Thank you so much Gavin. as a matter of fact, indeed m32 is not needed if I am making the compilation already in a 32b machine, and putting <cxxflags>-std=gnu++0x has really made the difference. Thank you so much for your support! Best Patricia El 22/01/2015, a las 06:26, Gavin Lambert escribió:

On 20/01/2015 22:41, Patricia Mendez Lorenzo wrote:

...

Dear Boost experts,

I'm definitely not an expert, but...

...

since a while I am trying to compile Boost from the source in a 32b linux machine with c++11 enabled. The compiler version used is gcc4.8.1

This is performed through cmake as follows:

--> Configuration: bootstrap.sh --with-python= --with-toolset=gcc --> Build: /b2 -j4 variant=release --layout=versioned --debug-configuration --toolset=gcc --user-config=<path to a user-config.jam>

Where user-config.jam contains the following information: using gcc : : : <linkflags>-m32 <compileflags>-m32 -std=gnu++0x

also tested it with -std=c++11

And finally installed which basically contains the same options as the build adding --prefix=<installation dir> The libraries built following this procedure does not seem to have been created with c++11.

If it looks like the main difference is that the -std flag was omitted, did you try using this instead:

using gcc : : : <linkflags>-m32 <compileflags>-m32 <cxxflags>-std=gnu++11

If you're compiling 32bit Boost in a 32bit machine you probably don't need the -m32 options though, so just:

using gcc : : : <cxxflags>-std=gnu++11

should do the trick. (Use gnu++0x instead of gnu++11 if you have an older version of gcc.)

_______________________________________________ Boost-users mailing list Boost-users@lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users