[12:11] <lifeless> kiko: whats the benefit of putting a branch up for review before its finished ?
[12:11] <kiko> it's practically finished, and it's a large branch that has been in waiting for months, I'd appreciate somebody giving it a once-over
[12:11] <kiko> (before carlos spends another week on it and then reviewers say it needs to all change)
[12:12] <lifeless> ok, so a once over, not a detailed review at this point ?
[12:12] <kiko> all that is missing is a few fixes to links in pagetests. your call as to how to handle it
[12:12] <kiko> I just think it's better to get it up earlier rather than later
[12:13] <lifeless> I worry that whats reviewed doesn't match what is up at that point, which leads to double handling or unreviewed code
[12:14] <lifeless> anyhow, its saturday morning, so I'm going to zone out. was a bad week
[12:21] <kiko> I see your point, but I don't think that'll happen here
[12:48] <ddaa> good night everybody
[12:48] <ddaa> or whatever time it's in your part of the world
[12:51] <kiko> night
[01:57] <mdke> spiv: around?
[02:51] <dilys> Merge to devel/launchpad/: [r=SteveA]  doc/bazaar: developer documentation for TheBazaar (r1834: David Allouche)
[06:44] <dilys> Merge to devel/launchpad/: [r=lifeless]  branch-scanner.py renaming and ScriptsAndDaemons compliance (r1835: David Allouche)
[02:16] <Znarl> stub, lifeless : gandwana - Launchpad Apps Server [1/2] 
[02:21] <Znarl> stub, lifeless : Restarted Apps servers on gandwana.
[02:53] <dilys> Merge to devel/launchpad/: [rs=bjornt]  Block external email delivery from Launchpad Dogfood application. (r1836: Celso Providelo)
[03:26] <lool> hi people, I've found a bug in launchpad, it doesn't escape some input prior to stuffing it in the URL
[03:26] <lool> I wouldn't go as far as calling that XSS cause it's on account names only :)
[03:26] <lool> the problematic page is https://launchpad.net/people/+requestmerge?field.dupeaccount=lool+launchpad
[03:27] <lool> the dupeaccount should be "lool+launchpad", not "lool launchpad"
[03:27] <kiko> good one
[03:27] <lool> so it should have been dupeaccount=lool%2Blaunchpad
[03:28] <lool> should I file a bug?  what's the place to do that?
[03:28] <kiko> yes, please do
[03:28] <kiko> https://launchpad.net/products/launchpad/+filebug
[03:28] <lool> thanks
[03:28] <kiko> please assign it to salgado
[03:28] <kiko> and severity critical
[03:29] <kiko> or tell me the bug number
[03:29] <lool> ok
[03:29] <lool> should I check security?
[03:30] <lool> (I didn't)
[03:31] <kiko> you could have
[03:31] <kiko> it won't matter so much because we don't yet have a contact for it
[03:31] <kiko> we need to get around to setting up an alias
[03:38] <lool> bye
[03:40] <kiko> matsubara, only 3 oopses, cool :)
[03:42] <matsubara> :)
[03:43] <matsubara> I'm writing the analysis, do you want to help me with the 3rd RequestExpired one?
[03:43] <kiko> yeah
[03:43] <kiko> well which one?
[03:43] <kiko> +translations and +translate time out
[03:43] <kiko> there are bugs for those
[03:43] <matsubara> RequestExpired: SELECT COUNT
[03:43] <kiko> okay
[03:43] <kiko> one hint I should give you
[03:44] <kiko> RequestExpired does not refer to a specific SQL query
[03:44] <kiko> it just means that the page was slow
[03:44] <kiko> RequestStatementTimedOut means a specific query was slow
[03:44] <kiko> so essentially the issues are:
[03:44] <kiko> +translations and +translate time out
[03:45] <kiko> you should look at the oopses and the times the queries took
[03:45] <kiko> because that will make it clear to you
[03:45] <kiko> it wasn't the SELECT COUNT that timed out
[03:45] <kiko> it was just that we issued a long query when the page started
[03:45] <kiko> and then a lot of smaller faster queries
[03:46] <kiko> and then boom we timed out during those faster queries
[03:46] <kiko> is that clearer?
[03:46] <kiko> SRTs are similar
[03:46] <kiko> they are an indication that the page is slow (but not slow enough to crash yet)
[03:46] <matsubara> a lot clearer
[03:46] <kiko> so if you look at the list
[03:46] <kiko> +translate is slow (news at 11 ;)
[03:47] <kiko> distroreleaselang is slow
[03:47] <kiko> seems list that page can benefit from prejoining!
[03:48] <kiko> or hmmm, maybe it's double-dotted prejoining missing
[03:48] <kiko> can you find a bug for that page and assign to me?
[03:48] <kiko> yeah, it is double-dotted prejoining which we need there
[03:49] <kiko> the +milestone/... crash is actually caused by TeamParticipation not being cached (seen the oops?)
[03:50] <kiko> no clue about the @@ SRTs
[03:51] <kiko> and the +cve SRTs are actually prejoin issues it appears
[03:51] <matsubara> hmm I removed the SRT from the report...
[03:51] <kiko> you could have kept it
[03:52] <kiko> and used my analysis above..
[03:53] <kiko> and there is an interesting additiona NotFound bug
[03:53] <kiko>    7 NotFound: Object: <Product at INSTANCE-ID>, name: u'gnomebaker'
[03:53] <kiko>     0% from search bots, 0% referred from local sites                  
[03:53] <kiko>        7 https://launchpad.net/rosetta/products/gnomebaker/gnomebaker
[03:53] <kiko>         OOPS-118A125, OOPS-118B102, OOPS-118B103, OOPS-118C101, OOPS-118D116
[03:53] <Ubugtu> https://chinstrap.ubuntu.com/~jamesh/oops.cgi/118A125
[03:53] <Ubugtu> https://chinstrap.ubuntu.com/~jamesh/oops.cgi/118B102
[03:53] <Ubugtu> https://chinstrap.ubuntu.com/~jamesh/oops.cgi/118B103
[03:53] <Ubugtu> https://chinstrap.ubuntu.com/~jamesh/oops.cgi/118C101
[03:53] <kiko>    6 NotFound: Object: <Product at INSTANCE-ID>, name: u'wordpress-1.5'
[03:53] <Ubugtu> https://chinstrap.ubuntu.com/~jamesh/oops.cgi/118D116
[03:53] <kiko>     0% from search bots, 0% referred from local sites                  
[03:53] <kiko>        6 https://launchpad.net/rosetta/products/wordpress/wordpress-1.5
[03:53] <kiko>         OOPS-118B149, OOPS-118B159, OOPS-118B210, OOPS-118B45, OOPS-118C121
[03:53] <Ubugtu> https://chinstrap.ubuntu.com/~jamesh/oops.cgi/118B149
[03:53] <Ubugtu> https://chinstrap.ubuntu.com/~jamesh/oops.cgi/118B159
[03:53] <Ubugtu> https://chinstrap.ubuntu.com/~jamesh/oops.cgi/118B210
[03:53] <Ubugtu> https://chinstrap.ubuntu.com/~jamesh/oops.cgi/118B45
[03:53] <Ubugtu> https://chinstrap.ubuntu.com/~jamesh/oops.cgi/118C121
[03:53] <kiko> both those seem to indicate that the traversal to products under /rosetta is kinda wonky but we have broken links
[03:53] <matsubara> the wordpress one is a broken link in wordpress
[03:54] <matsubara> I warned them in #wordpress
[03:54] <kiko> oh okay
[03:54] <matsubara> wordpress website i mean
[03:54] <kiko> they shouldn't link to /rosetta
[03:55] <matsubara> they link to launchpad.ubuntu.com still
[03:55] <matsubara> !
[03:55] <kiko> ave maria
[03:56] <matsubara> and I think the gnomebaker one might be the same case or a old bookmark. because it comes from launchpad.ubuntu.com
[03:56] <kiko> ok then
[03:56] <kiko> I thought we had broken links
[03:56] <matsubara> but there's one interesting
[03:56] <matsubara> 6 NotFound: Object: <canonical.launchpad.webapp.publisher.RootObject object at INSTANCE-ID>, name: u'Launchpad_files'
[03:56] <matsubara> >     0% from search bots, 100% referred from local sites
[03:56] <matsubara> >        6 https://launchpad.net/Launchpad_files
[03:57] <matsubara> this one is refered from launchpad.net
[03:57] <matsubara> OOPS-118D207
[03:57] <Ubugtu> https://chinstrap.ubuntu.com/~jamesh/oops.cgi/118D207
[03:57] <kiko> that looks like a virus or worm
[03:58] <kiko> matsubara, will you add the SRTs that I analyzed at least (and file the bugs?)
[03:58] <matsubara> yes.
[03:58] <kiko> thanks.
[03:58] <kiko> matsubara, another thing you need to do daily is take control of unconfirmeds
[03:59] <kiko> matsubara, have you been using /projects/launchpad/+bugs ?
[03:59] <kiko> it's very useful
[03:59] <matsubara> and what about the Retry exception should I point it to bug 41739?
[03:59] <Ubugtu> Malone bug 41739 in launchpad "Increase number of Retry attempts" [Normal,Confirmed]  http://launchpad.net/bugs/41739
[03:59] <kiko> yes
[03:59] <kiko> it should be made major importance
[04:00] <kiko> https://launchpad.net/projects/launchpad/+bugs?field.searchtext=&search=Search+bugs+in+The+Launchpad&orderby=-priority%2C-severity&field.status%3Alist=Unconfirmed&assignee_option=any&field.assignee=&field.owner=&field.omit_dupes.used=&field.omit_dupes=on&field.has_patch.used=
[04:00] <kiko> use that bookmark
[04:00] <kiko> that number needs to be zero
[04:00] <kiko> it is currently 283
[04:00] <kiko> not good!
[04:01] <matsubara> well, I use the scrape.py page
[04:02] <kiko> well /projects/launchpad/+bugs is probably as good as
[04:07] <kiko> matsubara, are Exceptions 1 and 3 known issues?
[04:07] <kiko> the latter one is a bug in zope3
[04:08] <matsubara> yes, bug 41138 and bug 31382
[04:08] <Ubugtu> Malone bug 41138 in malone "+viewstatus crashes when opening a remote bug" [Normal,Confirmed]  http://launchpad.net/bugs/41138
[04:08] <Ubugtu> Malone bug 31382 in launchpad "OOPS: unicode object is not callable" [Normal,Confirmed]  http://launchpad.net/bugs/31382
[04:08] <kiko> I see
[04:08] <kiko> is 41138 not trivial?
[04:08] <matsubara> don't know. I reported it but didn't investigate too deeply
[04:08] <kiko> I think it is
[04:20] <kiko> matsubara, I have a question about bug 31382.
[04:20] <Ubugtu> Malone bug 31382 in launchpad "OOPS: unicode object is not callable" [Normal,Confirmed]  http://launchpad.net/bugs/31382
[04:20] <kiko> matsubara, I am unable to get an oops here. does the oops happen /after/ the 404 is generated?
[04:22] <matsubara> hmm maybe it was fixed in zope 3.2
[04:22] <kiko> matsubara, well, we got the oops YESTERDAY :)
[04:23] <kiko> aha 
[04:23] <kiko> it is a different codepath
[04:24] <kiko> the bug seems to be in zope.app.traversing.namespace
[04:24] <kiko> matsubara, why don't you update the bug description with the new oops and assign to bjornt?
[04:24] <kiko> I'll add a test for the old problems
[04:25] <kiko> oh how odd
[04:25] <kiko> shipit/@@ is a blank page
[04:25] <kiko> as you reported yesterday
[04:25] <kiko> weird!
[05:07] <matsubara> kiko: How do you know when it's a prejoining issue? When you see lots of repeated SQL statements? Is that what the prejoin is used for? Caching the SQL statements?
[05:07] <kiko> prejoining is an interesting optimization to sqlobject
[05:07] <kiko> what it does is pre-fetches objects
[05:08] <kiko> so that when you do foo.bar.baz it doesn't need to look up bar -- it has already looked it up when looking up foo
[05:08] <kiko> this happens very often in pages which display listings
[05:08] <kiko> for instance a CVE listing
[05:09] <kiko> matsubara, if you look at https://launchpad.net/distros/ubuntu/+cve
[05:09] <kiko> matsubara, and the oops
[05:09] <kiko> what is each row item in terms of objects in SQL?
[05:09] <kiko> and for each row field, do we need to issue additiona SQL queries?
[05:10] <matsubara> and CVE, an Bug and a SP
[05:10] <matsubara> s/and/an/
[05:10] <kiko> okay
[05:10] <kiko> bug what does each row represent?
[05:10] <kiko> s/bug/but
[05:13] <matsubara> ok, we have a BugCVE object. and it represents the relation between the a bug and a cve, right?
[05:14] <kiko> right
[05:14] <kiko> so have you concluded that each row in the table is actually a bugcve?
[05:14] <kiko> look at the template to be sure
[05:18] <matsubara> hmm actually it's a bugtask related to a cve
[05:20] <kiko> yeah.
[05:20] <kiko> so look at the tempalte
[05:20] <kiko> template
[05:20] <kiko> it does:
[05:20] <kiko> bugtask/bug/cves
[05:20] <kiko> bugtask/sourcepackagename
[05:20] <kiko> and then:
[05:20] <kiko> cve/displayname
[05:21] <kiko> packagename/name
[05:21] <kiko> this means that for each row:
[05:21] <kiko> - we will issue one query for bug
[05:21] <kiko> - we will issue one query for cves associated to that bug
[05:21] <kiko>   - we will issue N queries for each cve
[05:22] <kiko> - we will issue one query for the bugtask's sourcepackagename
[05:22] <kiko> look through the oops report and identify these queries
[05:22] <kiko> what we can do when prejoining is: 
[05:23] <kiko> a) fetch bug and sourcepackagename together with the bugtask
[05:23] <kiko> b) fetch all CVEs when fetching cves associated to that bug
[05:23] <kiko> that would leave us with only one query per row
[05:23] <kiko> still not ideal
[05:23] <kiko> but much better
[05:25] <kiko> SteveA?
[05:26] <matsubara> hmm thanks for the explanation
[05:55] <kiko> dude pqm is as slow as a turd
[05:55] <kiko> what the hell is up
[06:38] <dilys> Merge to devel/launchpad/: Fix https://launchpad.net/products/blueprint/+bug/41032 (+specs for a team links to +addspecs which doesn't exist) r=kiko (r1837: Diogo Matsubara)
[06:39] <kiko> yeay
[06:40] <matsubara> kiko: do you want two separate bugs for $distro/+cve and $distrorelease/+cve ?
[06:41] <kiko> matsubara, I think so
[06:41] <kiko> but let's have lunch
[06:41] <matsubara> 2 min turkish
[09:09] <ozamosi> Is it possible to use the launchpad logins on another site?
[09:10] <cbx33> the ubuntu wiki
[09:10] <cbx33> :p
[09:10] <ozamosi> Well, I'm thinking of the ubuntu women wiki :/
[09:14] <ozamosi> I heard that "the launchpad devs wont provide that for php", which seems to imply that it is possible to do for other (python based) sites too.
[09:14] <welshbyte> the mplayer bugzilla doesn't appear to be in the list of bugtrackers i can add a watch to.. can it be added?
[09:37] <mdke> ozamosi: yes, it's possible to provide authentication on other sites
[09:37] <mdke> for example, the Ubuntu wiki, shipit, etc
[09:37] <ozamosi> mdke: If I want that, what do I do?
[09:37] <mdke> welshbyte: if you file a bug, I'm sure someone will deal with it
[09:38] <welshbyte> mdke: cheers
[09:38] <mdke> ozamosi: I'll dig out the documentation. however, can we chat about why you don't use the main wiki? perhaps in a more relevant channel, though
[09:41] <mdke> https://wiki.launchpad.canonical.com/AuthServerAPI <-- the page says to contact the LP team, so emailing the list would be a good idea if you decide you'd like to use it
[09:42] <ozamosi> mdke: ok, thanks!
[09:42] <mdke> ozamosi: so is that a no to my question?
[09:44] <ozamosi> mdke: Not a no, more of a "I'm looking for reasons, since it was decided a while ago, and my job is just to implement stuff"
[09:45] <mdke> ozamosi: did the people taking the decision consult with the wider community? all other (major) Ubuntu teams use the main wiki for their coordination. It's a vital part of getting a team well integrated and recognised by the main community
[09:46] <ozamosi> mdke: no idea...
[09:46] <mdke> ozamosi: who can I contact about it?
[09:47] <mdke> i notice there is already a pretty nice wiki page on the main wiki
[09:47] <ozamosi> svaksha@gmail.com is the person who I'm talking to. She's away now, I think.
[09:48] <mdke> right. Thanks
[10:25] <dilys> Merge to devel/launchpad/: [r=stevea,trivial]  make it possible to use zope.testbrowser in pagetests. a 'browser' object with handleErrors set to False is automatically set up in the pagetests to give more useful error messages. (r1838: Bjorn Tillenius)
[10:49] <oohlaf> Does launchpad scan registered branches which are in knit format and part of a shared repository?
[10:49] <oohlaf> I haven't seen any updates in my branches section lately
[10:49] <oohlaf> Is it because of knits?
[11:04] <SteveA> there's some knit support work going on right now
[11:04] <SteveA> so, yes, that would well be it
[11:05] <oohlaf> ok, just wondering
[11:06] <SteveA> if you see spiv around, ask him about it
[11:06] <SteveA> he's doing the work on it right now
[11:07] <oohlaf> nah, that's probably the cause, same happened when I changed my branches over to shared repositories
[11:07] <oohlaf> it will work eventually
[11:07] <oohlaf> :)
[11:08] <sivang> oohlaf: what does knits enable bzr to do?
[11:09] <oohlaf> smaller repository size, faster on remote branches
[11:10] <SteveA> it's an improved format over weaves, basically
[11:10] <SteveA> a lot of work and thinking went into knits
[11:10] <SteveA> and martin pool says he thinks that while it may be not the final standard format,
[11:10] <SteveA> it will be a good format for quite some time
[11:14] <sivang> SteveA: I see, so waves/knits/etc are the names given to the underlying bzr storage format?
[11:15] <sivang> *weaves
[11:15] <oohlaf> jup, type bzr info and you will see the formats used
[11:18] <sivang> hmm, bzr is real nice.
[11:19] <sivang> oohlaf: indeed I could see all nice stas about my branch
[11:45] <lifeless> moining
[11:45] <lifeless> SteveA: its almost certainly not the final format.
[11:46] <oohlaf> hi, you finally got some sleep ;)
[11:47] <lifeless> SteveA: http://www.robertcollins.net/talks/2006slug-vcsstorage.odp
[11:47] <SteveA> i want my repositories on the original vinyl
[11:47] <SteveA> it has a warmer sound
[11:48] <lifeless> kiko-afk: its reweaving. These branches have conflicting history metadata.