How speedy are patches to be downported to 11?

[Andrew Haley]

On 17/02/2021 07:43, Thomas Stüfe wrote:
> what is the general policy of downporting changes from 17 to 11? I see
> sometimes changes downported rather quickly which could have done with a
> little more baking time. Especially since 17 is not even released yet.
> 
> Are there guidelines to follow? I did not find anything in the Wiki. Eg
> minimum time to let a patch stew in the head release?
> 
> Related to that, how are possible compatibility issues to be handled? Eg. a
> patch fixing the output of a command may be correct but may still throw off
> parsers.
> 
> I suspect the answer will be "its up to you", and depends
> on patch complexity and bug severity. But I still am curious how you handle
> this in practice.

It completely depends on severity and urgency, and to some extent on
the age of the release. As JDK 11 gets older, we should be increasingly
reluctant to patch it: users have a reasonable assumption that a
release becomes increasingly stable over time.

It's always better to allow a patch to bake in head for some time, but
often there's no good reason to delay. In addition to what Aleksey said,
some complex and intrusive patches need a whole release cycle, more
than six months.

-- 
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