FWIW, I’m also willing to work on the following release (let’s say sometime in May?) so I can better understand everything.
I’d also like to speak to Alex on the phone before I come to an opinion here because I feel like some of the finer points have been getting lost in email. I appreciate all the work you and Chris have been doing to try to improve the release process. Thanks, Harbs > On Apr 1, 2020, at 12:50 PM, Harbs <[email protected]> wrote: > > Before we vote on something like this, I need to understand how it’ll work > and what the ramifications would be. Having two different release processes > sounds like it’ll make things worse instead of better. > > Life has been a bit hectic lately (as I imagine it has been for many), and I > don’t have clear in my head some of the finer points of the release process. > > I was hoping Yishay would do the next release which would give me an > opportunity to discuss with him how it went using the current process. (When > we can finally spend some face to face time…) ;-) > > I think we’re all in agreement that we want the release process to be easier, > but I’m not sure it’s totally clear to all of us how to best go about that. > > Please cancel the vote thread because I think you “jumped the gun” on that. > > My $0.02, > Harbs > >> On Apr 1, 2020, at 12:39 PM, Carlos Rovira <[email protected]> wrote: >> >> Hi All, >> >> after many days of going in circles here's a propossal thread as other PMCs >> requested and will be followed by a Vote thread. >> >> So the proposal is the following: >> >> Apache Royale needs a reliable, easy process to release often. We still >> don't have that, so we propose to release as the rest of Apache projects do >> [1] by using a tested Maven process [2]. >> >> There's as well another current way [3] that RMs can choose as well. So use >> [2] or [3] depending on what works for them better. >> >> About [2] we tested full release and the output did work in any tested IDE >> (Moonshine and VSCode). The source-bundles compiled without any issues in >> Ant and Maven and the Maven artifacts worked perfectly in a test project. >> So technically speaking seems ready to go (although as worked in real could >> expect some little changes). >> >> About [3], it was the system used to release 0.9.6, but trying to do 0.9.7 >> was not possible by some volunteers for that reason some of us want more >> options and propose [2]. >> >> But we want [3] continue to be an option for all volunteers that want to >> investigate and invest time on it. >> >> Main ponts: >> >> 1.- One process (taking about [2] and [3]) should not be over the other. >> 2.- It's up to the RM to choose one or the other, whatever he wants to do >> to get the release in best conditions and the work done faster, the better, >> and do the release with that choose. >> 3.- Release process can't block enhancements to build process, since more >> changes in the future are expected in build systems (even add more), so >> it's up to people choosing [2] or [3] invest the time to adapt to new >> changes, as any other thing in an ASF project "not to expect others do it >> for you, do yourself". >> >> The objective is not block any of the two process and releases doing with >> one or the other will be valid and could be voted as official Apache Royale >> releases. >> >> So will request your vote to any Apache Royale PMC, Commiter and user in >> this list. >> >> PMCs are binding votes. >> >> Thanks >> >> Carlos >> >> >> [1] https://twitter.com/ChristoferDutz/status/1240219399551934473 >> [2] >> https://github.com/apache/royale-asjs/wiki/Release-Manager-(already-tested) >> [3] https://github.com/apache/royale-asjs/wiki/Release-Manager-Notes >> >> -- >> Carlos Rovira >> http://about.me/carlosrovira >
