[00:42] robru: seems to be some problem with the publishing on silo 4 [00:42] with notes-app? [01:18] bfiller, oh hey, sorry, was running some errands. indeed it looks like nobody bothered to build notes-app. building it now [01:18] robru: cool, thanks [01:19] bfiller, you're welcome [01:34] but queuebot I just built notes-app! [01:54] is it just me, or is jenkisn broken on arm? [01:56] dobey, which jenkins are you referring to? [01:56] robru: https://jenkins.qa.ubuntu.com/job/pay-service-utopic-armhf-ci/25/console [01:57] looks like maybe the vm/chroot it's using got left in an unclean state? [02:00] dobey, hmmm maybe fginther knows more about that one, sorry ^ [02:04] === trainguard: IMAGE 111 building (started: 20140703 02:05) === [03:34] === trainguard: IMAGE 111 DONE (finished: 20140703 03:35) === [03:34] === changelog: http://people.canonical.com/~ogra/touch-image-stats/111.changes === [03:39] robru: publish for silo 4 yet? [07:06] http://people.canonical.com/~ogra/touch-image-stats/111.changes is 0 bytes long.. ogra_ ? [07:08] popey, it's a stable image :p [07:09] haah [08:06] plars: psivaa: health-check tests are not run anymore? dont see results on mako for last two images [08:07] plars: there is one ongoing now. i saw one of the earlier builds failed to flash with some network error between our server and system-image-server [08:07] thanks for checking this out .... not sure if we could improve the dashboard to always have an entry for health-check [08:07] so we always spot if the test doesnt run etc. [08:07] thanks [08:08] psivaa: was that for me? :) [08:08] asac: yea, it was. sorry [08:08] hehe [08:08] ok cool [08:08] still waking up :) [08:08] me too :P [08:09] asac: i'll ask if it's possible include a placeholder for health check in the dashboard [08:09] popey, fixed, sorry [08:13] stgraber, only core-devs have access to build images iirc ... [08:13] thanks ogra_ [08:18] ugh, the browser4 does not look happy [08:18] -4 [08:19] cant find its toolbar it seems [08:27] Yeah, he's not only not happy, he looks pissed off real nice [08:28] sil2100, it's really frustrating. *still* not reproducible locally [08:29] sil2100, ever [08:29] ;/ [08:29] At least we know that Mir didn't cause any of those regressions, because we had that Mir-specific image that was rather fine [08:30] will be late, doorbell [08:30] sil2100, 110? [08:30] Yes [08:38] robru: thank you :) [08:38] ogra_: liar liar pants for hire :D [08:39] plants* [08:40] heh [08:41] oh, crap,, i didnt write the endorsement ... [08:41] * ogra_ blushes [08:42] ogra_: :) [08:58] sil2100: when can we expect silo 4 packages published so that we are really unblocked to build the other silos [09:00] Saviq: I understand you had a super new landing with new suru icons for some of the system apps? [09:00] thostr_, once the clicks have been produced and landed for the apps in there [09:00] Saviq: the apps which are click packages need uploading to the store, because some of your changes aren't going to land in the phone image otherwise. [09:00] popey, yeah I know [09:01] popey, it only landed (not completely yet) this morning [09:01] ok. xnox can probably help you upload the clicks. [09:01] ah okay ☻ [09:01] super [09:01] popey, I'll leave it to bfiller, not sure how they build the clicks [09:03] Saviq, hmm, we wanted to build an image with the changes before landing other stuff ... [09:03] ogra_, I really dunno how they build the clicks :| [09:03] (i guess we need to re-work the landing process a bit that clicks land alongside in the future) [09:04] Saviq, there should be a jenkins build somewhere afaik [09:04] (now dont ask me where :P ) [09:04] ogra_, where? [09:04] hah [09:05] ogra_, anyway [09:05] ogra_, the app changes are minimal [09:05] ogra_, just the icons [09:05] well, do they still work with the old ones after the trheme switch ? === vila_ is now known as vila [09:06] ogra_, yeah, the old theme is a fallback for the new one [09:06] sil2100, ^^^ [09:06] Saviq: i know where ☻ [09:06] so i guess we could start an image build and just live with the old icons til the clicks land [09:06] oh [09:06] http://s-jenkins.ubuntu-ci:8080/view/click/ [09:06] which specific apps and I'll get you clicks [09:07] popey, https://launchpad.net/~ci-train-ppa-service/+archive/landing-004 [09:07] {address-book,camera,dialer,gallery,messaging,notes}-app [09:07] not all of those are clicks [09:07] only camera, gallery, notes [09:07] popey, yup, we only need those [09:08] ok, will get them for you [09:08] thostr_: we're working on getting them migrated stilll, but I suppose around noon it should be all ok [09:08] rsalveti, we got a regression in autopkgtest for uitk-gles [09:08] oh, hang on [09:08] these changes haven't landed in trunk? jenkins wont do that [09:08] needs bfiller.. [09:08] ogra_, Saviq: good to know this [09:09] popey, yeah, the silo didn't land yet [09:09] sil2100, hmm, didnt you add code that prevents landing of unmerged changes ? [09:09] popey, blocked in proposed [09:09] oh just not merged back [09:09] k [09:09] what is blocking it ? [09:10] ogra_: I added code that prevents landing unapproved changes [09:10] sil2100: https://bugs.launchpad.net/ubuntu/+source/urfkill/+bug/1337227 is the rfkill crash bug [09:10] Ubuntu bug 1337227 in urfkill (Ubuntu) "urfkilld crashed during security tests on touch image" [Undecided,New] [09:10] psivaa: thank you! [09:10] yw :) [09:11] ogra_, uitk-gles [09:11] rsalveti, ^^^ [09:11] ogra_, it's uploaded already [09:12] oh ok [09:12] ogra_, but there's a regression in one of the autopkgtests [09:12] but from what I can see it could just use a re-run if possible [09:12] 'cause it's some timing issue (and passed on i386) [09:12] right i see that [09:13] who can re-run these ? release team ? [09:13] pitti seems to have already done so [09:13] http://d-jenkins.ubuntu-ci:8080/view/Utopic/view/AutoPkgTest/job/utopic-adt-ubuntu-system-settings-online-accounts/76/console [09:13] oh, cool [09:13] yeah, it's a flaky test [09:13] autopkgtests and flaky tests is nasty [09:14] definitely :/ [09:15] fail again :| [09:16] sigh [09:16] psivaa, what are you going to modify, so i know what to give you. is it what's under /home/phablet/autopilot/? [09:17] brendand: let me check where webbrowser tests are copied to in the device when running the tests [09:20] brendand: /usr/lib/python3/dist-packages/webbrowser_app/tests/ [09:20] SUCCESS [09:20] ogra_, bug #1336650 btw [09:20] bug 1336650 in ubuntu-system-settings-online-accounts (Ubuntu) "some autopilot tests seem to be flaky" [Undecided,Confirmed] https://launchpad.net/bugs/1336650 [09:21] Saviq, https://code.launchpad.net/~elopio/ubuntu-system-settings-online-accounts/clean_tests/+merge/225437 [09:21] perhaps ? [09:22] Saviq, those aren't flaky, they're downright broken [09:22] brendand, yeah, sure :) [09:22] Saviq, me and elopio were looking at them yesterday [09:23] brendand, just looking at elopio's MP [09:26] psivaa, ok you need to replace the __init__.py there with this: http://paste.ubuntu.com/7741088/ [09:27] brendand: ok, let me first run webbrowser on a separate device and then copy this over [09:39] Saviq: it seems it's all sorted out already, right? [09:40] rsalveti, yes [09:40] rsalveti, as you were [09:41] looks like it should migrate real soon === psivaa changed the topic of #ubuntu-ci-eng to: Ubuntu CI Engineering Team | Vanguard: psivaa | CI Train Status: #106 promoted | CI Train Sheriff: robru | Known issues: The spreadsheet is extremely slow, http://people.canonical.com/~rbpark/citrain/ knows the statuses before the spreadsheet does. Long mako wait times on s-jenkins. [10:14] brendand: the tests failing still with the ^ __init__.py : http://q-jenkins.ubuntu-ci:8080/job/psivaa-utopic-touch-mako-smoke-daily-webbrowser/4/console [10:18] psivaa, ok [10:59] sil2100, i reproduced a web browser failure! [10:59] brendand: HOLY SH..! [10:59] brendand: how? [10:59] sil2100, it happens if the orientation is wrong [11:00] ah ha! So it's related to the thing ogra_ mentioned? [11:00] sil2100, the phone was flat but when AP launched the browser it was landscape [11:00] Ah [11:00] hm [11:00] sil2100, didn't he say it only stayed like that temporarily? [11:00] it flicks for a moment [11:00] but i'm holding it in portrait while that happens [11:02] ogra_, no this is lying flat, and it's stuck the whole time on landscape [11:02] * popey notes the FirefoxOS lab has all their phones upside down it seems. http://imgur.com/a/eOIeu [11:02] right, might be the same but i'm not seeing it because i hold it in portrait where it immediately switches to [11:02] which is odd [11:03] ogra_, when i ran it locally in the morning it stayed in portrait. i wonder why not it suddenly wants to be in landscape? [11:04] ogra_, except where i hold it up in front of me [11:04] ogra_, then it stays portrait [11:04] right [11:04] and for me it sometimes flicks half way into landscape and flicks back immediately [11:04] but if i lay it flat, right now it is always in landscape [11:05] so why can't the toolbar be revealed in landscape? surely that's a bug [11:05] as i said, ricmm knows about it and will look after the sprint ... not sure if we can do much on our side here [11:05] because where the toolbar would be the shell has either the launcher or the app selector [11:06] ah yes [11:06] and overrides ... [11:06] ogra_, so we need to talk to the lab people and try and make sure the devices are standing [11:06] and not take any excuses about 'we might not catch orientation related bugs' [11:07] well, we need ricmm to make the rotation use the device default on app startup [11:07] as the proper fix [11:07] but yeah, having them upright might work around it for now [11:07] ;) [11:08] sil2100, ogra_ - at least we know why now :) [11:08] yeah [11:08] Yeah... [11:08] i'll raise a webbrowser-app bug === MacSlow is now known as MacSlow|lunch [11:13] seb128, is there anything else blocking silo 4 in proposed? [11:14] * Saviq has no idea where to look [11:14] https://bugs.launchpad.net/webbrowser-app/+bug/1337284, for the landing email [11:14] Ubuntu bug 1337284 in webbrowser-app "Many autopilot tests will fail if the orientation of the app is landscape on startup" [Undecided,New] [11:14] I just checked and it seems to make ubuntu-touch uninstallable [11:14] Saviq: ^ [11:14] so hmmm [11:14] I still feel like there might be some other issues with webbrowser app, but we need to get that one out of the way to make it easier to diagnose the others [11:14] Saviq, http://people.canonical.com/~ubuntu-archive/proposed-migration/update_output.txt still lists ubuntu-touch unhappy in i386/amd64 [11:15] let me try in a pbuilder [11:17] pbuilder? you're so 2010 [11:18] * sil2100 is also using pbuilder [11:18] What is the new hip way? [11:18] ;) [11:20] Saviq, bah [11:20] sil2100, schroot/sbuild [11:21] Saviq, the issue is unity8 [11:21] + qtdeclarative5-ubuntu-ui-toolkit-plugin (>= 0.1.48), [11:21] on i386 qtdeclarative5-ubuntu-ui-toolkit-plugin-gles Provides qtdeclarative5-ubuntu-ui-toolkit-plugin [11:21] but provides are not versionned [11:21] seb128, ugh [11:21] so | [11:22] so it can't resolve the >= requirement [11:22] seb128, can you upload a fixed package directly? I'll merge back to trunk [11:22] Saviq, yep, seems to be what rsalveti did to other components [11:22] Saviq, unity8? sure [11:23] seb128, ubuntu-settings-components has the same [11:23] Saviq, k, doing that as well [11:23] * Saviq looks through the other merges [11:24] brendand: so i'll release the device to the pool that i allocated to webbrowser testing? [11:25] psivaa, i don't know what that means :) [11:26] ;) [11:26] psivaa, which device, which pool? [11:26] seb128, yeah, that's it [11:26] brendand: ohh? [11:26] brendand: i was allocating a device to run this specific webbrowser test with your changes, [11:27] brendand: that's what i said in the meeting [11:27] psivaa, oh yes - we don't need that anymore [11:27] brendand: so out of curiousity.. if lying flat is causing the failure.. how comes you are not able to reproduce? [11:27] psivaa, i am now [11:27] ok [11:27] psivaa, it's not lying flat itself. it's running in landscape mode [11:28] psivaa, but for some reason, when lying flat, it will launch in landscape mode some/most of the time [11:29] Saviq, http://paste.ubuntu.com/7741516/ ? [11:29] brendand: ok, was wondering how missed that before :) [11:29] seb128, yeah [11:29] Saviq, looks good to you? [11:29] great, uploading [11:30] psivaa, yeah i'm not sure either [11:31] psivaa, here's what's stranger. when i launch it from the shell it launches in portrait mode, but the test launches in landscape [11:33] brendand: ohh? i assume AP does not enforce that? [11:33] Saviq, u-s-c uploaded as well, let's see how that goes [11:33] seb128, thanks [11:33] yw! [11:33] psivaa, i think i might have an idea why. still doesn't explain everything though [11:34] brendand: ack, leaving it to you :), good that you're able to repro locally [11:34] my thoery was wrong :) [11:48] ogra_, if you sit the device even a little bit up, like resting it on the edge of the laptop (mine is quite thin anyway), then it never opens in landscape [11:48] ogra_, so literally they just need to put a little phone-pillow in the lab for each device :) [11:51] heh [11:51] well, for me it still does half a rotation sometimes [11:53] ogra_, when you sit it up? [11:53] when i hold it upright and start a webapp [11:54] or the browser [11:54] ogra_, i haven't gone near web-apps yet [11:54] it doesnt happen every time, but often enough to notice throughout the day === alan_g is now known as alan_g|lunch === psivaa changed the topic of #ubuntu-ci-eng to: Ubuntu CI Engineering Team | Vanguard: cihelp | CI Train Status: #106 promoted | CI Train Sheriff: robru | Known issues: The spreadsheet is extremely slow, http://people.canonical.com/~rbpark/citrain/ knows the statuses before the spreadsheet does. Long mako wait times on s-jenkins. === psivaa is now known as psivaa-lunch === MacSlow|lunch is now known as MacSlow [12:37] mind if I assign a silo for line 44? should be able to land it fairly quickly [12:43] Is there a way to perform conflict-free merge of the changelog? I.e. Developers mod the changelog at/around the same location, and when their MPs get merged to trunk, they could step on each others' toes. Any pointers? [12:51] cjwatson: if you give me ~10min I can fix the branch for the hook test otherwise I can assign it [12:51] mvo: oh, sure, go ahead. (and I can assign too :) ) === psivaa-lunch is now known as psivaa [12:52] mvo: hey! Maybe you could assign a silo for cjwatson? :) [12:53] And there seems to be one package for publishing which you could try [12:53] cjwatson: I need to learn it, I'm on landing team duty now [12:55] mvo: aha [12:57] cjwatson: branch updated [12:57] sil2100: let me find out which one that is :) [12:58] cjwatson: ah! You didn't set it to Ready: Yes! ;p [12:58] sil2100: how do I do that? just type in the third column? [12:58] but that says "auto-updated, do not touch" [12:58] cjwatson: no no, there's column I entitled "Ready(...)" [12:58] cjwatson: you need to set it to 'Yes' there :) [12:59] Once it's set like that, the landing team will be able to see it (and will get pinged by the CI Train choo choo as well) [13:00] sil2100: oh, away off to the right [13:02] Yeah, hidden from evil eyes [13:02] prodded [13:02] mvo: could you try assigning a silo? [13:03] sil2100, seb128 qofono has landed, can we move on Ken's branch now? [13:03] \o/ [13:03] sil2100: sure, givem me sec [13:03] pmcgowan: where's the branch? [13:03] pmcgowan, we can yes, it just needs somebody with free slots to pick that up I guess [13:03] e.g not me today [13:03] but I can have a look tomorrow === Ursinha changed the topic of #ubuntu-ci-eng to: Ubuntu CI Engineering Team | Vanguard: Ursinha | CI Train Status: #106 promoted | CI Train Sheriff: robru | Known issues: The spreadsheet is extremely slow, http://people.canonical.com/~rbpark/citrain/ knows the statuses before the spreadsheet does. Long mako wait times on s-jenkins. [13:04] mvo: I'll just merge that before I hit build then [13:04] cjwatson: cool, thanks, I assign a silo now [13:04] seb128, we are off in the US tomorrow otherwise I could test, but I ran his branch for two weeks prior [13:05] pmcgowan, ok, thanks, that's useful testing info [13:07] sil2100, could I get you to look at https://code.launchpad.net/~bregma/cupstream2distro/lp-1321755/+merge/225467 -- that bug is very high priority for us [13:07] cjwatson: your silo is assigned now [13:08] bregma: ACK, taking a look now === alan_g|lunch is now known as alan_g === josepht changed the topic of #ubuntu-ci-eng to: Ubuntu CI Engineering Team | Vanguard: josepht | CI Train Status: #106 promoted | CI Train Sheriff: robru | Known issues: The spreadsheet is extremely slow, http://people.canonical.com/~rbpark/citrain/ knows the statuses before the spreadsheet does. Long mako wait times on s-jenkins. [13:09] mvo: ta [13:15] sil2100, does the publication-migration job consider newer packages that go into destination, or does it require == version? [13:17] Saviq: it's looking for == version... [13:17] sil2100, ok, so we'll have to force M&C [13:17] "accepted: unity8" [13:17] Saviq: ah, since you published a new unity8, right? [13:17] YES [13:17] Yeah [13:17] :> [13:18] sil2100, yeah, we pushed fixed packaging [13:18] directly to distro [13:18] didn't want to delay any longer [13:18] Right [13:20] sil2100, seb128, it says "Pending" into release here https://launchpad.net/ubuntu/+source/unity8/7.90+14.10.20140701.2-0ubuntu2/+publishinghistory [13:20] sil2100, can I M&C? [13:20] Saviq, you can m&c [13:21] WHOO HOO HOO HOOOOO [13:21] Saviq: I would say it's a good time now for this, yes! [13:21] Saviq, well done ;-) [13:21] I can push a button! I can push a button! [13:22] * Saviq needs therapy [13:22] popey, stuff's pushing to trunk now, will jenkins click pick them up automagically? [13:22] josepht: did I start vanguard earlier than I should? [13:23] Saviq: if not, I can poke it to build them [13:23] Ursinha: I think so [13:23] popey, it's pushed, so you might as well [13:23] josepht: oops :) doing this almost everyday is confusing hehe, sorry [13:23] Ursinha: no worries :) [13:24] Saviq: all of them? camera, gallery, notes? [13:24] popey, yes [13:25] Saviq: all building now [13:26] popey, rad [13:26] sil2100: in your next email can you move the ringer bug out of the blocker section and into the other section, and the dialer app issue I hav e confirmed I can not reproduce on 111 so might of been a glitch in an image but seems fine now :) [13:26] the funny thing in "rad" is it means "happy" in PL as well ;) [13:27] only not as extremely [13:28] o/ hi, can i have a silo for line 45? [13:28] I did think we'd stepped into the 80's. [13:29] popey: I've seen you cartoon and music tastes you never left the 80's, same as me ;) [13:30] heh [13:31] dbarth: sure [13:34] dbarth: you should have landing-018 === sil2100 changed the topic of #ubuntu-ci-eng to: Ubuntu CI Engineering Team | Vanguard: josepht | CI Train Status: #106 promoted | CI Train Sheriff: sil2100, mvo | Known issues: The spreadsheet is extremely slow, http://people.canonical.com/~rbpark/citrain/ knows the statuses before the spreadsheet does. Long mako wait times on s-jenkins. [13:45] bfiller: I confirmed I couldn't reproduce the issue with the dialer app in 111 so marked it invalid [13:45] davmor2: good [13:49] bregma_: ok, I'm checking out the branch right now and it seems more or less ok, although I think there can be multiple commiters and authors, yes? [13:49] bregma_: Didier was splitting that by ',' [13:50] But I'm not sure now if that's a real case in bzr [13:53] that's a nice settings landing ;-) [13:53] sil2100, mm, I had thought about that case and I see I forgot about it in my solution, let me go back and think for a bit again [13:58] mvo: this seems somehow appropriate https://www.youtube.com/watch?v=we7UfubNeqM === pete-woods is now known as pete-woods-lunch [14:10] Saviq: http://people.canonical.com/~alan/clicks/ [14:11] Saviq: if those three clicks match the bzr revs you want uploaded, then ping balloons and he can upload to store for you ☻ [14:15] davmor2: hehe, yeah, also we play different music now it seems :) [14:15] popey, balloons, it's good, please upload [14:15] ogra_, do we need to upload anywhere else or is store enough for preinstalled? [14:16] Saviq, sergiusens usually cares for landing clicks ... [14:16] iirc he needs to update a list [14:16] Saviq: ogra_ I don't take care of the daily landing of clicks; each owner does that now [14:16] mvo: well r 'n' b and rock 'n' roll they both got 'n' s in the middle so they can't be that different right? [14:17] I only add/remove from the image and help when it comes to it. [14:17] sergiusens, ah, so click_list is updated automatically ? [14:17] davmor2: :) [14:17] sergiusens, ok, I'll ping bfiller [14:17] Saviq, popey which clicks? [14:17] ogra_: if it's approved in the store; yes [14:17] balloons, http://people.canonical.com/~alan/clicks/ [14:17] Saviq: what do you need? [14:17] the three there [14:17] balloons: ^ [14:17] sergiusens, ↑↑ [14:17] ↑ [14:17] Saviq: ah, balloons is taking care [14:18] Saviq: we'll be release gallery and camera clicks today, so no need for you guys to do it [14:18] about to land some changes on those [14:18] heh [14:18] balloons: ignore us. [14:19] popey, done.. :-) [14:19] balloons: you can do notes-app if you don't mind [14:19] I'm not touching that one [14:20] bfiller, FWIW ogra_ wanted an image with just the suru switch [14:20] well, the landing team did :) [14:21] i only spoke up :) [14:21] ;) [14:22] Saviq, ogra_ : up to you guys, we are going to land some silos today regardless and don't really want to delay [14:22] popey, "You don't have permission to access /~alan/clicks/com.ubuntu.notes_1.4.272_armhf.click on this server." [14:22] bfiller, I don't think it affects you in any way [14:23] bfiller, it's just a question whether there will be the interim (new icon only) versions in the next image or not [14:23] ogra_, ↑ your call guys [14:24] * Saviq hates queuebot for resizing my window all the tie [14:24] time [14:24] Saviq, bfiller, well, i dont reallly care for that specific image, what i do care for is that our landing process is broken, the silo should never have landed without the clicks being available alongside, we need to redefine the policy here ... i suspect the icon updates wont cause issues anyway [14:25] balloons: hang fire [14:26] Saviq: I just dropped printing the old status so that should save some space, I'll also switch it to printing less stuff and shorter status messages using the new state values introduced by sil2100 yesterday [14:28] ogra_, that's agreed === Ursinha changed the topic of #ubuntu-ci-eng to: Ubuntu CI Engineering Team | Vanguard: Ursinha | CI Train Status: #106 promoted | CI Train Sheriff: sil2100, mvo | Known issues: The spreadsheet is extremely slow, http://people.canonical.com/~rbpark/citrain/ knows the statuses before the spreadsheet does. Long mako wait times on s-jenkins. [14:28] stgraber, it's actually just the name "-queuebot/#ubuntu-ci-eng-" that causes the handle column to resize [14:29] stgraber, so it's probably xchat-gnome's fault [14:29] Make your client handle notices better, indeed :) [14:29] bfiller, Saviq, the impression in the landing meeting this morning was that everything was landed so we decided to have a stop gap image build ... when we decided that we didnt know something was stuck in proposed and the clicks werent there, i dont think that image is overly important [14:30] Saviq: ah, right, some IRC clients do that for notices [14:30] Saviq, just ise a non-castrated xchat then ;) [14:30] Mine has a similar display for notices, but it doesn't try to align the start of everybody's text to the same left margin [14:30] ogra_: yes we really need clicks to be released by the silo [14:30] Saviq: irssi doesn't have that problem (by not printing a user list) :) [14:30] and plain xchat has an option for this ;) [14:31] cjwatson is way too fast for me today... :) === alan_g is now known as alan_g|tea [14:40] sil2100: silo 2 ready for publishing [14:41] o/ [14:41] mvo: want to land this as well ^ :) [14:41] sil2100: sure [14:41] Thanks! === alan_g|tea is now known as alan_g === greyback_ is now known as greyback|post === pete-woods-lunch is now known as pete-woods [15:20] robru, password auth is disabled ... nontheless my ssh key works everywhere if i ever attached via phablet-shell (even on different devices) [15:21] ogra_, what?phablet-shell copies your ssh key down when you connect, but it's not magical across devices. not sure what you maen [15:21] i.e. my laptop has never connected to the device in my basement directly ... only to my flo via phablet-shell ... nontheless i can ssh into a freshly flashed test device i never had attached to [15:22] ogra_, define 'freshly flashed', your ssh key is stored in /home/phablet/.ssh so it won't get wiped in a flash unless you choose --bootstrap --wipe [15:22] robru, thats surely another bug ... nontheless logging out of phablet-shell should not leave sshd runnign ... an dphablet-shell shouldnt fire up a global ssh daemon [15:23] robru, i flash with --bootstrap ... but that device was never connected to my laptop i still can ssh in [15:23] ogra_, well your ssh key got in there somehow [15:23] that device never left my basement in the last year ... and i never connected directly to it [15:24] anyway, this is not about keys [15:24] but about sshd [15:24] ogra_, I don't know how to make sshd only listen on local interface, there doesn't seem to be an option for that [15:25] you use the cmdline options for it and start it directly instead of using upstart [15:25] ogra_, please show me the commandline option that says 'only connect on lo' because i don't see it in the man page [15:26] ogra_, looks like I need to munge sshd_config or something [15:26] robru, dunno, i know there is a "listen address" option you can use in the config [15:26] so there should be an equivalent on the cmdline [15:26] $ adb shell /usr/sbin/sshd -4 -o 'Port 22' -o 'ListenAddress 127.0.0.1' [15:26] /etc/ssh/sshd_config line 5: ports must be specified before ListenAddress. [15:27] ogra_, ^ it doesn't work super well [15:27] use -p [15:27] $ adb shell /usr/sbin/sshd -4 -p 22 -o 'ListenAddress 127.0.0.1' [15:28] robru, also dont use 22 please [15:28] Could not load host key: /etc/ssh/ssh_host_ecdsa_key [15:28]