You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Do you think it would be possible to tag a new release according to semver? Some go module based dependencies have to rewrite their go.mod to get the latest features in. 馃槆 Thanks in advance! :)
The main thing I'd like to see be ready before a new runtime-spec release is the resolution of #1008 -- which will allow runc to have a 1.0 release. (Tagging releases, especially spec releases, in OCI is a bit of a hassle.)
#1008 has been merged a week ago. Is all that's necessary a PR with a tag? I'm guessing this is better left to the maintainers, but according to several comments all over the place, runtime-spec 1.0.2 should be ready?
#1026 failed due to not meeting the 2/3 quorum of maintainers; since then, two inactive maintainers have stepped down, meaning the quorum could now be reached with the LGTMs from the last vote alone. Since a nit** was brought up by @kad on the vote thread anyway, I thought it'd be good to double-check the state of open PRs - not to further delay the release, but to pick up already finished changes, since runtime-spec releases are few and far between. In descending order of approval:
@kad: There is one questionable statement introduced by PRs #988 and #932 about Intel RDT closID state, which in theory should require small clarification:
One case that is not covered in the description is
"If closID is set and neither l3CacheSchema nor memBwSchema is set".
In this case, I think, runtime should check if closID exists in the resctrl and assign container to it.
If this closID does not exist in resctl - error.
Only two PRs left that could realistically still squeeze into 1.0.2. If anyone except @tianon (who already LGTM'd) could look at #1034#1035, there'd be nothing left but to actually release it. :) @crosbymichael@mrunalp@vbatts@dqminh@hqhq
I don't want to presume that a release-PR from a non-maintainer would be welcome, but in case it is, I have a branch ready that updates #1026 with the changes that have happened since then.
Activity
cyphar commentedon May 15, 2019
The main thing I'd like to see be ready before a new runtime-spec release is the resolution of #1008 -- which will allow runc to have a 1.0 release. (Tagging releases, especially spec releases, in OCI is a bit of a hassle.)
h-vetinari commentedon Jul 18, 2019
It seems the resolution of #1008 is held up mainly by waiting for someone to review the PR. ;-)
onlyjob commentedon Sep 11, 2019
Formal release is long overdue... Because of that some projects (namely runtime-tools) already deviated from good practice of vendoring by tag...
vbatts commentedon Dec 17, 2019
to echo @h-vetinari, we are waiting on #1008 for a next release of the spec, and of runc-1.0.0 馃槵
h-vetinari commentedon Jan 8, 2020
#1008 has been merged a week ago. Is all that's necessary a PR with a tag? I'm guessing this is better left to the maintainers, but according to several comments all over the place,
runtime-spec 1.0.2should be ready?vbatts commentedon Jan 9, 2020
vote for release is posted
h-vetinari commentedon Feb 3, 2020
#1026 failed due to not meeting the 2/3 quorum of maintainers; since then, two inactive maintainers have stepped down, meaning the quorum could now be reached with the LGTMs from the last vote alone. Since a nit** was brought up by @kad on the vote thread anyway, I thought it'd be good to double-check the state of open PRs - not to further delay the release, but to pick up already finished changes, since runtime-spec releases are few and far between. In descending order of approval:
config-linux: add SHOULD to linux.namespaces.type聽#1025; 2 LGTMs(done)implementations: add railcar聽#1003; 2 LGTMs(obsolete)config-windows: Added support for network namespace聽#978; 1 LGTM(obsolete)config-linux: describe more about rootfs mount propagation聽#975; 1 LGTM(done in config-linux: describe more about rootfs mount propagation (cont'd)聽#1035)remove unneeded indent聽#937; 1 LGTM(done in Rebase #937聽#1031)Fix typo in RELEASES.md聽#1033: 1 LGTM(done)Clarify case with pre-configured Intel RDT closID聽#1034: 1 LGTM (Fix for @kad's nit**)(done)For completeness, the remaining open PRs are probably not ready / necessary for a last-minute merge:
MUSTs: config: Do not allow runtimes to ignore properties defined by the spec聽#680, add RFC requirement for mountLabel聽#934, runtime.md: only required properties are MUST applied聽#935, add definite limit for username mappings聽#939, runtime: Document state annotations as a copy of config annotations聽#946, maskedPaths and readonlyPaths: skip non-existent paths聽#982RELEASE: document how to do the release聽#1027, PullApprove: No need for 3 DCO checks聽#1029**
vbatts commentedon Feb 4, 2020
Good list. Would want to try your hand at carrying over #937 ?
h-vetinari commentedon Mar 4, 2020
Updated the list.
h-vetinari commentedon Mar 4, 2020
All the open PRs that could still be considered for the next release are exceedingly trivial and could be finished if two devs took half an hour each.
In any case, now that two months have passed since the first quorum, it's really time to get this release done, IMHO.
h-vetinari commentedon Mar 7, 2020
Only two PRs left that could realistically still squeeze into
1.0.2. If anyone except @tianon (who already LGTM'd) could look at #1034 #1035, there'd be nothing left but to actually release it. :)@crosbymichael @mrunalp @vbatts @dqminh @hqhq
h-vetinari commentedon Mar 8, 2020
All done; thanks @vbatts! 馃帀
Any ETA on sending out the vote?
h-vetinari commentedon Mar 15, 2020
Another week, another ping (I'm sorry for apparently talking to myself here 馃槄)
h-vetinari commentedon Mar 20, 2020
I don't want to presume that a release-PR from a non-maintainer would be welcome, but in case it is, I have a branch ready that updates #1026 with the changes that have happened since then.
vbatts commentedon Mar 24, 2020
vote posted!