Also, install the manual in the lib install directory.

Indeed, "opam install camlidl" does put dllcamlidl.so into stublibs.

also, rename call-via-socket.ml inti call_via_socket.ml to avoid ocaml warning.

also, rename call-via-socket.ml inti call_via_socket.ml to avoid ocaml warning.

just a commit to increment the version number to make sure it works

Really, it is the job of lutin to set the seed.

Also, unplug the "calling lutin from lustre and C" tests, as they are
actually broken since the move to oasis/ocamlbuild.

There is still one problem: the rdbg and the lurette mode do not
generated the same sequence. The event number is the same. The
sequence of Random.init (performed when the event number is
incremented) is exactly the same too. Weird.

Note that this divergence is a regression

Also correct a lot of bugs.

Indeed, this module was not tested when I initially wrote it, because
of the sig faults pbs.

Not that it is still not working properly, as the CPS transformation
breaks Pascal's way of handling Lutin choice points backtracking, that
is based on ocaml exception. Indeed consider this exemple:

try
   call_to_a_func_that_traverse_the_lutin_FT ()
with Deadlock _ ->
   the_backtrack code

now, if call_to_a_func_that_traverse_the_lutin_FT returns an event,
the ocaml try is deleted !!!

because i'd like to use that name for the ocaml/lutin lib.

The C binding does not work yet (hence the test in examples/C and
examples/lustre are broken).

(53d82954)

Add the lutin tuto.

as it can confuse Genlex (and hence the Rif reader) from times to
times (e.g., 42e_blabla raises a float_of_string exception).

cf http://stackoverflow.com/questions/24111981/how-can-i-achieve-bash-exit-trap-when-exec-ing-another-binary

sim2chro arount supports this option (Sigh).

Tx to Nicolas Berthier.

It was due to an uninitialized local var ; the bug was revealed by a change in
caml/mlvalues.h::CAMLparam() done in 4.02.0

nb : it also fixes (i don't really know why) a performance bug in the rabbit simu.