Hi Julian, I didn't read this proposal as being in tension with apache releases. It sounds like the intention is to hold a vote every two weeks to verify a release artifacts? But maybe I misread or missed something. Were do you think the tension lies? Is it also producing the signed source artifact?
Since votes last for at least 72 hours this does seem like a lot of overhead every two weeks, but it seems that is something for Rust maintainers to decide and adjust. -Micah On Saturday, May 1, 2021, Julian Hyde <[email protected]> wrote: > (Removing user@ from cc. I think this is mainly a dev@ issue.) > > I believe there are some tensions between this process and the Apache > process. In particular, Apache releases tend to be a signed source > distribution (tarball) that at least three PMC members download and > verify. I totally understand why, as Rust developers, you might find > that an onerous process and might want to operate in a different way. > It makes sense, and I believe we can solve it. Perhaps by using a word > other than "release" for high-frequency snapshots. > > It is likely that other projects have already run into this problem > and have solved it. Therefore I have sent an email to comdev asking > for advice [1]. Feel free to join the thread. > > Julian > > [1] https://lists.apache.org/thread.html/rf12538ef0f60f7257e63391e5d496 > 2a6156564020c99d3dfb193f4d7%40%3Cdev.community.apache.org%3E > > On Sat, May 1, 2021 at 4:01 AM Andrew Lamb <[email protected]> wrote: > > > > I propose regularly releasing, every 2 weeks, minor and patch releases > of > > the arrow-rs crate, following the semver versioning scheme used by the > rest > > of the Rust ecosystem. I have written a proposal[1] describing how this > > might work. > > > > Feedback and comments most welcome. > > > > Andrew > > > > [1] > > https://docs.google.com/document/d/1QTGah5dkRG0Z6Gny_ > QCHmqMg7L2HmcbEpRISsfNEhSA/edit >
