=== ChanServ changed the topic of #ubuntu-uos-core to: Currently no events are active in this room - http://summit.ubuntu.com/uos-1511/core/ - http://irclogs.ubuntu.com/2015/11/04/%23ubuntu-uos-core.html === dholbach_ is now known as dholbach === hikiko is now known as hikiko|ln === hikiko|ln is now known as hikiko [14:49] slangasek, do you know who's running the node.js/libv8 session? === ChanServ changed the topic of #ubuntu-uos-core to: Track: Core | node.js and libv8 for 16.04 | Url: http://summit.ubuntu.com/uos-1511/meeting/22590/nodejs-and-libv8-for-1604/ [14:55] slangasek, ping? [14:55] alecu, do you maybe know who's running the next session? [14:58] dholbach: it says "Created by: Matthias Klose" [14:58] ok... but he's not online [14:58] slangasek, ^ [14:59] I can set the session up [14:59] but I'm not sure what to say................. [15:00] you could show a slideshow of nice V8 cars :P [15:01] dholbach: I guess somebody from distro and the people in security team that are marked as attending the session need to be in that [15:01] setting it up now [15:01] dholbach: I'm joining because we are using node.js and v8 for the new scope bindings. [15:01] ok, give me another sec or two [15:02] can chrisccoulson and I join the hangout? [15:02] sure [15:02] https://plus.google.com/hangouts/_/hoaevent/AP36tYeYeH42x7l09uefcnXCEwsjsCdU5h69oxDhB-7mgeYodV05bA [15:02] great [15:03] dholbach: sorry, was working on getting it set up; I'll join that one [15:04] we'll start in a minute [15:07] Notes are up here: http://pad.ubuntu.com/uos-1511-nodejs-and-libv8-for-1604 [15:12] yes node embeds v8 it does not depend on a shared v8 [15:14] we are embedding node into our own binary [15:14] as it is not available in universe or main [15:15] our executable embeds node and is shipped within a scope click [15:16] we take on the responsibility of keeping the node we embed up to date [15:18] we update node, we fix our internals, and our API/ABI to remain stable [15:18] marcustomlinson, if you want to join the hangout, let me know [15:19] (sorry I have bronchitis, voice is broken) [15:19] ok, no worries [15:19] I relayed your comment [15:27] yes, v8 code is litterally part of the node source tree [15:28] https://github.com/nodejs/node/tree/master/deps/v8 [15:28] they pull it in statically, no dynamic linking [15:28] get to feeling better, marcustomlinson :) [15:29] tyhicks: will do [15:29] :) [15:29] yeah, all the best! [15:50] doko, do you know if barry is running the next session? [15:51] dholbach, we will be both there, he created https://blueprints.launchpad.net/ubuntu/+spec/foundations-x-python3-only [15:51] ok... is anyone of you setting up and running the session? [15:51] https://wiki.ubuntu.com/UDS/Sessions [15:52] hey barry [15:52] are you setting up and running the session? [15:52] dholbach: hi [15:52] https://wiki.ubuntu.com/UDS/Sessions [15:52] yeah, i guess i should go to the other room now (we're sprinting [15:53] ok cool [15:53] let me know if you need help [15:55] dholbach: technically it's doko's session :) === ChanServ changed the topic of #ubuntu-uos-core to: Track: Core | Python3 Only on the images | Url: http://summit.ubuntu.com/uos-1511/meeting/22568/python3-only-on-the-images/ [15:55] ok... [15:56] I asked doko earlier [15:56] I wasn't quite sure who was running it [15:58] https://plus.google.com/hangouts/_/hoaevent/AP36tYdzlG929T3_3v8ifrDi3Hp2h0Tz8wZVRcdk07fcQA61nJdA6g?hl=en&authuser=0 [15:58] I think the time of the session was set wrong [15:59] the video currently says 9:00:45 [15:59] barry: ^^ [15:59] but maybe it doesn't matter and you can just hit start and it'll just work [15:59] dholbach: hmmm I said 'starting now' :) [15:59] I don't know <3 [15:59] oh ok [15:59] that's on http://summit.ubuntu.com/uos-1511/meeting/22568/python3-only-on-the-images/? [16:00] 8:59:50 now [16:00] dholbach: when creating the hangout [16:00] so anyway, I don't know [16:01] maybe just start the hangout and see if it works [16:02] https://bugs.launchpad.net/ubuntu/+bugs?field.tag=python3 [16:02] https://blueprints.launchpad.net/ubuntu/+spec/foundations-x-python3-only [16:02] ok, seems to be working :) [16:03] yes [16:03] can hear you [16:03] go go go [16:03] :) [16:03] https://docs.google.com/spreadsheets/d/1PnPLZRM2hu7ucv1yuq7Fd5hjNw_DYupHIYZpoP4-xHk/edit?pli=1#gid=1918227221 [16:05] landscape-client-ui-install can be purged. It is unsupported today [16:05] smadden_landscap: oh great :) [16:06] smadden_landscap: did you want to join us on the hangout, or are you just delivering the good news? [16:06] sure, I'd like to join [16:06] smadden_landscap: //plus.google.com/hangouts/_/hoaevent/AP36tYdzlG929T3_3v8ifrDi3Hp2h0Tz8wZVRcdk07fcQA61nJdA6g?hl=en&authuser=0 [16:06] Trevinho, any ideas on what Python2 is needed by U7? [16:08] willcooke: it's needed by Autopilot, but not by unity itself I think [16:08] let me check whether the unity script needs, it but I don't think so [16:08] or the migration scripts.. [16:08] willcooke: tools/unity.cmake [16:09] the unity scope depends on software-center in my testing (or at least apt daemon) [16:09] willcooke: LP: #1512909 [16:09] (there's a patch for that but it doesn't build yet for unrelated reasons i think) [16:10] Trevinho, ^^ [16:13] you can authenticate to the rnr server and or the store without using the ubuntu-sso-client fairly easily [16:13] it's a few oauth calls [16:14] unity scope applications gets its data from the software-center database though.. [16:15] thanks demmot [16:15] seb128, I'll speak to robert_ancell about that ^^ [16:15] willcooke, k, thanks [16:22] rdepends also shows python-apt [16:22] hi [16:22] yes I think that g-s uses xapian [16:23] via libappstream [16:23] you can see they are linked [16:23] seb128: slangasek [16:23] Laney: thanks [16:25] Laney, thanks [16:26] https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=781913 [16:27] ? [16:27] https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=781913 [16:34] barry, doko: we seem to have skipped over system-config-printer on the desktop? [16:34] sorry, my hangouts connection failed, so the broadcast will have been interrupted [16:36] hey [16:36] oh.. to come back to it. [16:36] cloud images were free of python2.7 [16:36] about checkbox [16:36] but not libpython2.7 [16:36] python2 is only needed by one test [16:36] there was a vim bug /dep on libpython2.7 [16:36] we can drop that test from the desktop [16:36] it's OpenCV related AFAIR [16:37] https://bugs.launchpad.net/bugs/1512642 [16:37] https://bugs.launchpad.net/ubuntu/+source/checkbox/+bug/1512641 [16:37] we can easily separate that as checkbox and the tests are separate packages [16:37] https://bugs.launchpad.net/ubuntu/+source/checkbox/+bug/1512642 [16:37] checkbox and plainbox are pure python3 [16:37] https://bugs.launchpad.net/bugs/1512641 [16:37] smoser, we could just build vim with python3 bindings [16:37] https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=729924 [16:37] we requested the removal of the checkbox package to replace it with a new stack that uses SDK apis [16:37] infinity had some comments on that (vim and python3) [16:37] so please stay reassured that checkbox will by python3 clean :) [16:38] zyga: thanks [16:39] :-) [16:39] https://bugs.launchpad.net/ubuntu/+source/libpeas/+bug/1440504 [16:40] Well, from a kubuntu side we can look at things ourselves, but we're also affected by stuff like apt-xapian-index or ubuntu-sso-client, so some coordination place would be nice [16:42] barry, there's delay [16:43] (i've verified and updated etherpad. 15.10 server install with default ends up with only libpython2.7-*, and that is rdpended on only by vim) [16:43] http://paste.ubuntu.com/13102639/ [16:43] cloud image now has lxd -> lxc -> cloud-image-utils . and smoser will drop/fix cloud-image-utils need of python2 [16:46] plan is to drop ubuntu software center really soon? [16:47] nah, the bugs are probably enough [16:47] we can just mark those as affecting it [16:47] yofel: cool thanks [16:47] *affecting us [16:47] lubuntu bug - https://bugs.launchpad.net/ubuntu/+source/lubuntu-meta/+bug/1465313 === abc is now known as Guest34615 [16:48] samba and system-config-printer seem like the big ones that are cross flavor issues [16:49] https://gist.github.com/smoser/8904199bb8f00a90dd04 [16:52] are you familiar with this technique? http://paste.ubuntu.com/13102740/ [16:53] (we may have talked about this before, smoser ?) [16:53] i did see that. [16:53] slangasek, yeah. [16:53] its interesting. [16:53] curtin has a solution too === ChanServ changed the topic of #ubuntu-uos-core to: Currently no events are active in this room - http://summit.ubuntu.com/uos-1511/core/ - http://irclogs.ubuntu.com/2015/11/04/%23ubuntu-uos-core.html [17:04] > === ChanServ changed the topic of #ubuntu-uos-core to: Track: Core | EFI Capsule Update and Fallback | Url: http://summit.ubuntu.com/uos-1511/meeting/22637/core-1511-efi-capsule-and-fallback/ [17:59] http://youtu.be/I4ci-GeJIQw [17:59] ^ for anyone who wants to follow, we'll watch questions on IRC too of course [17:59] who is here for this session [17:59] ? [18:00] cyphermox: can we share the hangout url here too for anyone else who wants to join the discussion? [18:00] yeah I was just grabbing that [18:00] https://plus.google.com/hangouts/_/qmal77cb6ezen6vpzuovvs3lkua [18:00] ^ the hangout URL [18:01] I won't make it broadcast unless we aren't the only ones here for the session though [18:01] cmagina: hi, here to talk about capsule updates? [18:01] nespresso ! [18:02] ogra_: ? [18:02] (no, wait , different capsules) [18:02] heh, nespresso capsules [18:02] :) [18:02] eep. [18:02] I'm here :) but not on hangouts/whatever [18:02] slangasek: curious about it :) [18:02] cyphermox: turns out I can't watch it anyway due to some video compat problem :P [18:02] sorry this capsule is not signed with the right key [18:02] alright, let's give it a go then [18:02] haha [18:02] pjones: shoulda run Ubuntu [18:02] ;) [18:03] Interested in this since I've had some experience with capsule updates - painful [18:03] slangasek: oh yeah, that's probably the issue :P [18:04] http://pad.ubuntu.com/uos-1511-core-1511-efi-capsule-and-fallback [18:05] I looked at the blueprint; i'd suggest the capsule path be /usr/share/uefi/capsule/ or similar, rather like /usr/share/misc/, rather than /opt// [18:06] https://blueprints.launchpad.net/ubuntu/+spec/core-1511-efi-capsule-and-fallback [18:07] https://blueprints.launchpad.net/ubuntu/+spec/foundations-w-uefi-capsule-update [18:09] https://secure-lvfs.rhcloud.com/ [18:10] is the URL [18:10] Also, with the fwupdate tool, some easy-to-use mechanism to delegate the update work to a BIOS vendor's own tooling, such as H2O-FFT [18:10] fwupd puts it the appstream data in /var/cache [18:11] (also https://beta-lvfs.rhcloud.com/ may be more interesting) [18:12] and of course http://www.fwupd.org/ being most interesting [18:14] gnome-software [18:14] has support for fwupd [18:14] already [18:14] once fwupd clears NEW in debian we were planning to enable it in Debian with gnome-software too [18:19] the vendor specific tooling isn't supposed to be activated until UpdateCapsule() is called. it's all a vendor specific implementation at that point [18:22] TJ-: do you have more questions? [18:22] quigley: and you? [18:22] Idea rather than question: Being able to preload a portable image with many capsules and use it on multiple systems should be a use-case to cover. Enterprises may not want to risk their users handling/invoking a firmware update, but have it done either remotely, by PXE boot image, or USB mass-storage [18:23] mass storage is already supported [18:23] fwupdmgr install blah.cab [18:23] I don't see why that wouldn't just work [18:23] you could already do that, as long as the device loads the efi binary for fwupdate [18:23] TJ-: from an upstream perspective, we've been assuming the enterprise will have the ability to remotely schedule it without the user involved, but it still runs through the local userland software [18:24] Good. just didn't see anything specific mentioning it [18:24] at least today most BIOS'es don't support PXE in UEFI mode [18:25] you have to load a CSM [18:25] (pxe booting it won't *quite* work yet; we need a local filesystem currently. but it's on my todo without a lot more work to do on it.) [18:25] but it can be done [18:25] superm1: lots do support dhcp+tftp (as opposed to the bios "pxe" api), but yeah. [18:25] oh right, because we look locally for the firmware images? [18:25] cyphermox: right; right now we don't support network device paths [18:25] yeah [18:25] yeah, but that's also assuming a working network stack in the BIOS at that time too [18:26] I have like 80% of that in my local tree from this friday. [18:26] superm1: yes, it is. [18:26] The only thing that occurs to me, based on the Whiteboard... is there a Process: Option 0 = a check-script that looks up the system ID against available updates and lets the user know... which would lead to Options 1/2 [18:27] in other words, prior to a download action by the user, to make them aware without them needing to initiate [18:29] TJ-: right now you need to run fwupdmgr refresh to look for updates, it could be cronned [18:30] but in the future we might want to get that done graphically/automatically some other way [18:30] I think if you're using gnome-software it'll do that for you? [18:30] ie. gnome-software that we mentioned earlier [18:30] well, and in fact if this is going to be using the appstream service and using gnome-software, there will need to be integration for that [18:30] * pjones isn't really sure about that end of the stack [18:30] today we use a cronjob + update-notifier to display the status of update availability [18:30] superm1: hey, I'm curious about pxe and uefi mode, why is that not supported? [18:31] slangasek: it's not like we couldn't do what software updates do already, look every once in a while and ask the user ifthey want to apply the new updates [18:31] I don't know in the gnome-software world what polls for updates, but that would surely need to poll this service [18:31] zyga: many firmwares don't support it, and those that do typically only recognize the addressing if they've been configured to boot via the network [18:32] zyga: so even if they know about it, using the network to fetch the image without actually /booting/ from it won't work without a lot more code. [18:32] right [18:32] I see [18:32] zyga: so basically because hardware vendors believe netboot is a value add and don't support it on their low-end machines. [18:32] zyga: traditional PXE support as we all know and love comes from the CSM generally (legacy boot mode) [18:33] UEFI netboot operates a bit differently in what you'll actually be booting [18:33] I'm interested in using the UEFI network stack [18:33] to do $stuff [18:34] where stuff is related to replacing what I'm booting [18:35] zyga: at the risk of being too helpful to ubuntu people, https://fedoraproject.org/wiki/QA:Testcase_UEFI_pxeboot :) [18:36] except that was written for our hacked up grub 0.97 efi binary; substitute a grub.cfg for the obvious difference. [18:36] pjones: interesting, I was thinking about something slightly different but I will check it out to lear more about it [18:36] pjones: I was thinking about using the network stack to boot one or other local disk [18:37] pjones: not about complete netboot [18:38] cron-job + update-notifier approach would be good especially for headless hardware, and to support non-gnome flavours [18:39] TJ-: yeah a script around fwupdmgr refresh and grep fwupdmgr get-updates would be sufficient for that. when you want to do an update it's just fwupdmgr update [18:40] longer-term, something like the debian/control Modaliases: might be useful too, to search archive packages using ubuntu-drivers [18:42] TJ-: superm1: if either of you wanted to help with that in fwupd , a "give me esrt info for this system" and a "look up info for this arbitrary ESRT dataset" call would let you split that up pretty nicely... [18:42] Good to see the progress on this; seeing far too much firmware-bug induced user issues the past 12 months - that combined with ACPI firmware bugs [18:42] it would be easier to teach ubuntu-drivers to parse appstream data in my opinion [18:43] so you could connect to dbus with whatever remote agent you have and ask for details, and then grab stuff on a common server end. [18:43] than to repackage all the data into dummy debian packages [18:43] superm1: right, just thinking about the common tool and how it currently works [18:44] Well, write what you like. I just hope we can avoid too much unneeded duplication of effort. [18:44] Yeah agree === ChanServ changed the topic of #ubuntu-uos-core to: Currently no events are active in this room - http://summit.ubuntu.com/uos-1511/core/ - http://irclogs.ubuntu.com/2015/11/04/%23ubuntu-uos-core.html [18:59] secureboot sign for ubuntu EFI [19:00] is there any method to implements signature for EFI with ubuntu EFI file?