| *** bhavik1 has joined #zuul | 04:04 | |
| *** bhavik1 has quit IRC | 05:18 | |
| *** lennyb has quit IRC | 06:41 | |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Add stats for executor and merger count https://review.openstack.org/513940 | 07:44 |
|---|---|---|
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Report initial pipeline changes under correct key https://review.openstack.org/513949 | 07:49 |
| openstackgerrit | Andreas Jaeger proposed openstack-infra/zuul feature/zuulv3: Fix typo, it's branches https://review.openstack.org/514040 | 08:17 |
| tobiash | jeblair: the result data change (504808) introduced a race in the data_return test | 08:23 |
| tobiash | jeblair: I hunted it down to ' If more than one parent job returns the same variable, the value from the later job in the job graph will take precedence.' being incorrect as the JobGraph doesn't have such a knowledge | 08:25 |
| tobiash | jeblair: it only has this knowlege in terms of alphabetically sorted job names (but doesn't expose this) | 08:25 |
| tobiash | jeblair: so I think there are two options | 08:26 |
| tobiash | jeblair: first is to drop this guarantee, add a note that multiple parent jobs returning the same variable lead to races (and adapt the test case to not make use of this guarantee) | 08:27 |
| tobiash | jeblair: second is to change this guarantee in 'later job in terms of alphabetically sort order', fix the getParentJobsRecursively and eventually the test case | 08:28 |
| *** lennyb has joined #zuul | 08:38 | |
| tobiash | jeblair: hm, alphabetically sorted is bad as this could lead to unexpected behavior: | 08:43 |
| tobiash | jeblair: for jobs b -> a -> c and a and b returning the same variable the user would expect the variable of a to be taken so we would need a topological sort here | 08:45 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul feature/zuulv3: Make multiple parent jobs returning same var undefined behavior https://review.openstack.org/514044 | 08:51 |
| tobiash | jeblair: (easy) option 1 ^ | 08:51 |
| tobiash | jeblair: thinking of option 2 (topological sort) can get somewhat tricky as we have a graph, not a tree | 08:55 |
| tobiash | jeblair: I'll try a finish time based sort order (which I think is what a user would expect) that naturally leverages that child jobs will be able to overwrite vars returned by their parent jobs in a multi layer graph | 08:56 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul feature/zuulv3: Use end time based sorting for applying parent data https://review.openstack.org/514049 | 09:42 |
| tobiash | jeblair: time base sorting as option 2 ^ (which I'd prever over option 1) | 09:42 |
| tobiash | jeblair: using a topological sort as option 3 is probably slightly more work | 09:43 |
| tobiash | jeblair: and a sort oder by definition in config is probably not really defined by definition as the graph is potentially constructed by configs in several repos | 09:44 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul feature/zuulv3: Fix undefined sort order when applying parent data https://review.openstack.org/514056 | 10:39 |
| tobiash | jeblair: option 3 to fix the order to the documented order ^ | 10:40 |
| *** bhavik1 has joined #zuul | 11:03 | |
| *** bhavik1 has quit IRC | 11:05 | |
| tobiash | jeblair: I'm in favor of the last one as it just fixes the behavior such it is complient with the doc | 11:06 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul-jobs master: Have tox-siblings edit requirements files instead https://review.openstack.org/514058 | 11:11 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Fix doc typo https://review.openstack.org/513932 | 11:15 |
| *** harlowja has quit IRC | 11:28 | |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul-jobs master: Have tox-siblings edit requirements files instead https://review.openstack.org/514058 | 11:46 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul-jobs master: Have tox-siblings edit requirements files instead https://review.openstack.org/514058 | 12:00 |
| *** yolanda has joined #zuul | 12:09 | |
| *** yolanda has quit IRC | 12:13 | |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul-jobs master: Have tox-siblings edit requirements files instead https://review.openstack.org/514058 | 12:18 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul-jobs master: Have tox-siblings edit requirements files instead https://review.openstack.org/514058 | 12:46 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul-jobs master: Have tox-siblings edit requirements files instead https://review.openstack.org/514058 | 12:54 |
| *** yolanda has joined #zuul | 13:11 | |
| openstackgerrit | David Moreau Simard proposed openstack-infra/zuul-jobs master: Persist iptables rules https://review.openstack.org/513943 | 15:19 |
| *** __zeus__ has joined #zuul | 16:25 | |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Fix typo, it's branches https://review.openstack.org/514040 | 19:08 |
| *** __zeus__ has quit IRC | 19:40 | |
| *** __zeus__ has joined #zuul | 21:05 | |
| *** __zeus__ has quit IRC | 23:23 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!