is moving at glacial speed through the code review process
[11:46] <mpt> bradb: I think we should have a general "What you can do next" section at the bottom of most pages
[11:46] <mpt> with contents that vary depending on the context
[11:46] <SteveA> i'm sure that we want some kind of navigation portlet, to show the context of where you've been, but we might want to make it work differently
[11:46] <mpt> so going to bugs in another product would fit well there.
[11:47] <bradb> yeah, i love menus that say, literally "Here's what you can do"
[11:50] <mpt> Oh.
[11:50] <mpt> SteveA: I was misreading the header results. It's not Launchpad setting everything's cacheability to 0; it's Firefox doing that.
[11:51] <mpt> Presumably because it's HTTPS.
[11:51] <SteveA> mpt: ok
[11:51] <mpt> So style sheet bandwidth does matter.
[11:54] <Kinnison> SteveA: fillum good
[11:58] <mpt> sabdfl? Sorry to bother you, but if it's ok, I'd like to land it before the bzr switchover if possible
[11:58] <mpt> because of the time that will take, and the difficulty of resolving CSS conflicts
[12:02] <sabdfl> mpt: can you bring this up somewhere i can look at it?
[12:04] <mpt> sabdfl: Sure, not that there's much to see :-)
[12:04] <mpt> It looks identical to Launchpad currently