=== Ursinha is now known as Ursula === Ursula is now known as Ursinha [13:20] benji and bac, hi. are either of you looking for a bug to tackle? [13:20] gary_poster: i will be shortly [13:21] bac, cool. benjim welcome to the other channel :-P [13:21] gary_poster: tag, you're it [13:21] :-) [13:21] :P [13:23] bac, benji, just forwarded you an email [13:23] * bac just made a rule to zap vcs-import mail and can see clearly again [13:23] heh [13:23] good idea [13:24] yeah, my rule got 400+ messages when I made it; that seems like something we should fix [13:29] gmb, for bug 772609, question for the call lest I forget: what does status of bug 772609 mean to deployment? Is it worthwhile talking about whether we can make the fix on top of the branches Danilo and I are working on for after the db deployment? [13:29] bac benji danilos gmb: kanban now, call really close to now [13:29] gary_poster: The fix has been rolled back and we can happily make the fix on top of yours and Danilo's branches. [13:29] ok cool [13:30] cause I think I might have ripped out some or all of the bits you are talking about in the most recent traceback [13:30] Ah, okay. [13:31] that Skype experience was weird [13:31] benji, there's actually a bug number in there [13:32] Nice, skype won't let me sign in... [13:32] hmmmmm.... [13:32] Skype is not happy.... [13:32] * gmb restarts skype [13:32] yeah, my skype isjust spinning [13:32] Glad it's not just me... [13:33] I can't log on now... [13:33] uh, danilo, how's the mumble going :-D [13:33] my skype launched but says none of my contacts are online [13:34] I wonder how it would work on Windows :-P [13:34] bac benji danilos gmb, everyone join me on mumble who can, please [13:35] * benji installs mumble [13:35] * danilos tries and hopes for the best [13:35] meh [13:35] ok. haven't mumbled since upgrading to natty [13:36] gary_poster, nope, it's blocked [13:36] I see danilos, but he is muted and deafened [13:36] gary_poster, what it used to do was kill my entire audio stack until I reboot (maybe playing with unloading/loading modules might work, but I never spent time to debug that) [13:37] gary_poster, hum, let me try something [13:39] nope, it seems to have problems with localized sound card names that it seems to get from pulseaudio [13:39] benji, (1) how goes mumble, (2) did you happen to run test suite with your "rip out feature flag" branch? [13:39] "The remote host closed the connection." [13:40] 2) I ran -m lp.bugs [13:40] danilos we can hear you [13:43] mumble is not connecting for me === Ursinha is now known as Ursula === Ursula is now known as Ursiha === Ursiha is now known as Ursinha [13:51] empathy? [13:52] benji, try skype again? [13:52] i think you can do empathy voice-only calls but i've not tried [13:53] ok [13:53] if not i have some electrical tape i can put over my camera [13:53] lol [13:53] trying skype [13:53] k [13:53] working on email incedent too [13:53] benji, well...that's one of the things I want to talk to you about :-) [13:54] gary_poster: if need be we can have a POTS call [13:55] benji...that's unthinkable! [13:56] heh [13:56] benji, I'm logged into Skype, but I'm guessing either you can't, or they have some weird split thing going on. [13:57] mine keeps spinning like it's trying to connect [14:23] my skype seems to be fine, and I can see Gary and gmb [14:23] though, I haven't logged off since yesterday [14:23] I see you but that's it [14:23] I've logged on and off several times, and for awhile there was noone at all [14:24] interesting, have you tried calling someone even if they appear offline? [14:55] not lately [14:57] danilos, when you get a chance, would you please take a look at https://devpad.canonical.com/~gary/mockup.png and give me your opinion? I prefer B and B-1 for consistency with the structural subscription options. Option A variants are slightly more usable because of fewer clicks, but then my original was even easier in that regard. I'll ask mrevell too. [14:57] anyone else is welcome to chime in too, btw :-) [14:58] * danilos looks [15:00] gary_poster: I think the A family is slightly better because I think it'll be easier for people to choose from a linear set of options instead of a branching set (i.e., the check box) [15:02] benji, I agree about usability, what about consistency? [15:02] gary_poster, I kind of like the explicit "or" in B-1 combined with a linear set of options (i.e. in A-1 "Remove your subscription" doesn't seem distinct enough) [15:02] danilos, A-3? [15:03] gary_poster, right, it's a very wide image, haven't noticed it so far :) [15:03] :-) [15:03] danilos, the only diff is that A-3 has the current option as a radio button [15:03] while A-1 has the current option as text [15:04] gary_poster: Isn't option A consistent with how we ask about direct subscriptions? [15:04] A-3 seems the best of the A-line [15:04] I don't think we need the "or"; all of them are "or" options [15:04] gary_poster, A-3 also has the separation between "unsubscribe" and "choose level" options, no? [15:05] benji, but they are clearly actions of the distinct nature, and one might simply be looking for "unsubscribe" [15:05] danilos, yeah, sorry, I meant, A-1 with "or" is still different than A-3 because of the radio button thing [15:05] not distinct per se, but it kind of feels as if unsubscribe is the most common action [15:05] benji is arguing for the A line I think, not A-1 in particular [15:06] gary_poster, right, understood [15:07] gary_poster, fwiw, I don't mind the B-1 either, I just find it a bit confusing and hard to clearly present, but then there's the consistency argument [15:07] == hard to clearly present available options [15:07] benji, what I have right now in the branch under review (and that is regarded as confusing) is consistent with "edit your bugs": action links. I can put up a mockup in case inspiration might strike. Option B is consistent with structural subscription options. option A I'd argue is consistent with not much :-) [15:08] gary_poster, hehe, A is consistent with static +subscribe page :) not that it's relevant [15:08] danilos, lol :-P [15:09] gary_poster, I think the only problem I have with B-1 is that I'd prefer the checkbox to read "Exclude comments" if the choice above it is "everything" [15:09] maybe it's because I haven't had my coffee yet, but option A looks a lot like what I see if I visit a bug and click "Subscribe" (the words are a little different and the options are in a different order, but that's fixable) [15:09] benji, right, but that's what we are replacing :-) [15:10] benji, fwiw: https://devpad.canonical.com/~gary/PastedGraphic-1.png https://devpad.canonical.com/~gary/PastedGraphic-2.png [15:10] well then, that changes things :) [15:10] benji, because that's coming from the static +subscribe page by loading the form using ajax [15:11] gary_poster, benji: I think we should just agree that B-3 is to everybody's taste and go with that [15:11] danilos: exclude comments: yeah, that's a "too many cooks" issue we have from structural subscriptions. Jono wanted the checkbox as it is there too, because he wants checkboxes to be positive choices, not exclusions [15:11] danilos: +1 :-) [15:12] I really like the links-as-actions approach here, especially PastedGraphic-2. [15:13] heh, I did too :-P [15:13] danilos was confused though [15:14] and Huw agrees that radio buttons should rule the world :-) [15:14] (not so strongly that we couldn't override, I think, but still) [15:15] The link approach is also faster to use, as danilo pointed out [15:15] and it is consistent with "edit your bugs" [15:16] Edit bug mail, I mean [15:16] benji, radio boxes clearly indicate to anyone who's been using computer UIs that there's one selected option of all the potential choices === Ursula is now known as U === U is now known as Ursinha [15:16] benji, for instance, in PastedGraphic-2, if I didn't know the underlying implementation details, I don't think I'd ever have guessed that the middle text indicates the selected option [15:17] gary_poster, benji: perhaps it'd be much better if the "choice" (plain text) was moved to the top [15:18] That would be easy to do, and I think an improvement. [15:18] (I had it where it is so you could easily see relative subscription levels) [15:19] gary_poster, right, so my preferences would be A-3, B-1 and then that :) we know which one is the simplest to implement ;) [15:19] heh [15:20] benji, do you care to do a ranking? [15:25] gary_poster: A-1, A-2, A, B, A-3, B-3, B-2 [15:26] (my guess as to how they will perform in user testing from best to worst)) [15:27] benji, where does PastedGraphic-2 fit into that? [15:27] hmm... [15:27] and did you really mean B-3, the absent choice? :-) [15:28] yeah, you can't rank it that low! [15:28] Or do the last two have an off-by-one error [15:28] gary_poster: fixed ranking: PastedGraphic-2, A-1, A-2, A, B, A-3, B-2, B-1 [15:29] :-) ok got it, thanks [15:29] based on that, you two average out to me doing nothing at all! woo hoo! ;-) [15:30] I'll mention this to mrevell. I'm inclined to do the "move the text to the top of PastedGraphic-2 and move on" as you might expect. [15:55] hi gary_poster -- looking at the MP mbp did (https://code.launchpad.net/~mbp/launchpad/788874-mail-oops/+merge/63343) i'm not sure why it was rejected. i think lifeless was wrong in his assessment. [15:56] gary_poster: well, i *do* understand why you rejected it...i just think it is a viable solution afterall [15:58] bac, ok. yeah, I was just going on the conversation. You are the engineer on it, and I'm happy for you to make a call. I'm also happy to dig into it if you'd like me to. [15:59] gary_poster: there have been other suggested actions to take but i think martin's approach makes sense [16:07] CHR! [16:12] gary_poster: have you been running the lp2kanban script? i never see my cards updated. am i doing something wrong? [16:14] bac, no, I haven't. :-P benji said I had to change something or other in order to point to the proper board, and while I'm sure that is super easy, it has kept me from pursuing it yet. Sorry. :-( I'll get on it. [16:14] gary_poster: my mumble problems look like a networking problem; I did a packet capture and see lots of retransmissions and duplicate ACKs [16:14] gary_poster: was just curious [16:14] beji, weird [16:14] bac, cool [16:35] ok, chr done [16:35] running to anniversary lunch :-) [16:35] biab [16:57] * danilos -> out [19:24] gary_poster: you have a moment to do an inflight sanity check / pre-review? [19:24] bac, heh, sure [19:24] http://pastebin.ubuntu.com/621121/ [19:24] it should be pretty much complete....modulo me not understanding option 1 mbp presented in the bug [19:25] k [19:25] i'd like to get your interpretation of his #1 and robert's response [19:26] k, lemme get bug up [19:29] bac, I understand poolie's option 1 by omitting ", or the in-application limit up,". I'm guessing that is a mistake. [19:30] ah, so unify them at 10MB or 50MB [19:30] yeah [19:30] ok, so that is outside the scope of this bug fix...and our app should be robust enough to live in a messed up world like we currently have [19:31] that sounded more philosophical than i intended. [19:31] bac, one thought I had so far was wondering about 64K as a good default (maybe bigger), and whether it ought to be implemented in terms lpconfig [19:31] :-) [19:31] agree, though [19:32] bac, my question is a "wondering," not an opinion [19:32] actually for returning a message that caused an OOPS i think 64K should be sufficient [19:33] and i can easily make it a config variable though i don't know what it buys us since we'd not want to tune it across environments [19:33] in doc, "outgoing message will be too big" I'd maybe ike a parenthetical ref to this bug, or the incident, or at least an explanation of what is actually enforcing our limit [19:33] i mean lp.dev vs production [19:33] config var: only to change quickly [19:34] ah, right, they are handled in a different more easily deployed branch. i always forget that [19:34] OK, well if you think the branch is sound i'll make that change and get it rolling [19:35] 64K sufficient: I have no idea...I don't have a clear idea of all the ways this is used tbh [19:35] thanks for the thoughts [19:35] but I'm happy to go along with it [19:35] sure, np. it looks good to me! thank you [19:45] gary_poster: have we told whomever is doing this DB deploy to turn on the feature flags for everyone after the deploy? [19:46] benji, robert knows, and he also wants to wait an hour even with the email, which was fine w/me [19:46] even with email change I mean [19:47] you know, we could have made +subscriptions redirect to the main page if the feature flag was off [19:47] cool; in that case I'll send him a pointer to the "what to do to turn off the message" wiki page [19:47] cool thanks [19:48] yeah, that would have been good, if we'd thought of it [20:00] * benji does CHR. [20:18] gary_poster: is there a way to QA email processing before it hits production? === Ursinha is now known as Ursula [20:26] bac, was [20:27] or even, yes [20:27] looking [20:27] we talked about it in our weekly notes [20:28] bac, oh, you mean *incoming* processing [20:28] I suspect there is [20:28] yeah [20:28] I think it is something like this [20:29] gary_poster: can you do the actual review for https://code.launchpad.net/~bac/launchpad/bug-788874/+merge/63750 ? [20:29] 1) find something on production that you could email to [20:29] sure [20:29] make sure that it is also on staging [20:29] convert the prod address [20:29] by adding staging. at the start [20:29] (of domain of course) [20:30] 2) run the pertinent cronscript with -vv and hope that it is hooked up to give you the detail you need [20:30] (the one we used for our last feature was) [20:30] The End [20:33] bac, why don't we need code to rip out the original message? I only see where we add the truncated one, not where we delete the original [20:33] delete from where? [20:34] you'll note i delete it from the mailbox before doing the send-oops [20:34] line 139 [20:34] gary_poster: ^^ [20:34] bac, when we assemble the message to send, leading up to line 62 of diff [20:35] line 60 redefines original_msg to be the truncated version [20:35] * bac admits that is a little confusing [20:35] oh, duh, got it [20:36] renaming truncated_msg to truncated_msg_str *might* have helped me. [20:36] but probably not :-P [20:37] approved bac [20:38] good idea [20:38] thanks [20:38] np [21:51] benji, you around? [21:52] gary_poster: yep [21:52] hey [21:52] bug 772609 [21:52] <_mup_> Bug #772609: bug subscription mute link is not shown for membership in a team with a direct or structural subscription

< https://launchpad.net/bugs/772609 > [21:52] I take it gmb suggested that you not focus on it right now? [21:53] benji: ^^ [21:54] gary_poster: I don't recall a conversation like that. Oh, I remember, I claimed it in LP and he already had put a card up for it previsouly. [21:54] benji, you are not assigned it in LP either [21:54] you are working on it? [21:55] I need to know fairly quickly--maybe quick Skype call? [21:55] benji ^^ [21:55] nope, I unassigned myself and he assigned himself [21:55] oh [21:55] I'm up for a call. [21:55] (but I think that's the extent of my knowleg (sp) on it) [21:56] and you aren't on Skype :-P [21:56] ok