Boost logo

Boost Users :

Subject: Re: [Boost-users] Boost.MSM cc1plus.exe:: error: out of memory allocating 65532
From: Christophe Henry (christophe.j.henry_at_[hidden])
Date: 2013-02-28 18:05:10


> Hey!
>
> Thanks for your fast and helpful response!
>
> One point regarding the "favor_compile_time" approach: I've alrady used it
> for the outermost and all enclosed submachines already. Unfortunately I
> couldn't detemine such a deep performance advantage. But one comment
> makes
> me a bit sceptical: two days avo you have suggested another user to also
> use
> favor_compile_time in context with the eUML front-end and kleene events,
> but
> according to your comment it has a disadvantage: the statemachines will
> lose
> the hierarchical support. Will this also happen for the functor-frontend?
> Maybe I've misunderstood something...
>
> Ciao,
>
> Rafael

There are limitations but it usually compiles a bit faster, plus as the fsm
is divided between different TUs, with -j9 (8 submachines + outer) you
should see some difference ;-) (supposing you really use several cpp files)

The limitation is not the fsm hierarchy, which of course still works. What
you lose is the event hierarchy, the capability to use in a transition an
event, which is the base class of the event you actually process. Little
known feature but useful, unfortunately lost in this mode.
And yes, it will happen in all front-ends :(

Cheers,
Christophe


Boost-users list run by williamkempf at hotmail.com, kalb at libertysoft.com, bjorn.karlsson at readsoft.com, gregod at cs.rpi.edu, wekempf at cox.net