When doing [JDK-8326960](https://bugs.openjdk.org/browse/JDK-8326960), I missed the case when sysroot would be restored from the cache. This would skip the configure and build steps, because it would only consult the status for create-sysroot job. This is only a problem when PR gets tested the _second_ time, the first time we would always create a sysroot. I only noticed this after merging the master into existing PRs.
This is a regression that might hide build issues on platforms, and so it should be fixed ASAP. Additional testing: - [x] GHA first run: cross-compiled builds run - [x] GHA second run: cross-compiled builds run ------------- Commit messages: - Empty commit to trigger second GHA run - Fix Changes: https://git.openjdk.org/jdk/pull/18475/files Webrev: https://webrevs.openjdk.org/?repo=jdk&pr=18475&range=00 Issue: https://bugs.openjdk.org/browse/JDK-8328948 Stats: 2 lines in 1 file changed: 0 ins; 0 del; 2 mod Patch: https://git.openjdk.org/jdk/pull/18475.diff Fetch: git fetch https://git.openjdk.org/jdk.git pull/18475/head:pull/18475 PR: https://git.openjdk.org/jdk/pull/18475
