[02:55] care to do some code reviews? https://code.launchpad.net/~dholbach/snapcraft/1496392/+merge/271319 [02:55] elopio, ^ [02:55] elopio, and maybe this https://code.launchpad.net/~ted/snapcraft/python-pip/+merge/270959 [02:56] elopio, and not sure you know how, but tedg (and I scarcely tried as well) is trying to get rid of the shebang replacement. [02:56] that setuptools does on 'scripts' [02:56] well, not replacement, generation (we want the env python3 magic) [06:51] good morning [07:07] good morning [07:23] hello fgimenez [07:23] fgimenez: I haven't found any blockers. Here's what I've been doing: http://pad.ubuntu.com/snappy-testing-2015-09-16 [07:24] I couldn't find the right data to send to package POST and config PUT. Could you please ask Chipaca when he wakes up and test those two methods? [07:25] hey elopio, ok thanks a lot, let me send you the link to the rest api doc [07:25] ah, that would have been useful. [07:32] elopio, sorry, i didn't recall yesterday you would need it. i'll continue from here [07:32] elopio, I merged your dep8 branch into https://code.launchpad.net/~dholbach/snapcraft/1496392/+merge/271319 so it can benefit from both our work - one test is still failing - this is just a headsup - I know you're busy with testing right now [07:33] fgimenez: well, looking at the doc I still don't fully get how the config works :p [07:33] I'm not sure it would have been different. [07:33] dholbach: hey, I saw your branch. [07:33] actually, branches. Thanks a lot for your help. I hope I'll be fully on snapcraft tomorrow. [07:34] dholbach: on my branch I also had one or two failures. This is a good time to dig on that. Thanks for merging them too. [07:34] I'll add you as a branch reviewer :) [07:34] ah, you already are :) [07:35] fgimenez: your snapcraft branch will be affected by that too ^ [07:35] elopio, ok thanks i'll take a look [07:36] uggh, I've just tried to roll back from #9 to #11 on the bbb, and it ended back on #11. [07:38] now it worked. I should have gone to bed like 2 minutes ago :D [07:39] elopio, go to bed anyway! i'll take care of that [07:42] yes, the issue seems to have disappeared after closer inspection 8-) [07:42] I'm leaving now. See you again soon o/ [07:54] good night elopio! [08:12] good night elopio and thanks for the testing [08:32] elopio: fgimenez: ehlo. config works fine =) [08:33] elopio: fgimenez: but if you could elaborate your question a little bit more, i might be able to give you a more helpful answer [08:35] elopio: fgimenez: method 1 of 2 for POST to packages: POST -H "X-Allow-Unsigned: 1" http://localhost:8080/1.0/packages < /tmp/x.snap [08:36] Chipaca, great thanks :) i'll try manually the put and post methods, will ping you if needed [08:36] elopio: fgimenez: method 2 of 2 for POST to packages: POST -H "Content-Type: multipart/form-data" http://localhost:8080/1.0/packages < /tmp/mime-multipart-form-data-file [08:38] elopio: fgimenez: alternative to method 2 of 2 for POST to packages: http://pastebin.ubuntu.com/12412149/ [08:38] elopio: fgimenez: make yer browser do the work ;) [08:38] elopio: fgimenez: or you can have netcat listening on that port and copy the document body from the request [08:39] elopio: fgimenez: PUT to config: echo -e "config:\n ubuntu-core:\n autopilot: false\n" | PUT http://localhost:8080/1.0/packages/ubuntu-core.ubuntu/config [09:10] Good morning all; happy Thursday, and happy Apple Dumpling Day! 😃 [09:17] Chipaca, i'm getting this from the post to packages http://paste.ubuntu.com/12436434/ with the X-Allow-Unsigned: 1 header [09:18] Chipaca, i've built the package from config-example in snappy-examples, let me try with another one [09:19] fgimenez: could you try with an html based multi-part upload as well instead [09:19] * ricmm finds it [09:19] ricmm: linked him to it already [09:19] oh ok [09:19] * ricmm too slow [09:19] http://paste.ubuntu.com/12436497/ anyways... [09:19] ricmm, yep, i'll try that now [09:20] fgimenez: just tried it here and it worked [09:20] * Chipaca tries something else [09:21] so, yes it worked here [09:21] let me try with curl [09:22] curl gives me the same error as you get fgimenez [09:26] oh [09:26] fgimenez: curl doesn't read post data from stdin [09:26] fgimenez: try --data-binary @filename [09:27] fgimenez: or @- for stdin [09:27] Chipaca, ok let me check [09:28] Chipaca, yes, {"result":{"created_at":"1442482081293797","may_cancel":false,"output":"go-example-webserver","resource":"/1.0/operations/09cf305f-1a0d-19ea-a98d-7e7331e1e46a","status":"succeeded","updated_at":"1442482081729617"},"status":"OK","status_code":200,"type":"sync"} [09:29] silly curl :) [09:29] Chipaca, yep :) [09:48] fgimenez: also this might come in handy for your testing: [09:48] fgimenez: r=$( echo -e '{"action": "enable"}' | PUT http://localhost:8080/1.0/packages/hello-dbus-fwk/services | jq -r .metadata.resource ); while ! (GET http://localhost:8080$r | jq .metadata.status | egrep 'succeeded|failed'); do echo .; sleep 1; done; GET http://localhost:8080$r | jq . [09:49] fgimenez: snippet from my bash_history; adjust as needed :) [09:51] Chipaca, ok thanks a lot. not only for today, it can also be adapted to the automated integration tests as well [09:52] fgimenez: well.. perhaps. When the above fails during interactive testing it's not a big deal :) [09:52] e.g. if the server crashes, it'll get stuck [09:52] if it's for automated testing you need to guard against that [09:56] Chipaca, yes that'll be covered, i mean the way to test a put to the services resource and how to check the status after that [09:56] ah. ok :) === bigcat is now known as kengyu [14:39] fdisk displays 4 partitions on the RPi2 image [14:39] but they don't add up to the total size of my sdcard, 8gb [14:40] how do I expand mmcvlk0p4 ? [14:46] ogra_, did your writable expansion virus land in 15.04? [14:46] sergiusens, yes [14:47] but the RPi image is still one version behind [15:01] is there a listing of snappy packages that can be installed? [15:09] https://code.launchpad.net/~dholbach/snapcraft/1496392/+merge/271319 [15:09] https://code.launchpad.net/~dholbach/snapcraft/1496789/+merge/271444 [15:16] atXyc0_, point your browser to http://webdm.local:4200/ [15:16] the store UI is running on the Pi [15:25] fgimenez: fginther wanted to talk about the jenkins launchpad plugin. [15:25] fgimenez: what you changed on your branch where only some branches that haven't landed yet, right? [15:25] elopio, fgimenez, right. I'd just like to know what problems you ran into while setting it up [15:25] so that we can determine if this can be utilized for future jenkins deployments [15:26] elopio, fginther yes, i needed to make some changes, let me check [15:26] fgimenez, thanks! [15:26] fgimenez, I already found a couple, including the need to add the build token MP [15:27] fgimenez: we have one big problem, but we can easily change it. Currently we have only one job that runs the test, not downstream jobs. So the comment that gets to launchpad is not too acurate. [15:27] elopio, could you point me to an example? [15:28] fginther: umm, maybe? Let me try to find it. [15:28] fginther: and the other big problem I can see in our future is that we will have git branches. [15:30] fginther: https://code.launchpad.net/~chipaca/snappy/schmideload/+merge/268498/comments/675267 [15:31] elopio, yes, but if we are going to split job for parallel execution we are going to have these downstream jobs anyway, what do you think? [15:31] elopio, when it comes to git branches, there is likely some better support via jenkins plugins and jenkins-launchpad-plugin may no longer be necessary (but that's a bit of speculation) [15:31] the $BUILD_URL issue is https://github.com/jenkinsci/docker/issues/138 [15:32] elopio, if we want to have a single complete execution it seems that we need to do some tweaking with the job definition for having that info in the messages with the current code [15:32] elopio, have you tried setting the public url in jenkins? [15:33] (it's on the main configuration page if you don't know where to look) [15:36] fgimenez, were there any other setup and installation issues you encountered? [15:38] fgimenez: yes to all you said. :) [15:39] fginther: and yes, the url issue is that we need to manually save the config from the web. No way to do a clean deploy and get everything ready to go. [15:39] still no big issue. [15:40] fginther: we got it working, which is good. If we are going to stick with it, we can do some tweaks and upstream them. [15:40] but I saw some branches ready to land that were never approved. [15:40] are you maintaining this project? [15:40] fginther, take a look at the branch history http://bazaar.launchpad.net/~fgimenez/+junk/jenkins-launchpad-plugin/changes/130?start_revid=130 [15:40] fgimenez: something left to test on snappy, or should I switch asap to snapcraft? [15:40] https://lists.ubuntu.com/archives/snappy-devel/2015-September/001041.html [15:41] elopio, we're maintaining it for now, trying to get it in better shape to use it with the jenkins-as-a-service deployments [15:41] elopio, i think we are done, didn't have time to reproduce tutorials/guides, but all seems to be fine [15:41] elopio, the current effort is to fix any install and compatibility issues encountered [15:42] fginther, thanks for that branch [15:42] errrr! [15:42] fgimenez, thanks for that branch [15:45] \o/ [15:56] fginther, yw :) let me know if there's anything unclear, that's all the changes i needed to make it work in our deployment (which includes a separate tarmac) [15:59] fgimenez, thanks again. I'll look through those changes in detail and let you know if I have any questions === JamesTai1 is now known as JamesTait [17:03] elopio, mind looking at https://code.launchpad.net/~sergiusens/snapcraft/clean/+merge/271487 [17:03] sergiusens: ok. [17:03] John reviewed on irc with me, but I guess he left already :-) [17:04] sergiusens: what about adding it to the integration tests? [17:07] elopio, like for every test possible? [17:07] sergiusens: no, just one execution of clean. [17:07] elopio, as in, for every test run in there? [17:07] elopio, ok, I don't mind that :-) [17:14] sergiusens: the parts directory is not removed. [17:14] is that on purpose? [17:15] elopio, yes; inside parts you can also have local plugins [17:15] elopio, it's sort of a safe haven for things [17:15] sergiusens: what about removing it if it's empty? [17:15] elopio, so instead of removing everything I thought it be safer this way [17:15] elopio, yeah, that can be arranged [17:16] just feels weird to leave behind only one empty dir. [17:23] tedg, found a small issue, which brings me to think an integration test for setup.py pypy would be good :-) [18:08] elopio, clean branch up to date [18:52] I fixed my issue [18:52] it was OSX [18:52] either dd, the userland filesystem driver, or the sdcard port on my MBP [18:52] either way, ubuntu saved ubuntu [18:54] hah, nice [19:10] elopio, https://code.launchpad.net/~sergiusens/snapcraft/pypy-config-test/+merge/271528 [19:10] elopio, btw, do we have adt enabled for our snappy-dev ppas? [19:11] elopio, scratch that, here you go https://code.launchpad.net/~sergiusens/snapcraft/pypy-config-test/+merge/271529 [19:17] sergiusens: I'm not sure how it works for ppas, don't they just run the test if they are present? [19:17] or mabye we need something on the recipe. [19:18] elopio, the feature is there, we just need to activate it [19:18] for the PPA [19:18] elopio, I'll be back in a bit; I hope you get to review by then ;-) [19:18] :-) === ahayzen_ is now known as ahayzen [20:11] sergiusens: I don't get where are you using pypy in this branch. [20:14] elopio, the pyyaml dep goes t pypy to fetch it [20:14] oh [20:14] sergiusens: to pypi? [20:14] elopio, yeah :-P [20:14] * sergiusens renames [20:15] I felt stupid looking at the config code trying to understand that part :D [20:16] sergiusens: maybe you can use the yaml summary and description to explain what's the purpose of the package. [20:17] or a comment in jobs.pxu. Just pypi-config might not be clear enough to understand the intention. [20:18] elopio, ok done, running tests and uploading [20:19] Chipaca, btw ^ [20:20] elopio, btw, we won't find the same line as in the yaml; it is wrapped :-) [20:24] sergiusens: yeah, just check for what you expect. [20:24] that's good because we check also the wrapping. [20:24] maybe all the env vars are not necessary to check. [20:27] sergiusens: I think we might need to add something like config helpers. Map the snappy yaml into native formats, ini, xml, gsettings, etc. [20:28] sergiusens: + description='config for webcam-webui', [20:28] that's from the other example. [20:31] elopio, yeha, but it is indeed a config for that and only that used here [20:31] tedg, I agree and I want to be smart about it and do it in wiki parts [20:34] elopio, branch pushed [20:34] sergiusens: I think they're gonna have to be plugins, eh? [20:35] sergiusens: Have to be able to override the config key in the meta directory [20:35] tedg, I think they will, yeah; which is what demotivates me [20:36] tedg, I don't want a swarm of plugins ;-) [20:36] sergiusens: Good luck, the goals are only to snap all the software in the world. I'm sure it's fairly consistent :-) [20:36] tedg, I can dream, can't I [20:37] tedg, the world will be written in go soon and this will be a non issue ;-) [20:37] * tedg hands sergiusens a newborn so he can't sleep and therefore dream ;-) [20:37] tedg, too late, already happened ;-) [20:38] tedg, btw, did you see my pip branch comment? [20:39] sergiusens: Yeah, not sure what is going on, it is odd to me that is complaining that the directory isn't in the path, and it is. [20:42] tedg, this is the suspect, right? http://bazaar.launchpad.net/~ted/snapcraft/python-pip/revision/151 ? [20:43] tedg, I'll hack in an 'env' inside that self.run to see how the env looks like [20:43] sergiusens: I don't think so, since it would have gotten the variable anyway through the env=env part. [20:43] sergiusens: I think it probably broke with the setuptools branch. [20:44] tedg, hmm, but that webcam-webui is based on the setuptools branch [20:45] Let me check trunk. [20:52] It runs in trunk but seems to have the same environment. [20:53] That was just an attempt to clean things up a bit. Nothing critical. [20:53] I'm surprised it would have any effect. [20:54] tedg, well now there is a simpler tests run during the integration tests [20:55] K, werid though. [21:17] tedg, PYTHONPATH is setup correctly [21:27] sergiusens: Your test fails in the same way. [21:27] Good I guess. [21:29] elopio: Is there a way I can run one of the integration tests without running all of them? [21:31] tedg, it is rather complicated with plainbox; you need to create a test suite :-/ [21:34] * tedg deletes all the other tests [21:47] tedg, removing def env from python3.py makes it all work [21:47] sergiusens: Adding it to the self.run() command? [21:48] tedg, yeah, like it was before [21:48] tedg, I guess it means we have a broken env story [21:49] Err, that's fine to get things working. But it doesn't quite explain it. They should be identical. [21:49] tedg, well none of your changes made sense of breaking per se ;-) [21:49] Wonder if there's some other env var that gets lots. [21:49] lst [21:49] lost [21:53] sergiusens: Are we doing the os.envir.copy() to start out? [21:54] yeah [21:54] tedg, but that alone isn't enough [22:10] tedg, the env gets setup twice and is overriden [22:11] tedg, export PYTHONPATH=/home/sergiusens/source/launchpad.net/snapcraft/integration-tests/data/pypi-config/parts/python3/install/usr/lib/python3/dist-packages [22:11] export PYTHONPATH=/home/sergiusens/source/launchpad.net/snapcraft/integration-tests/data/pypi-config/parts/python3/install/usr/lib/python3/dist-packages [22:13] tedg, http://paste.ubuntu.com/12441978/ (simple hack) [22:14] sergiusens: But why would assemble_env() return different things each time? [22:15] tedg, dunno, it's a global var [22:15] That's not good... [22:15] Both that it's a global and returning different things. [22:15] tedg, in snapcraft.common [22:16] sergiusens: Okay, I'm gonna have to come back to this, need to go start dinner. [22:16] sergiusens: I'll get a revision up later. [22:16] tedg, sounds good [22:22] tedg, in any case, this resolves it http://paste.ubuntu.com/12442059/ [22:23] tedg, ensuring the local part can override any setting done by its dependencies [22:23] tedg, but we need a better architecture for setting up the environment [23:46] problem trying snappy and docker: The command "docker build -t my-container . " gives me the following error "Error checking context is accessible: 'can't stat '.''. Please check permissions and try again." [23:47] https://bugs.launchpad.net/snappy/+bug/1412343 ... I tried building the docker image in $HOME/apps/docker but it also failed. [23:47] Launchpad bug 1412343 in Snappy "Command `docker build` is broken" [Medium,Fix released] [23:48] So, how do I get the fix? [23:49] I used "sudo snappy install docker" [23:56] Launchpad ticket 1412343 says "status: Confirmed → Fix Released" I assume the fix went into ubuntu-core, since it seems to be appmor-related. Now: [23:56] psanchez_, have you seen the last comment https://bugs.launchpad.net/snappy/+bug/1412343/comments/5 [23:56] Launchpad bug 1412343 in Snappy "Command `docker build` is broken" [Medium,Fix released] [23:56] $ snappy list [23:56] Name Date Version Developer [23:56] ubuntu-core 2015-09-17 5 ubuntu [23:56] docker 2015-09-17 1.6.2.003 canonical [23:56] webdm 2015-09-17 0.9 canonical [23:56] generic-amd64 2015-09-17 1.4 canonical [23:59] Yes, the laucnhpad ticket says "Fix Released," but where is it? I believe I am running the latest ubuntu-core, where I thought the fix is located.