fix(cpp): fix memory leaks and null dereference crashes

[ColinLeeo]

Summary

Fix 4 classes of memory/crash bugs discovered during a systematic scan of the C++ codebase.

Changes

1. ChunkReader double-free (src/reader/chunk_reader.cc)

reset() frees the wrapped buffer via mem_free(file_data_buf) but does not clear the wrapped pointer in ByteStream. When destroy() is subsequently called, get_wrapped_buf() returns the already-freed pointer and mem_free(buf) is called again → double-free crash.

Fix: Call in_stream_.clear_wrapped_buf() immediately after mem_free() to null out the stale pointer. destroy() already has its own null-guard, so this prevents the double-free.

2. ChunkReader read-failure memory leak (src/reader/chunk_reader.cc)

load_by_meta() allocates file_data_buf via mem_alloc(), then calls read_file_->read(). The original code only freed the buffer on the partial-success path (IS_SUCC(ret) && ret_read_len < MIN_SERIALIZED_SIZE). If read() itself returns an error, the buffer is never freed and never passed to wrap_from() → memory leak for every failed chunk load.

Fix: Add an early-exit path after read(): if !IS_SUCC(ret), free the buffer and return immediately.

3. LZ4Compressor error-path leaks (src/compress/lz4_compressor.cc)

compress() allocates compressed_buf_ via mem_alloc(), then has two error paths that do not free it:

• LZ4_compress_default returns ≤ 0 → compressed_buf_ leaked
• mem_realloc returns nullptr → compressed_buf_ leaked

Fix: Add mem_free(compressed_buf_); compressed_buf_ = nullptr; on both error paths.

4. C wrapper string getter null dereference (src/cwrapper/tsfile_cwrapper.cc)

Both tsfile_result_set_get_value_by_{name,index}_string() call r->get_value<common::String*>() which delegates to Field::get_string_value(). For non-STRING/TEXT/BLOB column types, this returns nullptr. The code then unconditionally dereferences ret->len_ → SIGSEGV.

Fix: Add a null check before the dereference, returning nullptr (which the Python caller already handles correctly, returning None).

Verification

• All 721 C++ tests pass (TsFile_Test with ASan/Debug)
• clang-format applied via spotless:apply

Files Changed

File	+/−
cpp/src/reader/chunk_reader.cc	+7/−1
cpp/src/compress/lz4_compressor.cc	+4/−0
cpp/src/cwrapper/tsfile_cwrapper.cc	+6/−0

[codecov-commenter]

Codecov Report

❌ Patch coverage is 0% with 10 lines in your changes missing coverage. Please review.
✅ Project coverage is 60.47%. Comparing base (a2fdf71) to head (b3a0002).

Files with missing lines	Patch %	Lines
cpp/src/compress/lz4_compressor.cc	0.00%	4 Missing ⚠️
cpp/src/reader/chunk_reader.cc	0.00%	2 Missing and 2 partials ⚠️
cpp/src/cwrapper/tsfile_cwrapper.cc	0.00%	0 Missing and 2 partials ⚠️

Additional details and impacted files

@@             Coverage Diff             @@
##           develop     #879      +/-   ##
===========================================
- Coverage    60.48%   60.47%   -0.02%     
===========================================
  Files          744      744              
  Lines        49014    49021       +7     
  Branches      7784     7787       +3     
===========================================
- Hits         29646    29643       -3     
- Misses       17966    17972       +6     
- Partials      1402     1406       +4     

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[Copilot]

Pull request overview

This PR fixes several C++ robustness issues in TsFile’s reader/compression/C-wrapper layers by addressing concrete memory-management bugs (double-free/leaks) and preventing a null dereference when requesting string values via the C API.

Changes:

• Prevent double-free in ChunkReader::reset() by clearing the wrapped buffer pointer after freeing it.
• Fix a read-failure leak in ChunkReader::load_by_meta() by freeing the allocated buffer on read() error and returning early.
• Fix LZ4Compressor::compress() error-path leaks by freeing compressed_buf_ when compression or realloc fails.
• Add null guards in C-wrapper string getters to avoid dereferencing a null common::String*.

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated no comments.

File	Description
cpp/src/reader/chunk_reader.cc	Avoids double-free on reset and plugs a buffer leak when chunk-header reads fail.
cpp/src/compress/lz4_compressor.cc	Frees the internal compression buffer on LZ4 failure paths to prevent leaks.
cpp/src/cwrapper/tsfile_cwrapper.cc	Returns nullptr instead of crashing when a non-string column is requested as string.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.