Hi Bryan, I tried sending the following earlier today, but it appears I was having issues with the mailing list.
I started looking into the issue myself. I bisected one of the issues to a change in https://github.com/apache/arrow/pull/8533 and asked the author about the check that seemed to be causing the issue (https://github.com/apache/arrow/pull/8533#issuecomment-721815501). Benjamin replied back as you can see in the thread and it looks like the `DCHECK_NE` line can be removed. Locally, I'm also seeing another issue I'm starting to debug, but the following diff "fixes" the builder for me locally: diff --git a/cpp/src/arrow/io/memory.cc b/cpp/src/arrow/io/memory.cc index 1cde5e64e..88f006fe0 100644 --- a/cpp/src/arrow/io/memory.cc +++ b/cpp/src/arrow/io/memory.cc @@ -54,7 +54,7 @@ BufferOutputStream::BufferOutputStream(const std::shared_ptr<ResizableBuffer>& b Result<std::shared_ptr<BufferOutputStream>> BufferOutputStream::Create( int64_t initial_capacity, MemoryPool* pool) { // ctor is private, so cannot use make_shared - DCHECK_NE(pool, nullptr); auto ptr = std::shared_ptr<BufferOutputStream>(new BufferOutputStream); RETURN_NOT_OK(ptr->Reset(initial_capacity, pool)); return ptr; diff --git a/java/flight/flight-core/src/test/java/org/apache/arrow/flight/TestDoExchange.java b/java/flight/flight-core/src/test/java/org/apache/arrow/flight/TestDoExchange.java index b7e7a20be..04383c476 100644 --- a/java/flight/flight-core/src/test/java/org/apache/arrow/flight/TestDoExchange.java +++ b/java/flight/flight-core/src/test/java/org/apache/arrow/flight/TestDoExchange.java @@ -303,7 +303,7 @@ public class TestDoExchange { final FlightStream reader = stream.getReader(); reader.cancel("", null); // Cancel should be idempotent - reader.cancel("", null); } } The second change is due to the test suite throwing an exception about the test having a memory leak, and could also be related to https://github.com/apache/arrow/pull/8533 since that updates memory allocation. Update: Since then I've identified that the nullptr is coming from https://github.com/apache/arrow/blob/3eb6e69d9c1ff34978d138bb8848a36cb938cec6/cpp/src/jni/orc/jni_wrapper.cpp#L229 and I can both remove the check which was causing the segfault, but also propose not using a nullptr explicitly and use `arrow::default_memory_pool()` which is used elsewhere in the JNI wrapper. I was just running the tests with the adjusted JNI wrapper and going to see if that fixed the other error I was seeing about the memory leaking. -Terence On 2020/11/03 20:53:59, Bryan Cutler <cutl...@gmail.com> wrote: > There seems to be a Github check error with the Java JNI tests for the ORC > adapter that is affecting a lot of recent PRs, see > https://github.com/apache/arrow/pull/8577/checks?check_run_id=1346780145. > From the log, it looks like some env setting, but I can't tell what's > wrong. Anyone else know? > > Thanks, > Bryan >
