=== Ursinha is now known as Ursinha-afk [11:53] * gmb lunches [13:21] danilos, hey. Can I take "UI: unmute does not show dialog, just unmutes"--that is, is the kanban board accurate? If so, what branch should I use as a base--is "Provide BugMute table and use it for muting bug subscriptions" in db-devel or is the kanban accurate again, and I need to get a branch? [13:22] Ah-ha [13:22] looks like I need to branch https://code.launchpad.net/~danilo/launchpad/proper-bug-muting [13:22] gary_poster, you could, but it'd be even better to branch off my nothing branch [13:22] ...uh...wha? [13:22] gary_poster, I've got one test to take care of there and it's going up for review [13:22] oh, your nothing branch! [13:23] heh, got it now [13:23] lp:~danilo/launchpad/bug-772763-remove-nothing [13:23] gary_poster, heh, yeah :) [13:23] cool thanks [13:24] gary_poster, I hit a bunch of test failures in the proper-bug-muting branch, fwiw, some of them related to person merging [13:24] ah, yeah, that's a pain [13:24] but resolved now, it looks like [13:24] gary_poster, yeah, up in ec2 again (just in case) [13:24] :-) cool [13:26] bac benji danilos gmb, kanban now-ish, skype in 4 [13:33] https://bugs.launchpad.net/launchpad/+bug/783948 [13:33] <_mup_> Bug #783948: Entering a comment returns an error < https://launchpad.net/bugs/783948 > [13:33] https://bugs.launchpad.net/ubuntu/+source/bash-completion/+bug/752193 [13:33] <_mup_> Bug #752193: Installation of the acroread package causes completion to treat directories like files with some commands

< https://launchpad.net/bugs/752193 > [13:58] gary_poster, benji: Am I right in thinking that a LocationError from TAL basically means "You're trying to access an attribute that doesn't exist"? [14:00] gmb: not normally; it's more often a problem finding out where an object is in the url space; e.g., it doesn't have a __parent__ [14:00] yeah [14:00] Hmm. [14:01] Here's the salient bit of the OOPS: [14:01] LocationError: (, 'current_user_mute_class') [14:01] I'm trying to figure out what that actually means in this case. [14:01] is there a traceback as well, or is that eaten somehow? [14:01] gary_poster: It's all here: https://lp-oops.canonical.com/oops.py/?oopsid=OOPS-1963BC361 [14:03] gmb, your initial description was correct [14:04] Ah. [14:04] Now that's very odd. [14:04] as I expect you saw, [14:04] attr = getattr(subject, name, _marker) [14:04] if attr is not _marker: [14:04] return attr [14:04] if hasattr(subject, '__getitem__'): [14:04] try: [14:04] return subject[name] [14:04] except (KeyError, TypeError): [14:04] pass [14:04] raise LocationError(subject, name) [14:04] Saw, but didn't pay enough attention to :) [14:04] :-) [14:05] Hmm. [14:06] this may be one of those transitive attribute errors...sometimes they explain surprises like this [14:06] if current_user_mute_class is an @property [14:07] * gary_poster looks forward to having Python use mwh's wrapped exceptions, from whatever version of Python that is in [14:09] gary_poster: That wouldn't explain why the page still OOPSes for Gavin and I but not wgrant, though. [14:09] gmb, ForbiddenAttribute and Unauthorized both subsclass AttributeError, IIRC. Maybe pertinent? [14:10] Ah, true. [14:10] Hmm. [14:10] * gmb pokes [14:10] They would be eaten by the code above [14:10] Yes. [14:10] and converted into LocationError [14:10] Quite so. [14:10] So, let's see if I can break it. [14:10] Muhahaha, etc. [14:11] heh [15:19] danilos, I'm ready anytime for our call. This is 11 minutes early, so no rush. ;-) Give me a Skype call when you are ready. [15:26] gary_poster, sure thing, I am ready myself [15:27] cool [15:28] gary_poster: I haven't been able to break anything and I've now been sucked into an OCR / helping out a contributor conversation. I think it might be a good time to roll back my last revision since that seems (to me) like a likely culprit. [15:52] gmb, ok, please do [15:53] On it now. [16:44] gary_poster: I can confirm that the rollback fixed the problem. [16:45] Now, that doesn't mean that it was actually my revision that caused the problem (we rolled back from 13050 to 13045 for logistical reasons) but it seems likely. [16:57] gmb, just saw this, but mostly good. :-) [16:57] Yeah. [16:57] Trouble is reproducing it. [17:35] gmb, thanks for the review [17:35] * danilos -> off [17:35] np [18:21] hi gary_poster, benji: on this page https://bugs.launchpad.net/ubuntu/+subscriptions there is a help icon next to the "Stop my emails" link. but we go to great pains to make it appear and disappear on hover. why? [18:21] i think that link only appears for team subscriptions and perhaps ones for teams you don't administer [18:24] bac, I don't remember specifically why I did it that way; I think it was to reduce confusion because of the number of other icons visible [18:24] bac: why: the number of icons were becoming problematic. The hover was a problematic solution that benji prototyped up and that we felt was an improvement. [Link should appear for team subscriptions, unless team has a contact address (that last part is in progress).] [18:25] ok. fwiw it seemed like a browser bug when i first encountered it. [18:25] heh [18:26] making it appear gradually would reduce that impression, but not remove the underlying discoverability problem. [18:37] gary_poster: bug 778323 [18:37] <_mup_> Bug #778323: On the subscription page, the help popup button opens up a 404 page. < https://launchpad.net/bugs/778323 > [18:37] gary_poster: so the problem is the help file only exists for bugs.lp.net [18:37] bac, good one. Sorry, I forget to do the bug review for you :-( [18:37] bac, is that one that we do a symlink for? [18:37] gary_poster: but i think people are hacking urls to go to http://launchpad.net/ubuntu/+subscriptions [18:38] gary_poster: yes, i can solve it by putting a symlink in registry/help...but if they use code.lp.net or answer.lp.net it'll break. :( [18:38] so, perhaps the right thing to do is to force a redirect to the bugs rootsite? [18:38] oh! [18:39] interesting [18:39] or change the way +help works [18:39] admittedly, coming via one of the other rootsites is pretty rare, i'd guess [18:39] bac, I'm ok with whichever is easier. The redirect would be what I would tackle, probably. [18:40] yeah, so how would you do that? [18:40] the zcml seems to only configure that page for the bugs rootsite, so i'm a bit confused why the others work [18:41] Well, I would have expected to simply di a redirect during the initialize stage [18:41] do [18:42] I'm not sure about the zcml. I can stare at it with you, if you like; it's LP zcml machinery, I think, not ye olde Zope stuff [18:43] that's ok. i'll bug you if i get stumped [18:44] cool [19:29] bac, ready for call whenever you are [19:29] gimme a call when you want [19:29] gary_poster: ok. just a sec. [19:29] np