Re: Upcoming stable point release (9.8)
On Fri, 2019-02-01 at 03:20 +, Ben Hutchings wrote: > On Thu, 2019-01-31 at 20:15 +, Adam D. Barratt wrote: > > Hi, > > > > The next point release for "stretch" (9.8) is scheduled for Saturday, > > February 16th. Processing of new uploads into stretch-proposed-updates > > will be frozen during the preceding weekend. > > We currently have version 4.9.144-2 in proposed-updates while the > stretch branch in git is based on 4.9.150. > > There is a regression in ceph, introduced in 4.9.144 and fixed in > 4.9.150 (commit 90b7f9481243). We should fix this before the point > release. > > There is another regression in spinlocks on x86, introduced in 4.9.147 > and fixed in 4.9.153 (commit 4b527f25a4ac). We should fix this if we > update to 4.9.150. > > As there isn't much time left before the point release, I am inclined > to make one more upload based on 4.9.144 that just fixes the ceph > regression and any other known recent regressions. I've now done this. Ben. > Does anyone want to argue for updating further? > > Does anyone know of other regressions between 4.9.130 and 4.9.144? -- Ben Hutchings Knowledge is power. France is bacon. signature.asc Description: This is a digitally signed message part
Re: Upcoming stable point release (9.8)
Ben Hutchings schrieb: > As there isn't much time left before the point release, I am inclined > to make one more upload based on 4.9.144 that just fixes the ceph > regression and any other known recent regressions. +1 And FWIW, I've tested 4.9.144-1 on a number of systems and didn't run into any regressions (not using ceph). Cheers, Moritz
Re: Upcoming stable point release (9.8)
On Fri, 2019-02-01 at 10:42 +0100, Yves-Alexis Perez wrote: > On Fri, 2019-02-01 at 03:20 +, Ben Hutchings wrote: > > As there isn't much time left before the point release, I am inclined > > to make one more upload based on 4.9.144 that just fixes the ceph > > regression and any other known recent regressions. > > > > Does anyone want to argue for updating further? > > Hi, > > I'm usually in favor of updating, but here we have a large delta between > proposed-updates and current (.150/.153). There are 319 commits between .144 > and .150 (and 431 for .153) which would be nice to have since they likely fix > quite some stuff (although they might as well introduce regression and we > don't have a lot of time to test it). > > > If we go for .144 + ceph fix, how do we handle the git part (since stretch is > already at .150)? We create a stretch-9.8 branch or something? I don't think the d-k-prerelease and d-k-tag scripts will accept that. What I did for 4.9.144-2 was to move my local stretch branch around: - Check out the stretch branch - Check that I have no commits to push - Reset to the new branch point: git reset --hard debian/4.9.144-1 - Commit my changes and tag - Reset to upstream: git reset --hard salsa/stretch - Merge the new tag: git merge debian/4.9.144-2 > In any case, I'd say we try to upload .153 (or .154) to proposed-updates just > after the point release. Yes, absolutely. Ben. -- Ben Hutchings Life is what happens to you while you're busy making other plans. - John Lennon signature.asc Description: This is a digitally signed message part
Re: Upcoming stable point release (9.8)
-BEGIN PGP SIGNED MESSAGE- Hash: SHA256 On Fri, 2019-02-01 at 03:20 +, Ben Hutchings wrote: > As there isn't much time left before the point release, I am inclined > to make one more upload based on 4.9.144 that just fixes the ceph > regression and any other known recent regressions. > > Does anyone want to argue for updating further? Hi, I'm usually in favor of updating, but here we have a large delta between proposed-updates and current (.150/.153). There are 319 commits between .144 and .150 (and 431 for .153) which would be nice to have since they likely fix quite some stuff (although they might as well introduce regression and we don't have a lot of time to test it). If we go for .144 + ceph fix, how do we handle the git part (since stretch is already at .150)? We create a stretch-9.8 branch or something? In any case, I'd say we try to upload .153 (or .154) to proposed-updates just after the point release. Regards, - -- Yves-Alexis -BEGIN PGP SIGNATURE- iQEzBAEBCAAdFiEE8vi34Qgfo83x35gF3rYcyPpXRFsFAlxUFJIACgkQ3rYcyPpX RFswvggAxEpFBnG68D2yTkNRlkwPVcmT+YfAurt5Z7ghCFNou7EfXU/YV8Q94qUd 4nmaZJIovv5qIWRK8k1nFkvpV+9R9lZb02K/n/FB1L/KBHjXOQWx3gj50vXiwAoR oQvlhv/fXi6WLTqxfeV+bxDpWpUxlmGWV1ZcuAZs8P7aXEdrcjyes6IoCXcCHkby 8hbMHkeW7LX26f105qFS9XmKK7bNogiP/LsGdrsgAVFY4ghxkpZ+wwEcpiYl8F3c cpo0RNHhcqffyHziSmDZMd0MUSdUfMy33d3q9FjQKzZndjVBm+dlO6WnAsDgs7xd 3Gxrx1BwHrbwkEIuetCDkpC42xDmRg== =pZ73 -END PGP SIGNATURE-
Re: Upcoming stable point release (9.8)
On Thu, 2019-01-31 at 20:15 +, Adam D. Barratt wrote: > Hi, > > The next point release for "stretch" (9.8) is scheduled for Saturday, > February 16th. Processing of new uploads into stretch-proposed-updates > will be frozen during the preceding weekend. We currently have version 4.9.144-2 in proposed-updates while the stretch branch in git is based on 4.9.150. There is a regression in ceph, introduced in 4.9.144 and fixed in 4.9.150 (commit 90b7f9481243). We should fix this before the point release. There is another regression in spinlocks on x86, introduced in 4.9.147 and fixed in 4.9.153 (commit 4b527f25a4ac). We should fix this if we update to 4.9.150. As there isn't much time left before the point release, I am inclined to make one more upload based on 4.9.144 that just fixes the ceph regression and any other known recent regressions. Does anyone want to argue for updating further? Does anyone know of other regressions between 4.9.130 and 4.9.144? Ben. -- Ben Hutchings Life is what happens to you while you're busy making other plans. - John Lennon signature.asc Description: This is a digitally signed message part
Upcoming stable point release (9.8)
Hi, The next point release for "stretch" (9.8) is scheduled for Saturday, February 16th. Processing of new uploads into stretch-proposed-updates will be frozen during the preceding weekend. Regards, Adam
