[01:11] <dilys> Merge to devel/launchpad/: Make the mirror prober honour the http_proxy env var, change it to issue batches of 50 requests instead of all at once and some other small things. r=spiv (r1829: Guilherme Salgado)
[01:13] <nictuku> can I ask rosetta questions here?
[01:15] <mpt> nictuku, yes, though you'll be more likely to get an answer when either carlos or jordi are awake
[04:33] <mpt__> Goooooooooooooood afternoon Launchpadders!
[04:36] <jamesh> mpool: ping?
[04:54] <dilys> Merge to devel/launchpad/: [trivial]  Fix removal of unvalidated team emails (https://chinstrap.ubuntu.com/~jamesh/oops.cgi/2006-04-26/D293) (r1830: Guilherme Salgado)
[05:12] <mpool> jamesh: ping?
[05:13] <jamesh> mpool: I was working on an improvement for the launchpad bzr branch syncing code, and had a few questions about the bzrlib APIs
[05:14] <jamesh> the first one is whether bzrlib guarantees that the items in the parent_ids list for a revision will be unique
[05:14] <lifeless> jamesh: revision.parent_ids you mean ?
[05:14] <jamesh> lifeless: yeah
[05:15] <lifeless> jamesh: are you asking if len(parent_ids) always == len(set(parent_ids))
[05:15] <jamesh> lifeless: yeah
[05:15] <lifeless> it *should* but it probably may not. If not I think we should de-dup in bzrlib
[05:16] <lifeless> the only semantic in the ordering of that list is the left-most entry
[05:16] <lifeless> the rest is treated as a set
[05:16] <jamesh> lifeless: ddaa had some code to remove duplicate entries dating back to October last year (I think).  I was wondering whether it was necessary still
[05:17] <lifeless> I dont recall any bugs filed on bzr
[05:17] <lifeless> so we haven't done anything to fix that 
[05:17] <jamesh> are you saying that bzr won't necessarily maintain the order of the later parent ids?
[05:18] <lifeless> no, I'm saying that it doesn't care what the order of the later ids is.
[05:19] <jamesh> okay.
[05:20] <jamesh> the second question was about testing overwriting the branch history with a shorter version that shared some history.  Branch.set_revision_history() looks appropriate, but I was wondering if it is intended for public usage
[05:20] <lifeless> we are normalising history now
[05:20] <lifeless> what situation you want to create?
[05:22] <jamesh> the one ddaa mentioned was the original history being something like "r1 -> r2 -> r3 ->r4" and the new history being "r1 -> r3 -> r4" (r3 having r1 and r2 as parents)
[05:22] <lifeless> right
[05:22] <lifeless> nowadays the pull that creates that will normalise it to r4->r4[0] ->r4[0] [0] ->r4[0] [0] [0]  etc
[05:23] <lifeless> to test it though - why do you need a branch like that, why not just inject the revision history you want to use into the relevant method ?
[05:24] <jamesh> that's what I was thinking of doing, and wondering if Branch.set_revision_history() is the appropriate method
[05:24] <jamesh> or whether it was considered internal only
[05:24] <lifeless> for now it is. I'm suggesting modifying the scanner code to allow (for instance), Scanner.process_revision_history([r1, r2, r3, r4] )
[05:24] <lifeless> that is, make the code easier to test.
[05:28] <jamesh> that should be possible with the current code by overwriting the BzrSync.bzr_history attribute
[05:28] <jamesh> I might try that instead
[05:29] <lifeless> with robust tests, its better to test the specific thing that you are changing rather than environmental aspects.
[05:30] <lifeless> i.e. we dont want to test bzr, we want to test what the scanner does.
[07:27] <welshbyte> morning
[08:38] <dilys> Merge to devel/launchpad/: [r=salgado]  Fix bug 36058 (Bug tasks need more dates) (r1831: Brad Bollenbach)
[09:16] <giftnudel> who is responsible for security related bugs in launchpad?
[09:21] <BjornT> giftnudel: we don't have any specific person for that atm, so if you report a security related bug, please subscribe the 'launchpad' team.
[09:22] <giftnudel> ok, I will do that
[09:34] <carlos> morning
[09:37] <giftnudel> BjornT: done, https://launchpad.net/products/launchpad/+bug/41835
[09:37] <BjornT> thanks giftnudel 
[10:13] <SteveA> morning
[10:35] <carlos> SteveA: good morning
[10:37] <TomaszD> hi, is this a lauchpad support channel? I have a question regarding rosetta
[10:37] <TomaszD> *launchpad
[10:38] <TomaszD> why gnome-cups-manager template still isn't imported to dapper?
[10:38] <TomaszD> we have like two weeks to go
[10:38] <TomaszD> and one of the main desktop apps just isn't there
[10:38] <TomaszD> how come?
[10:43] <jordi> carlos: which missplaced upload?
[10:43] <carlos> jordi: is a bug, don't worry
[10:45] <jordi> carlos: oh, ok
[10:55] <ddaa> hello
[10:55] <ddaa> spiv: welcome back
[10:56] <carlos> TomaszD: hi, sorry, didn't see your request
[10:57] <carlos> TomaszD: this is the right channel to ask about this kind of issues
[10:57] <TomaszD> oh
[10:57] <TomaszD> ok, so? :)
[10:57] <carlos> let me check...
[10:57] <carlos> TomaszD: it's there: https://launchpad.net/distros/ubuntu/dapper/+source/gnome-cups-manager/+pots/gnome-cups-manager
[10:58] <TomaszD> oh my
[10:58] <TomaszD> how come...
[10:58] <TomaszD> Jesus Christ
[10:58] <carlos> It was imported on 2006-03-20
[10:59] <TomaszD> I'm pretty sure the launchpad.net/rosetta points to hoary...
[10:59] <carlos> TomaszD: how where you looking for it?
[10:59] <carlos> s/where/were/
[10:59] <carlos> oh, that link...
[10:59] <carlos> mpt__: around?
[10:59] <carlos> mpt__: ^^^
[10:59] <SteveA> good morning carlos
[11:00] <TomaszD> anyhow, thanks carlos !
[11:00] <carlos> mpt__, jordi: I think we should show the product list with translations from productseries to prevent this kind of confusion (talking about launchpad.net/rosetta/)
[11:00] <carlos> TomaszD: you are welcome
[11:01] <carlos> TomaszD: please, use https://launchpad.net/distros/ubuntu/dapper/+translations to locate dapper translations
[11:01] <TomaszD> carlos, ok thanks!
[11:01] <carlos> TomaszD: you are welcome
[11:04] <SteveA> BjornT: hello.  do you know if the email bug comment notification format is as it should be in production?
[11:04] <SteveA> i want to file a bug on a change, but i won't do so if there are still changes to land.
[11:05] <BjornT> SteveA: there are no changes pending, so please do file a bug.
[11:05] <SteveA> ok
[11:06] <jordi> carlos: what if the prodcut doesn't have series translations?
[11:06] <carlos> jordi: gnome-cups-manager, evolution and others
[11:07] <carlos> jordi: oh, sorry, I didn't read your question correctly
[11:07] <carlos> jordi: the suggestion is that we don't add such products on that list
[11:07] <carlos> people have access to it from the distro pages
[11:08] <jordi> nod
[11:08] <carlos> so the product list at launchpad.net/rosetta would became the list of products that agree to use Rosetta
[11:12] <carlos> SteveA: How could I define a variable with tal that gets an integer as value?
[11:13] <carlos> I'm using: <tal:limit define="limit 3">
[11:13] <carlos> but I get: Warning: zope.tal.taldefs.TALError: Invalid variable name "3" in expression u'3', at line 176, column 15
[11:13] <SteveA> carlos: first... why do you want to do this?
[11:13] <SteveA> this sounds like programming in a page template
[11:13] <SteveA> and we try to not do that
[11:14] <SteveA> the answer anyway is that any plain expression in TALES is interpreted as a "path expression"
[11:14] <SteveA> like context/foo/bar/baz
[11:14] <SteveA> or view/bar/baz/spoo
[11:14] <carlos> SteveA: It's for suggestions code, we are sharing macros between two translation forms
[11:14] <SteveA> so, you've asked for a context called "3", rather than "view" or "context"
[11:14] <carlos> one has no limit and the other has a limit
[11:14] <SteveA> so it doesn't work
[11:15] <carlos> and both uses the same view
[11:15] <SteveA> macros have access to the view class
[11:15] <mpt> carlos, what do you mean by "the product list with translations from productseries"?
[11:15] <mpt> What's an example of the page you're talking about
[11:15] <carlos> mpt: https://launchpad.net/rosetta as a list of products with translation resources
[11:16] <carlos> mpt: but many of them come from ubuntu so people think they need to look there for ubuntu translation resources
[11:16] <SteveA> carlos: you can say <tal:define tal:define="limit python:3">
[11:16] <carlos> SteveA: but the view has no way to know the kind of template we are rendering
[11:16] <SteveA> carlos: actually it does
[11:16] <mpt> carlos, by "them" do you mean products, or people?
[11:16] <carlos> well... looking into the requested URL 
[11:16] <SteveA> and i'm going to make this a lot better in the near future
[11:17] <SteveA> with some better view support for launchpad
[11:17] <carlos> mpt: translation resources
[11:17] <SteveA> but meanwhile, i guess you can do this in TALES, although i do not like it
[11:18] <mpt> carlos, sorry, I don't understand the problem. Is it that people are scrolling down as far as the "Translations in Ubuntu", but not as far as "Translatable applications"?
[11:18] <carlos> SteveA: If you prefer to check the URL to set that value, I could do it too
[11:18] <SteveA> no, i don't like that either
[11:19] <SteveA> use TALES and please file a bug on me about it, indicating exactly what page template files and view classes are involved
[11:19] <carlos> mpt: They ignore the 'Translations in Ubuntu' section and look at 'Translatable applications' because we show there too the translations from Ubuntu
[11:19] <SteveA> and that it is for one view class and two templates, but with these different "limit" values
[11:19] <carlos> mpt: if we don't have a link there people think we didn't import it yet
[11:19] <SteveA> what is the "limit" for anyway?
[11:20] <SteveA> is that a good variable name?
[11:20] <mpt> That is messed up\
[11:20] <carlos> SteveA: ok
[11:20] <mpt> "The recommended target for current translation activity is Hoary gaim." Why?
[11:21] <carlos> SteveA: is the number of entries we are going to render, it controls the number of entries from a list that we are going to render, it acts as the 'limit' of entries
[11:21] <carlos> mpt: because no one is maintaining the links between products and distros
[11:21] <carlos> mpt: that's one of the problems, we talked about at London, do you remember?
[11:21] <mpt> carlos, why are we suggesting they translate distro packages at all?
[11:21] <SteveA> mpt: menus questions abound
[11:21] <carlos> mpt: because Mark asked us to implement it that way
[11:22] <SteveA> carlos: maxentriesrendered
[11:22] <carlos> SteveA: ok
[11:22] <SteveA> would be more descriptive than "limit"
[11:22] <SteveA> or just maxentries
[11:22] <SteveA> because "limit" can apply to *anything*
[11:22] <SteveA> so it is almost meaningless
[11:22] <SteveA> mpt: so please join the appropriate channel
[11:23] <carlos> mpt: I think that feature is good to have, but that page should render only applications that upstream maintains using Rosetta
[11:23] <carlos> SteveA: ok
[11:24] <mpt> carlos, if upstream uses Rosetta, that's precisely when we should be steering people to upstream translation, not distro translation!
[11:24] <mpt> Fall back to distro translation if the upstream developers don't use Rosetta
[11:24] <carlos> mpt: that's my point, I'm suggesting to show there only the products that have the Rosetta flag set
[11:25] <mpt> carlos, agreed
[11:25] <carlos> ok
[11:25] <carlos> I hope people start using the distro links after that change is done
[12:03] <carlos> SteveA: https://launchpad.net/products/rosetta/+bug/41858
[12:03] <Ubugtu> Malone bug 41858 in rosetta "<tal:limit define="maxentries python:3"> should be moved to python code" [Normal,Confirmed]  
[12:05] <SteveA> ta
[12:06] <mpt> SteveA, lost you
[12:07] <TomaszD> carlos, how about dapper xchat? https://launchpad.net/distros/ubuntu/dapper/+source/xchat/+translations
[12:08] <carlos> TomaszD: Rosetta has only packages from main
[12:08] <carlos> TomaszD: xchat is from universe
[12:08] <TomaszD> oh
[12:09] <TomaszD> hmm, so to get the translation in place, I'd have to ask for an upload to xchat's cvs then
[12:09] <carlos> TomaszD: either that or ask its ubuntu maintainer
[12:09] <SteveA> mpt: oh well
[12:09] <SteveA> irc it is
[12:09] <TomaszD> carlos, ok thanks
[12:09] <carlos> we were asked by MOTUs to not import those packages
[12:09] <TomaszD> why?
[12:10] <carlos> because we don't have language packs for universe and is hard to them to track the updates and do new uploads so those translations are not lost
[12:10] <TomaszD> oh I get it now, xchat was in main, that's why it had a translation before (in breezy)
[12:11] <TomaszD> OK I'll see about getting some feedback from the ubuntu maintainer
[12:19] <carlos> TomaszD: well, we imported universe for breezy and hoary, but the problem is the same, those translations are not being used in Ubuntu :-(
[12:58] <ddaa> hey stub
[12:59] <stub> Morning
[12:59] <ddaa> I'm looking at the person merging stuff for branches
[01:00] <ddaa> I do not think you actually read the doctest:
[01:00] <ddaa> +Branches with greater IDs will get greater numerical suffixes, so we create
[01:00] <ddaa> +"peoplemerge" first, and it will be renamed "peoplemerge-1". The extant
[01:00] <ddaa> +"peoplemerge-1" branch will be renamed "peoplemerge-2".
[01:01] <ddaa> I agree that as I described it, it's difficult to implement in sql, but the annoying bit is that now the doctest is contradicting itself.
[01:03] <stub> ok. So the text needs to be updated. I'm not too fussed if the renaming isn't particularly obvious - it will be extremely rare for the case to occur.
[01:06] <ddaa> how much longer are you around today?
[01:14] <ddaa> stub: for id, product, name in list(cur.fetchall()):
[01:14] <ddaa> any reason you use an explicit list there?
[01:15] <ddaa> it will end up using the iterator interface anyway
[01:19] <stub> ddaa: I'll be around for another 5 hours I think
[01:19] <stub> ddaa: I use list() there because I reuse the cursor in the loop
[01:19] <stub> (I think)
[01:19] <ddaa> good point
[01:19] <ddaa> thanks
[01:30] <ddaa> stub: please review https://chinstrap.ubuntu.com/~dsilvers/paste/filebtWwvB.html
[02:08] <Keybuk> psycopg.OperationalError: server closed the connection unexpectedly
[02:08] <Keybuk>         This probably means the server terminated abnormally
[02:08] <Keybuk>         before or while processing the request.
[02:08] <Keybuk> ; used connection string 'dbname=launchpad_prod user=queued host=jubany.ubuntu.com'
[02:08] <Keybuk> cprov !!! help :p
[02:09] <cprov> Keybuk: uhm .. this ramdom DB unavailability again, I can't do much for you. Which tool ?
[02:09] <Keybuk> queue
[02:10] <Keybuk> so if I retry, it should work?
[02:10] <cprov> Keybuk: yes
[02:11] <cprov> Keybuk: I wonder what is going so wrong latelly with jubany, maybe the obvious, high concurrence
[02:12] <carlos> see you later
[02:12] <Keybuk> cprov: now it's just being weird
[02:13] <Keybuk> I'm getting "Accepting ubuntu/dapper (NEW) 0/18"
[02:13] <Keybuk> then just "0/18 total"
[02:13] <Keybuk> oh, maybe I'm using it wrong
[02:14] <cprov> Keybuk: there is 18 items in NEW and your filter isn't selecting anyone for acceptation
[02:14] <cprov> there are ...
[02:14] <Keybuk> what's the filter supposed to be?
[02:15] <Keybuk> I put "language-pack-*" for that
[02:15] <cprov> name matching , i.e., queue -Q NEW accept foo
[02:15] <Keybuk> that didn't work
[02:15] <Keybuk> if I put the queue ids in, it works
[02:15] <Keybuk> (one at a tiem)
[02:15] <cprov> Keybuk: don't use '*'
[02:15] <Keybuk> oh, filter implied wildcards to me
[02:15] <stub> Keybuk: you got that just now??
[02:15] <Keybuk> stub: yeah
[02:16] <cprov> Keybuk: sorry, I didn't land wildcard support yet. Should be simple but I haven't time to investigate
[02:17] <Keybuk> ahh
[02:18] <cprov> Keybuk: use substring, as in a SQL LIKE %<>%
[02:18] <Keybuk> s'ok, I just did the ids :)
[02:18] <stub> cprov: Is this tool connecting as user 'fiera'?
[02:19] <cprov> stub: queued
[02:20] <stub> queued@launchpad_prod/29258 2006-04-28 13:15:55 BST LOG:  unexpected EOF on client connection
[02:20] <stub> Nothing else in the logs relevant
[02:20] <cprov> stub: config.uploadqueue.dbuser (ftpmaster config)
[02:20] <cprov> stub: have you seen it before in our env. ?
[02:20] <stub> Nope
[02:21] <stub> Possibly network issues, possibly something odd like the ident daemon on drescher not responding quickly enough and some timeout being triggered in the auth process.
[02:22] <stub> Doesn't look like anything wrong at the PostgreSQL level though, at least as far as the logs let me know.
[02:23] <cprov> stub: can you file a bug for this ? then we can track the same behavior in other situation like the publisher or the slave_scanner (I've seen the same in those lands)
[02:23] <stub> I've got other 'unexpected EOF' messages for the fiera user
[02:24] <cprov> stub: it's the slave_scanner
[02:24] <stub> And some others, like bugnotification. I suspect it is normal for a psycopg connection that isn't explicitly closed and a false alarm.
[02:24] <cprov> stub: or queue_builder (both are buildmaster children)
[02:27] <cprov> stub: maybe, if the process has already finished DB tasks it's ok, but it also happens in the middle of some intensive processes, i guess the transaction is just aborted since I've never get inconsistenses.
[02:27] <stub> transaction will always be rolled back
[02:28] <cprov> speaking about which, once a while publisher fails on a "serialization" problem, have you observed the same issue in other applications ?
[02:29] <stub> serialization exceptions are normal if you are using serialized transaction isolation level, which is the default with psycopg 1. Either change the transaction isolation level, or catch the exceptions and retry the transaction.
[02:29] <stub> Retrying the transaction is the documented way of handling this expected case.
[02:30] <stub> I'd consider dropping the transaction isolation level unless the lack of read consistency will cause your program problems.
[02:30] <cprov> stub: how do I retry the transaction ? shouldn't we have support for this in the commit method ?
[02:31] <stub> cprov: You need to redo what you where trying to do. commit can't do it automatically, as the transaction might need to do something different the second time round as the data in the database may have changed.
[02:31] <cprov> stub: publisher will probably failed on pending writes (publising & domination, both huges most of the time)
[02:31] <stub> In fact, it almost certainly have changed as the serialization exception was triggered.
[02:32] <cprov> stub: wait, did you mean changed by other transaction ?
[02:32] <stub> changed by another process I mean.
[02:33] <stub> Your script has requested the highest possible transaction isolation level. Some other process has mucked with the data in such a way that that transaction isolation level can no longer be guaranteed. So PostgreSQL raises the exception and expects you to deal with it, generally by trying whatever you were doing again.
[02:33] <cprov> stub: I see, the exception will be raised when I commit or anywhere after the transaction has been started ?
[02:34] <stub> I can't remember - whenever I've needed to catch it I've always wrapped the entire thing in a try: except: so haven't needed to worry.
[02:35] <cprov> Keybuk: did queue work for you ? 
[02:36] <Keybuk> cprov: seems to be working ok
[02:37] <cprov> stub: right, will find sometime to play around it during the weekend. I guess a good way to reproduce it should be invoking another script to mess with DB data in question after start a transaction. Anyway will figure out
[02:38] <cprov> Keybuk: ok, sorry for unsupported/undocummented feature, feel free to file a bug on missed wildcards functionality, it's also missing in the web UI implementation.
[02:39] <Keybuk> nah, s'ok -- was more a bug in my own docs of how to drive the tool I suspect :p
[02:39] <Keybuk> this was the first time I'd tried to accept more than one package at a time
[02:40] <stub> cprov: it is a pita to reproduce, as you need to mutate the data in the right way - if it is possible for PostgreSQL to cope, it copes. lib/canonical/launchpad/webapp/ftests/test_adapter.txt has some tests for Z3 handling of it. Most scripts I've just reduced the isolation level though - generally you don't need SELECTS to guarantee the same results on repeated queries, and sometimes this behavior even makes things more difficult.
[02:42] <cprov> stub: I'm scared now :(
[02:43] <stub> cprov: Its not really scary. Most db's don't even support this high a transaction isolation level at all.
[02:45] <cprov> stub: I wonder if a better approach should be identify which scripts are conflicting horrible and lockfile them.
[03:35] <stub> ddaa: https://chinstrap.ubuntu.com/~dsilvers/paste/filebtWwvB.html r=stub
[03:46] <doko> carlos: Could you give me translations "Get help Online" and "Translate this application" in all languages?
[04:07] <carlos> doko: aren't you able to use the translations from gtk+ or whatever it's stored for GNOME?
[04:07] <doko> carlos: OOo ...
[04:08] <carlos> where are the gtk/gnome ones stored? the source code that contains the english string
[04:14] <doko> carlos: lunchpack-integration
[04:16] <carlos> doko: https://launchpad.net/distros/ubuntu/dapper/+source/launchpad-integration/+pots/launchpad-integration/+export
[04:17] <carlos> then, there you have the translations we have 
[04:18] <doko> carlos: breezy does have more, but ok, that probably better to check this location regularily
[04:19] <doko> carlos: there's one bug: I cannot download the files in GSI format ;-P
[04:19] <carlos> doko: I need to migrate breezy's translations to dapper, that's the reason
[04:20] <carlos> doko: btw, aren't you going to add those strings to oo.org GSI files?
[04:23] <doko> carlos: hmm, I should, but it seems, they are currently not ...
[04:25] <doko> carlos: indeed, they are
[04:39] <salgado> stub, around?
[04:39] <stub> salgado: yes
[04:39] <salgado> great
[04:39] <salgado> stub, so, I need some help with the migration script for shipit
[04:40] <salgado> basically, I need to insert empty RequestCDs objects for the new flavours
[04:40] <salgado> s/objects/rows
[04:40] <salgado> not really empty, but with quantity=0 and quantityapproved=0
[04:41] <salgado> so, for every existent RequestCDs row, I need to insert two new ones, changing only the flavour column
[04:41] <kiko> GOOD MORNING WORLD
[04:41] <SteveA> yo
[05:03] <stub> salgado: That sounds dodgy. Why store the rows with quantity=0 and quantityapproved=0 at all? Tripling the data we store doesn't sound a good solution when slightly more intelligent code would handle it fine.
[05:08] <salgado> stub, yes, I agree with you... the problem is that I have to pass these RequestedCds objects from browser to database code in a few different places. and I'm afraid it'll be quite some work to change all these places to handle variable numbers of RequestedCDs objects
[05:09] <SteveA> the API can still do that
[05:09] <SteveA> doesn't mean it needs to be that way in the database
[05:09] <kiko> salgado, I still think stub is right -- otherwise, what happens when you want to add more archs and products?
[05:09] <kiko> listen to SteveA 
[05:09] <SteveA> then, you can consider refactoring later
[05:12] <salgado> I was just afraid it'd make the code even more complex. I have to admit I didn't investigate it well enough, though
[05:13] <salgado> maybe you can help me with that this afternoon, kiko?
[05:13] <kiko> sure!
[05:13] <kiko> thanks for asking
[05:13] <kiko> remind me
[05:13] <stub> I suspect offering three flavours instead of one will more than double the number of requested cds we sent out (and double the number that end up as drinks coasters). Even with hindsight though, we will have no way of knowing for sure (as we will also get an unknown increase due to increased publicity and public awareness).
[05:14] <kiko> stub, we only offer one flavor per site
[05:14] <kiko> i.e. 
[05:14] <kiko> the guy needs to go to shipit.edubuntu.org for edubunto CDs
[05:14] <kiko> edubuntu argh
[05:14] <kiko> so it's extra work to get more CDs
[05:14] <kiko> we also changed the workflow for custom requests
[05:15] <kiko> salgado, I'm pretty sure the checkbox is going to be important. make it a UI-only Javascript thing?
[05:17] <salgado> kiko, you mean, only enable the text box when the checkbox is clicked? (I'm assuming this is what we're going to use unless someone has a better idea)
[05:17] <kiko> yeah
[05:17] <kiko> do it
[05:22] <SteveA> hi tom
[05:22] <th1a> Hi SteveA.
[05:22] <kiko> yawns 
[05:22] <th1a> We're having a little LaunchPad problem.
[05:22] <SteveA> what's up?
[05:23] <bradb> kiko, BjornT, SteveA: Any thoughts on https://wiki.launchpad.canonical.com/BugDateLastUpdated ? I was planning on putting it up for review soon.
[05:23] <th1a> SteveA: jinty is the "registrant" for SchoolTool, but he can't seem to elevate my role on the project.
[05:24] <SteveA> this will allow us to do kinda roundup things, bradb 
[05:24] <bradb> (btw, by "it", I mean the code.)
[05:24] <kiko> bradb, isn't there a general "bug-modified" event?
[05:24] <kiko> bradb, BjornT: if not, shouldn't there be?
[05:24] <kiko> I think an update_bug_date_last_updated event is nonsensical
[05:25] <bradb> There isn't such an event now.
[05:25] <kiko> well that'd make a lot more sense
[05:25] <SteveA> th1a: what role is he trying to give you?
[05:26] <bradb> kiko: That's a pretty quick conclusion. :)
[05:26] <kiko> bradb, I mean, well, what do YOU think? :)
[05:26] <SteveA> th1a: https://launchpad.net/products/schooltool/+reassign ?
[05:26] <th1a> SteveA:  Well, I should be able to do just about everything.  I'm not sure what you call that.
[05:26] <bradb> It's not really that simple. There are various events that can happen to consider a bug being "modified", like adding a comment, adding a watch, changing the status/assignee/etc.
[05:26] <SteveA> th1a: maybe you want to create a team with you and brian in it
[05:27] <kiko> all of these, IMO, should trigger an update to the date, bradb 
[05:27] <th1a> SteveA:  I'm not sure what he did.
[05:27] <SteveA> and then use that page to make tha tteam the owner?
[05:27] <bradb> Different objects being modified, and even only specific fields on those objects.
[05:27] <jinty> SteveA: I get permission denied when I try that
[05:27] <matsubara> SteveA: looks like bug 41639
[05:27] <SteveA> jinty: when you try to use +reassign
[05:27] <Ubugtu> Malone bug 41639 in launchpad "Product owner should be able to reassign ownership to another user." [Normal,Confirmed]  http://launchpad.net/bugs/41639
[05:27] <bradb> kiko: someone new being subscribed to a bug causing it to bump the bug up the sort order? ew. :)
[05:27] <jinty> SteveA: to be exact, I can view it, but not change it
[05:28] <SteveA> i see
[05:28] <kiko> bradb, well, you could argue it either way..
[05:28] <SteveA> who do you want to be the product owners?
[05:28] <SteveA> do you have a team, or should it just be tom?
[05:28] <SteveA> thanks matsubara 
[05:30] <bradb> kiko: The point of this functionality is to help steer devs into spending their time on things where something has changed, instead of having them polling bug reports. It seems highly unlikely to me that adding a subscriber (or, for example, changing a bug nickname, perhaps) is important enough to bring the bug to the devs attention.
[05:30] <th1a> SteveA:  A team is probably better.
[05:30] <bradb> s/attention/attention again/
[05:31] <kiko> bradb, you're somewhat right, but where do you draw the line?
[05:31] <bradb> kiko: At a subscriber being added, for one. :)
[05:31] <SteveA> why?
[05:31] <SteveA> that shows that at least one more person is interested
[05:32] <bradb> SteveA: And...?
[05:32] <SteveA> well, you fix a bug depending on the number of people it affects, at least to some extent
[05:32] <SteveA> if it affects 0 people, you need not fix it
[05:32] <SteveA> if it affects 100% of people, you must fix it
[05:33] <SteveA> and subscribing is an indication of interest in the resolution
[05:33] <SteveA> which implies being affected
[05:34] <SteveA> th1a: do you have a team in mind?
[05:34] <kiko> bradb, #cm
[05:35] <th1a> SteveA:  It could be jinty and I.  In theory sabdfl should be an owner, too ;-)
[05:35] <carlos> stub: hi, I think staging mirror is broken atm...
[05:35] <SteveA> you should be able to create such a team for yourselves in launchpad
[05:36] <SteveA> and give it a name like schooltool-owners
[05:36] <SteveA> let me know when you've done that
[05:36] <SteveA> and i'll transfer ownership of the schooltool product
[05:38] <stub> ZConfig.ConfigurationSyntaxError: 'warn_about_no_published_uploads' is not a known key name (line 95 in file:///srv/launchpad.ubuntu.com/staging/launchpad/configs/staging/launchpad.conf) 
[05:38] <stub> Any reason why that is required?
[05:40] <salgado> stub, I dropped this config option, but aparently I forgot to remove it from staging's config. sorry for that
[05:40] <stub> ahh... yes. I had it backwards :-)
[05:47] <carlos> kiko: I have the PoMsgSetPage implemented already, I'm fixing tests now
[05:47] <kiko> YES!
[05:48] <SteveA> but please note the correct way to do this is...
[05:48] <SteveA>  "I have the PoMsgSetPage tests implemented already.  I'm fixing the implementation now"
[05:48] <kiko> well, the page has impact on existing tests, SteveA 
[05:49] <SteveA> of course
[05:49] <SteveA> which can be changed first too :-)
[05:49] <kiko> and some changes are rather unpredictable 
[05:49] <SteveA> but, you'd find those needed changes as you worked on the implementation
[05:49] <carlos> well, I'm getting test failures that should not be failing in first place...
[05:50] <SteveA> BjornT: what's the story on getting us able to use the new test facilities in zope3.2 ?
[05:50] <carlos> I'm reviewing the diff to know why are they failing
[05:50] <SteveA> if you're getting failing page tests, it may be easier to rewrite them in the new style...
[05:50] <SteveA> if we have the new style available
[05:51] <carlos> is not a problem with pagetests
[05:51] <ddaa> okay finished stuffed pqm
[05:52] <carlos> but with doctests
[05:52] <SteveA> okay
[05:52] <ddaa> going watch a movie, I'll probably be back afterwards
[05:52] <SteveA> carlos: ping me if you need help with fixing the tests
[05:52] <BjornT> SteveA: well, there's one rt ticket needing to be resolved, then test.py needs a small patch. after that it's possible to use it.
[05:53] <carlos> SteveA: will do, thanks
[05:55] <SteveA> BjornT: what rt number, and can the patch be made conditional so it may be applied now?
[05:57] <BjornT> SteveA: actually, the rt ticket might have been resolved already, i can't find it. the patch to test.py can be applied anyway, it just removes '.' from PYTHONPATH
[05:57] <SteveA> please get that patch into PQM
[05:58] <kiko> do it
[05:58] <SteveA> pop the trunk
[06:03] <SteveA> BjornT: can't find the ticket in current or resolved RT tickets for launchpad
[06:03] <SteveA> it may be in the "pending classification" queue, which I cannot see
[06:08] <ozamosi> When I have a bazaar-branch added to launchpad, how do I merge it back? Is there some easy way, or is that manual?
[06:08] <BjornT> SteveA: ok. does https://chinstrap.ubuntu.com/~dsilvers/paste/filecOpT0g.html look ok? it's copy&pasted from Zope3's test.py
[06:09] <kiko> ozamosi, can you clarify what yo are asking? merge it back.. from where? to where? to accomplish what?
[06:13] <SteveA> BjornT: say something more descriptive than "this directory"
[06:13] <SteveA> like "this module's directory" (is that right?)  or "the current working directory" (don't think that's it)
[06:14] <SteveA> why is it important to remove that directory from the module search path?
[06:15] <ozamosi> kiko: (newbe on bazaar) I want to develop a webpage with bazaar, and I can easily add a branch to launchpad. But how do changes actually get to the server? Would that be done outside of launchpad?
[06:15] <BjornT> SteveA: ok. i'll also add a comment as to why we remove the directory. it's because testbrowser does 'from test import pystone'
[06:16] <BjornT> right, it's the module's directory which is removed
[06:16] <kiko> ozamosi, you push your branch to the supermirror, generally. have you asked on #bzr?
[06:17] <uws> Ehm... the supermirror pushes from my http branch
[06:17] <uws> I don't push it to launchpad
[06:17] <ozamosi> kiko: actually, no. I just wondered if there was something launchpad did for me.
[06:17] <uws> argh, what I meant to say is: "the supermirror PULLS from my http branch"
[06:18] <kiko> uws, the supermirror can do both, but new branches can just as easily push 
[06:18] <kiko> without requiring registration
[06:18] <uws> kiko: Are you talking about ubuntu stuff or just any 3rd party branch that is registered in lp?
[06:19] <kiko> any branch
[06:19] <SteveA> BjornT, carlos, bradb, stub: karl has installed the needed python packages on the servers.  We can start using the new test stuff.
[06:20] <carlos> cool
[06:20] <bradb> sweet. I await your code review!
[06:20] <uws> kiko: How do I push to lp? I don't have a ssh/sftp/ftp account ?
[06:20] <kiko> uws, just register your ssh key 
[06:20] <uws> kiko: And how does LP know it was me and not some evil hacker/
[06:20] <BjornT> SteveA: ok. should i set up a 'browser' object in the pagetest setup?
[06:20] <uws> kiko: Ah, ok. that makes sense ;)
[06:20] <kiko> uws, you can then easily push to the sftp server
[06:20] <SteveA> sure, make it nice
[06:26] <kiko> BjornT, are you in the middle of tasks?
[06:27] <BjornT> kiko: yeah. what's up?
[06:30] <kiko> BjornT, well, mdz is desperate to get bug 29752 fixed, and I'm looking for takers
[06:30] <Ubugtu> Malone bug 29752 in malone "If a bug is marked as a duplicate, its subscribers should be notified when the duplicate bug changes" [Critical,Confirmed]  http://launchpad.net/bugs/29752
[06:30] <kiko> salgado, could you perhaps do it?
[06:33] <kiko> or bradb?
[06:35] <bradb> I can do it.
[06:37] <kiko> mdz would love you and then I would love you too
[06:41] <bradb> love is the answer
[06:41] <kiko> to many questions
[06:50] <mdz> bradb: if you fix that bug, I may kiss you
[06:50] <kiko> mmmm kissing
[06:50] <ruffneck> on the butt? ;)
[06:51] <bradb> wow
[06:51] <bradb> friendship...with benefits
[06:54] <kiko> no way!
[06:54] <kiko> he makes me so envious this bradb character
[07:18] <dilys> Merge to devel/launchpad/: [r=stub]  improve branch renaming on people merging (r1832: David Allouche)
[07:19] <SteveA> the macadamia is a kind of nut
[07:20] <SteveA> large, nutritious.  grows in australia
[07:23] <sabdfl> hi all
[07:23] <sabdfl> what's the right magic to get RO access to the staging db from mawson?
[07:23] <sabdfl> psql -h asuka.ubuntu.com -U ro -d launchpad_staging
[07:24] <sabdfl> is giving me "permission denied for relation bug" errors
[07:25] <carlos> sabdfl: seems like the mirror done today failed
[07:26] <carlos> at least that's the usual error while the mirror is being done
[07:26] <sabdfl> ok, could you ping stub about that monda?
[07:27] <carlos> I did it already
[07:28] <sabdfl> thanks carlos
[07:28] <carlos> hmm, he left already
[07:28] <sabdfl> import/export looking much healthier, well done, go for 100% ok?
[07:29] <carlos> not sure if we was aware of the problem
[07:29] <carlos> sabdfl: yeah, that's the plan. Thanks
[07:33] <jsgotangco> hey guys, in malone is it really necessary to have the "subscribe me" checkbox for comment when I'm already part of the team responsible for the package? we're already subscribed in the first place...
[07:37] <kiko-fud> jsgotangco, good point. easy fix too
[07:38] <jsgotangco> ahh i guess the backend is smart enough to know im part of the team...that's good thanks kiko-fud
[07:39] <kiko-fud> just check inteam
[07:39] <kiko-fud> easy
[07:39] <kiko-fud> dude my merge is taking hours
[07:41] <salgado> kiko-fud, inTeam() is not the right thing in this case, I think
[07:41] <salgado> I think we need to check only for direct memberships
[07:41] <kiko-fud> mmmm?
[07:42] <salgado> indirect members won't get mail notification if a team has no contact address, IIRC, so they won't receive bugmail that is sent to that team and thus they might want to subscribe
[07:45] <carlos> SteveA, BjornT: Do we have any documentation about how to use the new testing system?
[08:00] <mdke> jordi, carlos: around?
[08:00] <carlos> mdke: yes
[08:02] <mdke> carlos: I uploaded a pot with just a few string changes, apparently hundreds of strings have been marked as fuzzy, any idea why?
[08:03] <carlos> the .pot upload doesn't set/remove the fuzzy flag
[08:03] <carlos> mdke: the .po upload is the one that affects that flag
[08:03] <mdke> well, I didn't upload any po files :)
[08:04] <mdke> so, I'm not culpable?
[08:04] <mdke> ;)
[08:05] <carlos> don't think so
[08:05] <mdke> good. thanks
[08:05] <carlos> unless we have a unknown bug...
[08:05] <carlos> mdke: where did it happen?
[08:06] <mdke> carlos: in french, apparently
[08:07] <mdke> no more specific than that
[08:07] <mdke> someone posted to -translators
[08:08] <carlos> mdke: then I don't think it's a bug in our side, at least not with your .pot upload
[08:08] <mdke> good, I'll let them know. Thanks
[08:08] <carlos> it should affect all languages
[08:08] <carlos> or none
[08:08] <carlos> it sounds much more like a new fr.po upload
[08:09] <mdke> yeah
[08:10] <mdke> thanks carlos 
[08:10] <carlos> you are welcome
[08:22] <jordi> weird things happen lately
[08:22] <kiko> yes?
[08:24] <jordi> like frenchs getting fuzzy strings :)
[08:33] <cprov> BjornT: ping
[08:38] <sfllaw> Well, that was silly of me.
[08:38] <sfllaw> https://launchpad.net/products/nautilus/+bug/34375
[08:38] <Ubugtu> Malone bug 34375 in nautilus "Multi-mode (audio+data) CDs always opened by Sound Juicer" [Unknown,Unknown]  
[08:38] <sfllaw> See the remote bug watches.
[08:38] <sfllaw> Ah.  Fixed it.
[08:38] <BjornT> cprov: pong
[08:38] <ruffneck> bong
[08:40] <cprov> BjornT: do you know why dogfood conf external email was reenable in LP ? 
[08:43] <BjornT> cprov: no, no idea. i'd guess someone enabled it by mistake, since i'd think it should be disabled to avoid spamming people.
[08:43] <cprov> BjornT: maybe Zope 3.2 migration process ?
[08:51] <BjornT> cprov: i'm looking in configs/dogfood in rf, and it seems that mail-configure.zcml (which enables sending mail) hasn't been touched for quite a while. maybe you disabled it locally only, and then some rollout procedure overwrote mail-configure.zcml to the one in rf?
[08:52] <cprov> BjornT: or maybe it wasn't really disabled for the UI, only zopeless.
[08:52] <cprov> BjornT: does it fix the problem > https://chinstrap.ubuntu.com/~dsilvers/paste/filemyhNUC.html ?
[08:58] <BjornT> cprov: almost. you need a <mail:stubMailer> as well. look how it's done in packages-include/mail-configure-normal.zcml
[08:58] <cprov> BjornT: ok, one sec
[09:00] <cprov> BjornT: again -> https://chinstrap.ubuntu.com/~dsilvers/paste/fileLKCzSk.html , please
[09:01] <AlinuxOS> carlos, great new rosetta :) it rocks
[09:02] <AlinuxOS> and I enjoy Ubuntu in Georgian :D
[09:02] <carlos> AlinuxOS: thank you ;-)
[09:02] <kiko> hey carlos 
[09:02] <carlos> AlinuxOS: but if you talk about any UI improvement, say thank you to mpt ;-)
[09:02] <kiko> how's PMSP?
[09:02] <carlos> kiko: hi
[09:02] <carlos> kiko: running tests
[09:02] <AlinuxOS> :)
[09:02] <kiko> and also, bug 41653, which you need to look at
[09:02] <Ubugtu> Malone bug 41653 in rosetta "Importation of po file goes in the wrong template" [Major,Confirmed]  http://launchpad.net/bugs/41653
[09:02] <carlos> I think I fixed all tests already
[09:02] <AlinuxOS> no I love evetything...
[09:02] <AlinuxOS> :D
[09:02] <kiko> and also this weird bug reported by mdke 
[09:03] <kiko> and finally, the oops bug :)
[09:03] <AlinuxOS> carlos, imports are more faster..I like this
[09:03] <carlos> kiko: well, person.txt keeps failing, but I guess is not my fault so I will take a closer look to it if PQM rejects the merge due that test
[09:03] <carlos> AlinuxOS: :-D
[09:03] <kiko> person.txt? where does it fail? paste?
[09:03] <carlos> kiko: dude, I cannot do so much things... :-)
[09:03] <kiko> yes you can
[09:04] <kiko> stop stopping waiting for tests, paralelize!
[09:04] <BjornT> cprov: yeah, that should work. i'm not sure it's a good idea to spam root, though, is it?
[09:04] <carlos> kiko: https://chinstrap.ubuntu.com/~dsilvers/paste/filevEsytx.html
[09:05] <carlos> kiko: I paralelize!, but I have also a huge backlog of mail from launchpad and rosetta-users
[09:05] <kiko> carlos, forget mail, focus on those 4 tasks :)
[09:05] <carlos> I'm now uptodate with launchpad mail
[09:05] <kiko> mail is less important
[09:05] <cprov> BjornT: it should not send any email, isn't it ? they will remain in the specified queue, right ?
[09:05] <kiko> so carlos that is an odd crash, and one that I thought stub and ddaa had fixed in RF tip yesterday. are you fully remerged?
[09:06] <carlos> I merged yesterday
[09:06] <carlos> should I get a new update?
[09:06] <carlos> and a better question.. how is that pqm accepted that merge?
[09:07] <kiko> I think a DB constraint was added first and then the code was fixed
[09:07] <kiko> IIRC
[09:07] <carlos> ok
[09:08] <kiko> carlos, do you have RF revno 1824?
[09:09] <carlos> I branched from 3504
[09:09] <AlinuxOS> carlos, for example if I want a 456 string of gnome-panel,can I access more faster ?
[09:09] <AlinuxOS> or if I know a name of a string in english
[09:09] <AlinuxOS> can I acces faster to translate it ?
[09:10] <carlos> you can change the offset variable to jump to it
[09:10] <carlos> +translate?offset=450
[09:10] <BjornT> cprov: no, stubMailer does send the email to the to_addr. we have no mailer which simply drops the email. maybe we should add one, it's not hard to do.
[09:10] <AlinuxOS> carlos, and if I know a string name ?
[09:11] <carlos> AlinuxOS: no way to search (yet)
[09:11] <AlinuxOS> I I like this (yet)
[09:11] <AlinuxOS> :D
[09:11] <cprov> BjornT: ok, could be.
[09:11] <AlinuxOS> the important thing is this (yet)
[09:11] <AlinuxOS> so you willl improve.
[09:12] <carlos> yeah
[09:12] <carlos> and it became a priority this month
[09:12] <carlos> so I guess we will have it soon
[09:14] <bradb> kiko: do you have time to drive-by the fix for bug 29752?
[09:14] <Ubugtu> Malone bug 29752 in malone "If a bug is marked as a duplicate, its subscribers should be notified when the duplicate bug changes" [Critical,In progress]  http://launchpad.net/bugs/29752
[09:14] <kiko> yeah
[09:14] <kiko> unless it's too long
[09:14] <kiko> BjornT is a better candidate though
[09:14] <bradb> kiko: it's easy: https://chinstrap.ubuntu.com/~dsilvers/paste/fileO7fPOK.html
[09:15] <AlinuxOS> carlos, and if I would like to translate main rosetta interface in my language ?
[09:15] <kiko> bradb, that's half a patch, hmmm. the end-user may be confused about getting the email
[09:15] <AlinuxOS> example Georgian :)
[09:16] <kiko> I thought you were going to copy subscribers over, bradb?
[09:16] <bradb> hm, that seems a bit dodgy to me, dunno
[09:16] <bradb> particularly when unduping
[09:16] <carlos> AlinuxOS: we cannot do it yet
[09:17] <AlinuxOS> carlos, but still yet :)
[09:17] <carlos> AlinuxOS: we are changing a lot our UI interface and that would trash a lot of strings from translators
[09:17] <AlinuxOS> so it will be in the future :)
[09:17] <carlos> yes, we did some testing already
[09:17] <kiko> bradb, well, how would you make it clear that these people were getting email?
[09:17] <AlinuxOS>  great
[09:17] <bradb> kiko: That the end user is confused about getting the email is, I think, a separate problem (because we've already had end-users report confusion about getting mail.)
[09:17] <AlinuxOS> my pleasure to translate it :)
[09:18] <kiko> bradb, well, you're certainly not making it any better. what do you propose?
[09:18] <AlinuxOS> I really like your job guys..it's changed a lot...and changed in a righ way!
[09:18] <AlinuxOS> great :)
[09:18] <bradb> kiko: Ideally, a footer that says exactly why you're getting this email.
[09:19] <kiko> bradb, no, that's not a very good answer
[09:19] <kiko> well
[09:19] <kiko> okay, maybe it's an answer
[09:19] <kiko> but if I look at the bug ideally I know what Persons we're mailing
[09:19] <kiko> and the fix you did does not contemplate that
[09:19] <kiko> nor are you raising that issue
[09:19] <carlos> AlinuxOS: glad to know it ;-)
[09:20] <cprov> BjornT: changed the to_addr to launchpad@ (active admin account on mawson). Do I have rs=bjornt on it ?
[09:21] <bradb> kiko: Subscribing them to the dup target would be somewhat more confusing, IMHO.
[09:21] <kiko> it's how bugzilla does it, and what the bug commenter suggested
[09:21] <kiko> but you're not being helpful. suggest a way forward
[09:21] <AlinuxOS> so I exit :) to much bear that waits!!!
[09:21] <AlinuxOS> :D
[09:21] <AlinuxOS> god bless you all..
[09:24] <BjornT> cprov: hmm, don't know :) i have no idea where the email should go on mawson. you should set the from_addr to that address, though.
[09:25] <kiko> bradb?
[09:25] <cprov> BjornT: come on ;) changed also from_addr to launchpad@
[09:25] <kiko> I have a suggestion perhaps
[09:26] <bradb> kiko: For the display side of things, I'm thinking of a "Subscribers of duplicates of this bug:" section in the subscriptions portlet
[09:26] <kiko> which would be including dupe subscribers in the portlet
[09:26] <kiko> maybe just "From duplicates:" ?
[09:26] <bradb> Yeah, trying to thinking of how to make it concise
[09:26] <kiko> there's a problem with doing that though
[09:26] <kiko> and that is that your notificationRecipientAddresses already pulls emails
[09:27] <BjornT> cprov: well, if you think that's the right thing to do, sure rs=bjornt.
[09:27] <bradb> yeah, imp. details. no big deal.
[09:27] <kiko> you'd need to factor that into notificationRecipients and then use that in the portlet and from notificationRecipientAddresses
[09:27] <kiko> does that sound crash-prone?
[09:27] <cprov> BjornT: right, thx dude
[09:27] <kiko> will it work when there are no dupes? when there are dupes with no subscribers?
[09:28] <bradb> When there are no subscribers from dupes (including meaning there may be no dupes at all), it just wouldn't be shown, I think.
[09:28] <kiko> okay
[09:28] <kiko> that sounds like a good start
[09:29] <kiko> making it dynamic but visible
[09:29] <carlos> I need to leave now...
[09:29] <carlos> have a good weekend
[09:29] <bradb> thinking about the email itself though, hmm...
[09:30] <kiko> bradb, yeah. but maybe we can leave the email itself to be solved together with the footer?
[09:30] <bradb> sounds good
[09:30] <kiko> fine by me then
[09:30] <bradb> thanks
[09:30] <kiko> test looks good
[09:30] <carlos> kiko: I will ask for a review for the PoMsgSetPage branch tomorrow, I need to update a couple of pagetests for some broken links I detected
[09:31] <kiko> carlos, why don't you place it for review now?
[09:31] <kiko> carlos, pagetest link updates can either be reviewed separately or not at all if it's really trivial
[09:32] <carlos> ok
[09:34] <kiko> hey
[09:34] <kiko> why did revision numbers drop all of a sudden?
[09:35] <carlos> I think it's a bzr bug
[09:35] <kiko> okay, sent mail to lifeless
[09:35] <carlos> that mpt told me that lifeless said that would be fixed automatically when he fixes the bug that caused it
[09:35] <carlos> but I'm not sure I understood it correctly
[09:37] <kiko> weird
[09:41] <carlos> kiko: my branch is now added to pendingreviews page
[09:41] <carlos> see you!
[09:41] <kiko> what about the bugs!
[09:42] <carlos> dude, I cannot look into them atm
[09:43] <kiko> you did say you were going to look into that bug today
[09:43] <carlos> I'm going out to have dinner and is already 21:45 here..
[09:43] <kiko> yeah, but next time, try setting aside some time for analyzing fixing oopses -- even if that means having an hour less of PMSP
[09:44] <carlos> I know... but I was concentrated with PoMsgSetPage and decided to keep on track to kill it
[09:44] <kiko> anyway, night
[09:44] <carlos> ok
[09:50] <carlos> kiko: you are a lucky man, I got a call and my friend will be 30 minutes late
[09:50] <kiko> heh
[09:50] <kiko> I was born lucky
[09:50] <carlos> kiko: https://launchpad.net/products/rosetta/+bug/41943 <- This is a transaction problem that should be reviewed by stub
[09:50] <Ubugtu> Malone bug 41943 in rosetta "SQLObjectNotFound in +translante page" [Normal,Unconfirmed]  
[09:50] <carlos> if you look at the OPPS, it has the INSERT for the missing object that raised the exception
[09:50] <kiko> oh no
[09:51] <carlos> so we have  arace condition
[09:51] <kiko> one of those again
[09:51] <carlos> well, either stub or spiv
[09:58] <carlos> kiko: and confirmed, https://launchpad.net/products/rosetta/+bug/41653 is not an obvious bug in our side
[09:58] <Ubugtu> Malone bug 41653 in rosetta "Importation of po file goes in the wrong template" [Major,Confirmed]  
[09:58] <kiko> hmmm
[09:58] <kiko> carlos, the database looks good?
[09:58] <carlos> kiko: is either a cache problem or a bad upload from the user
[09:58] <carlos> kiko: yeah, I was confused because we didn't get the associated pofile automatically after the upload
[09:59] <carlos> now that I checked the source code
[09:59] <kiko> and why not?
[09:59] <carlos> I see that we link to the potemplate, not the pofile
[09:59] <kiko> ah
[09:59] <carlos> and checked again the import queue
[09:59] <carlos> https://launchpad.net/rosetta/imports/+index?status=APPROVED&type=all&start=75&batch=75
[09:59] <carlos> look for my name there
[09:59] <kiko> how could it be a cache problem?
[09:59] <carlos> and you will see that my import is attached to the right pofile
[10:00] <kiko> oh I see
[10:00] <carlos> kiko: well, I it's just an option
[10:00] <carlos> is a bit hard
[10:00] <carlos> as the link to the potemplate is done when we create a new object that represents an entry on the queue
[10:01] <carlos> entry = TranslationImportQueueEntry(path=path, content=alias,
[10:01] <carlos>                 importer=importer, sourcepackagename=sourcepackagename,
[10:01] <carlos>                 distrorelease=distrorelease, productseries=productseries,
[10:01] <carlos>                 is_published=is_published, potemplate=potemplate)
[10:01] <carlos> so I guess that if the object is created.... the cache should not be a problem at all
[10:01] <carlos> and thus, we only have the option that the user did a mistake with the upload...
[10:01] <kiko> could he have?
[10:03] <carlos> hmm... is the only option I can think on
[10:03] <carlos> because we have the potemplate information
[10:03] <carlos> and the guessing code, in that case does:
[10:03] <carlos> pofile = self.potemplate.getPOFileByLang(
[10:03] <carlos>             language.code, variant=language_variant)
[10:04] <carlos> and is 100% impossible that we get a POFile from another potemplate
[10:04] <kiko> could there be a conflicting potemplate and we are picking the worst one?
[10:04] <carlos> conflicting potemplate?
[10:04] <kiko> well, another one in the same path
[10:04] <carlos> no, we don't take the path in account in this case
[10:05] <kiko> I se
[10:05] <carlos> he choosed the potemplate and the pofile
[10:05] <kiko> e
[10:05] <carlos> hmmm, I found a possible option that would produce this problem if he didn't choose the pofile
[10:06] <kiko> ahm
[10:06] <carlos> but is a side effect of another know bug that allows two pofiles from different potemplates and the same productseries or sourcepackagename being in the same path
[10:06] <carlos> but that would mean too that he didn't use the URL he gave us
[10:08] <carlos> kiko: bug updated
[10:09] <carlos> the other bug I just commented is related to the poimport breakages that I'm able to fix using our web interface
[10:09] <carlos> and this increases the urgency to fix it
[10:09] <carlos> I updated the bug with my debugging conclusions
[10:10] <carlos> and now I need to leave
[10:10] <carlos> cheers
[10:13] <kiko> mmm
[10:13] <kiko> ok
[11:02] <dilys> Merge to devel/launchpad/: [r=spiv]  baz2bzr, importd produces bzr branches via baz (r1833: David Allouche)
[11:03] <kiko> ddaa?
[11:36] <ddaa> kiko: ?
[11:37] <ddaa> just came back
[11:38] <ddaa> kiko: if you are wondering about that merge, it's a branch that has been pending merge for about two months
[11:39] <ddaa> finally decided to work around the last problem that prevented merging, otherwise it was blocked by bzrtools/smallfixes, which is blocked by a bunch of sourcecode fixes from spiv
[11:44] <kiko> ddaa, what is it, fundamentally? 
[11:44] <kiko> oh, this code has already been rolled out, is that it?
[11:45] <ddaa> been out and running since importd has been producing bzr branches
[11:45] <kiko> ah
[11:45] <ddaa> this branch provides exactly that feature
[11:45] <kiko> okay
[11:46] <ddaa> I had an completely insane branch backlog which is quickly resorbing
[11:46] <kiko> cool
[11:46] <kiko> cool
[11:46] <ddaa> echo?
[11:46] <ddaa> echo?
[11:46] <kiko> vf
[12:00] <bradb> kiko: that patch got a little more complex, so I put it up for review
[12:02] <kiko> that's okay
[12:02] <kiko> thanks though