Re: Core dump while building with CPP

om: wesmck...@gmail.com >> Date: Mon, 1 Aug 2016 13:07:11 -0700 >> Subject: Re: Core dump while building with CPP >> To: dev@arrow.apache.org >> >> See from flatbuffers.h >> >> // The wire format uses a little endian encoding (since that's efficient fo

RE: Core dump while building with CPP

Didn't help much, I actually rebuilt whole arrow cpp with export CMAKE_CXX_FLAGS="-DFLATBUFFERS_LITTLEENDIAN=0" and then tried to build simple debug, still it's failing in same step. Thanks,Sanjay > From: wesmck...@gmail.com > Date: Mon, 1 Aug 2016 13:07:11 -0700 >

Re: Core dump while building with CPP

This looks like somehow flatbuffers (a third party library) isn't getting setup correctly for big-endian systems. According to the documentation [1] it should be able to work. It would take some investigations to figure why it isn't. [1] https://google.github.io/flatbuffers/md__internals.html O

Re: Core dump while building with CPP

See from flatbuffers.h // The wire format uses a little endian encoding (since that's efficient for // the common platforms). #if !defined(FLATBUFFERS_LITTLEENDIAN) #if defined(__GNUC__) || defined(__clang__) #ifdef __BIG_ENDIAN__ #define FLATBUFFERS_LITTLEENDIAN 0 #else #def

Core dump while building with CPP

cmake -DARROW_BUILD_BENCHMARKS=ON .make[ 0%] Running flatc compiler on /u01/hadoop/src/arrow/arrow/cpp/../format/Message.fbsAssertion failed: *reinterpret_cast(&endiantest) == FLATBUFFERS_LITTLEENDIAN, file /u01/hadoop/src/arrow/arrow/cpp/thirdparty/flatbuffers-1.3.0/include/flatbuffers/fl