On Thu, Mar 24, 2016 at 7:26 AM, Rene Rivera wrote:

On Wed, Mar 23, 2016 at 10:00 PM, Tom Kent wrote:

...

On Wed, Mar 23, 2016 at 2:29 PM, Vladimir Prus wrote:

...

Second set of release candidates for 1.61.0 beta 1 are now available:

<snip>

...

The changes from RC1 are:

* Top-level directory in the archive matches the name

So in the past the top level directory matched what we were going to eventually release as the beta. E.G. we would have boost_1_XX_0_b1_rcY.tar.bz which would contain the directory boost_1_XX_0_b1 (a similar thing for releases, without the _b1). That way once we have found a -rcY that everyone is happy with, we just need to re-name the archives and push them to sourceforge.

Was it intentional that the directory within is "boost_1_61_0_b1_rc2"? If so, does that mean that we will have to do a separate build (with changes?!?!) when we move from our last rc to the beta release?

It was not intentional. I just wasn't aware of the naming. We could do a new build. Though that's easy and repeatable as it's done by the CI system now. But we could also just unpack/rename/repack.

I think the unpack/rename/repack option could be ok this time, but for the future, it seems like we'd want the system to perform a build where we can simply rename the archive and upload. (For beta and release). Tom

On 24 Mar 2016, at 06:32, Oliver Kowalke wrote:

how can we deal with that? I don't own a Mac, so I can't test it. I assume it's caused by some configure quirks, e.g. in context/build/Jamfile.v2 evaluates ../../config/checks//cxx11_hdr_mutex. depending on the result of this check, boost.thread is linked or not.

It seems that it doesn’t link to boost::system which is required for boost::thread. Performing configuration checks - 32-bit : no - 64-bit : yes - arm : no - mips1 : no - power : no - sparc : no - x86 : yes Building the Boost C++ Libraries. - symlinks supported : yes - C++11 mutex : yes - lockfree boost::atomic_flag : yes - has_icu builds : no warning: Graph library does not contain MPI-based parallel components. note: to enable them, add "using mpi ;" to your user-config.jam - zlib : yes - iconv (libc) : no - iconv (separate) : yes - icu : no - icu (lib64) : no - compiler-supports-visibility : yes - compiler-supports-ssse3 : yes - compiler-supports-avx2 : no - gcc visibility : yes - long double support : yes warning: skipping optional Message Passing Interface (MPI) library. note: to enable MPI support, add "using mpi ;" to user-config.jam. note: to suppress this message, pass "--without-mpi" to bjam. note: otherwise, you can safely ignore this message. - zlib : yes darwin.link.dll bin.v2/libs/context/build/darwin-4.2.1/release/threading-multi/libboost_context.dylib Undefined symbols for architecture x86_64: "boost::system::system_category()", referenced from: __GLOBAL__sub_I_stack_traits.cpp in stack_traits.o "boost::system::generic_category()", referenced from: __GLOBAL__sub_I_stack_traits.cpp in stack_traits.o "boost::thread_detail::enter_once_region(boost::once_flag&)", referenced from: boost::context::stack_traits::is_unbounded() in stack_traits.o boost::context::stack_traits::page_size() in stack_traits.o boost::context::stack_traits::default_size() in stack_traits.o boost::context::stack_traits::maximum_size() in stack_traits.o "boost::thread_detail::commit_once_region(boost::once_flag&)", referenced from: boost::context::stack_traits::is_unbounded() in stack_traits.o boost::context::stack_traits::page_size() in stack_traits.o boost::context::stack_traits::default_size() in stack_traits.o boost::context::stack_traits::maximum_size() in stack_traits.o ld: symbol(s) not found for architecture x86_64 clang: error: linker command failed with exit code 1 (use -v to see invocation) "g++" -dynamiclib -Wl,-single_module -install_name "libboost_context.dylib" -o "bin.v2/libs/context/build/darwin-4.2.1/release/threading-multi/libboost_context.dylib" "bin.v2/libs/context/build/darwin-4.2.1/release/threading-multi/asm/make_x86_64_sysv_macho_gas.o" "bin.v2/libs/context/build/darwin-4.2.1/release/threading-multi/asm/jump_x86_64_sysv_macho_gas.o" "bin.v2/libs/context/build/darwin-4.2.1/release/threading-multi/asm/ontop_x86_64_sysv_macho_gas.o" "bin.v2/libs/context/build/darwin-4.2.1/release/threading-multi/posix/stack_traits.o" "bin.v2/libs/context/build/darwin-4.2.1/release/threading-multi/execution_context.o" -headerpad_max_install_names -Wl,-dead_strip -no_dead_strip_inits_and_terms -arch x86_64

2016-03-24 8:00 GMT+01:00 Thomas Trummer : note the check of C++11 mutex in your output

Building the Boost C++ Libraries.

- symlinks supported : yes

- C++11 mutex : yes

that means that boost.context will/should use std-thread library instead of boost.thread the question is why boost.build evaluates the feature-check of C++11-thread to true and still compiles/linkes with boost.thread. maybe you need a complete clean-checkout of the boost-repo or remove <boost-root>/bin.v1/libs and run b2 with --reconfigure option. at least that's my guess.

On 24 Mar 2016, at 08:27, Oliver Kowalke wrote:

the question is why boost.build evaluates the feature-check of C++11-thread to true and still compiles/linkes with boost.thread.

maybe you need a complete clean-checkout of the boost-repo or remove <boost-root>/bin.v1/libs and run b2 with --reconfigure option. at least that's my guess.

I’m actually using the current beta archive. Independent of boost::context using boost::thread or std::thread, it still still requires linking to boost::system, right? "boost::system::system_category()", referenced from: __GLOBAL__sub_I_stack_traits.cpp in stack_traits.o "boost::system::generic_category()", referenced from: __GLOBAL__sub_I_stack_traits.cpp in stack_traits.o Thomas

On 3/25/2016 5:33 AM, Tom Kent wrote:

...

I would appreciate if folks confirm that this information is sufficient to validate checksums.

I was able to get your key and validate that it had signed the checksums file. gpg --import vladimir.prus.gpg gpg --verify boost_1_61_0_b1_rc1.sums.asc

I was then able to verify that the checksums were correct for the downloaded files. tomkent@bart:~/tmp$ sha256sum -c boost_1_61_0_b1_rc2.sums.asc boost_1_61_0_b1_rc2.7z: OK boost_1_61_0_b1_rc2.tar.bz2: OK boost_1_61_0_b1_rc2.tar.gz: OK boost_1_61_0_b1_rc2.zip: OK sha256sum: WARNING: 21 lines are improperly formatted

Looks like we finally have a verified (rc) release! Thanks for getting that working.

Thanks for confirming, Tom! - Volodya -- Vladimir Prus http://vladimirprus.com