HI, 
thanks for looking into this!

i actually use pd 0.47 on RPi B+ / jessie

in any case i found a way to use the suggested workaround with [iemnet/udpsend]
i send [disconnect, connect IP PORT( only as when a message is received with [netrecieve -u -b]
and i prevent any messages beeing sent thru [udpsend] if the socket is closed (a zero on the outlet of udpsend)

so for the last couple of hours i did not have any errors like “iemnet: socket-shutdown: transport endpoint is not connected”
even if i switch off the wlan router - which caused the error before.

now i really hope that this solves the pd-freeze issue, since it might be that the closed socket and the pd-freeze coincided per accident…or that there is some other OS weirdness going on .



On 17 Feb 2019, at 23:26, Lucas Cordiviola <lucarda27@hotmail.com> wrote:

Hi,

I've been testing [netsend -u] for a couple of hours on Pd 0.49-0 on windows. I had no trouble if there is no receivers. I used a windows tablet and a linux box for listening. I have switch them off for a while and then re-enter to the LAN. Didn't get any "Socket operation on non-socket (88)".

A couple of things you could check:

+ Are you using Pd 0.49-0 on your raspberry?

+ Could it be that there is something on the OS that is closing "Socket connections" (or network adapter) to save battery power or something like that?



Mensaje telepatico asistido por maquinas.
On 2/17/2019 5:23 AM, michael strohmann wrote:
i am sending OSC messages with [netsend -u -b]

there is obviously an issue with [netsend] closing the socket if nobody is listening.
which it should not do in UDP mode.

but this:

netsend: Socket operation on non-socket (88)

is lethal to my patch. immediate freeze
unfortunatly it occurs quite often.

i also know next to nothing about Sockets
is it a pd problem or does it stem from the os (unix jessie in my case, on raspberry) ???

also due to Christof Ressi:
I can confirm that both [iemnet/udpsend] [mrpeach/udpsend] (from Pd extended) close the socket on Linux but only [mrpeach/udpsend] gives a warning and sends a 0. I'll do a bug report.

in the meantime, are there any options to send osc via udp regardless of the existance of a receiver?

cheers!
michael


On 16 Feb 2019, at 20:06, Lucas Cordiviola <lucarda27@hotmail.com> wrote:

Did you tried sending every second an empty message [send ( instead of
"reopening a connection" ?


Mensaje telepatico asistido por maquinas.

On 2/16/2019 3:57 PM, michael strohmann wrote:
i finally could caught pd in the act of freezing:

the console said:

recv: Connection refused (111)
netsend: Socket operation on non-socket (88)
warning: 33 removed from poll list but not found


i connect every second to mobile phones that run Lemur to send and receive OSC
usually if a phone leaves the network, nothing bad happens, the console says:


recv: Connection refused (111)
netsend: Bad file descriptor (9)
warning: 32 removed from poll list but not found


So it seems to me that this non-socket(88) problem manages to freeze pd
(or is it this number: 88, which has a bad connotation in austria ?)


Thanks!
michael




_______________________________________________
Pd-list@lists.iem.at mailing list
UNSUBSCRIBE and account-management -> https://lists.puredata.info/listinfo/pd-list
_______________________________________________
Pd-list@lists.iem.at mailing list
UNSUBSCRIBE and account-management -> https://lists.puredata.info/listinfo/pd-list