On Tue, Dec 03, 2019 at 10:20:51AM -0700, Simon Glass wrote: > Hi Tom, > > On Tue, 3 Dec 2019 at 10:07, Tom Rini <tr...@konsulko.com> wrote: > > > > On Tue, Dec 03, 2019 at 06:00:11PM +0100, Marek Vasut wrote: > > > On 12/3/19 5:52 PM, Simon Glass wrote: > > > > Hi Tom, > > > > > > > > On Mon, 2 Dec 2019 at 20:11, Tom Rini <tr...@konsulko.com> wrote: > > > >> > > > >> Hey all, > > > >> > > > >> It's release day and here is v2020.01-rc4. Yes, I'm still working on > > > >> fixing all of the issues that pop up as I get the MTD clean-up series > > > >> ready to go. In fact, what I need to do at this point is grab the > > > >> handful of size reduction patches that this has shown are worthwhile, > > > >> then I can do the MTD series. Then we're down to just fixing up > > > >> misconversions where things got turned off. > > > >> > > > >> Once again, for a changelog, > > > >> git log --merges v2020.01-rc3..v2020.01-rc4 > > > >> and as always, I ask for more details in the PRs people send me so I > > > >> can > > > >> put them in the merge commit. > > > >> > > > >> I'm planning on doing -rc5 on December 23rd with the release scheduled > > > >> on January 6th. Thanks all! > > > > > > > > Speaking of next year, what is the plan for the release. Is there any > > > > chance we might move back to a release every two months? The > > > > three-month process is very slow... > > > > > > I am very happy with this 3-month release cycle, it's less stressful and > > > I think the quality of the releases is higher too. > > > > I thought I said this a release or so ago, sorry. I believe we'll be > > sticking with the 3-month cycle and I hope to find time to more actively > > use a -next branch myself to help with some of the delay, or at least > > slower feedback cycle. > > That would certainly help, but it hasn't proved possible so far. Let's > see how it goes with the next release.
Well, it's also about custodian stress and time levels. The longer cycle seems to be better for that. > Do you think if we can improve the testing (e.g. with more mini-labs > attached to gitlab) we might resolve these stability problems? Ah right, I needed to reply to that thread, sorry. I get the feeling that the answer overall is that folks that have spent time with LAVA need to chime in, or barring that, I need to play with LAVA as that's how a lot of the "how do you reserve boards" and so forth problems are handled, and would also make an easier case for vendors to get more U-Boot testing done (as that gets you to kernelci too, or if you already have kernelci going, you can add U-Boot with just a few more steps). -- Tom
signature.asc
Description: PGP signature