Boost logo

Boost-Build :

From: Rene Rivera (grafikrobot_at_[hidden])
Date: 2006-08-27 18:46:56


Matt England wrote:
> As per:
>
> http://svn.cleversafe.org/moxie/trunk/thirdparty-builder/scripts/boost.sh
>
> The stanza starting at line 45 is commented out such that (from what we
> understand) the succeeding stanza will build and install all in one step.

Next time please include the relevant piece of code in emails. First;
some people may not have immediate access to the web while reading
email. Second; it makes it much easier to comment if we don't have to go
guessing where in the file you pointed to as in this case there are no
line numbers visible for reference. To make sure is this the code you
are referring to:

for LIB in ${BOOST_LIBS}; do
   echo "Building library ${LIB}..."
# bjam -sTOOLS=gcc -sBUILD="multi" --with-${LIB} \
# --layout="system" stage 2>&1 || \
# fail "error: could not build boost library ${LIB}"
   bjam -sTOOLS=gcc -sBUILD="multi" --with-${LIB} --layout="system" \
       --prefix=`pwd`/../../${INSTALL_DIR} install 2>&1 || \
      fail "error: could not install boost library ${LIB}"
done

??

> Which is right/preferred/better? Why does the script without the
> uncommented version not work?

Neither seems correct to me. By specifying '-sBUILD="multi"', which is
not a legal BUILD value, you'll end up getting some imaginary "multi"
variant getting built which is most likely never going to work. The
maximal build, which is the default, uses a BUILD value of "-sBUILD=
debug release <runtime-link>static/dynamic <threading>single/multi". If
you don't want the maximal set of libs built I suggest you remove some
of the options to get what you want. For example "-sBUILD=release
<runtime-link>dynamic <threading>multi".

-- 
-- Grafik - Don't Assume Anything
-- Redshift Software, Inc. - http://redshift-software.com
-- rrivera/acm.org - grafik/redshift-software.com
-- 102708583/icq - grafikrobot/aim - grafikrobot/yahoo

Boost-Build list run by bdawes at acm.org, david.abrahams at rcn.com, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk