=== tds8 is now known as tds [01:59] blackboxsw: Sounds good, I'll +1 the branch [02:01] blackboxsw: it looks like there's a package build failure related to pytest not in the build-reqs ... [02:01] do we need to sort that out ? [02:33] rharper: we do unfortunately [02:33] I'm trying to work that now [02:37] rharper, I just tested on the 3 BSD, if two interfaces use DHCP, the system uses the default route of the first NIC. [02:41] rharper: yeah I think we also need to cherry-pick the tox pytest deps changes [02:42] cherry-pick 986f37b01 [02:42] redoing that branch now to test [02:52] rharper: finally repushed PR 303. and added new manual steps to the PR description. basically I cherry-pick the already released 6600c642, then the tox pytest 986f37b01, then the rest of the picks [03:08] hrm I think the build fails again the issue is that sbuild "Install cloud-init build dependencies (apt-based resolver)" gets dependencies direct from test-requirements.txt before the cpick is applied to drop nose and add pytest. So, it's not present in the sbuild image. [03:09] even though the end result has a proper test-requirements.txt file once patches are applied [03:09] bummer, may have to fix this tomorrow. === tds1 is now known as tds === tds4 is now known as tds === hjensas|afk is now known as hjensas [14:44] blackboxsw: I think the github CI build is just a package/bddeb; so I doesn't have any patches applied, so I think it's a funny state; === hjensas is now known as hjensas|afk === hggdh is now known as hggdh_ === hggdh_ is now known as hggdh__ === hggdh__ is now known as hggdh-msft === eggbean is now known as paradroid === paradroid is now known as delta2 === delta2 is now known as eggbean === eggbean is now known as paradroid_ [16:57] blackboxsw: https://github.com/canonical/cloud-init/pull/306 is up [17:14] ahh that makes sense rharper [17:14] maybe we can change that too in a separate pr against cloud-init [17:14] looking over 306 [17:15] When using cloud-init multi-part MIME with multipass, I receive "error loading cloud-init config: yaml-cpp: error at line 5, column 13: illegal map value". [17:17] 1 Content-Type: multipart/mixed; boundary="===============8832723955843212272=="2 MIME-Version: 1.03 4 --===============8832723955843212272==5 Content-Type: text/cloud-config; charset="utf-8"6 MIME-Version: 1.07 Content-Transfer-Encoding: base64 [17:23] hi PodioSpaz, we only currently support plain user data in the --cloud-init argument [17:25] there's a handful of feature requests for more complex scenarios on our github [17:25] https://github.com/canonical/multipass/issues/557 [17:25] https://github.com/canonical/multipass/issues/1295 [17:26] https://github.com/canonical/multipass/issues/1040 [17:26] https://github.com/canonical/multipass/issues?q=is%3Aissue+is%3Aopen+cloud-init [17:26] please file one for multi-part if none of those fits your use case [17:27] Thanks [17:30] rharper: approved 306, needs a rebase against master [17:30] and can land [17:30] rharper: I'll rebase and push. [17:34] blackboxsw: thanks! [17:35] ci running on it. will merge it when I see that green [17:35] almost have new manual steps for you on release (it'll take into account your PR which we will land momentarily [17:36] OK [17:58] rharper: just landed https://github.com/canonical/cloud-init/pull/306 [17:58] I'll get you the steps for focal release [18:11] okay rharper new manual steps and push for focal release to include your 306 https://github.com/canonical/cloud-init/pull/303 [19:06] blackboxsw: ok, running them now [19:06] rharper: might want to refresh the pull request UI I updated the description [19:06] yes [19:06] thx [19:06] building package now to test [19:06] I had missed a couple things in adding my additional quilt patch [19:10] what's the ./picks.sh ?? is that our wrapper you're working on ? [19:11] rharper: that is just a script to automate the first half of the manual operations to avoid having to cut and paste each step [19:11] so you should be able to run that directly as is.... and manually continue with = cd cloud-init and beyond [19:11] hrm [19:11] cd /tmp; bash ./picks.sh [19:11] why do I need it? I've created all of the cpicks; then I push them on [19:12] fix up the unittest [19:12] etc. [19:12] I don't see why I need it? (or is it automating all of the cpick work I did following your steps?) [19:12] sorry, I see it now [19:12] if you've already created the cpicks. then you only need additionally to: [19:12] cherry-pick 09fea85fd1f6fd944f4cdd8b97e283090178ae88 # your PR 306 [19:13] and the quilt push -a .... .steps below [19:13] gotcha [19:13] the problem you have with your branch is that you've already updated and commited for release in debian/changelog. so you might was to get reset HEAD~1 to back off of that [19:13] You forgot an EOF [19:13] that's why [19:13] ahh bad paste. [19:13] I reset to upstream/ubuntu/devel first [19:14] so I was replaying all of the picks from the beginning but I didn't notice it was a heredoc ... [19:14] no worries. I'm on it now [19:14] did you delete the trailing whitespace after removing the unittest ? [19:14] oh there's the EOF ... hehe [19:14] rharper: my EOF is middle of that description as I kicked it out during the quilt header as I figured there'd be some other wrangling [19:14] picks calls itself =) [19:16] rharper: picks doesn't call itself the EOF was above that just before calling bash ./picks.sh [19:16] I'll update the description and include the entire script within the picks.sh [19:18] your last git commit -am pulls in local cruft, lemme try again [19:29] blackboxsw: I'm pretty close; the only delta I have is in the last quilt patch (how you describe editing the unittest file) [19:30] https://paste.ubuntu.com/p/PDmfQ2HQwv/ [19:30] blackboxsw: ^ [19:30] rharper: I just finally pushed again [19:30] and updated description [19:30] ok, let me replay pick.sh [19:31] it does speed things up =) [19:31] rharper: man , lintian is still comiing up with an error on the branch :/ [19:31] E: cloud-init changes: inconsistent-maintainer Chad Smith (changes vs. source) Ubuntu Developers [19:31] E: cloud-init: mismatch-translated-choices cloud-init/datasources choices-c [19:31] eww [19:31] I'll try applying that Author instead of me [19:32] when editing the header [19:34] oh, I bet you missed a file [19:34] the datasources choices breaks when the distro values aren't matching [19:34] we pulled out netbsd/openbsd from the unittst [19:34] but where else are they ? [19:35] oh, datasource isn't distro, but when we add/remove a datasource; isn't there some manual release branch updates ? [19:36] blackboxsw: ok, diff looks good, once you sort out the lintian failure [19:38] testing out my source package build right now [19:41] great rharper, I'm rebuilding now with Author: Ubuntu Developers in the dep3 header [19:41] to check [19:42] that's what I've got in mine now after your update [19:43] rharper: looks like this https://bugs.launchpad.net/ubuntu/+source/lintian/+bug/1862787 [19:43] Ubuntu bug 1862787 in lintian (Ubuntu) "inconsistent-maintainer error not applicable to Ubuntu" [Undecided,Fix released] [19:44] ah, yes [19:45] hrm, what does this 'mean'. does it mean our focal images in sbuild don't yet have that released version of lintian? [19:47] I'll ask in ubuntu-devel [19:51] not sure [19:51] Lintian: fail [19:51] Status: successful [19:51] Version: 20.1-10-g71af48df-0ubuntu13 [19:51] [19:52] the package builds just fine [19:52] lintial errors are non-fatal to package builds [19:52] I know I recently asked about this when I was testing out probert packaging changes [19:52] blackboxsw: so, I can reproduce your branch, git diff is OK, and sbuild is OK from my perspective; [19:53] is your build-and-push sbuild completing with Success ? [19:53] ahh right. ok cool. I was worried this'd break things [19:53] me too [19:53] rharper: ok. then I think we are good [19:54] I will build-and-push [19:54] one sec [19:54] I have commented those operations out [19:54] are you going to leave the author as you or ubuntu ? [19:54] ok holding. (and reperforming changes to track Author: chad.smith@c.com [19:54] I'd like to get a clean diff against your branch and I'll comment with that [19:54] ok [19:54] I'll replay and diff [19:54] yeah that was unrelated change that didn't fix anything [19:54] same [19:55] the issue was related the the generate *dsc file I think not the dep3 header on the patch [19:55] sooo many cherry-picks [19:59] rharper: force pushed and updated PR description [19:59] building oone more time [20:03] rharper: +1 successful build. with latest changes. [20:03] I won't build-and-push until you compare diffs against what I've pushed [20:04] blackboxsw: same here, I'm approving ; but you push from cli, not from gui. Do we mark approved or something else in the UI ? [20:04] rharper: yep we push releases from cli, not github [20:04] is marking approve OK? it won't autoland [20:04] to perform dput, tag and push to origin/ubuntu/devel [20:05] correct, no auto-land in github [20:05] ... YET. muhahhaa [20:05] heh, also CI will fail due to the no-patch-bddeb [20:05] so that'll block merging via UI [20:05] ok, I've Approved it [20:05] thanks for working on this [20:07] schweet. rharper thanks for the counsel here. also, I'm thinking I'll put up a tiny separate PR against ubuntu/devel now to add the debian/cloud-init-cherry-picks so we can use the uss-tableflip 'cherry-pick --pop-all' to fix daily builds [20:08] nice [20:16] merged thx rharper https://github.com/canonical/cloud-init/pull/303 [20:18] powersj: rharper [ubuntu/focal-proposed] cloud-init 20.1-10-g71af48df-0ubuntu3 (Waiting for approval) [20:23] \o/ [20:23] sweet