> On 18 May 2015, at 12:44, Manfred Kröhnert <mkroehner...@googlemail.com>
> wrote:
>
> Hello Sven,
>
> thanks for your fast reply.
>
> On Mon, May 18, 2015 at 11:41 AM, Sven Van Caekenberghe <s...@stfx.eu> wrote:
> Hi Manfred,
>
> Did you see/study the example/test UDPSocketEchoTest ?
>
> unfortunately I did not see this class, yet.
> But I did look at the Socket class itself and Socket>>timeTestUDP (in Pharo3).
We removed #timeTestUDP in Pharo 4 and higher, I believe, and replaced it with
proper tests/examples.
> Also, UDP is very primitive: it is send and forget, returns no errors, and is
> totally asynchronous. Listening should best be done in a separate process.
>
> Yes, UDP is primitive but the SSDP specification requires using UDP.
>
> Here is how I would write your send part:
>
> | message udpSocket |
> message := String crlf join: #(
> 'M-SEARCH * HTTP/1.1'
> 'HOST:239.255.255.250:1900'
> 'MAN:"ssdp:discover'
> 'ST:ssdp:all'
> 'MX:1'
> '').
>
> udpSocket := Socket newUDP.
> udpSocket sendData: message toHost: #[239 255 255 250] port: 1900.
> udpSocket waitForSendDoneFor: 5.
> udpSocket closeAndDestroy.
>
> Thanks for this snippet.
> For some reasons I do not understand yet, this gives me the expected result
> in contrast to my own version.
>
> I modified it like this to print the results on the Transcript:
>
> | message udpSocket |
> message := String crlf join: #(
> 'M-SEARCH * HTTP/1.1'
> 'HOST:239.255.255.250:1900'
> 'MAN:"ssdp:discover"'
> 'ST:ssdp:all'
> 'MX:1'
> '').
>
> udpSocket := Socket newUDP.
> udpSocket sendData: message toHost: #[239 255 255 250] port: 1900.
> udpSocket waitForSendDoneFor: 5.
> udpSocket waitForData.
> [udpSocket dataAvailable ] whileTrue: [ | buffer |
> buffer := ByteArray new: 1024.
> udpSocket receiveUDPDataInto: buffer.
> Transcript show: 'Received: ', buffer.
> [udpSocket waitForDataFor: 1] on: Error do: [ ].
> ].
> udpSocket closeAndDestroy.
>
> Currently, this will work for me, since I am not interested in continuously
> detecting new devices.
> But you mentioned reading should in general be done in a separate process.
> Would you reuse the existing udpSocket instance in this case or create a
> separate one for reading?
I think if it is only for a one time use, it is OK like you are doing it.
> Maybe nothing came back because the line ends where wrong ?
>
> The lineendings I used are required by the SSDP specification.
I meant that writing '\r\n' has no effect in Pharo, this is not processed (you
could do this manually, of course). This is why I did it with the #join:
That is the difference and why it works now, I guess.
> HTH,
>
> Sven
>
> Thanks,
> Manfred
>
> > On 18 May 2015, at 11:29, Manfred Kröhnert <mkroehner...@googlemail.com>
> > wrote:
> >
> > Hello everyone,
> >
> > I am currently trying to get a list of UPnP device via SSDP (Simple Service
> > Discovery Protocol).
> >
> > A minimalistic working NodeJS example looks like this:
> >
> > var dgram = require('dgram');
> > var message = new Buffer(
> > "M-SEARCH * HTTP/1.1\r\n" +
> > "HOST:239.255.255.250:1900\r\n" +
> > "MAN:\"ssdp:discover\"\r\n" +
> > "ST:ssdp:all\r\n" +
> > "MX:1\r\n" +
> > "\r\n"
> > );
> >
> > var client = dgram.createSocket("udp4");
> >
> > client.on("message", function (msg, rinfo) {
> > console.log("server got: " + msg + " from " + rinfo.address + ":" +
> > rinfo.port);
> > });
> > client.send(message, 0, message.length, 1900, "239.255.255.250");
> > client.close();
> >
> > I tried to recreate the same functionality in Pharo with the following
> > piece of code:
> >
> > | message udpSocket host |
> > message := 'M-SEARCH * HTTP/1.1\r\n' ,
> > 'HOST:239.255.255.250:1900\r\n' ,
> > 'MAN:"ssdp:discover"\r\n' ,
> > 'ST:ssdp:all\r\n' ,
> > 'MX:1\r\n' , '\r\n'.
> >
> > udpSocket := Socket newUDP.
> > host := (NetNameResolver addressFromString: '239.255.255.250').
> > udpSocket sendData: message toHost: host port: 1900.
> > "udpSocket waitForData."
> > Transcript show: 'Received: ' , udpSocket receiveData.
> > udpSocket closeAndDestroy.
> >
> > Unfortunately, this script hangs both in Pharo3 and Pharo4 (OS X 10.10,
> > Pharo 30856 and 40611).
> > When I interrupt with Cmd+. the debugger shows the method
> > Socket>>waitForDataIfClosed:
> > and the hang is apparently in the line
> > self readSemaphore wait
> >
> > Could anyone give me some advice if I should to do this differently or what
> > the issue with the Pharo script is?
> >
> > Thanks in advance,
> > Manfred
>
>
>
