[12:11] <dilys> Merge to devel/launchpad/: [trivial]  Change 'beginning of May' with 'around mid-May' on the instructions for people wanting to order Dapper CDs (r1853: Guilherme Salgado)
[02:10] <lifeless> SteveA: are you awake ?
[02:13] <mpool> he was a moment ago
[02:13] <mpool> just said he was going to bed
[02:45] <lifeless> jamesh: ping
[03:53] <jamesh> lifeless: pong
[03:54] <lifeless> hi
[03:54] <lifeless> 20  jamesh  david/bzrtools/smallfixes
[03:54] <lifeless> still shows as needs-review
[03:56] <jamesh> on it.
[03:56] <lifeless> thank you
[03:57] <lifeless> separately, have you seen https://launchpad.net/products/launchpad-development-infrastructure/+bug/42366
[03:57] <Ubugtu> Malone bug 42366 in launchpad-development-infrastructure "Update pending-reviews script to show time in current state" [Normal,Unconfirmed]  
[03:57] <jamesh> yeah
[03:57] <lifeless> cool
[04:03] <dilys> Merge to devel/launchpad/: [trivial]  Enable XML-RPC on staging (r1854: Stuart Bishop)
[06:34] <fabbione> morning guys
[06:34] <fabbione> are all the LP apps frontends ok???
[06:34] <fabbione> malone seems horribly slow this morning
[06:35] <fabbione> s/seems/i
[06:35] <fabbione> s/seems/is
[06:35] <sfllaw> fabbione: Morning.
[06:35] <fabbione> morning sfllaw 
[06:37] <ajmitch> hi
[06:47] <fabbione> mdz: ping?
[06:49] <fabbione> can somebody at least confirm i am not the only one with this problem?
[06:49] <stub> fabbione: Seems snappy to me, and all the appservers are responding.
[06:50] <fabbione> it seems very very very slow
[06:51] <fabbione> to the point i need to reload pages a few times before i actually get it
[06:51] <spiv> fabbione: are you getting error pages?
[06:51] <fabbione> no
[06:52] <fabbione> i get FF timeouts
[06:52] <spiv> What URLs are you looking at?
[06:52] <fabbione> https://launchpad.net/distros/ubuntu/+source/xserver-xorg-input-evdev/+bugs
[06:52] <fabbione> for example
[06:52] <fabbione> i had to reload it 3 times before getting anything back from LP
[06:52] <fabbione> ff was stalling on Waiting for launchcrack.net
[06:52] <spiv> Perhaps it's a connectivity issue where you are?
[06:52] <fabbione> spiv: mtr looks good
[06:53] <spiv> That page loads in under 3 seconds for me here, and I'm in .au :)
[06:53] <fabbione> .dk
[06:53] <jamesh> surely .dk has a thicker piece of string than .au
[06:54] <spiv> jamesh: I bet they get a whole *rope*!
[06:55] <jamesh> fabbione: thanks for the discover-data update for the XGI graphics cards
[06:56] <fabbione> jamesh: no problem at all...
[06:56] <fabbione> discover1-data is a piece of hutter shit
[06:57] <jamesh> I got a new laptop yesterday, so will see how well the install goes on that
[06:58] <jamesh> it just has intel graphics, which I assume are covered already :)
[06:58] <fabbione> meh
[06:59] <fabbione> jamesh: before you open any bug on X please read the 3 gazillions already there for i810
[06:59] <jamesh> fabbione: sure.  I usually search before filing bugs
[06:59] <fabbione> jamesh: ehehe
[07:11] <mpt> Eventually I'd like to put the "Report a bug" link at the end of the search results page, and nowhere else
[07:11] <mpt> Would that be horrible?
[07:12] <jamesh> mpt: yeah.
[07:16] <fabbione> ARRRRRGHHHHH
[07:17] <fabbione> another tons of conf/logs copy paste in the comments
[07:17] <fabbione> also please make LP refuse! to accept M$ word documents in attachment
[07:19] <mpt> wow
[07:19] <mpt> How was that even possible?
[07:19] <mpt> They were running Word under wine?
[07:20] <jamesh> mpt: do you think it would be worth trimming long comments with a "read rest of comment" link?
[07:20] <jamesh> mpt: you can create word documents with OpenOffice
[07:20] <mpt> oh, true
[07:21] <mpt> hmm
[07:21] <mpt> jamesh, that's maybe a good idea
[07:21] <mpt> trim them at ten lines :-)
[07:21] <jamesh> 10 lines is a bit short
[07:23] <mpt> oh, speaking of ten lines
[07:23] <mpt> Do you have a half-finished activity report open for tdoay, jamesh? :-)
[07:23] <mpt> today, even
[07:42] <spiv> jamesh: pending-reviews seems broken atm.
[07:43] <jamesh> mpt: not right now.  I'll start it.
[07:43] <jamesh> spiv: yes it does.  I'll take a look
[07:43] <spiv> jamesh: thanks
[07:44] <jamesh> mpt: as far as trimming long comments, trimming long comments to 10 lines is probably okay, but trimming all comments longer than 10 lines would be annoying
[07:44] <jamesh> maybe only trim comments longer than 30 or 50 lines?
[07:46] <jamesh> spiv: a bunch of InstallFailed exceptions from bzrlib ..
[07:47] <jamesh> InstallFailed: Could not install revisions:
[07:47] <jamesh> pqm@pqm.ubuntu.com-20060505020254-e26fc45197d86b65
[07:49] <lifeless> jamesh: a fixing
[07:50] <jamesh> ?
[07:50] <lifeless> jamesh: there was data corruption in chinstraps copy of rocketfuel
[07:50] <jamesh> ah.
[07:50] <lifeless> its now detectable, because of recent fixes to bzrlib
[07:50] <lifeless> when the current push to chinstrap finishes, it will all come good immediately
[07:51] <jamesh> okay
[07:51] <lifeless> | [[07:51] <jamesh> spiv: so it should sort itself out on the next run
[07:51] <jamesh> spiv: https://chinstrap.ubuntu.com/~jamesh/pending-reviews.old/ <- last run
[07:52] <spiv> jamesh: Ah, thanks.
[07:52] <spiv> jamesh: Cool, that has the diff I want, perfect :)
[07:56] <jamesh> spiv: btw, the funky "|=" usage in the patch you reviewed was in the program already -- I was just following the pre-existing style
[07:57] <jamesh> spiv: I can switch to a more explicit style if you want though, since it only seems to be used in that one file
[07:57] <spiv> jamesh: Yeah, reviewing tends to uncover lots of old sins :)
[07:58] <spiv> jamesh: Sorry for misattributing it, it's good that you were being consistent.  But yeah, I'd like it if you could fix that.
[08:00] <spiv> If I saw it in C code, it wouldn't bother me, but I don't like it in Python. :)
[08:06] <lifeless> jamesh: add author annotations to the diff output :)
[08:07] <lifeless> jamesh: push complete
[08:07] <jamesh> lifeless: which bzrlib function do I call to do that? :)_
[08:11] <lifeless> annotate.py  
[08:11] <lifeless> annotate the before and after file, then diff the annotated files :)
[08:12] <jamesh> at the moment I'm just doing bzrlib.diff.show_diff_trees
[08:15] <lifeless> ah
[08:15] <lifeless> will need refactoring then
[08:16] <jamesh> will annotate give correct output for a tree with uncommitted pending merges?
[08:18] <lifeless> it should now
[08:18] <lifeless> hmm, maybe
[08:18] <lifeless> there was a bug
[08:19] <jamesh> I suppose I'd need to do "--all" style annotation too
[08:19] <jamesh> to remove spurious diff output when an insertion occurs in the middle of a block of lines from another revision
[08:25] <lifeless> yeah
[08:25] <lifeless> so, the pending reviews time-in-state thing
[08:25] <lifeless> where is that on your TODO list ?
[08:26] <jamesh> lifeless: I'm getting the bzrsync history rewriting branch landed, and then finishing off moving the supermirror branch-pull-list to the authserver (to get rid of InternalHTTPServer)
[08:26] <jamesh> then I'm looking at the times thing
[08:29] <spiv> jamesh: Interesting, what's the benefit of moving the branch pull list to the authserver?  Just simplifying things by removing InternalHTTPServer?
[08:30] <jamesh> spiv: the details are here: https://launchpad.net/bugs/40383
[08:30] <Ubugtu> Malone bug 40383 in launchpad "remove InternalHTTPLayer" [Normal,Confirmed]  
[08:31] <spiv> jamesh: Fair enough.  Thanks.
[08:32] <lifeless> jamesh: ok. Is there any change I can get you to move the times thing up one ? It would be extremely useful
[08:33] <lifeless> re 40383, has stevea seen that ? I ask because he suggested internalhttpserver in the first place
[08:34] <jamesh> lifeless: SteveA actually suggested removing the InternalHTTPLayerstuff completely after moving the pull-list stuff to authserver
[08:36] <jamesh> lifeless: as far as the times thing, do you just want a "time in current state", or something more?
[08:51] <lifeless> jamesh: I just want to add a new field in the output
[08:51] <lifeless> jamesh: showing how many days its been waiting in that state
[08:53] <jamesh> lifeless: okay.  Currently there is no state persisted between runs (other than the bzr repository used when doing the diffs), so was wondering what to store
[08:56] <lifeless> state, date it entered the state IMO
[08:57] <jamesh> yep
[09:08] <pef> hello
[09:08] <pef> does someone has upload logs access ?
[09:09] <carlos> morning
[09:13] <spiv> lifeless: I sent you mail with a request for a commit to our Twisted snapshot, when you have time.
[09:14] <lifeless> spiv: will it break existing tests ?
[09:15] <lifeless> i.e. after I commit it, are we fucked 
[09:15] <spiv> lifeless: Not that I know of!  I'll double-check locally and let you know.
[09:15] <lifeless> thanks
[09:16] <SteveA> hi
[09:17] <spiv> stub: I'd like to chat to sync up on AuthServerCaching, and also about some wiki stuff.  Is now a good time?
[09:20] <carlos> BjornT: hi, around?
[09:21] <BjornT> hi carlos 
[09:21] <carlos> BjornT: about your review
[09:22] <carlos> BjornT: that branch reuses a lot code from the main +translate form, it adds a single message view reusing a lot of code that is already tested in the main form
[09:22] <carlos> that's why I added pagetests and nothing more
[09:22] <carlos> to test the specific part of that view
[09:28] <BjornT> carlos: that's understandable. still, the view class is far from trivial, so it deserves more testing than just page tests. it would been one thing if you added/changed only template code, but if you add code you should test it properly and not rely on pagetests only. if your view class is broken.
[09:28] <carlos> ok
[09:41] <mpt> sfllaw, ping
[10:18] <jordi> hrm
[10:18] <jordi> bubulle from debian wants me to speak about rosetta in Debconf
[10:19] <jordi> everyone seems to assume I'll be there.
[10:19] <SteveA> where is debconf this year?
[10:20] <jordi> Mexico
[10:20] <jordi> In 1.5 weeks
[10:20] <jordi> 1.3 weeks now. :)
[10:20] <jordi> People think I'll take a last minute flight or something.
[10:21] <SteveA> from my point of view, a talk would be nice if you were there anyway
[10:21] <SteveA> but not really sending you there to give the talk
[10:22] <jordi> oh it's not a talk just about rosetta
[10:22] <jordi> it's a i18n bof
[10:23] <jordi> and as rosetta is a new player in the Debian i18n infrastructure, they want me to discuss how to integrate it in a non-disruptive way
[10:23] <jordi> ie, clash of svn vs. rosetta translations for d-i etc
[10:24] <jordi> sigh, I don't know what to do
[10:24] <jordi> even asking my boss here would be interesting... "can I go to Mexico DF... next friday?"
[10:27] <SteveA> i see
[10:28] <SteveA> i think we'd be better seeing if any debian i18n people can come to the UDP meeting in paris
[10:36] <jordi> bubulle lives in the area :)
[10:39] <carlos> jordi: a last minute flight to Mexico doesn't sound like a cheap flight ;-)
[10:43] <stub> flights to LA are regular and there would be no trouble for short hops to mexico - I think it is better than you suspect.
[10:44] <SteveA> i think mark may be at debconf at some point
[10:44] <SteveA> sabdfl: will you be able to meet with people interested in rosetta at debconf?
[10:46] <jordi> carlos: the prices haven't changed much in 3 months :)
[10:55] <jordi> SteveA: if you manage to get ahold of sabdfl on this, tell me
[10:59] <sabdfl> SteveA: sure
[10:59] <sabdfl> though i won't pimp LP for debian usage because it will drive some folks nuts to do so, i'm happy to answer questions
[11:06] <mpt_> SteveA, can I get an rs= on replacing the old hmenu.js with the new one? The diff is the complete file because of the reindenting, but I've tested it works in Firefox, Opera, Konqueror, and Internet Explorer
[11:12] <mdz> fabbione: pong
[11:13] <fabbione> mdz: solved thanks :)
[11:13] <mdz> fabbione: glad to hear it
[11:15] <SteveA> mpt_: i'd prefer the minimal diff, but this is okay if you have confirmed that there is just that small change.
[11:15] <SteveA> mpt_: in that case, rs=me
[11:16] <SteveA> lifeless: hello
[11:17] <SteveA> lifeless: thanks for the email to the list about metaweave format.  can you confirm that to ensure they have the correct bzr version, folks should use the bzr that is in dapper, and ensure they've done an update/upgrade recently.
[11:17] <mpt_> SteveA, no, I haven't, because the diff is the entire file
[11:18] <mpt_> There's a comment specifying which is the extra code, but I can't be 100% sure that there have been no other changes
[11:18] <mpt_> hmm, I suppose I could just make the same changes to the unindented version
[11:18] <SteveA> here's how to tell....
[11:18] <lifeless> SteveA: yes
[11:18] <lifeless> SteveA: confirmed
[11:18] <mpt_> diff -wwithavengeance?
[11:19] <SteveA> lifeless: on the list
[11:19] <SteveA> ?
[11:19] <lifeless> SteveA: reply and ask :)
[11:19] <SteveA> mpt_: yeah something like that.  i expect jamesh would know the exact incarnation
[11:19] <ddaa> hello
[11:20] <ddaa> lifeless, care to remind me what was your concern with publishing the importd logs to the launchpad devels?
[11:20] <lifeless> ddaa: cvs passwords
[11:20] <ddaa> stored where?
[11:20] <lifeless> launchpad vcs details, emitted in logs
[11:20] <ddaa> your mean stuff in the ~/.cvspass
[11:21] <lifeless> no
[11:21] <ddaa> okay, so if I check the contents of ProductSeries and it comes clear, we can publish that?
[11:21] <lifeless> :pserver:user:password@host
[11:21] <lifeless> ddaa: if you change the form to not allow entry of passwords
[11:21] <ddaa> okay, can the same information be stored in ~/.cvspass?
[11:21] <lifeless> it can
[11:21] <SteveA> jamesh: ping
[11:21] <ddaa> lifeless: in any case, if there are passwords there, they are visible to the whole world in the web UI
[11:22] <ddaa> so fixing the form validation is a low priority IMO, the important thing is fixing the extant data
[11:22] <lifeless> ddaa: I thought the VCS Details were owner only on the seris
[11:23] <ddaa> they are listed in extenso on the series page
[11:24] <ddaa> good... I looked at error report generated by jamesh script yesterday, and I basically fell over myself in despair
[11:24] <lifeless> ddaa: look, I don't really care either way : its not a blocker. 
[11:24] <lifeless> but you asked for what the concern was. Thats what it was.
[11:24] <ddaa> even better
[11:24] <ddaa> I had the impression you considered it a blocker
[11:25] <ddaa> I personally could not care less
[11:35] <jamesh> SteveA: pong
[11:37] <SteveA> jamesh: hi.
[11:37] <SteveA> mpt needs to do a diff between two .js files
[11:37] <SteveA> one has had most whitespace removed
[11:37] <SteveA> one is nicely indented
[11:38] <SteveA> would you help him work out a good way to diff them please?
[11:38] <jamesh> okay
[11:38] <SteveA> i'm not sure whether the normal diff -w (or whatever) will be good enough
[11:39] <jamesh> mpt_: if the javascript files are relatively C-like (i.e. not relying on newline based statement termination), I'd try running "indent oldfile.js -o oldfile-indented.js" and the same for newfile.js
[11:39] <jamesh> mpt_: then try doing a diff of the consistently indented js files
[11:40] <jamesh> indent is designed for C code, so it might make a mess
[11:40] <jamesh> but it would be a good place to start
[11:44] <SteveA> jamesh: i was thinking of some sed or awk to kill the whitespace
[11:45] <Kinnison> and then some use of wdiff or similar?
[11:45] <jamesh> SteveA: given that diff is generally line based, I'd think having consistent newlines would be important
[11:45] <lifeless> diff -w -B
[11:46] <lifeless> if its only whitespace, not where { etc are, diff -w -B should be enough
[11:46] <jamesh> SteveA: I'm not suggesting that the results of indent be committed -- just used to compare the logic of the two files
[11:46] <SteveA> jamesh: right, cool
[11:47] <jamesh> lifeless: the file is lib/contrib/dynarch-resources/hmenu.js -- essentially no newlines
[11:48] <jordi> sabdfl: ok
[11:48] <jordi> sabdfl: if you want me to talk to bubulle about this, let me know
[11:50] <jordi> sabdfl: the scope of the bof is general debian i18n and l10n, but he wanted to tackle how rosetta translators can coordinate with the regular debian l10n teams
[11:50] <lifeless> astyle might do better
[11:54] <ddaa> lifeless: do you have any idea what is the piece of code responsible for restricting the permissions of the log directories in botmaster?
[11:54] <ddaa> did some grepping and testing, and apparently it's not what mkdir just do, but I cannot find the responsible chmod
[11:55] <lifeless> twistd probably
[11:55] <lifeless> as it does setuid etc
[11:55] <ddaa> mh
[11:56] <ddaa> you mean it would override the umask??
[11:57] <ddaa> bingo!
[11:57] <ddaa> CRAAAAAACK!
[11:58] <jordi> carlos: ping
[11:58] <jordi> carlos: can you have a look at the two posts I pointed you at yesterday?
[11:58] <carlos> jordi: sure, could you remind me it?
[11:59] <ddaa> I _guess_ I could reset the umask in master.cfg...
[12:03] <jordi> carlos: Subject: Re: bad merge?
[12:04] <jordi> the other was the plural forms thingy
[12:04] <jordi> I'll mail you and stub now
[12:05] <carlos> jordi: ok, thanks
[12:05] <carlos> jordi: the bad merge mail is already solved using irc and bug report
[12:05] <carlos> jordi: but I will answer anyway
[12:05] <jordi> ok
[12:17] <zakame> hi all :D
[12:18] <zakame> is bradb around? =)
[12:37] <SteveA> zakame: not for another hour or two
[12:42] <zakame> SteveA: ah, ok, I'll probably lurk for a while then :)
[01:01] <AlinuxSOS> https://launchpad.net/distros/ubuntu/dapper/i386/ttf-freefont
[01:01] <AlinuxSOS> howto file a bug against this package?
[01:02] <AlinuxSOS> I can't find an available "file a bug" link. 
[01:02] <AlinuxSOS> can someone help me?
[01:03] <zakame> hmm, ubuntu/+source/ttf-freefont mayhaps?
[01:03] <AlinuxSOS> zakame, thank you found! :)
[01:03] <zakame> AlinuxSOS: a guess, really :) no problem :)
[01:04] <mpt_> AlinuxSOS, you've just experienced a variant of bug 3152 and bug 36645
[01:04] <Ubugtu> Malone bug 3152 in malone "Prominently link between product bug listings <-> equivalent package bug listings" [Normal,Confirmed]  http://launchpad.net/bugs/3152
[01:04] <Ubugtu> Malone bug 36645 in malone "Link prominently from distro release bugs pages to distribution bugs pages" [Normal,Unconfirmed]  http://launchpad.net/bugs/36645
[01:05] <AlinuxSOS> ah
[01:05] <zakame> should 36645 be confirmed then? :)
[01:06] <mpt_> Well, it's not *quite* the same bug :-P
[01:07] <zakame> ah, heh
[01:11] <zakame> I suppose that's another consideration worth taking :)
[01:13] <mpt_> "Mark Shuttleworth does not have members, because Mark Shuttleworth is not a team."
[01:13] <mpt_> That's an interesting error message
[01:13] <lifeless> rotfl
[01:13] <lifeless> 404 would be better
[01:13] <ddaa> mpt_: luckily it's a plural!
[01:13] <zakame> LOL
[01:14] <ddaa> "has no member" would be much, much worse!
[01:14] <lifeless> ddaa: "Mark Shuttleworth does not have members, because Mark Shuttleworth are not a team." ?
[01:17] <mpt_> bug 43086
[01:17] <Ubugtu> Malone bug 43086 in launchpad "Distribution members page is silly when distro team is a person ("Mark Shuttleworth does not have members")" [Normal,Unconfirmed]  http://launchpad.net/bugs/43086
[01:19] <mpt_> But apart from that, the results for http://www.google.com/search?q=%22mark%20shuttleworth%22%20site:launchpad.net are absolutely beautiful
[01:19] <mpt_> Look at those page titles
[01:20] <ddaa> the "for" bits are a bit ugly
[01:20] <ddaa> should be "by", "assigned to", "related to", etc...
[01:20] <mpt_> In a couple of cases they should be "involving", yes
[01:26] <jamesh> mpt_: did you see my suggestion for comparing the new and old versions of that javascript?
[01:29] <mpt_> jamesh, yes, thanks
[01:29] <mpt_> gedit is hanging on the unindented version :-/
[01:29] <mpt_> max CPU eatage
[01:32] <sabdfl> ping
[01:32] <sabdfl> my member was itching, is someone talking about me?
[01:33] <mpt_> sabdfl, https://launchpad.net/distros/fink/+members
[01:34] <zakame> lol
[01:35] <sabdfl> mpt_: iiiinteresting
[01:35] <sabdfl> mpt_: those google hits are purrrrfect
[01:35] <mpt_> That page isn't expecting any distro to have a person as its members
[01:36] <zakame> I just love the new qprocd! :D
[01:36] <zakame> Accepted kdevelop3 (source) :)
[01:37] <mpt_> zakame, URL?
[01:37] <zakame> mpt_: nope, mail :) probably hitting dapper-changes soon
[01:37] <zakame> strike one more for the DhIconCacheChanges :)
[01:46] <sfllaw> mpt_: Pong.
[01:49] <mpt_> sfllaw, are you doing QA for Launchpad?
[02:04] <sfllaw> mpt_: Not that I know of.
[02:08] <mpt_> I wonder who kiko-zzz was talking about then ...
[02:10] <sfllaw> I'm the QA person for Distro, as far as I can tell.
[02:10] <sfllaw> I'd love to help with LP as well, but that's sort of secondary as we have 26 days to release.
[02:10] <sfllaw> :P
[02:11] <bradb> mpt_: matsubara, i thought
[02:14] <mpt_> hmm, possibly
[02:14] <cprov> good morning, hackers
[02:18] <matsubara> good morning!
[02:23] <welshbyte> hello
[02:23] <welshbyte> did everyone's karma get halved?
[02:28] <sfllaw> welshbyte: In LP?  It looks like it.
[02:29] <welshbyte> sfllaw: oh good, not just mine then :)
[03:06] <carlos> kiko: hi
[03:06] <kiko> morning carlos 
[03:07] <carlos> kiko: Do you want to talk about PoMsgSetPage now?
[03:08] <kiko> I need an hour first
[03:08] <kiko> but at 11 sure
[03:08] <kiko> 11 my time
[03:09] <carlos> kiko: ok, ping me when you have time, please
[03:32] <zakame> hi all
[03:32] <zakame> bradb: ping
[03:34] <kiko> bradb, carlos: it's true that PQM is taking forever, is it not?
[03:34] <carlos> kiko: it's being slow
[03:34] <carlos> I think lifeless said it's due rewave actions
[03:34] <kiko> ok
[03:35] <carlos> it should be fixed when we start using the new format
[03:35] <kiko> ok
[03:35] <kiko> or so we pray
[03:36] <kiko> BjornT, SteveA: do you have a moment to discuss something matsubara raised yesterday?
[03:37] <SteveA> ok
[03:37] <kiko> so look at matsubara's last analysis
[03:38] <kiko> OOPS-123B242
[03:38] <Ubugtu> https://chinstrap.ubuntu.com/~jamesh/oops.cgi/123B242
[03:38] <kiko> let's look at the traceback together, shall we?
[03:38] <kiko> if you look at it, we're doing a Snapshot()
[03:38] <kiko> now this causes us to cache /all/ attributes the object has 
[03:39] <kiko> including properties
[03:39] <SteveA> and... is this causing them to be loaded unnecessarily?
[03:39] <kiko> well, "unnecessarily" is questionable
[03:39] <kiko> however
[03:40] <kiko> it does mean that anything that does heavy computation as a property will be triggered when running an +editform
[03:40] <SteveA> heavy computation in properties is of course bad news anyway
[03:40] <kiko> yes, of course, PEP-8 says
[03:40] <kiko> but this adds some extra depth to the problem
[03:40] <SteveA> do we need all properties to be snapshotted?
[03:40] <kiko> well, that's the hard part
[03:41] <bradb> zakame: pong
[03:41] <SteveA> could it use names from a particular interface?
[03:41] <kiko> I'm not sure
[03:41] <kiko> it does
[03:41] <kiko> it does use names from an interface
[03:41] <kiko> however that interface includes properties ;)
[03:41] <SteveA> i see
[03:41] <SteveA> we should change those things not to be properties i think
[03:41] <bradb> kiko: "Forever" would be an improvement, actually.
[03:41] <kiko> that's one way out, which I think is possibly the best one
[03:41] <kiko> I just wanted to raise the issue
[03:42] <SteveA> i mean, we could have expensiveproperty as a descriptor type
[03:42] <SteveA> but that's kinda stupid
[03:42] <kiko> sounds a bit silly
[03:42] <kiko> and really
[03:42] <kiko> what do properties buy us?
[03:42] <kiko> less 2 chars typing?
[03:42] <SteveA> well
[03:42] <SteveA> more than that
[03:42] <kiko> oh?
[03:42] <SteveA> because we should use verb-style words for methods
[03:42] <zakame> bradb: hi, I was referred to you by mdz, re: LaunchpadSupportApp :)
[03:43] <SteveA> but even so
[03:43] <kiko> SteveA, okay, so potentially 6 chars. :)
[03:43] <SteveA> i prefer more chars if it makes things clearer
[03:43] <SteveA> and, being clear about the potential expensiveness of the operation
[03:43] <SteveA> is important
[03:43] <kiko> yeah.
[03:43] <kiko> all right.
[03:43] <kiko> can you email reviewers/launchpad to raise this issue and keep an eye out for this?
[03:43] <kiko> I'll do a sweep through database/ code.
[03:44] <bradb> zakame: Oh, hm, yeah, that brings back memories.
[03:44] <SteveA> there's a reviewers' meeting early next week.
[03:44] <SteveA> i'll put it on the agenda there
[03:44] <kiko> thanks SteveA 
[03:44] <zakame> bradb: ooh, how come? =)  Actually I'm interested in doing something similar as a project for the SoC
[03:45] <bradb> zakame: FWIW, the ticket tracking system in LP now is maintained mostly by BjornT.
[03:45] <kiko> SteveA, and email?
[03:46] <SteveA> an email will come out of the reviewers meeting
[03:46] <kiko> k
[03:46] <SteveA> based on the review policy decided at that meeting
[03:47] <bradb> "Our support tracking sysem, Pebcak..." -- that has a nice ring to it :P
[03:48] <zakame> lol
[03:51] <sivang> zakame: what is this app about? :)
[03:52] <zakame> siv	it's at the LP wiki
[03:53] <zakame> bradb: has there been work on user tool yet included in the Ubuntu distro itself, and not just in LP?  I figure that such an app leads to some interesting possibilities...
[03:54] <sivang> zakame: yeah, found it. this is going to be something installed in ubuntu right?
[03:55] <bradb> zakame: not to my knowledge
[03:55] <sivang> bradb: we did had plans to include another item into lp-integration, that would redirect a user's browser to the support tracker, right?
[03:56] <zakame> sivang: yes
[03:56] <sivang> (along with "Translate" and "Report bugs")
[03:58] <bradb> I think the current "Get Help Online..." would be that.
[03:58] <zakame> sivang: actually I'm meaning to do something like a better bugreporting tool that would automatically include the proper information for debugging certain pkgs, as well as help reduce duplicates
[03:59] <sivang> zakame: I see, so do check out https://wiki.ubuntu.com/AutomatedProblemReports
[03:59] <sivang> zakame: I think this is somewhat at that way
[04:01] <zakame> indeed, looks interesting
[04:03] <kiko> salgado, can anyone register a mirror?
[04:04] <salgado> yes
[04:04] <kiko> for any distro?
[04:04] <salgado> yes
[04:05] <kiko> mpt_, ping?
[04:05] <salgado> kiko, did you see nightly.sh's output?
[04:05] <kiko> not yet
[04:05] <kiko> what happened?
[04:05] <salgado> the prober raised some exceptions
[04:07] <kiko> ha ha
[04:07] <salgado> kiko, https://chinstrap.ubuntu.com/~dsilvers/paste/fileaCbVlg.html is the failing code
[04:07] <kiko> what's the traceback?
[04:08] <salgado> and it fails because there's no BinaryPackageFile for a given binarypackagerelease in a publishingrecord
[04:08] <kiko> ha ha ha
[04:08] <kiko> cprooooov
[04:08] <kiko> okay
[04:08] <kiko> here's a point
[04:08] <salgado> or at least not a .deb file
[04:08] <kiko> I think it is actually possible that a binary upload not include a .deb
[04:08] <kiko> can you do some querying to find out?
[04:09] <salgado> only on mawson
[04:10] <kiko> oh, not on staging/production?
[04:10] <salgado> no, I only have access to mawson
[04:11] <kiko> well
[04:11] <kiko> consider the BinaryPackageFileType dbschema
[04:12] <kiko> what options do we have apart from .DEB?
[04:12] <kiko> salgado, my gut feeling is I think you need to skip publishing records for BinaryPackageReleases that don't have .debs associated.
[04:15] <matsubara> kiko: salgado's machine hung while make checking.
[04:15] <kiko> hung hung or just very slow?
[04:16] <kiko> what's the box's name?
[04:16] <matsubara> kiko: to the point of being unusable, he's coming down
[04:16] <matsubara> canario
[04:24] <SteveA> carlos: incoming...
[04:25] <carlos> SteveA: talking about Tim's email?
[04:25] <SteveA> yes
[04:25] <SteveA> no need to drop everything, but we should tell him what the plan is to look into this
[04:26] <carlos> SteveA: I think the cause is that we imported a new .pot file
[04:27] <carlos> so no data was lost, let me talk with doko
[04:27] <SteveA> ok
[04:27] <carlos> he provide me the new files
[04:27] <SteveA> so, you mean upstream changed?
[04:27] <SteveA> or rather, the ubuntu package changed?
[04:27] <SteveA> and that changed the english translations in the pot file
[04:28] <carlos> SteveA: yeah, doko told me that openoffice.org was going to be backported to breezy and I guess he provide me the new .pot file instead the old 1.x template we had in breezy...
[04:28] <SteveA> if that's so, it's going to have an effect on all translation teams for OOO
[04:29] <carlos> SteveA: yes, and Rosetta is not implementing the fuzzy algorithm that gettext has so we don't create new fuzzy strings in this situation, that's bad... but when we implement it, our data would be updated automatically
[04:29] <carlos> SteveA: right, but most of them have full translations from upstream
[04:29] <carlos> esperanto is a special case
[04:29] <carlos> they are not yet part of upstream
[04:29] <SteveA> i see
[04:30] <carlos> like Kurdish
[04:34] <bradb> BjornT: ping
[04:34] <BjornT> hi bradb 
[04:34] <bradb> hey BjornT. I had a very strange exception raised while using z.tb.B. I was wondering if you could have a look.
[04:35] <kiko> hey BjornT 
[04:36] <BjornT> bradb: sure. what was it?
[04:36] <BjornT> hi kiko 
[04:36] <bradb> Somebody remind me, what's the cmd line tool to paste the clipboard contents into the shell?
[04:36] <kiko> shift-insert?
[04:37] <kiko> BjornT, is Mark's request hard to handle?
[04:37] <bradb> kiko: that's not a cmd line tool :)
[04:38] <kiko> bradb, really, I don't know apart from that and middle-click. do you need it non-interactively?!
[04:38] <carlos> SteveA: doko just confirmed this, current templates are for dapper's OpenOffice 2.x
[04:38] <carlos> SteveA: and he said the many things changed
[04:38] <bradb> kiko: Something I can pipe to utilities/paste
[04:38] <carlos> so it's 100% normal that amount of changes
[04:39] <BjornT> kiko: it's trivial to change the formatting.
[04:40] <kiko> BjornT, could you do it right now then and reply to him? :)
[04:40] <kiko> bradb, cat | utilities/paste and then shift-insert into cat?
[04:40] <kiko> ^D afterwards?
[04:41] <bradb> BjornT: https://chinstrap.ubuntu.com/~dsilvers/paste/file1zPwpy.html
[04:41] <BjornT> kiko: i'd rather not do it now :), i was planning to head out soon. but i could do it tomorrow if you want.
[04:42] <bradb> yeah, i guess i could have done that. i swear there was a cmd line tool that just did this though...i even thought it was you who suggested it to me a while back.
[04:42] <kiko> BjornT, keeping good QoS for mark is always good.. ok.
[04:45] <BjornT> bradb: hmm. in the test, do you log in using the login form, or by sending an Authorization header?
[04:46] <bradb> BjornT: auth header
[04:47] <kiko> spiv, ping?
[04:47] <kiko> or sabdfl, ping?
[04:47] <kiko> I want to know what orderBy on a RelatedJoin means
[04:47] <kiko> does it mean ordering by the target or by the intermediary table?
[04:48] <kiko> (or what it should mean!)
[04:48] <spiv> kiko: pong
[04:48] <kiko> ^^^
[04:48] <spiv> kiko: Excellent question, I'll take a look.
[04:48] <kiko> thanks.
[04:48] <BjornT> bradb: put a 'print text' right before 'c = SimpleCookie(text)' (line 399) in zope/publisher/http.py and see what it says.
[04:49] <spiv> kiko: Well, there are *zero* tests of RelatedJoin with orderBy in SQLObject SVN.
[04:50] <kiko> he he he
[04:50] <spiv> (assuming my grepping is accurate)
[04:50] <spiv> So the glib answer is that it's undefined :)
[04:51] <kiko> that does not reassure me one bit, because we are currently using it :)
[04:51] <bradb> BjornT: ''
[04:51] <bradb> (in pdb)
[04:51] <spiv> kiko: So we must have tests for it then ;)
[04:53] <bradb> BjornT: er, u'launchpad=C384MK2LmUEUUF1C8YMuMew5U-4g7ayVwhtR.EUsa3SU50KnLF94rQ' the second time it hits it
[04:53] <spiv> Wow, there is a seriously scare amount of code involved in RelatedJoin's orderBy stuff considering it has no tests I can find.
[04:54] <kiko> correct
[04:54] <kiko> which is why I was asking spiv :)
[04:54] <kiko> anyway, bbias, rebooting
[04:55] <BjornT> bradb: ok. seems like SimpleCookie doesn't like unicode strings...
[04:57] <bradb> BjornT: I don't get it. 40-show-subscribers-from-dupes.txt seems to work fine. What's the diff?
[04:58] <bradb> (i.e. it's another test where I used Browser)
[05:00] <BjornT> bradb: can you paste the two tests so that i can see what's different?
[05:02] <bradb> BjornT: sure, one sec, just trying to see if I can reduce it to a minimal test case
[05:02] <spiv> kiko: It appears to be sorting by the target.
[05:03] <spiv> Or "otherClass" in its terminology.
[05:03] <bradb> BjornT: madness!
[05:03] <kiko> okidok
[05:03] <kiko> thanks duder
[05:03] <bradb> BjornT: If I throw in a:
[05:03] <bradb> >>> browser.open("http://localhost:9000/products/firefox/+bug/4")
[05:03] <bradb> the test works
[05:04] <bradb> (i.e. add that right before browser.addHeader)
[05:05] <bradb> That's the only thing the other test is really doing differently: opening a url /before/ it calls .addHeader("Authorization". ...)
[05:06] <BjornT> bradb: what does the test do after the addHeader call?
[05:12] <bradb> BjornT: https://chinstrap.ubuntu.com/~dsilvers/paste/fileYiZ4j6.html
[05:12] <BjornT> kiko: how does https://chinstrap.ubuntu.com/~dsilvers/paste/fileNV4xRd.html look?
[05:14] <bradb> BjornT: if you add a "browser.open("http://localhost:9000/bugs/1")" above .addHeader, the test works
[05:14] <Keybuk> hi guys
[05:14] <Keybuk> will we have Debian-in-Launchpad and an XML-RPC way of querying and submitting bugs in, oh, four weeks time?
[05:14] <kiko> BjornT, r=kiko
[05:15] <BjornT> thanks
[05:15] <kiko> Keybuk, the former is doable, mainly a matter of pushing for it. the latter I'm not so sure about; I assume email and/or the text output isn't enough for you?
[05:16] <Keybuk> kiko: mom, still needs to know what bugs it's already reported
[05:16] <Keybuk> "the text output" ?
[05:16] <kiko> Keybuk, I suggested using the nickname field for that, have you tried?
[05:16] <kiko> Keybuk, yeah, ?format=text or something. bradb?
[05:18] <bradb>  /bugs/$id/+text
[05:19] <bradb> e.g. https://launchpad.net/bugs/1/+text
[05:19] <Ubugtu> Malone bug 1 in Ubuntu Dapper "Microsoft has a majority market share" [Critical,Confirmed]  
[05:19] <Keybuk> kiko: I thought the "nickname field was going"
[05:19] <Keybuk> bradb: can you get a bug listing in text?
[05:19] <kiko> Keybuk, it didn't, I kept it for you.
[05:19] <bradb> Keybuk: not atm, sorry
[05:31] <salgado> kiko, https://chinstrap.ubuntu.com/~dsilvers/paste/fileKkpPbb.html
[05:31] <salgado> that's the fix for the prober failure on production
[05:41] <kiko> "fix"
[05:44] (spiv/#launchpad) salgado: That loop would be simpler and clearer as https://chinstrap.ubuntu.com/~dsilvers/paste/filebyC90F.html, I think.
[05:44] (spiv/#launchpad) salgado: And it's not a fix if it doesn't have a test ;)
[05:44] (salgado/#launchpad) spiv, that fix is actually wrong
[05:45] <kiko> I think salgado has some crack upstairs
[05:56] <dilys> Merge to devel/launchpad/: [trivial]  Fixes bug 33743 (E-mail addresses box and product page have links to unavailable pages), bug 39934 (Support requests have no heading), bug 42519 (Use title capitalization, or at least be consistent on Launchpad front page), and bug 42788 (Merge account page shows the header twice). Adds text to the error pages indicating that the launchpad-users@ mailing list requires subscription. (r1855: Matthew Paul Thomas)
[05:58] <bradb> BjornT: Is it me, or were widgets fixed to not include a <div> tag when rendering themselves?
[05:59] <BjornT> bradb: afaik widgets have never included a <div> tag when rendering themselves.
[06:03] <bradb> hm, i swore they did before. oh well.
[06:06] <BjornT> bradb: btw, i found out why you got a ValueError in the page test. work around for now, i'm not sure yet what the best fix is.
[06:08] <kiko> carlos, has the kdegantt issue been solved?
[06:08] <kiko> kdgantt sorry
[06:08] <carlos> kiko: yes, but we need to solve it with code
[06:08] <kiko> ok
[06:08] <carlos> kiko: I asked Riddell about kde having two translation domains with the same name
[06:08] <kiko> ah
[06:09] <carlos> kiko: in the mean time, I renamed one of them and disabled it
[06:09] <carlos> as it's broken, not only for Rosetta but for any distribution
[06:09] <carlos> because two packages would provide the same file
[06:11] <kiko> yeah I see
[06:11] <bradb> BjornT: Zope bug?
[06:11] <ddaa> behold my fearsome sql foo https://launchpad.net/products/launchpad/+bug/43118
[06:24] <carlos> kiko: do we need to do things like 'person.id != foo.id' or 'person != foo' would work now?
[06:25] <kiko> hmmmm.
[06:25] <kiko> I think the latter works
[06:25] <kiko> but what are you trying to compare?
[06:25] <carlos> two Person objects
[06:26] <carlos> the code I'm changing atm is not using the .id part
[06:26] <kiko> one sec
[06:26] <kiko> I will test for you
[06:26] <carlos> but I remember that in other parts of code I used it
[06:26] <carlos> kiko: well, the thing is that I think sometimes will be the same an others will be different unless we fixed that
[06:27] <kiko> well
[06:27] <kiko> if you are holding objects across transaction boundaries it will fail
[06:27] <kiko> ah, one sec.
[06:28] <kiko> so yeah
[06:28] <carlos> is not the case here, all objects come from the same transaction
[06:28] <kiko> then they /should/ be the same
[06:28] <BjornT> bradb: actually, i'm not sure anymore... i thought it was a zope bug, but it could very well be a mechanize or ClientForm bug. i can't really find what is setting the Host header to u'localhost'. i'll try asking someone who might know more.
[06:28] <kiko> so look at this:
[06:31] <kiko> https://chinstrap.ubuntu.com/~dsilvers/paste/fileVnNdZe.html
[06:31] <kiko> carlos, that's a case where it breaks
[06:31] <kiko> I think .id is safer :-(
[06:34] <carlos> kiko: well, but that implies that the transaction failed and you keep using the object, right?
[06:34] <kiko> right.
[06:34] <carlos> then I don't need to use .id
[06:34] <carlos> kiko: we should not use objects after an abort
[06:34] <carlos> we need to refetch them
[06:35] <carlos> so I don't think it's an issue
[06:36] <carlos> kiko: thanks for the confirmation
[06:36] <kiko> carlos, I am not so sure it's safe though
[06:36] <kiko> it may depend on whether they all come from the same transaction or not
[06:36] <kiko> I'd do .id myself
[06:36] <carlos> ok, then I will do it too
[06:56] <sivang> I have the 'upbackup' product registered in LP, now I have changed the name, how can I change the name that is shown in the url part of the object path in LP?
[06:56] <sivang> (will I have to create a new product for that..?)
[06:58] <sabdfl> sivang: i can do that for you
[06:58] <sabdfl> want name do you want?
[06:58] <matsubara> sivang: I think, currently only LP admins can change the name of a product
[06:58] <sabdfl> sivang: the branch url will move
[07:16] <sivang> sabdfl: yay, thanks - change it to hubackup :)
[07:16] <sabdfl> h?
[07:16] <sabdfl> gotit
[07:17] <sivang> it's a nice short of home user backup :)
[07:18] <sivang> sabdfl: yay, thanks, I can already access it using the new object name. Do you do that straight in the db or is there an admin interface for that ?
[07:18] <sivang> (trying to know where the magic takes place)
[07:19] <sabdfl> there's an admin interface
[07:19] <sabdfl> i don't have db rw
[07:19] <sivang> :)
[07:19] <sivang> k, noted.
[07:19] <sabdfl> if you really know the registry system well id be happy to make yu a registry admin and you could do that yourself
[07:23] <sivang> sabdfl: I will ping you back once I've learned enough of that...I'd be great to be able to answer to people with those similar requests while others are away or on different timezone.
[07:23] <sabdfl> you need to understand projects, products, product series, and releases
[07:23] <sabdfl> and links to the distro packages
[07:24] <sabdfl> if you grok all of that, ping me again
[07:30] <sivang> sabdfl: I will, thanks!
[07:40] <sivang> hmm, I already get the mirrored url http://bazaar.launchpad.net/~sivan/hubackup/devel-main , but this seems to not exist yet. Is it created together with the scheduled cron that scans for branches over the night?
[07:40] <sivang> (it displayed on the branches list for my product in LP)
[07:40] <sivang> err, rather under the main development branch details.
[07:59] <salgado> kiko-fud, https://chinstrap.ubuntu.com/~dsilvers/paste/filerpF1nC.html
[08:00] <salgado> I still need to check how many tests these sampledata changes are going to break
[08:53] <salgado> kiko, so, r=kiko with the []  => () change
[08:53] <kiko> well
[08:59] <salgado> that should give me around 16h to get the stamp, considering pqm's current queue
[08:59] <salgado> :)
[08:59] <kiko> heh
[09:04] <salgado> come on kiko. approve that so I can free this branch and start working on the changes we just discussed
[09:04] <kiko> I thought you were going to do them together
[09:08] <salgado> I'm not sure it's a good idea. mainly because this is a somewhat urgent fix while the other can wait if I get carried away by shipit
[09:10] <kiko> ok
[09:10] <kiko> you know tests are going to fail anyway...
[09:16] <salgado> kiko, what should be the status of a mirror for which there's no publishing records at all?
[09:16] <kiko> kindof a corner case -- perhaps unknown?
[09:16] <kiko> the easiest thing for you
[09:18] <salgado> that's what I do today. it'll always happens for Backports and the distrorelease that is in development
[09:18] <kiko> yeah, it's okay
[10:26] <sivang> guys, refuling now inside a branch form about 2 months ago should work?
[10:26] <sivang> (given there are bzr changes taking place, or planned)
[10:40] <kiko> I'd reget unless you have something you want from the branch
[10:41] <kiko> salgado, the nice thing about this change
[10:41] <kiko> is that you can probably use that method from your test
[10:42] <kiko> it will avoid you needing to do all the horrible removeSecurityProxy() stuff 
[10:42] <kiko> right?
[10:44] <salgado> only if I define that method in the interface. and I don't think I should do that
[10:45] <salgado> actually, not even if I do that
[10:49] <kiko> salgado, huh?
[10:51] <salgado> I can't use that method at all
[10:52] <kiko> really now
[10:53] <sivang> kiko: can I just scp in order to get a branch from scratch, to overcome rsync overhead ? (since it has nothing to compare against while getting)
[10:53] <carlos> sivang: yes, it would work but it's really really slow
[10:53] <sivang> carlos: ah :) so rsync is better?
[10:54] <carlos> sivang: if you have changes that you want to keep, get a fresh checkout and merge your old branch into the new one, that's really fast
[10:54] <carlos> sivang: aren't you using rsync already?
[10:54] <carlos> sivang: we use a local mirror fetch using rsync
[10:54] <carlos> and then do the merge
[10:55] <carlos> sivang: I think you should take a look to the scripts at https://wiki.launchpad.canonical.com/RocketFuelSetup
[10:55] <carlos> rocketfuel-get and rocketfuel-refresh
[11:00] <sivang> carlos: I already used those scripts, but they all use rsync, I thought maybe with scp it would be faster ? (since rsync does a 'comaprison' phase with existing data). also, I don't have anything I like to keep in the old branch.
[11:01] <carlos> sivang: the scp code needs to fetch 100MB each time to know what to download...
[11:01] <carlos> that's why we use rsync
[11:01] <carlos> sivang: I guess that will change with the new backend file format that is being developed, I think it's called knits 
[11:02] <carlos> sivang: that 100MB is the inventory file
[11:02] <carlos> for launchpad
[11:03] <sivang> carlos: okay, then I will just use rocketfuel-get :)
[11:05] <carlos> :-)
[11:06] <sivang> laters all
[11:30] <carlos> night