On Mon, 2019-03-18 at 10:30 -0400, Jonathan Rajotte-Julien wrote: > Note I did not cc the mailing list since > I think it is moving the discussion away from the main focus.
It did make the list and FWIW I think it is important to ensure people understand why we're dropping certain patches or proposing changing some of our guidelines... lttng wasn't the only reason I wanted to change this, you were however a timely example of the confusion the current wording causes as I do think lttng has a "stable" policy which fits with ours. > ----- Original Message ----- > > From: "Burton, Ross" <[email protected]> > > I don't really have a strong opinion on the mechanics behind > > documenting the behaviour, but how do we decide what upstreams have > > a > > suitable stable release? We don't want to trust every upstream > > that > > says they've a 'stable branch' until they can demonstrate that they > > are competent at the job. > > Yeah this is not an easy thing to demonstrate on both side of the > equation. > > I'll pitch in my "upstream" opinion for the lttng* projects. > > So far, most of the problem we had with yocto/OE were either custom > patches on top of very important projects (glibc [1], gcc [2]) > or custom patches applied on the lttng-* projects without any regards > to the "upstream" opinion [3] on the matter. > That is when the original contributor of the patch even bother > sending it to our mailing list or bug tracker. > > We are more then fine with distro changing the "DEFAULT" values but > changing code is something else. Its hard for us to tell which things are "upstream are working on it" or "upstream didn't respond" compared to "upstream rejected this because XXX". Sometimes we might even disagree with XXX (gnupg and myself have a long running disagreement about pkg-config). As you'll likely have gathered, I was horrified about that glibc patch and want it removed ASAP which we're working on even in the stable branches. The gcc issue was a while ago and we were likely mislead by a backport tag :(. Mistakes do happen, what I want to ensure is that when we find issues we do the right things. I do also place a lot of value on the upstream opinions. > This is were the mechanism that Richard proposed seems to fit. > > We are slowly looking more at how lttng* is packaged in the > embedded/lean distro world (yocto, alpine etc.) > and trying to help the package maintainer as best as we can so our > users can pick up yocto/whatever and have a good experience. I really appreciate this and want to try and help users have that good experience too! > I understand not all projects are the same. Back to the original topic, this is very hard for us to figure out and it does change over time as people's availability does vary amongst other factors. Mathieu Desnoyers and I have talked in a bit of detail about this too. Another metric is to understand what would happen if one of these stable updates broke things somehow. I think in the lttng case there would be interest and likely assistance in figuring out why breakage happened and how to resolve it. To me, that also is a good sign we're doing the right things. We'd not get that from all projects. Cheers, Richard _______________________________________________ Openembedded-architecture mailing list [email protected] http://lists.openembedded.org/mailman/listinfo/openembedded-architecture
