[Live-devel] GroupsockHelper.cpp and weird problems with long startuptimes for applications

[Morgan Tørvolt]

On 17/02/2008, Ross Finlayson <finlayson at live555.com> wrote:
> >It seems from the trace that the ourSourceAddressForMulticast is the
> >sinner. Firstly, what does multicast have to do with the RTSP server?
> >Is that not purely unicast?
>
> The name of that function is an anachronism; I should rename it to
> "ourSourceAddress()"...
>
> Apart from perhaps renaming it, I have no current plans to change
> this function.
>
> (A reminder: Serious professionals do not use "@gmail.com" email addresses :-)
> --

Ross, we had this discussion before, both in the mailing list and
outside of it I believe ;-)

As for the real issue, do you not see the fact that it slows down the
startup process as a real problem? What about the part where it is
resolving a "wrong" IP address? It seems obvious to me that the
hardcoded test IP and port numbers is a unfortunate way of doing this
IP check. I suggest checking the routing table, seeing which interface
gets the default route, and then getting the IP of that interface. If
you do not bother though, fine. I will keep my little "temporary" fix
=)

-Morgan-