2011/06/10/#launchpad-yellow.txt

=== Ursinha is now known as Ursula
=== Ursula is now known as Ursinha
=== Ursinha is now known as Ursula
=== Ursula is now known as Ursinha
=== Ursinha is now known as Ursula
[13:15] <gary_poster> bac, benji, objectives today if possible (especially since I'm going on vacation)
[13:16] <gary_poster> also, is it safe to run lp2kanban?  Should I just run it with
[13:16] <gary_poster> bin/py src/lp2kanban/bugs2cards.py -e [email protected] -p <yourpassword>
[13:16] <gary_poster> or do I need to specify the board?
[13:17] <gary_poster> It looks like the ones I've done as test cases have not yet been synced
[13:17] <benji> gary_poster: I did mine yesterday... hopefully correctly.
[13:17] <gary_poster> cool benji I'll look.  thanks
[13:18] <gary_poster> benji, bac, oddly, I get
[13:18] <gary_poster> Traceback (most recent call last):
[13:18] <gary_poster>   File "src/lp2kanban/bugs2cards.py", line 5, in <module>
[13:18] <gary_poster>     from argparse import ArgumentParser
[13:18] <gary_poster> ImportError: No module named argparse
[13:19] <benji> gary_poster: did you run the buildout? that's a new dependency
[13:19] <gary_poster> I did, but maybe before I updated.  lemme retry
[13:20] <benji> if that doesn't work, perhaps it wasn't added to the install requires entry
[13:20] <gary_poster> benji, bin/buildout did nothing, and retrying failed again.  We can look into it later
[13:24] <gary_poster> benji, updating setup and buildout made it work.  I'l commit in a sec.  Looks like it is updating ythe test board?
[13:24] <benji> gary_poster: no idea, I haven't touched it lately
[13:24] <gary_poster> looks like -b yellow should work...
[13:26] <bac> hi gary_poster
[13:26] <gary_poster> hey bac
[13:26] <gary_poster> got it to work
[13:26] <bac> gary_poster: yes, yellowtest is the default.  safer that way
[13:26] <gary_poster> agreed
[13:26] <bac> gary_poster: and cards must be marked to get synced
[13:26] <gary_poster> but no cards are reported as updated (I mean, that's what the script reports).  I'm seeing what the cnstraint is...
[13:27] <gary_poster> ah ok
[13:27] <bac> title.startswith("sync:") or a tag in the description
[13:27] <gary_poster> coo;
[13:27] <gary_poster> cool, I see that
[13:28] <gary_poster> trying
[13:28] <gary_poster> bac benji danilos, call in 2
[13:28] <bac> ok
[13:28] <danilos> ack
[13:29] <benji> gary_poster: for a second there I thought Julia had taken over your IRC client
[13:30] <gary_poster> bac, benji, ta-dar! :-)  I synced my two bugs
[13:30] <benji> yay!
[13:30] <gary_poster> ga ga goo goo
[13:30] <danilos> fwiw, does anyone find that leankitkanban is much less responsive lately
[13:30] <bac> gary_poster: ooo, try again
[13:30] <danilos> maybe that was while the script was running :))
[13:30] <gary_poster> heh
[13:30]  * benji wonders if it's onomatopoeia like a pirate day.
[13:30] <gary_poster> bac, running
[13:31] <gary_poster> I'll start the call when it is finished
[13:31] <gary_poster> note you have to reload the board
[13:31] <bac> nice
[13:31] <gary_poster> it's done
[13:31] <bac> no
[13:31] <bac> it reload automatically
[13:31] <gary_poster> oh I see
[13:31] <gary_poster> I just wasn't patient enough the first time
[13:31] <gary_poster> sweet!
[13:31] <danilos> yeah, I got a notice how the board is updated
[13:32] <benji> yeah, in my experience it behaves the same as when people update the board
[13:32] <bac> gary_poster: how did you resolve the argparser issue?
[13:32] <gary_poster> bac, just updated setup.py and buildout.cfg.  will commit
[13:33] <bac> why did it work for me?  is that required for things that are in the standard distribution?
=== Ursula is now known as Ursinha
[13:39] <benji> bac: my psychic powers tell me that you're running Python 2.7 which has argparse and Gary is running 2.6 which doesn't
[13:40] <bac> benji: hmm, i thought i'd used 2.6
[13:41] <benji> bac: that or you're using a python with argparse already installed
[13:41] <bac> benji: i guess.  i don't recall installing it.  /me looks
[13:42] <danilos> gary_poster, lp:~danilo/launchpad/bug-772754-other-subscribers-remove-cruft
[13:42] <bac> benji: i guess i'm wrong!  it is in /usr/lib/pymodules so i guess that is something i installed?
[13:43] <benji> I can't tell from here. ;)
[13:43]  * benji supresses a system-python-is-evil rant.
[13:44] <bac> yep, i'd installed python-argparse at some point
[13:46] <bac> hey benji i was thinking it might be a good idea to save off the initial state of the kanban board in case we mess it up in some way.  i assume there is nothing in their API that would allow recovery.
[13:47] <benji> yeah, I don't know of a way to set the entire board state (although, as you know, they let you retrieve it)
[13:48] <bac> benji: i forgot to tell you that you and i swapped #launchpad help contact days this week!  i covered for you on monday.  you mind picking it up today?
[13:49] <benji> bac: sure
[13:50] <benji> I swear!  Now the Gnome shell is crashing on me.  Maybe it wan't Unity's fault.
[13:56] <benji> gary_poster: after looking at bug 697735 for a little while, I've realized that either I don't think it should be considered a bug or I don't understand what it should do instead -- if you hand craft odd URLs you'll get wierd errors, what's wrong with that?
[13:56] <_mup_> Bug #697735: ValueError raised submitting bogus query string <lp-foundations> <oops> <Launchpad itself:Triaged by benji> < https://launchpad.net/bugs/697735 >
[14:01] <gary_poster> sorry, just saw that; looking
[14:02] <gary_poster> benji, from a high-level perspective, the problem is that OOPS reports are supposed to be actionable
[14:02] <gary_poster> benji, it's fine for the user to get an error
[14:03] <gary_poster> benji, but as you say, there's nothing for us to fix--so we should not get an OOPS
[14:03] <gary_poster> That's my understanding
[14:03] <benji> makes sense, I was missing that angle
[14:03] <gary_poster> Of course, fixing it may be a little exciting :-/
[14:04] <benji> yeah, I'm starting to wonder if I made a bad pick.  I'll timebox at least the first stab at fixing it to see where it goes.
[15:02] <gary_poster> anyone, what date did you choose for your allhands objectives' target dates?  I have responded to all of them so I can't see them anymore  :-P
[15:03] <gary_poster> Oh Apr 12
[15:03] <gary_poster> nm
[15:05] <danilos> gary_poster, something random for me, mostly in March/April next year
[15:05] <gary_poster> :-)
[15:05] <gary_poster> cool
[15:05] <gary_poster> dragnob said in her email today april 12
[15:05] <gary_poster> I dunno if that is real or not
[15:06] <gary_poster> but whatever, it's in the ballpark :-)
[15:06] <gary_poster> thanks
[15:07] <danilos> gary_poster, fwiw, all the MPs are up and linked to the bug, in order of how they are in my pipeline (and in order of their dependencies; -lp-names is the only one which is relatively independent)
[15:08] <danilos> gary_poster, -sections, -subscribers, -activity, and -loading should be most interesting
[15:08] <gary_poster> danilos, awesome.  I got your last branch and I think it is built now; I'll look at the UX, and the diffs, soon.
[15:09] <danilos> gary_poster, fwiw, anything after -loading (inclusive) but before the last one (exclusive) will have two subscribers lists :)
[15:09] <gary_poster> danilos, heh, ok, good to know :-)
[15:09] <danilos> gary_poster, I wouldn't mind you reviewing a branch or two while I add tests for -actions :))
[15:10] <gary_poster> danilos, ok cool.  I should be able to start in about 10 min.
[15:12] <danilos> gary_poster, yeah, no worries, you don't have to review them (especially not all of them, that's why I split them up :)
[15:12] <danilos> gary_poster, I am also asking on #launchpad-dev, so if are starting a review, please claim it
[15:12] <gary_poster> ack
[15:16] <benji> gary_poster: what is an "informational" oops?
[15:16] <Ursinha> an oops that's not an oops? :P
[15:17] <gary_poster> benji, yeah
[15:17] <gary_poster> we track them separately
[15:18] <gary_poster> we don't need to drive them to zero
[15:18] <benji> gary_poster: so would inappropriate use of :int be an informational oops?
[15:18] <gary_poster> but it is convenient to use the OOPS machinery to collect the data
[15:18] <gary_poster> eh
[15:18] <gary_poster> no
[15:18] <gary_poster> not as far as I know
[15:18] <benji> but we don't really want to collect anything
[15:19] <gary_poster> because we don't want to track itright
[15:22] <gary_poster> benji, do you know how to generate api docs from a lazr.restful app, like api.launchpad.net/+apidoc ?  This is a question from ISD
[15:23] <benji> gary_poster: not off the top of my head, but I'm sure I could figure it out quickly enough
[15:24] <gary_poster> benji, if that's ok, please do; my stack is getting too deep for last day before vacation :-)
[15:24] <benji> heh, ok
[15:37] <benji> gary_poster: there's no super-straightforward way, but it's not too bad.  The best thing would be for them to look at utilities/create-lp-wadl-and-apidoc.py in LP.
[15:38] <gary_poster> benji, ok thanks
[15:42] <gary_poster> benji, I told Ricardo Kirkner, "pindonga," that he could ping you if he had questions.  I suggest that the first time he asks you for help, you give it to him as soon as convenient; subsequent times give him focus no more than once or twice a day if possible, unless, in your judgement, you think you ought to give him more.  Good relations are good, and they are doing important things.
[15:43] <benji> k
[16:02] <danilos> gary_poster, fwiw, I found one of the problems "generic" webkit JS engine (eg. epiphany) borks on: "class" as a variable name, or more frequently, a function parameter :/
[16:04] <gary_poster> danilos, interesting!
[16:05] <gary_poster> good to file away
[16:05] <gary_poster> and unfortunate :-/
[16:05] <danilos> gary_poster, do you happen to know if that's a problem for Safari as well? I find it interesting because webkit JS should be identical in both, minus the version differences
[16:06] <gary_poster> I don't know off-hand danilos.  I mosty use Chrom these days, which has a different JS engine AFAIK
[16:06] <danilos> gary_poster, yeah, it does
[16:12] <danilos> gary_poster, looking at the code, it seems your unlanded branch is the only one which tries to use it, though :) http://paste.ubuntu.com/623491/ if you want to apply it to your branch directly
[16:13] <gary_poster> danilos, cool, thanks, will do
[16:14] <danilos> gary_poster, ftr, it seems to be a problem with safari as well: http://blog.solution10.com/2010/02/use-of-class-in-javascript
[16:16] <gary_poster> cool
[16:16] <gary_poster> thanks
[16:30]  * danilos -> off, pushing branches up took longer than I expected, so still tests missing for the final branch
[16:31]  * danilos needs to add task cards to the kanban board as well... :/
[16:59]  * gary_poster lunches
[17:01] <gary_poster> bac, I want to reorder the sections in danilos' final branch.  I don't know if it affects yours.  I think that "Maybe notified" should be last, as it is in the mockup (and I wonder if we should have a help icon with the explanation for what "Maybe notified" means).
[17:01] <gary_poster> I suspect that is in a later branch
[17:02] <gary_poster> but wanted to share now, in case it is in that one
[17:02] <gary_poster> I'll dig in more after lunch
[18:24] <bac> ok
[18:31] <gary_poster> I'll just put the comment in the last MP and not worry about it, I think.
[18:41] <benji> gary_poster: I have a pretty good solution to the bug we discussed earlier (bug 697735) but I wanted to get a consult to be sure it doesn't go too high on the hack-o-meter: https://pastebin.canonical.com/48394/
[18:41] <_mup_> Bug #697735: ValueError raised submitting bogus query string <lp-foundations> <oops> <Launchpad itself:Triaged by benji> < https://launchpad.net/bugs/697735 >
[18:41] <gary_poster> k, looking
[18:41] <benji> it scores a 3/10 on my unit here, but mine is known to be unreliable
[18:41] <gary_poster> :-)
[18:46] <gary_poster> benji, mm, BrowserRequest.__processItem is so tasty.
[18:47] <benji> heh, how so?
[18:48] <gary_poster> sarcasm, just that it is so big and dense that copying and modifying is really unappealing, and then combine that with the __* munging and you have a really great reason to do the monkeypatch, which is a shame
[18:48] <gary_poster> benji, did you consider having a marker interface for the request instead?
[18:49] <gary_poster> we do that kind of thing a lot
[18:49] <benji> oh, indeed
[18:49] <benji> re. marker; well, it'd be a marker on the exception, right?  That'd be fine with me.
[18:50] <gary_poster> benji, wouldn't it have to be on the request?
[18:50] <benji> the reason I did it this way is because it mirrors the way we did something similar for the web service
[18:50] <gary_poster> since _isIgnoredException takes strtype
[18:50] <gary_poster> not the actual instance
[18:50] <benji> it also (after I changed it) takes the exception instance
[18:51] <gary_poster> oh, duh, yeah
[18:51] <gary_poster> so yeah, on the exception sounds good
[18:51] <gary_poster> that way you could do it on the fly or as a general statement in code
[18:52] <gary_poster> ("this kind of exception is not oops worthy")
[18:52] <benji> sounds good
[18:52] <gary_poster> cool
[18:52] <gary_poster> benji, the only thing that really irks me is the monkeypatch, and I see why you did it.  Overriding __processItem does not seem like a nice option
[18:53] <gary_poster> well...
[18:53] <benji> right, it seemed like the lesser of the two evils; the other option would be to modify zope.publisher to make this easier
[18:57] <gary_poster> I looked for the word "monkeypatch" in the code, and I only saw mailman do it, which is not regarded as one of our high marks :-)
[18:59] <gary_poster> benji, I'd get bac's opinion, which will be easy since he'll probably be your reviewer.  I think I'd trust his opinion more than mine.  The only other thought I had was that you could consider moving the monkeypatch to lib/lp_sitecustomize.py.  It is where we do other similar sorts of things, and is where people will expect to find stuff like this AFAIK
[19:00] <gary_poster> benji, other than that, I think the interface-on-the-exception method of squelching OOPSes will be quite welcome, and should be announced.
[19:00] <benji> k; I'll look in lib/lp_sitecustomize.py, and add a comment there that hopedully would have helped me find it when I was looking for a home for this thing
[19:00] <gary_poster> cool
[19:00] <benji> k
[19:01]  * benji orders bright white, embossed announcement cards.
[19:01] <gary_poster> :-)
[20:05]  * benji is reminded by his phone to do CHR tasks.
[20:12] <benji> hmm, we seem to be behind in project/license review (or I'm doing it wrong)
[20:28] <benji> gary_poster: your lost soul is still lost: https://support.one.ubuntu.com/Ticket/Display.html?id=2433  What should I do for him?  Ask a LOSA to change his account's email address?
[20:31] <gary_poster> benji, I was bad (and/or swamped in other things) today and didn't do CHR, and we are missing gmb, so you are getting a bunch.  Sorry. :-(  bac, is there any chance you could give some thoughts on help we could give to that guy--the oops that he mentions?
[20:31] <gary_poster> I *think* we just need to set a preferred email, but I'm not really sure, and going int the raw SQL feels a bit like diving into the deep end which might have some hungry sharks hanging around.
[20:31] <benji> heh
[20:32] <gary_poster> So I'm hoping you have some insight on "tried and true" things to do to help that guy, bac
[20:32] <gary_poster> (from your registry days)
[20:32] <benji> k, I'll set his email address; how do we know that we're not being soccially engineered into giving control of someone's account to someone else?
=== Ursula is now known as Ursinha-lunch
[20:36] <bac> gary_poster: i'll look
[20:37] <gary_poster> benji, we contacted him at this email address
[20:37] <gary_poster> thanks bac
[20:37] <gary_poster> getting kids from school, biab
[20:38] <benji> Hmm, I don't understand that.  He said he doesn't have access to the account's current email address any more.
[20:41] <bac> benji: i'm confused too
[20:57] <gary_poster> benji, bac, it may have been because it was compromised?  The situation was that his account was sending spam
[20:57] <bac> benji: so here is my take.  the dude was suspended for spamming.  it is unclear whether he confirmed that he resolved his spam problem, i.e. changed password, got rid of browser issue
[20:58] <bac> i think when we reactivate his account he has to go to the 'forgot my password' dance to get reinstated
[20:58] <bac> is that right, gary_poster?
[20:58] <gary_poster> (he said there was no browser issue)
[20:58] <bac> if so, he's currently stuck as he won't be able to retrieve the link
[20:58] <gary_poster> forgot my password: well...we suspended his LP account
[20:59] <gary_poster> not his SSO acount AFAIK
[20:59] <bac> gary_poster: then why can't he log in now?  why is his account still showing as suspended (grey avatar)?
[20:59] <gary_poster> I don't know.  I think it is because of an LP bug
[20:59] <gary_poster> if you look at the state in the administartion page
[21:00] <gary_poster> then I think you'll see that he is actuve
[21:00] <gary_poster> actve
[21:00] <gary_poster> active
[21:00] <gary_poster> (he said, typing slowly)
[21:00] <gary_poster> what I thought we needed
[21:00] <gary_poster> was to set a preferred email
[21:01] <gary_poster> because that's what the OOPS points to
[21:01] <gary_poster> but maybe logging in via SSO would work?
[21:01] <gary_poster> Should we tell him to make a new account, and then merge the new into the old?
[21:02] <gary_poster> if there's no accessible email associated with the old account, that won't work either
[21:02] <bac> oh, i missed that there was an OOPS
[21:02] <bac> no, a merge will not work
[21:03] <gary_poster> ok
[21:03] <bac> gary_poster: what is the OOPS?
[21:03] <bac> i don't see it
[21:03] <bac> i don't think lifeless understands my silly attempts at humor.
[21:04] <gary_poster> bac, argh, sorry.  This is not the most pitiful of souls, nor the one I was thinking about or talking about.  I meant https://support.one.ubuntu.com/Ticket/Display.html?id=2120
[21:04] <gary_poster> who has been unable to log in for two weeks now
[21:04] <bac> aha!
[21:04] <gary_poster> bac, OOPS-1973H90
[21:05] <gary_poster> bug 793670
[21:05] <_mup_> Bug #793670: User account missing preferred email after suspension/reactivation <oops> <Launchpad itself:Triaged> < https://launchpad.net/bugs/793670 >
[21:05] <benji> gary_poster: what's the bug number for the redoing of the bug subscription overlay you were working on?
[21:05] <benji> (I can't find it on the board.)
[21:06] <gary_poster> benji bug 772754 (Feature Work: (BIG BUGS))
[21:06] <_mup_> Bug #772754: After better-bug-notification changes, list of bug subscribers is confusing <qa-ok> <story-better-bug-notification> <Launchpad itself:In Progress by gary> < https://launchpad.net/bugs/772754 >
[21:06] <benji> doh!  I've gotten so used to not looking in there that I totally missed it.
[21:07] <gary_poster> :-)
[21:08] <benji> I'm tempted to be "conversational" when chaning bug statuses or marking them duplicates, i.e., adding a comment about what I'm doing and why, but that doesn't seem to be the culture so I'm abstaining
[21:09] <gary_poster> benji, dunno.  I do it when I think it adds value
[21:10] <benji> the only "value" I was feeling was along the lines of softening the "they just marked my bug as Low, they must hate me" reaction
[21:10] <gary_poster> I have about one hour before vacation...I could help with CHR or read about CoffeeScript.  Which one sounds like more fun?  Which one should I actually do?
[21:10] <gary_poster> Yeah
[21:10] <gary_poster> I've written those
[21:11] <gary_poster> I've become more...blunt lately with the new standards
[21:11] <gary_poster> So I have not
[21:11] <gary_poster> but do what feels right
[21:13] <bac> gary_poster, benji: i have not idea what to do for amichai2
[21:13] <bac> gary_poster, benji: short of asking a LOSA to do a SQL mod to give him a preferred email address
[21:13] <benji> gary_poster: I just finished CHR (whew, it took more than an hour), so I doubt there's anything to do
[21:14] <gary_poster> benji, I was pondering https://support.one.ubuntu.com/Ticket/Display.html?id=2514
[21:14] <benji> oh, feel free handle that one ;P
[21:14] <gary_poster> bac, that's what I was thinking.  Would you feel comfortable assembling such SQL?
[21:15] <bac> gary_poster: will it fulfill my SQL goal?  :)
[21:15] <gary_poster> bac, lol
[21:15] <gary_poster> bac, it would help
[21:16] <bac> sure, i'll see what i can do
[21:16] <gary_poster> cool
[21:16] <bac> am i going to need a TL or higher to sign off on my losa request?
[21:17] <gary_poster> bac, yeah, I think TL can do it
[21:19] <gary_poster> benji, FWIW, I assigned it to Matthew Revell and sent a message to the requester and Matthew to that effect
[21:20] <benji> cool
[21:25] <gary_poster> benji, fun: today Danilo alerted me that webkit falls over if you use "class" as an argument because it is a non-functioning reserved keyword in JS.  All other browsers are fine with it.  coffeescript compilation (and repl) clearly disallows.  another win.
[21:25] <benji> heh, cool
[21:27] <benji> gary_poster: I assume that goes for all the other reserved, but unused words.  There are lots of them.
[21:27] <gary_poster> benji, I'd hope so. dunno
[21:27] <gary_poster> I'd expect so tbh.  this obviously has been around the block
[21:32] <bac> gary_poster: it looks like there is enough info at https://wiki.canonical.com/InformationInfrastructure/ISD/Docs/SSO/ErrorSettingPreferredEmail for a losa to work with
[21:33] <gary_poster> bac, oh interesting
[21:33] <gary_poster> bac, +1 I guess...assuming he has an email at all, which I hope he does
[21:33] <gary_poster> I mean, registered in LP
[21:37] <bac> benji: did you see statik talking about playing with coffeescript?
[21:37] <benji> nope, I'll look in my logs
[21:45] <bac> benji: it was only 15 minutes ago or so
[21:45] <bac> gary_poster: is there any chance we messed up with that user?  should he have gone from suspended to unactivated?
[21:45] <bac> which would require following an activation link?
[21:45] <bac> i can find no docs
[21:46] <gary_poster> bac, I went from suspended to activated, and that failed
[21:46] <gary_poster> then
[21:46] <gary_poster> sinzui said we'd had truble with that in the past
[21:46] <gary_poster> and suggested I go from deactivated to activated
[21:46] <gary_poster> so I did that
[21:46] <gary_poster> and that failed too
[21:47] <bac> did you try unactivated?
[22:03] <bac> gary_poster: the guy's email address was marked 2-validated.  we forced it to preferred
[22:03] <bac> https://launchpad.net/~amichai2
[22:03] <bac> looks much better
[22:03] <bac> Go Banana Slugs!
[22:12] <gary_poster> yay bac!
[22:12] <gary_poster> thanks
[22:13] <bac> gary_poster: the other gent was in the same boat.  i think we've got a suspend/reanimate issue here
[22:13] <gary_poster> bac, huh
[22:13] <gary_poster> maybe make a comment on the related bug?
[22:14] <bac> this guy: https://support.one.ubuntu.com/Ticket/Display.html?id=2433
[22:14] <bac> ok
[22:14] <gary_poster> yeah, I figured that was who you meant
[22:14] <bac> if the bug is not critical it should be
[22:14] <gary_poster> I think I marked it critical
[22:14] <gary_poster> bug 793670
[22:14] <_mup_> Bug #793670: User account missing preferred email after suspension/reactivation <oops> <Launchpad itself:Triaged> < https://launchpad.net/bugs/793670 >
[22:15] <gary_poster> yeah
[22:22] <gary_poster> bye
[22:22] <gary_poster> tty;
[22:23] <gary_poster> l
[22:23] <gary_poster> need to get better at this typing while walking thing :-)
=== Ursinha-lunch is now known as Ursinha