| 文件 | 最后提交记录 | 最后更新时间 |
|---|---|---|
upload source code | 8 个月前 | |
chore: consolidate sandbox follow-up fixes Signed-off-by: mhsong2 <songminhui2@huawei.com> | 26 天前 | |
fix(test): restore functionsystem test execution Recent litebus API changes, distro hardening defaults, and the test harness runtime environment had drifted apart enough that the repo-level test entrypoint either failed before execution, discovered zero cases, or hung in several InstanceCtrl paths. This change realigns build flags, runtime env wiring, and the affected tests so the requested repo workflow is trustworthy again. Constraint: User-required verification path is the real repo entrypoint bash run.sh test -j8 Constraint: Memory pressure requires keeping verification capped at -j8 instead of full parallelism Rejected: Switch test away from Debug | hides the real compiler/toolchain conflict instead of fixing it Rejected: Keep -fstack-check and force-disable clash protection | weakens the host toolchain hardening defaults Rejected: Only patch the hanging tests | leaves discovery, link, and remote-signal regressions unresolved Confidence: high Scope-risk: moderate Reversibility: clean Directive: If run.sh test regresses again, verify discovery counts and serial-suite handling before chasing individual test failures Tested: ./functionsystem/build/bin/functionsystem_unit_test --gtest_filter=InstanceCtrlTest.KillInstanceWithCreating; ./functionsystem/build/bin/functionsystem_unit_test --gtest_filter=InstanceCtrlTest.ForwardCustomSignalSuccess; ./functionsystem/build/bin/functionsystem_unit_test --gtest_filter=InstanceCtrlTest.*; bash run.sh test -j8 Not-tested: Full test run above -j8; non-Linux toolchains beyond the currently validated environment Signed-off-by: yuchaow <wangyuchao12@huawei.com> | 2 个月前 |
| 文件 | 最后提交记录 | 最后更新时间 |
|---|---|---|
| 8 个月前 | ||
| 26 天前 | ||
| 2 个月前 |