Thanks, I am ok with this.
But I am not sure I understand the overall sentiment. KIPs are supposed
to be completed by feature freeze, so if we want to give more time for
KIPs to get completed, we should have extended feature freeze deadling
instead (and we already effectively gave two more weeks, so we adding a
third one now)
And we should use the time between feature freeze and code freeze to
revert changes of non-completed KIPs as necessary (ie, public API
changes or similar -- we opened two PR just today now that we have a 4.4
branch from our side)
Since feature freeze, actually only bug-fixes and minor KIP follow ups
[of completed KIPs] and new tests should get merged into AK 4.4 branch,
to stabilize the release branch before code freeze.
After code freeze, we should only accept blockers, not regular bug-fixes
or minor improvements any longer. Maybe there is some confusion about
how the deadlines are intended to work?
-Matthias
On 8/11/26 2:22 AM, Chia-Ping Tsai wrote:
hi Matthias
Given that there are still 7 incomplete KIPs and most of them are almost there,
I think it makes sense to extend the code freeze deadline by one week (to Aug
19th) to give the authors a bit more time.
However, after this extension, we should strictly freeze it. Reverting a KIP
from 4.4.0 is absolutely fine if the author does not have the bandwidth to
complete it by the new deadline.
I will review all tickets marked for 4.4.0 soon to ensure we don't miss any
critical bugs.
Best,
Chia-Ping
Matthias J. Sax <[email protected]> 於 2026年8月11日 下午2:41 寫道：
Does it stay as-is
