[01:10] Anybody know how to install Ubuntu Touch on a Samsung Galaxy Fascinate? [01:12] !devices [01:12] You can find the full list of devices, official images, community images, and works in progress at https://wiki.ubuntu.com/Touch/Devices [01:46] morning to all [01:50] lotuspsychje: morning! [01:50] nhaines: hello mate [01:50] nhaines: ive tested the reset function to factory settings on channel=devel [01:51] working nicely and smooth [01:52] lotuspsychje: great! :) I think I tried it out a month ago and it worked pretty well. [01:55] nhaines: are you gonna try rtm or stick to devel? [01:57] I'm sticking with devel. There's no reason for me to try RTM until... well :) Until it's actually RTM. I'm mostly just doing promoted images anyway. [01:57] I'm on a Nexus 5 anyway. [01:57] devel will have stability of rtm also no? === Ursinha-afk is now known as Ursinha [04:06] nik90: awake? :) === ecloud is now known as ecloud_wfh [07:00] good morning [07:15] dholbach: good morning. :) [07:16] hi nhaines [07:16] hi dholbach. [07:18] This Ubuntu phone thing is starting to look pretty good. :) [07:18] it is :-D [07:21] I ran it on my N5 on Saturday during the LUG meeting. A lot of people were pretty interested. And I may give a presentation at a local Android user group in November. :) === tarman is now known as zyga [07:56] bah [07:57] where did the calendar app go? [07:57] it's not in the dash anymore since the update to rtm58 on krillin [08:08] Good morning all; happy Punctuation Day! :-D [08:29] Elleo, morning! do you have an example of the download you are trying to deflate, else I'll be testing with a random zip [08:36] mandel: I'm testing with a 7digital album download (which you'd need to purchase) but I expect any zip would give the same results [08:36] Elleo, I'll test with the python download then :) [08:37] Elleo, will post my results in a few mins [08:38] pete-woods, hey, do you know what code displays the error page in the click-scope (like if you try to install a click without having an u1 account configured) [08:39] mandel: I'm thinking it probably only applies that are reported as being encoded as deflate/gzip/zip/etc by the server, rather than downloads with zip mimetypes [08:40] Elleo, correct, but what I could see in the bug report is that the header from 7digital are correctly set to do so [08:40] mandel: http://en.m.wikipedia.org/wiki/HTTP_compression [08:40] seb128: I think it's a preview, with an arrangement of widgets to make that screen up [08:41] pete-woods, ok, my issue is that the error label doesn't wrap and it's too long to fit on screen in french, what component should I open the bug against? [08:43] seb128: if the issue is that the bold part doesn't wrap, perhaps unity8? [08:43] mandel: are you sure? https://bugs.launchpad.net/webbrowser-app/+bug/1365993/comments/4 <-- shows the transfer encoding as "Chunked" rather than "zip" or similar [08:43] Launchpad bug 1365993 in webbrowser-app "Support download of albums as zip files from 7Digital" [Critical,In progress] [08:44] pete-woods, http://people.canonical.com/~seb128/unity.jpg [08:44] mandel: it just has the content-type as application/zip which I don't think is enough to be treated as a deflatable thing [08:44] pete-woods, screenshots are easier I guess ;-) [08:44] pete-woods, the "subtitle" line [08:44] Elleo, true, application/zip is not enough [08:45] mandel: http://pastebin.ubuntu.com/8412245/ <-- and those are the headers I got, which are pretty similar [08:45] ogra_: i've been told you might know how does the android layer decide what is the local timezone when calling the localtime() from libc. it does not seem to think it's the same timezone that ubuntu thinks it is [08:45] Elleo, will test with other file, in the case in which we do not have the correct header we will call the trusted helper.. I don't see other possible way [08:46] mandel: okay [08:46] seb128: yes. okay, that might be intentional. we'd need to see what the unity8 guys say about that widget. I'm a relative latecomer to scopes, so don't really know the designs [08:48] Elleo, is kinda ugly if you ask me, but doable I suppose [08:50] Elleo, the other way is for you to pass the posdownload command with the unzip [08:51] pete-woods, so bug against unity8? [08:52] seb128: I think that's a starting point. then if u8 guys say no, need to go to unity-scope-click, and try there? [08:53] pete-woods, well, it really depends who renders the ui, the string comes from unity-scope-click/libclickscope/click/preview.cpp:411 [08:54] but that only seems to return the text [08:54] not sure where the rendering is done [08:54] mandel: sending a command to execute some string from the browser sounds like something security might get rather worried about to me [08:54] seb128: well the rendering is certainly happening in unity8. my point it just that it may be argued that the subtitle widget is behaving according to spec. so the click scope guys will need to use a different widget [08:55] Elleo, yes, I'll deal with the use of the trusted helper and be done with it or unzip myself [08:55] hence my suggestion to try reporting u8 bug first, and seeing what is said [08:55] pete-woods, that makes sense, going to do that, thanks! [08:55] mandel: okay, thanks [08:55] Elleo, I dont see the diff between the trusted helper and udm, I'll talk with jdstart about it and you will have a fix this afternoon [08:56] mandel: okay cool [08:57] Elleo, udm is getting huge hehe [08:57] heh [08:57] pete-woods, ok, I filed https://bugs.launchpad.net/ubuntu/+source/unity8/+bug/1373315 [08:57] Launchpad bug 1373315 in unity8 (Ubuntu) "click scope errors don't always fit on screen" [Undecided,New] [09:03] seb128: hopefully the response will be positive :) [09:03] yeah, let's see === chihchun is now known as chihchun_afk [10:35] ogra_: ping [10:37] rsalveti: do you know how the android layer decides what is the local timezone when calling the localtime() from libc. it does not seem to think it's the same timezone that ubuntu thinks it is [10:45] nerochiaro, why would that matter ? [10:45] (it gets it from bionic i guess) [10:46] ogra_: because when we ask android to take a picture it will save in the EXIF tags a timestamp that is different from the current time on the phone [10:46] bah [10:48] ogra_: well i have a bug assigned asking to fix that, and either we keep the tz info in android the same as in ubuntu, or i have to fix the exif tag before saving the jpeg to disk [10:48] right [10:48] ogra_: not a big problem but if there's an easy solution in android i'll take that [10:48] ogra_: if not, i already have a branch that does the exif fixing [10:49] well, seems we shuld be able to just set the persist.sys.timezone property to a proper TZ value [10:49] the question is if the naming of bionic is the same here ... if not it will be tricky [10:49] (for the timezone names i mean) [10:50] ogra_: right, any pointers on how to check that ? [10:52] phablet@ubuntu-phablet:~$ sudo lxc-console -t0 -nandroid [10:52] [sudo] password for phablet: [10:52] ... [10:52] root@mako:/ # date [10:52] Wed Sep 24 10:51:11 GMT 2014 [10:52] ... [10:52] root@mako:/ # setprop persist.sys.timezone Europe/Berlin [10:52] root@mako:/ # date [10:52] Wed Sep 24 12:52:15 CEST 2014 [10:52] ... [10:52] seems it uses the same notation [10:53] (but to verify 100% one would likely need to check the source of bionic or so [10:53] ) [10:54] nerochiaro, probably copy the above into your bug ;) [10:55] pitti, do you think it would be possible to make timedated set that property alongside setting the timezone ? [10:55] (assuming this it what sets it) === MacSlow is now known as MacSlow|lunch [10:57] dpm: can you please have a look at this .desktop i18n MR ? https://code.launchpad.net/~phablet-team/notes-app/inline-desktop-i18n/+merge/235785 [10:58] nerochiaro, oh, and https://gist.github.com/arpit/1035596 [11:06] ogra_: oh, first time I hear about it; what is using that property? Why can't we just use /etc/timezone? [11:08] ogra_: yes, I guess it can be put into timedated (we already have that hideous hack to support /etc/writable), but NB that this is by far not the only way to set/change the time zone [11:08] pitti, we use bionic inside the android container and libc outside ... seems bionic need the property above set ... the camera service runs completely inside the container [11:09] pitti, i would like to wait for rsalveti's feedback, though, he usually has a better overview of the container stuff ... [11:32] zbenjamin: when you modified apps to be more compatible with qtcreator did you also do that for notes ? [11:40] mpt: hi! There is the case of some account providers (U1 and facebook) for which we should limit to 1 the number of accounts created [11:41] mpt: from the system settings, that is easy: I just grey out the option to click on that provider, when there is already one account configured [11:42] mpt: but when the request comes from an app, there is at least one case in which I need your decision: [11:42] mpt: if the app has already one account authorized, and it requests a new one, what should happen? [11:42] mpt: I guess we need to show a dialog with an error message [11:48] mardy, I don’t see how that could work. If I used my Facebook account in an app, but then I told the app to forget about it, and then later I used it again … To Online Accounts that would look like the app is trying to use a second account, when it isn’t [11:49] Why does Online Accounts need to do anything special for that case at all? It’s up to the app to make sure it isn’t trying to use multiple Facebook accounts. [11:52] mardy: what is the UI presented in the application in that scenario? [11:52] nerochiaro: have the bug in hands? [11:53] rsalveti: https://bugs.launchpad.net/ubuntu/+source/qtubuntu-camera/+bug/1371390 [11:53] Launchpad bug 1371390 in qtubuntu-camera (Ubuntu) "pictures taken with camera have wrong timestamp" [High,In progress] [11:54] rsalveti, i suspect a wrong tz in the contaainer can have other issues too (timestamps on radio packaets etc) [11:54] nerochiaro: ogra_: yeah, this is a problem because we abstract the camera by using the camera-service [11:54] mpt: we need to reflect the terms of the service at the system level; if an app wants to use another FB account, it can do so on its own [11:54] and not only the HAL [11:55] so indeed, not something I can fix on the hybris side [11:55] rsalveti, well, seems it luckily isnt to hard to set (with a simple property) and the TZ naming in bionic seems to match the libc one [11:55] so the workaround in qtubuntu-camera looks good, but ideally it'd also be good to set the right timezone inside the container as ogra_ suggested [11:55] mpt: just not the system one [11:55] so we could just have timedated set the property alongside [11:56] ogra_: yeah, that would already help [11:56] nerochiaro, add a timedated task then i guess [11:56] mpt: i guess, irrespective of the terms & criteria discussion, we're looking for a solution for accounts which need to be unique, like U1 [11:56] (systemd source package afaik) [11:56] mpt: so, the goal is to have at most one facebook account configured in System Settings === gnu is now known as Guest80678 [11:57] ogra_: rsalveti: so you suggest for now fixing the EXIF and later on removing that fix once we set the right tz in the container ? [11:57] nerochiaro: right [11:57] really ? why not just fix the TZ and be done ? [11:57] mpt: we do know which accounts an app is using, because that information is in the System Settings (that list of enabled apps per each account) [11:58] dbarth, mardy, that’s fine. App X says to Online Accounts, “Hey I need to use a Facebook account.” OA puts up the prompt asking you if it’s ok for X to use it. You accept, and OA gives X what it needs. Sometime later, X says to OA, “Hey I need to use a Facebook account.” OA gives X what it needs, because you already granted permission earlier. [11:58] ogra_: well, if we can get someone to fix the TZ, even better [11:58] but I don't mind having the change in qtubuntu-camera if that helps unblocking a critical bug already [11:58] well, i'm not sure how hard it will be to make timedated depend on the property service ... that will need some clever checks [11:58] since it need to operate without as well [11:58] right [11:59] mpt: not really: when the app starts up later on, if it needs just 1 facebook account, it won't talk to OA, because it already has one [11:59] the setprop call itself should eb a one liner [11:59] *be [11:59] yeah, ideally just checking if setprop is available, if so, run the one liner [11:59] right [11:59] mardy, this is in the case I described where you told the app to forget about the Facebook account, or it forgot because of a bug, or whatever [12:00] mpt: when the app starts up, it enumerates the accounts which are available to it; it doesn't (or shouldn't) store this information anywhere [12:00] lol [12:00] rsalveti: ogra_: what project is timedated in LP ? [12:00] apt-get source systemd ... the amount of patches that applies on unpack exceeds my terminal scrollback buffer ! [12:00] ogra_: help; where can i enable debugging for a device not showing up in adb devices? [12:01] mpt: once you have authorized an app to use an account, that app will see that account, until you go and disable the app from the System settings (or you delete the account) [12:01] i have adb enabled on the phone, according to google-gadget-service [12:01] dbarth, hmm, is the UI session running ? and is the password *not* identical with the user name ? [12:01] they are different yes [12:01] unity8 is running if that's what you mean [12:02] yeah ... well, adbd looks for lightdm ... [12:02] do you see adbd in the processlist in the termianl app ? [12:03] mardy, sure. So for an account type where you can have only one, the OA prompt will be just “Allow”/“Don’t Allow”, without the “Add Another…” button. Do we need to change anything other than that? [12:03] mpt: this is fine in case the app hasn't been authorized to use that account before, but we need something more for the case where the app requests another account of the same type [12:04] mpt: indeed, the app should not do that [12:04] mpt: but maybe the app does not know that this account type is limited to one account [12:04] ogra_: yeah, changing systemd is no fun [12:05] mpt: we might even decide to change this in the future (maybe one day we'll decide to limit twitter to one account, for example) [12:05] timedated.c looks ppretty straight forward though ... [12:05] mpt: so yes, I think it's a bit of a corner case [12:06] mpt: we can silently fail, maybe. That is, in that case we always return "declined" to the app, without actually showing any UI [12:06] ogra_: where do i get that ? [12:08] mardy, I was thinking that OA would always return "accepted", giving the account that the app has already been granted access to, on the grounds that the app has forgotten about it during runtime [12:08] (probably) [12:08] nerochiaro, apt-get source systemd [12:08] nerochiaro, then src/timedate/timedated.c [12:08] mardy, but I guess if the app hasn’t forgotten, and is expecting to handle multiple accounts when it can’t, that would be a problem [12:09] nerochiaro, i guess adding some snippet to write_data_timezone() should be the way to go [12:09] mardy, so yeah, always returning “declined” makes sense [12:10] mpt: OK, thanks [12:10] ogra_: whoa, the list of patches is mind boggling indeed [12:10] yeah, entertaining :) [12:15] ogra_: rsalveti: to be perfectly honest i'm a bit scared to put my hands down there in systemd. i'd rather have the patch to qtubuntu-camera in for now and leave that to someone who knows what they are doing at that system stuff better than I do [12:16] mardy, specification updated. [12:16] nerochiaro, well, all you need to do is to check if the setprop binary exists in $PATH ... and if it does call "setprop persist.sys.timezone" pointing to the new tz [12:16] nerochiaro, and pitti will surely happily review a patch from you [12:17] mpt: excellent, thanks! [12:17] ogra_: i'm having an headache just trying to get that stuff compiled, but ok, i'll give it a shot. where does the project live in lp ? [12:18] nerochiaro, i dont know, ask pitti [12:18] pitti: ^ [12:18] but i guess a debdiff will be simplest === dandrader is now known as dandrader|afk [12:27] hmm, when did the reboot option come back ? [12:28] ogra_, where? I don't have it on rtm 59 [12:28] seb128, on my utopic device [12:28] oh, hum [12:28] which still runs 251 [12:29] will soon be in rtm too i guess [12:29] https://launchpad.net/ubuntu/+source/unity8/8.00+14.10.20140918.3-0ubuntu1 [12:29] * Re-add restart button for power menu. (LP: 1358197) (LP: #1358197) [12:29] Launchpad bug 1358197 in Ubuntu UX "Power off dialog box no longer shows restart" [High,Fix committed] https://launchpad.net/bugs/1358197 [12:32] ogra_, it's in https://launchpad.net/~ci-train-ppa-service/+archive/ubuntu-rtm/landing-006 [12:32] sweet ! [12:35] nerochiaro: https://code.launchpad.net/~zeller-benjamin/notes-app/sdkcompat/+merge/233892 [12:35] seb128: could you install and test a bit glib from rtm silo #15 please? [12:35] Laney, sure [12:35] ty [12:35] yw! [12:35] did you get any feedback from qa on it yet? [12:35] I think QA don't want to test it ... [12:36] they like to have a plan [12:36] k [12:36] let me install it on my killin, reboot and use it for a bit [12:36] ty [12:36] lunch, brb === nik90 is now known as nik90|Lunch [12:39] Hi all! I have a quick question. When I clicking the camera icon at Messaging App, a new dialog popup and says No app installed or similar. What app is that? So were are can find this string? [12:39] it is also in the address-book-app === _salem is now known as salem_ [12:42] screenshot: http://oi59.tinypic.com/15zqvew.jpg [12:42] Can somebody help? [12:42] mpt, ping [12:43] blake_r, i think thats the content-hub [12:43] shrug, unity8 is hanging on the sudoku start screen [12:43] can't swipe from any edge or anything [12:44] seb128, starts fine here [12:44] oh, back, I guess it segfaulted [12:44] with apport blocking it for a minute [12:44] yeah, check for crashes [12:44] ogra_: Yes, that is it :) Thanks [12:44] right [12:45] bah, I wish design would make the "unlock dialpad" screen use the lock-screen image, rather than the purple default one [12:45] * seb128 wonders if he can hack that up locally easily === dandrader|afk is now known as dandrader [12:47] seb128, oh, gimme the patch if you find a way ! [12:47] :) [12:47] i bet simply replacing the default image would do though ... [12:47] (til the next OTA) [12:47] yeah [12:48] it's just that I don't get what they are thinking [12:48] yeah [12:48] you get like 3 images [12:48] lock, unlock dialpad, dash [12:48] which unlock dialpad is the purple image which you might really dislike [12:48] * ogra_ would also love if it would immediately lock [12:48] buut only after a few seconds [12:48] it wouldn't hurt to use the lockscreen one, maybe with renforced constrast to see the number or something [12:49] yeah === dholbach_ is now known as dholbach [13:08] Chipaca, hi, is/was there a bug report about getting rid of the notification bubbles for system updates? [13:11] mpt, hi, can you change the reset launcher spec[1] to work with a reboot action, which is required? [13:13] jgdx, yes, is the bug assigned to me? [13:14] Hi guys. I do have a question regarding the list of supported devices etc [13:14] mpt, no, and I don't have perms to assign it to you [13:14] jgdx, what’s the bug number? [13:14] I noticed for example Xperia Z is on the list of supported devices, however the "mobile modem" seems not to be working. Is that statement linked to official/legal issues and can be bypassed by coding your own call software? === MacSlow|lunch is now known as MacSlow [13:22] Is it a taboo question? Googled it wrong? [13:30] if anyone who likes to look at phablet-tools has a moment, I would appreciate a review of https://code.launchpad.net/~pwlars/phablet-tools/network-no-sudo/+merge/235689 === nik90|Lunch is now known as nik90 [13:31] ogra_: someone told me you already had something in progress for phablet-tools landing but I don't see it, so maybe it made it through already? [13:34] * mpt crashes indicator-network by opening it [13:37] jgdx, I’m looking at your fix to bug 1364366… [13:37] bug 1364366 in ubuntu-system-settings (Ubuntu) "Wi-Fi networks in System Settings lack strength or authentication icons" [High,Fix released] https://launchpad.net/bugs/1364366 [13:37] jgdx, I’m really impressed at how the fix was just one line, but somehow the appearance in System Settings is so different from in the indicator menu :-) [13:38] mpt, yeah, it was an impressive fix [13:41] mpt, right, there's difference. But AFAIK it's very hard/impossible to make the separation between the currently active Wi-Fi network and other inactive ones. [13:42] jgdx, oh, I’m not referring to that, but to the icons being four times larger, the padlocks being a different color, and the padlocks being a different position [13:43] mpt, right, I sped ahead. So if you move the padlocks, they are going to be besides the checkbox? [13:43] How is that even possible if it’s just reusing the indicator code? [13:43] It uses the data from the indicator but lays it out itself [13:44] Ahh, ok [13:44] Worst of both worlds :-) [13:44] to a point [13:44] we should use unity-settings-components [13:44] *from a point [13:45] seb128, if you write the backend, I'll happily do the FE [13:45] lol [13:45] that's not a backend [13:45] Hi people, I do have a question prior to buying a phone from the compatibility list, anyone cares to provide me guidance? [13:45] it's replacing custom qml by shared components [13:45] no? [13:46] moooooop: the only really supported phone atm is nexus 4 [13:46] I noticed on the "compatibility list" that for some phone * is okay, however "wireless modem" is greyed out. Is there any way to bypass that through custom software? [13:46] Like what they did on the pi phone etc? [13:46] thank you cwayne [13:47] jgdx, anyway, sorry to gripe. I was just dealing with a bug report about the white padlocks and wondered why that icon existed [13:47] seb128, oh [13:48] moooooop: there's a community build for the nexus 5 that works quite well also, but not officially supported [13:48] I guess i dont mind for the support, thank you [13:48] official that is ;) [13:49] seb128, where can read about this magic? [13:50] seb128, right, but using shared components fixes not all issues with the wifi list [13:51] jgdx, I didn't say it would, but it would be less custom code [13:51] jgdx, dunno if it's documented, ted might know, he did work on that back then iirc [13:55] cwayne: thanks a lot. [14:03] ahayzen, Were you able to confirm whether UAL caused the issue? [14:04] tedg, literally reflashing back to #243 at the moment! [14:04] ahayzen, Hah, cool. [14:04] tedg, when it flashes back what is the best way to upgrade just the UAL packages? [14:04] ahayzen, apt-get install ubuntu-app-launch libubuntu-app-launch2 [14:04] tedg, cool thanks [14:05] ahayzen, You'll have to reboot so that unity will reload the lib. [14:05] ok cool [14:15] kenvandine, hi, after I recovered from the black-screen system update, I went back to following your instructions, and realized that I don’t know how to *add* a silo [14:16] kenvandine, citrain device-upgrade upgrades everything, right? Whereas I just want to install specific packages [14:16] mpt, it could pull in some other updates as well [14:17] but that's not usually much [14:17] you could just add the ppa and apt-get install [14:17] plars, yeah, yesterday [14:17] plars, land away :) [14:17] sudo apt-get install ubuntu-system-settings-wizard ubuntu-system-settings libsystemsettings1 [14:17] mpt ^^ [14:23] * tedg still can't get used to the sudo password for phablet not being phablet [14:24] kenvandine, I need to add the silo before I do any of those things, right? [14:24] mpt, yes [14:25] you can always just use add-apt-repository on the device if you like [14:25] kenvandine, so, how do I do that? [14:25] sudo add-apt-repository ppa:ci-train-ppa-service/landing-001 [14:25] kenvandine: only if you are not on rtm [14:25] then apt-get update [14:25] davmor2 yes :) [14:25] mpt, phablet-config writable-image [14:27] mpt, unfortunately... there is now a newer version in the archive... so you need to specify the version :) [14:27] apt-get install {ubuntu-system-settings-wizard,ubuntu-system-settings,libsystemsettings1}=0.3+14.10.20140923.1-0ubuntu1 [14:27] mpt, oh so many steps :) [14:27] Thanks tedg [14:29] * mpt tries not to wonder why phablet-config pings launchpad.net [14:29] kenvandine, sorry, a newer version of what? [14:30] mterry, hey, do you have an issue about the issue I described on https://code.launchpad.net/~mterry/ubuntu-system-settings/flickable-passwd-page/+merge/235117 ? [14:30] system-settings [14:30] just landed in utopic-proposed [14:30] mpt, It's the most reliable server on the Internet. :-) [14:30] seb128, no I couldn't reproduce [14:30] seb128, thought I mentioned that on IRC, sorry [14:30] mterry, you might have and I might have missing it [14:30] kenvandine, I have a few USS branches I'd love you review if you have time today [14:30] mterry, planning to :) [14:30] * mterry hugs kenvandine [14:31] kenvandine, “E: Version '0.3+14.10.20140923.1-0ubuntu1' for 'ubuntu-system-settings-wizard' was not found” (etc) [14:32] mpt: sudo apt-get update after you added the ppa [14:34] thanks davmor2 :-) [14:34] tedg, ...ok so flashed back to #243...ran ap test worked fine.... updated ubuntu-app-launch libubuntu-app-launch2 .. restarted... reran ap tests and mocking is broken! [14:35] ahayzen, Uhg, okay. Do you know which module is getting an error? [14:36] tedg, i have seen no errors that is why we were stumped what was causing it....but maybe now we know UAL is causing it we can dig a bit easier [14:36] tedg, does UAL have a log somewhere? [14:37] ahayzen, Well, the problem is that it's not a thing itself, it's more a way to use upstart. So it doesn't have its own log. [14:37] :/ [14:37] kenvandine: can you comment on https://bugs.launchpad.net/ubuntu-system-image/+bug/1370586/comments/2 ? i'd like to take a crack at this soon [14:37] ahayzen, So can you walk me through how the test works? [14:37] Launchpad bug 1370586 in Ubuntu system image "Add synchronous method to determine if there are known updates" [High,Triaged] [14:38] ahayzen, You setup the media-scanner, start the app? Do you set anything? (env vars, etc) [14:38] tedg, erm i can try :) ... yeah we do [14:39] tedg, the magic happens in here http://bazaar.launchpad.net/~music-app-dev/music-app/trunk/view/head:/tests/autopilot/music_app/tests/__init__.py [14:39] tedg, so for a click env... we build a fakeenv fixture [14:39] set that as HOME [14:39] barry, commented! [14:40] tedg, then go and copy the music.... live patch the ms2 db [14:41] ahayzen, So then who knows what $HOME is? [14:41] tedg, and then launch the app via ap's launch_click_package() [14:41] kenvandine: thanks. my big question is what to do in the "we-don't-know" case. should the target_build_number key just be missing? should it return a non-number like "none", or should it return a nonsense number like -1 (or 0?). my initial preference would be to omit the key, but i don't know how much trouble the different options would cause the client [14:41] ahayzen, Do you restart Unity in the new $HOME? [14:41] barry, i think i'd prefer -1 [14:41] tedg, no ... this is where part of reason we are patching the db but we're looking to improve that [14:42] kenvandine: thanks, i will copy and paste that into the comments [14:42] tedg, i think only python and the app itself are affected by the new HOME? [14:42] tedg, balloons probably has more knowledge of how it all works ;) [14:43] I'm a little disturbed that you're trying to replace $HOME for half the world, without telling the other half. [14:43] heh [14:43] Honestly, shocked it worked before :-) [14:43] $HOME is where the heart is though [14:43] :) [14:43] it works well :) until ms2 change their db schema [14:44] ogra_, Yes, the problem with that with two $HOME's for your heart, well, that never works out well ;-) [14:44] yeah, that might have some circulation issues [14:44] :) [14:44] mv ♥ $HOME [14:44] So my guess is that the environment variables aren't getting set correctly. [14:44] They changed slight in the new UAL as part of the performance optimizations. [14:45] tedg, this is where it is set currently http://bazaar.launchpad.net/~music-app-dev/music-app/trunk/view/head:/tests/autopilot/music_app/tests/__init__.py#L162 [14:45] tedg, yeah so then it is seeing the wrong db cache? [14:45] ahayzen, So I think what probably needs to happen is that also needs to call a "setenv()" to set it for the test process as well. [14:45] yeah [14:46] this will only affect things started by upstart otherwise [14:46] tedg, ok i'll try that and report back [14:46] ahayzen, Cool, thanks! [14:47] Just for lurkers, basically we changed that a bunch of the environment variables are built into one big structure in libual. [14:47] So that way we don't have to do it in the job. [14:47] ah ok [14:47] That got us under the time limit. [14:47] So that environment processing would be in the test, as it's the libual user in this case. [14:48] we have a time limit ? [14:48] ogra_, I was given a budget. [14:48] 100ms for UAL stuff. [14:48] kenvandine, ok, now I’m trying to “sudo nano /etc/system-image/channel.ini”, but after I change the version number, it won’t save [14:48] ah [14:48] tedg, do you mean $ initctl set-env ? or a different way just to check [14:48] mpt, in "adb shell" or phablet-shell ? [14:48] @find noein 15 [14:49] ahayzen, I mean what ever the python equiv of "setenv" is. I'm guessing system.setenv('HOME', tempdir) ? [14:49] sorry wrong channel [14:49] mpt, note that adb doesnt use termcap ... which means control chars wont work properly in some cases [14:49] tedg, oh so >> os.environ["HOME"] = tmpdir [14:49] mpt, use phablet-shell for that it works around that issue [14:49] ogra_, ah.. thx :) [14:49] ahayzen, Sure! :-) [14:49] \o/ i'll try that [14:49] ahayzen, Just make sure it's not only for Python, as the C lib needs to see it as well. [14:50] ahayzen, Not sure if that sets the actual environment or just the Python copy. [14:50] hmm [14:50] * balloons is listening [14:50] balloons, o/ [14:50] balloons, any comments/pointers? [14:50] kenvandine, oh, btw, i kind of stepped on your toes with a u-s-s landing that went into utopic ... if you do the next rtm landing, can you make sure https://code.launchpad.net/~ogra/ubuntu-system-settings/system-settings-fix-devmode-page-refresh/+merge/235778 doesnt get forgotten ? [14:52] ogra_, phablet-shell says “ssh_exchange_identification: Connection closed by remote host” [14:52] uh [14:52]