| *** larainema has quit IRC | 01:58 | |
| *** caphrim007 has joined #openstack-jjb | 02:53 | |
| *** caphrim007_ has joined #openstack-jjb | 03:20 | |
| *** caphrim007 has quit IRC | 03:24 | |
| *** hashar has joined #openstack-jjb | 06:53 | |
| *** electrofelix has joined #openstack-jjb | 07:58 | |
| ssbarnea | zxiiro: is nothing, by monday we should consider this done. I was considering writing a gerritcop that does rebase, asks for refresh and eventually abandons outdated reviews. there are few of them for github but none for gerrit (or i failed to find it) | 09:19 |
|---|---|---|
| ssbarnea | a project that keeps CRs updated is always better | 09:20 |
| ssbarnea | but as always, there are not enough hours in a day | 09:20 |
| openstackgerrit | Sorin Sbarnea proposed openstack-infra/jenkins-job-builder master: conditional_publisher: add support for Matrix Aggregation https://review.openstack.org/460898 | 10:53 |
| openstackgerrit | Merged openstack-infra/jenkins-job-builder master: wrappers: fix XML node name for scrip-env's only-run-on-parent https://review.openstack.org/560684 | 10:54 |
| openstackgerrit | Merged openstack-infra/jenkins-job-builder master: gitlab trigger: add support for different MR events https://review.openstack.org/513712 | 10:54 |
| ssbarnea | electrofelix: one last eye on https://review.openstack.org/#/c/560842/ please. | 10:55 |
| openstackgerrit | Merged openstack-infra/jenkins-job-builder master: Add helper tag to join arbitrary lists into strings https://review.openstack.org/265926 | 11:00 |
| electrofelix | ssbarnea: do we need to inspect REQUESTS_CA_BUNDLE or CURL_CA_BUNDLE? or are we able to just leave that entire up to requests? my understand was the problem was setting the session.verify value to False would fail to override env settings | 11:34 |
| electrofelix | so the wrapper just needs to check if the session verify is explicitly set to False, and if so, ensure the passed through value to the merge method is also false, otherwise the default behaviour works as desired? | 11:35 |
| ssbarnea | electrofelix: sadly no because current version of request is buggy regarding this aspect. In the future, once they solve it, we can bump reqs version and remove our code. | 11:37 |
| ssbarnea | electrofelix: i did spend most of the yesterday on investigating a fix for that and this was the only solution I got. i had to test with all combinations of: cert(global/present-in-bundle/not-present-or-invalid), env var defined or not, verify defined or not.... a total pleasure. | 11:40 |
| electrofelix | can you elaborate on what use cases that block is needed? | 11:40 |
| electrofelix | guess I'm trying to work out why the following code in requests doesn't take care of that scenario? https://github.com/requests/requests/blob/master/requests/sessions.py#L683-L685 | 11:40 |
| ssbarnea | electrofelix: mainly they will fix it in request 3.x as they decided not to break current behaviour in current version. | 11:42 |
| ssbarnea | i will test one more time with that section removed | 11:45 |
| electrofelix | ssbarnea: looking at the code in python-jenkins, if the code to test REQUESTS_CA_BUNDLE and CURL_CA_BUNDLE is dropped, the call to super(WrappedSession, self).merge_environment_settings(...) results in performing the same test on the value of verify being True or None and then inspecting the two env variables at https://github.com/requests/requests/blob/master/requests/sessions.py#L683-L685 so it looks like it just duplicates what would happen any | 11:45 |
| ssbarnea | indeed, but i remember that it didn't work, give me 5mins to check. | 11:45 |
| electrofelix | if not, then we probably need something in the commit message or a comment to remind us of the fact that the behaviour is different and explain why | 11:46 |
| electrofelix | unless we're doing something to unset trust_env on the session? | 11:47 |
| openstackgerrit | Sorin Sbarnea proposed openstack/python-jenkins master: Improve SSL support with requests lib https://review.openstack.org/560842 | 11:52 |
| ssbarnea | electrofelix: you were right, those lines were not needed. i guess i added them before implementing the _send workaround. | 11:53 |
| ssbarnea | i removed them and also improved the description | 11:53 |
| ssbarnea | now i remember that the merge..() was not working, but it was not working because _send() was not using it, by requests design decisions. | 11:54 |
| electrofelix | good to get it confirmed | 11:54 |
| ssbarnea | electrofelix: yep last two days make me learn few new things about requests lib, used to like it a lot before knowing more :D | 11:55 |
| electrofelix | nearly always the case when you have to go deep | 12:02 |
| openstackgerrit | Akshat Tandon proposed openstack/python-jenkins master: Adds support for multi-select build parameters https://review.openstack.org/517361 | 12:40 |
| *** caphrim007_ has quit IRC | 14:30 | |
| *** frayedknot has joined #openstack-jjb | 14:43 | |
| frayedknot | I have another question that relates to my parameter question yesterday - it seems like I'm having trouble to get interpolation of parameters within parameters working correctly: https://pastebin.com/qA8WJQ7K | 14:49 |
| frayedknot | Rather than getting a default value of 'feature/prj-123-test' I get '{feature-branch}-test' | 14:49 |
| *** openstackgerrit has quit IRC | 14:50 | |
| zxiiro | frayedknot: you cannot nest variables in JJB | 14:55 |
| frayedknot | dang, ok! | 14:55 |
| frayedknot | I misinterpreted this from the documentation: "JJB also supports interpolation of parameters within parameters." | 14:58 |
| frayedknot | via https://docs.openstack.org/infra/jenkins-job-builder/definition.html?highlight=interpolation | 14:58 |
| zxiiro | frayedknot: yeah so I think what it means is you pass the result parameters from outside of the template into the template | 15:00 |
| zxiiro | frayedknot: you see they do a 'downstream: {var1}_{var2}_bar' thing | 15:01 |
| zxiiro | frayedknot: but inside the template it's just "{downstream}" | 15:01 |
| zxiiro | frayedknot: so if you want to do that then from your Project section you'd pass in "BRANCH: feature/prj-123-test" | 15:02 |
| zxiiro | frayedknot: oh wait, you need it do it from a job-group | 15:02 |
| frayedknot | it appear to work if I do it in the project | 15:03 |
| zxiiro | it wouldn't work from project section because you cannot declare variables in projects | 15:03 |
| zxiiro | really? ok | 15:03 |
| *** openstackgerrit has joined #openstack-jjb | 15:48 | |
| openstackgerrit | Merged openstack/python-jenkins master: Improve SSL support with requests lib https://review.openstack.org/560842 | 15:48 |
| openstackgerrit | Sorin Sbarnea proposed openstack/python-jenkins master: enable cross testing with jenkins-job-builder https://review.openstack.org/561260 | 15:49 |
| *** caphrim007 has joined #openstack-jjb | 16:05 | |
| zxiiro | electrofelix: that {OS_USERNAME|my_user} is something I really wish I knew earlier. Seems like it's undocumented? I couldn't find it in docs... I can look into documenting it if it's not already somewhere and I couldn't find it. | 16:07 |
| electrofelix | zxiiro: oops, I guess I forgot to document it at the time it was added | 16:13 |
| electrofelix | should probably be part of https://docs.openstack.org/infra/jenkins-job-builder/definition.html#variable-references | 16:13 |
| zxiiro | electrofelix: cool. Yeah let me look into updating that. That made my day, it's something we've been struggling with for our jjb template repo. | 16:14 |
| *** hashar is now known as hasharAway | 16:24 | |
| openstackgerrit | Merged openstack/python-jenkins master: enable cross testing with jenkins-job-builder https://review.openstack.org/561260 | 16:39 |
| openstackgerrit | Sorin Sbarnea proposed openstack/python-jenkins master: Adds support for deleting builds https://review.openstack.org/289257 | 17:23 |
| zxiiro | electrofelix: So I think there's a bug in that feature. If I define {var|default} in a macro, and then call the macro in a job-temlpate without overriding anything. It results in the literal "{var|default}" in the resultant job rather than the expected "default" | 17:31 |
| zxiiro | I'm gonna document the feature from a job-template perspective but I think we need to figure out why it's not parsing correctly in macros. | 17:31 |
| electrofelix | zxiiro: possibly that relates to the same issue where we're not expanding macros at the same time as templates? | 17:32 |
| zxiiro | electrofelix: oh yeah possibly, I had forgotten about that one. | 17:33 |
| *** electrofelix has quit IRC | 17:38 | |
| openstackgerrit | Thanh Ha proposed openstack-infra/jenkins-job-builder master: Document use of | character for variable defaults https://review.openstack.org/561278 | 17:38 |
| openstackgerrit | Trevor Bramwell proposed openstack-infra/jenkins-job-builder master: Fix Postbuildscript Plugin Info Lookup https://review.openstack.org/561279 | 17:40 |
| *** caphrim007_ has joined #openstack-jjb | 17:48 | |
| *** caphrim007 has quit IRC | 17:52 | |
| *** caphrim007_ has quit IRC | 18:07 | |
| *** caphrim007 has joined #openstack-jjb | 18:07 | |
| *** smyers has quit IRC | 18:30 | |
| *** smyers has joined #openstack-jjb | 18:33 | |
| *** caphrim007_ has joined #openstack-jjb | 20:58 | |
| *** caphrim007 has quit IRC | 21:02 | |
| *** frayedknot has quit IRC | 21:30 | |
| *** caphrim007_ has quit IRC | 21:54 | |
| *** caphrim007 has joined #openstack-jjb | 21:55 | |
| *** hasharAway has quit IRC | 22:58 | |
| *** caphrim007_ has joined #openstack-jjb | 23:18 | |
| *** caphrim007 has quit IRC | 23:21 | |
| *** caphrim007 has joined #openstack-jjb | 23:21 | |
| *** caphrim007_ has quit IRC | 23:21 | |
| *** caphrim007 has quit IRC | 23:26 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!