JDK 11u policy notice: no more enhancements

[Andrew Haley]

On 10/26/21 14:54, Andrew Haley wrote:
> Now that 17u is the current stable OpenJDK release, we need to remind
> contributors of the rules for 11 updates. In particular, there are some
> proposals for backports of performance improvements which we should not
> allow. 11u is a stable branch to be maintained above all else.

One thing I should have said, but I didn't:

   ** This policy is up for discussion, like everything else. **

It has been suggested to me that we should not have a hard stop at this point,
but taper over time. That seems reasonable, and we should pick a time period
over which to do that.


However, doing so means risking breakage in the meantime.

Consider the 8u bug

   8270533: AArch64: size_fits_all_mem_uses should return false if its output is a CAS

It was a fix for a bug introduced by

   8267689: [aarch64] Crash due to bad shift in indirect addressing mode

Now, this fix for 8267689 looks perfectly reasonable, and works in head, but
it has a bug that's latent in head but is very much not latent in 8u. As
Java Release X gets older, head moves further and further from Release X
and such problems become increasingly likely.

-- 
Andrew Haley  (he/him)
Java Platform Lead Engineer
Red Hat UK Ltd. <https://www.redhat.com>
https://keybase.io/andrewhaley
EAC8 43EB D3EF DB98 CC77 2FAD A5CD 6035 332F A671