| paladox | Fungi I do wish that gerrits ACL supported both “block” and “allow” in the same section. | 00:12 |
|---|---|---|
| *** ianychoi has quit IRC | 00:12 | |
| *** ianychoi has joined #zuul | 00:14 | |
| *** jeliu_ has joined #zuul | 00:14 | |
| *** ianychoi has quit IRC | 00:20 | |
| *** ianychoi has joined #zuul | 00:21 | |
| *** sgw has quit IRC | 00:22 | |
| *** ianychoi has quit IRC | 00:30 | |
| *** ianychoi has joined #zuul | 00:31 | |
| *** jeliu_ has quit IRC | 00:35 | |
| *** ianychoi has quit IRC | 01:03 | |
| *** ianychoi has joined #zuul | 01:04 | |
| *** mgoddard has quit IRC | 01:40 | |
| *** mgoddard has joined #zuul | 01:48 | |
| *** swest has quit IRC | 01:53 | |
| *** igordc has quit IRC | 01:58 | |
| *** swest has joined #zuul | 02:10 | |
| *** ianychoi has quit IRC | 02:14 | |
| *** ianychoi has joined #zuul | 02:16 | |
| *** ianychoi has quit IRC | 02:38 | |
| *** ianychoi has joined #zuul | 02:40 | |
| *** bhavikdbavishi has joined #zuul | 03:34 | |
| *** bhavikdbavishi has quit IRC | 03:38 | |
| *** bhavikdbavishi has joined #zuul | 03:46 | |
| *** raukadah is now known as chandankumar | 03:52 | |
| *** swest has quit IRC | 04:01 | |
| *** jhesketh has quit IRC | 04:19 | |
| *** jhesketh has joined #zuul | 04:19 | |
| *** mgoddard has quit IRC | 05:54 | |
| *** mgoddard has joined #zuul | 06:00 | |
| *** pcaruana has joined #zuul | 06:26 | |
| *** ianychoi has quit IRC | 06:34 | |
| *** swest has joined #zuul | 06:37 | |
| *** ianychoi has joined #zuul | 06:41 | |
| *** themroc has joined #zuul | 07:10 | |
| *** tosky has joined #zuul | 07:16 | |
| openstackgerrit | Dirk Mueller proposed zuul/nodepool master: Switch to openSUSE 15.1 https://review.opendev.org/667532 | 07:33 |
| *** hashar has joined #zuul | 07:45 | |
| openstackgerrit | Mark Meyer proposed zuul/zuul master: Extend event reporting https://review.opendev.org/662134 | 07:50 |
| openstackgerrit | Matthieu Huin proposed zuul/zuul-jobs master: install-nodejs: add support for RPM-based OSes https://review.opendev.org/631049 | 07:57 |
| *** ianychoi has quit IRC | 07:59 | |
| *** ianychoi has joined #zuul | 08:01 | |
| *** hwangbo has quit IRC | 08:48 | |
| *** ianychoi has quit IRC | 09:07 | |
| *** ianychoi has joined #zuul | 09:10 | |
| *** hashar has quit IRC | 09:39 | |
| *** bhavikdbavishi has quit IRC | 09:52 | |
| *** badboy has quit IRC | 10:08 | |
| *** bhavikdbavishi has joined #zuul | 10:59 | |
| *** tosky has quit IRC | 11:15 | |
| *** tosky has joined #zuul | 11:16 | |
| *** shachar has quit IRC | 11:18 | |
| *** saneax has joined #zuul | 11:39 | |
| *** ianychoi has quit IRC | 11:40 | |
| *** ianychoi has joined #zuul | 11:45 | |
| openstackgerrit | Andreas Jaeger proposed zuul/zuul-jobs master: Use PDK to build puppet module https://review.opendev.org/627534 | 11:55 |
| openstackgerrit | Andreas Jaeger proposed zuul/zuul-jobs master: Prepare: Use PDK to build puppet module https://review.opendev.org/667590 | 12:07 |
| openstackgerrit | Andreas Jaeger proposed zuul/zuul-jobs master: Use PDK to build puppet module https://review.opendev.org/667591 | 12:07 |
| openstackgerrit | Andreas Jaeger proposed zuul/zuul-jobs master: Prepare: Use PDK to build puppet module https://review.opendev.org/667590 | 12:09 |
| openstackgerrit | Andreas Jaeger proposed zuul/zuul-jobs master: Use PDK to build puppet module https://review.opendev.org/627534 | 12:09 |
| *** ianychoi has quit IRC | 12:14 | |
| *** swest has quit IRC | 12:15 | |
| *** ianychoi has joined #zuul | 12:22 | |
| openstackgerrit | Andreas Jaeger proposed zuul/zuul-jobs master: Prepare: Use PDK to build puppet module https://review.opendev.org/667590 | 12:22 |
| openstackgerrit | Andreas Jaeger proposed zuul/zuul-jobs master: Use PDK to build puppet module https://review.opendev.org/627534 | 12:22 |
| *** rlandy has joined #zuul | 12:25 | |
| openstackgerrit | Andreas Jaeger proposed zuul/zuul-jobs master: Prepare: Use PDK to build puppet module https://review.opendev.org/667590 | 12:40 |
| openstackgerrit | Andreas Jaeger proposed zuul/zuul-jobs master: Use PDK to build puppet module https://review.opendev.org/627534 | 12:40 |
| pabelanger | this is new | 12:49 |
| pabelanger | 2019-06-26 12:28:46,778 DEBUG zuul.AnsibleJob: [e: 9c4bc610-9809-11e9-8257-656dd3ede3b9] [build: db73073592cb49549a8f9df01b9497ac] Updating playbook or role ansible/ansible-zuul-jobs | 12:50 |
| pabelanger | we seem to now log 2 IDs in log files | 12:50 |
| pabelanger | tobiash: I noticed https://review.opendev.org/666972/ over the weekend about left over bwrap bindmounts when we migrated to multi-ansible support | 12:54 |
| *** sgw has joined #zuul | 13:01 | |
| *** rf0lc0 has joined #zuul | 13:18 | |
| *** rfolco has quit IRC | 13:21 | |
| pabelanger | tobiash: also, today seems to be the day I am migrating to prepare-workspace-git :) | 13:22 |
| pabelanger | trying to debug failure with prepare-workspace from a specific executor pushing repos | 13:22 |
| tobiash | pabelanger: the e id is the event id that triggered the action | 13:23 |
| flaper87 | Is there a way to set the `ansible_shell_type` for a specific image? I would like to connect to my windows nodes using ssh: https://docs.ansible.com/ansible/latest/user_guide/windows_faq.html#can-i-connect-to-windows-hosts-over-ssh | 13:23 |
| flaper87 | but I'd need to set the shell to `powershell` | 13:23 |
| pabelanger | flaper87: you can set the connection-type in nodepool | 13:26 |
| pabelanger | https://zuul-ci.org/docs/nodepool/configuration.html#attr-providers.[openstack].diskimages.connection-type | 13:26 |
| pabelanger | but, ansible_shell_type should just be a job variable | 13:27 |
| flaper87 | pabelanger: I don't want to set the connection-type, I want it to be ssh | 13:27 |
| flaper87 | oh, didn't think about setting ansible_shell_type as a job variable | 13:28 |
| flaper87 | will do | 13:28 |
| pabelanger | ah, yah, that is the default | 13:28 |
| pabelanger | so, job vars should be all that is needed | 13:28 |
| * flaper87 tries | 13:28 | |
| flaper87 | thanks | 13:28 |
| *** bhavikdbavishi has quit IRC | 13:28 | |
| *** swest has joined #zuul | 13:32 | |
| flaper87 | pabelanger: mmh, doesn't seem to be working. In ansible, you'd set this var in the inventory, I wonder if it's being picked at all | 13:33 |
| flaper87 | oh, hold on | 13:33 |
| *** jeliu_ has joined #zuul | 13:33 | |
| flaper87 | this is only available on ansible 2.8. How can I make sure that ansible 2.8 is used for this job? | 13:34 |
| flaper87 | assuming there's a way to do that | 13:34 |
| pabelanger | job.ansible_version = 2.8 | 13:35 |
| *** jeliu_ has quit IRC | 13:38 | |
| *** rf0lc0 has quit IRC | 13:42 | |
| *** rf0lc0 has joined #zuul | 13:42 | |
| pabelanger | so, this is odd, I'm seeing 'file has vanished' from rync in prepare-workspace | 14:05 |
| pabelanger | https://logs.zuul.ansible.com/90/58390/cb7832ea4c51e7894e3de1c95d7e6437dcb40c27/third-party-check/ansible-test-network-integration-iosxr-python36/6b3baa0/job-output.html#l122 | 14:05 |
| pabelanger | that just started happening 2 hours ago | 14:06 |
| openstackgerrit | James E. Blair proposed zuul/nodepool master: WIP test ubuntu builds https://review.opendev.org/667624 | 14:06 |
| *** bhavikdbavishi has joined #zuul | 14:07 | |
| *** jpena|off is now known as jpena | 14:10 | |
| *** jpena is now known as jpena|mtg | 14:10 | |
| *** saneax has quit IRC | 14:13 | |
| *** saneax has joined #zuul | 14:14 | |
| *** bhavikdbavishi has quit IRC | 14:18 | |
| pabelanger | does anybody know why files would be missing (vanished) on a zuul-executor work_root directory? | 14:25 |
| fungi | your work_root is in /tmp and you have tempreaper installed? two builds ended up colliding over the same work_root? | 14:28 |
| * fungi can think of some even more outlandish causes probably | 14:28 | |
| pabelanger | yah, /tmp | 14:32 |
| pabelanger | but, don't think I have tempreper installed | 14:33 |
| *** themroc has quit IRC | 14:33 | |
| *** mattw4 has joined #zuul | 14:35 | |
| pabelanger | https://review.opendev.org/665186/ landed recently, which changes it from tmp | 14:39 |
| pabelanger | let me do that too | 14:39 |
| *** bhavikdbavishi has joined #zuul | 14:40 | |
| fungi | zbr|ruck: in 652708 you suggested that ansible-lint has softened its positions on things since becoming an official ansible project... in your opinion is it going to be easier now to uncap it again? or are the new opinions which caused us to add the <4 cap in test-requirements.txt still in effect? | 14:44 |
| zbr|ruck | fungi: I would not go so far. We use it as pre-commit hook and is pinned, we update it manually each time. this protects us from surprises on new versions. | 14:44 |
| zbr|ruck | usually is better to be 1-2 versions behind on it, is very similar experience with flake8 | 14:45 |
| openstackgerrit | James E. Blair proposed zuul/nodepool master: WIP test ubuntu builds https://review.opendev.org/667624 | 14:45 |
| fungi | zbr|ruck: do you think they're likely to reassess and remove or default disable some of the new rules they enabled in v4? | 14:45 |
| zbr|ruck | fungi: yes, i know for sure that one or two rules were disabled by default after proved to cause too many false positives. | 14:46 |
| fungi | good to know | 14:47 |
| fungi | i'm trying it out now to see what the result might look like | 14:47 |
| zbr|ruck | i raise few bugs myself, but other than this is ok. i can make a POC change if you want. | 14:47 |
| zbr|ruck | just tell me which repo. | 14:47 |
| fungi | when v4 was first released it looked like they had added new rules which were only relevant for roles targeting distribution on galaxy | 14:48 |
| pabelanger | same issue, after moving it from /tmp | 14:48 |
| pabelanger | so, must be something with brwap, or file-system | 14:48 |
| fungi | pabelanger: yeah, also unlikely to have been tmpreaper anyway because that usually only removes files older than some configurable timeframe (typically on the order of days) | 14:49 |
| zbr|ruck | yep 4.0 was really controversial one, but they learnt from it. we do neither use galaxy with tripleo roles, but got over it easily, even without excludes. | 14:49 |
| pabelanger | leaning towards file-system, since this just started happening today | 14:49 |
| zbr|ruck | a minimal meta.yml file helps documenting things, even if you never plan to publish. | 14:49 |
| pabelanger | but, maybe something in ansible 2.8 also changed | 14:50 |
| fungi | zbr|ruck: cool, i'm just trying on zuul-jobs with the <4 cap removed just to see how rough it might be | 14:50 |
| zbr|ruck | fungi: what I did on other repos was to temporary introduce excludes to ease migration and slowly removing them later. so people have time to adapt. this avoids wide changes. | 14:51 |
| AJaeger | if we investigate ansible-lint, let's look at https://review.opendev.org/532083 for a job... | 14:52 |
| fungi | huh... trying to get a baseline run before removing the cap, and ansible-lint doesn't find the roles directory in the zuul-jobs repo | 14:53 |
| fungi | when it gets to checking the playbooks, it can't find roles referenced in them which are included in the repo | 14:53 |
| fungi | does it work locally on current master branch state of zuul-jobs for anyone else? | 14:54 |
| AJaeger | fungi: did you set ANSIBLE_ROLES_PATH ? | 14:56 |
| fungi | oh, i guess we don't set that in tox? | 14:57 |
| AJaeger | there's a comment in tox.ini about it | 14:57 |
| fungi | aha | 14:57 |
| fungi | thanks AJaeger! | 14:58 |
| *** michael-beaver has quit IRC | 15:00 | |
| AJaeger | fungi: working now? | 15:02 |
| fungi | not sure yet, but will know shortly | 15:03 |
| fungi | i tried setting it to a relative path first but that didn't work, so being canonical with it now | 15:03 |
| fungi | but takes a while before it starts to go through the playbooks | 15:03 |
| openstackgerrit | Mark Meyer proposed zuul/zuul master: Extend event reporting https://review.opendev.org/662134 | 15:05 |
| fungi | oh, duh, i was trying to assign it in passenv rather than exporting it in the calling environment | 15:05 |
| *** mattw4 has quit IRC | 15:05 | |
| corvus | tobiash: https://review.opendev.org/666982 lgtm; do you want to update with clarkb's suggestion? | 15:20 |
| tobiash | I can do that if you want (later this evening) | 15:25 |
| corvus | tobiash: however you want to handle it is fine with me | 15:25 |
| tobiash | then I'd suggest a followup as this is an undocumented switch anyway | 15:26 |
| *** hashar has joined #zuul | 15:29 | |
| *** swest has quit IRC | 15:31 | |
| clarkb | ya I +2'd because it isn't an advertised option so happy for it to go in as is | 15:34 |
| *** tosky has quit IRC | 15:37 | |
| openstackgerrit | Fabien Boucher proposed zuul/zuul master: URLTrigger driver time based https://review.opendev.org/635567 | 15:37 |
| *** sshnaidm is now known as sshnaidm|afk | 15:40 | |
| pabelanger | tobiash: okay, no issues with prepare-workspace-git I see, promote it to our base jobs now | 15:48 |
| pabelanger | then will start caching repos with DIB | 15:48 |
| openstackgerrit | Merged zuul/zuul master: Add missing docker daemon installation for Ubuntu/Debian https://review.opendev.org/665520 | 15:55 |
| pabelanger | fungi: welp, rsync is working again, so have to stop debugging for now | 15:56 |
| pabelanger | but was limited to a single executor | 15:56 |
| *** mattw4 has joined #zuul | 15:57 | |
| *** igordc has joined #zuul | 15:59 | |
| fungi | pabelanger: oh, that's definitely good to know that it was only one executor. what's the filesystem backed by? | 16:01 |
| pabelanger | fungi: ceph | 16:01 |
| fungi | cephfs or block storage on ceph? | 16:02 |
| *** jpena|mtg is now known as jpena|off | 16:04 | |
| openstackgerrit | Merged zuul/zuul master: Return store object instead of a function https://review.opendev.org/665642 | 16:12 |
| pabelanger | fungi: block storage, this is in vexxhost | 16:16 |
| pabelanger | file-system is ext4 | 16:16 |
| clarkb | pabelanger: fwiw I'm in the middle of a gitea06 rebuild because its fs got corrupted after/during a live migration there | 16:17 |
| clarkb | pabelanger: possible your host is in the same situation? | 16:17 |
| pabelanger | clarkb: maybe, but I don't see anything warning in kernel or dmesg | 16:18 |
| pabelanger | the system looks healthy | 16:18 |
| fungi | pabelanger: yeah, if there were problems with the backend then i would expect filesystem errors and block device errors in dmesg, so sounds like there really is something less obvious at work | 16:21 |
| pabelanger | for now, I've switched to prepare-workspace-git, to get more inline with others here. And start working to setup git cache for dib | 16:22 |
| pabelanger | that _should_ reduce the amount of stuff we need to rsync | 16:22 |
| *** hashar has quit IRC | 16:24 | |
| *** bhavikdbavishi has quit IRC | 16:27 | |
| openstackgerrit | Merged zuul/zuul master: Add missing start-message in pipeline config schema https://review.opendev.org/665936 | 16:30 |
| *** sanjayu_ has joined #zuul | 16:33 | |
| *** saneax has quit IRC | 16:35 | |
| *** bhavikdbavishi has joined #zuul | 16:43 | |
| openstackgerrit | Merged zuul/zuul master: Add missing doc for pipeline start-message https://review.opendev.org/665930 | 16:53 |
| *** hwangbo has joined #zuul | 17:12 | |
| openstackgerrit | Merged zuul/zuul master: Add support for item.change for pipeline start-message formater https://review.opendev.org/665968 | 17:12 |
| openstackgerrit | James E. Blair proposed zuul/nodepool master: WIP test ubuntu builds https://review.opendev.org/667624 | 17:15 |
| *** zbr|ruck is now known as zbr | 17:18 | |
| *** rf0lc0 is now known as rfolco | 17:30 | |
| *** panda has quit IRC | 17:31 | |
| openstackgerrit | Merged zuul/zuul master: Allow config projects to override allowed-projects https://review.opendev.org/666733 | 17:33 |
| *** panda has joined #zuul | 17:35 | |
| *** pleia2_ has joined #zuul | 17:49 | |
| openstackgerrit | James E. Blair proposed zuul/nodepool master: WIP test ubuntu builds https://review.opendev.org/667624 | 17:49 |
| *** tflink_ has joined #zuul | 17:50 | |
| *** sanjayu_ has quit IRC | 17:51 | |
| *** klindgren_ has joined #zuul | 17:52 | |
| fungi | what does {{- varname -}} do? i can't seem to find any reference to that syntax in the jinja2 docs, but we do it in some roles like https://opendev.org/zuul/zuul-jobs/src/branch/master/roles/multi-node-firewall/tasks/main.yaml#L26 | 17:53 |
| *** pleia2 has quit IRC | 17:53 | |
| *** tflink has quit IRC | 17:53 | |
| *** sileht has quit IRC | 17:53 | |
| fungi | maybe a cleaner example is https://opendev.org/zuul/zuul-jobs/src/branch/master/roles/multi-node-hosts-file/tasks/main.yaml#L9 | 17:53 |
| *** klindgren has quit IRC | 17:53 | |
| pabelanger | fungi: I think strips whitespaces | 17:55 |
| pabelanger | but I haven't see that before | 17:55 |
| *** altlogbot_1 has quit IRC | 17:55 | |
| fungi | i thought that's what {{%- varname -%}} did | 17:56 |
| pabelanger | yah | 17:56 |
| pabelanger | that's what I've seen before | 17:56 |
| pabelanger | {{- foo -}} is new to me | 17:56 |
| fungi | we do it in a few places in zuul-jobs so trying to figure out whether it's an actual part of the syntax or a mistake | 17:57 |
| pabelanger | a quick google say it is valid | 17:57 |
| pabelanger | and related to whitespace | 17:57 |
| *** bhavikdbavishi has quit IRC | 17:58 | |
| fungi | huh, okay. is there a more correct syntax? | 17:58 |
| openstackgerrit | Merged zuul/zuul master: Strengthen the caution about allowed-projects https://review.opendev.org/667241 | 17:58 |
| fungi | ansible-lint rule 206 considers it a problem because there's no space after the {{ or before the }} | 17:58 |
| pabelanger | oh | 17:58 |
| pabelanger | ha | 17:58 |
| pabelanger | https://github.com/ansible/ansible-lint/issues/499 | 17:58 |
| pabelanger | that is what I was looking at | 17:59 |
| pabelanger | likely related | 17:59 |
| pabelanger | fungi: I've usually used jinja2 filters to trim or strip | 17:59 |
| fungi | got it, so 4.1.1 will fix it | 17:59 |
| fungi | yeah, that's exactly what i ran across | 18:00 |
| *** altlogbot_3 has joined #zuul | 18:00 | |
| *** sileht has joined #zuul | 18:01 | |
| *** altlogbot_3 has quit IRC | 18:01 | |
| fungi | thanks! | 18:03 |
| *** altlogbot_1 has joined #zuul | 18:03 | |
| *** pwhalen has joined #zuul | 18:06 | |
| openstackgerrit | Jeremy Stanley proposed zuul/zuul-jobs master: Advance ansible-lint cap to test with 4 https://review.opendev.org/667695 | 18:11 |
| openstackgerrit | Jeremy Stanley proposed zuul/zuul-jobs master: Make all lines less than 160 characters long https://review.opendev.org/667696 | 18:11 |
| openstackgerrit | Jeremy Stanley proposed zuul/zuul-jobs master: Don't compare to literal True/False https://review.opendev.org/667697 | 18:11 |
| openstackgerrit | Jeremy Stanley proposed zuul/zuul-jobs master: Be consistent about spaces before and after vars https://review.opendev.org/667698 | 18:11 |
| fungi | zbr: ^ that's my attempt at convincing myself that ansible-lint is worth keeping, and to serve as a new conversation starter on the subject | 18:11 |
| openstackgerrit | Sorin Sbarnea proposed zuul/zuul-jobs master: POC: Run linters via pre-commit https://review.opendev.org/667699 | 18:14 |
| zbr | fungi: yeah, that's the idea. sadly regarding 206 there is still a pending bug, but using "# noqa 206" works well. | 18:15 |
| AJaeger | great work, fungi ! Thanks! | 18:17 |
| openstackgerrit | Merged zuul/zuul master: Stop bind mounting zuul dir into bwrap https://review.opendev.org/666972 | 18:20 |
| fungi | if we want this, we can just wait to merge the change which un-excludes 206 when they tag 4.1.1, which hopefully won't be much longer? | 18:21 |
| openstackgerrit | Jeremy Stanley proposed zuul/zuul-jobs master: Don't compare to literal True/False https://review.opendev.org/667697 | 18:23 |
| openstackgerrit | Jeremy Stanley proposed zuul/zuul-jobs master: Be consistent about spaces before and after vars https://review.opendev.org/667698 | 18:23 |
| corvus | fungi: i'm really uncomfortable with a line length limit. because of all the possible syntax values, it's very difficult for me to know exactly how to safely break lines. | 18:27 |
| fungi | i concur. that's the ugliest of the lot to me | 18:29 |
| fungi | this is also why i broke each rule into its own change, so we can decide to omit some of those and keep them excluded forever if we prefer | 18:30 |
| fungi | please -1 any you disagree with (including the first in the series for that matter) | 18:30 |
| fungi | as i said, this was exploratory work to evaluate whether we want to keep using ansible-lint | 18:31 |
| corvus | yeah, i left that as a -1 | 18:31 |
| fungi | i didn't feel i could justifiably disagree with continuing to use it unless i actually checked how useful the new rules it's been adding actually arte | 18:32 |
| fungi | s/arte/are/ | 18:32 |
| openstackgerrit | Alex Schultz proposed zuul/zuul master: Additional note about branches for implied-branches https://review.opendev.org/667415 | 18:32 |
| corvus | i've got a -1 on that, and a 0-question on the boolean one | 18:33 |
| fungi | cool | 18:33 |
| fungi | thanks! | 18:33 |
| fungi | corvus: wow, https://github.com/ansible/ansible-lint/commit/a02e6669cd73e0057a846854190c6b9fb207d5cf#diff-005e27ce9f611f6b5d67b2a0172432d9R15 even set that to severity "high" | 18:37 |
| openstackgerrit | Merged zuul/zuul master: Fix build leak with paused jobs https://review.opendev.org/667072 | 18:39 |
| fungi | seems https://github.com/ansible/ansible-lint/pull/384 included very minimal discussion and only rational about avoiding != | 18:39 |
| fungi | er, rationale | 18:39 |
| corvus | okay, so the rule allows "is True" it only disallows "== True" | 18:43 |
| fungi | i guess "is not False" could also work there? | 18:43 |
| corvus | yeah that should be fine too | 18:43 |
| corvus | that's better than what i thought the rule was | 18:43 |
| fungi | or similar | 18:43 |
| clarkb | I'm guess the rule is there to encourage use of truthy values? | 18:44 |
| fungi | also happy to rework the edits for that change to use explicit bool matches without equivalence operators | 18:44 |
| *** michael-beaver has joined #zuul | 18:44 | |
| corvus | fungi: i think at the least one would need to go through and make sure that the equality (or identity) test wasn't necessary | 18:45 |
| fungi | though "foo == True" is generally served by "foo" while "foo is True" is much more specific as the type for foo must also be boolean, right? | 18:45 |
| corvus | fungi: yeah | 18:45 |
| corvus | "foo" and "foo is True" are really very different statements. | 18:46 |
| corvus | "foo == True" is somewhere in the middle :) | 18:47 |
| fungi | so replacing "foo == True" with just "foo" should be identical. if we actually wanted "foo is True" then we should have been using it | 18:47 |
| fungi | because "foo == True" doesn't require that foo be a boolean | 18:47 |
| fungi | oh, huh, actually arbitrary strings don't evaluate == True | 18:49 |
| fungi | for some reason i thought they did | 18:49 |
| corvus | right, so i think "foo == True" is much closer to "foo is True" | 18:49 |
| corvus | it's in the negation where things get weirder | 18:50 |
| fungi | so `if "xyzzy":` doesn't get you the same as `if "xyzzy" == True:` | 18:50 |
| corvus | correct | 18:50 |
| corvus | '' == False is also false | 18:50 |
| fungi | there is apparently a ternary condition for boolean evaluation of non-boolean types in python that i hadn't realized until now | 18:51 |
| fungi | if you cast with bool() you get the behavior i was thinking the equivalence operator also used | 18:52 |
| fungi | how was i unaware of this? gonna need an extra drink tonight to cope with this new reality i've landed in | 18:52 |
| *** tflink_ is now known as tflink | 18:55 | |
| clarkb | [DEPRECATION WARNING]: Use errors="ignore" instead of skip. This feature will be removed in version 2.12. Deprecation warnings can be disabled by setting deprecation_warnings=False in ansible.cfg. | 18:55 |
| clarkb | just got that from our producition ansible | 18:56 |
| clarkb | does the linter catch those deprecated things too? I could see how those would be useful at least | 18:56 |
| fungi | sounds like from what zbr was saying in #openstack-infra ansible-lint implements its own (rudimentary) ansible parser and doesn't actually get ansible to parse the files | 18:57 |
| corvus | if folks are interested in improving the testing situation for zuul-jobs, there is some significant work that needs to be done to move testing from openstack-zuul-jobs into zuul-jobs -- that's an opportunity to set up a better model for testing roles in zuul-jobs. it's a blocker for moving zuul-jobs to the zuul tenant, and i'm happy to help anyone interested get started. | 18:57 |
| openstackgerrit | Merged zuul/zuul master: Split ansible output into its own sub logger https://review.opendev.org/667131 | 18:59 |
| corvus | i think we could get to a point where we could easily have full functional testing for any role in zuul-jobs | 19:00 |
| clarkb | hrm re the bool thing there is a deprecation warning from ansible saying you need to | bool prior to 2.12 | 19:01 |
| clarkb | in which case == True should be fine | 19:01 |
| zbr | ansible2.8 introduce a huge amount of deprecations and writing code that works without warnings on multiple versions of ansible became a real challenge. | 19:01 |
| clarkb | (my theory on using truthy values is wrong then) | 19:01 |
| zbr | if you are curious, look at how ugly was one of the workarounds https://github.com/ansible/molecule/pull/2124/files | 19:02 |
| zbr | clarkb: i did not find any issues with conditionals, to me "when: foo.exists" is the pythonic way of doing it. the linter is already going in the same direction: make it pythonic. | 19:03 |
| clarkb | zbr: yes but now you must do foo.exists | bool | 19:04 |
| clarkb | in which case foo.exists | bool == True is fine | 19:04 |
| zbr | hmm,... i did not observe the need for this. | 19:04 |
| clarkb | I would agree with you if | bool wasn't being forced on us | 19:04 |
| clarkb | zbr: ansible 2.12 will require | bool according to the warning I just saw | 19:04 |
| zbr | lucky for us 2.12 is many months away, this may change. | 19:05 |
| zbr | i would not be the first time when I see a minor patch undoing some deprecation, based on user feedback | 19:05 |
| zbr | it would make sense to require |bool if you do it inside jinja2, but on a when clause which does not use jinja {{ }} I would make it implicit | 19:07 |
| fungi | i think corvus's suggestion is very compelling. rather than have some style checker act as a proxy for whether we've written valid ansible, actually execute those jobs within a job and let the real ansible parser (every version of it we care about) tell us whether it works | 19:12 |
| zbr | you cannot replace linting with unit/functional testing. Ansible will not be able to catch style issue, also running quickly it would be impossible. why do we have so many linters in python? | 19:14 |
| fungi | because there are plenty of projects where people care more about what the source code looks like than how it runs | 19:14 |
| openstackgerrit | Merged zuul/zuul master: Remove non working tests/base.py ZuulTestCase.getPipeline method https://review.opendev.org/667351 | 19:14 |
| zbr | also ansible has zero support to treat warnings as errors. unless you write some output parser, you would be able to catch deprecations. | 19:14 |
| fungi | style checkers force people to write things legibly, they don't actually test whether the written thing works properly | 19:15 |
| zbr | fungi: true. i did not say that you do not also need to test them. | 19:15 |
| fungi | the main argument i've heard so far for keeping ansible-lint is that it catches mistakes in ansible syntax, not that people have been writing illegible ansible | 19:16 |
| zbr | that's why I find molecule quite useful as I can run roles on lots of platforms easily, and catch bugs, | 19:16 |
| zbr | nobody is willing to recognize he is writing poor code. and one example is the lack of names on tasks. | 19:17 |
| fungi | the test of upgrading to ansible-lint 4(.1.0) indicates that the new rules they've added have identified no syntactic bugs in zuul-jobs, but complain about three conditions one of which we'll almost certainly ignore, one of which is of dubious value and one of which is still rather broken | 19:17 |
| zbr | ansible is perfectly normal with no names on tasks, but this makes debugging execution quite hard. | 19:17 |
| zbr | fungi: linters help in avoiding discussion about code styling during reviews, which could become really stressful, I seen reviewer-A requesting a change, and one day later reviewer-B asking for the other one. With a linter you make the choice impersonal. | 19:19 |
| fungi | maybe we want ansible-lint with only rules we care about turned on? is that an option without playing whack-a-mole with new exclusions every time there's a release? | 19:20 |
| zbr | sure, picking which rules matter for you is choice of each project. | 19:20 |
| zbr | bending the rules a *bit* is not bad. | 19:21 |
| fungi | just wondering if it's possible to tell ansible-lint to load no rules except the ones we want (i.e. whitelisting a few rules instead of blacklisting a majority of them) | 19:21 |
| fungi | for example, if all we wanted to do was make sure tasks all have names | 19:22 |
| zbr | especially with codebases with long history is not resonable to make a total switch, too much effort and conflicts. | 19:22 |
| zbr | i am not sure about whitelisting but by pinning it you can obtain the same predictable behavior. | 19:22 |
| zbr | there are two projects which I would always ping: flake8 and ansible-lint (based on experience) | 19:23 |
| zbr | pin | 19:23 |
| fungi | pinning works until we need a newer ansible-lint because the python version it worked on is no longer available to us | 19:23 |
| zbr | fungi: that's true for other jobs too, including sphinx 3. is just maintenance | 19:24 |
| fungi | yes, and then the rule exclusions would need to be updated at the same time if blacklisting is the only option | 19:25 |
| zbr | with the scheduled death of py2 we will see more of these. | 19:25 |
| zbr | fungi: yeah, usually doing the bumping involves a combination of fixes on new issues and new blacklists. not really a big deal. probably done twice a year. | 19:26 |
| fungi | just questioning whether that is worth the hassle if we only actually want to enforce a handful of style consistencies (e.g. naming tasks) | 19:27 |
| zbr | i hope you will, i find it easy to search for a task by name, not to mention that without tasks names is impossible to read the performance reports from callbacks like timer/... | 19:34 |
| zbr | i am referring to the profile_tasks callback which I use quite often, https://sketchingdev.co.uk/blog/profiling-ansible-playbooks-to-csv.html | 19:35 |
| *** panda has quit IRC | 19:39 | |
| openstackgerrit | Merged zuul/zuul master: Parallelize github event processing https://review.opendev.org/666982 | 19:39 |
| *** panda has joined #zuul | 19:40 | |
| *** altlogbot_1 has quit IRC | 19:45 | |
| *** altlogbot_0 has joined #zuul | 19:48 | |
| *** hwangbo has quit IRC | 19:55 | |
| *** armstrongs has joined #zuul | 19:58 | |
| openstackgerrit | Merged zuul/zuul master: Differentiate between queued and waiting jobs in zuul web UI https://review.opendev.org/660878 | 19:59 |
| openstackgerrit | Dirk Mueller proposed zuul/nodepool master: Switch to openSUSE 15.1 https://review.opendev.org/667532 | 20:09 |
| openstackgerrit | Dirk Mueller proposed zuul/nodepool master: Add Python 3.7 testing https://review.opendev.org/667720 | 20:09 |
| *** armstrongs has quit IRC | 20:09 | |
| pabelanger | fungi: from what I've seen, there has been a large push on ansible-lint as the way to 'score' roles on galaxy. I can see wanting to have common readability there. | 20:14 |
| *** altlogbot_0 has quit IRC | 20:15 | |
| openstackgerrit | Merged zuul/zuul master: [doc][monitoring] Fix the wait_time parent attribute https://review.opendev.org/667342 | 20:18 |
| *** altlogbot_2 has joined #zuul | 20:20 | |
| *** armstrongs has joined #zuul | 20:21 | |
| *** armstrongs has quit IRC | 20:25 | |
| *** altlogbot_2 has quit IRC | 20:43 | |
| *** altlogbot_3 has joined #zuul | 20:46 | |
| *** altlogbot_3 has quit IRC | 21:00 | |
| fungi | pabelanger: this i can understand, but grading zuul jobs the same way probably doesn't make sense | 21:00 |
| pabelanger | I'd agree | 21:02 |
| fungi | it may mean that the intended audience of ansible-lint is drifting significantly farther from what we need for zuul jobs as time goes on | 21:02 |
| *** tosky has joined #zuul | 21:02 | |
| pabelanger | I've been usually a lot of yamllint recently, to help ensure yaml syntax is valid | 21:03 |
| pabelanger | but still have some roles using ansible-lint, but also pinned | 21:03 |
| *** pcaruana has quit IRC | 21:05 | |
| *** altlogbot_3 has joined #zuul | 21:06 | |
| fungi | just noting that pinning ansible-lint forever isn't really an option | 21:07 |
| fungi | either we use supported versions of things or we replace them with supported versions of other things or we live without | 21:07 |
| *** armstrongs has joined #zuul | 21:09 | |
| armstrongs | Quick question with the common zuul roles that are provided. I would like to provide my own common roles repo in the same format. Will I be able to provide these as common roles that can be used in projects? | 21:13 |
| armstrongs | Is it as simple as just putting an untrusted project with the roles in it? | 21:14 |
| fungi | armstrongs: yes, it needs to be in the list of untrusted (at least) projects in your zuul configuration, *and* it needs to be in the "roles" list for jobs which are trying to reference its roles from another project | 21:17 |
| armstrongs | Thanks | 21:18 |
| *** ianychoi has quit IRC | 21:23 | |
| *** hwangbo has joined #zuul | 21:25 | |
| *** armstrongs has quit IRC | 21:25 | |
| *** ianychoi has joined #zuul | 21:28 | |
| *** panda has quit IRC | 21:42 | |
| *** panda has joined #zuul | 21:45 | |
| openstackgerrit | James E. Blair proposed zuul/nodepool master: Switch functional testing to a devstack consumer job https://review.opendev.org/665023 | 21:55 |
| *** michael-beaver has quit IRC | 22:04 | |
| *** rlandy is now known as rlandy|bbl | 22:20 | |
| *** mattw4 has quit IRC | 22:37 | |
| *** armstrongs has joined #zuul | 23:07 | |
| *** armstrongs has quit IRC | 23:08 | |
| *** sanjayu_ has joined #zuul | 23:08 | |
| *** tosky has quit IRC | 23:12 | |
| *** threestrands has joined #zuul | 23:15 | |
| *** sgw has quit IRC | 23:16 | |
| *** igordc has quit IRC | 23:25 | |
| *** threestrands has quit IRC | 23:29 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!