BradBSteveA: Sources seem pretty cool, but indeed it turns out that some things break with an upgrade (7 tests fail.) Do we battle through an upgrade, or reimplement sources in our Z3?12:04
sabdfllet's stay closer to upstream if possible12:05
BradBI'd like to too.12:06
BradBsabdfl: Should I proceed ahead with this locally upgraded Z3, and hold off committing until my changes are ready and jblack/lifeless get the newest Z3 imported?12:14
BradBproceed ahead as in spend time fixing what's broken in Malone12:14
sabdflBradB: your call12:15
dilysMerge to rocketfuel@canonical.com/launchpad--devel--0: introduce buttress (patch-833)12:17
=== BradB takes a look to see if the brokenness is fairly shallow
BradBSteveA: ping12:26
carlosBradB: he went to sleep about one hour ago12:32
BradBnow up to 41 failures and 2 errors. ugh.12:48
BradBdoes anything special have to be done to upgrade the zope version? i just made sourcecode/zope a symlink to ~/Zope312:49
BradBseems like many of these errors are just not finding names.12:49
=== carlos [~carlos@69.Red-80-33-181.pooles.rima-tde.net] has joined #launchpad
=== stub [~stub@dsl-] has joined #launchpad
BradBstub: yo02:29
BradBstub: i'm trying to upgrade Z3 locally. to do that i simply moved sourcecode/zope aside and made a symlink to my ~/Zope3 checkout.02:30
BradBbut now when i visit a URL, e.g. http://localhost:8086/, I get a Error type: zope.publisher.interfaces.NotFound02:30
BradBNotFound: Object: <canonical.publication.RootObject object at 0x30d0c090>, name: u'index.html'02:31
BradBany obvious alarms going off as to why that may be?02:31
BradB(zope starts up without incident, and if i put an error in the directive that defines that page, zope craps out, so i know it's still loading that confi)02:32
BradBconfig, even02:32
dilysMerge to rocketfuel@canonical.com/launchpad--devel--0: Rosetta project view improvements (patch-834)02:32
BradBI'm thinking it's *possible* that the publishing internals have changed, and a Z3 upgrade b0rks our publisher.02:36
BradB(I'm wanting to upgrade so that I can use ISource et al.)02:36
stubI think for you to get that error message, lib/canonical/configure.zcml is not being loaded02:37
BradBdude, it is02:38
stubTry introducing an obvious XML error?02:38
BradBi mentioned that already02:38
=== BradB decides to debug through queryView
stubAnd root index contains no TAL, so it isn't something else raising a NotFound02:39
stubI'm not sure if we still have local patches against our Z3 tree. You would need to ask SteveA or lifeless02:40
stubYou could be right about our publisher - if I remember it tries to lookup a few things regarding our custom suburl directives and whatnot before falling back to a normal view, and the KeyError or AttributeError could be caused in there. Need pdb to trace it.02:43
stubI think we should ask lifeless to get the Z3 SVN repository mirrored at arch.ubuntu.com - we could then trivially upgrade since we could specify any revision we like.02:45
BradBAs sabdfl mentioned earlier, I think it's definitely a good idea for us to try to keep as close to upstream as possible, perhaps particularly in this still relatively early stage. The last thing we want is to end up wedged in on a non-release version of Zope 3. So yeah, the easier it is to upgrade, the better.02:47
=== stub [~stub@dsl-] has joined #launchpad
=== netjoined: irc.freenode.net -> tolkien.freenode.net
=== sabdfl [~mark@host217-37-231-28.in-addr.btopenworld.com] has joined #launchpad
=== daf [daf@muse.19inch.net] has joined #launchpad
=== daf [daf@muse.19inch.net] has joined #launchpad
!alindeman:*! One of our main rotation servers just split from the network there. We're working to mitigate further effects. Thanks for using freenode and sorry for any inconvenience.05:20
SteveABradB: So, you can't upgrade Zope3 at the moment, because some stuff breaks when you do.08:29
SteveAstub: definitely a good idea to get the Zope 3 SVN repo mirrored.08:32
SteveAstub, BradB: we have no local patches against the Zope 3 SVN tree.08:32
SteveAit sounds to me like a problem with layers and skins.  I can take a look at this a bit later on today.08:34
SteveABradB: probably not worth you spending much time on it.  I can imagine it would be easy to lose hours looking at it.08:34
sabdflmorning all09:33
sabdflstub: i took some liberties in grabbing 4-14, is that ok?09:33
stubBad Mark, but I'll retroactively approve it since we had already discussed this change ;)09:56
=== lulu [~lu@host217-37-231-28.in-addr.btopenworld.com] has joined #launchpad
=== sabdfl [~mark@host217-37-231-28.in-addr.btopenworld.com] has left #launchpad []
lulukinnison:morning :o)10:38
Kinnisonlulu: I'm happier this morning10:40
lulukinnison: glad to hear it mate :o)10:40
dilysMerge to rocketfuel@canonical.com/launchpad--devel--0: show buttress in launchpad (patch-835)10:41
lifelesscan someone give bob2 a hand debugging why launchpad doesn't let him add projects ?10:52
stubbob2: I can have a look, but I'm not familiar with that code. Whats up?11:02
=== salgado [~salgado@200-206-134-238.async.com.br] has joined #launchpad
bob2stub: I don't seem to have some of the needed tables, like project and person.11:08
bob2(after a "make" in the db schema dir)11:08
Kinnisonbob2: You're missing the ability to create plpythonu or plpgsql in your db11:09
Kinnisonbob2: either make yourself a postgres superuser or add them to template111:09
bob2which is easier?11:09
Kinnisonyour own box?11:09
Kinnisonthe former11:09
Kinnisonsudo -u postgres psql template111:10
Kinnisonalter user bob2 with createdb createuser;11:10
bob2hrm, did that, then re-made the db, and I seem to be getting a similar error11:15
bob2hwo can I make sure I'm a superuser?11:19
stubpsql -d template111:20
stubselect * from pg_user;11:20
bob2 usename  | usesysid | usecreatedb | usesuper | usecatupd |  passwd  | valuntil | useconfig 11:20
bob2 rob      |      100 | t           | t        | t         | ******** |          |11:20
stubSo that isn't the problem11:21
stubpsql -d launchpad_dev11:22
stub >   \d person11:22
bob2template1=# \d person11:22
bob2Did not find any relation named "person".11:22
stub\i trusted.sql11:23
stub(after running psql -d launchad_dev in the database/schema directory)11:23
stubAny errors? (ignore the SET: lines)11:24
bob2erk, it says the db doesn't exist11:26
bob2shouldn't the schema Makefile make it?11:27
stubErm... launchpad_dev, not launchad_dev.11:27
bob2hah, right11:27
bob2no errors11:27
stubSo it isn't a plpythonu problem11:28
stubselect valid_name('hello');11:28
stub(still in psql)11:28
bob2 valid_name 11:29
bob2 t11:29
stubYou will need to email me the output of 'make' in database/schema11:30
stubOr one of those paste site thingies11:32
=== Kinnison hurrahs an initial success on the old multiarch gina import front
=== Kinnison commits
Kinnisonstub: next pqm merge from me contains another db patch11:37
stubbob2: According to that output, everything worked just fine. No errors in the create tables, the sample data loaded just fine, and the comments were added.11:43
dilysMerge to rocketfuel@canonical.com/launchpad--devel--0: Multiarch gina (stage1) and a db patch for binarypackage constraints (patch-836)11:47
Kinnisonstub: if you fancy an easy task; the patch in that merge is a two line constraint change ;-)11:49
stubyup. In a tick11:50
Kinnisonbob2: is sabdfl around?11:50
bob2Kinnison: he's out atm11:51
stubbob2: psql -d launchpad_dev11:51
=== kiko [~kiko@200-206-134-238.async.com.br] has joined #launchpad
Kinnisonbob2: okay; ta11:51
bob2stub: can you reproduce it on rocketfuel@canonical.com/launchpad--devel--0--patch-83411:51
stubbob2: Then 'select current_database();'11:51
bob2 current_database 11:51
bob2 launchpad_dev11:51
stub\d person11:52
bob2lots of stuff11:52
stubOk - it is there now. 11:52
stubPerhaps the first time you ran make, another process was still connected to launchpad_dev so the final stage of the makefile failed.11:53
bob2the doap and malone stuff still throws exceptions11:55
bob2"make" in the schema dir then "make build run" in the root should be enough, right?11:55
stubWhat exceptions are you seeing?11:59
stubbob2: Table not found ones?11:59
bob2  File "/home/rob/projects/warthogs/launchpad/launchpad/lib/canonical/publication.py", line 88, in publishTraverse12:00
bob2    raise NotFound(self.context, name)12:00
bob2exception.txt on chinstrap12:00
stubbob2: Ignore that one - it is just saying some browser requested favicon.ico and it isn't there.12:02
stub(Just a 404 error)12:02
bob2er, wrong one, sorry12:02
bob2    *  Module zope.publisher.publish, line 132, in publish12:03
bob2      publication.beforeTraversal(request)12:03
bob2    * Module canonical.publication, line 238, in beforeTraversal12:03
bob2      p = prin_reg.authenticate(request)12:03
bob2    * Module canonical.launchpad.webapp.authentication, line 91, in authenticate12:03
bob2      return self._authenticateUsingBasicAuth(credentials)12:03
bob2    * Module canonical.launchpad.webapp.authentication, line 65, in _authenticateUsingBasicAuth12:03
bob2      principal = login_src.getPrincipalByLogin(login)12:03
bob2    * Module canonical.launchpad.webapp.authentication, line 184, in getPrincipalByLogin12:03
bob2      person = getUtility(IPersonSet).getByEmail(login)12:03
bob2    * Module canonical.launchpad.database.person, line 132, in getByEmail12:03
bob2      return results[0] .person12:03
bob2    * Module sqlobject.main, line 1194, in __getitem__12:03
bob2      return list(self.clone(start=start, end=start+1))[0] 12:03
bob2    * Module sqlobject.main, line 1198, in __iter__12:03
bob2      return conn.iterSelect(self)12:03
bob2    * Module sqlobject.dbconnection, line 507, in iterSelect12:03
bob2      select, keepConnection=True))12:03
bob2(when hitting http://localhost:8086/doap/projects/+new)12:03
stubYou are missing the actual exception (that is just the traceback). That page works for me with that revision from rocketfuel.12:08
bob2TypeError: iteration over non-sequence12:08
bob2erk, sorry12:08
stubAt a guess, there is no entry in the database for the email address you are attempting to log in as (you need to exist in the sampledata), and that case isn't being handled correctly.12:08
bob2I wasn't logged in12:09
stubpsql -d launchpad_dev,   'select * from emailaddress where email ='stuart.bishop@canonical.com' except with your email address12:09
bob2foo.bar@canonical.com/test, right?12:09
stubThink so. The email address is right anyway12:11
stubbaz switch rocks ;)12:12
stubbaz switch rocketfuel@canonical.com/launchpad--devel--0--patch-834 just worked ;)12:12
bob2hah, nice12:14
bob2ok, I get an error while logging in, too12:14
bob2    *  Module canonical.launchpad.database.person, line 132, in getByEmail12:14
bob2      return results[0] .person12:14
bob2    * Module sqlobject.main, line 1194, in __getitem__12:14
bob2      return list(self.clone(start=start, end=start+1))[0] 12:14
bob2    * Module sqlobject.main, line 1198, in __iter__12:14
bob2      return conn.iterSelect(self)12:14
bob2    * Module sqlobject.dbconnection, line 507, in iterSelect12:14
bob2      select, keepConnection=True))12:14
bob2TypeError: iteration over non-sequence12:14
stubThat vaguely reminds me of an exception I would expect to see if you had an outdated release of sqlobject12:15
bob2oh god12:16
=== bob2 prepares the knives
bob2the config is still broken12:17
lifelessstub: so did you roll out the db upgrade ?12:18
lifelesscould you please ?12:19
bob2    *  Module sqlobject.dbconnection, line 194, in _executeRetry12:21
bob2      return cursor.execute(query)12:21
bob2    * Module zope.app.rdb, line 258, in execute12:21
bob2      return self.cursor.execute(operation)12:21
bob2ProgrammingError: ERROR: relation reference "potemplate" cannot be used as a select-list entry HINT: Write "potemplate".* to denote all the columns of the relation. SELECT POTemplate.id, POTemplate.changeset, POTemplate.owner, POTemplate.description, POTemplate.copyright, POTemplate.title, POTemplate.priority, POTemplate.branch, POTemplate.product, POTemplate.messagecount, POTemplate.path, POTemplate.name, POTemplate.license, POTemplate.datecreated, POTemplate.iscur12:21
stubDo we have someone who can restart the authserver if it doesn't survive the db bouncing?12:34
=== carlos [~carlos@69.Red-80-33-181.pooles.rima-tde.net] has joined #launchpad
lifelessstub: dunno.12:38
SteveAGood question.  I think only andrew has restarted it before.  When it didn't survive a restart last time, andrew added code to make it survive the next time.12:38
SteveAI think elmo would be able to find out.12:38
dilysMerge to rocketfuel@canonical.com/buildbot--devel--0: set the product link in imported branches (patch-74)12:39
stubWe need to start setting up procedures for the production systems. And I thought I'd gotten out of all that sysadmin shit ;)12:40
SteveAout of the old shit, and into some that is shiny and new12:40
carlosstub: when do you sync dogfood server with latest launchpad code?12:43
stubIt depends. Last one was yesterday morning (about 34 hours ago).12:43
carloswe need to do small fixes to rosetta and a sync will be needed so the betatesting server is fixed 12:45
stublifeless: I assume you want that patch in now? I might just apply yours since I don't want to risk taking down the authserver.12:45
stubcarlos: I'll be doing an update shortly (after I've sorted lifeless and Kinnison)12:45
carlosstub: ok, could you ping me before that, I'm going to finish the needed bits before your sync12:46
lifelessstub: I'll be doing a full code drop12:46
lifelessso ...12:46
lifelessI'd much rather the lot.12:47
lifelesscan elmo bounce the authserver ?12:47
stubelmo: ping12:47
=== Kinnison -> town, brb
dilysNew Malone bug #64: "Nagios authserver plugin", submitted by Stuart Bishop12:53
stubBradB: Did you change the Product vocabulary back to use the id as the token rather than '$project.name $product.name' ?12:58
bob2stub: hrm, seems to have just been the wrong sqlobject; after a rebuild it works great now01:01
bob2stub: thanks for your help01:01
lifelesswhat does 'fogo na bomba' mean ?01:02
kikofire in the hole more or less01:03
carlosSteveA: how could I know from rosetta if they are visiting the site from launchpad.ubuntu.com or rosetta.shuttleworthfoundation.org/ ?01:04
lifelessso its not like, forget the bombs?'01:04
carlosSteveA: the PATH_INFO is the same01:04
lifelessstub: is that done ?01:07
dilysMerge to rocketfuel@canonical.com/launchpad--devel--0: update the arch model to allow setting the project link in Branch from imports (patch-837)01:08
stublifeless: No - waiting on elmo or someone who knows about the authserver to arrive01:17
SteveAcarlos: look in the request.URL01:17
carlosSteveA: request.URL? don't see it01:18
carlosCHANNEL_CREATION_TIME: 1101211740.29 GATEWAY_INTERFACE: CGI/1.1 HTTP_ACCEPT: text/xml,application/xml,application/xhtml+xml,text/html;q=0.9,text/plain;q=0.8,image/png,*/*;q=0.5 HTTP_ACCEPT_CHARSET: ISO-8859-1,utf-8;q=0.7,*;q=0.7 HTTP_ACCEPT_ENCODING: gzip,deflate HTTP_ACCEPT_LANGUAGE: es,en-gb;q=0.7,en;q=0.3 HTTP_HOST: localhost:9020 HTTP_MAX_FORWARDS: 10 HTTP_REFERER: https://launchpad.ubuntu.com/ HTTP_USER_AGENT: Mozilla/5.0 (X1101:18
carlos; U; Linux ppc; en-US; rv:1.7.5) Gecko/20041116 Epiphany/1.5.1 (Ubuntu) (Ubuntu package 1.0-2ubuntu3) HTTP_VIA: 1.1 launchpad.ubuntu.com HTTP_X_FORWARDED_FOR: HTTP_X_FORWARDED_HOST: launchpad.ubuntu.com HTTP_X_FORWARDED_SERVER: launchpad.ubuntu.com PATH_INFO: /++vh++https:launchpad.ubuntu.com:443/++/rosetta REMOTE_ADDR: REQUEST_METHOD: GET SCRIPT_NAME: SERVER_NAME: rosetta SERVER_PORT: 9020 SERVER_PROTOCOL: H01:18
carlosTTP/1.1 SERVER_SOFTWARE: zope.server.http (DebugLayerHTTP)01:18
carlosSteveA: I was thinking on use request.HTTP_X_FORWARDED_SERVER01:19
SteveAuse request.getURL() 01:21
carloslet me try..01:21
SteveAI guess you can use PATH_INFO too01:22
(elmo/#launchpad) stub: hmm?01:22
carlosSteveA: yes, I just noticied the difference between launchpad and rosetta servers01:23
(elmo/#launchpad) does the authserver still need restarted?01:24
stubelmo: No - just wanted to make sure someone who could restart it was around before I bounced postgres to do a db upgrade01:25
(elmo/#launchpad) oh, ok01:25
stubI'll take postgres@emperor down now01:26
(elmo/#launchpad) shout when I should apply the big hammer01:28
=== kiko is now known as kiko-afk
stublifeless: Done01:35
stubelmo: In theory, it is now working. It is a just in case thing because it is an important service and we havn't pulled the rug out from under its feet enough times to trust it yet ;)01:36
stubelmo: I don't see any connections...01:51
(elmo/#launchpad) ... restart it?01:52
(elmo/#launchpad) yeah, it's screwed01:53
(elmo/#launchpad) restarted01:53
dilysNew Malone bug #65: "Production authserver not surviving database outages", submitted by Stuart Bishop02:01
carlosstub: I just requested a merge of the needed fixes from Rosetta02:03
carloslunch time02:10
=== lulu [~lu@host217-37-231-28.in-addr.btopenworld.com] has joined #launchpad
dilysMerge to rocketfuel@canonical.com/launchpad--devel--0: Ugly hack to fix the rosetta betatesting template (patch-838)02:24
lifelessstub thanks02:31
stubcarlos: eh?02:33
carlosstub: from rosetta we need to apply some changes to dogfood code base to fix the links at https://rosetta.shuttleworthfoundation.org/02:38
carlosthe changes are now at rocketfuel and it's interesting to have them merged into dogfood server as soon as possible so our betatesting server is useful for our users02:39
stubAhh... this is the stuff you want rolled out to dogfood ;)02:49
=== sabdfl [~mark@host217-37-231-28.in-addr.btopenworld.com] has joined #launchpad
sabdflme refuels... love my daily code fix02:55
sabdflerm, /me02:55
carlosstub: yes ;-)02:56
=== kiko-afk is now known as kiko
kikosabdfl, SteveA, salgado was asking me about our plan for moving person-related code into FOAF03:03
kikodo you have some minutes to discuss this with me?03:03
sabdflsounds good03:03
kikowith us, more properly03:03
sabdflfire away03:03
kikocurrently the login service is inside webapp03:03
kikoand create new person is (AFAICT) limited to soyuz03:03
kikowe display person-related information in all applications03:03
kikoand check permissions in all applications03:04
kikowe also have application-implemented preferences interfaces 03:04
kikowhat would we like to see centralized in FOAF? note that this could mean FOAF as infrastructure, not just UI, if we like.03:04
kikosabdfl, do you have an idea of at least a set of pages we'd like in FOAF?03:08
kikoI was thinking of moving the Soyuz Person branch *entirely* into FOAF03:08
kikobut that needs some approval.03:08
sabdflkiko: can you give me a url to check out the SoyuzPerson stuff?03:08
kikothat could mean bringing down soyuz one level, in fact03:08
sabdfli've found the skins mechanism to be a*very* nice way to integrate foaf / doap with the other web apps03:09
sabdflfor example, I use skins to show all the bugs related to a project, and a product03:10
sabdflhere's how it works03:10
sabdflthe standard setup is for a ProjectSet to traverse to Project, which traverses to Product03:10
sabdfland each of these objects has a page called +index with a corresponding template03:10
sabdflthat template shows the object "neutrally"03:10
sabdflthen I create another page on the object, called +malone-index03:11
sabdfland I add a defaultView *for the Malone skin* of +malone-index03:11
sabdflthen, inside Malone, I just traverse to a ProjectSet03:11
sabdflbut now you don't see the neutral page, you see the +malone-index03:11
sabdfland that template renders the same object from a bugs perspective03:12
kikothat's awesome03:12
sabdfl(it actually shows a list of products and a summary of the new/accepted/rejected/fixed bugs in different severities)03:12
sabdflnow, i think foaf should work the same way03:12
sabdfldefine a core person object, with different defaultviews03:12
sabdflbut the basic traversal stuff would remain the same for a Person object03:12
sabdflthis allows you to get rid of DoapPerson and SoyuzPerson and MalonePerson03:13
sabdflyou just have a Person03:13
kikoso this must mean you agree to moving the person branch into foaf, and having soyuz just skin it?03:13
sabdflwhich you will display differently based on the skin03:13
BradBSteveA: Have you had a chance to look at upgrading Z3? I think it's something about our custom publisher that may now be b0rked with the upgrade.03:16
sabdflhiya BradB03:20
BradBsabdfl: we were going to walkthrough the dump this morning. when do you want to do that?03:21
sabdflnow is good03:21
BradBok, reloading it, one sec03:21
Kinnisondoes python have the ? :  operator?03:23
BradBKinnison: nope, it was specifically rejected after lengthy debate :)03:24
Kinnisonwhat's the python idiom for a?b:c then?03:24
salgadoKinnison, a and b or c03:25
Kinnisonsalgado: what if b evaluates to falsehood in a logical context?03:25
Kinnison(lua has this same issue)03:25
BradBKinnison: it's just if/else.03:25
KinnisonBradB: thanks dude03:26
salgadoKinnison, in this case this didn't work03:26
KinnisonI was tempted to try and add an infix structure to lua to do it. Of the form: "b if a else c" but that interfered horribly with statement parsing03:27
=== Kinnison will work it out eventually
dilysMerge to rocketfuel@canonical.com/launchpad--devel--0: New database schema baseline from production and binarypackage constraint fix for Kinnison (patch-839)03:27
Kinnisonthanks stub03:32
BradBsabdfl: ok, had to do a bit of tla undo'ing and linking back to our older version of Z3. ready to go now.03:34
sabdfli'm sort of in the same boat03:35
BradBsabdfl: you tried upgrading03:35
dilysMerge to rocketfuel@canonical.com/dists--devel--0: add production-4 config (patch-33)03:37
sabdfli turned add_missing_from on in my db and now it's crashing ;-)03:40
BradBoh, heh03:40
stubsabdfl: I'll give people plenty of warning before switching that to an error  :-)03:41
sabdflwell, debonzi's been working at in in Soyuz i thought i'd do the same for malone03:41
stubBradB: Did you update the Product vocabulary to use integers for tokens again? Or have I lost that work somewhere?03:41
kikosabdfl, so do you say yea, on with the person-branch-move?03:42
BradBstub: yeah, i think i did, because product names aren't unique03:42
sabdflok bradb let's start i'll fix these things as they occur03:42
kikosabdfl, soyuz/person -> foaf/person ?03:42
sabdflproduct names are unique within a project, projects are unique03:43
stubBradB: It was using project name + product name, which is unique. If that was the only problem, I'll switch it back03:43
BradBstub: it was? oh.03:43
stub(With a short circuit if the projectname == product name, just to be confusing ;)03:43
BradBstub: IDs are what everything else uses, why be different?03:43
sabdflkiko: yes go ahead03:44
kikosabdfl, you da man!03:44
sabdflkiko: no, YOU rock03:44
kikosalgado, we got signal03:44
salgadokiko, fogo na bomba03:44
stubBradB: Because it makes the UI sucky. When I put it back, you can just type 'launchpad' into the text box and it will work. Or you can type 'gnome' into the text box and you get a select when the page reloads. So you don't have to use the popups at all.03:44
BradBstub: eek :) the other selects already uses IDs though, as i say.03:45
sabdflBradB: i'd rather expose project/product than id03:45
sabdflBradB: let's do this walkthrough03:46
BradBi'm ready whenever you are :)03:46
stubSourcepackage is still under debate (it needs to be fixed - we just havn't worked out how), and the others are selects so it isn't exposed. I want to fix binarypackage selector next.03:46
sabdflok, /malone/packages/03:46
sabdflstub: sourcepackage is the bitch03:46
sabdflbecause id is the *only* unique way to refer to it03:47
kikosabdfl, I have a surprise for our fellow lunchpadders this mataro session03:47
kikojust you wait03:47
SteveABradB: I haven't looked at the zope3 yet.  I'll look at it a bit later.  I think it is to do with registering views for layers, and which is considered the default layer, but I can't be sure yet.03:47
stubWhich we need to fix, either by not refering to it or sorting out constraints so we can. Something for the conference.03:48
BradBSteveA: ok, that'd be great. without being able to upgrade, i don't have a clear path with how to proceed on implementing what we want for filtering vocabs.03:48
BradBsabdfl: heh, yeah, that screen needs paging :)03:48
BradBsabdfl: and, well, searching, of course03:49
sabdflstub: yes, well, we also have to reflect the reality out there ;-)03:49
sabdflbradb yes03:49
sabdflgo to mpg32103:49
BradBthis'll be another application of the super widget03:49
BradBsabdfl: ok, there03:50
sabdfllooks pretty good to me03:50
sabdflwould be nice to see # of bugs03:50
sabdflalso, i think we need to be able to filter bugs by status, severity03:50
BradBsabdfl: like "10 bugs found", you mean?03:50
sabdflneed a one-click accept / reject03:50
BradBsabdfl: actually, this would take the user to the main search screen03:51
BradBsabdfl: by which i mean it should, but doesn't yet.03:52
sabdflmain search screen?03:52
sabdfli disagree03:52
BradBsabdfl: yeah, the main bug listing.03:52
sabdfli don't think we want a one-size-fits-all search page03:52
BradBsabdfl: why duplicate effort?03:52
sabdflbecause different people will use malone differently03:52
BradBsabdfl: you can see all this information on the main search screen03:52
BradBsabdfl: i'm referring specifically to this functionality though.03:52
sabdflyes, but then you end up with bugzilla03:52
sabdfltoo many damn options to search on03:53
BradBsabdfl: this is simple though. we just filter on the package assignment.03:53
BradBsabdfl: and when we decide there's another thing that needs to be included on a bug listing, you only have to pay a developer to do it once, instead, of, say 3 times03:54
sabdflbut think of having bits of malone embedded INSIDE Soyuz and Doap03:54
sabdflin other words, i'm looking at a package, and click, now i see its bugs03:54
sabdflthis page will end up inside soyuz03:54
BradBsabdfl: sure, that's different from talking about not duplicating effort on this specific page.03:55
BradBsabdfl: what you're talking about involves ensuring we have a good, modular set of templates, to make it easy to reuse bits in other apps (e.g. expose macros for soyuz geeks to easily do bug listings on what are otherwise soyuz-like pages)03:56
BradBand the links *will* have to point to real bugs, even in soyuz.03:57
sabdfli *really* don't like the idea of a big search page, the way the "show all bugs" page is headed03:57
sabdflthe reality is that malone is many bug systems, aggregated03:57
sabdfltrying to show a god view of ALL the bugs is just going to blow people's minds03:57
BradBsabdfl: why don't you like having just one place to go to find out what you need to know about bugs?03:58
BradBsabdfl: seeing the god view will be extremely rare.03:58
BradBsabdfl: you only see 20 bugs at a time, and they'll always be filtered relative to your context. you'll only see them all if you say "show me ALL the bugs".03:58
BradBsabdfl: e.g. it might be mind-blowing that each of the 10 bugs assigned to you have also be assigned to four other packages, which would make for a huge listing, but you won't care, because you'll know nothing about those other assignments, because they'll be automatically filtered03:59
sabdflBradB: i'm debating changing the bugsourcepackageassignment to assign to a tuple of ( sourcepackagename, distro )04:01
sabdflon the basis that (a) these are pieces of information any user would be able to figure out04:01
BradBsabdfl: i mentioned this to you in london. you said not to bother, because at the start we'll just have ubuntu :)04:01
sabdflyes, now we are starting to deal with it because I'm playing with the debbugs sync04:02
BradB(spn, distro) is an inevitable change.04:02
sabdflthe problem is, it's not exactly accurate04:03
sabdflsourcepackage is in fact more accurate04:03
sabdflbecause you might get two different sourcepackages in the same distro with the same name, and different bugs04:03
sabdfland (spname, distro) does not allow you to differentiate04:03
sabdflyou might also get the same package in TWO distro's, but spname on its own does not allow you to filter it04:04
sabdflthe truest system would be ( sourcepackage, distro )04:04
=== BradB looks at the schema. maybe i meant (sp, distro), but maybe that can already be determined.
stubIs sourcepackage actually that accurate? If you want accuracy you need to specify a sourcepackagerelease.04:05
BradBstub: not for an assignment dude04:06
BradBstub: there may be some automagic to pick the affected version which auto-creates the relevant infestation though.04:06
BradBstub: additionally, we've also talked about having it so that when a new /infestation/ is added, there's a little checkbox that says "notify the maintainer of this package", which'll automagically add an assignment, if there wasn't already one, so you're covered on both ends.04:07
salgadosabdfl, person related pages should be inside foaf/ or foaf/people/ ?04:08
sabdflactually, foaf/people/ gives you more room for other things04:08
sabdflso run with that04:08
salgadosabdfl, ok. 04:09
stubWith infestations, I think the (sourcepackagename,distro) is a good option. There is no way for a user to specify the sourcepackage to assign to - the only information they have is the name. If they specify the sourcepackagename, we can spawn the relevant infestations for maintainers to trim.04:10
BradBer, sorry, the "automagic" i referred to in picking the affected version; i meant that the user would pick the affected version and we'd automatically add the infestation.04:11
BradBi.e. we'd modify the bug add form to allow choosing a relevant version number04:11
carlosstub: is there any problem with carlos-launchpad-language-preference.sql? I'm asking because it's not applied yet to the database04:12
sabdflthe issue with name + distro, is that we need to know how to deal with bugs when the source package changes04:12
stubcarlos: Oh. I applied a similar fix quite some time ago I think (after the discussion on the mailing list).04:12
sabdflbut... in many cases, name + distro will be canonical04:12
sabdfland in fact, in many cases it will be the best way to tell people where t  find the fix04:12
BradBthere's still a lot more metadata we're missing through. e.g. when a bug is reported on a product, we keep no information about the environment in which the user is able to reproduce the problem, which'll make life much harder for the people trying to fix it.04:12
sabdfl"ah, it's fixed in ubuntu / apache2"04:13
sabdflBradB: that almost needs to be a set of data around the assignment04:13
carlosstub: hmm, didn't saw it :-P04:13
stubcarlos: Check '\d person' and confirm04:14
carlosstub: yes, it's there04:14
BradBsabdfl: yeah, a set of data around something very soonish, that's for sure :)04:14
carlosstub: I'm going to kill the patch then. Thanks04:14
BradBsabdfl: though i could see the environment in which a bug is reproducable be /different/ for different versions (e.g. bug shows up on ppc and i386 in 1.17, but only on ppc by 1.20)04:15
sabdflBradB: i think that will be described in the messages04:16
sabdfli don't want to try to get too deeply into modelling every permutation and combination04:16
BradBsabdfl: that's not very useful though, because you can't search for stuff like that.04:16
sabdflpeople love debbugs despite its lack of structure, perhaps because of it04:16
sabdflseacrhing for arcane permutations and combinations is hugely overrated04:16
BradBit's not a major concern right now. time will tell.04:17
sabdflcan you click on one of the sync'd bug messages?04:17
sabdfllet's check out the display of the bug04:18
BradBeeeg, i really dislike the formatting of bug messages :/04:18
carlossabdfl: I have a patch for the raw po/pot files as you requested, could you confirm you agree with it before I ask for the database merge?04:18
BradBsabdfl: what do you mean by click on one of the messages though?04:18
BradBi'm looking at one of the sync'd bugs, if that's what you meant04:19
sabdflwe need to clean up the message display04:22
sabdflwe need to figure out how best to store messages on a bug where the message is genuinely an email message04:22
BradBsabdfl: it was clean before, save for the formatting-of-incoming-email-reports issue.04:23
sabdflcprov, Kinnison: are there functions in Gina to create a person given an email address? "sdfsfd sdfsdf <dsfsd@dfsd.com>" ?04:23
dilysMerge to rocketfuel@canonical.com/launchpad--devel--0: Second try to be able to store raw po/pot files into our database (patch-840)04:24
stubFor the record, Roundup is using pre because "people have tried all sorts of wierd ways of formatting the issues, but the only thing that works consistently is <pre>"04:25
sabdflthis is outside my knowledge of html etc04:26
sabdfli need to confirm that we are happy just to put the full email (headers and body and attachments) in the "contents" field of Message table04:26
sabdflthereafter i'll leave it to bradb and stub to figure out display issues04:27
BradBsabdfl: as long as we have some way of identifying type, yeah, i'd want to dump the whole message in there.04:27
kikosabdfl, yes, there are.04:28
kikosabdfl, (functions in gina)04:28
kikosabdfl, (to create a new person -- I should know, I wrote it!)04:28
kikostub, <pre> is really the only way to go with comments.04:28
stubDo you actually have a real email message from debbugs, or something munged? Character encoding issues would mean you can't just dump the raw email-as-received onto a webpage.04:28
kikootherwise you can't do ASCII art.04:28
kikoand ASCII art is mandatory04:28
stubkiko: (_0_)04:29
BradBstub: not onto the web page; into the db04:29
kikostub, 3:|>04:29
=== stub really shouldn't use an uppercase O with that one...
kikothanks for the consideration04:30
BradBsabdfl: anything else? are you satisified with our idea of having a "google" in malone?04:31
BradBactually, google integration would be pretty cool too. top 5 google links for the bug title or something.04:32
BradBconfigurable portlets, etc. but i digress...04:32
sabdflBradB: i also want to find one of these reports where the title is so long it makes the table bigger than the screen04:33
BradBsabdfl: yeah, we'll do eliding there.04:33
sabdflcan you do that in TAL?04:33
BradB"a huge problem that will entirely destroy your machine if you..."04:34
BradBsabdfl: i doubt TALES is that helpful. should be easily done with a bit of python though.04:34
stubIt would be trivial to add the eliding formatter to the fmt: namespace - checkout webapp/tales.py04:41
stubtal:content="whatever/title/fmt:elide80" or tal:content="whatever/title/fmt:shorten/40" or something similar04:42
sabdflthat would be perfect04:43
sabdflBradB: other than that i think we're in good shape04:43
BradBSteveA: any news on upgrading Z3? if not, i think it's more effective for me to return bug-fixing, in the hopes that we'll have this problem solved by the end of todayish.04:46
=== kiko is now known as kiko-fud
sabdflstub: can i send you a fragment of sql to strip cve ref's into a separate table?04:53
SteveABradB: I'm still working on other stuff right now.  I will look at it today.04:54
BradBok, thanks04:54
=== BradB wonders if dilys is alive
stubsabdfl: send away.04:56
sabdflso guys, if i strip CVE references into a separate table, what's a good name for the BugExternalRef table if it's just URL's?04:56
stubI think the name is fine04:56
sabdflstub: yowser, though you'd be in dreamland by now!04:57
sabdflBugExternalRef stays that?04:57
=== stub has no life
SteveAis there a bug filed on the fmt:shorten/60 idea?04:57
sabdflbut can i at least rename the "data" to "url" at this point?04:57
stubNope, but somebody should do it before I do the dogfood codedrop...04:58
stubsabdfl: Sure04:58
sabdfland description to title?04:58
=== lulu [~lu@host217-37-231-28.in-addr.btopenworld.com] has joined #launchpad
BradBsabdfl, stub: did one of you guys undo my change to "expand" the comment add widgets by default?04:58
SteveAstub: when are you doing the dogfood codedrop?04:59
stubYes. Description makes no sense. One day we might want a shortdesc, but it is probably WHUI04:59
stubBradB: Not me04:59
SteveAwe haven't use it04:59
stubErm... YAGNI04:59
sabdflBradB: yes, i did04:59
SteveAwhat a yagni becomes if you allow it into the code04:59
BradBsabdfl: urgh, why? :)04:59
stubI'm getting ahead of my self ;)04:59
sabdflbecause the form is ugly, and most people don't comment on bugs they browser04:59
BradBsabdfl: yeah, that's why i moved it to the bottom04:59
sabdflBradB: erm... it needs to be at the top05:00
BradBsabdfl: why?05:00
sabdfleveryone hates having to scroll to the bottom of bugzilla to comment05:00
sabdfland at the top, it's in the "most recent" position05:00
kiko-fudit does help make sure people read comments though.05:00
stubSteveA: As soon as my last patch filters through pqm I was hoping to drop. There will be another tomorrow too.05:00
sabdflcomments should start with the most recent05:00
kiko-fudbut there's always food for thought there05:00
=== kiko-fud runs off to lunch
(elmo/#launchpad) heh, debbugs use to do that and changed05:00
SteveAstub: I can do this in 10 mins.05:01
(elmo/#launchpad) it's the classic no-win situation, some people hate most-recent-at-the-top, others love it05:01
BradBI expect tree widgets to be used for trees, but oh well.05:01
sabdflbradb and i have been in a dance, moving it top to bottom and back, and collapsed and back05:01
sabdfli don't want a threaded display there, at least not for a while05:02
stubOhh... I see user preferences comming sooner rather than later. <div tal:condition="user/comment_box_up_top_dammit">...</div>05:02
BradBsabdfl: me neither.05:02
stubpage test that!05:02
sabdflelmo: i suspect most of the people who love it at the bottom do so because they want to change other people's behaviour05:02
kiko-fudwell, indeed. they want people to read comments instead of ignore them :)05:03
sabdfllet's just sit tight where it is, we can make these sorts of changes much later05:03
BradBthe way it is currently i've got to click my mouse an extra time and do a bit more pointer-fu to get what i want. perhaps i'm just that lazy that i notice extra-effort-required like that.05:04
BradBokie doke05:04
=== salgado is now known as salgado-lunch
BradBi wonder why i didn't put next and back links on the paging. i'll have to change that now so i can close the paging bug.05:05
BradBdilys: @#*$@&05:07
dilysMerge to rocketfuel@canonical.com/launchpad--devel--0: IHugeVocabulary vocabs should have human readable tokens for funky UI goodness (patch-841)05:10
sabdflBradB: most people who view a bug do not comment on it05:10
sabdflthose who do, have to do pointer-fu in any event05:11
sabdflBradB: looks like she's listening to you05:11
BradBshe's not telling me about bugs that have been fixed though05:11
dilysNew Malone bug #66: "Distro sourcepackage search navigation broken", submitted by Brad Bollenbach05:18
dilysMerge to rocketfuel@canonical.com/launchpad--devel--0: added fmt:shorten/60 feature to tales (patch-842)05:26
dilysNew Malone bug #67: "Bug assignments search too slow", submitted by Brad Bollenbach05:38
BradBdaf: ping05:40
=== SteveA pings daf and carlos
carlosSteveA: pong05:46
=== Kinnison pats gina on the rump and sends her to try *again*
=== Kinnison sighs
SteveAhi carlos.  how's the work on po and pot importing over the web going?05:47
stubKinnison: I'm about to kill the librarian.05:47
carlosSteveA: new database proposal in rocketfuel05:48
carloswaiting for mark's review05:48
Kinnisonstub: on zhongshan?!05:48
carlosSteveA: daf started yesterday the submit form05:48
stubKinnison: Oh - sorry ;)05:48
stubKinnison: Mawson ;)05:48
Kinnisonstub: *grin*05:48
Kinnisonstub: I'd be everyone's least favourite coder if I were doing these tests on mawson05:49
KinnisonI keep wedging the database for a start05:49
carlosand I'm going to look at the final import script/process, but need to discuss somethings with daf05:49
SteveAwill the new database stuff go into stubs next dogfood drop?05:50
Kinnisonalthough the good news is that I almost have gina to the stage we need her at for continual dogfooding of all of hoary05:50
carlosSteveA, sabdfl: well, I could also ask you about it, what should import the pot/po from the raw field in the database? a cron like script or from a website by hand?05:50
carlosSteveA: if mark agrees it should05:51
SteveAright now, just write it so that there is a URL to go to that processes one item from the queue05:51
SteveAonce you have all this working, we can talk about the exact mechanics of making it work05:51
=== salgado-lunch is now known as salgado
sabdflcron like script05:53
sabdflKinnison: rock05:53
sabdflSteveA, carlos, there isn't really a queue is there?05:53
sabdfljust a list of potemplates and pofiles which have an unprocessed uploaded file05:53
carlossabdfl: right05:53
carlosthe queue is the answer of a query05:53
sabdflcarlos: is your db patch in rf?05:53
carlossabdfl: yes05:53
SteveAessentially, that's a queue.  we want to take one item from it at a time.05:53
Kinnisonsabdfl: I wanted to ask you. How's Oscar?05:53
SteveAthe order of it doesn't really matter05:53
Kinnisonsabdfl: I'm starting to think of useful tests for him to do05:54
carlossabdfl:  carlos-raw-po-pot-files.sql05:54
carlosat launchpad/database/schema/pending05:54
dafBradB: pong05:54
dafSteveA: pong05:54
SteveAhi daf05:54
SteveAhow's the upload forms going?05:54
sabdflguys, we are going to ROCK the Mataro Sessions ;-)05:55
dafhi Steve05:55
SteveAneed any help with the forms?05:55
dafnot yet05:56
carlossabdfl: yeah!05:56
BradBdaf: when will dilys tell us when bugs are fixed again?05:56
dilysMalone bug #44 fixed for product Malone: Bad destination after editing a bugassignment05:56
dafdilys didn't know about product assignements, only package assignments05:57
dafand she didn't know about fixed bugs, only closed ones05:57
BradBshe might also want to know about rejected bugs05:58
stubcarlos: drop done05:58
sabdflstub: doh, there's a silly bug in that fragment05:59
carlosSteveA: any URL suggestion for the moderation queue form?05:59
sabdfli need to DELETE FROM bugexternalref WHERE bugreftype = 1;05:59
SteveA rosetta/+uploadqueue 05:59
sabdflso we get rid of the CVE things from BugExternalRef after we've created them in CVERef05:59
carlosstub: perfect, thanks!!!05:59
carlosSteveA, daf: The dogfood server is now 100% functional like the old alpha server06:00
carlos(or it should) :-P06:00
carlosdaf: and your UI improvements are soooo cool :-P06:02
dafthanks :)06:02
sabdflstub: should send you an updated copy, with comments?06:02
=== carlos leaves for some minutes
stubsabdfl: Or just check it into the pending directory. With comments of course.06:10
sabdflstub: the pending dir always gets me stuck06:10
sabdflbecause i can't commit it once i start working on the code that will use06:11
sabdflright now, i can't commit to rf, because my code expects the new structure CVE / ext ref06:11
sabdflthe tests will fail, unless that patch is in the schema/patch- file06:11
stubok. And you can't just commit a file you just added, which will hopefully be fixed in baz sometime sooner rather than later06:12
SteveAthat's the point of using a branch.  but, branches are hassle.06:12
dafstub: I thought that bug had been fixed06:12
dafI'm sure I did that recently06:12
stub'baz switch' should make it much nicer now. 'baz tag -S launchpad--dbpatches--0; baz switch launchpad--dbpatches--0'06:13
stubOoh... coolies.06:13
sabdflif i was really smart i would know in advance what db changes i'll need :-)06:13
sabdflbut alas, i'm not that smart06:14
sabdfland nonetheless, if the db change requires the existing code to change  you have the catch2206:14
=== stub orders another batch of prescience
SteveAyou can end up writing dodgy "if" statements all over...06:14
SteveAbut that is obviously bad news06:14
=== daf workraves
stubsabdfl: Those comments on the way, or do you promise to update comments.sql?06:17
dilysMerge to rocketfuel@canonical.com/buildbot--devel--0: dont conflict on _branch in versions (patch-75)06:17
sabdflstub: updated one in mail in 30 seconds.06:17
sabdflincludes the delete from06:17
Kinnisonsabdfl: did you see my question about Oscar?06:19
sabdflKinnison: nope?06:19
Kinnisonsabdfl: I asked how it was going because I have started coming up with ideas for tests I want oscar to run06:19
=== kiko-fud is now known as kiko
sabdflKinnison: in sooth, i've not begun06:20
sabdflstub: on the wire06:20
sabdflstub: how best to handle this one?06:20
Kinnisonsabdfl: Heh, I guessed you were way busy by the huge number of changes coming from your camp ;-)06:20
sabdflif you approve it and give me a patch number, i'll move it in and commit it at the same time as the code to handle the changes06:20
dilysMerge to rocketfuel@canonical.com/launchpad--devel--0: dont conflict on _branch in versions (patch-843)06:23
dilysMerge to rocketfuel@canonical.com/launchpad--production--1.4: cherry pick _branch overload fix (patch-1)06:26
stubsabdfl: Approved. patch-5-02-0.sql.06:27
sabdflstub: great! any changes?06:27
sabdflstub: lifeless says it's all good06:28
stubsabdfl: Need sample data regenerated unless you want me to handle that post commit06:28
sabdflstub: could you?06:28
sabdflhmm... i will06:28
sabdflno prob06:29
stubsabdfl: Are CVE's integers or text? I thought they were integers.06:30
=== Kinnison watches gina very carefully
sabdflive kept them as text just in case, am happy to change them now06:31
stubI think integer would be better - it is easier to make it text later if necessary than vice-versa.06:32
(elmo/#launchpad) they're not integers?06:32
(elmo/#launchpad) they're {CVE,CAN}-[0-9] {x4}-[0-9] {x4}06:32
(elmo/#launchpad) in some madeup pseudo glob, pseudo regex06:33
Kinnisonsabdfl: Looks like I got my first correct from-scratch import of warty main,restricted for i386,powerpc,amd64 with proper shared arch:all stuff06:33
sabdflgreat work Kinnison06:33
Kinnisonsabdfl: I'm knackered. Gina is a tough mistress06:33
stubelmo: Ta.06:33
sabdfllike takiing a 16 year old to a mature escort...06:34
=== Kinnison gets on with isolating the publishing code to make it a secondary action :-)
stubsabdfl: so scrap the integer, but feel free to comment BugExternalRef if you have your editor open ;)06:34
Kinnisonsabdfl: If you say so. *sir* :-)06:34
sabdflstub: willdo06:34
KinnisonSounds like the best way to do CVE validation will be a nice python function06:34
dilysNew Malone bug #69: "Need a CVE id validator", submitted by Stuart Bishop06:37
dilysMerge to rocketfuel@canonical.com/launchpad--devel--0: Fix huge-select search result default option (patch-844)07:14
carlosstub: dude, how many hours your days have?07:15
carlosSteveA, sabdfl,daf : ping07:40
sabdflcarlos: hi07:40
carloshow could we handle the development of the po/pot import until the database changes are applied to rocketfuel?07:40
carlosbecause we need some changes to databas/pofile.py07:40
carlosbut if we change that file without changing the database launchpad will be broken07:40
carlosand both, daf and I need those changes to do the work07:41
sabdflthat's the glitch in our dba method at the moment07:42
sabdflin theory, make a branch, work on the branch07:42
dafI'm planning to temporarily use the ZODB07:42
sabdfldaf: please don't07:42
dafwhy not?07:42
sabdflbecause it will just require more fiddling to get running properly07:42
sabdflcarlos: let me look at the db patch07:43
sabdfli can probably make a few comments, then stub can approve it first thing in his morning07:43
sabdfland you can commit tomorrow07:43
sabdflare you and daf both going to be working on the same codebase?07:43
lulunight all :o)07:43
carlosto work from a branch I suppose daf or I should tag from our personal archive to prevent the star break, right?07:43
=== lulu [~lu@host217-37-231-28.in-addr.btopenworld.com] has left #launchpad []
carlossabdfl: daf will do the submit part07:44
carlosand I will do the final import part07:44
carlostwo different forms07:44
sabdflif daf makes the core change (sql patch, interface and db objects), then07:44
sabdflyou tag off daf into a new working branch / dir07:45
sabdflthen you can work on separate aspects of the problem07:45
sabdfldaf can merge your work regularly07:45
sabdflwhen you are ready, daf commits07:45
carloshope arch works as it should ;-)07:46
sabdfldaf, does that sound workable?07:46
dafit sounds possible07:46
sabdflthe main thing will be that when daf merges from carlos, daf must merge from the NEW BRANCH07:46
dafI suspect it might be overkill07:46
=== Kinnison -> Huntingdon for dinner
carlossabdfl: perhaps we could wait some days to implement this feature07:47
sabdflthe problem is that to get started on this you both need the same set of changes07:47
carlosuntil the database change is done....07:47
sabdflcarlos: yes, if you can wait till tomorrow, then we can get the db stuff squared away07:47
carlosdaf: what do you think?07:47
carlosis just that I thought it's a big priority task07:47
carlosthat's why I asked07:48
dafis there anything stopping you from working on the database code in your own archive?07:48
carlosdaf: the pofile.py changes07:48
carlosto use the new database rows07:48
carlosI suppose you need it also, right?07:48
dafyes, I will need it07:49
carlosso you decide, I tag from your archive and go ahead with this or wait for stub and mark's review tomorrow07:50
dafok, let's try branching07:51
carloswill you work on your main archive ?07:52
carloswow, it's still tagging your archive..08:04
dafmaking the base-0, I expect08:05
dafperhaps I should put cacherevs in my archive08:05
carlosit's checking the gpg sign of all your patches 08:06
carlosthis is insane08:16
carlos20 minutes and it's still tagging the archive08:16
carlosdaf: what does cacherevs exactly?08:17
dafpristine revisions in an archive, I think08:17
dafsaves on the patching08:18
carlosand consumes more harddisk, right?08:18
dafyep :)08:18
carlosI just understood the need of fooo--200408:18
dafyeah, doing a lot of tagging once every year is not so bad08:19
carlosdaf: the problem is when someone else needs to tag your archive :-)08:19
carlosat the end of the year08:20
carlosand we have only less than 6 months of work...08:20
dafI wonder how you deal with that...08:20
carlosI cannot imagine it at the end of 2005....08:20
daflifeless: thoughts?08:21
carlosin fact, I don't want to imagine it ;-)08:21
carlos"only" 99 patches to go...08:22
=== carlos works on other tasks
=== daf cacherevs daf@canonical.com--2004-MIRROR/launchpad--devel--0--patch-299
carlosdaf: wait for the patch-300, it's prettier08:24
dafactually, I could have commit hook that automatically makes a cacherev if patch % 50 == 008:24
(mdz/#launchpad) BradB: here?08:26
BradBmdz: yeah08:26
carlosdaf: feel free to share it with me ;-)08:26
(mdz/#launchpad) BradB: I'd like for pitti to start using Malone to track security bugs, if you're up for it08:27
=== pitti [~martin@] has joined #launchpad
(mdz/#launchpad) BradB: should be a reasonable small-scale real-world test08:27
dafhi Martin :)08:27
pittiHi daf!08:27
(mdz/#launchpad) BradB: should we set up a separate instance for that, or what?08:27
pittiNow I finally know where all the other guys are hanging out :-)08:27
BradBmdz: sure...does this involve us having to import anything?08:27
BradBmdz: i think we should use the same instance.08:28
=== BradB is currently suffering through trying to optimize the bug listing for 7000+ bugs.
BradBwhich is still tinyish, gragh08:28
(mdz/#launchpad) BradB: we can start with an empty db, should be fine08:28
(mdz/#launchpad) pitti can import a couple of things by hand if he has open issues to track08:29
(mdz/#launchpad) we just ran into a situation where we released an advisory, and then a new cross-reference appeared08:29
pittican I import Debian bugs?08:29
(mdz/#launchpad) and we don't have anyplace to record that information08:29
pittior shall I create new ones?08:29
(mdz/#launchpad) pitti: sabdfl was talking about that, I think we have that capability08:29
=== pitti never saw malone up to now
carloswe are invaded by Ubuntu!!!08:29
sabdflmdz, pitti, we'll bring this on stream for production use during es-conf08:30
pittisabdfl: a very good idea. Can we have kind of a tutorial for the distro folks?08:30
BradBpitti: do you know how to access malone and all that then?08:30
pittiBradB: as I said, I never saw it08:30
pittiBradB: do I have/need an account?08:30
BradBpitti: you have one08:31
(mdz/#launchpad) sabdfl: does this sound OK as an alpha test?08:31
BradBpitti: whatever your ubuntulinux.org u/p is08:31
pittiah, nice08:31
BradBpitti: you also need the client certificate08:31
BradBpitti: the url is https://launchpad.ubuntu.com/malone/bugs08:31
sabdflmdz: sure08:31
BradBpitti: the client cert was mailed to the list a while back, called launchpad.p12, i think08:32
pittiBradB: hmm, 40308:32
BradBpitti: you need the cert :)08:32
pittiBradB: to which list?08:32
BradBlp@ i think08:32
pittiBradB: http://lists.ubuntu.com/archives/lp does not exist08:33
pittiBradB: can you please simply mail me the cert?08:33
carlospitti: launchpad08:33
BradBpitti: what's your email?08:33
pittiBradB: martin.pitt@canonical.com08:33
=== pitti is looking forward to get at look into the future :-)
pittiBradB: got the mail now08:35
carloslifeless, daf: wow. 40 minutes took the tag!!08:36
pittiBradB: sorry for sounding dumb, but what shall I do with this file now?08:36
(mdz/#launchpad) pitti: point firefox at it08:37
(mdz/#launchpad) pitti: or perhaps go into the certificate preferences and import it there08:37
pittimdz: hmm, it offers me to download it08:37
pittimdz: I already looked for a cert pref dialog08:37
dafcarlos: :-(08:38
pittiah, here08:38
(mdz/#launchpad) pitti: preferences, advanced, certificates, manage certificates08:38
pittimdz: yeah, got it08:38
pittiBradB: I set my own password, now it asks me for the password that was used to encrypt this certificate08:39
pittioh, that wasn't a typo :-)08:39
pittiseems to work08:39
pittiBradB: thanks08:40
BradBno prob08:40
BradBpitti: if you find bugs in malone, you can report them against the malone product08:41
pittiI just got the bug confirmation from Debian08:41
pittican I import this bug or shall I create a new one?08:41
kikopitti, dude, that is ONE BIG FRIGGIN PATCH08:41
kikocan't you escape the autotools crud?08:41
pittikiko: mysql?08:41
BradBpitti: there's no debian import support08:42
BradBpitti: there will be, but there isn't yet.08:42
pittikiko: the mysql debdiff does not touch autotools files08:42
kiko--- mysql-dfsg-4.0.20.orig/bdb/dist/configure08:42
kiko+++ mysql-dfsg-4.0.20/bdb/dist/configure08:42
pittikiko: however, every hunk is present twice, due to the "odd" build system08:43
kikosorry, I looked at the wrong file.08:43
pittikiko: debdiff mysql-dfsg_4.0.20-2ubuntu1.dsc mysql-dfsg_4.0.20-2ubuntu1.1.dsc|lsdiff | grep configure08:43
pittikiko: empty08:43
pittiBradB: hmm, I tried with my u.o website login, does not work08:44
pittiBradB: I do not get any error, just the login page agein08:44
BradBpitti: have you changed your p/w recentlyish?08:46
pittikiko: maybe now you have an idea why the update is pending for so long :-)08:46
pittiBradB: recently == about a month ago08:46
kikopitti, it's scary08:46
BradBpitti: you'll have to use the older one08:46
pittikiko: agreed :-)08:46
pittikiko: not the patches I'm used to for other packages08:46
pittikiko: however, this fixes some design flaws,  three-liners don't work there08:47
pittiBradB: argh, after registering to plone I immediately changed it to my own08:47
pittiBradB: I don't know the original one08:47
pittiBradB: I'll look into my mail archives08:47
BradBpitti: can you confirm that your login works on ubuntulinux.org right now too?08:48
=== pitti tries
pittiit did an hour ago08:49
pittiBradB: yes, confirmed08:49
BradBhm, well, if you're positive it doesn't work on the dogfood site, then yeah, i'd try the older one.08:49
pittiBradB: I found my original password, but it doesn't work neither :(08:52
kikoyeah, I admire you for it, but ooof08:52
BradBpitti: you've only had those two p/w's on ul.org ever?08:52
pittiBradB: yes08:53
=== BradB boggles
pittiBradB: the autogenerated one after registering08:53
pittiwhich I immediately changed to my own08:53
BradBI dunno, you'll have to ask stub then, unfortunately.08:53
pittiBradB: can you reset it somehow?08:53
BradBpitti: you've got mail dude08:54
BradBlet's try this08:55
pittinot yet arrived08:55
pittiah, now08:55
BradBi'm not sure if the link is correct in that email, but you should be able to figure out the correct one if it isn't08:55
pittihey, what's this?08:56
pitti" Sorry, there has been a problem in resetting your password. Please try again, using the Forgotten password  form"08:56
BradBwhat url?08:57
pittiI just used the form to try again08:57
pittiI used the URL you sent me08:57
BradBwhat url?08:57
BradBas i say, the url might be incorrect08:58
BradBand it is08:58
pittiI just got a new one08:58
BradByou don't want to touch that08:58
BradBchange the link in the first email you go to launchpad.ubuntu.com08:58
pitti" A system error occurred."08:58
BradBoh yeah? crap, i was worried that it still hadn't been fixed. :/08:59
pittiwell, this error was for the u.o link08:59
pittiI'm trying the launchpad one now08:59
pitti Error type: zope.publisher.interfaces.NotFound09:00
pittiNotFound: Object: <canonical.publication.RootObject object at 0x41002b4c>, name: youforgottenpassword'09:00
sabdflSteveA: around?09:00
dafthat's a 404 in Zopespeak :)09:00
dafwhat's the URL?09:00
sabdfldaf, bradb, we're converting code from somewhere else to work in zope309:00
BradBpitti: https://launchpad.ubuntu.com/ubuntulinux/forgottenpassword/5brqqfj0059xnk4kfbns854gdsr5gxvfdnr9lzqm09:01
sabdfland seeing the problem where NONE of the object attributes are visible to Z309:01
sabdflI have an interface, which seems correct09:01
sabdfla content declaration in the zcml09:01
pittithis gives a right plone form09:01
dafhmm, I've seen some problems writing to attributes, but not with viewing them09:01
pittiI enter my stuff and then09:01
sabdflpermission=zope.Public AND set_schema=interfacename09:02
pitti Error type: exceptions.NameError09:02
pittiTraceback (innermost last):09:02
pitti    * Module zope.publisher.publish, line 138, in publish09:02
pitti      result = publication.callObject(request, object)09:02
pittisorry, dudes :-/09:02
pittiNameError: global name 'getUtility' is not defined09:02
dafhmm, interesting09:02
BradBpitti: eck, yeah, that is really unfortunate to see on a near-production machine. :/09:02
BradBsabdfl: yeah, you saw me mention that right/09:03
pittiBradB: do you want the full trace?09:03
BradBsabdfl: that's what i spend a couple hours trying to figure out yesterday, to no avail09:03
BradBpitti: no, i can see it here09:03
BradBpitti: you'll have to ask stub to get your p/w sorted out then09:04
pittiI will09:04
pittihe's not online now, I try it tomorrow morning09:04
pittii. e. in about 10 hours09:04
BradBsounds good, thanks09:04
pittithanks to you, too09:04
pittinice to be an alpha tester :-)09:04
BradBsabdfl: SteveA was going to look into it today.09:05
pittimdz: I'm afraid this experiment has to wait until tomorrow, then09:05
dafBradB: pitti's looks like a missing import to me09:05
sabdflBradB: where did you see this?09:05
daf* pitti's error09:05
sabdflhold on, can one of us not just reset pitti's passwd to something simple for the purposes of this run, on dogfood?09:06
BradBdaf: yes09:07
BradBsabdfl: when i tried upgrading zope 3 locally09:07
kikoit's a missing import, surely.09:07
=== BradB tries something
pittigo, BradB, go! :-)09:09
=== daf sings the BradB theme song
sabdflBradB: i haven't done any z3 update09:09
BradBoh, i misread your earlier comment09:10
sabdflive seen this before09:11
sabdflspend hours staring at it09:11
sabdflit turns out to be a tiny bit of zcml somewhere09:12
sabdflbasically, the zope security proxy is not allowing ANY reading of the attriutes09:12
(mdz/#launchpad) pitti: ok09:13
BradBpitti: your email in the system is martin@piware.de09:18
BradBpitti: anyway, i've reset your p/w for now. i'll /msg you09:19
pittihey, now it shows "You're logged in" :-)09:20
pittithanks, BradB09:20
pittimdz: I'm in09:21
(mdz/#launchpad) great09:21
pittimdz: so now I create a new report, tag the CAN stuff at it09:21
BradBpitti: cool09:21
pittimdz: then I close it again and mark the versions?09:21
(mdz/#launchpad) do we have a way to add references to USNs?09:21
BradBsabdfl: what error are you seeing exactly?09:21
(mdz/#launchpad) pitti: sounds appropriate09:21
sabdflmdz: we have a general URL link09:22
sabdflwe could easily add a USN09:22
BradBsabdfl: that's almost surely a missing interface declaration.09:22
BradBsabdfl: e.g.09:23
BradB    <content class="canonical.launchpad.database.Bug">09:23
BradB        <allow interface="canonical.launchpad.interfaces.IBug" />09:23
BradB        <require permission="zope.View" set_schema="canonical.launchpad.interfaces.IBug" />09:23
BradB    </content>09:23
sabdfl<content class="canonical.launchpad.database.Branch">09:23
sabdfl        <allow interface="canonical.launchpad.interfaces.IBranch" />09:23
sabdfl        <require09:23
sabdfl            permission="zope.View"09:23
sabdfl            set_schema="canonical.launchpad.interfaces.IBranch"09:23
sabdfl        />09:23
sabdfl    </content>09:23
BradBwhat's the attribute that's forbidden?09:23
sabdflall of them09:23
sabdfldir(subject) gives [] 09:24
sabdflid in particular09:24
dafwhat's the difference between zope.View and zope.Public?09:24
BradBsabdfl: you're positive that config's getting loaded? e.g. what if you change it to IBranchadljasdfj?09:25
sabdfli think bradb just hit the nail on the head09:26
sabdflBradB: big, big, big brownie points09:26
pittimdz: I just had a second look09:27
pittimdz: the new CAN-2004-1015 does not apply to the Ubuntu and Debian versions09:27
(mdz/#launchpad) pitti: heh, ok09:29
(mdz/#launchpad) pitti: thinking of it, we should probably have a way to represent that information09:30
(mdz/#launchpad) like Debian's nonvulns list09:30
(mdz/#launchpad) sabdfl, BradB: is this possibly a Malone thing, or part of a hypothetical security database app?09:30
pittimdz: right now I wrote into the changelog that we aren't vulnerable to CAN-2004-101109:30
BradBmdz: if it's an important external source related to a bug (e.g. an id that links to something on a security site) we probably want to integrate it into malone...if that's what you're asking. :)09:31
BradBmdz: like CVEs09:32
dilysNew Malone bug #70: "several exploitable buffer overflows", submitted by Martin Pitt 09:33
(mdz/#launchpad) BradB: what we're describing is the bit of information which says "I investigated the vulnerability referred to by this CVE name, and it doesn't apply to source package versions x, y and z"09:33
(mdz/#launchpad) BradB: I guess that's a separate bug, with its own CVE reference, and affected version informatino09:33
(mdz/#launchpad) information09:33
pittijust submitted a#7009:34
BradBhm, you'd have to clarify the semantics of CVEs with sabdfl. he's in the midst of doing some data-model-fu with it, so he'd be more familiar with how that's intended to work.09:35
pittiand triggered another exception when searching for bugs :-(09:35
sabdflpitti: i'm just splitting out the cve stuff from other references09:36
sabdfli think we could easily build in an CVE-unaffected list09:36
sabdflso malone could track who is affected *and* who is not affected09:36
pitti Error type: exceptions.ValueError09:37
pittiValueError: invalid literal for int(): createdby09:37
BradBpitti: that screen's getting crufty and going away soonish09:37
pittiI cannot search for ubuntu bugs neither09:37
BradBpitti: by adding another search widget to the main bug listing, so everything's in front of you at once.09:37
BradBpitti: they're all ubuntu bugs :)09:37
BradBbut yeah, i think i know what you mean, and hm, i'm not sure if there's a way to do that easily yet09:38
pittiBradB: is there a reason why #70 appears in the list twice?09:38
pittihttps://launchpad.ubuntu.com/malone/bugs/+index?batch_start=40&batch_end=50 has two entries for #7009:39
BradBpitti: each row you're seeing is an assignment to a package or a product.09:39
BradBpitti: in that case, you should have only filed the bug against the package, not the "ubuntu" product09:40
pittiBradB: ah; bugzilla requires to specify both, that's why I did that09:41
BradBpitti: currently, it's implicit that a package belongs to ubuntu09:41
BradBpitti: we'll need to improve this UI though, of course, but for now that's how it works.09:41
BradBpitti: you can "Reject" the assignment to ubuntu to clean things up, if you want09:42
pittiBradB: I clicked on source package infestation; however, cyrus21-imapd does not exist :-(09:43
pittiBradB: rejected09:44
pittiBradB: nice, "Martin Pitt" appears twice in the Assignee list09:45
=== pitti searches for his other instance
pittidarn, it rejects assigning to BinaryPackageValue cyrus21-imapd (although I selected it from the list)09:47
BradBpitti: can you file a bug if there are releases missing that prevent you from filing the appropriate infestation?09:51
pittiBradB: in fact I only see one release09:52
BradBpitti: also for the duplicate assignee thing09:52
pittiBradB: 2.1.16-1009:52
pittiBradB: but the warty, warty-security and the latest Hoary releases are missing09:52
BradBpitti: and the binarypackagevalue thing09:52
pittiBradB: product malone and no source package?09:56
pitti"canonical malone", that is?09:57
pittibtw, the product selection could be a bit easier09:57
BradBjust the product malone, and yes, i'm working on improve the selects (though waiting on getting Z3 upgraded before continuing)09:57
dilysNew Malone bug #71: "available source versions missing", submitted by Martin Pitt 09:58
pitti^^ is this automated somehow?09:59
BradByes, somehow :)10:00
dilysNew Malone bug #72: "cannot assign #70 to cyrus21-imapd", submitted by Martin Pitt 10:10
dilysNew Malone bug #73: ""Martin Pitt" is in the assignee list twice", submitted by Martin Pitt 10:13
pittinew Pitti bug #1: "Martin Pitt submits too many bug reports", submitted by the caretaker10:14
=== mdz_ [~mdz@69-167-148-207.vnnyca.adelphia.net] has joined #launchpad
BradBgargh i wish bug assignments had one main table, with only the columns that differ per assignment type stored in the child tables. batching assignment results is exceedingly complicated.10:21
BradBwhich means that batching any kind of aggregate results set like assignments is going to be exceedingly complicated.10:22
=== BradB is now known as BradB|lunch
=== mdz_ is now known as mdz
SteveAsabdfl, BradB|lunch: still here11:37
sabdflhey SteveA11:37
sabdflwe had a long dig for a problem with interface permissions11:38
sabdflbradb finally asked if we were oading the zcml11:38
sabdflwhich we weren't :-)11:38
sabdflproblem solved11:38
SteveAholger and armin wanted to pay their "rent" for staying at my and aiste's place during the pypy sprint, so we and the rest of the pypy folks just went out for dinner.11:38
SteveAwas this in a script?11:39
SteveAoh, you mean you didn't include the zcml from another zcml file?11:39
SteveAI really need to improve the "forbidden" page.  I know how to make it give decent disgnostics.  Sorry I didnt get round to it so far -- could have saved you some time.11:40
SteveAI'd like to chat about this in london -- see if I can see what error you were getting, and what it would have been good to show you.11:41
SteveAdaf: don't use zope.View11:41
SteveAthe only permissions we should be using at the moment are zope.Public and launchpad.AnyPerson11:42
SteveA(I keep trying to get zope.Public changed to just "public", but Jim doesn't agree with me on that one_11:42
dafSteveA: I'm not, I just noticed that sabdfl was11:44
sabdflsince it's late and i'm highly medicated for this flu...11:44
SteveAsabdfl: don't use zope.View.  Use only zope.Public and launchpad.AnyPerson.11:44
sabdfldaf, is that lp:url code still alive somewhere?11:44
sabdflstevea: got you, will update my zcml11:44
SteveAwhat does lp:url do?11:44
dafI don't know11:45
=== BradB|lunch is now known as BradB
=== SteveA checks out getting launchpad running with the latest Zope3
SteveAwhat would you like lp:url to do?11:45
SteveAsabdfl: not good about the flu.  Want me to bring some russian vodka with me?  it's one of the local cures, I think11:46
sabdflglobal cures ;-)11:47
sabdflda! spasibo!11:47
sabdfli'm thinking of cross-referencing the lp:url list from daf with the pagetests to get a report of pages we don't test11:47
=== sabdfl puts on janitorial suit
=== sabdfl banishes zope.View
SteveAwhen I saw lp:url, I was thinking of a TALES extension11:49
SteveAgetting a report of untested pages would be cool.  a bit like a code coverage report11:49
SteveAoh, I see, lp:url in zcml files11:50
SteveAI'd forgotten about that, for some reason11:50
=== SteveA awards himself a "Duh!"
BradBif it printed out the titles of the pages that weren't being tested that would be like AI...11:50
SteveAAI ?11:50
BradBArtificial Intelligence!11:51
BradBThings you're not testing:11:51
SteveAwe might be better off loading the zcml, and looking for page directives11:51
BradB  - Bug Listing11:51
BradB  - Product Search Page11:51
BradB  - etc...11:51
SteveAlike, the "foo.html" page for an IBug11:51
BradBSteveA: you remember how i vouched for having parent/child tables in London? i.e. one "parent" table that holds all the common columns, and then child tables for each thing that has to store something slightly specialized.11:53
BradBSteveA: i can give you one example of a very good reason why.11:54
SteveAyeah, I'm familiar with the pattern11:54
BradBhere's why we we're hurting without it in launchpad:11:54
BradBbug listing => batching results11:54
BradBbug listings are a list of assignments11:54
BradBthere are two kinds of assignments: product and package11:55
BradBeach has their own table11:55
BradBso, we need a special batching class that knows how to batch *two* different result sets into one thing11:55
BradBso, imagine:11:55
SteveAdaf: ping11:55
BradBpackages is a SelectResults containing 3000 packages, products is the same, but for products11:56
BradBassignments *must* be aggregated and order by bug id11:56
dafSteveA: pong11:56
BradBso, the user wants to see results 100-120....how in the /heck/ do you return that? :)11:56
SteveAdaf: I want to make a change to rosetta/configure.zcml, but I'm not sure if it is right.  Also, I want you to look at mark's XXX comment in there, and see if it is still pertinent.11:57
BradBs,return that,batch that,11:57
SteveABradB: sorry, can't pay attention just now, as I'm working on the zope3 upgrading11:57
SteveAcan you email it?11:57
SteveAmaybe to the whole list?11:57
sabdflBradB: that's why my report separated these thiings out11:57
dafSteveA: I'm not sure if that comment is still extant11:57
SteveAif it is not pertinent, please remove it11:58
SteveAthere is always kudos available for removing XXX comments11:58
SteveAI want an XXX report to run on checkins11:58
BradBsabdfl: one sec, i'll tell you why that's no good either, but i'm on the phone :)11:59
SteveAor maybe a daily run, sending results to the mailing list, along with "blame" for the XXXes11:59
SteveAwe should not really have XXXs in production code.11:59
=== SteveA pastes daf a diff

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