[wasphin]

What problem does this PR solve?

Problem Summary:

The JSON memory pool dereferences a failed chunk allocation and assumes a successful allocation when growing a buffer.

What is changed and the side effects?

Changed:

• Return nullptr from pool allocation and reallocation when the backing allocator fails.
• Preserve existing chunks and contents after a failed allocation.
• Add a failing backing-allocator regression to CMake and Bazel.

Side effects:

• Performance effects: Negligible checks after allocating a chunk.
• Breaking backward compatibility: No public signature changes. This fixes the allocator contract; DOM consumers still need their own allocation-failure handling.

---

Check List:

• Changes compile locally.
• Added focused test coverage.
• Please follow Contributor Covenant Code of Conduct.

[Copilot]

🟢 Approval recommended

The localized checks preserve pool state on failure, and the regression covers failure and recovery paths; tests were not executed during this review.

No open issues 🎉

What changed in this PR

This PR makes bRPC’s bundled RapidJSON memory pool handle backing-allocation failures without dereferencing null pointers or altering existing pool contents.

Changes:

• Return null when chunk allocation or buffer growth fails.
• Add regression coverage for allocation failures, state preservation, and successful recovery.
• Register the regression test in CMake and Bazel.

File	Description
test/​rapidjson_allocator_unittest.cc	Tests failure handling, preserved contents, and recovery.
test/​CMakeLists.txt	Adds the regression to the CMake test sources.
test/​BUILD.bazel	Adds the regression to the Bazel test sources.
src/​butil/​third_party/​rapidjson/​allocators.h	Checks allocation results before modifying chunks or copying data.

_{Review effort: Balanced}