=== stub [~stub@dsl-] has joined #launchpad
=== BradB [~bradb@modemcable202.193-131-66.mc.videotron.ca] has joined #launchpad
stubsabdfl: Should the bugattachment table reference Message or BugMessage?05:35
=== stub thinks Message and it needs renaming
lifelessBradB: I don't know .. should I ?11:36
sabdflstub: good question12:47
sabdfli think Message, if the attachment came from a Message12:47
=== carlos [~carlos@69.Red-80-33-181.pooles.rima-tde.net] has joined #launchpad
=== kiko [~kiko@200-206-134-238.async.com.br] has joined #launchpad
=== kiko [~kiko@200-206-134-238.async.com.br] has joined #launchpad
lifelesshey ki01:51
sabdflhey kiko01:53
kikohello sabdfl 02:04
kikohey lifeless 02:04
kikolifeless, did you get a minute to sort out salgado's key?02:04
lifelesskiko: I haven't seen the email.02:11
lifelesslet me check again.02:11
lifelessah I think I have it.02:12
lifelessguilherme ?02:12
lifelessyeah, I have it. will be done in ~ 10.02:12
kikoI may be out of touch, but why is Soyuz leading me to notready.ubuntu.com?02:15
lifelessdogfood.canonical.com is where you want to be02:16
lifelesslaunchpad is now the production install.02:16
kikoand the old addresses are dead, then.02:16
kikoI see02:16
kikolifeless, they run the same database, I assume?02:16
kikosabdfl, we have some "views" we'd like implemented in malone -- a way to limit a list of bugs per distribution, and per distribution release02:17
kikoI am wondering today if there is enough information to tie all this together, however02:17
lifelesskiko: salgados key is registered, he can register his archive and merge now.02:18
sabdflkiko: i'd rather not address that till we've solved some other problems02:18
kikosabdfl, okay, because that's one of the pending soyuz items we have02:18
lifelesssorry it took so long02:18
kikono problem02:18
sabdflwe've got solid answers to how bugs relate to package, and package release, but not yet the higher level distrorelease and distro02:18
sabdflthis is stuff we can hammer out in spain02:19
kikookay, let me note that02:19
sabdflright now the goal is to clean up what we have so we can show it to people in spain02:19
kikosabdfl, does this mean hiding what is unimplemented?02:19
kikoI was meaning to get a few features landed this week02:19
sabdflalso, i am find a a LOT of low-hanging fruit02:19
sabdflin terms of links between pieces of the system that ARE working02:19
lifelesssabdfl just wanted to hear the ping02:19
sabdflso i'm starting to add portlets to project and product to show related bits elsewhere in the system02:20
kikowell, I have the current tasklist open for this week:02:20
kiko- getting pending finally displayed in the interface (required DB schema work, kinnison/cprov in charge)02:20
kiko- Karma infrastructure (salgado)02:20
kiko- SQL DistroAnnotations (sabdfl to approve, debonzi)02:21
kiko- PackagesSubcription (sabdfl to define)02:21
kiko- activate Librarian links (debonzi)02:21
kiko- Links to distro and distro release bugs (debonzi)02:21
kiko- Report a bug on displayed source package (sabdfl to advise, debonzi)02:22
sabdfldistro annotations? in general, don't work on annotations stuff, it's fluff for the moment02:22
kiko- Fix edit distro information02:22
kiko- Fix permissions on edit distro (cprov)02:22
kikoSo there's a lot that is blocked by SEP02:23
kiko(someone else's problem)02:23
sabdflkiko: i'm sure there is a lot that could be done in terms of cleanups02:23
sabdflbut my most important goals are:02:23
kikosabdfl, code cleanups?02:23
sabdflkinnison / cprov: get the package workflow from uploaded, through pending, to installed, and published in the archive working02:24
sabdflsalgado: make sure that we have the general karma framework up, with specific karma for one (even tiny) part of the system02:24
sabdfldebonzi: make sure soyuz is using the same layout and standards as malone02:25
sabdflzcml and naming and file structure02:25
kikoby cleanups and layout and standards you mean code, not UI. 02:25
kikoI see.02:25
kikowe like to deliver more than is expected!02:25
sabdfli'm nervous about things like canonical.soyuz.browser02:25
sabdfl(from memory)02:26
kikomost of that is going away in some debonzi-patches02:26
sabdflright now, i'm starting to beat up on rosetta for that stuff02:26
sabdflRosettaProjectSet, RosettaProduct etc MUST GO (TM)02:26
kikowe've gotten rid of those for us at least02:27
sabdfli will take a wander through soyuz this afternoon, ok>02:27
kikoI need to sit down with daf and carlos to sort out some rosetta infrastructure for us, but I'll do that in mataro02:27
kikoit's difficult to answer questions like "What is the translation status for my distribution", "What is the translation status for release Hoary", "How was the translation status for release Warty (in the past)"02:28
kikoI don't think rosetta has a concept of history, which limits how useful it is to soyuz right now, AIUI02:28
kiko(I could be analyzing this all wrong, but I'd like to see someone else do better <wink>)02:29
carloskiko: I think it's better if you send us your needs at the moment so we can look at them before Matar and prepare the "interview"02:29
kikocarlos, look three lines up, the questions we want to answer are there.02:29
kikothat is a good summary.02:29
kikowe would also appreciate having the "translation status" made into a good framework of statistics to display02:30
kikoI know you are involved in the mechanics of things, but we want at least some reporting02:30
carloswe have them for projects/products02:30
sabdfli must say thngs are starting to come into focus02:31
carlosI suppose it's a matter to do an agregation function for distributions02:31
kikocarlos, we need *history*, which is an issue.02:31
sabdflit's *starting* to be possible to move through the Launchpad efficiently02:31
kikoproducts/projects are timeless02:31
carloskiko: we don't have history yet02:31
carloskiko: mark asked us to do it later02:31
kikoyes, I understand, but i just want to point out that that is a limiting factor in soyuz-rosetta integration.02:32
carloskiko: why?02:32
kikoI'm not meaning to sound anxious, just state the [perhaps obvious]  fact, to be addressed whenever02:32
carlosyou can get the current status02:32
carlosthat's all02:32
kikocarlos, because those questions can't be answered without history02:32
kikobut soyuz has a moving concept of current02:33
sabdflfocus on the present02:33
kikothe distro release hoary is today current, tomorrow old02:33
carloskiko: dude, that feature is not something critical for soyuz02:33
carlosit's critical for Rosetta :-)02:33
sabdflthe linkages from soyuz take you to where you CAN TRANSLATE02:33
carloskiko: but that will be still valid02:33
kikoI agree, I agree02:33
sabdflproviding a transaltion for a prior release may still be helpful and useful02:33
kikodon't shoot me down for just pointing out the obvious :)02:33
kikobut when we do a roundup of the pages and missing bits, that's one of the things that show up, which is why I bring it up.02:34
kikoI understand of course prioritization02:34
kikoI just want acknowledgement that it's "to be dealt with" to avoid people frowning because it was supposed to have been solved already!02:35
=== daf [daf@muse.19inch.net] has joined #launchpad
sabdflgood point02:37
sabdflmaybe i should prepare a shopping list for next week, for each of the teams02:37
sabdflhey daf02:37
kikoenjoy it02:37
sabdfltaking lifeless out to lunch to talk bazaar02:38
sabdflbaz is going to ROCK02:38
lifelessthat is, a bazaar lunch02:38
sabdflalready, it's much better02:38
kikoit's killer indeed02:39
kikoworks perfect on our diskless boxes02:39
sabdfldaf, carlos, that pqm merge is on its way02:41
carlossabdfl: ok, will review it after lunch. Thanks02:41
=== dilys [daf@muse.19inch.net] has joined #launchpad
dilysNew Malone bug #93: "Filing a new bug takes a *long* time", submitted by Christian Robotton Reis02:55
dilysNew Malone bug #94: "No confirmation of bug ID or success when a bug is filed", submitted by Christian Robotton Reis02:58
dilysNew Malone bug #95: "Could we have an automatic-ish duplicate-matching system?", submitted by Christian Robotton Reis03:00
dilysNew Malone bug #96: "Activate librarian links for packages", submitted by Christian Robotton Reis03:03
dilysNew Malone bug #97: "Bug count is bogus?", submitted by Christian Robotton Reis03:07
dilysNew Malone bug #98: "Fix edit distro permissions", submitted by Christian Robotton Reis03:08
dilysNew Malone bug #99: "double-spacing in bug comments?", submitted by Christian Robotton Reis03:13
dilysNew Malone bug #100: "Assignee should be settable from new bug", submitted by Christian Robotton Reis03:14
dilysMerge to rocketfuel@canonical.com/launchpad--devel--0: rosetta and sourcesource cleanups, and project portlets (patch-886)03:37
dilysMerge to rocketfuel@canonical.com/launchpad--production--1.5: snarf devel (patch-3)03:37
dilysMerge to rocketfuel@canonical.com/launchpad--devel--0: fix syntax error in broker.py (patch-885)03:37
dilysMerge to rocketfuel@canonical.com/launchpad--production--1.5: merge from devel (patch-2)03:37
dilysMerge to rocketfuel@canonical.com/launchpad--devel--0: improve sourcesource review process (patch-884)03:37
dilysMerge to rocketfuel@canonical.com/buildbot--devel--0: ensure generated data is not unicode (patch-79)03:37
dilysBug 2014 resolved: Create a findByPeople() method on SourcePackages and BinaryPackages03:44
dilysBug 1998 resolved: Missing package maintainer information.03:44
carlosdilys is going to ask a higher salary, she's working too much..03:52
sabdfldilys ROCKS03:57
sabdflcarlos: is RosettaProjectSet definitely needed, or could it be merged to ProjectSet?04:15
carloswhere is it?04:16
carlosooh, domain.py...04:17
carlossabdfl: ask Steve04:17
carlosit was a sample code for the domain thing 04:17
carlosthat's what it's still there04:17
sabdflso do you guys use it at all?04:18
carlossabdfl: I think so04:18
carlosbut it's not really needed04:18
carlosI mean, it could be merged04:18
sabdfli'm going to try to get rid of it now04:18
carlosI think the functionality is already merged, it's just a matter of doing some .zcml changes04:19
dilysMerge to rocketfuel@canonical.com/launchpad--devel--0: rosetta zcml cleanups (patch-887)04:27
carlosdaf: ping?04:45
carlossabdfl: should we move from canonical/rosetta/browser.py to canonical/launchpad/browser/ ?05:23
sabdflcarlos: yes, but let me finish my walkthrough first05:23
sabdflyou can work on that next week05:23
carlosok, I'm adding new pages and saw the new layout05:24
carlosthat's why I'm asking05:24
sabdflcarlos: yes, for new pages, please put them in the new zcml files, definitely05:24
carlosnew zcml files and new browser location?05:25
carlosor only zcml files05:25
dilysMerge to rocketfuel@canonical.com/launchpad--devel--0: banish RosettaProjectSet (patch-888)05:29
=== robertc [~robertc@host217-37-231-28.in-addr.btopenworld.com] has joined #launchpad
carlossabdfl: we need a place to add/edit Licenses05:48
carlossabdfl: where do you think it's the best place to do it?05:48
carloswe should stop using facked values...05:49
sabdflcarlos: add a LicenseSet and Licence class, traverse to them from rosetta for the moment05:49
sabdflwe will probably add that to DOAP05:49
sabdflbut i don't think that's critical to get done before mataro05:50
carlosok, so I implement it into Rosetta and it will be moved later.05:50
carlosso just use License.id = 1 and that's it05:50
carlossabdfl: how could I restrict the access to a concrete page inside a browser:pages directive?06:25
carlosI have this:06:25
carlos<browser:pages for="canonical.launchpad.interfaces.IProduct"06:25
carlos        permission="zope.Public"06:25
carlos        class="canonical.rosetta.browser.ViewProduct"06:25
carlos        lp:url="Rosetta/projects/$Project.name/$Product.name">06:25
carlos        <browser:page name="+rosetta-index"06:25
carlos            template="../templates/rosetta-product-index.pt"06:25
carlos            lp:url="Rosetta/projects/$Project.name/$Product.name/index" />06:25
carlos        <browser:page name="+newpotemplate"06:25
carlos            template="../templates/potemplate-new.pt"06:25
carlos            lp:url="Rosetta/projects/$Project.name/+newpotemplate" />06:25
carlos    </browser:pages>06:25
carlosand I want that <browser:page name="+newpotemplate"06:25
carlosis only available to authenticated users06:25
carlossorry: /s/is/be/06:31
sabdflcarlos: restrict access, you mean have a specific permission on that page?06:47
sabdflso you want to have a pages directive, then inside that 10 different pages, each with a custom permission?06:48
sabdfli also had that problem06:48
sabdflit makes sense to me you could say "pages" and specify the view class, then have a bunch of "page" elements whcih specify permission and template06:49
sabdflbut i don't think zcml allows that just yet06:49
carlosyea, that06:50
sabdflstevea can probably knock it up in a few minutes though :-)06:50
carloswill ask for alternatives to Steve06:50
carlosis there?06:50
sabdflright now you need a different pages directive, which sets the permission06:50
sabdfland each page inside that must use the permission, and class of the pages directive06:51
carloswill try it, thanks06:52
carloshmmm I need then also two different classes06:55
dilysMerge to rocketfuel@canonical.com/launchpad--devel--0: add advanced project searching (patch-889)07:02
=== BradB [~bradb@modemcable202.193-131-66.mc.videotron.ca] has joined #launchpad
sabdflhave had lots of fun today07:12
sabdflin the good sense07:12
dilysMerge to rocketfuel@canonical.com/launchpad--production--1.5: merge from devel (patch-4)07:12
BradBsabdfl: heh07:15
BradBWhat did you think of the search UI changes? Way off the mark, or an improvement?07:16
BradBThey're kind of in a halfway house, to be honest. 1. I was hinting at the fact that we should aim to make our widgets less ugly in our various forms (i.e. "normal" HTML submit buttons and input widgets are so unappealing) and 2. I'm halfway to hiding the rest of the "filter criteria", by default.07:17
sabdflhow do you make the button so slick?07:18
BradBsabdfl: CSS. The HTML is messy at the moment (it was a quick prototype to see if you think it looks good, next up is create a page test for that accepted functionality and then be able to clean it up a bit)07:20
sabdflbut dude, are you addicted to this page?07:20
sabdfli'm free-ranging on the launchpad finding tons of low hanging fruit that pulls it all together nicely07:20
sabdflthere are lots of separate pieces of functionality that can be interlinked now07:21
BradBsabdfl: I have to be, because it's 80% of what Malone is.07:21
sabdflno it isn't07:21
sabdflyou're still thinking of malone in a traditional bug collector sense07:21
BradBsabdfl: E.g. the next thing that /really/ needs help is making it simple to go from that list to editing assignments.07:21
sabdfland yes, we have to get that stuff right07:21
sabdflbut we have some time to do it07:21
BradBsabdfl: The quick search mods I made were just an attempt at making life easier for the user. The bug id/text search was a need-to-have though. I go nuts trying to report bugs on dogfood without being able to usefully search to see if they've already been reported.07:22
sabdflyes, that was definitely needed07:23
BradBThe one-click mods collectively save a load of time for Malone users too.07:23
sabdflwe need a google-style approach, and you are on your way there07:23
sabdflcheck the new project search stuff i did today for the style i prefer though07:24
sabdflat least, when i can merge it07:24
=== BradB has to get a wireless NIC today...gotta be prepped for .es
elmoBradB: don't let it stop you, but lamont always brings enough spares for the team +their family07:26
BradBelmo: That work on PPC?07:26
elmoyes, same card as I have07:26
elmothat's why I bought it - I knew from borrowing LaMont's that it works :)07:27
BradBWhat was the exact model again?07:27
elmoCisco Aironet 350 Series07:28
BradBnoted, thanks07:29
carlosDo we have an easy way to check if a form has the required fields filed?07:29
BradBsabdfl: got a moment to brainstorm on making assignment editing not suck?07:29
carlosor should we do all checks before using them always?07:29
elmohey, sweet, hotplug got fixed07:30
BradBelmo: does that mean i can plug in a NIC and have it Just Work?07:30
elmooh, you could always do that - unplug use to be broken tho, at least for this card/my machine, so you couldn't plug, [...] , unplug it and then plug it back in plug07:31
elmobut I just pulled it out to get the model off the back of the card for you, and plugged it back in afterwards07:31
elmo[and it worked] 07:31
BradBI've noticed that playing music sounds like arse on my machine, more or less.07:32
elmoarse how?  too fast?07:33
elmoif so, try using direct output rather than alsa07:33
BradBSometimes it seems a bit fast (maybe I'm imagining it), but it definitely seems to bleep and skip.07:33
elmoyeah, try !alsa07:34
elmoworks fine for me with xmms07:34
dilysBug 2165 resolved: How do I download a binary package?07:36
kikoBradB, yo?07:37
dilysBug 2046 resolved: Implement paging for package pages07:38
=== kiko pokes BradB
kikois it possible/easy to file a bug for a certain package via a URL?07:41
kikoif you could hook that up for us we'd get automagic bug-filing links from soyuz07:41
BradBkiko: Yeah, sabdfl's already done something like that for Malone.07:41
BradBkiko: Just look at the request that gets sent on the add form.07:41
kikoBradB, can you detail hmmmm 07:42
kikobut that would require creating a custom view/template?07:42
BradBkiko: heck no, that's too much work.07:42
kikothen you mean a form post? :)07:42
BradBer, well, depending on what you mean. hang on, i'm search for sabdfl's stuff.07:42
=== kiko is guessing games
sabdflgrep for filebug in zcml/07:42
sabdflhey kiko07:43
kikohey sabdfl 07:43
sabdflthere is a custom View class07:43
kikosabdfl, could I have one for a source package?07:43
kikois it indeed?07:44
sabdflhmm... if your stuff traverses to a SPECIFIC source package then actually, not tricky07:44
BradBkiko: just use the add form dude.07:44
dilysMerge to rocketfuel@canonical.com/launchpad--devel--0: break cross merge deadlock (patch-890)07:44
kikoyes, a specific source package --- I'm in soyuz!07:44
dilysMalone bug #61 fixed for product The Rosetta Translation Portal: Create a new table to store temporary po/pot files to be imported07:44
BradBkiko: and...? :)07:44
kikowell, I'd rather use a link than a button, but..07:44
sabdflkiko: let me do it now it will give me a chance to poke around in soyuz a bit more07:45
carlosBradB: the bug edit form changes after you submit it :-?07:45
BradBkiko: You're looking for a URL to use to add a bug to a source package or a product. That URL is http://localhost:8086/malone/bugs/+new with the appropriate params.07:45
kikosabdfl, cool.07:45
sabdflBradB: no, it so is NOT :-)07:45
kikoBradB, that means using a <submit> button :)07:45
BradBsabdfl: why not?07:46
sabdflBradB: you are still thinking in terms of malone as a monlithic bug system07:46
BradBsabdfl: No, I'm thinking in terms of not duplicating effort.07:46
sabdfltry to imagine that the launchpad will actually be highly fragmented07:46
sabdfllots of small communities, all using it, all just seeing their bugs, their packages07:46
sabdflthe fact that its all in one system gives us great opportunities for collaboration07:47
sabdflbut where possible we still want people just to see information that is really relevant to them07:47
BradBsabdfl: I've already imagined that. :) Why is it good to solve the same problem more than once here?07:47
sabdfllaunchpad.dia.org should in general only see bugs related to dia07:47
sabdflbecause kiko wants to add a bug ON A SPECIFIC PACKAGE07:47
BradBsabdfl: that can be done with the link i provided.07:48
sabdflso, he's imagining a scenario where the guy is already looking at the package, and all he wants is to add a bug to that07:48
BradBsabdfl: he'd have to provide the appropriate params to that link, but that's a given.07:48
sabdflyour method would take him to the MASTER BUG ADD PAGE with a zillion knobs and buttoms07:48
sabdflmost of which are unrelated to HIM07:48
BradBsabdfl: no.07:48
kikothe issue is passing parameters via HTTP GET 07:48
BradBsabdfl: he could do what i've already done and submit the URL with .js.07:48
kikoor via a JS POST07:49
sabdflhttp://localhost:8086/malone/bugs/+new doesn't know anything about sourcepackage07:49
sabdflunless you add paramteres07:49
kikoboth of which I think are limited..07:49
sabdflmuch better to use the context, luke07:49
BradBsabdfl: that's /precisely/ what i'm saying :)07:49
BradBsabdfl: the "context" is something the caller is smart enough to pass in to us.07:49
BradBsabdfl: Remember, you already went off and solved this problem again, which ended up breaking things, because you rewrote code for which I've already written code that solves the problem, and in your rewrite left out important things that BugFactory already does (e.g. auto-subscribing the submitter)07:50
BradBsabdfl: duplicating work is error prone and costs more money.07:50
BradBIf one /really/ /really/ insists on creating their own form for this, then it's not that big of a deal, they just have to be certain to use BugFactory as their content factory.07:51
sabdflparameters are suckier than context07:52
sabdflbut the end result is that they guy's browser suddenly shoots off to some master page, which has a ton of knobs on it07:52
sabdfland yes, maybe your parameters pre-set the knobs in a ncie way for him, but there are still a daunting number of knobs07:52
sabdfli'm going to keep teasing you about this till you get it ;-)07:52
kiko<sabdfl> parameters are suckier than context07:52
kikothat is soooo true07:52
sabdflit doesn't require duplicating work07:53
kikoit should be in a style guide somewhere07:53
BradBI get it. I've already provided you an example of why my suggestion works better than yours too, with the example of what happened when you went and reinvented this. :)07:53
kikosabdfl, indeed, if done correctly <wink>07:53
sabdfli have refactored my +filebug to use your BugFactory ;-)07:53
sabdflBradB: what do you mean, went and reinvented this?07:53
BradBsabdfl: i.e. when you didn't use BugFactory :)07:53
sabdflthe solution was simply to use bugfactory from my specialised addform07:54
sabdflthat's no problem07:54
BradBsabdfl: as i said above though, if people /really/ insist on creating a separate addform for this, fine, not that big of a deal, as long as they use BugFactory as their content factory.07:54
sabdflbut the user experience should not be to be redirected to one central master page07:54
BradBIOW, I think we agree.07:54
BradBsabdfl: I'm not as insistent on the URL as on the content factory.07:54
sabdfldude, i'm not going to be happy till you drop the /really/ attitude07:54
sabdflthis is important07:54
sabdflwe need to think of lots of *little* projects using malone, rosetta, the bounty system, support tracker07:55
sabdflthey should each see it as their system07:55
sabdflthey don't want to collide with a zillion other knobs, and data from a zillion other projects07:55
BradBOkay, not /really/ then. It makes perfect sense for people to create their own fancy context-aware bug add forms as long as they use BugFactory (which is why I filed a bug when somebody went off and didn't use BugFactory.) Creating a form is otherwise reasonable, though ideally our schemas would be setup to properly inherit from one another, so we don't have duplicate schema defs, etc.07:56
sabdflthis url would be better as launchpad.canonical.com/rosetta/bugs07:57
sabdflthe schema recreation thing is what i was bitching about to stub the other day07:57
sabdflcreating a BugAddSchema is crazy07:57
BradBkiko's original question asked me for a url though, hence why i told him about the existing all-purpose add form that one could trivial call behind-the-scenes. but i can happily live with new forms that reuse existing code.07:58
sabdfli then have to copy the title and description and required fields from the main schema into the AddFormSchema, and remember to update them07:58
sabdflthat's why i want to be able to create a form, and specify fields that come from different schemas08:00
BradBsabdfl: Reuse is a general problem we're seeing in Malone. e.g. think about it, if we decide to use icons or do some other fancy things to the bug listing to make it real-estate sensitive, we're going to have to then go in and modify the places where this was reinvented in the pages malone has for source packages and products.08:01
sabdflbtw, lifeless is digging through sqlobject and could use your insight08:01
BradBsabdfl: so when i say that the bug listing is 80% of malone, that's what i mean. i don't necessarily mean the link at "See the complete bug list" is 80% of Malone, but rather that a cool, easy-to-grok-and-get-to-where-i-want bug listing is an /integral/ part of Malone.08:02
sabdflwe really DO need the big form with all the master knobs08:03
sabdfli totally get that08:03
sabdflbut it's not a huge priority *right now*08:03
sabdfli want to be able to show off malone for a small upstream project in mataro08:03
sabdfland right now, there are lots of places it could do with polish08:04
sabdflthat have nothing to do with the big master form08:04
BradBsabdfl: my current focus is on making editing assignments not suck. i want to reject bugs, and i can't even explain why i'm rejecting them.08:04
dilysMerge to rocketfuel@canonical.com/launchpad--devel--0: advanced project searching and selection (patch-891)08:04
sabdflthere's some really nice code behind the ProjectSet.search function now08:04
BradBand also, i want to get to the assignment in one-click from the bug listing, not two.08:05
=== BradB does another star-merge
carlossabdfl: or there is a bug with arch -> mail gateway  or you forgot to add any change with your last commit08:05
robertcBradB: do you know the sqlobject & sqlos innards ?08:05
carlosthe mail report does not have any new/modified file08:06
BradBrobertc: Depends on what you want to know :)08:06
BradBrobertc: I don't know anything specific about SQLOS though.08:06
BradBer, sqlos08:06
robertcwe have a problem where new transactions are not getting the ConnectionAdapter registered with them, using initZopeless08:07
robertcthis leads to everything after the first commit() in a single thread failing.08:07
BradBThis sounds like neither an sqlobject, nor an sqlos problem.08:08
BradBrobertc: do you have a unit test demonstrating the failing behaviour? that would be a starting point as to figuring out what's going wrong (and, of course, making sure it doesn't come back.)08:09
robertcheh. I can write if I can guarantee a database etc etc.08:09
kikoskippin off for a bit08:10
robertcits really quite easy though :  initZopeless(), + select something + change + commit. then select + change + commit.08:10
BradBrobertc: Practicality beats purity. You can guarantee a database.08:10
kikosabdfl, let me know what comes out of it08:10
dilysMerge to rocketfuel@canonical.com/launchpad--devel--0: Added forms to create and edit a potemplate and reenabled the poexport test (patch-892)08:12
robertcSQLBase.begin() is wrong. it makes a new txn without joining it to the connection.08:16
robertci.e. it should be :08:16
robertc    def begin(self):08:17
robertc        txn = self.manager.begin()08:17
robertc        txn.join(self._dm())08:17
robertc(untested )08:17
robertcyep, thats it. FIXED.08:20
sabdflcarlos: that your merge? cool08:21
sabdflrobertc: you ROCK08:21
BradBrobertc: Just please make sure a test comes along proving that your fix works. :)08:21
carlossabdfl: yep, we are really close to be able to import files from the web without any manual process from daf or I08:22
robertcBradB: normally, I'd be the test-nazi on this, but frankly, after 4 hours on it, + 2 days of rob wier & a day of stevea, I'm over it.08:23
sabdflcarlos: excellent, thank you!08:32
sabdfli'm still refactoring the rosetta zcml08:32
dilysMerge to rocketfuel@canonical.com/launchpad--devel--0: Fix 'initzopeless second commit and thereafter' (patch-893)08:54
dilysMerge to rocketfuel@canonical.com/launchpad--devel--0: fixed Submit a Bug button to go to the bug add page (patch-894)09:19
=== lifeless [~robertc@dsl-] has joined #launchpad

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!