On Thu, 12 Sep 2024 06:39:59 GMT, David Holmes <dhol...@openjdk.org> wrote:
> Aside: we do an awful lot of Java code execution in the launcher these days
> before we even get to load the real main class. I have to wonder how all this
> affects startup?
In addition to regular regression testing, I also ran existing internal startup
benchmarks against this change. I noticed no observable regression or
improvement with these changes.
> src/java.base/share/classes/sun/launcher/LauncherHelper.java line 607:
>
>> 605: if (mainAttrs == null) {
>> 606: abort(null, "java.launcher.jar.error3", jarname);
>> 607: }
>
> Why do we no longer need a null check?
The null check is still done, but it's now done before calling this private
method. It's done here
https://github.com/openjdk/jdk/pull/20929/files/abc3100078c2ab522054d1074d1cdc53c2879388#diff-108a3a3e3c2d108c8c7f19ea498f641413b7c9239ecd2975a6c27d904c2ba226R848
-------------
PR Comment: https://git.openjdk.org/jdk/pull/20929#issuecomment-2345411767
PR Review Comment: https://git.openjdk.org/jdk/pull/20929#discussion_r1756234772
