| openstackgerrit | Thomas Goirand proposed openstack/deb-heat: Merge tag '7.0.0_rc1' into debian/newton https://review.openstack.org/374586 | 07:48 |
|---|---|---|
| openstackgerrit | Thomas Goirand proposed openstack/deb-keystone: Merge tag '10.0.0_rc1' into debian/newton https://review.openstack.org/374588 | 07:51 |
| openstackgerrit | Thomas Goirand proposed openstack/deb-manila: Merge tag '3.0.0_rc1' into debian/newton https://review.openstack.org/374592 | 07:54 |
| openstackgerrit | Thomas Goirand proposed openstack/deb-mistral: Merge tag '3.0.0_rc1' into debian/newton https://review.openstack.org/374593 | 07:55 |
| openstackgerrit | Thomas Goirand proposed openstack/deb-neutron: Merge tag '9.0.0_rc1' into debian/newton https://review.openstack.org/374596 | 07:58 |
| tlbr | zigo, could you please push master-dfsg branch here https://anonscm.debian.org/cgit/pkg-javascript/libjs-tv4.git/ ? | 08:02 |
| zigo | tlbr: Hi. | 08:02 |
| tlbr | hi! | 08:02 |
| zigo | tlbr: Can't you push it?!? | 08:02 |
| zigo | I don't get it. | 08:02 |
| tlbr | zigo, it isn't there anymore =\ | 08:02 |
| zigo | tlbr: I don't have it. | 08:02 |
| zigo | tlbr: I only took the debian/unstable branch. | 08:03 |
| zigo | tlbr: Thorsten rejected libjs-tv4. | 08:03 |
| zigo | please mention all licenses/copyright information in your debian/copyright. | 08:03 |
| zigo | For example libjs-tv4-1.2.7+dfsg1/component.json mentions MIT, | 08:03 |
| zigo | libjs-tv4-1.2.7+dfsg1/try/ace/theme-tomorrow.js and others are under BSD3, | 08:03 |
| zigo | libjs-tv4-1.2.7+dfsg1/try/ace/mode-rdoc.js and others are under AGPL-3 | 08:03 |
| tlbr | that is why i'm working on it | 08:03 |
| zigo | Ok. | 08:04 |
| zigo | tlbr: Let me know. | 08:04 |
| zigo | tlbr: FYI, I'm pushing RC1 for all packages. | 08:04 |
| zigo | Probably you saw it already... | 08:05 |
| tlbr | yep | 08:05 |
| zigo | tlbr: We're stuck with Horizon until libjs-tv4 enters unstable, I guess, right? | 08:05 |
| tlbr | not sure | 08:06 |
| zigo | tlbr: BTW, you shouldn't merge 10.0.0.0rc1, you should retag that as 10.0.0_rc1, push that tag to gerrit, and merge that tag. | 08:06 |
| tlbr | you mean remove tag 10.0.0.0rc1 and replace it with 10.0.0_rc1 ? | 08:07 |
| zigo | You can't remove tags, we don't have gerrit rights for that. | 08:07 |
| zigo | But you can do: | 08:07 |
| tlbr | i mean remove it locally | 08:07 |
| zigo | git tag -s 10.0.0_rc1 10.0.0.0rc1 | 08:07 |
| zigo | git push gerrit --tags | 08:07 |
| tlbr | that is actually what i did | 08:08 |
| zigo | That's what I've been doing all the way ... | 08:08 |
| tlbr | without pushing tags.. | 08:08 |
| zigo | Well, here, you've merged the wrong tag, no? | 08:08 |
| zigo | https://review.openstack.org/#/c/374435/ | 08:08 |
| zigo | I don't think it makes a difference though. | 08:08 |
| zigo | Since you're core in the team, you can push (signed only) tags to Gerrit. | 08:08 |
| tlbr | ok | 08:09 |
| openstackgerrit | Thomas Goirand proposed openstack/deb-nova: Merge tag '14.0.0_rc1' into debian/newton https://review.openstack.org/374608 | 08:15 |
| openstackgerrit | Thomas Goirand proposed openstack/deb-openstack-trove: Merge tag '6.0.0_rc1' into debian/newton https://review.openstack.org/374609 | 08:18 |
| openstackgerrit | Thomas Goirand proposed openstack/deb-sahara: Merge tag '5.0.0_rc1' into debian/newton https://review.openstack.org/374610 | 08:22 |
| openstackgerrit | Thomas Goirand proposed openstack/deb-senlin: Merge tag '2.0.0_rc1' into debian/newton https://review.openstack.org/374611 | 08:25 |
| zigo | Ok, done for all packages we could. | 08:29 |
| openstackgerrit | Thomas Goirand proposed openstack/deb-zaqar: Merge tag '3.0.0_rc1' into debian/newton https://review.openstack.org/374615 | 08:29 |
| zigo | Just Horizon still missing... | 08:29 |
| zigo | I'll make sushis waiting for the CI ! :0 | 08:29 |
| zigo | tlbr: I'm still stuck with python-pandas taking more than 1h30 to build, but hopefully the timeout is now 120 (the patch in project-config was merged) and it will go through. | 08:32 |
| zigo | Then we can build gnocchi. | 08:32 |
| zigo | Evertyhing else is kind of under control! :) | 08:32 |
| zigo | tlbr: Are you soon done with libjs-tv4? | 08:55 |
| tlbr | soon soon | 08:55 |
| *** yatinkarel has quit IRC | 08:58 | |
| *** ig0r_ has joined #openstack-pkg | 09:22 | |
| tlbr | zigo, i've pushed the changes | 10:43 |
| *** zigo has quit IRC | 11:54 | |
| *** zigo has joined #openstack-pkg | 11:58 | |
| *** zigo is now known as Guest83601 | 11:59 | |
| *** Guest83601 has quit IRC | 12:03 | |
| *** zigo_ has joined #openstack-pkg | 12:12 | |
| *** zigo_ has quit IRC | 12:17 | |
| *** zigo_ has joined #openstack-pkg | 12:18 | |
| *** zigo_ has quit IRC | 12:48 | |
| *** zigo_ has joined #openstack-pkg | 12:51 | |
| *** zigo_ is now known as zigo | 15:14 | |
| openstackgerrit | Thomas Goirand proposed openstack/deb-sahara: Merge tag '5.0.0_rc1' into debian/newton https://review.openstack.org/374610 | 15:21 |
| openstackgerrit | Thomas Goirand proposed openstack/deb-nova: Merge tag '14.0.0_rc1' into debian/newton https://review.openstack.org/374608 | 15:23 |
| zigo | tlbr: I wonder if we shouldn't reprocude all of the AGPL in tv4's debian/copyright. | 15:30 |
| tlbr | zigo, not sure, i guess what i did should be enough. May be Thorsten just greped by license word ? | 15:33 |
| zigo | tlbr: For sure he has tools to do that, yeah. | 15:33 |
| zigo | tlbr: I'm adding the full of the AGPL license in debian/copyright right now. | 15:33 |
| zigo | it wont take long. | 15:34 |
| tlbr | zigo, we should not do that! | 15:34 |
| tlbr | I've checked how other do that | 15:34 |
| zigo | tlbr: Why not? | 15:34 |
| zigo | tlbr: What others do isn't relevant. | 15:34 |
| tlbr | zigo, please check https://www.gnu.org/licenses/agpl-3.0.en.html | 15:34 |
| tlbr | "How to Apply These Terms to Your New Programs" section | 15:34 |
| zigo | The only case where we shoulnd't put the whole license is when it's available in /usr/share/common-license. | 15:35 |
| zigo | "You should have received a copy of the GNU Affero General Public License" ... | 15:35 |
| zigo | I never saw a package rejected because it contained TOO MUCH license text. | 15:35 |
| zigo | I did see some rejected, because NOT ENOUGH. | 15:36 |
| zigo | So better be safe. | 15:36 |
| zigo | It would have been easier to just remove that folder entirely... | 15:36 |
| tlbr | yes, easier but I'm not sure if it is the right thing to do in this case | 15:37 |
| zigo | I'm convinced it is. | 15:38 |
| tlbr | so we should remove it? | 15:38 |
| zigo | Remove what? The folder? | 15:41 |
| zigo | We that was a possibility yes. | 15:41 |
| tlbr | the folder, yep | 15:41 |
| zigo | Now, we've gone into another direction, so let's finish this way. | 15:41 |
| tlbr | oh, so you agree that we should leave as is | 15:41 |
| tlbr | ok | 15:41 |
| zigo | We have 2 choices: either copy the full of the AGPL in debian/copyright, or get rid of that file ... :P | 15:42 |
| zigo | The less you package, the less troubles ... | 15:43 |
| zigo | Sure, it's nice to have as much as possible. | 15:43 |
| zigo | But I'm not sure if it was really necessary in this case. | 15:43 |
| tlbr | ok :) | 15:46 |
| zigo | tlbr: Uploaded. | 15:47 |
| tlbr | zigo, thanks | 15:47 |
| zigo | tlbr: Can you validat all the RC1 patches? | 16:21 |
| tlbr | zigo, in all projects ? | 16:21 |
| zigo | tlbr: Yup. | 16:22 |
| tlbr | zigo, ok | 16:23 |
| zigo | deb-barbican, deb-ceilometer, deb-designate, deb-glance, deb-heat, deb-keystone, deb-manila, deb-mistral, deb-neutron, deb-nova, deb-openstack-trove, deb-sahara, deb-senlin, deb-zaqar | 16:23 |
| tlbr | ok, i'll check it | 16:24 |
| *** zigo has quit IRC | 16:41 | |
| *** zul has quit IRC | 16:44 | |
| *** zul has joined #openstack-pkg | 16:45 | |
| *** zigo has joined #openstack-pkg | 16:51 | |
| *** zigo is now known as Guest18656 | 16:52 | |
| *** Guest18656 has quit IRC | 16:56 | |
| *** zigo_ has joined #openstack-pkg | 16:59 | |
| *** zigo_ has quit IRC | 17:07 | |
| *** zigo_ has joined #openstack-pkg | 17:11 | |
| zigo_ | tlbr: Thorsten (the FTP master that rejected) confirmed: | 17:30 |
| zigo_ | <ta> zigo: yes, everything not in /usr/share/common-licenses must be in debian/copyright | 17:30 |
| tlbr | zigo_, ok, i agree with that. My only concern was whether we should put all the Licence text | 17:30 |
| zigo_ | That was a reply about that very question. | 17:31 |
| zigo_ | tlbr: Should I self-approve my rc1 merge patches? :) | 17:32 |
| tlbr | zigo_, sec, I'll check it | 17:36 |
| zigo_ | Also, any idea how we can add onovy as core? | 17:37 |
| zigo_ | I'm skeaming through the doc ... | 17:37 |
| onovy | i think i know how | 17:37 |
| onovy | you need to send review to one repo | 17:37 |
| tlbr | zigo_, I've added onovy | 17:38 |
| tlbr | zigo_, https://review.openstack.org/#/admin/groups/1136,members | 17:38 |
| zigo_ | How did you do that? | 17:38 |
| zigo_ | Oh ok. | 17:38 |
| zigo_ | Good ! :) | 17:38 |
| onovy | ah | 17:38 |
| tlbr | :) | 17:38 |
| zigo_ | onovy: Feel free to update Alembic! :) | 17:38 |
| onovy | what about gerrit/acls/openstack/<projectname>.config | 17:39 |
| onovy | https://github.com/openstack-infra/project-config/blob/master/gerrit/acls/openstack/deb-packaging.config | 17:39 |
| onovy | got it. group :)( | 17:39 |
| onovy | tlbr: zigo_: thanks | 17:40 |
| tlbr | np :) | 17:40 |
| zigo_ | onovy: Did you read how merge commiit work? | 17:40 |
| onovy | no | 17:40 |
| onovy | git merge + git review? | 17:40 |
| zigo_ | onovy: https://wiki.openstack.org/wiki/DEB-packaging#Releasing_a_new_upstream_release_package | 17:40 |
| onovy | so merge commit will edit changelog? | 17:41 |
| zigo_ | onovy: The point is, your merge commit must also include the packaging change for the new release. | 17:41 |
| onovy | so only one commit, ok | 17:41 |
| zigo_ | Yup. | 17:41 |
| zigo_ | Which makes a lot of sense, IMO. | 17:41 |
| zigo_ | Since every commit must continue to make the package build. | 17:41 |
| onovy | yep | 17:42 |
| onovy | git tag 0.8.8 rel_0_8_8 | 17:42 |
| onovy | this is same? | 17:42 |
| onovy | and git push gerrit --tags? | 17:43 |
| zigo_ | onovy: Almost! | 17:43 |
| zigo_ | onovy: Gerrit will only accept signed tags, so you got to: git tag -s 0.8.8 rel_0_8_8 | 17:43 |
| zigo_ | git push gerrit --tags | 17:43 |
| zigo_ | Let me add this to the wiki... | 17:43 |
| onovy | cool | 17:44 |
| onovy | dislike: changing b-d/d with same commit as "new upstream release" | 17:46 |
| openstackgerrit | Ondřej Nový proposed openstack/deb-alembic: Merge tag '0.8.8' into debian/newton https://review.openstack.org/374997 | 17:47 |
| onovy | zigo_: anyone review? | 17:48 |
| onovy | :) | 17:48 |
| zigo_ | onovy: Not until the package is built successfully ! :) | 17:48 |
| onovy | you can do +2+A | 17:48 |
| onovy | it will not merge if not built | 17:48 |
| onovy | if you do +2+A now, jenkins will finish first round, will add +1, than +0, final round, +2, merge | 17:49 |
| onovy | if first or second round fails, it will not merge at all | 17:49 |
| onovy | so it's safe to add +2+A for 'looks good changes' and forgot | 17:49 |
| openstackgerrit | Merged openstack/deb-manila: Merge tag '3.0.0_rc1' into debian/newton https://review.openstack.org/374592 | 17:51 |
| zigo_ | tlbr: Missing +2s: https://review.openstack.org/374610 https://review.openstack.org/374593 | 17:59 |
| onovy | zigo_: we are creating new jessie chroot for every jenkins jobs? | 18:02 |
| zigo_ | onovy: For the moment, yes. | 18:02 |
| onovy | huh, ok | 18:03 |
| zigo_ | onovy: We need to add an "sbuild" element to dib to fix that. | 18:03 |
| openstackgerrit | Merged openstack/deb-zaqar: Merge tag '3.0.0_rc1' into debian/newton https://review.openstack.org/374615 | 18:03 |
| zigo_ | onovy: I tried, and failed so far. | 18:03 |
| onovy | what is "dib"? | 18:03 |
| openstackgerrit | Merged openstack/deb-barbican: Merge tag '3.0.0_rc1' into debian/newton https://review.openstack.org/374398 | 18:03 |
| zigo_ | onovy: diskimage-builder. | 18:04 |
| zigo_ | onovy: https://github.com/openstack-infra/project-config/blob/master/tools/build-image.sh <--- The script from infra to build all of the images daily. | 18:04 |
| onovy | btw: i think we can do merge commit with dch and second commit with b-d/d changes. first commit will not build, but follow up will. so you will send two reviews, both with +2+A | 18:04 |
| zigo_ | onovy: The other issue is that infra is maintaining a useless Debian mirror using reprepro. It's useless because not signed ... | 18:04 |
| zigo_ | Then sbuild-createchroot cannot use it. | 18:05 |
| zigo_ | They need to switch to a real ftp-sync script. | 18:05 |
| openstackgerrit | Merged openstack/deb-senlin: Merge tag '2.0.0_rc1' into debian/newton https://review.openstack.org/374611 | 18:05 |
| zigo_ | onovy: I think it's best to just keep stuff in a single commit. | 18:05 |
| onovy | or modern varnish | 18:05 |
| onovy | zigo_: i make git commit history a bit noisy | 18:05 |
| onovy | it | 18:06 |
| onovy | i like 'as small commit as possible' | 18:06 |
| openstackgerrit | Merged openstack/deb-glance: Merge tag '13.0.0_rc1' into debian/newton https://review.openstack.org/374403 | 18:06 |
| openstackgerrit | Merged openstack/deb-ceilometer: Merge tag '7.0.0_rc1' into debian/newton https://review.openstack.org/374400 | 18:06 |
| openstackgerrit | Merged openstack/deb-designate: Merge tag '3.0.0_rc1' into debian/newton https://review.openstack.org/374401 | 18:06 |
| onovy | zigo_: alembic_0.8.8-1+66~bpo8+1_all.deb // that name is correct? | 18:08 |
| onovy | +66 commits? | 18:08 |
| zigo_ | onovy: Yes. | 18:08 |
| zigo_ | onovy: This includes commits from upstream. | 18:08 |
| zigo_ | It will be reseted once you tag the debian release. | 18:08 |
| onovy | ah, strange | 18:08 |
| onovy | ah, ok | 18:08 |
| zigo_ | Not an issue, really. | 18:08 |
| onovy | undo: strange ; got it now | 18:08 |
| zigo_ | That's just a counter ... | 18:08 |
| zigo_ | Please make sure you only tag debian/1.2.3 when you upload to Debian. | 18:08 |
| onovy | maybe it should be ~66 | 18:09 |
| zigo_ | Both would work ... | 18:09 |
| * zigo_ has a dinner break | 18:09 | |
| zigo_ | ttyl | 18:09 |
| onovy | because after i tag, i will became 0.8.8-1 which is << 0.8.8-1+66 | 18:09 |
| *** ig0r_ has quit IRC | 18:09 | |
| zigo_ | onovy: After you tag, you've just uploaded, so the next commit *MUST* include a dch -i | 18:09 |
| onovy | yep | 18:10 |
| zigo_ | So that's fine, next version will be higher ... :P | 18:10 |
| zigo_ | read the wiki | 18:10 |
| onovy | but package from infra before tag will be higher | 18:10 |
| onovy | hey i understand it :) | 18:10 |
| onovy | just think about it | 18:10 |
| zigo_ | I need to help for the dinner, or my wife will kill me | 18:10 |
| zigo_ | ttyl | 18:10 |
| onovy | ok :) i will wrote it here, reply later | 18:10 |
| onovy | infra built 0.8.8-1+66~bpo... package now | 18:11 |
| onovy | after i tag, it will built 0.8.8-1 | 18:11 |
| onovy | 0.8.8-1 << 0.8.8-1+66~bpo... which is WRONG | 18:11 |
| onovy | it's wrong because 0.8.8-1 is newer package than 0.8.8-1+66~bpo... | 18:11 |
| openstackgerrit | Merged openstack/deb-openstack-trove: Merge tag '6.0.0_rc1' into debian/newton https://review.openstack.org/374609 | 18:12 |
| onovy | so if we publish not tagged packages from infra... | 18:12 |
| onovy | so if you are saying. first commit after upload will bump version, it's logical to use ~ not + | 18:13 |
| onovy | because first commit after upload is saying target version, and every commit after is before target version | 18:13 |
| openstackgerrit | Merged openstack/deb-alembic: Merge tag '0.8.8' into debian/newton https://review.openstack.org/374997 | 18:14 |
| openstackgerrit | Merged openstack/deb-heat: Merge tag '7.0.0_rc1' into debian/newton https://review.openstack.org/374586 | 18:18 |
| openstackgerrit | Merged openstack/deb-nova: Merge tag '14.0.0_rc1' into debian/newton https://review.openstack.org/374608 | 18:18 |
| openstackgerrit | Merged openstack/deb-neutron: Merge tag '9.0.0_rc1' into debian/newton https://review.openstack.org/374596 | 18:22 |
| openstackgerrit | Merged openstack/deb-mistral: Merge tag '3.0.0_rc1' into debian/newton https://review.openstack.org/374593 | 18:29 |
| openstackgerrit | Ondřej Nový proposed openstack/deb-alembic: releasing package alembic version 0.8.8-1 https://review.openstack.org/375058 | 18:50 |
| openstackgerrit | Merged openstack/deb-alembic: releasing package alembic version 0.8.8-1 https://review.openstack.org/375058 | 19:09 |
| zigo_ | onovy: The thing that you don't get is that, once you've built 0.8.8-1+66~bpo... then you tag it, there can *NEVER* be any 0.8,8-1 build, the next commit should fix the changelog and do a 0.8.8-2. | 19:28 |
| zigo_ | So it's in fact a good thing that it's not publishing something we don't want. | 19:28 |
| onovy | hey i really get it | 19:28 |
| onovy | but still thing we should use ~ | 19:29 |
| zigo_ | It's unfortunately too late to do so. | 19:29 |
| zigo_ | If we do, then any build will be broken. | 19:29 |
| onovy | everything can be changes :] | 19:29 |
| zigo_ | Maybe we can do that for Ocata ... | 19:29 |
| onovy | *changed | 19:29 |
| zigo_ | Yeah, but not for Newton. | 19:29 |
| onovy | so lets plan it for ocata | 19:29 |
| zigo_ | This can be the first thing we do in Ocata. | 19:29 |
| onovy | newton will be released soon | 19:29 |
| zigo_ | Yup. | 19:29 |
| onovy | but i have different problem now :) | 19:30 |
| zigo_ | And it can live within deb-openstack-pkg-tools in the debian/ocata branch. | 19:30 |
| onovy | deb-swift repo already exists | 19:30 |
| *** zigo_ is now known as zigo | 19:30 | |
| onovy | and i want to sync it with swift from alioth | 19:30 |
| zigo | How is this a problem? | 19:30 |
| onovy | i need to add debian/mitaka branch + sync debian/newton branch | 19:30 |
| onovy | i can't push | 19:31 |
| onovy | and i can't edit last commit (because it's signed tag) | 19:31 |
| onovy | exactly it's not problem, i only don't know how | 19:32 |
| zigo | You can't just push. | 19:32 |
| zigo | You have to do it through a commit. | 19:32 |
| onovy | so create dummy commit with change-id + git review? | 19:32 |
| zigo | Just rm the debian folder, copy what you have in Alioth, and commit that. | 19:32 |
| onovy | i don't want to lose history | 19:33 |
| zigo | No choice ... | 19:33 |
| onovy | hey | 19:33 |
| zigo | You'll only loose a little bit of history (ie: the last commits). | 19:33 |
| zigo | Otherwise, you can just commit every patch one by one. | 19:33 |
| onovy | really don't want this | 19:33 |
| zigo | But that's very annoying to do ... | 19:33 |
| onovy | hey, that's broke | 19:33 |
| onovy | n | 19:33 |
| onovy | let's found better solution | 19:34 |
| zigo | :P | 19:34 |
| onovy | for example remove that repo | 19:34 |
| onovy | and import it again | 19:34 |
| zigo | The other solution would be to ask a gerrit admin to do the push for you. | 19:34 |
| zigo | You can ask in #openstack-infra. | 19:34 |
| onovy | you told me: remeber what changes you did and we will sync it later | 19:34 |
| onovy | you didn't say i will lose commit history for this changes | 19:34 |
| zigo | Oh, I understand... you don't have a debian/newton branch on Alioth, do you? | 19:34 |
| onovy | i have | 19:34 |
| onovy | i need to add debian/mitaka (it's not in gerrit now) and sync debian/newton from alioth | 19:35 |
| zigo | Hum... | 19:35 |
| onovy | than i can move my work to gerrit | 19:35 |
| zigo | Why do you need a debian/mitaka branch on Gerrit? | 19:35 |
| onovy | because i'm still developing debian/mitaka | 19:35 |
| zigo | Continue debian/mitaka on Alioth. | 19:36 |
| onovy | and how can i cherry-pick/merge from two differents repos | 19:36 |
| zigo | That's the general plan. | 19:36 |
| onovy | (simply) | 19:36 |
| zigo | git format-patch is your friend? :) | 19:36 |
| onovy | no. | 19:36 |
| onovy | (SIMPLY) | 19:36 |
| onovy | if we can't add debian/mitaka branch to gerrit, let's postpone swift migration after newton release | 19:36 |
| zigo | I really think your best bet here is to discuss this in #openstack-infra. | 19:37 |
| zigo | You can, I don't think that's a problem. | 19:37 |
| onovy | if plan is to not add mitaka branches, let's postpone it. it's simplest solution | 19:37 |
| zigo | Though it's going to be difficult for us to do it. | 19:37 |
| onovy | and than ask os infras to ff it | 19:37 |
| zigo | That was *MY* plan. | 19:37 |
| onovy | i'm fine with it | 19:37 |
| zigo | You're free to do differently if you like. | 19:38 |
| zigo | We can ask guys from #openstack-infra what's best way to re-import the deb-swift repo. | 19:38 |
| zigo | Maybe that's best way indeed, considering what you just wrote. | 19:38 |
| zigo | Should I ask for you? | 19:38 |
| onovy | jobs will not work for debian/mitaka correctly i guess | 19:38 |
| onovy | mmnt :) let's talk about it here first | 19:38 |
| zigo | Hum... | 19:38 |
| zigo | I didn't think about that. | 19:38 |
| onovy | we have "per branch" repos in infra? | 19:39 |
| zigo | I know it's possible to do no job at all for a specific branch if we need to. | 19:39 |
| zigo | No. | 19:39 |
| zigo | Single repo, multiple branch. | 19:39 |
| onovy | hmm | 19:39 |
| zigo | I was planning to ask to open debian/mitaka branches for everything after the release. | 19:40 |
| onovy | ? | 19:40 |
| zigo | But we can have different setup for different branches. | 19:40 |
| onovy | open mitaka branches after newton release? | 19:40 |
| zigo | In theory, we can add branch by ourselves. | 19:40 |
| zigo | Just using the gerrit web interface. | 19:40 |
| zigo | Sorry... | 19:40 |
| zigo | Ocata after newton release. | 19:40 |
| onovy | ah | 19:40 |
| onovy | question: | 19:40 |
| onovy | newton swift needs newton python-eventlet | 19:41 |
| onovy | we have single repo | 19:41 |
| zigo | Right. | 19:41 |
| onovy | after we open ocata branch and i upload python-event to ocata branch | 19:41 |
| onovy | i can't change newton swift | 19:41 |
| onovy | because it will use ocata eventlet | 19:41 |
| zigo | No. | 19:41 |
| zigo | Repos are per openstack release. | 19:41 |
| onovy | -> per branch :) | 19:41 |
| zigo | Yup. | 19:41 |
| zigo | But we don't have a Mitaka repo in infra. | 19:41 |
| onovy | onovy | we have "per branch" repos in infra? zigo | No. | 19:42 |
| zigo | Ah, sorry ! :) | 19:42 |
| zigo | repo = deb repo or repo = git repo | 19:42 |
| zigo | That's confusing ! :) | 19:42 |
| zigo | We have a single repo for git... | 19:42 |
| onovy | my fault | 19:42 |
| zigo | Multiple for openstack releases. | 19:42 |
| onovy | single git repo, git branches, multiple debian repo, one for every git branch | 19:42 |
| onovy | right? | 19:43 |
| zigo | After we're done with Newton, there's 2 things that will need to happen: 1/ create 2 a new deb repo for Ocata, importing all Newton packages 2/ make all new debian/ocata branches in Git. | 19:43 |
| zigo | Right. | 19:43 |
| onovy | cool | 19:43 |
| onovy | so back to my problem | 19:43 |
| onovy | let's wait for newton release | 19:43 |
| onovy | i will upload newton swift to unstable than | 19:44 |
| onovy | and move to gerrit | 19:44 |
| onovy | and forgot about mitaka branch | 19:44 |
| zigo | Well, considering swift, the release will be in a few days... | 19:44 |
| onovy | yep. that's reason why i will wait | 19:44 |
| zigo | John Dickinson wrote that on his channel. | 19:44 |
| zigo | Ok, got it. | 19:44 |
| onovy | it's simpler for me | 19:44 |
| zigo | Then you'll stop maintaining Swift for mitaka, right? | 19:44 |
| onovy | yep | 19:45 |
| zigo | ok | 19:45 |
| onovy | no reason to maintain it | 19:45 |
| zigo | Agree. | 19:45 |
| onovy | what about jessie-bpo branches? | 19:45 |
| zigo | I was also thinking: what will we do with jessie-backports ... | 19:45 |
| onovy | hey? :) | 19:45 |
| zigo | :) | 19:45 |
| zigo | Newton will be in stretch. | 19:45 |
| zigo | I don't see why we should do any effort maintaining Mitaka. | 19:45 |
| zigo | Or rather: it's too much a burden for me. | 19:45 |
| onovy | yep | 19:46 |
| zigo | BPO FTP masters will hate us though ... | 19:46 |
| onovy | they hate us every 6 mons | 19:46 |
| onovy | so who cares? | 19:46 |
| zigo | Right! :) | 19:46 |
| onovy | jessie-backports should be last stable | 19:46 |
| onovy | ->newton after newton release | 19:46 |
| zigo | What we could do is continue to maintain for *very serious* security issue until Stretch is out. | 19:46 |
| onovy | stable=openstack stable | 19:46 |
| zigo | Do you think we should upload Newton to jessie-bpo ? | 19:47 |
| onovy | yep | 19:47 |
| zigo | Then it's going to be in both Stretch, and jessie-bpo ... | 19:47 |
| onovy | as last openstack version for jessie i think | 19:47 |
| onovy | yep | 19:47 |
| zigo | How long are we suppose to maintain Newton in jessie-bpo then? | 19:48 |
| zigo | One year after Stretch is released? | 19:48 |
| onovy | it's simple | 19:48 |
| onovy | we will maintain it for strech | 19:48 |
| zigo | I really don't plan on doing LTS stuff with OpenStack. | 19:48 |
| onovy | so why not upload SAME package to j-bpo | 19:48 |
| zigo | Right. | 19:48 |
| zigo | That's not much of a work. | 19:48 |
| onovy | so as long as jessie exists | 19:48 |
| onovy | which is < strech exists | 19:48 |
| zigo | Also, people upgrading from jessie-bpo to stretch, it's nice, because they only upgrade the underlying OS, and OpenStack stays on the same version. | 19:48 |
| zigo | So that's a very nice upgrade path. | 19:49 |
| onovy | "same" version | 19:49 |
| onovy | 1-1~bpo.. -> 1-1 | 19:49 |
| zigo | Well, same without ~bpo ... :) | 19:49 |
| onovy | but it will be almost same package | 19:49 |
| onovy | *package content | 19:49 |
| zigo | Right. | 19:49 |
| zigo | So, it will upgrade, but config for openstack for example wont need to be touched. | 19:49 |
| onovy | yep | 19:50 |
| zigo | Ok, good we have a plan, I'm very happy we discussed that ! :) | 19:50 |
| onovy | which is cool | 19:50 |
| onovy | and we can't upload ocata to j-bpo | 19:50 |
| onovy | because we can't upload it to strech | 19:50 |
| onovy | so fixed :) | 19:50 |
| zigo | Right. | 19:50 |
| zigo | Also, .... BIKESHED ?!? :) | 19:50 |
| zigo | Hopefully, this will happen one day. | 19:50 |
| onovy | :] | 19:51 |
| zigo | onovy: https://review.openstack.org/374610 <--- +2a please? | 19:52 |
| zigo | onovy: Oh, there's one thing that doesn't work on your "maintain swift in Newton only after the release" | 19:53 |
| zigo | onovy: I need swift to be able to do Tempest stuff ... | 19:53 |
| onovy | when? | 19:53 |
| zigo | onovy: You mean, when will I start the tempest validation? | 19:53 |
| onovy | yep | 19:53 |
| zigo | onovy: Next week maybe? | 19:53 |
| onovy | and can you downloiad swift from experimental? | 19:54 |
| zigo | No. | 19:54 |
| zigo | Can't. | 19:54 |
| zigo | It's Jessie + what's in infra. | 19:54 |
| onovy | hmm, what about python-mock package? | 19:54 |
| zigo | I plan to write this tempest validation suite as a gating job too... | 19:54 |
| onovy | 2.0 version is only in experimental and it's maintainted in DPMT | 19:54 |
| zigo | The issue is that if we download it from experimental (and rebuild it), then it's going to be on 2 repository: jessie-newton and jessie-newton-backports... | 19:55 |
| zigo | I don't want this. | 19:55 |
| onovy | that's not the answer :) | 19:55 |
| onovy | there always be packages like python-mock | 19:55 |
| zigo | onovy: https://github.com/openstack/deb-auto-backports/blob/debian/newton/packages-list#L75 | 19:56 |
| zigo | That's where it comes from. | 19:56 |
| zigo | So yes, technically, we can do a backport by just adding "swift" in the package list. | 19:56 |
| onovy | add swift and you are done :) | 19:56 |
| zigo | Though that's really not what we want to do. | 19:56 |
| zigo | Because once it's in the jessie-newton-backports repo, it can't go away. | 19:56 |
| onovy | i'm sure i can migrate swift to gerrit during next week | 19:57 |
| onovy | if we can ff that repo | 19:57 |
| zigo | onovy: Oh, by the way, have you seen that? https://github.com/openstack/deb-auto-backports/blob/debian/newton/packages-list#L122 | 19:57 |
| onovy | only with debian/newton branch | 19:57 |
| onovy | huh, strange | 19:57 |
| zigo | I had to hack A LOT, to just download from official jessie-bpo, because webkit2gtk needs more resources than infra VMs can provide. | 19:57 |
| zigo | It needs more than 8GB of RAM !!! | 19:57 |
| onovy | :) | 19:58 |
| onovy | blame gcc | 19:58 |
| zigo | Here, it crashed when linking ... :P | 20:02 |
| zigo | onovy: Anyway, back to Swift, do you do a lot of cherry-pick from Mitaka ? | 20:03 |
| onovy | yep | 20:03 |
| onovy | my flow is: | 20:03 |
| onovy | 1. fix bug in mitaka | 20:03 |
| onovy | 2. upload to unstable | 20:03 |
| onovy | 3. merge to newton | 20:03 |
| onovy | 4. upload to experimental | 20:03 |
| onovy | so exactly i'm merging not cherry-picking it | 20:03 |
| onovy | http://metadata.ftp-master.debian.org/changelogs//main/s/swift/swift_2.9.0-2_changelog | 20:04 |
| onovy | * Rebased on 2.7.0-{7..10} | 20:04 |
| onovy | it's typical for me. i found something in our dev/test in swift mitaka j-bpo | 20:05 |
| onovy | i fix it in mitaka, because i need it for production | 20:05 |
| onovy | and i can't wait to newton release | 20:05 |
| onovy | for example: swift (2.7.0-10) unstable; urgency=medium | 20:05 |
| onovy | this was bug found just before we wanted to go to production | 20:06 |
| onovy | and it's really strange bug | 20:06 |
| * onovy have an idea | 20:09 | |
| onovy | i can add commits to gerrit repo by merge commit | 20:09 |
| onovy | so if i create merge commit for old+new repo, i can git review it | 20:09 |
| onovy | git merge --no-ff old_repo/HEAD inside new repo | 20:10 |
| onovy | zigo: what do you think? | 20:10 |
| zigo | You can't do that, it wont work. | 20:11 |
| onovy | why? | 20:11 |
| zigo | Gerrit wont accept it. | 20:11 |
| zigo | Gerrit only accept merge commits for branches it already maintains. | 20:12 |
| onovy | gerrit is accepting it now for upstream commits | 20:12 |
| onovy | ah... | 20:12 |
| onovy | damn | 20:12 |
| zigo | Which is why every deb-* repo have a track-upstream option. | 20:12 |
| zigo | What we could do however, would be a "track-upstream" option that would get stuff from Alioth... | 20:12 |
| zigo | But then how would we make it build? | 20:13 |
| onovy | i will ask @ infra | 20:13 |
| zigo | I don't think we have a workflow here ... | 20:13 |
| zigo | onovy: The main issue is having a debian/newton branch of Swift so we can test the rest, right? | 20:14 |
| zigo | onovy: Why don't we just push one, and build swift once, just for our tests? | 20:14 |
| zigo | You can sync from alioth later on ... | 20:14 |
| zigo | Don't you think? | 20:14 |
| zigo | I mean, just something barely useable is enough for us to do tempest things ... | 20:14 |
| zigo | Let's just do that ! | 20:15 |
| onovy | i'm lost | 20:15 |
| onovy | i don't want 'barely useable' i want to fix it :) | 20:15 |
| onovy | i asked in @infra | 20:15 |
| onovy | i need to sync alioth swift -> alioth deb-swift | 20:15 |
| openstackgerrit | Thomas Goirand proposed openstack/deb-swift: Add a .gitreview file https://review.openstack.org/375088 | 20:16 |
| zigo | That's possible, really ! :) | 20:16 |
| zigo | But that's not the issue we're having. | 20:16 |
| zigo | Let's first see if swift builds! :P | 20:16 |
| zigo | Oh, btw, to be able to build for jessie-bpo, we'll have the issue that we need to add extend-diff-ignore = ".gitreview" in debian/source/options of all packages. | 20:17 |
| zigo | Not a very big deal though ... | 20:18 |
| onovy | alembic builds fine without it | 20:18 |
| zigo | That's because I did a hack in deb-openstack-pkg-tools | 20:19 |
| zigo | onovy: https://github.com/openstack/deb-openstack-pkg-tools/blob/debian/newton/build-tools/pkgos-infra-install-sbuild#L246 | 20:20 |
| zigo | :P | 20:20 |
| onovy | so it's fixed | 20:21 |
| onovy | not needed to add it to exetend-diff-ignore, right? | 20:21 |
| zigo | onovy: Official jessie-bpo doesn't have the fix. | 20:21 |
| onovy | i'm lost again | 20:21 |
| zigo | onovy: So, to upload stuff prepared in infra to jessie-bpo, we'll have to add the extended-diff-ignore. | 20:21 |
| onovy | before migration to os infra, this problem doesn't exists | 20:21 |
| zigo | The issue is in libdpkg-perl ... | 20:22 |
| zigo | onovy: Ok, let me explain ... | 20:22 |
| onovy | i can build any package for jessie-bpo locally | 20:22 |
| onovy | yep, thanks :) | 20:22 |
| zigo | onovy: In every deb-* gerrit repo, we need a .gitreview to point at the correct place, you agree with that, right? | 20:22 |
| onovy | yep | 20:22 |
| zigo | onovy: Now, for deb-nova, this clashes with the .gitreview file of upstream nova gerrit repo: we have a diff with the upstream .gitreview file. | 20:23 |
| openstackgerrit | Merged openstack/deb-sahara: Merge tag '5.0.0_rc1' into debian/newton https://review.openstack.org/374610 | 20:23 |
| onovy | ah, got it | 20:23 |
| zigo | onovy: In Sid/Stretch, that's not a problem, because libdpkg-perl automatically ignores it. | 20:23 |
| zigo | In Jessie, libdpkg-perl is older ... | 20:23 |
| zigo | onovy: So, I've patched it in deb-opentsack-pkg-tools so that we don't need to bother. | 20:23 |
| zigo | onovy: Though for an upload in official jessie-bpo, we'll have to manually add the extended-diff-ignore stuff in debian/source/options | 20:24 |
| zigo | That's easy to automate though... | 20:24 |
| zigo | onovy: FYI, I opened a bug against dpkg to fix that, and the dpkg maintainer refused saying that this is a behavior change that he doesn't want to introduce in stable. | 20:25 |
| zigo | I can agree with this view, even if that's very annoying for us. | 20:25 |
| onovy | btw: i didn't change .gitreview in alembic repo | 20:25 |
| onovy | which is wrong | 20:25 |
| zigo | onovy: I did that before, which is why ... | 20:26 |
| onovy | did that for all repos? | 20:26 |
| zigo | onovy: Editing the .gitreview was the commit I used to trigger builds, so yes. | 20:26 |
| zigo | I have hundreds of "Add a .gitreview file" types of commits ! :) | 20:26 |
| onovy | https://github.com/openstack/deb-swift/commits/debian/newton | 20:27 |
| onovy | 404 not found ;) | 20:27 |
| onovy | zigo: from infra: don't merge anything to deb-swift | 20:30 |
| onovy | and they can ff it manually | 20:30 |
| onovy | so don't a | 20:32 |
| onovy | +A | 20:32 |
| zigo | onovy: https://github.com/openstack/deb-swift/commits/debian/newton <--- This isn't leading me to a 404 ... | 20:39 |
| onovy | yep, it's not HTTP 404 | 20:40 |
| onovy | it was joke. i can't found commit for .gitreview | 20:40 |
| zigo | onovy: Well, I haven't done it for swift (yet). | 20:40 |
| onovy | so don't do it now | 20:40 |
| onovy | i will wait for infra | 20:41 |
| zigo | onovy: Why? | 20:41 |
| onovy | because i need ff-only | 20:41 |
| onovy | you can commit it into alioth | 20:41 |
| zigo | onovy: Why is it so important that you have the full commit history? | 20:41 |
| onovy | but not into gerrit | 20:41 |
| onovy | it's important for me | 20:41 |
| onovy | i like to use git blame | 20:41 |
| zigo | onovy: Well, I want to have a build in infra !!! | 20:41 |
| onovy | if you +A any commit, it's not possible to merge it with my changes | 20:42 |
| zigo | There's no other way, unfortunately. | 20:42 |
| onovy | it's | 20:42 |
| onovy | wait. | 20:42 |
| onovy | few hours/days | 20:42 |
| zigo | Right. | 20:42 |
| onovy | infra will ff it | 20:42 |
| zigo | Until Swift for Newton is ready, right? | 20:42 |
| zigo | Oh ok. | 20:42 |
| zigo | I thought they wouldn't ! :) | 20:42 |
| onovy | no, earlier | 20:42 |
| zigo | Got it. | 20:42 |
| onovy | earlier... maybe. swift will be released tomorrow | 20:42 |
| onovy | and i want to release it throu gerrit | 20:43 |
| zigo | :) | 20:44 |
| onovy | i ff-ed deb-swift in alioth | 20:44 |
| onovy | https://anonscm.debian.org/cgit/openstack/deb-swift.git | 20:44 |
| onovy | so i only need to ff gerrit now | 20:44 |
| onovy | than we can continue | 20:44 |
| zigo | onovy: BTW, now that we have all of the RC1 built and available on infra, we need to build them and upload to experimental ... | 20:47 |
| zigo | onovy: Do you have time for it? :) | 20:47 |
| onovy | don't build it | 20:47 |
| onovy | only upload it | 20:47 |
| zigo | onovy: You mean source-only uploads? | 20:47 |
| zigo | :P | 20:48 |
| zigo | I really think that's dangerous. | 20:48 |
| zigo | And that we should build to make sure all unit tests passing in Sid too. | 20:48 |
| onovy | it's not | 20:48 |
| onovy | (dangerous) | 20:48 |
| onovy | if used correctly, it's perfect | 20:48 |
| onovy | i'm using almost only source-only | 20:48 |
| zigo | Well, the source-only upload, right. | 20:48 |
| zigo | But we should still check that unit tests are passing. | 20:48 |
| onovy | yep, there could be inconsitency between infra and deb | 20:49 |
| zigo | Hopefully, we'll have a (non voting?) check job with a Sid sbuild chroot in infra ... | 20:49 |
| onovy | hopefully we will have gate for debian upload? | 20:49 |
| zigo | LOL ! :) | 20:50 |
| onovy | think out of the box | 20:50 |
| zigo | I'm not sure the FTP masters will like the idea ... | 20:50 |
| onovy | of course you can't give anyone PGP keys | 20:50 |
| onovy | and ftp masters will not like this idea | 20:50 |
| zigo | I've been thinking about that, yes... | 20:50 |
| onovy | but we can found solution somewhere in middle | 20:50 |
| zigo | This will only be possible once we have reprocuded OpenStack infra within a DSA machine. | 20:50 |
| onovy | infra will prepare it, you will sign it, and upload it | 20:51 |
| onovy | but with source-only is this all useless | 20:51 |
| onovy | we should gate TAGS for sid schroot | 20:51 |
| onovy | if jenkins ack, i can upload it source-only | 20:51 |
| zigo | != jenkins | 20:51 |
| openstack | zigo: Error: "=" is not a valid command. | 20:51 |
| zigo | Jenkins is *GONE* ! :) | 20:51 |
| zigo | !help | 20:52 |
| openstack | zigo: (help [<plugin>] [<command>]) -- This command gives a useful description of what <command> does. <plugin> is only necessary if the command is in more than one plugin. | 20:52 |
| onovy | jenkins = openstack infra | 20:52 |
| openstackgerrit | Merged openstack/deb-auto-backports: Build pandas needed by gnocchi https://review.openstack.org/373021 | 20:52 |
| onovy | from now :) | 20:52 |
| zigo | YEAH !!! | 20:52 |
| zigo | PANDAS GOT IN !!! :P | 20:52 |
| zigo | Let's wait another 2 hours until the package is built ... | 20:52 |
| zigo | onovy: I had to increase the deb-auto-backports timeout from 1h30 to 2h to make it happen ... :) | 20:52 |
| onovy | :] | 20:53 |
| zigo | So, another 2 hours of POST commit job, and it will be in the jessie-newton-backports repo. | 20:53 |
| zigo | Then, I can build Gnocchi ... | 20:53 |
| onovy | 2 hours for python code... | 20:53 |
| zigo | That's mostly C code. | 20:54 |
| zigo | (in a Python module) | 20:54 |
| zigo | Lot's of unit tests, you know ... | 20:54 |
| zigo | The Sphinx doc takes also forever to build. | 20:54 |
| zigo | Gosh ... :( | 20:56 |
| zigo | http://logs.openstack.org/36/36b8f73bc5d5b4546d45ff971cb9f620ffffbdea/post/pkgdeb-publish-auto-backports/57bd4f4/console.html | 20:56 |
| zigo | :/ | 20:56 |
| zigo | So, from now on: forbidden to attempt any other backport until pandas is in. | 20:56 |
| openstackgerrit | Thomas Goirand proposed openstack/deb-auto-backports: Dumb commit to re-attempt pandas build https://review.openstack.org/375108 | 20:57 |
| onovy | maybe you could be more robust in build script | 20:57 |
| onovy | and try apt-get more than onetime | 20:57 |
| onovy | why you didn't just write "recheck" | 20:58 |
| zigo | onovy: The same issue could happen when doing the sbuild-createchroot | 20:58 |
| zigo | Or any apt-get during the build... | 20:58 |
| onovy | i know, it's not ideal | 20:58 |
| zigo | onovy: The thing that broke is the POST job, not the check or the gate. | 20:58 |
| onovy | ah | 20:58 |
| zigo | onovy: What's broken is that we're still using zuul 2.5, and not 3... :( | 20:58 |
| zigo | With zuul 3, we wouldn't need to build 3 times. | 20:59 |
| zigo | onovy: BTW, you can install python-git-os-job (from sid...), it's quite useful. | 21:00 |
| onovy | hmm | 21:00 |
| zigo | # cat /usr/bin/goj | 21:00 |
| zigo | #!/bin/sh | 21:00 |
| zigo | git-os-job $@ $(git log | head -n 1 | awk '{print $2}') | 21:00 |
| zigo | :P | 21:00 |
| onovy | ehm | 21:01 |
| onovy | git os-job HEAD? | 21:01 |
| onovy | i want to look to last commit, right? | 21:01 |
| zigo | No, because I want it to do it in my current branch. | 21:01 |
| zigo | Oh, that works too ... | 21:02 |
| onovy | git osb-job debian/newton | 21:02 |
| * zigo hides with shame ... | 21:02 | |
| onovy | but i think HEAD works too | 21:02 |
| zigo | It does. | 21:02 |
| onovy | because HEAD is per-branch | 21:02 |
| zigo | onovy: BTW, will you be in Barcelona? | 21:03 |
| onovy | yep, you already asked me | 21:03 |
| onovy | me+3 coworkers | 21:03 |
| zigo | I thought I did, indeed, but forgot. | 21:03 |
| zigo | Nice. | 21:03 |
| onovy | barcelona is cheap for us | 21:04 |
| onovy | btw: do you know mirantis bought tcp cloud from cz? | 21:04 |
| onovy | you are colleage with fpytloun :) | 21:04 |
| onovy | gn | 21:08 |
| zigo | onovy: Of course I know ... | 21:14 |
| zigo | onovy: I knew about it a way before the general announcement ! :) | 21:14 |
| *** zul has quit IRC | 21:36 | |
| *** strigazi has quit IRC | 21:37 | |
| *** onovy has quit IRC | 21:37 | |
| *** IgorYozhikov has quit IRC | 21:37 | |
| *** themill has quit IRC | 21:37 | |
| *** beisner has quit IRC | 21:37 | |
| *** adreznec has quit IRC | 21:37 | |
| *** IgorYozhikov has joined #openstack-pkg | 21:38 | |
| *** themill has joined #openstack-pkg | 21:38 | |
| *** beisner has joined #openstack-pkg | 21:40 | |
| *** adreznec has joined #openstack-pkg | 21:40 | |
| *** zul has joined #openstack-pkg | 21:40 | |
| *** strigazi has joined #openstack-pkg | 21:40 | |
| *** onovy has joined #openstack-pkg | 21:41 | |
| openstackgerrit | Merged openstack/deb-keystone: Merge tag '10.0.0_rc1' into debian/newton https://review.openstack.org/374588 | 22:15 |
| openstackgerrit | Thomas Goirand proposed openstack/deb-python-ironic-lib: Add a .gitreview file https://review.openstack.org/375143 | 22:31 |
| openstackgerrit | Thomas Goirand proposed openstack/deb-heat: Add python-aodhclient (>= 0.5.0) b-d https://review.openstack.org/375149 | 22:39 |
| openstackgerrit | Thomas Goirand proposed openstack/deb-heat: Add python-aodhclient (>= 0.5.0) b-d https://review.openstack.org/375149 | 22:40 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!