On Fri, 28 Aug 2026 08:48:03 GMT, GuoziElain <[email protected]> wrote:
>> `ExecutionEnvironment` verifies that the JDK launcher preserves >> `LD_LIBRARY_PATH` when starting a child VM. A RISC-V QEMU launcher wrapper >> needs to set this variable to locate the target libraries, so the wrapper >> overwrites the sentinel value before the child VM starts. The test therefore >> cannot verify its intended launcher property in this environment. >> >> This change excludes only RISC-V running under QEMU, using the same >> `vm.cpu.features` condition already used by other tests. Native RISC-V and >> all other supported Unix platforms continue to run the test. >> >> Testing: >> >> - Reproduced the original failure under `qemu-riscv64-static`: `FAIL: did >> not get <LD_LIBRARY_PATH=/Bridge/On/The/River/Kwai>`. >> - Verified that the test is not selected when `os.arch=riscv64` and >> `vm.cpu.features=qemu`. >> - Verified `test/jdk/tools/launcher/ExecutionEnvironment.java` passes on >> Linux x86_64 (`passed: 1`). >> >> --------- >> - [x] I confirm that I make this contribution in accordance with the >> [OpenJDK Interim AI Policy](https://openjdk.org/legal/ai). > > GuoziElain has updated the pull request incrementally with one additional > commit since the last revision: > > 8353495: Update copyright year Yes, I found one additional similar case: `tools/launcher/Test7029048.java`. It also sets and validates `LD_LIBRARY_PATH`. I reproduced it with QEMU 10.0.0, where the JVM CPU feature string correctly contains `qemu`. Before the update, `ExecutionEnvironment.java` was excluded as intended, while both `LIBJVM` cases in `Test7029048.java` failed with `expected 3 but got 1` because the QEMU wrapper replaced `LD_LIBRARY_PATH`. I added the same RISC-V qemu-user `@requires` condition to `Test7029048.java` in commit 58ea09a2506. With that change, both tests are excluded under RISC-V qemu-user, and `Test7029048.java` still passes on native Linux x86_64 (`passed: 1`). I also audited the other `LD_LIBRARY_PATH` references and did not find another launcher test that directly checks this preservation behavior. ------------- PR Comment: https://git.openjdk.org/jdk/pull/32553#issuecomment-5478473049
