[12:21] <kiko> darn
[12:21] <kiko> who is awake?
[12:22] <kiko> argh
[12:23] <kiko> https://launchpad.net/malone/bugtrackers/ubuntu-bugzilla
[12:23] <kiko> this is pretty cool
[12:25] <ajmitch> oh that's neat
[12:25] <ajmitch> how many other hidden features are there?
[12:26] <kiko> quite a few
[12:26] <ajmitch> would be nice if the debbugs status showed up on the bug watch :)
[12:27] <kiko> doesn't it?
[12:27] <kiko> oh, only bugzilla
[12:27] <allee> kiko: he! Nice! Can I dare to fix this myself: https://launchpad.net/products/malone/+bug/2877  or will my lauchpad account automaticly closed due to abusing it ;)
[12:28] <kiko> allee, you can fix it if you like, sure
[12:28] <kiko> https://launchpad.net/malone/bugtrackers/+newbugtracker
[12:28] <allee> kiko: 'k I found this already. let's try
[12:30] <kiko> do it!
[12:39] <dilys> Merge to rocketfuel@canonical.com/launchpad--devel--0: [r=stevea]  Buildd rescueBuildIfLost implementation, rescue slaves processing deleted or modified jobs. (patch-2580: celso.providelo@canonical.com)
[12:47] <allee> kiko: done https://launchpad.net/malone/bugtrackers/kde-bugs
[12:51] <allee> What is IMHO strange is that add link to external bug-tracker listen in lower right corner, while I somehow expected that in link to bugtracker one could choose 'upstream, distro: debian, suse, ...' and that it get's listed right at the top of the bug page
[01:24] <dilys> Merge to rocketfuel@canonical.com/launchpad--devel--0: r=kiko Fix for bug 2499: 'Unsubscribe from bug' page has incorrect documentation. Adds a section and conditionals for unsubscription. Patch by Diogo Matsubara <matsubara@async.com.br> (patch-2581: matsubara@async.com.br, christian.reis@canonical.com)
[01:30] <deb_user_ba> I cannot send my GPG key.. 
[01:35] <allee> Is there a way to register an already defined bug tracker (kde-bugs) with a project (KDE).  I always get the +newbugtracker form :(
[01:42] <deb_user_ba> Algum do Brasil??
[01:43] <kiko> allee, you can actually tie a bug task to a bug watch
[01:43] <kiko> bug you need to do it manually
[01:43] <kiko> deb_user_ba, do you have a sign-only key?
[01:43] <kiko> what is the error?
[01:43] <kiko> allee, I don't think projects and bug trackers can be associated
[01:44] <allee> kiko: well, https://launchpad.net/projects/kde  has a 'add a bug-tracker' link
[01:45] <allee> kiko: and it lead me to the same form I used to register kde-bugs from https://launchpad.net/malone/bugtrackers/+newbugtracker
[01:45] <kiko> oh
[01:45] <kiko> I see
[01:46] <lifeless> morning
[01:46] <deb_user_ba> Hi Kiko.. where are u from?? Do you speak portuguese??
[01:46] <allee> lifeless: late nite ;)
[01:47] <kiko> allee, can you file a bug: Impossible to associate existing bugtracker with project?
[01:47] <kiko> deb_user_ba, yeah, but I need to skip out -- can you please file a bug?
[01:47] <allee> kiko: yeap I now know how to do it ;)
[01:48] <kiko> thanks
[01:48] <deb_user_ba> kiko.. when I try send my GPG key.. 
[01:49] <deb_user_ba> i receive this msg:
[01:50] <deb_user_ba> Launchpad could not import GPG key, the reason was:HTTP Error 500: OK at http://keyserver.ubuntu.com:11371/pks/lookup?search=0x16BDCEF5&op=get.Check if you published it correctly in the global key ring (using gpg --send-keys KEY) and that you add entered the fingerprint correctly (as produced by gpg --fingerprint YOU). Try later or cancel your request.
[01:50] <deb_user_ba> GPG key is D652 32F3 7C85 6CB3 3363  36AC 92EE A639 16BD CEF5 ?!??!
[01:56] <deb_user_ba> kiko-afk??
[01:57] <allee> kiko: done https://launchpad.net/products/malone/+bug/2879
[03:28] <dilys> Merge to rocketfuel@canonical.com/launchpad--devel--0: [trivial]  Fix Bug#2182 (patch-2582: stuart.bishop@canonical.com)
[03:45] <salgado> stub: ping?
[03:47] <stub> salgado: pong
[03:47] <salgado> stub: I can't merge my shipit-exports branch because the cronscript runs with the shipit user, and so does its test
[03:48] <salgado> (and there's no shipit user on rocketfuel)
[03:48] <stub> Ahh... on chinstrap you mean. I sec...
[03:49] <stub> salgado: ok. Should be fine now.
[03:49] <salgado> yes, chinstrap. 
[03:49] <salgado> thanks, stub!
[03:49] <salgado> stub: also, what ranking constant should I use for the recipientdisplayname column? A?
[03:50] <stub> A in this case. It is the most important column when you are doing the searches.
[03:57] <salgado> stub: can we have another run of the shipit-exports script on staging, this time with the data migration patch?
[03:58] <stub> Sure. The staging update is running right now so I can merge your branch in as soon as that is done.
[04:04] <salgado> great. I just sent a merge request to pqm
[04:23] <salgado> thanks stub 
[04:38] <dilys> Merge to rocketfuel@canonical.com/launchpad--devel--0: Shipit exports and a lot more fixes. r=SteveA,stub (patch-2583: guilherme.salgado@canonical.com)
[04:44] <stub> lifeless: Another offering for mixing Python & C - http://www.scipy.org/documentation/weave/  
[04:45] <lifeless> stub: thanks
[06:54] <dilys> Merge to rocketfuel@canonical.com/launchpad--devel--0: [trivial]  Fix bugs in shipit data migration patch (patch-2584: stuart.bishop@canonical.com)
[07:48] <dilys> Merge to rocketfuel@canonical.com/launchpad--devel--0: [trivial]  Add new required sections to staging and production config files (patch-2585: stuart.bishop@canonical.com)
[09:02] <sivang> Good morning all!
[09:20] <jamesh> stub: ping?
[09:34] <carlos> morning
[09:34] <SteveA> hi
[09:43] <lifeless> morning
[09:59] <jamesh> stub: ping?
[10:00] <stub> jamesh: pong
[10:01] <jamesh> stub: was running some of the tests in my keyring-trust-analyzer branch, and ran into a new failure inside PersonSet.merge()
[10:02] <jamesh> stub: the problem seems related to the queries used to rename the merged account, which generate unicode query strings
[10:03] <jamesh> stub: since merge() is building its own queries, should it be responsible for converting to UTF-8?
[10:04] <stub> jamesh: Yes. But I don't see how that could possibly happen given that it should only be dealing with ASCII query strings and integer parameters.
[10:05] <jamesh> stub: it is creating a new name for the person based on from_person.name (which is unicode)
[10:05] <stub> (cursor.execute() requires UTF-8 strings and parameters, so anything talking at this level is responsible because nothing else can be)
[10:06] <jamesh> stub: so when it interpolates that name into the query string, you get a unicode string
[10:06] <jamesh> stub: this seems to be a fairly recent addition to the method
[10:06] <stub> jamesh: oic. It is ascii, but passing a unicode string (even if containing just ascii) futzes psycopg
[10:07] <jamesh> stub: okay.  I'll add an explicit conversion to unicode
[10:07] <jamesh> arg
[10:07] <stub> We should .encode('ascii') - merge() needs to be as paranoid as possible
[10:07] <jamesh> ^^ conversion to utf-8
[10:07] <jamesh> okay
[10:13] <stub> spiv: Did you get an email from the spec system asking for a review?
[10:15] <martin> servhelp
[10:16] <martin> Hello, what is the best channel to ask tech questions ( I want to setup a printer, for example)
[10:17] <carlos> martin, #ubuntu
[10:17] <martin> thanks carlos
[10:25] <zyga> morning
[10:37] <spiv> stub: I don't think so.
[10:39] <spiv> stub: https://launchpad.net/products/launchpad/+spec/librarian-garbage-collection ?
[10:39] <spiv> (at least it's pretty easy to find via my person page in launchpad)
[10:40] <WaterSevenUb> zyga, morning
[10:40] <zyga> WaterSevenUb: hello :)
[10:40] <zyga> WaterSevenUb: how are you 
[10:41] <lifeless> SteveA: how does launchpad detect one is logged in ?
[10:42] <WaterSevenUb> zyga, overloaded ;) 
[10:43] <zyga> WaterSevenUb: translating?
[10:43] <lifeless> how do we get specs listed  : https://launchpad.net/sprints/ubz
[10:45] <WaterSevenUb> zyga, I wish :) Preparing classes...
[10:45] <SteveA> lifeless: detect?
[10:45] <zyga> WaterSevenUb: you are a teacher?
[10:46] <SteveA> lifeless: there's a cookie that acts as a key to look up session data that is held in ram
[10:46] <lifeless> SteveA: I commonly have to log in N times
[10:46] <lifeless> SteveA: and many people complain about similar things
[10:46] <SteveA> lifeless: that sounds that some kind of problem
[10:46] <lifeless> do we set Vary ?
[10:46] <SteveA> almost certainly not
[10:46] <lifeless> i.e. Vary: Cookie
[10:46] <lifeless> then that is almost certainly the problem.
[10:47] <lifeless> erm, it might be varies. I'll check
[10:47] <SteveA> it's being served over https ... is that still relevant?
[10:47] <lifeless> yes
[10:47] <lifeless> there is a cache in your browser, and there are https proxies
[10:49] <lifeless> the -only- thing https gives you is confidence that noone the user didn't expect to see, sees the data you send and transmit
[10:49] <spiv> stub: looks good.
[10:49] <SteveA> so, Vary: * would fix these problems ?
[10:50] <SteveA> i can make the standard template and the shipit template set this
[10:50] <SteveA> as i don't want it to be set on images, css etc.
[10:51] <lifeless> SteveA: so. vary: * is a worst case catchall
[10:51] <SteveA> optimise later?
[10:52] <SteveA> a topic for UBZ ought to be opening launchpad to http where possible
[10:52] <SteveA> and that would only make such issues more significant
[10:52] <SteveA> in practice
[10:52] <lifeless> no, I think 'vary: cookie' is all that is needed
[10:52] <lifeless> if thats the only thing we use for deciding what we send, thats what we should emit
[10:53] <lifeless> if we vary on accept-language too, then very: cookie, accept-language
[10:53] <SteveA> does that need to be set for all pages, or just those when the cookie is sent
[10:53] <lifeless> it should be sent for any url where adding/removing or altering the listed headers may/will get you a different body
[10:54] <SteveA> it is bad to just set it for everything?
[10:54] <jamesh> SteveA: it should be set for any page where the cookie would alter how it is rendered
[10:54] <lifeless> not particiularly
[10:54] <jamesh> SteveA: which is pretty much everything, due to the user name in the header
[10:54] <SteveA> not images
[10:54] <SteveA> not css
[10:54] <jamesh> page header
[10:54] <lifeless> the things it would cause confusion on are in places we are not likely to encounter
[10:55] <lifeless> i.e. non conformant proxies from 1993
[10:55] <lifeless> putting it on 'static content' is not useful, nor is it harmful (especially if we support if-none-match)
[10:56] <SteveA> and it is set for all HTTP request methods?
[10:56] <lifeless> where we might send trivially different content, but we dont' *care* if the user gets different content, we don't need to set it either
[10:56] <lifeless> and yes, -all- methods on that url should emit vary
[10:56] <SteveA> ok
[10:56] <lifeless> and the vary string should be the same for -all- responses from that url, though it can be different across the site
[10:57] <SteveA> ok
[10:58] <SteveA> so, the probable cause of the problem you're seeing is that your browser, or some other cache, is cacheing the page rendered for anon. ?
[10:58] <SteveA> despite your having logged in
[10:58] <lifeless> this is not a panacea, we may find its not the root cause of the current issues, but it is likely contributing factor
[10:58] <SteveA> i don't see that Vary will help much in that case
[10:59] <lifeless> it will
[10:59] <SteveA> seeing as you can still (under some circumstances) have the same cookie
[10:59] <SteveA> but the data it points to says "no one is logged in"
[10:59] <SteveA> so, it doesn't necessarily vary
[10:59] <lifeless> do we not reset the cookie when we see that ?
[10:59] <SteveA> although, in practice, it will right now, i think
[11:00] <lifeless> ok, back for the lp meeting
[11:00] <SteveA> the default zope behaviour is to use the cookie at all times
[11:00] <SteveA> i think my hack is still in place
[11:01] <SteveA> that uses it (sets it) only when necessary
[11:01] <SteveA> but i may have removed that... not sure
[11:01] <lifeless> we should have a change visible to the client when login occurs
[11:01] <SteveA> so, i wonder if other no-cache headers are required
[11:01] <lifeless> i.e. new cookie, change the cookie, or set hard no-cache headers before login occurs.
[11:01] <SteveA> you want the cookie value to change when there's a change in session state
[11:01] <SteveA> where that change will affect the cacheability of the page
[11:02] <lifeless> yes
[11:02] <SteveA> okay, that makes sense to me
[11:02] <SteveA> i'll look into it
[11:02] <lifeless> because vary is a POST-hoc invalidation technique
[11:02] <SteveA> and
[11:02] <lifeless> no-cache is a foreknowledge technique
[11:02] <SteveA> you think we should use vary and variable cookies rather than no-cache etc.
[11:02] <lifeless> no-cache has its place
[11:03] <SteveA> okay, i think i know how to make the cookie session machinery do this
[11:03] <SteveA> this is stuff we should push upstream sometime
[11:03] <jamesh> it looks like the "fileresource" class handles if-modified-since style validation
[11:03] <jamesh> but doesn't do etags
[11:03] <lifeless> something that is expensive to generate but cheap to validate is a good candidate for vary
[11:03] <SteveA> thanks, see you at the baz meeting
[11:03] <lifeless> as are things that are cheap to validate but change after the fact
[11:04] <lifeless> SteveA: when is that, 2.5 hours from now ?
[11:04] <SteveA> thought it was 1 hr before the lp meeting. or is it 30mins before?
[11:04] <lifeless> oh, its one hour before, for 30 minutes
[11:04] <SteveA> ok
[11:04] <lifeless> so thats 2 hours yes ?
[11:04] <SteveA> yes
[11:04] <lifeless> till then, ciao
[11:04] <SteveA> iki
[11:06] <SteveA> so... the cookie value can be set as a combination of two things: the session key (HMAC encrypted and all that) and a session change index
[11:13] <ddaa> Nooooo! not the rubber duck!
[11:54] <stub> Is there a way to 'baz switch' a hardlinked tree without losing the hardlinkingness ?
[11:57] <jamesh> stub: "baz diff --link" should update hard links
[11:58] <stub> round about, but usable. Thanks )
[12:50] <carlos> ddaa, lifeless, jblack https://chinstrap.ubuntu.com/~dsilvers/paste/fileys8U6e.html
[12:51] <carlos> ddaa, lifeless, jblack any idea about how to solve that? I got it sometime ago, but forgot it until today that I did a new merge...
[12:51] <carlos> I can merge from rocketfuel, but I cannot commit the changes
[12:52] <ddaa> stub: nope
[12:52] <ddaa> but in practise it's not that much of big deal unless you are switching between trees that have a very old
[12:52] <ddaa> * that are very old
[12:53] <ddaa> you reap most of the performance benefits by inode comparison of patchlogs
[01:03] <carlos> ddaa, any idea about the problem I'm having?
[01:22] <ddaa> carlos: did somebody fixed your problem while my network was broken?
[01:22] <carlos> ddaa, no
[01:23] <ddaa> random try: delete the cached ancestry file in the broken revision
[01:23] <ddaa> figuring out the exact issue is a bit longish, and I'm in a meeting
[01:23] <carlos> ddaa, where is that cached ancestry file?
[01:24] <ddaa> sorry pal, after meeting
[01:24] <cprov> morning hackers
[01:24] <carlos> ok
[01:24] <carlos> stub, could you remove the -q flag from the poimport script?
[01:25] <carlos> stub, I don't know why but since yesterday is not importing anything
[01:31] <stub> carlos: Done
[01:31] <carlos> stub, thanks
[01:39] <ddaa> carlos: -> #bazaar
[01:45] <SteveA> developers meeting here, 15 mins
[01:45] <SteveA>  /msg me any meeting items
[01:45] <cprov> okay
[01:46] <lifeless> interesting ..
[01:46] <lifeless> https://launchpad.net/products/bzr/+spec/bzr-gpg-keysigning/+deptree
[01:46] <lifeless> claims it both doesn't block others, and lists the ones it blocks ;0
[01:47] <jordi> SteveA: should I do my lines now? I won't be available in a few mins
[01:48] <SteveA> jordi: /msg me them
[01:50] <jamesh> garr.  pyme has a module called "import"
[01:50] <SteveA> no way
[01:50] <SteveA> that's insane
[01:50] <jamesh> pyme.constants.import
[01:52] <Kinnison> M-8 ?
[01:56] <SteveA> 5 mins to meeting
[01:56] <SteveA> take a workrave now
[01:56] <SteveA> and /msg me any items
[01:56] <SteveA> but not both at once, obviously
[01:59] <niemeyer> Good morning!
[01:59] <lifeless> morning
[02:00] <jblack> hi
[02:00] <salgado> SteveA, my shipit-searching branch is marked needs-reply on your queue. was it me who set the wrong status on it? (I ask because I didn't get any review of it)
[02:00] <ddaa> hello
[02:00] <SteveA> it's time for the meeting
[02:00] <matsubara> morning all
[02:01] <SteveA> who's here?
[02:01] <jamesh> me
[02:01] <jblack> me
[02:01] <mpt> me
[02:01] <salgado> me
[02:01] <matsubara> me
[02:01] <lifeless> I think I can
[02:01] <bradb_> me
[02:01] <bob2> I know I can?
[02:01] <niemeyer> me
[02:01] <stub> yo
[02:01] <cprov> me
[02:01] <SteveA> kiko-afk: ?
[02:02] <ddaa> hey, that was before the meeting started!
[02:02] <SteveA> spiv, mbp and jordi send apologies
[02:02] <SteveA> i have their three sentences
[02:02] <SteveA> carlos: ?
[02:03] <jordi> I'm here sortf
[02:03] <SteveA> bjorn is on vacation
[02:03] <jordi> sortof even
[02:03] <SteveA> == Agenda ==
[02:03] <SteveA>   - roll call
[02:03] <SteveA>   - agenda
[02:03] <SteveA>   - next meeting
[02:03] <SteveA>   - activity reports
[02:03] <SteveA>   - production / staging
[02:03] <SteveA>   - slow queries
[02:03] <SteveA>   - issues for sysadmins
[02:03] <SteveA>   - menus delivery
[02:03] <SteveA>   - launchpad on bzr status
[02:03] <SteveA>   - DST
[02:03] <SteveA>   - three sentences
[02:03] <SteveA> 
[02:03] <SteveA> next meeting, same time next week?
[02:03] <cprov> yup
[02:03] <bradb> sure
[02:04] <jblack> aye
[02:04] <carlos> yeah
[02:04] <lifeless> HAH
[02:04] <jamesh> sure
[02:05] <jblack> jamesh: can we talk after the meeting?
[02:05] <jamesh> jblack: sure.
[02:05] <SteveA> activity reports: who's doing better than i am currently?  (i have them in gtimelog, but i've been too slack to send them.  bad wicked me)
[02:05] <stub> I'm up to date
[02:05] <SteveA> salgado: seen kiko?
[02:05] <lifeless> PERFECTO
[02:05] <mpt> up to date except yesterday
[02:05] <niemeyer> I'm ok as well
[02:05] <salgado> SteveA, no. want me to call him?
[02:05] <jblack> SteveA: I am, but barely. Fell into the "I can remember and write later"
[02:05] <SteveA> salgado: please
[02:05] <salgado> I'm up to date, btw
[02:06] <SteveA> carlos: start again today
[02:06] <SteveA> jamesh: please send today's right now, even if it doesn't include times
[02:06] <carlos> SteveA, will do
[02:06] <ddaa> uptodate by the day
[02:07] <SteveA> anyone not spoken?
[02:07] <jblack> thought you had one. :)
[02:07] <SteveA>   - production / staging
[02:07] <jordi> I'm totally behind on activity reports, agreed with carlos to do 1/week
[02:07] <SteveA> thought i was one...
[02:07] <SteveA> jordi: okay
[02:07] <SteveA> stub: plans for production, status of staging please?
[02:08] <SteveA> niemeyer: got a canonical url?
[02:08] <SteveA> there's a weekend coming, and i could do some hacking
[02:08] <niemeyer> SteveA: http://labix.org/pybot
[02:08] <SteveA> ta
[02:08] <stub> Production will be tagged for rollout next tuesday from he current head, unless people have features to land tonight or tomorrow.
[02:08] <stub> Staging has been running the shipit-exports branch for testing exports
[02:08] <lifeless> SteveA: also, there is the twisted irc protocol, its quite nice
[02:09] <stub> staging logs are being logged to a seperate directory ready for mirroring to chinstrap
[02:09] <stub> I think that is it
[02:09] <SteveA> stub: i'll want to get some sessions improvements cherrypicked later.
[02:09] <stub> cool
[02:09] <SteveA> production logs are already on chinstrip.
[02:09] <SteveA> and being rsynced
[02:09] <SteveA> to there from production
[02:10] <SteveA>   - slow queries
[02:10] <SteveA> how's all that looking?
[02:10] <SteveA> what's the current settings for query/request timeout in production?
[02:11] <stub> I havn't heard anything. Jordi or Carlos would know if that means the issues are hiding, or are still biting people regularly
[02:11] <carlos> we have problems with Rosetta
[02:11] <stub> timeeout was increased to 15 seconds after the second launchpad backend came online
[02:11] <SteveA> carlos: what are the problems?
[02:11] <carlos> mainly the +lang/code page, it's huge and takes too much time 
[02:11] <Kinnison> (On dogfood)
[02:12] <Kinnison> I'm concerned about how the db will behave when we put the distro on it
[02:12] <carlos> SteveA, people complain about system erros there too often and it's always related with timeouts
[02:12] <SteveA> Kinnison: i guess we'll want some special staging time for that
[02:12] <SteveA> carlos: timeouts should have their own error page now
[02:12] <carlos> SteveA, there are other reports about the translation form, but that's a bit more difficult to debug and it does not happen so often
[02:13] <Kinnison> SteveA: hopefully it won't matter. The only transactions deadlocks I see are on the publishing table
[02:13] <carlos> SteveA, so people will know that they are getting timeouts?
[02:13] <Kinnison> SteveA: and I'm working to eliminate them
[02:13] <SteveA> carlos: yes
[02:13] <carlos> cool
[02:13] <SteveA> carlos: so, what can we do about the language code page?
[02:13] <carlos> that will help us a lot
[02:13] <SteveA> hi kiko
[02:13] <kiko> grrr, sorry
[02:13] <SteveA> how are your activity reports?
[02:13] <carlos> SteveA, I will take a look at that page just in case I can see a way to improve the performance
[02:13] <kiko> perfectly up-to-date
[02:14] <SteveA> carlos: can you give an example of a live url to that page?
[02:14] <stub> Kinnison: There is always the option of lowering the transaction isolation level. So far we haven't needed to change it from the default (and highest) level (although the Librarian might be the first, depending on how spiv is going with librarian/rosetta issues)
[02:14] <carlos> if it's not possible... I suppose we should add any kind of batching or split it showing, for instance, most translated, normal translated and non translated sourcepackages
[02:14] <Kinnison> stub: what does that do?
[02:15] <Kinnison> stub: and is it possible to lower the isolation level on a per-connection basis?
[02:15] <SteveA> carlos: okay, i think you should talk with kiko and mpt about that sometime later on
[02:15] <carlos> SteveA, https://launchpad.net/distros/ubuntu/breezy/+lang/fr
[02:15] <jamesh> Kinnison: yes
[02:15] <kiko> carlos, I have been unable to trigger a timeout on that page yet, though I have seen reports on it
[02:15] <Kinnison> jamesh: if I google for transaction isolation level will I get useful stuff to read about it?
[02:16] <carlos> kiko, I got some of them already, it depends on the load of the server
[02:16] <SteveA> that page is HUGE
[02:16] <lifeless> Kinnison: lowering the isolation level is a non trivial exercise
[02:16] <carlos> SteveA, I know
[02:16] <lifeless> Kinnison: I would consider much else first
[02:16] <SteveA> there has to be a better way of arranging that
[02:16] <Kinnison> lifeless: right, I'm currently pondering a "Fucking with the publishing records" lock on disk
[02:16] <stub> Kinnison: http://www.postgresql.org/docs/7.4/interactive/mvcc.html
[02:16] <SteveA> i think that page is too large to be really useful
[02:17] <kiko> ah, cool -- carlos, I just got the timeout error! :)
[02:17] <stub> Kinnison: connection.set_isolation_level() lets you set the transaction isolation level per connection
[02:17] <jamesh> Kinnison: http://www.postgresql.org/docs/7.4/interactive/sql-set-transaction.html
[02:17] <carlos> SteveA, I think so, but mark said that batching is not trivial and that is better that page that nothing
[02:17] <SteveA> carlos: i don't think batching is the answer
[02:17] <kiko> me neither
[02:17] <carlos> kiko, some people getting it...
[02:17] <SteveA> carlos: i think looking at the use-cases for that page is
[02:18] <SteveA> and doing something that meets those use cases better than just a huge list
[02:18] <carlos> ok
[02:18] <SteveA> let us move swiftly onwards
[02:18] <SteveA>   - issues for sysadmins
[02:18] <lifeless> 32k files
[02:18] <SteveA> anything that still hasn't been done by sysadmins that they need hassling about?
[02:18] <lifeless> waiting for elmo to return
[02:18] <jblack> yeah.
[02:18] <SteveA> lifeless: noted, do you want me to do anything about that?
[02:19] <SteveA> i'm aware of all the issues
[02:19] <kiko> SteveA, salgado has a request but I think it's stub's, not sysadmins
[02:19] <salgado> kiko, already sorted
[02:19] <lifeless> SteveA: if you dont see me heckle elmo today, please channel my sarcastic nature
[02:19] <jblack> planet.bazaar.canonical.com is held on info request since thursday.
[02:19] <SteveA> lifeless: i'll give him a phone call when he's around
[02:19] <kiko> rock
[02:19] <SteveA> jblack: okay
[02:20] <SteveA> jblack: can you get around that by pointing the admins to the places where they need to set these things in the baz archive of the planet?
[02:21] <SteveA>   - menus delivery
[02:21] <jblack> SteveA: Not yet. I need the places they are going to install to first, for the config.
[02:21] <SteveA> the infrastructure is in place.  there should be no portlet-actions left, except in shipit pages.  i still need to update the docs. any issues / questions about menus.
[02:22] <salgado> SteveA, I renamed all portlet-actions in shipit pages
[02:22] <SteveA> salgado: thanks
[02:22] <mpt> Some templates may be missing the application menu they should have
[02:22] <salgado> btw, you reviewed that. ;)
[02:22] <mpt> When my design fascism branch is landed, I'll be going through all the templates to check them
[02:22] <SteveA> there's still a coupling between having canonical urls, and displaying menus
[02:22] <SteveA> i'm removing that as part of my "navigation" work
[02:22] <kiko> SteveA, there's still a handful of locations that I will be fixing today -- gneuman has the patch
[02:23] <kiko> gneuman, you need to get me that patch with the updated tests
[02:23] <SteveA> kiko: okay, fine.  grepping for portlet-actions in /templates/ should give no results
[02:23] <SteveA> and all the pages where there were some should be pagetested
[02:23] <SteveA> thanks mpt
[02:23] <SteveA>   - launchpad on bzr status
[02:23] <SteveA> lifeless: can you say a few words?
[02:23] <gneuman> k
[02:23] <lifeless> quardle ardle oodle ardle, the magpie said
[02:24] <SteveA> you didn't menion PIE
[02:24] <kiko> he did
[02:24] <kiko> the magnetic sort
[02:24] <lifeless> hheh
[02:24] <lifeless> so
[02:24] <lifeless> this last week I've been helping martin with weave
[02:24] <lifeless> and am cutting rc1 of 0.1 right now
[02:24] <jblack> excellent results, btw.
[02:24] <ddaa> clear the area, he's about to handweave!
[02:25] <SteveA> "If it wasnae for the weavers what would we do We widnae hae clothes made o' woo"
[02:25] <lifeless> but thats done, so I'm back on track, no more distractions :)
[02:25] <lifeless> sftp support is in the pipeline now, via john mienel
[02:26] <lifeless> and I'm actively hacking on gpg support from tomorrow
[02:26] <lifeless> switch will come right after that
[02:26] <ddaa> yay!
[02:26] <lifeless> but I have a sinking feeling about PIE
[02:26] <jblack> is there a spec for gpg? 
[02:26] <lifeless> jblack: yes
[02:26] <lifeless> https://launchpad.net/products/bzr/+spec/bzr-gpg-keysigning
[02:26] <kiko> lifeless, dude, regardless of pie, you rock
[02:26] <kiko> great work
[02:26] <lifeless> kiko: thanks :)
[02:27] <SteveA> this is looking very promising
[02:27] <kiko> lifeless, more realistically, do you think we'll be on bzr for rocketfuel post-ubz?
[02:27] <kiko> I mean, november/decemberish?
[02:27] <lifeless> oh yes
[02:27] <lifeless> I mean, I think if I could make life uncomfortable for you for a bit, it could be before ubz
[02:27] <lifeless> but I'd rather not do that
[02:28] <kiko> what's the chance of showstoppers popping up? very high, high, medium, low?
[02:28] <lifeless> low
[02:28] <kiko> ok
[02:28] <SteveA> item from the agenda i'd forgotten:
[02:28] <SteveA>  - https://launchpad.net/products/launchpad/+specs
[02:28] <lifeless> :1
[02:28] <kiko> lifeless, what are the major bits missing? sftp and gpg, then?
[02:28] <lifeless> http://bazaar.canonical.com/CanonicalBzrDogfooding
[02:28] <lifeless> doco
[02:29] <lifeless> converter is DONE (for rocketfuel needs)
[02:29] <lifeless> jamesh script is DONE
[02:29] <lifeless> rollout scripts need adjusting, or maybe I can lever stub into that
[02:29] <lifeless> switch is needed
[02:29] <lifeless> gpg signing is needed
[02:29] <lifeless> commit-with conflict prevention is DONE in abentleys branch, will merge monday
[02:29] <lifeless> symlinks are DONE
[02:30] <bradb> do we need switch before starting to use it with rf?
[02:30] <kiko> is switch the same as baz switch?
[02:30] <lifeless> yes
[02:30] <SteveA> yes, we need switch
[02:30] <lifeless> switch is preserve what I have here, give me another branch in place, kthankxbye
[02:30] <kiko> ok
[02:30] <ddaa> bradb: needed no change launchpad version easily (because of nested trees)
[02:30] <lifeless> niemeyer: like a breeze :)
[02:31] <niemeyer> Anyway, later..
[02:31] <niemeyer> lifeless: :)
[02:31] <bradb> ddaa: right
[02:31] <jblack> if he can pull off gpg signing, he can do switch
[02:31] <SteveA> okay... gotta move on
[02:31] <niemeyer> jblack: I think so :)
[02:32] <SteveA> so, the launchpad specs that are on the wiki are appearing in the launchpad product's spec-tracker thanks to matsubara and kiko
[02:32] <bradb> I added tons of Malone ones too, btw ;)
[02:32] <SteveA> please register any new specs you write with teh spec tracker
[02:32] <SteveA> thanks bradb 
[02:33] <SteveA> this will help us to keep track of the bigger picture of what we need to do
[02:33] <bradb> I also fixed the bug that was preventing specs from being assigned to milestones.
[02:33] <SteveA> questions about the spec tracker?
[02:33] <SteveA> cool
[02:34] <carlos> no questions here...
[02:34] <SteveA>   - DST
[02:34] <SteveA> spiv said:
[02:34] <SteveA> The only thing I'd want to raise is that daylight savings time will soon be
[02:34] <SteveA> happening, so 1200 UTC will become 11pm for me rather than 10pm, so I'd be
[02:34] <SteveA> happy to see it shift to 1100 UTC in a few weeks.
[02:34] <SteveA> 
[02:35] <SteveA> how does that work for jblack and bradb and brazilians?
[02:35] <kiko> 8am here..
[02:35] <jblack> what is it relative to now? Earlier or later? 
[02:35] <lifeless> thats one hour earlier right ?
[02:35] <SteveA> yes
[02:35] <niemeyer> SteveA: He wants to shift the UTC time?
[02:35] <kiko> oh, well, we also enter DST soonish
[02:35] <SteveA> yes
[02:35] <jblack> Sure for me, if we can make the bzr meeting one hour later.
[02:36] <jblack> i.e. swap places
[02:36] <bradb> I think that would be 7am here, but whatever. I'm planning to start waking up earlier anyway.
[02:36] <SteveA> jblack: i think we'd just pick a new time for the bzr meeting
[02:36] <niemeyer> kiko: Yes, with our "completely predictable rules" :(
[02:36] <bradb> The morning hours are good.
[02:36] <SteveA> bradb: http://www.stevepavlina.com/blog/2005/05/how-to-become-an-early-riser/  ;-)
[02:36] <kiko> niemeyer, I discovered tz-brazil!
[02:36] <kiko> or rather
[02:36] <kiko> niemeyer, I discovered tz-brasil!
[02:36] <ddaa> morning is a creation of the devil
[02:37] <jblack> I absolutely detest very early morning meetings.
[02:37] <niemeyer> kiko: And what would be that?
[02:37] <bradb> SteveA: interesting blog
[02:37] <jblack> I'll do it for the team. :) but I hate it.
[02:37] <bradb> RSS is stealing my life
[02:37] <SteveA> jblack: apparently, it helps if you cut out the caffiene ;-)
[02:37] <kiko> niemeyer, a debian package that does everything for you
[02:38] <SteveA> bradb: i go to lunch with people who read blogs, so i get mine second hand :-)
[02:38] <jblack> I think you're trolling! Me? No caffiene? 
[02:38] <jblack> I'd sleep for a year
[02:38] <bradb> SteveA: heh
[02:38] <niemeyer> kiko: Nice! Thanks for pointing it
[02:39] <SteveA> okay, we don't need to make the decision today.  assuming the baz meeting is moved, would 1100 UTC be a disaster for anyone, once DST has changed?
[02:39] <jblack> no
[02:39] <cprov> kiko: thanks pzn .. the only crap stuff is the list of 'changing time days' download from somewhere ...
[02:39] <SteveA> okay... it is that time
[02:39] <SteveA>   - three sentences
[02:39] <Kinnison> SteveA: It'd be slightly irritating for us northerners, but hey :-)
[02:40] <SteveA> speak your brains
[02:40] <lifeless> DONE: weave-many tweaks, unifed text and metadata revision recording, reduce shotgun surgery, update baz2bzr, implement ghost support in weave, release weave rc1
[02:40] <lifeless> FIRST POST
[02:40] <lifeless> TODO: bzr transactions, gpg signing
[02:40] <lifeless> BLOCKED: None
 DONE: bzr 0.1 almost ready to go; many fixes towards that
 TODO: actually release it
 BLOCKED: no
 DONE: mainly reviewing
 TODO: reviews, AuthServerCaching
 BLOCKED: no
[02:40] <SteveA> DONE: reviews, management, navigation components
[02:40] <SteveA> TODO: land navigation components, implement Vary header properly
[02:40] <SteveA> BLOCKED: no
[02:40] <jblack> DONE: bzr support, documentation
[02:40] <cprov> DONE: rescuebuilder and buildUI fixes
[02:40] <cprov> TODO: builddUI spec split, investigate some way to reuse code
[02:40] <cprov> BLOCKED: none
 DONE: FAQ, new imports, mailing list replies.
 TODO: new import requests, new replies to rosetta@ email, cleanup "review-breezy*" templates
 BLOCKED: "review-breezy", as I can't edit templatyes. Carlos is working on it right now afaik. #2802
[02:40] <mpt> DONE: design fascism, text-to-html redesign, bug listing cleanup, bugfixes
[02:40] <mpt> TODO: template menu checking, bug 2194, branch landings, bugfixes
[02:40] <mpt> HINDRANCES: baz is very slow
[02:40] <ddaa> DONE: dusted off and finished importd-archivelocation
[02:40] <ddaa> TODO: land stuff in review, BranchDataStorage db changes and infrastructure fixes, finish LaunchpadBranches, upgrade to breezy
[02:40] <ddaa> BLOCKED: lifeless python patch, ext3 32k links limit for python import
[02:40] <Kinnison> DONE: buildd sequencer, protocol rework, auto-giveback, tidying publisher
[02:40] <Kinnison> TODO: Upload handler (still)
[02:40] <Kinnison> BLOCKED: Elmo being around to actually go through what's needed for dapper to open
[02:40] <jblack> TODO: documentation, packaging for pqm/cm
[02:40] <salgado> DONE: ShipIt, helping out matsubara and gneuman.
[02:40] <jblack> BLOCKED: None new.
[02:40] <stub> DONE: Database fixture updates, LibrarianGarbageCollection spec
[02:40] <stub> TODO: LibrarianGarbageCollection
[02:40] <stub> Blocked: Nothing
[02:41] <bradb> DONE: Landed lots of usability fixes. TODO: Land some more usability fixes. BLOCKED: No.
[02:41] <salgado> TODO; More shipit, more help to matsubara and gneuman, maybe some code review
[02:41] <niemeyer> DONE: Released svn2bzr, merged revisions logging on bzr, prefixed stores on bzr, SmartPM maintenance, updated IRSS, etc
[02:41] <salgado> BLOCKED: No
[02:41] <kiko> DONE: drag through bug triage and code reviews
[02:41] <niemeyer> TODO: Two more days on bzr, work on launchpad, etc
[02:41] <SteveA> oh, and TODO: review stub's browser messages branch
[02:41] <carlos> DONE: More language packs small cleanups, merge daf's branch about 'tab' higlights, and many user support and debugging
[02:41] <niemeyer> BLOCKED: Nope
[02:41] <kiko> TODO: more reviews, spec transferral, more triage
[02:41] <kiko> BLOCKED: no
[02:41] <carlos> TODO: New import/attach system, give jordi all needed rights to do his work
[02:41] <carlos> BLOCKED: no
[02:42] <kiko> mpt, you must do something special, because, honestly, baz never is the blocker for me
[02:42] <kiko> mpt, have you tried keeping two trees instead of baz switching all the time?
[02:42] <jordi> carlos: go carlos :D
[02:42] <jamesh> DONE: code reviews, integrate lifeless's bzr support into pending-reviews and fix a few bugs, look at sqlobject object expiry for zopeless
[02:42] <jamesh> TODO: code reviews, specs stuff, sqlobject expiry
[02:42] <jamesh> BLOCKED: no 
[02:42] <kiko> jamesh, lifeless: what about the gpg id scanner thingy?
[02:42] <mpt> kiko: no, I haven't tried that
[02:43] <kiko> mpt, I've only suggested it to you like 10 times by now
[02:43] <jamesh> kiko: was going to merge it today, but the new tests showed up a few unrelated problems
[02:43] <mpt> itym once
[02:43] <niemeyer> ddaa: "ext3 32k links limit for python import", is that about bzr?
[02:43] <jamesh> kiko: should get it merged tonight
[02:43] <mpt> in April or so
[02:43] <kiko> jamesh, cool
[02:43] <kiko> mpt, no, you just have a bad memory
[02:43] <SteveA> niemeyer: no, baz
[02:43] <ddaa> niemeyer: baz-importd import of python, linked to baz archive storage being demanding on the fs and ext3 sucking hard.
[02:44] <SteveA> okay, so i have various "elmo time" stuff that is blocking people
[02:44] <SteveA> but nothing else
[02:44] <SteveA> is that so?
[02:44] <carlos> SteveA, kiko I'm not blocked directly by this, but Rosetta usage starts to be blocked... we need gina run on production... forgot to add it as part of today's agenda, sorry
[02:44] <niemeyer> ddaa: Ahh, understood
[02:44] <kiko> carlos, that's blocked by the gpg script, which is why I'm asking jamesh 
[02:44] <carlos> ok
[02:44] <SteveA> carlos, kiko, jamesh: let's get this sorted out after the meeting
[02:44] <kiko> sure
[02:44] <SteveA> thanks
[02:44] <niemeyer> ddaa: The prefixed stores in bzr fixes exactly that
[02:44] <kiko> though apparently it's on track
[02:45] <SteveA> we have 45 seconds more.  final words anyone?
[02:45] <bradb> One small note about the specs in LP: sabdfl would probably be really happy if the appropriate specs were assigned to 1.0 (which I've already done for Malone)
[02:45] <niemeyer> Yes, thanks!
[02:45] <niemeyer> :)
[02:45] <ddaa> niemeyer: nah fixing a different problem. The problem is limit on link count, that hits dirs with many subdirs, not dirs with many files.
[02:45] <SteveA> yeah, good call bradb
[02:45] <mpt> kiko: And meanwhile you keep whining at me for the amount of disk space I'm using, so I wasn't taking the suggestion seriously :-)
[02:45] <SteveA> okay, it's a wrap
[02:45] <niemeyer> ddaa: Ack
[02:45] <SteveA> MEETING OVER
[02:45] <SteveA> thanks everyone.
[02:45] <SteveA> i'm off to get some lunch
[02:45] <salgado> SteveA, wait
[02:45] <lifeless> niemeyer, ddaa - cross purposes. ddaa is dealing with a 'baz C version' problem.
[02:46] <SteveA> salgado: i'm hungry, dude
[02:46] <salgado> SteveA, did you review my shipit-searching branch or it's only its status that's wrong in the PendingReviews page?
[02:46] <kiko> mpt, it's only an extra half-gig per tree, so I don't think it's a big deal
[02:46] <kiko> mpt, your huge revlib is a lot more worrisome
[02:46] <carlos> kiko, jamesh or are we going to have the meeting now, before jamesh goes to sleep?
[02:47] <SteveA> salgado: probably not.
[02:47] <salgado> SteveA, okay, just wanted to make sure
[02:47] <kiko> jamesh, carlos: do we need to discuss anything, or are we on track?
[02:48] <kiko> ddaa, mpt has woken up with a 2.5gig revlib. do you have any clue what we can do to reduce this?
[02:48] <ddaa> first thing would be using something else than macosx
[02:48] <mpt> ddaa: It's called "Ubuntu 5.04"
[02:48] <jblack> kiko: He trim older less interesting stuff
[02:48] <kiko> ddaa, he's on linux.
[02:48] <jamesh> kiko/carlos: note that my code for adding users is capable of merging existing accounts too
[02:48] <carlos> kiko, don't know, I only need gina run, if we can talk about a deployment plan....
[02:49] <ddaa> okay, looking for right link
[02:49] <kiko> jamesh, that's pretty cool
[02:49] <kiko> jamesh, does that need extra knobs or is it the default?
[02:49] <kiko> jamesh, I imagine you can merge people with no passwords, at least..
[02:49] <ddaa> mpt: https://wiki.launchpad.canonical.com/LaunchpadHackingFAQ#head-7ede78aa7deeb4b4c9e649ad7bf58422279436f6
[02:49] <jamesh> so it should be able to clean things up if gina creates more people than necessary
[02:49] <kiko> jamesh, are you making sure to ignore merged accounts?
[02:49] <kiko> ah, cool
[02:49] <jamesh> kiko: it will merge accounts if no more than one account has a preferred email
[02:50] <jamesh> kiko: the person lookup code skips merged accounts (which don't have associated email addresses anyway)
[02:51] <ddaa> mpt: better if using xfs, ext3 has very bad behaviour with revlibs, wasting tons of space with directories.
[02:51] <kiko> jamesh, hmmm, right.
[02:51] <kiko> ddaa, mpt: I'm considering the switch to xfs this week, stay tuned.
[02:51] <ddaa> but anyway 2.5G is not terribly large for a revlib
[02:51] <ddaa> mine is 1.5G
[02:51] <mpt> ddaa: ok, thanks
[02:52] <kiko> ddaa, it's 1G overweight
[02:52] <mpt> SteveA: Does https://launchpad.net/products/malone/+bug/2115 (POST form submission while not logged in) need a BoF?
[02:52] <jamesh> kiko: you could stick your trees on one of these: http://www.theregister.co.uk/2005/10/03/gigabyte_i-ram_ramdisk/ :)
[02:52] <ddaa> it's 0.6x excess... not 10x, is what I mean
[02:53] <kiko> mpt, why not? I've already outlined what we need to do once..
[02:54] <jbailey> salgado / kiko: When you had me poke at shipit.ubuntu.com to look at the LP integrated version, did my request get kept or discarded?
[02:54] <kiko> jbailey, if it was on that address, it was kept
[02:54] <jbailey> I think it wasn't, I think it was a staging site.
[02:54] <mpt> kiko: You mean we shouldn't because you already know what to do?
[02:54] <mpt> if so, cool
[02:54] <jbailey> I just realised that although I *think* I've alrady put in for my Ubuntu CDs that perhaps I haven't =)
[02:55] <Nafallo> jbailey: if you login you will see ;-)
[02:55] <jbailey> Nafallo: to shipit or launchpad?
[02:55] <salgado> jbailey, to shipit
[02:55] <kiko> mpt, well, nobody's stepped up to implement it, and it does look like a SteveA sort of thing.
[02:56] <jbailey> It doesn't mention anything.  I'll put it in again.
[02:56] <jbailey> Thanks!
[02:56] <salgado> stub, ping
[02:56] <stub> salgado: pong
[02:58] <carlos> kiko, I'm leaving to have lunch will talk later (if needed) ok?
[02:58] <kiko> sure, carlos 
[02:58] <salgado> stub, what fix was needed in the data migration script?
[02:58] <carlos> later!
[02:58] <kiko> jblack, ddaa: is there a script that cleans up old revisions from the revlib?
[02:59] <stub> salgado: There was a missing '%' sign creating a syntax error, and the data being retrieved from the database needed to be converted to Unicode before being passed to the name generation routine
[02:59] <jbailey> Is there a component in Malone for bugs in shipit?
[02:59] <ddaa> kiko: I just have small pybaz script here that removes evenything but the last revision of every branch
[02:59] <kiko> ddaa, that would be cool to run
[02:59] <salgado> jbailey, yes. https://launchpad.net/products/shipit/+filebug
[02:59] <jbailey> salgado: Tx
[03:00] <salgado> jbailey, or https://launchpad.net/products/shipit/+bugs to see the existing ones
[03:00] <salgado> stub, sorry for that, dude.
[03:00] <jbailey> salgado: Cool, thanks.  Looks like a new one. =)
[03:00] <ddaa> kiko: https://chinstrap.ubuntu.com/~dsilvers/paste/fileTPA2Qu.html
[03:00] <stub> no worries
[03:02] <salgado> stub, is it going to be possible to cherrypick it?
[03:03] <stub> salgado: I'll give it a go
[03:06] <kiko> stub, salgado: jane /really/ wants to be able to run an export this week. what if we told marilize to use staging today?
[03:10] <salgado> kiko, I'm not following you
[03:10] <kiko> salgado, to run a test export?
[03:11] <salgado> kiko, we already have a test export
[03:11] <kiko> oh
[03:11] <kiko> was I not copied on these emails?
[03:11] <salgado> shipit.staging.canonical.com
[03:11] <salgado> no emails
[03:11] <kiko> has marilize looked at it?
[03:11] <salgado> Jane looked at the previous one
[03:12] <salgado> and asked me to fix some things. now I fixed them and this new export was generated today
[03:12] <kiko> cool
[03:12] <kiko> over IRC, I imagine?
[03:13] <salgado> what?
[03:14] <jbailey> Err.  How do I delete an attachment in Malone?
[03:15] <kiko> salgado, this communication -- I saw no email on the topic
[03:16] <salgado> kiko, yes, it was over IRC
[03:16] <kiko> okay
[03:18] <wickedpuppy> hi guys ... i got problem importing ssh key and gpg key into launchpad ... any help ?
[03:18] <kiko> wickedpuppy, always. what happens?
[03:18] <wickedpuppy> i got error malformed keys
[03:18] <kiko> wickedpuppy, first, is it a sign-only key?
[03:18] <wickedpuppy> perhaps i am pasting wrong key ?
[03:19] <wickedpuppy> well ah ... how i know its signed only ?
[03:19] <kiko> wickedpuppy, hmmmm. perhaps -- did you follow the instructions?
[03:19] <wickedpuppy> gpg --gen-key 
[03:19] <kiko> wickedpuppy, what does gpg --list-keys say?
[03:19] <wickedpuppy> and i followed the instructions
[03:20] <wickedpuppy> should i paste ?
[03:20] <wickedpuppy> its more than 3 lines
[03:20] <wickedpuppy> ok 3 lines
[03:20] <wickedpuppy> lol
[03:20] <wickedpuppy> pub   1024D/B8C8C194 2005-10-06
[03:20] <wickedpuppy> uid                  billy am <wickedpuppy@gmail.com>
[03:20] <wickedpuppy> sub   2048g/47752BC1 2005-10-06
[03:20] <kiko> looks correct
[03:21] <wickedpuppy> so which is the key that i should paste into launchpad ?
[03:21] <wickedpuppy> i tried 1024D/B8C8C194 , B8C8C194 , 2048g/47752BC1 , 47752BC1 .. all error
[03:23] <wickedpuppy> eh ah guys ... which is the key figerprint ?
[03:23] <Nafallo> wickedpuppy: not there...
[03:24] <kiko> wickedpuppy, what are you pasting into the box, exactly?
[03:24] <wickedpuppy> i tried 1024D/B8C8C194 , B8C8C194 , 2048g/47752BC1 , 47752BC1 .. all error
[03:24] <wickedpuppy> billy@wicked1:~/programming/perl$ gpg --fingerprint billy am
[03:24] <wickedpuppy> pub   1024D/B8C8C194 2005-10-06
[03:24] <wickedpuppy>       Key fingerprint = 1DA6 FF19 7A20 78C4 53EE  1338 58A3 6668 B8C8 C194
[03:24] <Nafallo> 1DA6 FF19 7A20 78C4 53EE  1338 58A3 6668 B8C8 C194
[03:24] <Nafallo> paste that
[03:24] <wickedpuppy> i also tried the key fingerprint ..
[03:25] <kiko> that's what you need to paste in
[03:25] <kiko> I will improve the text on that page.
[03:25] <wickedpuppy> Launchpad could not import GPG key, the reason was:HTTP Error 500: OK at http://keyserver.ubuntu.com:11371/pks/lookup?search=0xB8C8C194&op=get.Check if you published it correctly in the global key ring (using gpg --send-keys KEY) and that you add entered the fingerprint correctly (as produced by gpg --fingerprint YOU). Try later or cancel your request.
[03:25] <wickedpuppy> got this
[03:25] <kiko> wickedpuppy, you need to publish your key to a public keyserver first. have you done that?
[03:26] <wickedpuppy> eh no ... how ?
[03:26] <wickedpuppy> i gotta get a front-end for gpg ...
[03:27] <Nafallo> (using gpg --send-keys KEY)
[03:27] <kiko> gpg --send-keys B8C8C194
[03:27] <kiko> that's all
[03:27] <wickedpuppy> alrighty
[03:27] <wickedpuppy> let me go back to launchpad
[03:28] <wickedpuppy> is it immediate ?
[03:28] <wickedpuppy> lol
[03:28] <cprov> wickedpuppy: it won't be immediatly available
[03:28] <wickedpuppy> ah k
[03:28] <wickedpuppy> thanks guys :P
[03:29] <wickedpuppy> now i go figure out my ssh-keygen
[03:29] <wickedpuppy> lol
[03:29] <kiko> heh
[03:32] <wickedpuppy> alrighty
[03:33] <wickedpuppy> done both ssh and gpg
[03:33] <wickedpuppy> thanks guys :P
[03:33] <kiko> salgado, ping?
[03:33] <Nafallo> fwiw, I think the instructions are were straightforward if one reads them ;-)
[03:34] <salgado> kiko, pong
[03:34] <Nafallo> s/were/very/
[03:35] <kiko> salgado, I'm going to override the ubuntu images, ok?
[03:35] <kiko> in shipit
[03:35] <kiko> via css
[03:35] <kiko> they are breaking the ssl padlock, I think
[03:36] <salgado> kiko, isn't it because we import the css from http://www.ubuntu.com?
[03:36] <kiko> salgado, I think the CSS may be okay, but the images now
[03:36] <kiko> not
[03:36] <kiko> I need to test
[03:36] <kiko> btw
[03:36] <kiko> Expression: <PathExpr standard:'view/processForm'>
[03:37] <kiko> ProgrammingError: ERROR: expression too complex DETAIL: Nesting depth exceeds maximum expression depth 10000. HINT: Increase the configuration parameter "max_expr_depth". SELECT COUNT(*) FROM
[03:37] <dilys> Merge to rocketfuel@canonical.com/launchpad--production--1.35: Charry pick steve.alexander@canonical.com/launchpad--production--1.35 into production (patch-7: steve.alexander@canonical.com)
[03:37] <kiko> ShippingRequest WHERE id in (... billions ...)
[03:37] <kiko> stub, we're in the charry picking business too?
[03:38] <stub> I suspect not and that there were rocketfuel or other merges into that branch
[03:39] <salgado> stub, what if I replay all my changes into a branch of production?
[03:39] <stub> salgado: That would work
[03:39] <stub> I can do it here too probably
[03:40] <salgado> stub, there'll be some conflicts
[03:41] <salgado> stub, what production branch are we running?
[03:41] <stub> rocketfuel@canonical.com/launchpad--production--1.35
[04:03] <kiko> Kinnison, cprov: what is auto-depwaiting?
[04:04] <bradb_> Do we have a specific(ish) date in mind for when we're going to move to Zope 3.1?
[04:04] <Kinnison> kiko: It's actually auto-un-depwait-ing
[04:04] <Kinnison> kiko: as it were
[04:04] <Kinnison> kiko: It's a tool for automatically re-trying builds which failed due to a missing dependency
[04:06] <kiko> ah, okay
[04:12] <kiko> Kinnison, one-line summary of "Build Daemon protocol rework. r=spiv"
[04:13] <Kinnison> kiko: Clean up the buildd master<->slave protocol to get rid of accumulated cruft and implement an important pair of features (namely Ogre and arch_indep control).
[04:14] <kiko> what are these features for?
[04:16] <kiko> bradb, can we make it so that +bug and +bugs traverse identically? having +bugs and +bug is the most crackful idea I've ever heard of, but it's particularly bad for URL-hacking when triaging
[04:16] <mpt> +1
[04:16] <bradb> kiko: It's nasty indeed.
[04:17] <kiko> we can then phase out +bug when nobody's looking
[04:17] <bradb> kiko: I wonder though...can we improve the UI in some way to make you not have to hack the URL?
[04:17] <bradb> kiko: I've been thinking of next/prev links on bug pages and things like that
[04:18] <kiko> bradb, well, I am reading commit logs and picking up bug numbers. the fastest way to get to them is using my browser history, but then I need to delete the trailing "s" from +bugs to get to the bug 
[04:18] <kiko> bradb, can we get a malone google search plugin?
[04:18] <Kinnison> kiko: 'ogre' is about making sure things only build with the right components available to them
[04:18] <kiko> Kinnison, "only build"?
[04:19] <bradb> kiko: meaning being able to search google from Malone, or...?
[04:19] <Kinnison> kiko: the arch_indep stuff is about building the _all.u?deb files on the right arch for a given distrorelease
[04:19] <bradb> or being able to search Malone with google?
[04:19] <mpt> bradb: Having a great everything-else interface is no excuse for having a bad URL interface
[04:19] <Kinnison> kiko: I.E. the apt sources in the chroot only turn on main for main, main+universe for universe, etc
[04:19] <kiko> bradb, s/google/firefox and it might make sense :)
[04:19] <bradb> mpt: I'm not defending the URLs, btw. ;)
[04:19] <kiko> I know you're not
[04:20] <kiko> Kinnison, ah, of course. thanks.
[04:20] <Kinnison> kiko: no probs
[04:20] <Kinnison> kiko: anything else I can tell you?
[04:20] <Nafallo> Kinnison: why did you name it ogre? :-)
[04:20] <Nafallo> my server is named ogre ;-)
[04:21] <kiko> Kinnison, there are many things, but this is a great start
[04:23] <Kinnison> Nafallo: I believe the name 'Ogre' comes from the fact that ogres are like onions
[04:23] <Kinnison> (They've got databases^Wlayers)
[04:23] <kiko> ogres have layers? that's news to me
[04:23] <Kinnison> kiko: have you not seen shrek?
[04:24] <kiko> oh-oh
[04:24] <Nafallo> ehm, have we talked about this before? half a year or so ago? :-P
[04:25] <kiko> Kinnison, what are _all.u?deb files?
[04:25] <Kinnison> kiko: debs and udebs which are architecture independant
[04:25] <Kinnison> I.E. the filename ends _all.deb or _all.udeb
[04:25] <kiko> why do they need to be built on the right arch?
[04:28] <Kinnison> because we only want them to build once, or there'll be clashes in the archive
[04:28] <Kinnison> so we have a nominated architecture which builds the _all as well as the _<arch> files
[04:28] <kiko> which architecture is that?
[04:28] <Kinnison> I believe for ubuntu it's i386
[04:29] <kiko> thanks
[04:29] <bradb_> argh
[04:29] <kiko> Kinnison, 
[04:29] <kiko> [r=stevea]  Buildd rescueBuildIfLost implementation, rescue slaves processing
[04:29] <kiko> +deleted or modified jobs
[04:29] <kiko> can you translate? :)
[04:30] <bradb_> mpt: Any idea how hard it would be to add a search box on the right side above the tabs inside the bugs facet, with a label like "Search ${context/targetname} bugs"?
[04:32] <kiko> bradb_, it should probably be under the tabs
[04:32] <kaliwanagan> hi all. is there a way to delete a product series?
[04:34] <bradb_> kiko: Hm, interesting point.
[04:35] <bradb_> maybe in a portlet instead?
[04:35] <bradb_> i.e. bugtarget-portlet-search.pt
[04:35] <bradb_> s/i.e./e.g./
[04:36] <kiko> bradb_, sure, that's a cool idea
[04:36] <Kinnison> kiko: If a slave is processing a deleted or otherwise modified job, it would previously be lost to the system. I.E. not IDLE, but not recorded as doing any work in the db
[04:36] <Kinnison> kiko: cprov's patch makes sure that slaves doing stuff we don't know about get cleaned up and reintroduced to the work pool
[04:37] <salgado> Kinnison, do you know about any upper limit in the number of results a subselect can return without breaking, in postgres?
[04:37] <bradb_> Right, must get the sorting stuff out of the way so I can DOIT. A search portlet would seem to have some pretty serious ROI. In adding a search portlet, I should be able to sneak in making +bugs work as well, but this portlet should hopefully avoid you wanting to hand-hack URLs.
[04:37] <kiko> Kinnison, thanks.
[04:39] <Kinnison> salgado: Nup
[04:39] <Kinnison> salgado: I thought a subselect was effectively merged into the query when run anyway
[04:41] <salgado> Kinnison, stub once told me about a 10000 results limit, but I'm not sure in what circunstances it apply
[04:46] <kaliwanagan> if not delete a product series, then at least any product release?
[04:46] <kaliwanagan> https://launchpad.net/products/openkore seems to have some duplicate releases
[04:46] <kaliwanagan> four infact, all leading to the same directory.
[04:46] <kiko> kaliwanagan, yeah, this is actually a dupe, I need to find the original bug and sic matsubara on it
[04:47] <kaliwanagan> oh i see. thanks for the heads up :)
[04:51] <mpt> bradb: As kiko suggested, "above the tabs" and "inside the bugs facet" would be a contradiction
[04:52] <mpt> bradb: What pages are you talking about adding this too?
[04:52] <Kinnison> lifeless: Erm, you're sending out {virus} warnings
[04:57] <SteveA> mpt, kiko: for the POST submission login thing, i need to consult with lifeless on exactly how to deal with the HTTP aspects
[04:59] <kiko> ah, perhaps a bof then
[05:00] <kaliwanagan> thanks all. I've received an email confirming that the bug has been assigned and is being worked on. :)
[05:10] <kiko> jordi, carlos?
[05:10] <carlos> kiko, hi
[05:10] <kiko> carlos, how do I associate a template name with a product?
[05:10] <kiko> http://localhost:8086/products/gnomebaker/+potemplatenames
[05:10] <kiko> doesn't seem to let me
[05:11] <carlos> kiko, you need a POTemplate
[05:11] <kiko> carlos, how do I create one of those?
[05:11] <kiko> (sorry, it's non-trivial to discover)
[05:12] <jordi>  /potemplatenames  to create
[05:12] <carlos> kiko, it's an admin task so I think we don't have it documented
[05:12] <jordi> .../+addpotemplate to a series
[05:12] <kiko> aha
[05:12] <carlos> kiko, https://launchpad.net/products/gnomebaker/+series/main/+addpotemplate
[05:13] <dilys> Merge to rocketfuel@canonical.com/launchpad--devel--0: keyring trust analyzer code, r=lifeless (patch-2586: james.henstridge@canonical.com, robert.collins@canonical.com)
[05:13] <Kinnison> kiko: the sushi here is good, but I *MISS* the chinese
[05:13] <jordi> gotta go
[05:15] <kiko> Kinnison, so do I :-(
[05:15] <kiko> carlos, I get a system error running that locally. 
[05:16] <kiko>     *  Module canonical.launchpad.helpers, line 691, in uploadRosettaFile
[05:16] <kiko>       raise RawFileAttachFailed(str(e))
[05:16] <kiko> RawFileAttachFailed: Server said: 500 Internal server error
[05:16] <kiko> carlos, is this a librarian bug or?
[05:16] <kiko> 	  File "/home/kiko/devel/rocketfuel2/launchpad/lib/canonical/librarian/storage.py", line 128, in _move
[05:16] <kiko> 	    raise DuplicateFileIDError(fileID)
[05:16] <kiko> 	canonical.librarian.storage.DuplicateFileIDError: 35
[05:16] <carlos> kiko, breezy?
[05:16] <Kinnison> kiko: but you live in the town of the chinese
[05:16] <carlos> kiko, that smells like a librarian bug, yes
[05:16] <kiko> ok
[05:17] <carlos> kiko, I'm getting too Internal server errors using breezy
[05:17] <kiko> I'm not using breezy
[05:20] <kiko> carlos, is it the same error? look at your server output
[05:21] <carlos> kiko, last time I ran the test I didn't get it so I don't have it
[05:21] <kiko> ah
[05:23] <carlos> kiko, I don't get it always
[05:23] <kiko> me neither
[05:37] <gneuman> does anyone know how to find a url for this template:template-page.pt?
[05:39] <kiko> gneuman?
[05:39] <kiko> template-page.pt? it's just a draft file, to be copied into others
[05:40] <gneuman> ok
[05:40] <gneuman> =] 
[05:40] <gneuman> better now
[05:40] <kiko> it's not referenced in any zcml that I am aware of
[05:41] <SteveA> the template-* files should just have those actions things removed
[05:41] <kiko> right
[05:41] <gneuman> ok
[05:50] <segfault> Rosetta encountered problems exporting the files you
[05:50] <segfault> requested. The Rosetta team has been notified of this
[05:50] <segfault> problem. Please reply to this email for further assistance.
[05:50] <segfault> Has anyone noticed this?
[05:50] <kiko> yes
[05:51] <segfault> what hapenned?
[05:51] <kiko> segfault, it's normally related to problems in the translation. what file was that? I can help you debug
[05:51] <segfault> Hoary's rpm package
[05:52] <kiko> segfault, what's your surname?
[05:52] <kiko> Santiviago?
[05:52] <segfault> Santiviago
[05:52] <segfault> yes
[05:53] <kiko> okay
[05:53] <kiko> the problem is that the brazilian portuguese translation is encoded using the ascii codec
[05:53] <kiko> it then fails because it encounters an  
[05:53] <kiko> can you change it to be UTF-8 or at least latin-1, segfault?
[05:54] <dilys> Merge to rocketfuel@canonical.com/launchpad--devel--0: [trivial]  Fix https://launchpad.net/products/shipit/+bug/2893 (Error page doesn't preserve the shipping request) (patch-2587: guilherme.salgado@canonical.com)
[05:54] <kiko> thanks salgado-lunch, extra thanks
[05:54] <segfault> kiko: well, i was trying to get the actual po file from Rosetta, i haven't touched it yet. But i can try to get another one and upload.
[05:55] <kiko> carlos, ping?
[05:56] <carlos> kiko, pong
[05:59] <kiko> carlos, can segfault change the encoding of the pofile in any way /bug/ uploading a new file?
[05:59] <kiko> carlos, also, I need you to explain to me what      [trivial]  Detected and notified when a .po file is imported and the
[05:59] <kiko> +msgid_Plural was changed in the .pot one + test
[05:59] <kiko> means :)
[05:59] <kiko> carlos, s/bug/but above btw
[06:00] <carlos> kiko, it's a workaround that I talked with Steve
[06:00] <carlos> kiko, that fixes, for instance, control-center imports
[06:02] <carlos> kiko, there is a corner case when a maintainer changes the msgid_Plural but don't change the msgid, in that case, the .po imports would fail because we cannot represent that in our db, we are only allowed to store one msgid pero translation domain and thus only a msgid_Plural for that msgid
[06:02] <carlos> kiko, the only way to change an enconding is reuploading a new file or manually doing that using psql
[06:03] <kiko> carlos, okay, thanks.
[06:03] <kiko> segfault, you'll need to reupload a file, sorry.
[06:03] <kiko> carlos, will the upload clobber any existing translations?
[06:03] <carlos> kiko, for your next question.. yes, we should add a way to change the encoding using the website
[06:03] <kiko> okay, thanks.
[06:03] <kiko> carlos, should be trivial, right? 
[06:04] <carlos> kiko, if the upload is exactly what we have imported as upstream and it's imported as "published", no we should not break any existing translation as all things are stored as UTF-8 in our database
[06:05] <segfault> kiko: ok, i'll try to do that. But what happens if i upload a PO file with more msgids than the current in Rosetta?
[06:05] <carlos> segfault, btw, please, update the po revision date adding 1 minute or your upload will be ignored, I have a fix for that but it's not yet merged
[06:05] <kiko> segfault, that's a question for carlos 
[06:05] <carlos> segfault, they will be added 
[06:06] <carlos> segfault, you can upload an update without problems, if it's older or with the same po revision date, it will be ignored, that's why is important that you update that field
[06:06] <carlos> kiko, the encoding editor? yes, it should be trivial
[06:06] <segfault> carlos: sure, i'll do that later.
[06:07] <kiko> carlos, just change a flag in the table, right?
[06:07] <carlos> segfault, thank you
 kiko, if the upload is exactly what we have imported as upstream and it's imported as "published", no we should not break any existing translation as all things are stored as UTF-8 in our database
[06:07] <carlos> kiko, no, parse the pofile.header field, update the dictionary and dump the new header back to pofile.header
[06:08] <kiko> carlos, the upload will not be exactly what we have imported, btw -- segfault said he was going to upload a new one
[06:08] <kiko> carlos, ah, right, of course.
[06:09] <carlos> kiko, it can be an update, it's just that you should be careful with the way you upload it, that's all, the easier fix is if we don't upload changes
[06:09] <carlos> but we will never lose information unless we upload the wrong .po file as not published, that would hide translations from Rosetta
[06:17] <mpt> carlos: I'm pulling out the translation form for a single string into its own template, to make pofile-translate.pt less scary. In pofile.zcml, what should be the for= value for that browser:page?
[06:18] <mpt> for="canonical.launchpad.interfaces.IPOMsgID" doesn't work
[06:18] <kiko> mpt, use a macro
[06:19] <mpt> kiko: That's what I'm doing, I think
[06:19] <kiko> mpt, it's for whatever context you are going to pull it out of
[06:19] <carlos> mpt, that's part of the review suggestions page that lets you navigate message by message, right?
[06:19] <mpt> copying from how malone-index.pt uses bugtask-listing-detailed.pt
[06:19] <carlos> mpt, in that case, I think you should use IPOMsgSet
[06:20] <mpt> carlos: No, it's the part of the translation page that lets you translate a single item
[06:20] <kiko> I thought you were just refactoring the template
[06:20] <carlos> mpt, isn't that page the one that shows you *all* suggestions available ?
[06:21] <mpt> carlos: Suggestions is a part of it
[06:21] <mpt> but not all of it
[06:21] <carlos> mpt, yeah
[06:21] <carlos> mpt, IPOMsgSet
[06:22] <mpt> KeyError: 'messageset' 
[06:22] <mpt> hum
[06:24] <kiko> mpt, would you allow me to remove padding: inherit from the <code> tag?
[06:24] <kiko> mpt, it causes issues when using it inside a validation or notification div.
[06:25] <mpt> kiko: sure
[06:25] <mpt> why did it exist?
[06:25] <mpt> oh, code has a dotted border or something hideous around it, doesn't it?
[06:25] <mpt> so get rid of that too
[06:25] <kiko> right
[06:25] <kiko> I think you removed that
[06:25] <kiko> but the original code has
[06:25] <kiko>     padding: 0 0.1em;
[06:26] <kiko> and I think you wanted to undo that
[06:27] <mpt> I don't know why that exists
[06:27] <kiko> hmmm
[06:31] <dilys> Merge to rocketfuel@canonical.com/launchpad--devel--0: [trivial]  XHTML validity fixes on front page (patch-2588: mpt@canonical.com)
[06:33] <bradb> SteveA: When are we planning to upgrade to Zope 3.1?
[06:35] <SteveA> bradb: before ubz
[06:35] <bradb> cool
[06:36] <bradb> SteveA: Did you say that it includes a mechanizer-type form testing framework?
[06:36] <mpt> carlos, do you have time to help met get unstuck? :-)
[06:37] <carlos> mpt, sure
[06:37] <mpt> carlos: In pofile-translate.pt, I have <tr tal:repeat="messageSet view/messageSets" tal:replace="structure messageSet/@@+translate-item" />
[06:37] <mpt> ok so far?
[06:38] <carlos> mpt, yes
[06:39] <mpt> then in pofile.zcml, I have <browser:page for="canonical.launchpad.interfaces.IPOMsgSet" name="+translate-item" template="../templates/translate-item.pt" permission="zope.Public" />
[06:40] <mpt> and translate-item.pt contains the <tr> for a translatable item
[06:41] <mpt> carlos: but then I get "NotFoundError: (<canonical.launchpad.browser.pomsgset.POMsgSetView object at 0xb584f44c>, '+translate-item')"
[06:41] <carlos> mpt, btw, permission should be launchpad.Edit as it's an editable field
[06:42] <carlos> mpt, messageSet is not an IPOMsgSet
[06:42] <mpt> ok
[06:42] <carlos> mpt, I think It's a dictionary
[06:43] <mpt> IDictionary?
[06:44] <mpt> hmm, no, there seems to be no such thing
[06:45] <carlos> mpt, no, I suppose we should rewrite that page to use IPOMsgSet directly
[06:47] <mpt> so there's no way to get this working now?
[06:49] <SteveA> what do you want to do mpt?
[06:50] <SteveA> what's the "IDictionary" issue?
[06:51] <mpt> SteveA: pull out the translation form (the <tr> including text fields, suggestions, etc) from pofile-translate.pt into its own translate-item.pt, in the same way as malone-index.pt uses bug-listing-detailed.pt
[06:51] <mpt> but I don't know how to work out the correct attribute values for the zcml
[06:51] <mpt> (at least, I think that's the problem)
[06:52] <SteveA> so, while you can register such a view on a dict, it isn't a good idea
[06:52] <SteveA> so, carlos is correct, i think
[06:52] <SteveA> the object doesn't carry enough information what what it actually is
[06:55] <mpt> I'll leave that, then, so as not to waste too much of other people's time
[06:55] <SteveA> mpt: file a bug on what carlos said
[06:55] <SteveA> or ask carlos to file it
[06:55] <carlos> mpt, I prefer if you can file it O:-)
[06:56] <mpt> ok
[06:59] <mpt> https://launchpad.net/products/rosetta/+bug/2907
[07:01] <carlos> mpt, thanks!
[07:09] <bradb> SteveA: Did you say that it includes a mechanizer-type form testing framework?
[07:30] <SteveA> bradb: that's one of the things we want
[07:30] <dilys> Merge to rocketfuel@canonical.com/launchpad--devel--0: [trivial]  Improve poparser exceptions and Fix for bug 2901: /malone/bugs/assigned causes Oops. Remove dead code and XXX. Improve the wording on the edit GPG key page, nuke bogus CSS entry (patch-2589: christian.reis@canonical.com)
[07:30] <SteveA> kiko-fud: ping
[07:40] <kiko-fud> SteveA, pong
[07:41] <SteveA> kiko: what do you think about arranging a regular launchpad users meeting?
[07:41] <SteveA> on irc?
[07:41] <kiko> SteveA, wow, that's an interesting idea
[07:42] <SteveA> we don't get a lot of participation in the developers meeting
[07:42] <kiko> no, we don't -- understandably
[07:42] <SteveA> but then again, we don't really make it relevant to our community
[07:42] <kiko> hmmm
[07:45] <bradb> SteveA: What's the full URL at which to access the Zope server inside a doctest? I'm trying to write a test with ClientForm, using http://localhost:9000/... (also tried 8085, 8086 and 8089) and keep getting connection refused.
[07:45] <SteveA> bradb: don't use ClientForm yet
[07:46] <bradb> hrmph. ok.
[07:46] <kiko> SteveA, let's say we did the meeting right after the developers meeting -- 13UTC on thursdays?
[07:46] <kiko> or on wednesdays and then used the feedback to discuss on thursdays?
[07:47] <SteveA> i can't do it right after the developers' meeting
[07:47] <SteveA> i need to have lunch sometime during the day
[07:47] <SteveA> and i've been in meetings 2 hours straight then
[07:51] <SteveA> jamesh: the pyme makefile uses python2.3 for something
[07:55] <niemeyer> SteveA: It's not really a packaging/dependency issue.. just a policy one..
[07:55] <SteveA> ok
[07:55] <SteveA> i've seen various zope/python projects go "batteries included" in the past
[07:55] <SteveA> and it has caused problems for packages
[07:55] <SteveA> and it has caused problems for packagers
[07:58] <niemeyer> SteveA: Indeed.. even projects like Twisted, which had no external dependencies, went like "aggregate aggregate aggregate.. oops.. split split split"
[08:00] <Kinnison> ciao all
[08:00] <niemeyer> Kinnison: See ya
[08:08] <carlos> see you
[08:25] <SteveA> bradb: ping
[08:26] <bradb> SteveA: pong
[08:27] <SteveA> hi
[08:27] <SteveA> https://chinstrap.ubuntu.com/~dsilvers/paste/file2mwoog.html
[08:28] <SteveA> please take a look at that
[08:28] <SteveA> i want to talk through it with you
[08:28] <SteveA> because i want to convert it to the new navigation format
[08:28] <SteveA> i have a few questions about it
[08:28] <bradb> ok, go ahead
[08:29] <SteveA> so, this is registered to traverse an IBugTask
[08:30] <SteveA> i see two sections to the function
[08:30] <SteveA> the first section deals with null bug tasks, and standard views on bugtasks
[08:30] <SteveA> the second section deals with traversing to utilities
[08:30] <SteveA> is that right so far?
[08:30] <bradb> yup
[08:31] <SteveA> is the ordering of the two sections important?
[08:31] <SteveA> could the order be reversed, and it would still work out okat?
[08:32] <bradb> well, it could be reversed, but the code would have to be slightly modified, of course
[08:32] <bradb> e.g. if utility_iface is None: return None, etc. wouldn't make sense if there was more name traversal code further down
[08:32] <SteveA> okay
[08:34] <SteveA> i don't really get the idea of the traversal to these utilities
[08:35] <SteveA> why is there a traversal from a bugtask (the context) to context/tasks/... ?
[08:35] <bradb> that's a bug
[08:35] <bradb> there shouldn't be a tasks traversal; i thought i removed that
[08:35] <SteveA> you mean, in the code?
[08:35] <SteveA> okay
[08:35] <SteveA> don't remove it ;-)
[08:35] <SteveA> you'll conflict, or i will
[08:35] <bradb> right, i won't remove it
[08:36] <SteveA> okay
[08:36] <SteveA> so, the idea is to handle context/watches/NUMBER urls?
[08:36] <bradb> yes
[08:36] <bradb> context/watches/..., more specifically
[08:37] <bradb> "..." can be more than just a number
[08:37] <SteveA> okay
[08:37] <SteveA> and it also needs to handle context/watches/ ?
[08:38] <bradb> or more than /just/ a number, that is. like watches/42/+edit, for example
[08:38] <bradb> SteveA: there's no default views for context/watches/ type URLs (or, if there is, it's only by accident)
[08:39] <SteveA>     if not nextstep:
[08:39] <SteveA>         return utility
[08:39] <SteveA> 
[08:39] <bradb> they usually redirect to the context page, IIRC
[08:39] <SteveA> so, they can become not found, then?
[08:39] <SteveA> or redirections?
[08:39] <bradb> either seems ok to me
[08:40] <SteveA> can you check that a redirection is what you want for attachments, referernces and watches?
[08:41] <bradb> SteveA: attachments/ 404s currently, references and watches redirect.
[08:41] <SteveA> thanks
[08:41] <SteveA> is that okay to keep it that way?
[08:42] <bradb> which one is correct is a matter of opinion, i think
[08:42] <bradb> SteveA: sure
[08:42] <SteveA> ok
[08:42] <SteveA> it will be easy to change once the stuff lands
[08:43] <SteveA> actually, i think i'll try making them all 404s
[08:43] <SteveA> it is simpler, and can be changed later if needed
[08:43] <bradb> sounds good
[08:50] <SteveA> kiko: darn... you edited traversers.py!
[08:50] <SteveA> kiko: pqm conflicted on me.  i've totally changed traversers.py with the Navigation changes.
[08:51] <kiko> SteveA, I fixed a bug there that mpt reported :)
[08:51] <kiko> don't regress it please
[08:51] <kiko> oh, darn
[08:52] <kiko> I didn't add a test for it
[08:52] <bradb> heh
[08:52] <SteveA> oh
[08:53] <SteveA> you just removed to =='assigned' thing
[08:53] <bradb> kiko: are you interested in doing a quick review of my sortwidget patch when baz is ready? it gets rid of those horrifying results headers on the +bugs listings
[08:53] <SteveA> that doesn't need a test
[08:53] <SteveA> and i removed it on my branch already
[08:53] <kiko> SteveA, yeah, well, notfound-traversals..
[08:53] <kiko> bradb, mpt: https://launchpad.net/products/malone/+bug/2075 is fixed, right?
[08:54] <bradb> kiko: no idea
[08:56] <SteveA>         # If the bug is not found, we expect a NotFoundError. If the
[08:56] <SteveA>         # value of name is not a value that can be used to retrieve a
[08:56] <SteveA>         # specific bug, we expect a ValueError.
[08:56] <SteveA> 
[08:56] <SteveA> i don't think having IBugSet.get(name) raise a ValueError here is useful
[08:56] <SteveA> it should be a NotFoundError or subclasses of it always
[08:56] <bradb> kiko: I believe that we've remedied #1 in that report, but #2 is still there.
[08:57] <bradb> maybe that makes it fixed though, hm
[08:58] <bradb> SteveA: I'm not too bothered either way, but NotFoundError might make it harder to detect "weird" bugs, in cases where unexpected values are getting passed.
[08:59] <SteveA> so, make it a specific subclass of NotFoundError
[08:59] <SteveA> like FormatOfNameIncorrect
[08:59] <SteveA> but still allow presentation code to catch NotFoundError
[09:00] <bradb> oh, right
[09:01] <bradb> for chrissake, WHY am i applying 117 revisions?
[09:04] <kiko> bradb, do you use a revision library?
[09:04] <bradb> kiko: yeah, but i have to keep deleting it
[09:05] <bradb> i've only got a couple gigs of free space on that partition
[09:05] <kiko> bradb, use library-relink and library-shrink
[09:06] <bradb> ah didn't know about those
[09:06] <bradb> sounds like what i'm looking for
[09:06] <bradb> thanks
[09:08] <kiko> don't delete your revlib
[09:08] <kiko> it won't make things better
[09:08] <bradb> that's for sure
[09:18] <bradb> where's library-shrink?
[09:18] <kiko> salgado?
[09:19] <salgado> yep
[09:19] <kiko> help bradb 
[09:20] <bradb> salgado: where's library-shrink?
[09:22] <salgado> bradb, the one I have here I got from http://mirrors.sourcecontrol.net/jblack@gnuarch.org--2004/tlacontrib--devo--1.2--base-0
[09:22] <bradb> salgado: is it possible to check it in?
[09:22] <bradb> since i doubt i'm the only one who wants it
[09:23] <salgado> you mean, put it in utilities/?
[09:24] <bradb> yeah
[09:25] <bradb> right beside library-relink
[09:25] <salgado> I think it's okay. let me check if it has no extra dependencies
[09:25] <salgado> the library-relink is already in utilities/?
[09:25] <bradb> yeah
[09:26] <bradb> kiko: are you interested in reviewing my patch that puts a proper sort widget on the +bugs listing pages, replacing that horrid results header?
[09:27] <kiko> sure sure
[09:28] <bradb> kiko: sent!
[09:35] <salgado> kiko, what about mine? ;)
[09:38] <kiko> sure sure
[09:39] <kiko> ajmitch_, can you get somebody to look at bug 2913 for me? 
[09:40] <ajmitch_> yeah, I can find someone who knows
[09:42] <kiko> thanks
[09:42] <ajmitch_> so it just needs the patch uploaded?
[09:42] <ajmitch_> I can handle that
[09:42] <kiko> well
[09:42] <kiko> I'd rather someone with discernment looked at it
[09:43] <kiko> it's a bit of a hack
[09:43] <dilys> Merge to rocketfuel@canonical.com/launchpad--devel--0: [r=jamesh, trivial]  initial Navigation landing, with some traversal refactorings. also some other minor changes, for example, to how tracebacks appear in production. (patch-2590: steve.alexander@canonical.com)
[09:43] <kiko> and there's another patch
[09:44] <ajmitch_> right, is this for apache or apache2?
[09:44] <ajmitch_> never mind, I saw 1.3
[09:44] <kiko> ajmitch_, it may apply to both, I'm not sure
[09:44] <kiko> mine is 1.3
[09:47] <kiko> SteveA, do you understand how the stubmailer works?
[09:47] <kiko> why/how it's activated?
[09:47] <SteveA> kiko: not without a bit of research
[09:48] <kiko> ok
[09:50] <bradb> kiko: It's hooked up in the ZCML that the test harness is configured with, I believe. It's just an object that holds the emails in memory (in a list), instead of, say, connecting to an MTA
[09:51] <bradb> er, unless you mean that one that is used when running the app (as opposed to running the tests). but i imagine it's basically the same idea.
[09:51] <kiko> bradb, how does it know it should be used when running on localhost?
[09:51] <kiko> bradb, right, that's what I mean
[09:51] <kiko> that's the stubmailer
[09:52] <bradb> kiko: the config file connects it like this:
[09:52] <bradb>     <!-- Development configuration to avoid spamming innocent people -->
[09:52] <bradb>     <mail:queuedDelivery name="Mail" permission="zope.SendMail"
[09:52] <bradb>         queuePath="/var/tmp/launchpad_mailqueue" mailer="stub" />
[09:52] <bradb> and:
[09:52] <bradb> (earlier)
[09:52] <bradb>     <mail:stubMailer 
[09:52] <bradb>         name="stub" from_addr="root@localhost" to_addr="root@localhost"
[09:52] <bradb>         />
[09:52] <kiko> hmmm!
[09:53] <bradb> from package-includes/mail-configure-normal.zcml
[09:53] <kiko> package-includes? what's that?
[09:53] <bradb> a directory full of zcml files
[09:53] <bradb> loaded in site.zcml
[09:54] <bradb>     <!-- package-includes/*-configure.zcml contains package configuration
[09:54] <bradb>         shared between a normal launchpad instance and the test suite.
[09:54] <bradb>         package-includes/*-configure-normal.zcml and 
[09:54] <bradb>         package-includes/*-configure-testing.zcml contain configuration
[09:54] <bradb>         that needs to be different between normal and test modes.
[09:54] <bradb>     -->
[09:54] <bradb>     <include files="package-includes/*-configure.zcml" />
[09:54] <bradb>     <include files="package-includes/*-configure-normal.zcml" />
[09:54] <kiko> I see
[09:54] <kiko> and the configs/* override this, bradb?
[09:55] <bradb> not sure, but I *think* those configure a different set of things
[09:56] <bradb> no, in looking at them, they appear to configure the same kinds of things.
[10:04] <SteveA> salgado: my point is this
[10:04] <SteveA> don't set a value of 10000 in the code, and then use a value of 15000 as the cutoff
[10:04] <SteveA> simply set a value of 15000 in the code
[10:05] <salgado> and if there's 150001 requests, I split them into one file with 10000 and other with 5001?
[10:07] <salgado> s/150001/15001
[10:09] <SteveA> and how is that different from setting the cutoff at 15000 ?
[10:09] <SteveA> your scheme just makes the functioning of the software more obscure
[10:15] <salgado> SteveA, I'm not seeing how that would be less obscure by setting the cutoff at 15000 but actually doing the split at 10000. I guess I don't understand what you're suggesting
[10:17] <SteveA> i'm sugesting that you set the cutoff at say 15000
[10:18] <SteveA> and then you do the cutoff at 15000
[10:18] <SteveA> see, the numbers are the same
[10:18] <SteveA> it is easier to understand
[10:18] <salgado> definitely
[10:18] <SteveA> but, you are arguing for setting the cutoff at 10000
[10:18] <SteveA> and doing the cutoff at 15000
[10:18] <SteveA> the numbers are different
[10:18] <SteveA> for no good reason
[10:19] <salgado> I have the feeling Jane thinks having a file with 1 or 10 or 20 requests is a good reason, and that's why I did this
[10:21] <salgado> of course, if she says that's no big deal, I'll remove that code for sure
[10:21] <bradb> SteveA: So, browser:navigation replaces browser:traverse?
[10:21] <SteveA> salgado: your scheme doesn't actually do that
[10:21] <SteveA> bradb: yes
[10:21] <bradb> ok, thanks
[10:21] <SteveA> bradb: i'm handling all the replacement
[10:21] <SteveA> ok
[10:21] <SteveA> browser:traverse and browser:suburl are going away
[10:22] <salgado> SteveA, doesn't do what?
[10:22] <SteveA> salgado: to do what you're saying, you'd need a cutoff and a minimum size for the last file
[10:22] <SteveA> this is the same is widow-orphan control in typesetting
[10:23] <salgado> the minimun size for the last file will always be 5000, and the maximun will be 15000
[10:24] <SteveA> i see.  you weren't saying that earlier.
[10:28] <SteveA> i think where you confused me was here:
[10:28] <SteveA> "The way I did it,
[10:28] <SteveA> if a shipping run would have more than LIMIT*1.5 requests then it's split
[10:28] <SteveA> into multiple shipping runs of at most LIMIT requests."
[10:28] <SteveA> what you mean was "it is split into multiple shipping runs of at least LIMIT / 2 requests"
[10:30] <salgado> yes, what I said is completely wrong. sorry for that
[10:31] <salgado> but at least the code does what I can't explain
[10:31] <bradb> kiko: how does the sortorder widget patch look?
[10:32] <jordi> damn i need carlos so much right now.
[10:32] <jordi> kiko: you may know, https://launchpad.net/products/control-center/+translations
[10:33] <jordi> why is the only template listed the hoary one? shouldn't there be a breezy template?
[10:34] <SteveA> haha -- browser:suburl is no more
[10:35] <bradb> sweet
[10:40] <dilys> Merge to rocketfuel@canonical.com/launchpad--devel--0: [trivial]  use better names for the files generated by the shipit exports. also adds the shrink-library script to utilities/ (patch-2591: guilherme.salgado@canonical.com)
[10:48] <Naunaud> hello ubuntu people :)
[10:55] <SteveA> bradb: (and anyone else who is listening), so i've handled all browser:suburl things.  i'll be doing browser:traverse tomorrow.  and then after that, browser:url will be folded into the navigation classes.
[10:55] <bradb> great.
[10:56] <SteveA> although i might do breadcrumbs next
[10:56] <SteveA> as that will have the greatest immediate benefit
[10:56] <SteveA> we'll see how it comes out
[11:05] <dilys> Merge to rocketfuel@canonical.com/cscvs--devel--1.0: [r=spiv]  samba fixes: directory copying and more robust patching (patch-112: david.allouche@canonical.com)
[11:18] <bradb> later
[11:38] <lifeless> Kinnison: no I'm not
[11:39] <dilys> Merge to rocketfuel@canonical.com/launchpad--devel--0: [trivial]  remove all uses of browser:suburl and nuke the suburl directive code. (patch-2592: steve.alexander@canonical.com)
[12:00] <lifeless> Kinnison: its my f*cking ISP.
[12:00] <lifeless> who do not know the first thing about mail standards