[omniORB] OmniORB 4.04 windows bug !

[Renzo Tomaselli]

As it is now, there is no way to have it running stable on Windows. The bad
point is that it seems do depend on local configuration getting socket n.
higher or lower 2048. Far from predictable, thus I vote for a new release.

Renzo Tomaselli

----- Original Message ----- 
From: "Duncan Grisby" <duncan at grisby.org>
To: "Renzo Tomaselli" <renzo.tomaselli at tecnotp.it>
Cc: "Omniorb list" <omniorb-list at omniorb-support.com>
Sent: Tuesday, August 17, 2004 5:12 PM
Subject: Re: [omniORB] OmniORB 4.04 windows bug !


> On Friday 13 August, "Renzo Tomaselli" wrote:
>
> >     ver. 4.04 cannot run properly on Windows.
> > Methods SocketCollection::is/clear/setSelectable got a new initial test
> > about socket n. not being higher than 2048 on Windows (sock >=
> > SELECTABLE_FD_LIMIT). This test is new (not in 4.03).
>
> The intention of the new code is to make sure that the sockets are
> actually selectable, otherwise you can end up waiting for ever for an
> event to occur.
>
> The problem is that Windows fd_sets work completely differently from
> Unix ones, so the new code broke on Windows. On Windows, FD_SETSIZE is
> the number of sockets that can be stored in the set; on Unix, it's the
> highest socket number that can be stored.
>
> I've fixed the code in CVS so it doesn't do the check on Windows. The
> question now is whether that warrants doing a new full release. What
> does everyone on the list think?  Is it worth doing 4.0.5 just for
> this?
>
> Cheers,
>
> Duncan.
>
> -- 
>  -- Duncan Grisby         --
>   -- duncan at grisby.org     --
>    -- http://www.grisby.org --
>
> _______________________________________________
> omniORB-list mailing list
> omniORB-list at omniorb-support.com
> http://www.omniorb-support.com/mailman/listinfo/omniorb-list
>