2020-09-09 04:17:26 UTC - Dominic Kim: hm.. I think this kind of change can be made at any time. It is also worth handling them in a backward-compatible way and #4970 looks reasonable, but in some sense, if we add this kind of code whenever messages change and any changes are applied I feel like codes would include more and more skeleton codes and be complicated. And it's not even possible in some cases. I am inclined to handle them differently. For example, we could handle such a mismatch by introducing another instance of Kafka. https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1599625046042300?thread_ts=1599603164.041400&cid=C3TPCAQG1 ---- 2020-09-09 04:26:24 UTC - Brendan Doyle: It’s not about always supporting every message type. You can deprecate old message versions and remove the code that supports them after some time, but you need to provide a path forward for people to upgrade or warn them that this is a major upgrade if it truly can’t be handled. It’s the whole purpose for semver so people can track how upgrades are going to affect them. Just introducing a completely non-backwards compatible minor serialization change that can be handled easily through options in my opinion is bad. https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1599625584047300?thread_ts=1599603164.041400&cid=C3TPCAQG1 ---- 2020-09-09 04:46:36 UTC - Dominic Kim: Yes we have been working in such a way. Whenever any breaking changes had introduced, we've shared them in dev list and let users know. We've provided ways to selectivley enable new features and introduced new features based on SPI as much as we can, etc. I am also not saying it's enough to introduce just non-backwards compatible change and let it done. But in some cases, (surely this time is not the case) it would introduce more complication and IMHO the upstream should evolve without being dragged by the downstream. We regularly release codes and breaking changes between each release can happen I suppose. This becomes an issue when downstream rebases codes whenever they need rather than taking released version. And I feel we lack such official processes to deprecate a certain version and remove them entirely from the code with some migration time for downstreams. It would be great to have it. https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1599626796047500?thread_ts=1599603164.041400&cid=C3TPCAQG1 ---- 2020-09-09 13:04:05 UTC - Rodric Rabbah: We should also cut more active releases :wink: This would allow transitions or breaking changes to be better communicated and allow operators to use the version they need to while they transition to new version. +1 : Brendan Doyle https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1599656645047700?thread_ts=1599603164.041400&cid=C3TPCAQG1 ---- 2020-09-09 16:53:14 UTC - Raghuveer Raavi: Hi team, I opened a draft <https://github.com/apache/openwhisk/pull/4972|PR> addressing #4798 issue.. I appreciate feedback on it.
side note: I’m looking to work on <https://github.com/apache/openwhisk/issues/4523|this> issue next! https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1599670394050100 ---- 2020-09-09 17:12:08 UTC - Brendan Doyle: ^ exactly what I was getting at. We need a more active release versioning process. Openwhisk is a large project now, not everyone using it might be following the dev list closely for small breaking changes https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1599671528050200?thread_ts=1599603164.041400&cid=C3TPCAQG1 ---- 2020-09-09 17:32:17 UTC - Rodric Rabbah: we are overdue for a release @Dominic Kim did you still want to be release manager for ow 1.0? https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1599672737050500?thread_ts=1599603164.041400&cid=C3TPCAQG1 ---- 2020-09-09 17:56:39 UTC - Tyson Norris: I’m still working on build issues with ContainerProxy, but in the meantime, I just started getting scala 2.13 errors from travis builds - did something change there? I will look into it more… https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1599674199051500?thread_ts=1599674199.051500&cid=C3TPCAQG1 ---- 2020-09-09 17:56:52 UTC - Tyson Norris: ```232.18s$ TERM=dumb OW_SCALA_VERSION=2.13 ./gradlew :tests:compileTestScala Build using Scala 2.13``` https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1599674212051600?thread_ts=1599674199.051500&cid=C3TPCAQG1 ---- 2020-09-09 18:11:51 UTC - Rodric Rabbah: nothing of note in the commits https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1599675111051800?thread_ts=1599674199.051500&cid=C3TPCAQG1 ---- 2020-09-09 18:11:59 UTC - Rodric Rabbah: should we revert the container proxy pr? https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1599675119052000?thread_ts=1599674199.051500&cid=C3TPCAQG1 ---- 2020-09-09 18:13:39 UTC - Rodric Rabbah: > nd I feel we lack such official processes to deprecate a certain version and remove them entirely from the code with some migration time for downstreams. It would be great to have it. to really do this we need to publish more releases - then downstream is really picking up releases instead of commits +1 : Brendan Doyle https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1599675219052200?thread_ts=1599603164.041400&cid=C3TPCAQG1 ---- 2020-09-09 18:13:58 UTC - Rodric Rabbah: @Dominic Kim are you ok with merging Brendan’s PR? https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1599675238052400?thread_ts=1599603164.041400&cid=C3TPCAQG1 ---- 2020-09-09 18:15:46 UTC - Rodric Rabbah: @Brendan Doyle please send a note to the Dev list about your PR. :pray: not everyone is on slack. https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1599675346052700?thread_ts=1599603164.041400&cid=C3TPCAQG1 ---- 2020-09-09 18:16:26 UTC - Brendan Doyle: will do when I get a chance thankyou : Rodric Rabbah https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1599675386052900?thread_ts=1599603164.041400&cid=C3TPCAQG1 ---- 2020-09-09 18:24:57 UTC - Rodric Rabbah: welcome @nitika https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1599675897053500 ---- 2020-09-09 18:28:33 UTC - Tyson Norris: let me see if I can find the reason for `OW_SCALA_VERSION` and get my latest PR tested; if that doesn’t work then yes let’s revert https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1599676113053600?thread_ts=1599674199.051500&cid=C3TPCAQG1 ---- 2020-09-09 18:29:55 UTC - Tyson Norris: oh I guess there has been a separate job for scala 2.13 for a while… https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1599676195053800?thread_ts=1599674199.051500&cid=C3TPCAQG1 ---- 2020-09-09 18:33:25 UTC - Tyson Norris: will see if restarting those jobs does anything… https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1599676405054000?thread_ts=1599674199.051500&cid=C3TPCAQG1 ---- 2020-09-09 18:35:01 UTC - Rodric Rabbah: @Markus Thömmes did that - so we don’t regress on being able to compile for 2.13 i think he shelved for a while the actual cut over from .12 to .13 https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1599676501054200?thread_ts=1599674199.051500&cid=C3TPCAQG1 ---- 2020-09-09 18:37:43 UTC - Markus Thömmes: yeah, I don't have a local system anymore so I wasn't able to debug why 2.13 based seemingly doesn't deploy :smile: https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1599676663054400?thread_ts=1599674199.051500&cid=C3TPCAQG1 ---- 2020-09-09 19:27:58 UTC - Brendan Doyle: "to really do this we need to publish more releases - then downstream is really picking up releases instead of commits" Big +1 on this. Having the releases allows the operator determine if they can upgrade just by looking at the version or if they can't, what they may need to do to upgrade in the release notes. And it adds a level of accountability for tracking what breaking changes are being made. https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1599679678054600?thread_ts=1599603164.041400&cid=C3TPCAQG1 ---- 2020-09-09 19:31:03 UTC - Rodric Rabbah: I believe we’re at least ready to cut a new 1.0 release. @Dave Grove can you confirm? I’ll wait for Dominic - he had wanted to do this. If he can’t one of the other committers will do it. https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1599679863054800?thread_ts=1599603164.041400&cid=C3TPCAQG1 ---- 2020-09-09 19:33:13 UTC - Dave Grove: was contemplating try to roll the Go 1.13/1.15 through all of the runtimes, but honestly I think we can proceed in parallel. I suspect we won’t be happy with the first release candidate for 1.0 core anyways and will need to iteratre https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1599679993055000?thread_ts=1599603164.041400&cid=C3TPCAQG1 ---- 2020-09-09 19:36:21 UTC - Brendan Doyle: Would you all be open to a release cadence proposal as well? I think that could be interesting https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1599680181055200?thread_ts=1599603164.041400&cid=C3TPCAQG1 ---- 2020-09-09 19:36:41 UTC - Brendan Doyle: at least for the main repo https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1599680201055400?thread_ts=1599603164.041400&cid=C3TPCAQG1 ---- 2020-09-09 19:40:41 UTC - Dave Grove: my 2 cents is that it would be nice to have a cadence, but we’d need people with the time to step up to actually make it happen. https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1599680441055600?thread_ts=1599603164.041400&cid=C3TPCAQG1 ---- 2020-09-09 19:43:10 UTC - Brendan Doyle: true I definitely agree there. I've never done release management for a project of this magnitude so I don't know how much work that will entail for people https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1599680590055800?thread_ts=1599603164.041400&cid=C3TPCAQG1 ---- 2020-09-09 20:10:40 UTC - Tyson Norris: oh looks like my latest change was simply not 2.13 compatible ( `Unit` not allowed, use `()`) :disappointed: will try another build… https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1599682240056000?thread_ts=1599674199.051500&cid=C3TPCAQG1 ---- 2020-09-09 20:11:00 UTC - Rodric Rabbah: ha https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1599682260056200?thread_ts=1599674199.051500&cid=C3TPCAQG1 ---- 2020-09-09 21:09:15 UTC - Tyson Norris: ok <https://github.com/apache/openwhisk/pull/4971> passed; @Markus Thömmes when you’re up, if you can take a look, or @Rodric Rabbah if you want to check it out https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1599685755056400?thread_ts=1599674199.051500&cid=C3TPCAQG1 ----
