=== chihchun_afk is now known as chihchun [05:23] Yay [06:44] mardy: qtbase fix has migrated to xenial-release! (among with my previous qtbase upload, which was also stuck in proposed) [06:44] it seems the mysql transition was finished [07:49] Mirv: excellent, thanks a lot! [08:47] ubuntu-qa: any chance silo 38 will be validated before EOW ? [08:56] oSoMoN, it will [08:57] thanks! === chihchun is now known as chihchun_afk [10:20] oSoMoN, I'm trying silo 38 and verifying bug 1534112. I downloaded 4 files, when I go to downloads I see them all [10:20] bug 1534112 in webbrowser-app (Ubuntu) "Can select multiple files in downloads view when only one expected" [Medium,In progress] https://launchpad.net/bugs/1534112 [10:21] oSoMoN, but when I follow the steps in the bug report, I see only one [10:21] oSoMoN, is there a restriction on which file is available for the selection? [10:22] jibel, probably because that webpage accepts only images, you’ll need to download at least another image (or use another webpage that presents a file picker that accepts all file types) [10:22] oSoMoN, right, only one was an image, just verified while I was typing :) sorry [10:23] nw :) === chihchun_afk is now known as chihchun [10:31] oSoMoN, when I enter the address bar with a keyboard shortcut should I be able to modify the content? [10:31] oSoMoN, I can only select an entry from the list [10:32] jibel, what do you mean? you press Ctrl+L to focus the address bar, and you cannot type a new address? [10:32] oSoMoN, for example in FF I type CTRL+L, the address bar gets the focus and I can modify the url [10:32] oSoMoN, I cannot do that with the webbrowser app [10:32] I cannot modify the url [10:32] just select one from a list [10:33] oSoMoN, it is with a BT keyboard [10:33] jibel, huh that’s unexpected indeed [10:34] oSoMoN, it is a regression with the sio [10:34] silo [10:35] jibel, I don’t have my bt keyboard handy (not at home atm), will test this afternoon, in the meantime I guess the silo should be marked blocked [10:35] oSoMoN, I verified on arale which as current version and I can update the address [10:35] has* [10:36] oSoMoN, blocked, waiting for your feedback [10:36] jibel, thanks (and sorry about that) [10:37] np === chihchun is now known as chihchun_afk [11:32] dbarth: https://bugs.launchpad.net/canonical-devices-system-image/+bug/1567377 [11:32] Launchpad bug 1567377 in webbrowser-app (Ubuntu) "Select text popup not localized" [Undecided,New] === alan_g is now known as alan_g|lunch [12:11] sil2100, treat my +1 as a community review :) === _salem is now known as salem_ [12:24] rvr: ack [12:38] jibel, I’m not seeing the issue you’re describing on https://trello.com/c/GxLM9Hj7/3032-1196-ubuntu-landing-038-webbrowser-app-osomon . I’m testing on arale with latest rc-proposed + silo 38, when I press Ctrl+L on my BT keyboard the address bar is focused, and if I then type any letter it is correctly being input in the address bar [12:38] or was the issue a different one? [12:38] oSoMoN, no it is the issue. I'll try again [12:39] oSoMoN, it was on krillin but I don't think it would make a difference [12:39] it really shouldn’t [12:45] oSoMoN, weird. now it works. [12:46] heisenbug [12:46] oSoMoN, I didn't have a cursor in the text field previously [12:46] oSoMoN, I'll finish the verification and see if it happens again [12:46] jibel, thanks [12:49] oSoMoN, ah I reproduced the problem [12:51] oSoMoN, CTRL+T to open a new tab, select a page in the recently visited pages, CTRL+L [12:51] oSoMoN, the address bar is not editable anymore then [12:56] oSoMoN, it doesn't happen with current version of the webbrowser [12:56] oSoMoN, can you confirm? === chihchun_afk is now known as chihchun [13:04] jibel, looking === alan_g|lunch is now known as alan_g [13:06] jibel, I’m seeing it indeed [13:06] jibel, let me check on rc-proposed without silo 38 [13:19] jibel, this looks very similar to bug #1545802 , I wonder if it could be the same issue [13:19] bug 1545802 in ubuntu-ui-toolkit (Ubuntu) "activeFocus not being forwarded to TextInput inside TextField" [Undecided,New] https://launchpad.net/bugs/1545802 [13:35] oSoMoN, indeed, it is very similar [13:36] jibel, I’ll look into it a bit later today [13:36] oSoMoN, thanks. I failed the silo [13:36] jibel, thanks for the thorough testing :) [14:08] trainguards, can you please retry https://launchpad.net/~ci-train-ppa-service/+archive/ubuntu/landing-071/+build/9547402 ? thanks [14:09] Saviq: done! [14:15] plars, looks like using ssh-over-adb is just not gonna work https://unity8-jenkins.ubuntu.com/computer/arale-01/builds https://unity8-jenkins.ubuntu.com/computer/krillin-07/builds - 1 of 2 seems to break the connection :/ [14:25] Saviq: is it always with that protocol fault error? [14:25] Saviq: I thought you hit something else yesterday [14:25] plars, it doesn't look exactly the same every time [14:26] most often adb doesn't say anything [14:26] since its output is hidden by adt-run I think [14:26] Saviq: I think alesage uses it in his tests also - alesage are you seeing errors like that in your test runs? [14:26] alesage: Saviq seems to be hitting issues now where adb seems to die in the middle of his test run [14:27] but almost every time it's "Connection to localhost closed by remote host" [14:27] so either adb forward dies, or device dies... [14:29] Saviq, rhuddie will remember more about this but we're having to retry a few times on initial connection to device [14:29] IIRC (similar to what happens from you laptop) [14:29] initial connection actually works for me most of the times, but it drops after a while [14:29] Saviq, o hmm--this is on arale? [14:30] we're also seeing cut-offs on arale [14:30] alesage, both arale and krillin [14:30] Saviq, ok then, is this recent? I haven't checked our runs closely over last 2 days [14:31] alesage, I only started paying more attention recently, but we've had similar issues before [14:31] you can see in https://unity8-jenkins.ubuntu.com/computer/krillin-07/builds how many of them are red - most often than not it means the test run was interrupted like that [14:32] Saviq, stand-up time, will come back and dig in a bit [14:33] Saviq: yeah, https://unity8-jenkins.ubuntu.com/job/test-0-autopkgtest/label=phone-armhf,release=vivid+overlay,testname=autopilot.sh/404/console is definitely before the adb update, and for some reason it looks like adt couldn't connect. But as you say, it's not really clear why [14:36] plars, yeah, it looks like the adb forwarding died [14:37] plars, could the devices have resolvable hostnames like arale-01, krillin-07? we could try and see if connecting over wifi helps [14:41] Saviq: unlikely - I'll have to check with IS. Right now they all just attach to a WAP that IS runs and I have no control over. So my guess is that they all just get a randomly assigned ip from it. Worse - it's probably on a different network, so I'm not sure if you can even reach it but I'll check [14:42] ugh [14:42] makes sense, security-wise... [14:44] Saviq: I'm not even sure if adt could handle that... doesn't it need to start ssh over adb first? [14:45] Saviq: and confirmed - I can't seem to reach the network the phones attach to with wifi [14:45] plars, well, I'd have to modify adt-virt-ssh-adb or use adt-virt-ssh directly after having set up ssh first [14:46] plars, on one hand I feel like the multitude of devices connected to the same adb server could make it go bad like that... but then again you guys have been doing that for quite a while already, before jenkaas was born [14:47] plars, whatif Saviq is just on an unlucky router, e.g.? [14:47] alesage: he's on the same one as you [14:47] or cable [14:47] OTOH maybe it wasn't using ssh-over-adb then, so didn't require a continuous connection [14:47] Saviq: yeah, I don't think we've had issues like that before. Which is why I was hoping to see if qa had problems like that [14:48] Saviq: both has been done in the past. For proposed migration, it all used adt [14:48] o yes we can't get a reliable run from our arales, sounds very much the same [14:48] plars, I'll add some debugging to my jobs - see if adbd's PID changed, what forwards are enabled (if any) etc. [14:48] Saviq: for image smoke tests - it was just over adb directly. We snarfed the return value of things another way and didn't rely on ssh over adb [14:48] Saviq: that would definitely be interesting to see [14:49] maybe it will give us some ideas (anything in particular would be useful?) [14:49] adb devices | grep $ANDROID_SERIAL; ps aux | grep adbd; adb forward --list [14:49] plars, Saviq what would an "adb connection health check" test look like? [14:49] maybe just running something repeatedly over adt-run and waiting for a failure? [14:49] Saviq: don't grep for adbd (that's the piece on the device) grep for fork-server instead [14:50] alesage, in theory if flashing completes, that's the best you can hope for [14:50] I mean you've pushed several hundred MBs over adb [14:50] Saviq, fair, but time may be a factor [14:50] trueth [14:50] plars, right [14:50] * Saviq will tail dmesg and syslog, too [15:11] jibel, I’ve confirmed that the regression you’re seeing with silo 38 is bug #1545802 , now I need to figure out a fix for it… [15:11] bug 1545802 in ubuntu-ui-toolkit (Ubuntu) "activeFocus not being forwarded to TextInput inside TextField" [Undecided,New] https://launchpad.net/bugs/1545802 [15:23] plars, hah, so here, for example https://unity8-jenkins.ubuntu.com/job/test-ppa-autopkgtest/label=phone-armhf,package=unity8,release=vivid+overlay,testname=autopilot.sh/76/console network provisioning failed - maybe I need to give it more time or something [15:26] Saviq: yeah, I seem to recall that network provisioning was always kinda flaky. I think we always made it retry a few times, and after that, we turned off the welcome wizard, enabled writable-image, turned off edges intro. One of those triggers a reboot I think too [15:27] Saviq: there's a LOT that has to happen to prepare things for automatically testing. I think adt is supposed to do some of that though [15:59] jibel, and in fact the exact same issue that you’ve observed with silo 38 already exists on tablets, silo 38 just makes it more obvious by exposing it to phones too [16:02] jibel, I could revert the one change that exposes the issue to phones, would that make the silo acceptable for landing (knowing that the issue still exists on tablets)? === chihchun is now known as chihchun_afk [16:03] oSoMoN, works for me [16:07] sil2100: meeting? [16:07] sil2100: meeting === alan_g is now known as alan_g|EOD [17:36] trainguards, can you please add dandrader to ci-train-users? thanks :) === marcusto_ is now known as marcustomlinson [18:10] Saviq: done. I'm afk, can you link him to the docs? [18:12] robru, ack === salem_ is now known as _salem