JDK 25.0.2 base repository

[Lindenmaier, Goetz]

Hi Aleksey, 

Maybe we could do that rollback,
but it would not be really nice.
But is this really necessary?
We don't need that for our process.

Anyways, I announced this before jdk25u-dev was
created: https://mail.openjdk.org/pipermail/jdk-updates-dev/2025-November/050191.html
Actually when I proposed this there was
a huge amount of outstanding backports, but 
in the end many were abandoned.

How do you deal with *.0.1?  Is that different 
from now in any way?  Less hassle?

Best regards, 
  Goetz.

> -----Original Message-----
> From: Shipilev, Aleksey <shipilev at amazon.de>
> Sent: Monday, December 15, 2025 4:48 PM
> To: Lindenmaier, Goetz <goetz.lindenmaier at sap.com>; jdk-updates-
> dev at openjdk.org
> Subject: Re: JDK 25.0.2 base repository
> 
> On 15.12.25, 15:45, "Lindenmaier, Goetz" <goetz.lindenmaier at sap.com
> <mailto:goetz.lindenmaier at sap.com>> wrote:
> > Yes, you can either sync below that change or pull jdk25.0.3+0, which is the
> tag that does not yet contain any 25.0.3 changes.
> > Basically it's like when testing a *.0.1 while changes for *.0.2 have already
> been pushed to the respective repo.
> 
> Right. It makes sense in pre-dev-split world.
> 
> The problem with this approach after jdk25u and jdk25u-dev split is that
> downstream vendors need to know that jdk25u is _not_ the base for a
> quarterly release. It was a long-standing behavior (even codified on Wiki) that
> downstreams can just pull from jdk${X}u and that would be sensible quarterly
> release. It would make sense for jdk${X}u to contain recent GA bits right
> before and immediately after the quarterly release. The downstream
> pipelines are likely pulling the jdk25u now for testing -- what they think --
> 25.0.2 EA, while in reality they get something else.
> 
> We have caught this in our downstream Corretto 25.0.2 pipeline: we
> automatically sync from openjdk/jdk25u for quarterly release baseline. And
> the stuff that was not supposed to be in 25.0.2 was still there. We can
> probably monkey-patch current pipelines to deal with it, but it is a hassle.
> Anyway, I suspect we are not the only ones who this affects, and without
> some sort of intervention, some vendors would accidentally release
> something that is not a real 25.0.2. I strongly believe this is the problem
> waiting to happen in only 4 weeks.
> 
> Arguably, this is "no good deed goes unpunished" kind of thing: accepting
> older PRs to jdk25u after the split was very convenient (thanks!), but now it
> bites us in our collective rears.
> 
> So, how do you feel about rolling back jdk25u to 25.0.2 state? I think jdk25u-
> dev was already merged from jdk25u, so all 25.0.3 stuff should be in jdk25u-
> dev already. It would be awkward if post-GA jdk25u -> jdk25u-dev merge
> accidentally reverts some 25.0.3 stuff, though. I suspect that can be helped if
> we merge from jdk25u immediately, preferring jdk25u-dev side.
> 
> Something like (untested, just for illustration):
>   a. jdk25u: git diff jdk25.0.3+0...HEAD > revert.patch
>   b. jdk25u: git apply revert.patch && git commit -m "Revert to 25.0.2 state"
>   c. jdk25u-dev: git merge upstream/jdk25u -s ours
> 
> Thanks,
> -Aleksey
> 
> 
> 
> 
> Amazon Web Services Development Center Germany GmbH
> Tamara-Danz-Str. 13
> 10243 Berlin
> Geschaeftsfuehrung: Christof Hellmis, Andreas Stieger
> Eingetragen am Amtsgericht Charlottenburg unter HRB 257764 B
> Sitz: Berlin
> Ust-ID: DE 365 538 597