[reSIProcate] epoll prototype patch

[Francis Joanis]

Hi,

Sounds like this would be easier with a more automated build system where
you can tell the configure phase to actually try to compile bits of code to
see if it works (as suggested, look at the configure.in file of libevent and
search for epoll_create).

I'm currently not a huge fan of autotools but I am having great success with
CMake on some of my projects. One thing I like about it is that it generates
the makefiles/project files (VS, XCode, ...) for you from a single set of
instructions (you don't need to keep on changing the Windows VisualStudio
project files by hands, for example).

Do we know if the current build scripts are all still in use and working (I
mean all the targets and options - I mainly only use the core ones myself)?
If it turns out only a subset of them are really required it will probably
not be a huge task to port the build system to something like CMake.

As for using libevent directly in reSIProcate, this could be very
interesting (since it looks like it is already doing the best event
"selection" abstractions per OS). We would need to see if it would easily
integrate within the threading model of the SipStack (I'm thinking
process/buildFdSet).

Regards,
Francis

On Tue, Nov 16, 2010 at 7:12 PM, Philip Kizer <pckizer at nostrum.com> wrote:

> On 2010/11/16, at 17:50, Kennard White wrote:
> > I would appreciate thoughts on the best way to determine if a given
> platform
> > supports epoll(). Should I add this as an option to the configure script?
> Is
> > there any more automatic approach?
>
> I'm not sure off-hand, but I'd look to other projects that make use of
> OS-offered event functions like libevent.
>
> Since I do not know if you can use libevent (<
> http://monkey.org/~provos/libevent/>) directly, perhaps you can at least
> see how that project does it and use similar mechanisms.
>
> If you could use libevent (or its mechanisms), you might even get support
> for kevents and other more efficient implementations along with it.
>
>
> -p
>
> _______________________________________________
> resiprocate-devel mailing list
> resiprocate-devel at resiprocate.org
> https://list.resiprocate.org/mailman/listinfo/resiprocate-devel
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://list.resiprocate.org/pipermail/resiprocate-devel/attachments/20101116/2b94c473/attachment.htm>