On Thu, 20 Mar 2025 09:52:02 GMT, Aleksey Shipilev <sh...@openjdk.org> wrote:
> See the bug for rationale. > > This goal for this improvement is to be easily backportable, so we can catch > up with update releases. As such, it does a few borderline-trivial changes, > and _does not_ change the jspawnhelper protocol. So the overwrite of > `jspawnhelper` with this new version would be "compatible" in a very weak > sense of the word. > > Additional testing: > - [x] Linux x86_64 server fastdebug, `java/lang/ProcessBuilder` Yes, _I agree_ with the _original_ point for the jspawnhelper version checks, not in the last part because I am the one who made those points an year ago :) But a year later, we have more reports -- from users! -- which prompts at least myself to reconsider. I was trying to explain this with a wall of text in the JBS issue ([JDK-8352489](https://bugs.openjdk.org/browse/JDK-8352489)) -- I suggest to keep the discussion on the rationale of the change there. Let's leave PR discussion for the technical comments about the implementation itself. ------------- PR Comment: https://git.openjdk.org/jdk/pull/24127#issuecomment-2740395613
