/srv/irclogs.ubuntu.com/2015/11/04/#snappy.txt

Sam_Hi03:41
=== ubott2 is now known as ubottu
=== Ursinha-afk_ is now known as Ursinha
=== Ursinha is now known as Guest82331
=== ev_ is now known as ev
=== inaddy is now known as tinoco
=== rsalveti_ is now known as rsalveti
=== cprov_ is now known as cprov
=== retrack is now known as Guest53865
=== plars_ is now known as plars
=== moul_ is now known as moul
=== Guest82331 is now known as Ursinha
=== Ursinha is now known as ursula1234
=== ursula1234 is now known as Ursinha
* Guest42341 such a beautiful day, today. great for science and such05:22
liuxg_where can find the output messages from a running snappy app? thanks07:39
fgimenezgood morning08:11
kgunnstgraber: ping08:19
=== dholbach_ is now known as dholbach
mvo_ogra_: you may need to rebase your livecd-rootfs ppa changes with my latest upload (should be trivial though)09:20
ogra_mvo_, will do (i hope i get done today, implementing this is hell ... (our chroot is so messed up after build due to removing half the packaging system))09:46
JamesTaitGood morning all; happy Wednesday, and happy Stress Awareness Day! 😃10:14
=== nessita_ is now known as nessita
tedgOh man, I didn't send out stress awareness day cards. Now I'm stressed about it, but at least aware.14:20
mcphailHi. Do we have a target date for the phone to transition to snappy?14:24
mvo_jdstrand: hey, just fyi - https://github.com/ubuntu-core/snappy/pull/55 is up mostly to get input from Chipaca but you are welcome to look/comment/reply of course14:35
jdstrandmvo_: ack, thanks15:04
tedgmcphail: No, there isn't a date set yet.15:24
elopiofgimenez: do you think subunit Status should return the number of bytes written?15:31
fgimenezelopio, doesn't feel quite well, maybe we could get rid of all the writer stuff15:32
fgimenezelopio, i think that the only useful part now is the multiwriter15:32
elopiofgimenez: I agree. Your pipes structure was cool, but now I don't find a use for it.15:33
elopiobut before I get rid of it, I want to make sure that we would be able to put this subunit module in gocheck.15:33
elopiomaybe I'll try that first before doing anything crazy on our branch.15:33
fgimenezelopio, ok makes sense15:34
mcphailtedg: thanks15:47
stgraberkgunn: pong15:47
snappy_getting a "UbuntuClientIntegration: connection to Mir server failed" error when trying to run mir app.17:35
snappy_any1 else run into this issue?17:35
mvo_jdstrand: I added the check for changes in "apparmor, ubuntu-core-security-*" around the snappy policygen --regenerate-all now, do you happen to remember if anything else is missing featurewise?18:19
sergiusensmvo_, is there a way to setup APT::Install-Recommends false using python3-apt?19:02
mvo_sergiusens: yes, apt.apt_pkg.config.set("apt::install-recommends", "false")19:02
sergiusensmvo_, neat19:03
mvo_python-apt is quite good :)19:03
sergiusensmvo_, it is :-)19:04
sergiusensmvo_, btw, did you figure out my question from the other day?19:04
mvo_sergiusens: uh, what was the question? about apt update files? sorry, I forgot again, could you help me?19:05
sergiusensmvo_, about having .fetch_archives tell me what it downloaded19:05
sergiusensor what it is about to download19:06
sergiusensmvo_, either is fine19:06
mvo_sergiusens: aha, so there is definitely a list of the urls in the acquire system, I just need to have a look how to access it19:06
sergiusensmvo_, great, this will allow me to download once per project if multiple parts require the same deb19:08
mvo_sergiusens: ok, so the best way is to create a "fetcher" via apt_pkg.Acquire(progress) and with fetch.list you can iterate over the list of items that got downloaded19:09
mvo_sergiusens: oh, if that is your use-case just use a common download dir19:09
mvo_sergiusens: apt will re-use existing debs, i.e. not fetch them again19:09
sergiusensmvo_, oh, but I need to unpack each to its own private location19:10
sergiusensmvo_, unless I just download once and unpack into the stage area and forget about putting it in part.installdir19:11
mvo_sergiusens: right, unpack them to any location, just keep the apt.Cache(rootdir=./something) to the same "something" for all parts19:11
sergiusensalso a possibility; then we get the same phantom dep bugs as in debian packaging when using multiple parts masking missed deps :-)19:11
mvo_sergiusens: i.e. just keep the debs and the list in the same dir, unpack and handle in whatever way yu want19:11
mvo_sergiusens: well, you can of course iterate over a fetcher.items, each item will have a uri and a description19:12
sergiusensmvo_, ok, I'll look into it19:13
mvo_sergiusens: I will be off a bit, but just mail me any questions19:13
mvo_sergiusens: or telegram19:13
sergiusensthanks19:14
jerryGChipaca: what changes need to be made to get sockets working with freerdp?19:14
ChipacajerryG: I don't know; what isn't working?19:33
jerryGChipaca: pastebin.com/ZvU8yFgL19:36
ChipacajerryG: oooh, nice :)19:36
ChipacajerryG: so, first, figure out what system call is being blocekd19:37
Chipacablocked*19:37
Chipacasecond, have your wrapper script set XKB_CONFIG_ROOT so it finds the xkb bits, if you need them19:38
ChipacajerryG: but your first problem is that system call, clearly19:38
ChipacajerryG: so. sc-logresolve is probably what you want. Or sudo journalctl | grep -i audit19:40
ChipacajerryG: where do sockets come into it?19:44
ChipacajerryG: also, do you need sudo?19:48
tedgzyga: I'm trying to run a single plainbox test in Snapcraft, do you know how to do that?21:10
zygatedg: yes, one sec21:34
zygatedg: have you seen ...21:35
zygatedg: https://code.launchpad.net/~zyga/snapcraft/plainbox-app21:35
zygatedg: (I need to return to that soon)21:35
zygatedg: this allows you to do all of that21:36
tedgzyga: So I can't do it today?21:41
zygatedg: with that example you can, with the current code it is also doable, just uglier21:41
zygaone sec21:41
tedgWhy do we have all these custom wrappers? Feels weird this functionality isn't just provided as part of plainbox.21:42
tedgWe shouldn't have code to support running the tool :-/21:42
zygatedg: because it was written against old version that was in the archive21:43
sergiusenszyga, since you are here, can we have 'silent mode unless errors' for plainbox commands spitting output?21:43
zygasergiusens: I believe exactly that is implemented in the branch above21:43
zygatedg: plainbox is primarily a library and now it's very easy to create all kinds of tailored tools on top21:44
zygatedg: like the one I wrote there21:44
zygasergiusens: it runs tests and shows the error for each failing test21:44
zygatedg: plainbox run -i regexp-matching-test-id21:45
zygatedg: that runs a single thing21:45
zygatedg: in integration_tests/runtests.sh you can see we call plainbox run -T $test_plan21:46
zygatedg: you can run plainbox run -i blah manually21:46
zygatedg: for quick testing that's actually pretty useful21:46
zygatedg: you can run manage.py develop21:46
zygatedg: then all of the snapcraft integration test cases are available directly21:46
zygatedg: plainbox run -i test_id21:46
tedgzyga: Can you give me an example, that wasn't working for me.21:47
zygatedg: sure, let's see21:47
zygatedg: eh :)21:52
zygatedg:  SNAPCRAFT=snapcraft plainbox run -i 2015.com.canonical.snapcraft::snapcraft/normal/bzr-tag21:52
zygatedg: everything failed because $SNAPCRAFT was unset, it is set by the wrapper script21:52
zygatedg: I think scripts could just default to ${SNAPCRAFT:-snapcraft}21:52
zygatedg: or really make it always snapcraft and rely on PATH being set21:53
zygatedg: but that's just script design21:53
zygatedg: that worked for me _after_ I ran "./manage.py develop" to register the snapcraft integration testing tests with plainbox21:53
zygatedg: actually running a provider "from source" is one of the last snapcraft driven APIs that I haven't improved, I have a branch for that but I haven't worked on it since holidays (overloaded with other stuff)21:54
zygatedg: there's a branch that adds a provider unit so that manage.py becomes obsolete (apart from being a convenience for running stuff interactively)21:54
zygatedg: and then plainbox run could take a path to the provider as an argument, to let it just run without any extra setup21:55
tedgOkay, I think that I was missing the ./manage.py develop step21:55
zygatedg: and anther, unrelated, branch that adds environment units so that a job can sensibly depend on an environment variable and things like unset SNAPCRAFT won't have bad UX21:55
tedgI seem to get a test now.21:55
zygatedg: I will return to them after I settle down in the snappy core team and start feeling some free time on Fridays21:56
tedgThe test fails, but at least it runs :-)21:56
zygatedg: hehe :)21:56
zygastuff I've mentioned is in https://code.launchpad.net/~zyga/checkbox/config-units21:56
zygaand in https://code.launchpad.net/~zyga/checkbox/provider-unit21:56
zygasergiusens: can you confirm my assertion21:58
zygasergiusens: if there is something to actually do there I'd love to know21:58
sergiusenszyga, how far back do I have to read? I am all for not needing manage.py21:59
zygasergiusens: just about the interactive output on error thing you asked about22:00
sergiusenszyga, if it is already there, that's fine; I recall it from Budapest now22:01
sergiusenszyga, ideally it would live in trunk/master ;-)22:01
zyga::)22:01
zygasergiusens: yeah, belive me, I wish I had 34 hours, not 2422:01
zygasergiusens: catching up with family, checkbox, snappy and UOS now22:01
zygasergiusens: and $commercial_project22:01
zygasergiusens: fun22:01
jerryGChipaca: no.  I don't need sudo to reproduce error.  DO u want me to send you journalctl output?23:02
jerryGChipaca: sockets are used to connect w/ remote host23:03

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!