=== chihchun_afk is now known as chihchun === chihchun is now known as chihchun_afk [08:02] trainguards, I would like to have https://requests.ci-train.ubuntu.com/#/ticket/548 landed to wily overlay [08:04] tvoss: if you set the destination ppa to the overlay you can click publish on it [08:05] tvoss: there's a dropdown menu you can select the overlay from, just double click in the field [08:05] robru|sick, ah, so wily + overlay ppa [08:05] cool, let me try to hit publish :) [08:05] tvoss: looks good [08:06] ah, I'm not authorized to upload :) [08:06] Ooooh [08:06] tvoss: you need a core dev then, I'm afraid i can't help you there [08:07] robru|sick, ack and thx [08:07] tvoss: you're welcome [08:07] robru|sick, and get well soon :) [08:08] tvoss: thanks, will sleep it off tonight === chihchun_afk is now known as chihchun [09:01] sil2100, re: translations, I don't think there's really any other way, translators should use rc (if we start using it as actual rc), we need a string freeze @ which point we release to rc, give translators a few days to do translations and start testing then [09:03] Saviq: yeah, but I don't want to force translators to have to use rc even at best... since even though it's safer than rc-proposed, the first images we build in rc will be untested, so they could have some issues on their main phones [09:03] sil2100, how else are they supposed to see new, untranslated, or changed, strings [09:03] sil2100, otherwise you release untranslated images to users [09:04] sil2100, I agree with Saviq there is no other way than using a pre-release image [09:04] sil2100, if we treat translation changes as "safer", you could update just the langpacks post-testing [09:05] we could also publish a list of new strings in the coming ota [09:05] but it would mean keeping track of string changes [09:05] seb128, well, should be easy to get .po diffs [09:05] seb128, but, it's not just about the strings [09:05] seb128: we already do something like that, since LP has the new strings earlier - the problem is that they do not know the context of some of the strings [09:06] exactly, or how much space they have [09:06] seb128, right but without context and the layout on the phones it's hard to figure out what the translation should be. The physical space on the screen is a hard constraint [09:06] there is not one phone and screen size though [09:06] so that's not really an argument [09:06] well, 40GU is our smallest target [09:07] right, so now you change from "test on your device" to "test on a 40GU" [09:07] ideally the emulator would make that easy [09:08] if it wasn't as slow as it is... but sure, I treat the emulator == device [09:09] but I'm probably slightly different than community folk just trying to help, not brick their phones [10:22] mardy: ping [10:30] rvr: pong [10:30] mardy: Hey [10:30] mardy: Silo 2 [10:30] mardy: «I have created a Google account. I have granted access in System Settings > Accounts to both YouTube and Contacts. Two issues: YouTube scope doesn't hide the "Log in YouTube" button» [10:31] mardy: «Clicking in "Import contacts from Google" opens the Google login dialog to introduce the credentials.» [10:32] rvr: brb, got to reboot [10:36] rvr: ok; the second one seems normal [10:36] rvr: what happens when you tap on the Log in Youtube button? [10:37] mardy: Even though the account has been created and permission granted? [10:37] rvr: wait, does it asks you to authorize a service, or to enter your credentials? [10:38] mardy: Yes [10:38] mardy: The button doesn't do anything, apparently [10:38] It reloads the page, but doesn't go away [10:42] rvr: can you please try without the silo? [10:45] brendand, kudos on picking up the behemoth silo ;), let me know please if you have any questions [10:45] Saviq, np i'm currently filling my armory [10:46] let's see, assault rifle, check. grenades, check. 8 inch hunting knife, check [10:46] body armour, check [10:47] :) [11:05] trainguards, https://requests.ci-train.ubuntu.com/#/ticket/445 somehow got SILO DIRTY for unity8, even though I rebuilt it yesterday and no one released unity8 in the mean time [11:06] oh ok, duflu merged trunk... bad timing [11:06] is this going to be a problem for publishing? [11:07] Saviq: silos now can also be marked as dirty if someone pushes a new commit to any of the branches - anyway, so did anyone release unity8 in the end, or is it that case right now? [11:07] sil2100, one of the branches got a trunk merge after it was built in the silo, so there's no change really [11:08] sil2100, we can safely ignore that commit [11:09] sil2100, just not sure what the result's gonna be when publishing (should I WATCH_ONLY?) [11:22] Saviq: hmm, good question, might cause some problems [11:23] Saviq: try watch-only, but with all the recent changes that Robert made I'm not sure if watch-only does what it was doing in the past anymore === _salem is now known as salem_ === olli_ is now known as olli [11:33] mardy: Same without the silo [11:33] rvr: that's kind of good news :-) [11:34] sil2100, it'll likely just go back to dirty the next time it does the checks... but let's try [11:34] rvr: rc-proposed? [11:34] mardy: Good news for the silo, but it's not working as expected [11:34] mardy: Yes [11:35] If I granted permissions to Contacts, why does it ask again? And YouTube scope seems broken. [11:35] rvr: since it's already landed, can you please file a couple of bugs? [11:36] rvr: one is the local permission to use the account, the other one is Google request to authorize the app -- but I'll check [11:36] sil2100, actually, watch_only didn't clear the flag, let's see what happens [11:37] mardy: In Contacts, the dialog to fill the Google account credentials is displayed [11:38] rvr: mmm... I tried just now: deleted the account, re-created, enabled contacts from system settings: [11:39] rvr: the transfers indicator showed an error, with "tap to retry" suggestion [11:39] rvr: I tapped, it retried, and it worked, with no question asked [11:39] rvr: can you tell me exactly your steps? [11:39] mardy: 1. System Settings > Accounts > Create Google Account. [11:40] mardy: 2. Edit account to grant permissions to Contacts and YouTube [11:40] mardy: 3. Open Contacts, click on Import contacts from Google [11:41] mardy: At this point, the Google login page appears [11:43] rvr: don't you get the transfer icon in the indicators, as soon as you enable Contacts in the account? [11:45] mardy: Yes [11:45] mardy: But no error === chihchun is now known as chihchun_afk [11:48] * Mirv freeing up 10GB of PPA space :) [11:52] rvr: that's weird, so there is a working sync after creating the account, but it doesn't work from the calendar app... [11:52] rvr: did you install the first version of buteo landing, which was later reverted? [11:53] mardy: I'm latest using rc-proposed [11:53] I'm using latest rc-proposed [11:53] rvr: yes, but I mean, if you once installed that (broken) silo, the accounts.db file might still contain wrong infos [11:54] rvr: unless you wiped the device [11:56] mardy: The device is wiped for each silo [11:59] rvr: ok, then please can you try again while running this: OAU_LOGGING_LEVEL=2 OAU_DAEMON_TIMEOUT=9999 online-accounts-service === alan_g is now known as alan_g|lunch [12:07] mardy: http://paste.ubuntu.com/12893637/ [12:08] mardy: That's without the silo [12:13] https://bugs.launchpad.net/ubuntu-system-settings-online-accounts/+bug/1508903 [12:13] Launchpad bug 1508903 in Online Accounts setup for Ubuntu Touch "[Contacts] Permission dialog shows in already granted account" [Undecided,New] [12:14] rvr: you know what? it's not using the same account, it's actually creating another account [12:15] mardy: I suppose it's not recognizing the existing one [12:19] rvr: can you also paste the output of "account-console show ", where the google account id is returned by account-console list? [12:21] mardy: https://pastebin.canonical.com/142430/ [12:23] rvr: thanks; looks like this is a regression in calendar, I'll comment on the bug [12:23] mardy: In calendar? [12:24] rvr: contacts, calendar, same thing ;-) [12:24] mardy: Ah, ok [12:33] mardy: The account tester app is failing. "Error 4" [12:34] mardy: signon-apparmor-extension is installed [12:39] rvr: are you testing the silo 2 now, or still on vanilla rc-proposed? [12:44] rvr: in any case, could you please try again with OAU_LOGGING_LEVEL=2 OAU_DAEMON_TIMEOUT=9999 online-accounts-service ? [12:55] mardy: Silo :) [12:57] mardy: Forget it, it was an expected error === alan_g|lunch is now known as alan_g [13:20] are these issues known about? http://paste.ubuntu.com/12893982/ [13:21] brendand: hm, no, I don't think so [13:22] brendand: is that with proper PPA pinning? [13:34] mardy: Approving silo 2 [13:34] mardy: I filled another bug https://bugs.launchpad.net/ubuntu-system-settings-online-accounts/+bug/1508935 [13:34] Launchpad bug 1508935 in Online Accounts setup for Ubuntu Touch "[Account tester] Test case password-query fails" [Undecided,New] [13:36] renatu: ping [13:36] rvr, pong [13:36] renatu: om26er has been trying to install silo 55 in OTA7 without luck [13:36] renatu: https://trello.com/c/Sof1bEKA/2397-540-ubuntu-landing-055-address-book-service-bfiller [13:37] renatu: Should be installed in rc-proposed instead? [13:37] OTA7 lacks dependencies [13:37] rvr, did we update evolution-data-server libraries? [13:37] renatu: I don't know [13:37] Version of libstdc++6:armhf on system is 4.9.2-10ubuntu13. [13:38] it is complaning about evolution-data-server libraries and stdc++ versions [13:39] rvr, I do not have controls over these libraries, I can try push a new build on the silo to check if it will fix the problems [13:39] renatu: I could install it in rc-proposed [13:39] renatu: But om26er was trying to do it in OTA7, as explained in the merge proposal [13:40] rvr, do I need to change something in the silo to create packages for OTA7? [13:40] jibel: hey. I don't see cachio here but request 531 / silo 014 was marked as Passed but the ticket was not updated. [13:41] rvr, I do not know why the silo is creating packages incompatible with the libraries on OTA7 [13:42] renatu: Hmm [13:42] renatu: I'm going to take look to it. Why is OTA7 a requirement? [13:42] Mirv: I see [13:42] rvr, OTA7 is not a requirement [13:42] Mirv: Maybe he has no bileto powers [13:42] Mirv: Approving [13:42] rvr, it will be land on OTA8 [13:43] rvr: thank you! possibly so. [13:45] rvr, ok I see the test plan ask to install OTA6. I will update that. [13:46] renatu: Ok [13:46] rvr, done [13:46] robru|sick: Can you add cachio to bileto (qa-team)? [13:47] rvr: I can do that if needed :) [13:48] sil2100: Yes, please :) [13:50] renatu: I'm reading the steps here https://code.launchpad.net/~renatofilho/address-book-service/fix-facebook-account-import/+merge/275076 [13:50] renatu: * Install OTA7 or older image on phone [13:50] renatu: Step '''Install the last version of the package "buteo-sync-plugins-contacts-google" and "address-book-updater"''' ... is this equivalent to install the silo packages? [13:51] rvr, I added a extra step: Flash proposed image; [13:51] :D [13:52] renatu: Same URL? [13:52] yes [13:54] Still see the same steps [13:54] rvr, for me it is showing: [13:54] Open system settings and create one account (Account D): [13:54] Make sure that contact sync is disabled for this account [13:54] Flash proposed image; [13:54] Install the last version of the package "buteo-sync-plugins-contacts-google" and "address-book-updater" [13:54] reboot the phone [13:55] renatu: Flash without wiping? [13:55] yes [13:56] let me update that [13:56] Weird, in https://code.launchpad.net/~renatofilho/address-book-service/fix-facebook-account-import/+merge/275076 I still see the old steps [13:56] om26er: You? ^ [13:57] rvr, yeah they are the same as they were previously [13:57] om26er: renatu has updated the steps, see above [13:58] rvr, ok sorry I have been updating the test plan page: https://wiki.ubuntu.com/Process/Merges/TestPlan/buteo-sync [13:58] rvr, I will update the MR description [13:58] ;) [13:58] renatu: Aaaaaahhhhhh [13:59] renatu, is buteo-sync-plugins-contacts-google not installed by default ? [13:59] om26er, yes they are on proposed [13:59] barry: ping [14:01] renatu, hmm, so I don't need to install it then ? or do I have to downgrade to the previous version of this package ? [14:02] om26er, you need to install the packages on silo [14:03] renatu, the silo does not contain buteo-sync-plugins-contacts-google it only contains address-book-service [14:05] renatu, https://launchpad.net/~ci-train-ppa-service/+archive/ubuntu/landing-055 [14:06] yes in this case just install the silo [14:06] I will update the description [14:47] popey: oh I like it [14:48] wat? [14:48] popey: I thought well that bug number one not fixed and then I tapped the today and up it scrolled very swish [14:48] \o/ [14:51] and the bottom one pops it's details up rather than down that's 2 out of 2 woohoo! now to find some new issues ;) [14:52] brendand, how's your armour, holding? need a med-pack? [15:00] popey: found a bug [15:01] popey: 19 kph is not 19 mph for wind [15:01] popey: 19kmh is about 12 mph [15:02] kph even [15:05] davmor2, not more issues! ;-) [15:05] ahayzen: flick between mph and kmh surely the numbers should change ;) [15:05] they do for me.. [15:06] kph even [15:06] which provider and is that the day delegate or the today info ? [15:06] davmor2, oh you mean the numbers would change good point! [15:06] ahayzen: both providers on the default location page for any day including today [15:07] ahayzen: indeed 19 kph != 19mph more like 12mph [15:07] davmor2, haha who wrote this! [15:07] * ahayzen hides [15:07] ahayzen: ermmmmmm I forget his name [15:08] bzr blame will tell us ;-) [15:08] ahayzen: I just blame popey ™ any way [15:09] davmor2, right investigating, you can report a bug and assign to me if you want :-) === Ursinha_ is now known as Ursinha [15:12] davmor2, haha, so if you change the temperature from C to F i bet the wind value will change ;-) [15:13] davmor2, have you seen brendand? I'm worried silo 22 kiled him... === Mirv__ is now known as Mirv [15:13] Saviq: bath is a long way from wolverhampton dude ;) [15:14] ahayzen: what muppet wrote that ;) [15:14] :-) [15:16] davmor2, not sure what it has to do with the fact you don't take regular baths [15:16] I work from home, too === Chipaca` is now known as Chipaca [15:17] Saviq: yes an I can't see you either :P === cjwatson_ is now known as cjwatson [15:18] * Saviq wipes the camera lens [15:19] Saviq: oh dude seriously put some clothes on no one needs to see that. /me buys some more mind bleach [15:26] slangasek: hey! I see the xenial series is up on LP, feel free to do the wily-overlay batch copy once you see the time is right and give us a sign to re-target the CI Train [15:27] sil2100, Hi! [15:27] sil2100: ah, we should probably do that asap [15:27] sil2100, I need to bless https://requests.ci-train.ubuntu.com/#/ticket/540 [15:28] sil2100, I would probably need permissions to do that, can you give me permissions ? [15:28] om26er: sure! === dbarth_ is now known as dbarth [15:29] Saviq: there you go resurrected brendand for you [15:29] Saviq: don't be killing him again I only have the one potion available till I complete the next quest [15:30] om26er: added! [15:30] rvr: oh and I almost forgot about your request, added that person now - sorry for the delay! [15:30] sil2100: Thanks [15:34] jibel, robru|sick, rvr, davmor2: anything to discuss besides celebrating wily release? [15:34] sil2100, nothing from me === ssweeny` is now known as ssweeny [16:01] davmor2: http://people.canonical.com/~alan/weather/com.ubuntu.weather_3.0.150_all.click for you - fixes your bug [16:02] rvr, why is 23 still blocked? [16:02] jibel: barry haven't pinged back [16:02] barry, ^ [16:02] hasn't [16:03] barry, there is a comment on this card https://trello.com/c/lGitnxvO/2373-514-ubuntu-landing-023-system-image-barry [16:03] ahayzen: https://bugs.launchpad.net/ubuntu-weather-app/+bug/1508998 [16:03] Launchpad bug 1508998 in Ubuntu Weather App "Numbers are not recalculated on unit change for wind speed" [Undecided,New] [16:04] popey: oh nice [16:04] davmor2, already fixed :-P [16:04] davmor2: added link to the citrain task [16:05] popey: thanks === dpm is now known as dpm-afk [16:22] davmor2: give us a ping if you find any more :) [16:26] jibel: i apparently can't log into trello any more. getting a pw reset [16:32] popey: will do [16:36] rvr: do you have a few minutes to talk about the test b failure? [16:36] barry: Sure [16:37] rvr: what does it mean "s-i-cli doesn't unset auto downloads?" [16:38] barry: system-image-cli is self-consistent, it gets and sets its autodownload status correctly [16:38] barry: but is not consistent with System Settings [16:39] barry: When I unset autodownloading with system-image-cli, System Settings says otherwise [16:40] rvr: i guess that's a bug in system-settings? [16:40] ahayzen, popey: that's better :) [16:40] davmor2, :-) [16:41] rvr: who owns system-settings these days? [16:41] barry: Ask kenvandine [16:41] ahayzen: how can tell if the refresh has happened? [16:42] davmor2, when the spinner stops spinning? or the bouncing stops bouncing ? [16:42] or if the data changes [16:42] davmor2, you could load you cache folder with the data from the autopilot tests [16:42] then you would see it all change [16:43] ahayzen: I mean for the refresh interval to test it is actually refreshing [16:43] ah hmm [16:43] change the time on your clock ? [16:43] rvr: okay, while waiting for kenvandine to respond, let me ask about the other problem i think you reported. you're not seeing updates because of another system-settings issue, or is that a system-image problem? [16:43] ahayzen: is there a timestamp on the data? [16:44] davmor2, not sure think that was hijacked from the old app, let me check [16:45] barry: Well, I don't know whether the problem is in System Settings or in system-image, but I did "Wait for it to say there is an update available" and in the main screen, I couldn't see the new image notification [16:46] rvr: did you explicitly check for updates? [16:46] barry: I followed the test plan [16:46] davmor2, hmm i'm not sure if that auto-refreshes or if when you restart the app it then uses that to determine if its loading from the cache or if it needs to go to the web [16:47] barry: https://wiki.ubuntu.com/Process/Merges/TestPlan/ubuntu-system-image#Test_B_--_Good_Path_Automatic_Download_Current_Packages [16:47] barry: As I didn't get the notification, I stopped to ask whether that was expected or not [16:47] ahayzen: okay I'll have a look at the data see if it has a timestamp and play a bit [16:48] rvr: note that for s-i 3 you only follow the first Test A and B. i suppose i should relabel those so it's not confusing [16:48] rvr: but i will also double check on my phone [16:48] davmor2, this is the bit of code it hits if its any use http://bazaar.launchpad.net/~ubuntu-weather-dev/ubuntu-weather-app/reboot/view/head:/app/data/WeatherApi.js#L771 [16:49] barry: Oh, right, don't worry I was using test plan for system-image 3 [16:49] It's clear [16:49] rvr: cool [16:49] ahayzen: thanks I'll have a look after [16:50] rvr: i'm on ubuntu-touch/rc-proposed/bq-aquaris.en r149 [16:50] rvr: let me install si 3.0.2 from the silo and test it [16:50] barry: Ok [16:55] ahayzen: yeap looks like it updates on close and reopen if it passes the time selected by the look of it but I'll have a confirmation in 8 minutes [16:55] davmor2, sweet :-) [16:58] barry, rvr: hey, mostly jgdx and myself [16:58] rvr: hmm. i see "Auto download Never" now [16:58] after i install s-i 3.0.2, and reboot [16:59] barry: Hmmm [16:59] barry, reminds me, we need to stop using Info if you backport s-i 3 to vivid+overlay, right? [16:59] jgdx: right. use .Information() [16:59] barry, I have an MP for that [17:01] jgdx: cool. rvr, jgdx, kenvandine one thing i am seeing though: if i set auto_download=0 via system-image-cli, system-settings doesn't see the change [17:01] i'm guessing it caches it and doesn't recheck [17:02] however, if i swipe up the s-s panel and relaunch, it sees the change [17:02] barry, while page open, or do you back->re-enter? [17:02] jgdx: back->re-enter [17:03] kenvandine, shouldn't be any caching between panel “sessions” should there? [17:04] it might not get the property changes [17:04] so won't see the value changed until it restarts [17:04] barry, that sends a dbus signal? [17:04] unless it was changed in settings [17:05] jgdx: yes, it should send a SettingChanged signal with the key and new value [17:05] okay, maybe it's not hooked yp [17:05] my guess is we don't handle that signal [17:05] jgdx, kenvandine swipe-up->relaunch does work [17:10] jgdx, kenvandine, rvr so the other thing i see is that on manual download, while i do get an offer for r155 (i'm on r149), i don't see any progress during the download. might want to check that you're handling UpdateProgress signals correctly too [17:11] * barry will update the trello [17:12] barry, mind filing a bug? [17:13] kenvandine: on both signal issues, or just UP? [17:13] both please [17:13] one bug for both is fine [17:13] kenvandine: ack [17:13] thx [17:14] ahayzen, popey, pmcgowan ^ [17:14] nice [17:15] popey: now just publish the right one ;) [17:16] brilliant news, thanks davmor2 [17:17] rvr, jgdx, kenvandine at least the --set issue isn't a bug in system-settings i think. i think the signal doesn't get set when system-image-cli is used [17:17] *sent [17:18] i'm still going to file the bug, but i'll add a bugtask to system-image for that, though 1) i'm not sure if i can make that work; 2) it's an odd corner case anyway since "normal people" don't use the cli [17:19] barry: I see [17:20] barry: Hmm [17:20] barry: I'm re-running test B to check the rest of it === charles__ is now known as charles [17:23] rvr: okay. jgdx, kenvandine LP: #1509022 [17:23] Launchpad bug 1509022 in ubuntu-system-settings (Ubuntu) "system-settings may not be responding to system-image D-Bus signals" [Undecided,New] https://launchpad.net/bugs/1509022 [17:24] barry: Great, thanks [17:26] rvr: thanks! [17:27] barry, thx [17:40] barry: Test B [17:40] barry: Open System Settings panel [17:40] barry: I go to Updates [17:41] barry: And, although I set system-image-cli --set auto_download=0, Updates says it downloads when wifi available [17:41] barry: Anyway, I installed that update [17:42] barry: After the phone is rebooted, I rm -f /etc/system-image/config.d/02_testing.ini [17:42] barry: Test plan says "Open System Settings panel, then Updates. It should report that your Software is up to date" [17:42] barry: But it downloads the image again [17:43] barry: cat: /etc/system-image/config.d/02_testing.ini: No such file or directory [17:43] sil2100, ping [17:46] cyphermox, hey can you do me a quick favor and push qtmultimedia-opensource-src from ppa:jhodapp/ubuntu/ppa to silo 15 please? [17:46] cyphermox, source package only [17:46] sure [17:46] rvr: i think i left off a step. after printf'ing to 02_testing.ini, system-image-dbus needs to be restarted. let me verify and update the wiki [17:46] thanks [17:46] davmor2: ahayzen published :) (the right one) [17:47] rvr: or you can wait for it to timeout, but that's no fun [17:48] barry: I see [17:49] jhodapp: stuff is building, do you still need the copy? [17:49] cyphermox, yeah, if the build is in the way feel free to cancel it [17:51] cyphermox, thanks for doing that, going to grab lunch quickly...bbiab [17:52] rvr: hey. sorry, forgot to approve the one branch in our iap silo. it's approved now :) [17:52] jhodapp: copied; it's building now [17:52] dobey: Great! [17:53] thanks! === robru|sick is now known as robru [17:54] sil2100: I'm here to enable xenial if you want to EOD [17:55] robru: hey! You feeling good enough? [17:55] sil2100: oh yeah, just slept a bit late, feeling fresh now [17:55] robru: ok, didn't start the switch yet, I'm deep in Qt debugging now [17:55] barry: Ok, killing system-image-dbus does the trick [17:55] sil2100: alright [17:55] I suppose you would do it much faster than me anyway [17:55] barry: System up-to-date [17:56] robru: Welcome back [17:57] rvr: thanks [17:58] rvr: cool. wiki page updated [18:04] sil2100: slangasek: ok bileto trunk is updated for xenial, just waiting on #webops to do a rollout (vanguard is on lunch) [18:04] robru: excellent, thanks! Now just the train needs to be switched back to not to land to the overlay for x [18:04] sil2100: oh right, heh, on it ;-) [18:05] robru: why did you leave wily+vivid as a target and remove trusty+precise? [18:05] robru: and where is the logic that tells it to land xenial to the main archive instead of the overlay ppa? I couldn't figure this out looking at the code [18:05] barry: Silo approved [18:06] slangasek: left wily+vivid for transitional purposes. weird results happen if we do a hard/instantateous cutover, eg, silos mid-publish won't be able to find the packages during the migration check. removed trusty+precise because that was more of just an example that I don't think anybody was actually using. [18:06] ok [18:11] tvoss: hi, just replied to your comment on LP: #1278780 [18:11] Launchpad bug 1278780 in apport (Ubuntu) "apport takes too long to write crash report, appears to lock up phone" [High,Triaged] https://launchpad.net/bugs/1278780 [18:12] slangasek, cool [18:13] tvoss: summary: I like the sendfile idea, but want to discuss further why you classify unmapping of graphics buffers as "quite dangerous" :) [18:14] slangasek, yup, just read through your comment. So the problem is: it is not as simple as unmapping a buffer, and we have no idea what an underlying driver might do if we call "free graphics buffer" [18:15] if it was an munmap, fine with me [18:15] tvoss: ok; why is it not that simple? :) [18:15] we've segfaulted. You don't get to do any other cleanup in this case [18:15] is it because you don't have enough info to call munmap()? [18:15] slangasek, we use gralloc, with a binaryblob providing us with the actual implementation [18:16] slangasek, we don't call munmap, but would call into a binary blob we don't own and have no idea what it is doing :) [18:16] slangasek, so "mapped graphics buffers" does not mean mmap'd in the general case [18:17] hmm [18:17] robru, ping [18:17] tvoss: de facto, if it's been mapped into the process's memory space, it should be possible to unmap it [18:18] alexabreu: pong [18:18] using munmap() [18:18] sil2100: slangasek: ok, xenial+vivid is live in bileto production [18:18] robru: did you see my second question, about where the logic lives to tell it that xenial is landing to the main archive and not the overlay? [18:19] slangasek: yep, that's fixed in cu2d and also in production already [18:19] slangasek: or sorry, you wanted to know *where* the logic is? [18:19] robru: ok, so I was just looking at the wrong repo - thanks [18:19] slangasek: yeah check the latest commit at lp:cupstream2distro [18:19] slangasek, sure, although I would like to make sure we at least have an idea of how things might go wrong. also: ashmem and pmem ftw, you never know what a driver uses [18:19] robru: I wanted to know why I couldn't find it, and I also wanted to know it was done ;) [18:20] slangasek: heh, looking at the scrollback, yeah I somehow missed your entire second message there [18:20] slangasek, don't get me wrong: minimizing the memory footprint is a great idea, but it needs further thought to minimize the risk of corrupting graphics driver state [18:20] robru, have a mns to check something w/ me? [18:20] alexabreu: sure, what's up? [18:21] robru, https://jenkins.qa.ubuntu.com/job/unity-js-scopes-wily-armhf-ci/23/console check at the bottom, ... the copyright errors, ... & the associated branch seems to be fine in terms of deb copy http://bazaar.launchpad.net/~abreu-alexandre/unity-js-scopes/doc/view/head:/debian/copyright [18:21] robru: the train code is in lp:cupstream2distro if anything ;) [18:21]