danilosgmb, hi, did you have a chance to take a look at accordionlayer branch Gary mentioned?09:17
gmbdanilos: Hi, sorry, missed your ping earlier as my connection was down. So, I've been looking at it and haven't got much further than you (judging by your email). However, I'm wondering if Gavin's changes to the way lp.js is build might have anything to do with it. Probably not but I'll ask him anyway.11:03
danilosgmb, right, thanks11:04
gmbdanilos: So that's a no, then. Hmm.11:09
danilosgmb, yep11:10
gmbdanilos: First up, I think we can drop the activate_collapsible_div code completely, since it's not referenced.11:19
gmbOf course, that doesn't solve our other problems, but at least it's one less thing to care about.11:19
danilosgmb, right, I did see activate_collapsibles somewhere11:19
gmbYeah, there's activate_collapsibles(), which works, and activate_collapsible_div(), which doesn't appear to exist.11:19
gmbThe latter is called in product-index.pt11:20
danilosok, so I assume they are unrelated then :)11:20
danilos(I was thinking it might have been a left-over typo from rename or something)11:20
gmbI think it might be that someone's reinvented the wheel and then backed out some of their changes.11:20
gmbBut since nothing broke they never noticed the error.11:20
gmbAnyway, I'll remove that code now.11:20
danilosgmb, sure11:22
danilosgmb, also, I wonder where is the link supposed to show up (so I know what I am looking for, since I haven't looked at the code that deeply :)11:23
gmbdanilos: It should appear under the "Administer" link in the top portlet, but only if you're logged in as Mark.11:23
danilosgmb, mark only, not if I am logged in as admin?11:24
gmbdanilos: Mark only, last I checked.11:24
gmbmark@example.com:test, IIRC11:24
danilosugh, /me hits himself over the head for not following the instructions as written :)11:25
danilosand launchpad.net/firefox oopses if you are not logged in :)11:25
gmbdanilos: Yeah, that's a fun one.11:27
danilosgmb, ok to commit the fix for that: https://pastebin.canonical.com/44657/11:28
gmbdanilos: Yes, go for it.11:30
gmbdanilos: You'll need to merge the latest ~yellow/launchpad/accordionoverlay11:30
danilosgmb, yep11:31
gary_posterany wondrous solutions? :-) I see the emails showing progress11:41
gary_posterbtw, danilos, I hope you weren't working on unsubscribe in anger.  I have some significant changes from your branch in a branch of my own.  If you have been, no worries--we can connect in a little while and compare notes11:43
danilosgary_poster, hey-hey11:43
danilosgary_poster, no, I haven't since I've seen the emails, and that was right away in the morning :)11:44
gary_posterhey!  cool.11:44
danilosgary_poster, so, I was enjoying the JS debugging :)11:44
gary_posterhow was the test?11:44
gary_posteralso, I was wondering if you and gmb have daylight savings time yet.  US chnged this weekend11:45
danilosgary_poster, we'll see, not too happy with it11:46
danilosgary_poster, hum, I think we're on only next week11:46
gary_postertest: :-( sorry to hear, but hopefully you'll be happily surprised11:47
gary_posterok, next week is not too far away :-)11:47
danilosgary_poster, actually, it's two weeks away: http://www.timeanddate.com/news/time/europe-starts-dst-2011.html :)11:47
gary_poster:-) ok11:48
gary_posterI'll ask around what everyone wants to do for the two weeks--move to the US version of 8:30 now (in 41 min) or wait till you guys have moved11:49
gary_postereither way is fine with me11:49
danilosgmb, fwiw, even if I ignore the issue of LP.links.me being undefined (and set it manually in the code to "/~mark" :), LP.cache.context causes the next error, so I guess loading order is what we need to fix11:51
gmbdanilos: Right. And here's a weird thing. If you do11:53
gmbThe first shows that LP.links.me is set correctly ("/~mark") but the second shows that it's undefined.11:53
danilosgmb, I explained that in the email, the first one stores a reference to the object, and you evaluate it only when you look at it11:54
gmbdanilos: Sorry, missed that message (or glazed over, one of the two)11:54
danilosit's generally a smart idea to ignore me (i.e. 9 out of 10 times), but sometimes I hit the nail on the head :)11:54
gary_posterI'm guessing you guys saw Gavin's email about rebuilding etc?11:55
gmbgary_poster: Yes. His work seems to have introduced some debugging noise into the JS console, but he's given me a workaround for that and is emailing the list about it.11:55
danilosgary_poster, right, I've seen it, and 'make jsbuild' has been functioning correctly for me (and I prefer the non-minified version, so I didn't need the min option)11:56
danilosgmb, I am looking forward to that email, but if it's easy enough, I wouldn't mind hearing from you directly about the workaround11:56
gmbdanilos: make clean_js; make JS_BUILD=min jsbuild_lazr; make JS_BUILD= -W jsbuild_lazr jsbuild11:57
gmbApparently, if you s/min/raw you get unminified JS, too.11:57
danilosah, nice11:58
danilosok, so the first thing I'll do is try to run setup() from non-domready and see how that works11:59
danilosyes, that works12:01
gmbdanilos: Great minds :)12:01
gmbAlthough it would have been quicker if I'd remembered that syntax matters.12:02
danilosso, this is https://pastebin.canonical.com/44658/ what will allow one to get a subscription link by clicking on a big ugly h2 in the middle of the page :)12:08
gary_postership it!12:10
danilosheh, yeah, "Click to load" is inviting enough :)12:11
danilosit should at least allow some of us to test the real stuff, while others try to figure out the load ordering problem :)12:12
danilosanyway, gary_poster, call still in 1h18 mins? (I want to get lunch before that then)12:12
gary_posterdanilos, you are the first to voice a preference, and we need to make a decision now, so sure, your preference wins. :-)12:13
* gmb grabs lunch too12:15
danilosgary_poster, heh, I am basing my "preference" on what works for me today, but sure, it's fine in general as well :)12:15
gary_posterbac and benji, gmb and danilos do not change DST for two more weeks.  For these two weeks, I suggest we let them keep their same times for the morning meeting, and we move to 9:30.  When they switch, we'll switch back.  They have lunch schedules to work with  and such.12:15
gary_postercool danilos :-)12:16
benjidaylight saving time is the devil12:16
gary_posterbenji, btw, I made some server side changes that affect your work, particularly in browser/structuralsubscription.py.12:17
gary_poster(they are pushed to accordion overlay12:18
gary_posterbac and benji, danilos and gmb are lunching.  are one of you tackling the JS problem , or both of you?  If you are not tackling that problem, note the pastebin danilos gave above to get something working so you can fix other stuff.12:50
bacgary_poster: i'm looking at it12:51
gary_posterok thanks bac12:51
benjiI'm blistfully ignoring the problem at the moment.12:51
gary_postersounds perfect12:51
gary_posterbac benji danilos gmb mumble/kanban in 1-ish13:29
benjigmb: I use a gmail filter to add a label to mail sent directly to me in a colo(u)rful way so it stands out from the sea of in-direct mail I get13:41
gmbbenji: As do I, but bugmail comes directly to me still. At least I think it does. I'll have a look.13:42
gmb(I know I need to fix my filters since everything went to launchpad rather than being in {malone, rosetta, soyuz, etc.} anyway)13:42
benjioh, that's why it works for me, I use a different address for bugmail so they don't get the colo(u)rful label13:46
gary_posterbenji, locally I now have a working (modulo the fix for the onload thing from Friday that's affecting all of us) page https://bugs.launchpad.dev/firefox/+subscriptions that shows subscriptions-per-target using the exact same code of yours as powers https://bugs.launchpad.dev/firefox/+bug/1/+subscriptions .  So, once the other fix is done, I'll commit this and we should be good there.14:01
_mup_Bug #1: Microsoft has a majority market share <iso-testing> <ubuntu> <Clubdistro:Confirmed> <Computer Science Ubuntu:Invalid by compscibuntu-bugs> <EasyPeasy Overview:Invalid by ramvi> <GNOME Screensaver:Won't Fix> <Ichthux:Invalid by raphink> <JAK LINUX:Invalid> <The Linux OS Project:In Progress> <metacity:In Progress> <OpenOffice:In Progress by lh-maviya> <Tabuntu:Invalid by tinarussell> <Tivion:Invalid by shakaran> <Tv-Player:New> <Ubunt14:01
gary_posterI did see one oddity that was pertinent to our discussion Friday14:01
benjiwhat's that?14:01
gary_posterSomehow I managed to make a subscription to  “mozilla-firefox” package in Ubuntu that does not have a filter14:02
danilosok, figured out the problem, and it's as simple as us not using a closure for the event handler, so the following fixes it: https://pastebin.canonical.com/44665/ (I fired a special lp-ready event when all the data was set up and still got undefined, which pointed me at the problem)14:02
gary_posterNot quite sue how14:02
gary_posterah, danilo!14:02
danilosgary_poster, fwiw, I might have forgotten to update the sampledata, so maybe that was just an existing one14:02
gary_postercool danilo, I'll verify14:03
gary_poster(sampledata I mean)14:03
gary_posterdanilo, that patch makes perfect sense now that I see it :-P14:03
danilosyeah, so, how do I proceed from here? land this on accordionoverlay branch directly?14:03
gary_posteryeah I think so.  We'll give bac a second to have an opinion?14:04
danilosgary_poster, heh, yeah, it's nice that JS/YUI lets you do stuff like this14:04
danilossure thing, bac, what do you think of the following fix for the JS load problem: https://pastebin.canonical.com/44665/14:04
bacsorry was afk for a bit.  reads back.14:04
bacdanilos: why does that work?14:05
gary_posterwhat we had before calls module.setup({content_box: "#structural-subscription-content-box"}) immediately and registers that result as the domready callback14:06
danilosbac, because previous version ran setup() when defining an event handler and set the return value (probably null) as the event handler14:06
bacland away!14:06
danilosbac, ack, off it goes14:06
gary_posterthanks danilos, great14:06
danilosit's very "nice" that: 1) YUI lets you do this, 2) this is so obvious from syntax that multiple of us spent hours on it; /me feels sad for JS taking over the world as a de facto programming language14:09
danilosanyway, landed, gary_poster, when you've got a minute, it'd be nice to chat about what do I do next :)14:10
gary_postercool danilos.  what to do next: would one of the two jobs benji listed for you be ok?14:11
gary_posteractually, gimme a sec14:11
gary_posterI'm applying your fix to those pages too14:11
danilosgary_poster, ok, let me read through the proposed tasks in the meantime14:12
gary_posterbenji, danilos, ~yellow/... has the code to fix up [bugtarget]/+subscriptions to work, and to make [bug]/+subscriptions to have the new JS fix14:16
gary_posterso now I need to know what to do too.  :-)14:16
danilosgary_poster, would one of the two jobs benji listed for you be ok? :P14:17
gary_posteractually, I'm going to go see if I can dupe making a structural subscription to a package that does not have a filter14:17
danilosok, never mind that then :)14:17
gary_posterdanilos, what I'm going to do is just verify.  Then I'll look into one of those two, probably, yeah.  So, tell me which one you are going to work on and I'll try to do the other one14:18
danilosgary_poster, just to confirm, that work should be based off ~yellow/.../accordionoverlay14:18
gary_posterdanilos, you want to chat about the unsubscribe in anger work now for a sec even though we are not focusing on it?14:18
gary_posteryes danilos14:19
danilosgary_poster, sure thing, that's fine as well14:19
gary_posterI'll push a branch for you then I'll be ready14:19
gary_posterdanilos, I don't know if we'll need to look at the branch or not, but it is lp:~gary/launchpad/bug728370 .  Also I made changes (mostly additions) to https://dev.launchpad.net/LEP/BetterBugSubscriptionsAndNotifications/DirectSubscriptionsOnBug14:22
bachurrah, danilos, my subscription link has returned!  thanks!14:22
danilosbac, cheers :)14:53
gary_posterdanilos, you're going to give benji the "it's broke" news? :-)14:53
danilosbenji, hey, I am looking at "Save Changes" button you proposed, but I can't get anything to happen when I click on "Edit subscription" button14:53
danilosgary_poster, already on it :)14:53
gary_posteryeah :-)14:53
danilosbenji, (ftr, I am using ~yellow/lp/accordionoverlay branch)14:54
danilosbenji, if I should use a different branch or debug the problem, just let me know14:54
benjidanilos: darn; it works in my checkout and I was just about to merge from ~yellow/lp/accordionoverlay and push my changes so I'll be able to look at it momentarily14:54
danilosgary_poster, oh, one thing: what do I do on the kanban board? :)14:55
danilosgary_poster, (because I am only assigned to stuff that I am not working on :)14:55
benjidanilos: if you don't want to wait you can work from ~benji/launchpad/yellow-accordionoverlay which should work for you14:56
danilosbenji, ok, sounds good, thanks14:56
danilosbenji, this is all feature work 2, right?14:57
benjidanilos: nope, feature work 114:57
danilosbenji, oh, right, thanks14:58
gary_postersorry danilos, yeah, feature 1.15:06
gary_posterbenji, danilos, I think foo.bar must have had a sample-data subscription, as danilos suspected.  I unsubscribed and resubscribed, and he had a filter.15:08
gary_poster(after the resubscription)15:08
benjiis this for the "“mozilla-firefox” package in Ubuntu" subscription?  if so, I believe that comes from sampledata15:09
gary_posteryes, benji15:09
danilosgary_poster, right, I've add a quick-jobs task to update the sample data (to avoid future confusion), if you don't think that warrants a task, just drop it15:10
danilosgary_poster, oh, btw, did the query improvement prove enough?15:10
gary_postertask: cool, danilos, sounds like a good idea unless benji decides we don't need to have a filter for every subscription after all15:10
benjinope, a filter in every pot^W subscription is here to stay15:11
gary_posterquery: it seems so.  on staging the two pertinent queries were 1 ms and 30 ms, respectively, so I declared it a victory.  the oops tool is aggregating these timeouts poorly so it's hard to be completely confident, but it looks pretty good15:12
gary_posterok good to know benji, thanks15:12
danilosgary_poster, cool, great job there!15:12
gary_posterthanks :-)15:12
gary_posterbenji, your branch is already merged in with ~yellow :-/15:18
benjigary_poster: I've merged it before, but since made more changes and plan on merging it again.  Is that problematic?15:19
gary_posterbenji, no I just meant that, at least as of the version of ~benji/launchpad/yellow-accordionoverlay that I just got, it's already been merged in with ~yellow/launchpad/accordionoverlay/ and the problem that danilos descibes exists there.  Therefore, merging doesn't show us what might be going wrong.15:20
gary_posterthe diff between those two branches is 36187 lines15:21
benjigary_poster: well, I'd hoped that merging would break my working copy (in hopefully the same way) so that I can fix it15:21
danilosbenji, you've been very productive if it's 36k lines :)15:22
gary_postermerging ~yellow into ~benji will probably break ~benji, yes15:22
gary_poster:-) I think the 36K is changes in devel since benji branched15:22
benjiI believe they are.15:23
gary_posteror, last merged, I should say15:23
benjidarn, this looks like something I've already fixed; maybe I fixed it wrong or another merge accidentally clobbered my last fix; I need tests15:29
* gmb -> run15:31
gary_posterbac, did you and jon agree on who will actually remove that bad JS function call, or is that still up in the air?  if it is still up in the air, I'm going to send him a reply thanking him and asking him to clarify if he is going to do it, or if he is just going to file a bug and we should do it.15:31
bacgary_poster: yes, i'm removing it in my gallery-accordion update branch.15:31
gmbgary_poster, bac: I removed the bad call from the ~yellow branch this morning.15:31
baci've just submitted it for review15:31
gary_postergreat, thank you bac and gmb15:32
baci think it has been removed about half-dozen times today15:32
gmbDistributed development FTW.15:32
* gmb -> really run15:32
benjidanilos, gary_poster: I just pushed my branch to ~yellow which also contains the fix15:35
gary_posteryay benji, thanks15:36
benjiwell, it's currently pushing15:36
gary_posterheh, SPEED!15:36
gary_posterbenji, I'm going to tackle the unsubscribe link15:36
benjishe can'a take no more capt'n15:36
danilosbenji, cool, danke!15:36
gary_posterYou said "I15:37
gary_posterbelieve the REST collection operations are the way to remove a filter,15:37
gary_posterbut I'm not sure how, exactly."15:37
gary_posterForgive the flail,15:37
gary_posterbut do you have any ideas on where to start?15:37
gary_posterMy thoughs:15:37
gary_poster1) ask leonard15:37
gary_poster2) grep for ... something.  find a collection and then grep js for that name?15:37
gary_poster3) (obvious winner) ask Benji!15:38
benjia hint: I would have been more specific if I knew the specifics ;)15:38
gary_posterlol, ok, proceeding with #2 and then #1 :-)15:39
gary_poster(I can also ask Graham when he returns)15:39
gary_posterdang, is the pull *still* going on?15:40
gary_posterpush I mean15:41
danilosbac, hi, it seems you've overwritten a few changes gmb and I put into ~yellow/launchpad/accordionoverlay (and then recommitted them, so no big harm done, but just saying)15:41
benjigary_poster: bzr: ERROR: These branches have diverged.15:41
benjitrying to fix now15:41
danilosfwiw, integration branches suck because we all have to behave as pqm15:41
gary_posterah, I think you and Brad got in a race, benji15:41
benjithat would explain it15:41
gary_posterack danilos.15:42
danilosthe proper way to merge something would be: (cd shared-accordionoverlay && bzr pull lp:~yellow/launchpad/accordionoverlay && bzr merge my-branch && bzr ci -m "Merge my branch doing foo-bar." && bzr push)15:43
gary_posterwell, that's still a race condition, just a shorter one, yeah>15:44
danilosand never, ever use push --overwrite :)15:44
danilosgary_poster, I am suspecting push --overwrite is what happened here15:45
gary_posterah :-/15:46
benjigary_poster: push succesful15:47
gary_posterbenji: "No revisions to pull."  Maybe wrong destination?15:49
gary_poster(I used "bzr pull :parent" and then "bzr pull lp:~yellow/launchpad/accordionoverlay/" just for apranoia)15:49
benjigary_poster: yep, wrong destination; I guess I should do it the way danilos described instead of pusning directly, so that'll take second15:49
danilosthat's the problem of an override in locations.conf, I'd love a solution to that (not even --remember helps :()15:50
benjigary_poster: this time for sure15:56
gary_posterbenji, seems to have worked :-)15:57
* danilos -> off, tty tomorrow16:30
gary_posterbye danilos, thank you16:34
bacdanilos: i didn't use overwrite so i figured i was ok16:55
benjigary_poster: do you know how much progress danilos made on the save button?  Are you doing the Unsubscribe link?17:01
gary_posterno, and yes, benji17:01
benjigary_poster: ok, I figured that if he didn't make much (or his work was publicly available) I would pick it up17:02
gary_posterI suspect he would have told us if it were publicly available.  I have no idea as to his progress. :-/  If you have something else valuable to tackle, that might make sense.  Otherwise I might be able to come up with ways to divide up mine if you wanted.17:03
gary_posterbenji, what is the minimum I have to do to get my changes to structural-subscription.js used?  rebuild the js somehow?  do I have to restart LP?17:12
benjigary_poster: make jsbuild17:13
benji(and reload the page of course)17:13
gary_posterI don't have to restart?17:13
gary_postercool thanks17:15
benjiI assume the resources are either served by apache or are "resources" in the zope sense and since the server is in developer mode they are reloaded on every request17:15
gary_posterright, cool17:16
gary_posterDoes thismean anything to you?17:16
_mup_Bug #1: Microsoft has a majority market share <iso-testing> <ubuntu> <Clubdistro:Confirmed> <Computer Science Ubuntu:Invalid by compscibuntu-bugs> <EasyPeasy Overview:Invalid by ramvi> <GNOME Screensaver:Won't Fix> <Ichthux:Invalid by raphink> <JAK LINUX:Invalid> <The Linux OS Project:In Progress> <metacity:In Progress> <OpenOffice:In Progress by lh-maviya> <Tabuntu:Invalid by tinarussell> <Tivion:Invalid by shakaran> <Tv-Player:New> <Ubunt17:16
gary_posteruncaught exception: [Exception... "Component returned failure code: 0x80004005 (NS_ERROR_FAILURE) [nsIXMLHttpRequest.setRequestHeader]"  nsresult: "0x80004005 (NS_ERROR_FAILURE)"  location: "JS frame :: https://bugs.launchpad.dev/+icing/rev12438/build/launchpad.js :: _setHeaders :: line 29595"  data: no]17:16
gary_posterLine 017:16
gary_posterbenji ^^^17:16
gary_posterrunning to lunch now though :-P17:16
benjinope; the client sets some X headers to pretend to be using PUT, that'd be my first guess.  Why it's not working I don't know.17:18
gary_posterk thanks18:06
benjido we expect the YUI tests to fail when we run "xvfb-run ./bin/test --layer=RegistryWindmillLayer -cvvt test_yuitests"?18:31
* gary_poster would expect not, but expects/hopes bac would have a better idea18:33
bacbenji: i would expect them to pass but i haven't tried in a while18:34
baci've just been loading up in a browser18:34
baci'll try now.18:34
benjiIt may not be a test failure as much as a test error: AssertionError: Test harness or js failed.18:34
benjigary_poster: I guess I need to recruit a new RM now; I suppose an email to lp-dev is the way to go.18:40
gary_posterbenji, sounds about right.  if there are no replies, Francis/the team leads will bustle about I expect18:41
bacbenji: that's a great CYA error message18:41
benjigary_poster: k18:41
benjibac: heh18:41
baci got it too18:41
bacbenji: it is pretty bad.  it means it got no yui test results18:43
benjibad indeed18:43
gary_posterbenji, if you have more than one subscription on the page, they are all numbered 0.  (e.g., edit-0, unsubscribe-0).  I'm looking into it, but giving you a chance now to tell me in case you've already fixed it or are working on it18:45
benjigary_poster: that's a new one18:45
benjigary_poster: is that on the bug subscriptoin listing?  I don't see that behavior.18:47
gary_posterbenji, yes.  the reason why is that you restart link_id to be 0 for each subscription_info, but you expect link_id to be unique in wire_up_edit_links.  I see why you did it, but it doesn't work well for finding thigs by id, the way it's done now18:48
benjiyeah, I'd be happy for us to find a better way18:50
benjiuntil then you could make the counter a global and not reset it18:50
gary_postercool, looking into it18:50
gary_posterbenji, so here's an odd one.  I am successfully sending "DELETE /api/devel/firefox/+subscription/name16/+filter/1" to the server, and getting back a 200, but they are not actually deleted. :-/18:54
benjibac: are you investgating the YUI test failure further or shall I take it on?18:54
gary_posterI guess I'll have to pdb in the publisher, 'cause I have no idea what else to do18:54
bacbenji: i'm looking righ tnow18:54
benjiI find the use of "constants" in structural-subscription.js overdone.  E.g., option.set(INNER_HTML, team.title);19:08
bacbenji: some i lifted and probably too many.  others are useful.19:13
benjiEXPANDER_COLLAPSED and     EXPANDER_EXPANDED = '/@@/treeExpanded',19:14
baci think the pattern is there b/c JS will silently let you use the wrong string literal19:14
benji oops19:14
benjiEXPANDER_COLLAPSED and EXPANDER_EXPANDED make sense as constants19:14
gary_posteryou can set an attr on a dom node and JS won't complain if you are setting innrHTML rather than innerHTML, I believe he means19:16
benjithat sounds a lot like how Python will let you set an attribute on an object and won't complain if you are setting innrHTML rather than innerHTML19:17
bacbenji: i had turned off output to the console for debugging.  that's why the testrunner died.19:46
benjibac: interesting; let me know when the fix is available on ~yellow; thanks19:47
bacbenji: i just pushed a new version to ~yellow but have not been able to verify it.  i'll be afk for a bit so if you could try it out that would be helpful19:47
benjicool, will do19:47
bacbenji: test just finished.  seems to work.19:48
benjiDoes anyone know of a way to run a subset of the YUI tests?20:20
gary_posterI think I've heard that you can't.  but you probably shouldn't go by my hearsay.20:21
benjiI hope it's not true.20:21
gary_posterderyck is gone by now...bac is the only person I'd know to ask who might be around at this hour20:22
gary_posterI *assume* deryck is gone now20:22
gary_postergary, source of assumptions and hearsay :-P20:22
bacbenji: no, you cannot run a subset within a group21:00
baci.e. you have to run all registry at once21:00
benjiok; thanks for the info21:00
bacgary_poster: next to deryck, sinzui has been most involved in yui-test stuff.21:05
gary_posterbac, ah good to know thanks21:21
bacgary_poster: plus sinzui works extra late now!21:22
gary_posterheh, even better21:31
gary_posternight all21:45

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