Am 07.07.2017 um 16:57 schrieb Alessandro Marzialetti via Boost-users:
Hello,
I'm working on a msm with a big transition table (120 elements)
This probably needs a lot of RAM when compiling... That is the reason why I am currently switching from Boost.MSM to (to-be-proposed-)Boost.SML. (You can find it at: https://github.com/boost-experimental/sml)
I already tried with #define BOOST_MPL_CFG_NO_PREPROCESSED_HEADERS then #define BOOST_MPL_LIMIT_LIST_SIZE <custom value> But I can't go over a size of 50 (including "boost/mpl/vector/vector50.hpp")
Any idea on how I can make a bigger table?
You must increase the size of the pre-processed mpl::vectors. For that there exists a python-script in "libs/mpl/preprocessed". It is called "boost_mpl_preprocess.py". Calling that script with the "--help" command should give you some hints on how to call it. You probably need to call it like this: python boost_mpl_preprocess.py --num-elements 120 Possibly, it would be enough to only increase the size of preprocessed mpl::vector. Then you could use the additional options "--no-list", "--no-set" and "--no-map" and thereby leave the size of the preprocessed mpl::list, mpl::set and mpl::map as is. For further background-information you might find this StackOverflow entry, and the conversation on the boost-mailing-list interesting: https://stackoverflow.com/questions/19493630/boostmpl-how-to-generate-pre-ge... http://boost.2283326.n4.nabble.com/MPL-Config-Can-no-longer-pre-generate-hea...
Thanks, Alessandro
<http://www.boost.org/doc/libs/1_63_0/libs/mpl/doc/refmanual/cfg-no-preprocessed-headers.html>
Hope that helps, Deniz