[11u]: RFR: 8238225: Issues reported after replacing symlink at Contents/MacOS/libjli.dylib with binary

[Thomas Stüfe]

Looks good.

..Thomas

On Sun, Feb 9, 2020 at 10:44 PM Langer, Christoph <christoph.langer at sap.com>
wrote:

> Hi,
>
> the fix for "8235687: Contents/MacOS/libjli.dylib cannot be a symlink"
> [0], which had been backported to 11.0.6, has introduced an issue for
> custom launchers that try to start a JVM from a MacOS bundle. This has been
> fixed and we should take this fix back to 11u as well.
>
> Please help to review:
> Bug: https://bugs.openjdk.java.net/browse/JDK-8238225
> Original Patch: https://hg.openjdk.java.net/jdk/jdk/rev/023df178388e
> Webrev: http://cr.openjdk.java.net/~clanger/webrevs/8238225.11u/
>
> The actual fix in src/java.base/macosx/native/libjli/java_md_macosx.c is
> quite straightforward, except that the file was renamed to
> src/java.base/macosx/native/libjli/java_md_macosx.m in head revision which
> caused the need for a little reshuffling. However, I also wanted to take
> the test along which didn't exist yet in jdk11u, so I had to borrow parts
> from JDK-8213362 [1], which introduced the test and subsequent changes
> JDK-8215000 [2] and JDK-8216265 [3]. These  changes however don't qualify
> for a full backport, so I decided for carveouts. I also had to account for
> the fact that in 11u libjli is located in subfolder lib/jli for non-windows
> platforms, so I made a few additional modifications in
> make/test/JtregNativeJdk.gmk and test/jdk/tools/launcher/JliLaunchTest.java.
>
> Thanks
> Christoph
>
> [0] https://bugs.openjdk.java.net/browse/JDK-8235687
> [1] https://bugs.openjdk.java.net/browse/JDK-8213362
> [2] https://bugs.openjdk.java.net/browse/JDK-8215000
> [3] https://bugs.openjdk.java.net/browse/JDK-8216265
>
>