Beyond Linux® From Scratch (System V Edition) - Version r12.2-615

Chapter 9. General Libraries

boost-1.86.0

Introduction to Boost

Boost provides a set of free peer-reviewed portable C++ source libraries. It includes libraries for linear algebra, pseudorandom number generation, multithreading, image processing, regular expressions and unit testing.

[Note]

Note

Development versions of BLFS may not build or run some packages properly if LFS or dependencies have been updated since the most recent stable versions of the books.

[Note]

Note

This package will extract to the boost-1.86.0/ directory.

Package Information

Additional Downloads

Boost Dependencies

Recommended

Which-2.21

Optional

ICU-76.1, NumPy-2.1.3, and Open MPI

Editor Notes: https://wiki.linuxfromscratch.org/blfs/wiki/boost

Installation of Boost

First, fix a problem with Boost and NumPy-2.1.3 that causes the build to fail: 

patch -Np1 -i ../boost-1.86.0-upstream_fixes-1.patch

Next, fix a build issue which occurs in the stacktrace library. This issue is specific to i686 systems. 

case $(uname -m) in
   i?86)
      sed -e "s/defined(__MINGW32__)/& || defined(__i386__)/" \
          -i ./libs/stacktrace/src/exception_headers.h ;;
esac

This package can be built with several jobs running in parallel. In the instructions below, <N> stands for the number of jobs. Install Boost by running the following commands: 

./bootstrap.sh --prefix=/usr --with-python=python3 &&
./b2 stage -j<N> threading=multi link=shared

To run Boost.Build's regression tests, issue pushd tools/build/test; python3 test_all.py; popd.

To run every library's regression tests, issue pushd status; ../b2; popd. A few tests may fail. They take a very long time (over 119 SBU at -j4) and use a very large amount of disk space (46 GB). You should use the -jN switch to speed them up.

[Note]

Note

Boost installs many versioned directories in /usr/lib/cmake. If a new version of Boost is installed over a previous version, the older cmake directories need to be explicitly removed. To do this, run as the root user: 

rm -rf /usr/lib/cmake/[Bb]oost*

before installing the new version.

Now, as the root user: 

./b2 install threading=multi link=shared

Command Explanations

threading=multi: This parameter ensures that Boost is built with multithreading support.

link=shared: This parameter ensures that only shared libraries are created, except for libboost_exception and libboost_test_exec_monitor which are created as static. Most people will not need the static libraries, and most programs using Boost only use the headers. Omit this parameter if you do need static libraries.

--with-python=python3: This switch ensures Python3 is used if Python2 is installed.

-jN: This switch may be added to the b2 command lines, to run up to N processes in parallel.

Contents

Installed Programs: None
Installed Libraries: libboost_atomic.so, libboost_charconv.so, libboost_chrono.so, libboost_container.so, libboost_context.so, libboost_contract.so, libboost_coroutine.so, libboost_date_time.so, libboost_exception.a, libboost_fiber.so, libboost_filesystem.so, libboost_graph.so, libboost_iostreams.so, libboost_json.so, libboost_locale.so, libboost_log_setup.so, libboost_log.so, libboost_math_c99.so, libboost_math_c99f.so, libboost_math_c99l.so, libboost_math_tr1.so, libboost_math_tr1f.so, libboost_math_tr1l.so, libboost_nowide.so, libboost_numpy313.so, libboost_prg_exec_monitor.so, libboost_program_options.so, libboost_python313.so, libboost_random.so, libboost_regex.so, libboost_serialization.so, libboost_stacktrace_addr2line.so, libboost_stacktrace_basic.so, libboost_stacktrace_noop.so, libboost_system.so, libboost_test_exec_monitor.a, libboost_thread.so, libboost_timer.so, libboost_type_erasure.so, libboost_unit_test_framework.so, libboost_url.so, libboost_wave.so, and libboost_wserialization.so
Installed Directory: /usr/include/boost

Short Descriptions

libboost_*.so

contains libraries that provide a variety of functions, including wrappers for NumPy, linear algebra, pseudorandom number generation, multithreading, image processing, regular expressions, unit testing, locales, JSON, logging, graphing, and debugging