test: changed the buffer offset

[thangktran]

to avoid problem with the new behaviour of new V8 BackingStore API. By
changing the offset, the base address of each test case will be
different.

Fixes: #31061

Checklist

• make -j4 test (UNIX), or vcbuild test (Windows) passes
• tests and/or benchmarks are included
• documentation is changed or added
• commit message follows commit guidelines

[thangktran]

/cc @addaleax@richardlau

Could someone please run a stress test on this PR to verify that the flake is fixed. Thank you.

[addaleax]

Stress test CI: https://ci.nodejs.org/job/node-stress-single-test/35/

[richardlau]

Since there’s no protection in the binding I’d suggest some sort of comment in the test to discourage new check()s being added in the future that result in duplicate base addresses.

[nodejs-github-bot]

CI: https://ci.nodejs.org/job/node-test-pull-request/28164/

[thangktran]

Since there’s no protection in the binding I’d suggest some sort of comment in the test to discourage new check()s being added in the future that result in duplicate base addresses.

fixed

[nodejs-github-bot]

CI: https://ci.nodejs.org/job/node-test-pull-request/28214/

[nodejs-github-bot]

CI: https://ci.nodejs.org/job/node-test-pull-request/28218/

[nodejs-github-bot]

CI: https://ci.nodejs.org/job/node-test-pull-request/28221/

[Trott]

Landed in 783f8c6

[sam-github]

Is https://ci.nodejs.org/job/node-test-commit-linux-containered/17217/ / https://ci.nodejs.org/job/node-test-commit-linux-containered/17217/nodes=ubuntu1804_sharedlibs_openssl111_x64/ also a (non-AIX) symptom of the same underlying problem?

00:09:40 not ok 2531 addons/buffer-free-callback/test
00:09:40 ---
00:09:40 duration_ms: 0.426
00:09:40 severity: crashed
00:09:40 exitcode: -4
00:09:40 stack: |-
00:09:40 00:09:40 00:09:40 #
00:09:40 # Fatal error in , line 0
00:09:40 # Check failed: result.second.
00:09:40 #
00:09:40 #
00:09:40 #
00:09:40 #FailureMessage Object: 0x7ffd3790cd20
00:09:40 1: 0x55cde33b5151 [out/Release/node]
00:09:40 2: 0x55cde41cabce V8_Fatal(char const*, ...) [out/Release/node]
00:09:40 3: 0x55cde37e679d v8::internal::GlobalBackingStoreRegistry::Register(std::shared_ptr<v8::internal::BackingStore>) [out/Release/node]
00:09:40 4: 0x55cde3505c9e v8::ArrayBuffer::GetBackingStore() [out/Release/node]
00:09:40 5: 0x55cde3320e7f node::Buffer::New(node::Environment*, char*, unsigned long, void (*)(char*, void*), void*) [out/Release/node]
00:09:40 6: 0x55cde332151c node::Buffer::New(v8::Isolate*, char*, unsigned long, void (*)(char*, void*), void*) [out/Release/node]
00:09:40 7: 0x7f13e05a7f89 Alloc(v8::FunctionCallbackInfo<v8::Value> const&) [/home/iojs/build/workspace/node-test-commit-linux-containered/test/addons/buffer-free-callback/build/Release/binding.node]
00:09:40 8: 0x55cde354b4b3 [out/Release/node]
00:09:40 9: 0x55cde354d3af v8::internal::Builtin_HandleApiCall(int, unsigned long*, v8::internal::Isolate*) [out/Release/node]
00:09:40 10: 0x55cde3da43f9 [out/Release/node]
00:09:40 ...

The daily build failure predates this PR, so maybe is fixed.

[richardlau]

@sam-github I believe it is -- it's failing the same check and the underlying issue that was fixed is not AIX specific.