Antoine Pitrou

> And yeah, I think the "stale" case is what [@pitrou](https://github.com/pitrou) was interested in -- he can motivate it further but IIUC the concern was that older projects don't always...

A small (probably innocuous) leak in reliably reported by the ASAN job. I am narrowing it down and trying to come up with a workaround: ``` ==13823==ERROR: LeakSanitizer: detected memory...

A more detailed leak trace (obtained using `LSAN_OPTIONS=fast_unwind_on_malloc=0:malloc_context_size=100`) is the following: ``` Direct leak of 968 byte(s) in 1 object(s) allocated from: #0 0x5600b66641e8 in calloc (/build/cpp/debug/arrow-azurefs-test+0x4c31e8) (BuildId: 5f129cb231bb2651f5fb432f4c8d3c2f40506da3) #1...

The leak manifests in `TestAzuriteFileSystem.OpenOutputStreamCloseAsync` and `TestAzuriteFileSystem.OpenOutputStreamAsyncDestructor`. However, `TestAzuriteFileSystem.OpenOutputStreamAsyncDestructorNoBackgroundWrites` is fine.

@OliLay There's already a skip related to a libxml2 leak with threaded operation, so I added another one. This might actually be fixed by https://github.com/Azure/azure-sdk-for-cpp/pull/5767 which was recently merged in...

Actually, no, we may need something like this hack from ClickHouse: https://github.com/ClickHouse/ClickHouse/blob/054b38d4ebeabc33010fb532766f937af2d44c03/programs/server/Server.cpp#L987-L998 It was added in this PR: https://github.com/ClickHouse/ClickHouse/pull/45796

@github-actions crossbow submit -g cpp

Hmm, I rebased for CI.

@github-actions crossbow submit -g cpp

macOS build failure is unrelated as it happened elsewhere: https://github.com/apache/arrow/actions/runs/10481257647/job/29030430952