Re: [certi-dev] CERTI 3.4.1 released!!

[Eric Noulard]

2011/10/25 Michael Raab <address@hidden>:
> Hi Eric,
>
> when fetching the recent CVS version of CERTI I found out that there are
> some uncommited changes in my local repository.
> Could you please have a look if it is resonable to commit some of them?
> Especially for the visibilty of the RTIA during execution I would propose to
> add some CMAKE flag to be able to turn visibilty on/off.
> What do you think? Patch is attached.

Hi Michael,

Yes we can definitely take some if not all of them, too bad
you did not send those before the release :-(

Would you be able to split them in order to make the review easier?

- One  for the RTIA console thing and attach it to
  https://savannah.nongnu.org/patch/?7401
  I [wrongly] thought I did merge this but in fact I didn't

  I'm OK with the CMake option for that with the default behavior
  being "invisible" (visibility = off)

- One for the TerminateProcess in ambassador destructor
  What was the purpose of that?
  Why is it enabled in 1516 and not in 1.3 ?

- And a last one for the modification in M_Classes.cc
  M_Query_Lookahead::M_Query_Lookahead()
   --> lookahead = 0.0

  However, you should never modify "[N]M_Classes.[cc|hh]" by hand besides for
  testing purpose. Those files are generated from the CERTI message
  specfication files:
   see
http://download.savannah.gnu.org/releases/certi/papers/CERTI-GenMsg-presentation-6june2011.pdf
   if you want to know how it works.

  The source of thoses files are:
    scripts/CERTI_Message.msg   --> for M_Classes.*
    scripts/CERTI_NetworkMessage.msg  --> for NM_Classes.h

  in your case your modification could be done as:
message M_Query_Lookahead : merge Message {
        required double lookahead {default=0}
}

instead of current:
message M_Query_Lookahead : merge Message {
        required double lookahead
}

   Don't bother with a patch for that last one, if you just explain me
why we need a default value
   I'll do it, or may be other messages of the same kind deserve a
default value as well so
   in that case a patch would be better in order to avoid forgetting one.


-- 
Erk
Membre de l'April - « promouvoir et défendre le logiciel libre » -
http://www.april.org