If not when will it be available ? :)
No.
> If not when will it be available ? :)
Probably never. SO_REUSEADDR accomplishes the same thing for TCP server
applications or when testing UDP clients and servers on the same computer.
--
Phil Frisbie, Jr.
Hawk Software
http://www.hawksoft.com
I don't think so.
When I have 2 applications listening on the same udp port only 1 application
will receive the udp messages
and the other will receive nothing.
Test it and you will see you are wrong.
I tested it (again).
And I was correct, only one application will receive the udp packets.
The other application might receive a few but that's it.
99.999 % of all packets are received by one application only.
Other stacks do have a SO_REUSEPORT option and if that would not be needed
like you say... then why do other stacks have it ?
So clearly you must be mistaken.
You say SO_REUSEADDR has the same effect as SO_REUSEPORT but I don't think
that is true.
Now it's your turn to test it :)
When you test it make sure to send many packets per second... not just one
or two.
Skybuck.
Actually, why _should_ more than 1 application receive UDP messages aimed at a
port? Only if the socket is bound to a multicast or broadcast address,
naturally, and ISTR that worked the last time I tested it.
Alun.
~~~~
[Please don't email posters, if a Usenet response is appropriate.]
--
Texas Imperial Software | Try WFTPD, the Windows FTP Server. Find us at
1602 Harvest Moon Place | http://www.wftpd.com or email al...@texis.com
Cedar Park TX 78613-1419 | VISA/MC accepted. NT-based sites, be sure to
Fax/Voice +1(512)258-9858 | read details of WFTPD Pro for XP/2000/NT.
Post a small code sample.
> Other stacks do have a SO_REUSEPORT option and if that would not be needed
> like you say... then why do other stacks have it ?
> So clearly you must be mistaken.
>
> You say SO_REUSEADDR has the same effect as SO_REUSEPORT but I don't think
> that is true.
>
> Now it's your turn to test it :)
I DID test it before my first response, both UDP broadcast and multicast. I did
not bother with unicast because why would someone do that?
> When you test it make sure to send many packets per second... not just one
> or two.
UDP is not reliable, not even on loopback!
> Skybuck.
During that test I tested UDP unicast.
>
> > Other stacks do have a SO_REUSEPORT option and if that would not be
needed
> > like you say... then why do other stacks have it ?
> > So clearly you must be mistaken.
> >
> > You say SO_REUSEADDR has the same effect as SO_REUSEPORT but I don't
think
> > that is true.
> >
> > Now it's your turn to test it :)
>
> I DID test it before my first response, both UDP broadcast and multicast.
I did
> not bother with unicast because why would someone do that?
This time I tested multicast and broadcast too and to my surprise everything
was working ok.
I tested windows 98 <-> windows xp and windows 98 <-> windows 98.
Some months or so ago it wasn't working... I wonder why... maybe I did
something wrong
like a typo with addresses or so... or maybe the firewall was not setup
right... pretty weird...
But I know me :) I thought I did do good testing back then... I wanted to
write some broadcast stuff
but I decided not too because I thought I did not work with multiple
applications on the same port...
Now it turns out multicast and broadcast is working just fine when two
applications or using the same port for receiving.
I noticed that on slower machines... one application might receive a little
bit more than the other but that's ok.
Anyway... why limited it to broadcast/multicast ? :)