Re: Independent JS patch releases

2021-11-18 Thread Dominik Moritz
I only want to have independent patch releases but still keep the main releases aligned with the other releases. I want to make sure I have the process right. When we want to make a patch release, do we still need to call a vote? When we actually want to make a release, what do we do to bump the

Re: Independent JS patch releases

2021-11-14 Thread Wes McKinney
I think the idea would be to make a source tarball out of the JS directory — we had a script for this a few years ago from the last time we made independent JS releases — if you want to resurrect this and modify it for the current state of the project, this seems fine. It would also be helpful to h

Re: Independent JS patch releases

2021-11-12 Thread Dominik Moritz
Thank you, Wes. Could you point me to what preparations we need to make now and how we can make a patch release in the future? On Nov 10, 2021 at 17:34:44, Wes McKinney wrote: > I don't see a problem with making JS-only patch releases. There's some > work to facilitate the release management b

Re: Independent JS patch releases

2021-11-10 Thread Wes McKinney
I don't see a problem with making JS-only patch releases. There's some work to facilitate the release management but if it's a source-only release it shouldn't be _that_ difficult. On Fri, Nov 5, 2021 at 9:50 AM Dominik Moritz wrote: > > Hi Neal, > > Great questions. We could potentially add an

Re: Independent JS patch releases

2021-11-05 Thread Dominik Moritz
Hi Neal, Great questions. We could potentially add an integration test for major bundlers. However, then we would also need a way to test these packages in different environments like browsers and that’s a lot of work that I’m not sure will be proportional to the benefit. The issue in my experien

Re: Independent JS patch releases

2021-11-05 Thread Neal Richardson
Not expressing an opinion on the original question, but if the problem is "not getting everything right the first time", what can be done to reduce the likelihood of getting things wrong? Other languages/implementations have extensive CI and nightly builds, some of which test different packaging sc