Le 22/01/14 17:43, John Maddock a écrit :
Does anyone happen to have an instance of this for a test that fails on Windows? That would let me test locally. Otherwise, I would need the bjam log from the test runner, since the usual source of this kind of error is a bug the log scrapper program and that is impractical to debug without the log itself.
The library build is failing, and the tool is picking a random source file to display. Unfortunately, it isn't the one that failed.
OK, some hunting around shows it's an issue with Boost.Thread not building on that platform: http://www.boost.org/development/tests/develop/developer/thread.html
Hi, yes there are surely some issues with Boost.Thread but here it is the report http://www.boost.org/development/tests/develop/developer/marshall-mac-thread... Test output: boost/bin.v2/libs/thread/build/clang-darwin-asan/debug/threading-multi Rev 3d21a16db5ada89d4c5875074a87ba81239048a0 / Wed, 22 Jan 2014 06:18:00 +0000 Compile [2014-01-22 12:13:47 UTC]: succeed "/Volumes/TwoTB/LLVM/bin/bin/clang++" -x c++ -fsanitize=address -O0 -g -Wextra -Wno-long-long -Wno-variadic-macros -Wunused-function -pedantic -O0 -fno-inline -Wall -pedantic -g -DBOOST_ALL_NO_LIB=1 -DBOOST_ATOMIC_DYN_LINK=1 -DBOOST_SYSTEM_DYN_LINK=1 -DBOOST_THREAD_BUILD_DLL=1 -DBOOST_THREAD_DONT_USE_CHRONO -DBOOST_THREAD_POSIX -I".." -c -o "/Volumes/TwoTB/boost/regression/master/results/boost/bin.v2/libs/thread/build/clang-darwin-asan/debug/threading-multi/future.o" "../libs/thread/src/future.cpp" With this report I don't know what to fix :( Best, Vicente P.S. BTW, the build on these targets work on my computer without any problem.