[01:53] mwhudson: if I ask nicely, could you land a branch for me ? [01:54] https://code.edge.launchpad.net/~lifeless/launchpad/merge [01:54] thumper: got a minute to talk package reviews? I want to flow a few high level things past you [01:55] lifeless: not right this instance, I've asked aaron to talk to james at uds [01:55] lifeless: perhaps in an hour or so [01:55] lifeless: ok [01:56] thumper: the bzr team is working on this problem/area too - we should talk together [01:56] james_w: ^ [01:56] mwthanks! [01:56] mwhudson: thanks! [01:56] thumper: ok, will poll you ~2:30 [01:56] lifeless: want to set a commit message? [01:56] doing so [01:58] done [02:01] lifeless: kicked off [02:01] thanks! [02:05] beuno: I bet your not really around are you? [02:05] I wish leonardr was around [02:16] thumper, I am. My brain, OTOH.... [02:50] bryceh: you're right, it doesn't turn up on UI [02:50] I assumed it came up because the srting was right for the UI [03:02] hmm... [03:02] is mailing list stuff foundations? [03:02] I guess it is [03:02] It's Registry. [03:04] wgrant: is it? [03:04] hmm... ok [03:04] thumper: Yes. Think where barry was, and think where teams are. [03:04] (and also look where all the bugs are) [03:04] * thumper moves the question again... [03:54] thumper: poll() [03:54] lifeless: here, but collecting girls from school in about 3 minutes [03:54] heh [03:54] after that ? [03:55] sure === Ursinha_ is now known as Ursinha [04:33] lifeless: here now [04:33] lifeless: if you don't mind listening to me munching [04:33] sure === Guest31208 is now known as AdamDV [04:54] Hi all [04:55] I'm installing launchpad on Ubuntu 10.04 and I'm facing a problem during the "make schema" [04:56] jonathan__: What's the issue? [04:56] jonathan__: pastebin the log you gt? [04:57] yes I am doing this [04:59] you can have the result of the "make schema" command here :http://paste.ubuntu.com/429292/ [05:01] jonathan__: sudo apt-get upgrade [05:01] Then make schema again. [05:01] ok let me try [05:01] You need a new version of python-tickcount from the PPA that rocketfuel-setup activated. [05:03] thumper: https://bugs.edge.launchpad.net/bzr/+bug/576778 [05:03] Bug #576778: easy multi-user server setup [05:04] +lots [05:20] wgrant: ok thanks it seems to work [05:22] jonathan__: Great, [06:17] spm: Is the current mailman problem known? It seems to not be (de)activitating lists or subscriptions. [06:20] wgrant: yes, and bug is filed [06:20] Ah, good. [06:33] wgrant: So, it looks like the Sources.gz on iron is out of date [06:34] wgrant: The pool has 1.0.4-3 in it, but the Sources.gz for sid only mentions 1.0.4-1 [06:34] (And hence, gina only tries to import 1.0.4-1) [06:36] and we're hoping to get some info out of debmirror in ~ 1.5 hours as to whats up with that [07:12] StevenK: Heh, that's why I wanted the extra debugging info. [07:12] wgrant: Yeah, I got it before lunch [07:13] spm: ... 1.5 hours? [07:13] Now less than 1 [07:13] What's taking so long? [07:14] We're waiting for an automatic run [07:14] scheduled outa cron. this is not so urgent a task it needs a manual force re-run on what may be a rather... aggressive task. unfreidnly to the remote system. [07:14] Ah, right. [07:15] If it's debmirror, it's just downloading a few megabytes each run, but OK. [07:15] oh? is that all? is that a fairly fast process? [07:15] Yes. It's not like rsync. [07:15] spm: debmirror reads the package metadata [07:15] bugger it. runing now. [07:15] It just grabs the indices, works out what it doesn't have, and grabs it. [07:15] spm: contrast with rsync :P [07:16] Haha [07:16] damn. very fast. [07:16] wooo. we have logs. [07:17] Hmmm. I can't see what's wrong tho. does this mean anything to you guys? [07:17] [0%] Getting: dists/lenny/Release... dists/lenny/Release failed 500 Can't connect to ftp.uk.debian.org:80 (connect: Connection refused) [07:17] Who broke the firewall *again*? [07:17] It's broken a few times in the last couple of years. [07:17] ahh. point. that may be pebkac.... [07:18] trying agin with proxy settings... [07:18] Hehe [07:18] nope. same. [07:18] Using the special debmirror proxy options? [07:19] It has its own, which suggests that it might not respect $http_proxy. [07:19] * StevenK checks that [07:19] bleh, lets give that a whirl [07:20] debmirror, how I hate thee [07:20] although; wget seems to be having issues as well; so the problem is looking firewallish... [07:22] And debmirror does respect http_proxy [07:22] StevenK: Did you try it, or venture deep into the lovely, lovely Perl? [07:22] The latter [07:22] I call access shenanigans. will create an RT to look into; no GSA's around atm. [07:23] And debmirror's Perl makes me vomit [07:23] StevenK: As I said: lovely, lovely Perl. [07:23] oh ffs. [07:23] (Coming from someone who used to code Perl for a job) [07:23] spm? [07:23] Heh, yes. [07:24] I exported the proxy settings, and it's working. [07:24] Excuse me while I cackle [07:24] Hahah. [07:24] Now, it'll probably get a hash or key mismatch or something. [07:24] I was quite sure you don't need to do that - but the way this script is being called... i suspect breaks that nicely... [07:25] Although the latter is less likely if the stuff's already in the pool. [07:25] http://paste.ubuntu.com/429339/ [07:26] Er. [07:26] Er what. [07:26] What. [07:26] What. [07:26] let's see what we have here. [07:29] Huh. [07:29] Sources is broken. [07:29] debmirror is perfectly correct. [07:29] sid's main Sources.gz refers to this file: [07:29] ebe76d2199e758df9aa7e3eaaf499190 9250730 kde-l1 [07:30] In which stanza? [07:30] kde-l10n 4:4.4.3-1 [07:30] In fact, the whole Files section is massively truncated compared to the Checksums-*. [07:31] I call broken a-f cache. [07:31] Yay, we proved Debian is broken [07:32] I'm not sure that's something to celebrate over :) [07:32] Are you sure? It's a problem with debmirror, and not gina [07:32] Not with debmirror. [07:33] Somewhere in dak! Even better. [07:33] Heh [07:33] Somewhere where *I* don't have to fix it [07:33] I see no ftp.d.o bugs, so I wonder if they know about it. [07:33] on the grounds I like my job; I refuse to believe dak could ever have problems [07:34] Haha [07:34] Heh. [07:34] spm: So, dak is just like TeX? It doesn't have bugs, it has "porting issues" ? [07:35] it has no issues. no further discussions will be entered into [07:35] right. added those proxy settings to the wrapper script and exported same. if THAT doesn't keep it working; I'll break down and cry, or maybe sniffle. [07:36] FSVO "working". [07:36] spm: And kept the logging? [07:36] yarp [07:36] Okay, sweet [07:37] So... someone should go and attack an ftpmaster, I guess. [07:37] StevenK: that we weren't seeing ANY emails is bad. At least with logging we get something. [07:37] I'll add some date stamps in there as well; make it useful. [08:09] Hi all (again) [08:10] Hi jonathan__. Did you get it working? [08:10] to run launchpad the correct command is make run, Am i right ? [08:10] That's right. [08:10] yes it worked [08:10] I restarted my virtual machine [08:10] and now I have a problem when I run 'make run' [08:10] You may need to start Apache manually, depending on your setup. [08:11] What's the problem? [08:12] wgrant: no it's not a problem of apache === almaisan-away is now known as al-maisan [08:13] the script tries to kill the process librarian and it can't find it [08:13] http://paste.ubuntu.com/429355/ [08:13] Ah. Perhaps remove the pid file manually. [08:14] You didn't stop Launchpad cleanly before you restarted it? [08:14] Unhappily you should have right ... [08:20] so how to restore/repair Launchpad if it was not stopped cleanly ? [08:21] Just remove the PID file. [08:21] It should do it itself, but apparently doesn't at the moment. [08:21] There's nothing actually broken. [08:22] the pid of launchpad ? [08:26] Yes. [08:27] noodles775: Morning. Can you please reEC2 that branch? It appears to have vanished. [08:28] Hi wgrant , sure. [08:28] wgrant: Yes, right, I deleted the development-launchpad and development-librarian pids. It's ok now [08:28] jonathan__: Great. [08:28] noodles775: Thanks. [09:01] Hello! === lifeless_ is now known as lifeless [09:08] wgrant: I'm having issues with ec2 land, so haven't sent it off again yet. I'll let you know if I do. [09:08] noodles775: Eep. OK. Thanks for trying. === mrevell is now known as mrevell-lunch === al-maisan is now known as almaisan-away === mrevell-lunch is now known as mrevell === lifeless_ is now known as lifeless [15:20] Shouldn't the topic say week # of 10.05? [15:20] StevenK, hello, is this the bug i am experiencing here? OOPS-1588ED588 - https://bugs.launchpad.net/soyuz/+bug/575426 [15:20] Bug #575426: SHA1-based copy checking breaks when there are expired sources in the target [15:58] cody-somerville, that's a grey area === mars changed the topic of #launchpad-dev to: Launchpad Development Channel | Week μ of 10.05 | PQM is open | https://dev.launchpad.net/ | Get the code: https://dev.launchpad.net/Getting | On-call review in irc://irc.freenode.net/#launchpad-reviews | Use http://paste.ubuntu.com/ for pastes === salgado is now known as salgado-lunch [17:09] bac or matsubara, ping, what command did you use when you experienced the ec2 suite hangs? [17:09] mars, ec2 land [17:11] matsubara, same here (I think). Thanks. [17:13] mars, np, when I used ec2 test it didn't fail for me but salgado reported that he had experienced the hang with ec2 test too === matsubara is now known as matsubara-lunch === almaisan-away is now known as al-maisan === beuno is now known as beuno-lunch === salgado-lunch is now known as salgado [18:07] https://edge.launchpad.net/+search?field.text=error+setting+MTRR ← does that page reference itself in the first hit? === gary_poster is now known as gary-lunch [18:11] jpds, it most certainly does :) [18:11] mars: Impressive. [18:11] jpds, we should not be indexing the search results page. That is a bug. === beuno-lunch is now known as beuno === matsubara-lunch is now known as matsubara === gary-lunch is now known as gary_poster [19:39] mars: ec2 test is what hung on me yesterday. i've seen it with land and test [19:39] bac, ok, thanks [19:57] abentley, how do source package branches get updated? does people commit to them or are they imported branches? if the former, how do we control who can commit to a given source package's branches? [19:58] salgado, it doesn't matter. All kinds of branches can be sourcepackage branches. [19:59] The owner can commit to a sourcepackage branch, (if it's not an import/mirrored branch) [20:00] salgado, for example, you could commit to ~salgado/ubuntu/lucid/bzr/my-change [20:01] abentley, my question was not clear, sorry. I was actually thinking about the "official" source package branch (e.g. lp:ubuntu/bzr?) but I guess that's just an alias to another branch, similar to the development focus of a project branch? [20:02] salgado, that's right. [20:02] I see. it makes sense now [20:02] thanks abentley [20:02] salgado, np === al-maisan is now known as almaisan-away [20:34] edwingrubbs, is there some way to use the storm equivalent of prejoins to reduce the number of database queries necessary to get all of an object's fields? [20:37] leonardr, yes, let me get you an example [20:37] EdwinGrubbs: i'm in a tricky situation because i want to do this from lazr.restful, with no special knowledge of the object's schema [20:39] just as a random example, it would be nice if bug.can_expire didn't make its own sql query [20:41] leonardr, hmmm, that will takes some introspection on the storm attribute definitions [20:42] edwingrubbs: let me state the actual problem rather than possible solutions [20:42] i benchmarked the creation of a lazr.restful json representation [20:43] it took 6 database queries to build that representation [20:44] two bug.bugtasks, one bug.can_expire, one bug.permits_expiration, two more i can't track [20:45] leonardr, it looks like the easiest thing would be to copy some of the logic that the fake sqlobject wrapper around storm uses to handle prejoins. It looks like prejoins handle complicated situations like loading bug.assignee.team_owner, so you could probably simplify it quite a bit. [20:48] edwingrubbs, could it be simplified to the point where lazr.restful could simply adapt an object to an IPreloadedObject (if such an adapter was available) and proceed in the knowledge that the right thing had been done? [20:51] leonardr, actually, it looks like you could take a storm object and just call SQLObjectResultSet(storm_object, prejoins=attr_list)._prepare_result_set() where you just populate the attr_list with the names of attributes that are instances of SimpleProperty. So, using an adapter should work. _prepare_result_set() will call find() and return a normal storm resultset to you. [20:52] edwin: spell it out for me. i have a bug, i make a SQLObjectResultSet with prejoins= all the fields that are published in the web service [20:52] and then i get what--the same bug? a new bug? a list of objects of some kind? [20:57] leonardr, if you are expecting a single item, you would call resultset.one(), and that will raise an exception if you get more than one item. If you are expecting a list, you can just iterate over resultset. [20:59] the problem is i don't know what to expect [21:00] i have an instance of Bug [21:00] i know which fields i'm going to access [21:00] and i want some other code that knows something about how Bug works to do all the queries at once, behind the scenes [21:00] looking at how properties like can_expire are implemented, this seems like a pipe dream [21:00] leonardr, let me try to mock something up for you. [21:01] ok [21:01] for the sake of argument, i have a bug and i want to access can_expire and permits_expiration [21:06] leonardr, I hadn't looked at permits_expiration and can_expire. Those are much more complicated than what prejoins let you do. can_expire actually calls permits_expiration first to try to avoid the really expensive query. [21:07] ok, how about is_complete? [21:14] leonardr, all three of those query all the bugtasks for a bug. storm caching uses the primary key to look up objects, so accessing BugTask.bug would use the cache, but Bug.bugtasks depends on the foreign key on the BugTask table. The only way to load all the necessary data in one query is to make an adhoc query with lots of subqueries. [21:15] edwin: i think this is getting complex enough that there's clearly not an easy win here (though there does seem to be a general performance problem if we're getting the list of bug tasks for a bug twice in a single user request) [21:19] leonardr, well, I think storm with cache the Bug.bugtasks attribute that is used both in permits_expiration and is_complete, but can_expire does a much more complicated custom query without actually using the Bug.bugtasks attribute. [21:21] edwingrubbs: i'm eod. gary, take a look at my conversation with edwin. let's talk briefly on monday about how far this is worth pursuing [21:22] leonardr: ack will do thanks. have a good weekend === salgado is now known as salgado-afk [22:17] maxb, ping