)]}'
{"/COMMIT_MSG":[{"author":{"_account_id":748,"name":"Armando Migliaccio","email":"armamig@gmail.com","username":"armando-migliaccio"},"change_message_id":"5579f891470e18f5b418a7ba89fb729def091ad5","unresolved":false,"context_lines":[{"line_number":6,"context_line":""},{"line_number":7,"context_line":"thirdparty-ci: Don\u0027t allow voting -1"},{"line_number":8,"context_line":""},{"line_number":9,"context_line":"Everyone ignores -1 votes from thirdparty CI, so lets simply"},{"line_number":10,"context_line":"specify the fact we don\u0027t want that happening anymore."},{"line_number":11,"context_line":""},{"line_number":12,"context_line":"Change-Id: Ie93bc1903bc8c0f2ca76701ebe9ae7232baa8f5e"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":1,"id":"3a50d1a3_25c61469","line":9,"updated":"2015-07-30 00:17:46.000000000","message":"Let\u0027s not make such harsh statements in commit messages. I am sure there\u0027s someone out there who cares, or at least who used to.","commit_id":"d32d3ca000da4af9d1005015be3772f0241e44a7"},{"author":{"_account_id":6524,"name":"Henry Gessau","email":"HenryG@gessau.net","username":"gessau"},"change_message_id":"707c0f47c9c3e27e1e42b8aae9c34444f7701f9e","unresolved":false,"context_lines":[{"line_number":6,"context_line":""},{"line_number":7,"context_line":"thirdparty-ci: Don\u0027t allow voting -1"},{"line_number":8,"context_line":""},{"line_number":9,"context_line":"Everyone ignores -1 votes from thirdparty CI, so lets simply"},{"line_number":10,"context_line":"specify the fact we don\u0027t want that happening anymore."},{"line_number":11,"context_line":""},{"line_number":12,"context_line":"Change-Id: Ie93bc1903bc8c0f2ca76701ebe9ae7232baa8f5e"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":1,"id":"3a50d1a3_30ef30eb","line":9,"in_reply_to":"3a50d1a3_25c61469","updated":"2015-07-30 00:40:45.000000000","message":"OK I\u0027ll admit I did not read the commit message at first, and I agree with Armando here.\n\nI certainly look at the results of the third-party CIs a lot, and I care very much that they post their results.","commit_id":"d32d3ca000da4af9d1005015be3772f0241e44a7"},{"author":{"_account_id":105,"name":"Kyle Mestery","email":"mestery@mestery.com","username":"mestery"},"change_message_id":"31cbefbe7b7b44f2b0f9f31de7d777003d486cb2","unresolved":false,"context_lines":[{"line_number":6,"context_line":""},{"line_number":7,"context_line":"thirdparty-ci: Don\u0027t allow voting -1"},{"line_number":8,"context_line":""},{"line_number":9,"context_line":"Everyone ignores -1 votes from thirdparty CI, so lets simply"},{"line_number":10,"context_line":"specify the fact we don\u0027t want that happening anymore."},{"line_number":11,"context_line":""},{"line_number":12,"context_line":"Change-Id: Ie93bc1903bc8c0f2ca76701ebe9ae7232baa8f5e"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":1,"id":"1a4dcd0f_8521dcfe","line":9,"in_reply_to":"3a50d1a3_30ef30eb","updated":"2015-08-10 17:40:34.000000000","message":"OK, fine, I\u0027ll update this before we merge this.","commit_id":"d32d3ca000da4af9d1005015be3772f0241e44a7"}],"doc/source/policies/thirdparty-ci.rst":[{"author":{"_account_id":748,"name":"Armando Migliaccio","email":"armamig@gmail.com","username":"armando-migliaccio"},"change_message_id":"5579f891470e18f5b418a7ba89fb729def091ad5","unresolved":false,"context_lines":[{"line_number":60,"context_line":""},{"line_number":61,"context_line":"The Neutron team encourages you to NOT vote -1 with a third-party CI. False"},{"line_number":62,"context_line":"negatives are noisy to the community, and have given -1 from third-party"},{"line_number":63,"context_line":"CIs a bad reputation. Really bad, to the point of people ignoring them all."},{"line_number":64,"context_line":"Failure messages are useful to those doing refactors, and provide you"},{"line_number":65,"context_line":"feedback on the state of your plugin."},{"line_number":66,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"3a50d1a3_259db466","side":"PARENT","line":63,"updated":"2015-07-30 00:17:46.000000000","message":"I think the problem stems from the fact that infrastructure failures are not  silently dropped as they should. It\u0027s not a hard problem to abort a build due to infrastructure failures.","commit_id":"de2dd7aafeb4de49103d48e6a56083173bd40973"},{"author":{"_account_id":5263,"name":"Jeremy Stanley","display_name":"fungi","email":"fungi@yuggoth.org","username":"fungi","status":"missing, presumed fed"},"change_message_id":"0827f5a7e37c02bb7ae9c6e20d59db5e0f2af99c","unresolved":false,"context_lines":[{"line_number":60,"context_line":""},{"line_number":61,"context_line":"The Neutron team encourages you to NOT vote -1 with a third-party CI. False"},{"line_number":62,"context_line":"negatives are noisy to the community, and have given -1 from third-party"},{"line_number":63,"context_line":"CIs a bad reputation. Really bad, to the point of people ignoring them all."},{"line_number":64,"context_line":"Failure messages are useful to those doing refactors, and provide you"},{"line_number":65,"context_line":"feedback on the state of your plugin."},{"line_number":66,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"fa1b9901_b6800c8d","side":"PARENT","line":63,"in_reply_to":"1a4dcd0f_5fec0629","updated":"2015-08-22 20:13:36.000000000","message":"There are absolutely changes you could propose to neutron which would result in a pip install or git fetch error. I\u0027m not saying they\u0027re changes to often-edited files in the repo but they are certainly possible to cause. I wouldn\u0027t want to see our upstream testing simply ignore those, and I have a hard time recommending holding external testing systems to standards we can\u0027t even apply to our official CI.","commit_id":"de2dd7aafeb4de49103d48e6a56083173bd40973"},{"author":{"_account_id":6635,"name":"John Davidge","email":"john.davidge@rackspace.com","username":"john-davidge"},"change_message_id":"e44a8e14b52d0151c7d31af8ad659697472381c6","unresolved":false,"context_lines":[{"line_number":60,"context_line":""},{"line_number":61,"context_line":"The Neutron team encourages you to NOT vote -1 with a third-party CI. False"},{"line_number":62,"context_line":"negatives are noisy to the community, and have given -1 from third-party"},{"line_number":63,"context_line":"CIs a bad reputation. Really bad, to the point of people ignoring them all."},{"line_number":64,"context_line":"Failure messages are useful to those doing refactors, and provide you"},{"line_number":65,"context_line":"feedback on the state of your plugin."},{"line_number":66,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"3a50d1a3_78158c99","side":"PARENT","line":63,"in_reply_to":"3a50d1a3_259db466","updated":"2015-07-30 11:37:31.000000000","message":"Another issue is that a -1 can be caused due to a problem with the third party plugin rather than the patch being tested. In this case it\u0027s important to know that there is an issue with the plugin (by posting results), but we don\u0027t want a negative vote on the patch if that\u0027s not where the problem is.","commit_id":"de2dd7aafeb4de49103d48e6a56083173bd40973"},{"author":{"_account_id":5263,"name":"Jeremy Stanley","display_name":"fungi","email":"fungi@yuggoth.org","username":"fungi","status":"missing, presumed fed"},"change_message_id":"28b7805c9be43303268d948e6a0d33e1fb5d2543","unresolved":false,"context_lines":[{"line_number":60,"context_line":""},{"line_number":61,"context_line":"The Neutron team encourages you to NOT vote -1 with a third-party CI. False"},{"line_number":62,"context_line":"negatives are noisy to the community, and have given -1 from third-party"},{"line_number":63,"context_line":"CIs a bad reputation. Really bad, to the point of people ignoring them all."},{"line_number":64,"context_line":"Failure messages are useful to those doing refactors, and provide you"},{"line_number":65,"context_line":"feedback on the state of your plugin."},{"line_number":66,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"3a50d1a3_a4ed2db9","side":"PARENT","line":63,"in_reply_to":"3a50d1a3_78158c99","updated":"2015-07-30 12:47:38.000000000","message":"As a counter, there isn\u0027t great tooling available today to handle this and distinguishing between an \"infrastructure failure\" and a bad patch is non-trivial (especially when the code being tested has root access to the worker and is making local network changes!). Unless by \"not a hard problem\" you mean it can be achieved by the CI operator inspecting the logs after a job failure and manually triggering an actual -1 in those situations where it\u0027s actually warranted... that\u0027s what Dan Prince used to do for Red Hat\u0027s Smokestack CI, so there is at least precedent.\n\nOur upstream CI doesn\u0027t do much of anything to tell these conditions apart either, and we have enough cases where a patch proposed (especially to neutron) can fail in ways that look just like a devstack worker becoming unreachable due to issues in the underlying cloud provider.","commit_id":"de2dd7aafeb4de49103d48e6a56083173bd40973"},{"author":{"_account_id":14287,"name":"Waldemar Znoinski","email":"wznoinsk@redhat.com","username":"wznoinsk"},"change_message_id":"407b99f286808de6a21fa5f7fee47ec426832e43","unresolved":false,"context_lines":[{"line_number":60,"context_line":""},{"line_number":61,"context_line":"The Neutron team encourages you to NOT vote -1 with a third-party CI. False"},{"line_number":62,"context_line":"negatives are noisy to the community, and have given -1 from third-party"},{"line_number":63,"context_line":"CIs a bad reputation. Really bad, to the point of people ignoring them all."},{"line_number":64,"context_line":"Failure messages are useful to those doing refactors, and provide you"},{"line_number":65,"context_line":"feedback on the state of your plugin."},{"line_number":66,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"1a4dcd0f_5fec0629","side":"PARENT","line":63,"in_reply_to":"3a50d1a3_81e56add","updated":"2015-08-10 16:23:17.000000000","message":"a long shot but a plugin written (if anyone would want to write one) or extended from existing one (log-parser to be able to exit with certaing code (0/1)) for Jenkins for parsing devstack builds could have logic of understanding certain errors and deciding of yes/no \u0027commenting/voting\u0027 should be send back to Gerrit or not (instead of error codes Jenkins itself would exit with), it could be possibly integrated with elastic-search/check with Jenkins. Was there a lot of work already done on moving away from Jenkins?\n...anything that could be included in the common-openstack-ci (downstream-puppet) to standardize error/exception catching would be great.\n\nI do agree with Armando, I think it\u0027s doable to have the errors worked out and find out whether that\u0027s an infra/code problem (to some extend - nothing is perfect) - there will always be noise. I\u0027m doing it \u0027by eye and hand\u0027 for last 6+ months on our CI and sounds like certain patterns of issues occur (or at least happen in certain moments that are purelly infra/not infra related - pip install/git fetch with exit code !\u003d 0 would be infra problem, where starting a service more likely a code etc. - the logic would be somehow big to cover most issues).\n\nIn short, I would see standardization of testing as the way to go.","commit_id":"de2dd7aafeb4de49103d48e6a56083173bd40973"},{"author":{"_account_id":748,"name":"Armando Migliaccio","email":"armamig@gmail.com","username":"armando-migliaccio"},"change_message_id":"53442b10aef562015c4e4647a6a4a6baf8c2228e","unresolved":false,"context_lines":[{"line_number":60,"context_line":""},{"line_number":61,"context_line":"The Neutron team encourages you to NOT vote -1 with a third-party CI. False"},{"line_number":62,"context_line":"negatives are noisy to the community, and have given -1 from third-party"},{"line_number":63,"context_line":"CIs a bad reputation. Really bad, to the point of people ignoring them all."},{"line_number":64,"context_line":"Failure messages are useful to those doing refactors, and provide you"},{"line_number":65,"context_line":"feedback on the state of your plugin."},{"line_number":66,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"3a50d1a3_e70a4d27","side":"PARENT","line":63,"in_reply_to":"3a50d1a3_a4ed2db9","updated":"2015-07-30 16:32:13.000000000","message":"Jeremy, you obviously make a valid point. However, in my experience inspecting these logs, I have seen many failures where, you can clearly pinpoint you\u0027re having underlying issues. More often than not, a CI system doesn\u0027t even get to do a stack.sh, but perhaps stalls and die whilst creating/connecting to workers. Different error codes could be used to set apart these from other more genuine errors and some automatic post-processing can also help scale way down the amount of spurious -1\u0027s.\n\nIf we all did that, a handful of occasional -1 on a patch is really not that big deal IMO.","commit_id":"de2dd7aafeb4de49103d48e6a56083173bd40973"},{"author":{"_account_id":5263,"name":"Jeremy Stanley","display_name":"fungi","email":"fungi@yuggoth.org","username":"fungi","status":"missing, presumed fed"},"change_message_id":"37c1759d4855c873f0bfdb0e83b0089cc6f15f43","unresolved":false,"context_lines":[{"line_number":60,"context_line":""},{"line_number":61,"context_line":"The Neutron team encourages you to NOT vote -1 with a third-party CI. False"},{"line_number":62,"context_line":"negatives are noisy to the community, and have given -1 from third-party"},{"line_number":63,"context_line":"CIs a bad reputation. Really bad, to the point of people ignoring them all."},{"line_number":64,"context_line":"Failure messages are useful to those doing refactors, and provide you"},{"line_number":65,"context_line":"feedback on the state of your plugin."},{"line_number":66,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"3a50d1a3_81e56add","side":"PARENT","line":63,"in_reply_to":"3a50d1a3_e70a4d27","updated":"2015-07-30 21:33:13.000000000","message":"I\u0027m definitely in favor of additional \"error codes\" but for CI systems reliant on Jenkins (such as the upstream OpenStack one), job results are reported back as either SUCCESS or FAILURE (ignoring for now the existence of states like UNSTABLE or ABORT which you can\u0027t directly force from within the job). Instead you need some separate system inspecting the logs themselves for known failure modes and following up on matches, which is why elastic-recheck evolved.\n\nMore granular failure indicators have been on our wish list for a while, and that\u0027s one of the reasons we\u0027re working toward removing Jenkins from our CI architecture altogether.","commit_id":"de2dd7aafeb4de49103d48e6a56083173bd40973"},{"author":{"_account_id":748,"name":"Armando Migliaccio","email":"armamig@gmail.com","username":"armando-migliaccio"},"change_message_id":"5579f891470e18f5b418a7ba89fb729def091ad5","unresolved":false,"context_lines":[{"line_number":58,"context_line":"Third Party CI Voting"},{"line_number":59,"context_line":"---------------------"},{"line_number":60,"context_line":""},{"line_number":61,"context_line":"The Neutron team does not allow you to vote -1 with a third-party CI. False"},{"line_number":62,"context_line":"negatives are noisy to the community, and have given -1 from third-party"},{"line_number":63,"context_line":"CIs a bad reputation. Really bad, to the point of people ignoring them all."},{"line_number":64,"context_line":"Failure messages are useful to those doing refactors, and provide you"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3a50d1a3_0592f84b","line":61,"updated":"2015-07-30 00:17:46.000000000","message":"I think this is the wrong thing to do. At this point, it begs the question: why do we even have 3rd party CI posting results at all? We might as well scrap the whole thing and call it \u0027disaster\u0027.","commit_id":"d32d3ca000da4af9d1005015be3772f0241e44a7"},{"author":{"_account_id":7448,"name":"Carl Baldwin","email":"carl@ecbaldwin.net","username":"carl-baldwin"},"change_message_id":"f3a05c411f930145c3e30b2df21070726680a62e","unresolved":false,"context_lines":[{"line_number":58,"context_line":"Third Party CI Voting"},{"line_number":59,"context_line":"---------------------"},{"line_number":60,"context_line":""},{"line_number":61,"context_line":"The Neutron team does not allow you to vote -1 with a third-party CI. False"},{"line_number":62,"context_line":"negatives are noisy to the community, and have given -1 from third-party"},{"line_number":63,"context_line":"CIs a bad reputation. Really bad, to the point of people ignoring them all."},{"line_number":64,"context_line":"Failure messages are useful to those doing refactors, and provide you"}],"source_content_type":"text/x-rst","patch_set":1,"id":"1a4dcd0f_4ceed48c","line":61,"in_reply_to":"1a4dcd0f_472fa1a1","updated":"2015-08-03 23:34:01.000000000","message":"I tend to think that stripping the ability to -1 from a third party CI is a little premature.  It could make it even more difficult for those that do care what their CI is voting to find patches that might be causing problems.  For example, if I had a CI with -1 voting, I\u0027d probably want a gerrit query for patches where it did so.  If there is a small number of patches with -1, then I can focus on those patches and figure out what is going wrong.  If all or many patches show up in my query, then I know something is wrong.\n\nI think a -1 from a CI should be interpreted differently than \"I\u0027d prefer you not merge this\" and maybe we should figure out what it does mean and find a better way to convey it before removing this.\n\nPossible ways to interpret a -1 from a CI:\n\n- this patch invalidates a bad assumption made by my plugin which I should fix\n- this patch really breaks something and my CI is the only one that caught it\n- my CI failed to setup up or run the test properly\n- my CI is angry and votes -1 everywhere\n- my CI is totally broken and will be fixed tomorrow\n- my CI is totally broken and won\u0027t be fixed for some time\n- other failure modes I didn\u0027t think of\n\nI don\u0027t think we can require a CI system to be able to automatically determine which of these happened.  Though I do tend to agree with Armando that we could get better at doing this in the common cases.\n\nI suppose each CI owner could get information on job failures on their own without voting -1 on patches but wouldn\u0027t it be nice to have a place where everyone could see these results in case the results do have bearing on the viability of the patch?\n\nI also tend to agree with Doug that we should take away -1 rights for CI systems that are not maintained to a standard.  Is there a gerrit or git notes query we can craft to show how a CI is doing?","commit_id":"d32d3ca000da4af9d1005015be3772f0241e44a7"},{"author":{"_account_id":14287,"name":"Waldemar Znoinski","email":"wznoinsk@redhat.com","username":"wznoinsk"},"change_message_id":"407b99f286808de6a21fa5f7fee47ec426832e43","unresolved":false,"context_lines":[{"line_number":58,"context_line":"Third Party CI Voting"},{"line_number":59,"context_line":"---------------------"},{"line_number":60,"context_line":""},{"line_number":61,"context_line":"The Neutron team does not allow you to vote -1 with a third-party CI. False"},{"line_number":62,"context_line":"negatives are noisy to the community, and have given -1 from third-party"},{"line_number":63,"context_line":"CIs a bad reputation. Really bad, to the point of people ignoring them all."},{"line_number":64,"context_line":"Failure messages are useful to those doing refactors, and provide you"}],"source_content_type":"text/x-rst","patch_set":1,"id":"1a4dcd0f_754ec3c2","line":61,"in_reply_to":"1a4dcd0f_4ceed48c","updated":"2015-08-10 16:23:17.000000000","message":"The situation of loosening the governance (BigT) and saying we\u0027ll care less about our testing (3rd party CI that cover what upstream Jenkins doesn\u0027t) is not a good way to go (long-run) in my opinion - at least not for a production ready product. Instead Openstack Foundation should either provide infrastructure to test more configurations (and all of these CIs would have voting permission) or rely and support the community to strenghten the 3rdparty testing land and promote these that adhere to rules.\n\nI do understand comments of a good 3rdparty CI are most valuable to the vendor of the platform/driver and the developer of tested code so there\u0027s not much of a value to Cores as it used to be. Without standardization of 3rdparty testing there will be less trust in Openstack/project and competition is not a sleep. It\u0027s somehow striking to me having \u0027free\u0027 3rdparty resources (nodes + ci operators) neutron cores doesn\u0027t want to leverage these for the Openstack benefit by standardizing the testing framework - or at least not going the opposite way.\n\nCarl - I think not yet official is patrickeast\u0027s dashboard - check with Infra/anteaya/patrickeast (freenode) that gives you an eye-balling ability to compare different CI systems. \n\n\nOn the commenting/voting difference to me became clear only votes are looked at to the developers and I would like a 3rd party CI (our intel-networking-ci or any other) to be able to show we follow the rules and test good hence we can vote so it benefits the developer improving their code (if CI fails) and Intel being recognized as doing a good job and caring about testing. I think comments are too much noise nowadays (too many 3rdparty flowers bloomed) (good to have there in case someone cares) but voting is the actual indication we which CIs are doing good job at testing.\n\nI think we should focus on standardizing CI testing (again downstream puppet could be a place to put logic of error/exception catching?) and monitoring it in an automated way. Is that something to be investigated by infra/qa teams? \n\nOn the other hand - I don\u0027t see any benefit of taking down only -1 rights, false positive can be as much (if not more) damaging as false negative if we decide to go that route.","commit_id":"d32d3ca000da4af9d1005015be3772f0241e44a7"},{"author":{"_account_id":5263,"name":"Jeremy Stanley","display_name":"fungi","email":"fungi@yuggoth.org","username":"fungi","status":"missing, presumed fed"},"change_message_id":"0827f5a7e37c02bb7ae9c6e20d59db5e0f2af99c","unresolved":false,"context_lines":[{"line_number":58,"context_line":"Third Party CI Voting"},{"line_number":59,"context_line":"---------------------"},{"line_number":60,"context_line":""},{"line_number":61,"context_line":"The Neutron team does not allow you to vote -1 with a third-party CI. False"},{"line_number":62,"context_line":"negatives are noisy to the community, and have given -1 from third-party"},{"line_number":63,"context_line":"CIs a bad reputation. Really bad, to the point of people ignoring them all."},{"line_number":64,"context_line":"Failure messages are useful to those doing refactors, and provide you"}],"source_content_type":"text/x-rst","patch_set":1,"id":"fa1b9901_9635c8f8","line":61,"in_reply_to":"1a4dcd0f_754ec3c2","updated":"2015-08-22 20:13:36.000000000","message":"The \"OpenStack Foundation\" is you, is me, is all of us. The only way the OpenStack Foundation \"provides\" infrastructure is by asking the board members (Intel and others) to make it available to our community.","commit_id":"d32d3ca000da4af9d1005015be3772f0241e44a7"},{"author":{"_account_id":6524,"name":"Henry Gessau","email":"HenryG@gessau.net","username":"gessau"},"change_message_id":"707c0f47c9c3e27e1e42b8aae9c34444f7701f9e","unresolved":false,"context_lines":[{"line_number":58,"context_line":"Third Party CI Voting"},{"line_number":59,"context_line":"---------------------"},{"line_number":60,"context_line":""},{"line_number":61,"context_line":"The Neutron team does not allow you to vote -1 with a third-party CI. False"},{"line_number":62,"context_line":"negatives are noisy to the community, and have given -1 from third-party"},{"line_number":63,"context_line":"CIs a bad reputation. Really bad, to the point of people ignoring them all."},{"line_number":64,"context_line":"Failure messages are useful to those doing refactors, and provide you"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3a50d1a3_f08c6830","line":61,"in_reply_to":"3a50d1a3_0592f84b","updated":"2015-07-30 00:40:45.000000000","message":"I don\u0027t see this as a disaster at all. I took this change to mean that we want CIs to post results but not vote. Maybe the wording could be more clear on that?","commit_id":"d32d3ca000da4af9d1005015be3772f0241e44a7"},{"author":{"_account_id":6635,"name":"John Davidge","email":"john.davidge@rackspace.com","username":"john-davidge"},"change_message_id":"e44a8e14b52d0151c7d31af8ad659697472381c6","unresolved":false,"context_lines":[{"line_number":58,"context_line":"Third Party CI Voting"},{"line_number":59,"context_line":"---------------------"},{"line_number":60,"context_line":""},{"line_number":61,"context_line":"The Neutron team does not allow you to vote -1 with a third-party CI. False"},{"line_number":62,"context_line":"negatives are noisy to the community, and have given -1 from third-party"},{"line_number":63,"context_line":"CIs a bad reputation. Really bad, to the point of people ignoring them all."},{"line_number":64,"context_line":"Failure messages are useful to those doing refactors, and provide you"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3a50d1a3_9875f8fb","line":61,"in_reply_to":"3a50d1a3_0592f84b","updated":"2015-07-30 11:37:31.000000000","message":"Posting results is absolutely helpful in some circumstances. It\u0027s only the negative votes that we want to remove as they aren\u0027t helpful a lot of the time.","commit_id":"d32d3ca000da4af9d1005015be3772f0241e44a7"},{"author":{"_account_id":10980,"name":"Doug Wiegley","email":"dougwig@parkside.io","username":"dougw"},"change_message_id":"3f1024f7fcdb31cebb10056596d448ecba82eff2","unresolved":false,"context_lines":[{"line_number":58,"context_line":"Third Party CI Voting"},{"line_number":59,"context_line":"---------------------"},{"line_number":60,"context_line":""},{"line_number":61,"context_line":"The Neutron team does not allow you to vote -1 with a third-party CI. False"},{"line_number":62,"context_line":"negatives are noisy to the community, and have given -1 from third-party"},{"line_number":63,"context_line":"CIs a bad reputation. Really bad, to the point of people ignoring them all."},{"line_number":64,"context_line":"Failure messages are useful to those doing refactors, and provide you"}],"source_content_type":"text/x-rst","patch_set":1,"id":"1a4dcd0f_472fa1a1","line":61,"in_reply_to":"3a50d1a3_9875f8fb","updated":"2015-08-03 16:18:00.000000000","message":"So, I\u0027m not in favor of this change, even though I run a CI which does NOT vote -1. To me, it\u0027s a bit of information that says, \"you should really really trust this negative result\", and if someone is confident enough in their system enough to actual trigger a -1 in gerrit, then I pay slightly more attention to it than I would for a FAILED result that doesn\u0027t vote.\n\nOn the flipside, however, if we provide this rope, we should be absolutely BRUTAL in taking the rights away from any CI that enables -1 voting and has a history of false negatives.","commit_id":"d32d3ca000da4af9d1005015be3772f0241e44a7"},{"author":{"_account_id":14287,"name":"Waldemar Znoinski","email":"wznoinsk@redhat.com","username":"wznoinsk"},"change_message_id":"982f7bae3a8f37d194291a28fae7e40420617825","unresolved":false,"context_lines":[{"line_number":58,"context_line":"Third Party CI Voting"},{"line_number":59,"context_line":"---------------------"},{"line_number":60,"context_line":""},{"line_number":61,"context_line":"The Neutron team does not allow you to vote -1 with a third-party CI. False"},{"line_number":62,"context_line":"negatives are noisy to the community, and have given -1 from third-party"},{"line_number":63,"context_line":"CIs a bad reputation. Really bad, to the point of people ignoring them all."},{"line_number":64,"context_line":"Failure messages are useful to those doing refactors, and provide you"}],"source_content_type":"text/x-rst","patch_set":1,"id":"da20952f_a507ce80","line":61,"in_reply_to":"fa1b9901_9635c8f8","updated":"2015-09-01 12:36:36.000000000","message":"I agree, there has to be the decision made by the OS governance/project to ask for such resources though. That would be the indication OS intends to provide more resources to cover more testing. On the other hand there are 3rd party CIs - (free to OS) resources for testing.","commit_id":"d32d3ca000da4af9d1005015be3772f0241e44a7"}]}
