On 26.08.2009, at 22:05, Philippe Marschall wrote:

2009/8/26 Andreas Raab andreas.raab@gmx.de:

...

Quick Unix question: How do I find out what the IP address for eth0/ eth1 is? I thought NetNameResolver>>localHostAddress was supposed to return this but it returns consistently 127.0.0.1 which is fine from a purity point of view but utterly useless from any practical perspective ;-)

I reported exactly this a while ago and was told that the problem is with me and not the code. Then I suggested that the code would be changed to directly return 127.0.0.1. That would achieve the same result more efficiently. Somehow that never got implemented though.

What got implemented though is a primitive to list all interfaces that are available. Hosts do not have IP addresses, interfaces do. And since hosts typically have multiple interfaces, it's not really up to the VM to pick one, but the application.

- Bert -

On 26.08.2009, at 22:27, Andreas Raab wrote:

Bert Freudenberg wrote:

...

What got implemented though is a primitive to list all interfaces that are available. Hosts do not have IP addresses, interfaces do. And since hosts typically have multiple interfaces, it's not really up to the VM to pick one, but the application.

What's the primitive? How do I use it? Any examples?

There are examples and tests in the image ... it's part of the IPv6 work, which did not only add IPv6 support but generalized the Socket support. Not sure exactly which method it would be.

- Bert -

NetNameResolver addressesForName: self localHostName

but that gives back a list of three address like 10.37.129.2(10.37.129.2),0(0)-inet4-stream-tcp

but lacking is information about WHICH one is actually usable, the 192.168.1.139 is the correct one for my office subnet but 10.211.55.2, and 10.37.129.2 have been setup by Parallels for os- x, If I pick 10.211.55.2, or 10.37.129 then things just won't work since computer B elsewhere in the office doesn't see the 10.*.*.* addresses.

Obviously the data coming back has to be a bit more helpful to let the smalltalk code pick the right answer.

ifconfig shows

en1: flags=8863<UP,BROADCAST,SMART,RUNNING,SIMPLEX,MULTICAST> mtu 1500 inet6 fe80::21b:63ff:fe02:d2db%en1 prefixlen 64 scopeid 0x6 inet 192.168.1.139 netmask 0xffffff00 broadcast 192.168.1.255 ether 00:1b:63:02:d2:db media: autoselect status: active supported media: autoselect

en4: flags=8963<UP,BROADCAST,SMART,RUNNING,PROMISC,SIMPLEX,MULTICAST> mtu 1500 inet6 fe80::21c:42ff:fe00:8%en4 prefixlen 64 scopeid 0x8 inet 10.211.55.2 netmask 0xffffff00 broadcast 10.211.55.255 ether 00:1c:42:00:00:08 media: autoselect status: active supported media: autoselect

en5: flags=8963<UP,BROADCAST,SMART,RUNNING,PROMISC,SIMPLEX,MULTICAST> mtu 1500 inet6 fe80::21c:42ff:fe00:9%en5 prefixlen 64 scopeid 0x9 inet 10.37.129.2 netmask 0xffffff00 broadcast 10.37.129.255 ether 00:1c:42:00:00:09 media: autoselect status: active supported media: autoselect

On 26-Aug-09, at 1:27 PM, Andreas Raab wrote:

Bert Freudenberg wrote:

...

What got implemented though is a primitive to list all interfaces that are available. Hosts do not have IP addresses, interfaces do. And since hosts typically have multiple interfaces, it's not really up to the VM to pick one, but the application.

What's the primitive? How do I use it? Any examples?

Thanks,

• Andreas

-- = = = ======================================================================== John M. McIntosh johnmci@smalltalkconsulting.com Twitter: squeaker68882 Corporate Smalltalk Consulting Ltd. http://www.smalltalkconsulting.com = = = ========================================================================

On 26.08.2009, at 23:57, Andreas Raab wrote:

John M McIntosh wrote:

...

NetNameResolver addressesForName: self localHostName

This doesn't work at all for me. In fact it's not even asking the right question ;-) Using the 3.10-6 VM with an Etoys-dev image I get this:

NetNameResolver localHostName => 'localhost.localdomain'

NetNameResolver addressesForName: NetNameResolver localHostName => 127.0.0.1(localhost),0(0)-inet4-stream-tcp

I'm not surprised with the result because you are asking DNS to resolve localhost.localdomain which is different from asking the question of what IP addresses are assigned to the various interfaces in your system.

I thought Bert said that there is a way to enumerate the IP addresses of the interfaces?

I thought it did, but I never had to use it yet ... grmph, I'm sure there is a way. Ian?

- Bert -

On 27.08.2009, at 00:18, John M McIntosh wrote:

Sure well

NetNameResolver localHostName gives me 'Otter-2.local'

then

NetNameResolver addressesForName: 'Otter-2.local' gives me back a collection with 6 addresses, but you are saying for your machine you only get back 1 address? Lucky you, but will your production machine give back one address, or 23?

I think here the problem is your machine is returning 'localhost.localdomain' Oddly on os-x asking for NetNameResolver addressForName: 'localhost.localdomain' gives you a primitive failure.

Also as you saw, my localHostName is 'Otter-2.local' which is the boujour name http://support.apple.com/kb/TA20999?viewlocale=en_US

Maybe the question is does your computer have a DNS assigned name?

This machine has two interfaces (see below).

NetNameResolver localHostName

gives

'fedora10'

but

NetNameResolver addressesForName: NetNameResolver localHostName

still only yields

an OrderedCollection(127.0.0.1(fedora10),0(spr-itunes)-inet4-stream- tcp)

That does not sound right indeed.

- Bert -

[bert@fedora10 ~]$ ifconfig eth0 Link encap:Ethernet HWaddr 00:0C:29:BF:6D:A7 inet addr:172.16.22.154 Bcast:172.16.22.255 Mask: 255.255.255.0 inet6 addr: fe80::20c:29ff:febf:6da7/64 Scope:Link UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 RX packets:223095 errors:194711 dropped:0 overruns:0 frame:0 TX packets:108938 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:1000 RX bytes:179764576 (171.4 MiB) TX bytes:76413825 (72.8 MiB) Interrupt:19 Base address:0x2024

lo Link encap:Local Loopback inet addr:127.0.0.1 Mask:255.0.0.0 inet6 addr: ::1/128 Scope:Host UP LOOPBACK RUNNING MTU:16436 Metric:1 RX packets:38 errors:0 dropped:0 overruns:0 frame:0 TX packets:38 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:0 RX bytes:6908 (6.7 KiB) TX bytes:6908 (6.7 KiB)

David T. Lewis wrote:

Assuming that you're not looking for an elegant solution, you could figure it out in the start script and pass it to Squeak as a command line parameter.

#!/bin/sh INTERFACE=eth0 IPV4_ADDR=`/sbin/ifconfig $INTERFACE | grep 'inet addr' | cut -c21-34` echo the IP address on interface $INTERFACE is $IPV4_ADDR /usr/local/bin/squeak myimage.image $IPV4_ADDR

Wow, without sed? I'm impressed ;-) And you know what, that's just what I'll do. Thanks for the snippet.

Cheers, - Andreas

Careful, on my macbook pro, the ethernet port is en0, which is rarely used. But en1 is wireless and mostly used so the INTERFACE=eth0 is going to be machine dependent

On 26-Aug-09, at 3:15 PM, Andreas Raab wrote:

David T. Lewis wrote:

...

Assuming that you're not looking for an elegant solution, you could figure it out in the start script and pass it to Squeak as a command line parameter. #!/bin/sh INTERFACE=eth0 IPV4_ADDR=`/sbin/ifconfig $INTERFACE | grep 'inet addr' | cut - c21-34` echo the IP address on interface $INTERFACE is $IPV4_ADDR /usr/local/bin/squeak myimage.image $IPV4_ADDR

Wow, without sed? I'm impressed ;-) And you know what, that's just what I'll do. Thanks for the snippet.

Cheers,

• Andreas

-- = = = ======================================================================== John M. McIntosh johnmci@smalltalkconsulting.com Twitter: squeaker68882 Corporate Smalltalk Consulting Ltd. http://www.smalltalkconsulting.com = = = ========================================================================

On Aug 26, 2009, at 3:15 PM, Andreas Raab wrote:

David T. Lewis wrote:

...

#!/bin/sh

Wow, without sed? I'm impressed ;-) And you know what, that's just what I'll do. Thanks for the snippet.

If you change your mind, your primitive could evolve trivially from the following.

On Darwin this picks up inet4 and inet6 interface addresses. SIOCGIFCONF seems to be broken on Linux and it only picks up inet4 addresses regardless of the address family of the socket.

#include <stdio.h> #include <unistd.h> #include <sys/socket.h> #include <sys/ioctl.h> #include <net/if.h> #include <netinet/in.h> #include <arpa/inet.h>

#ifndef s6_addr16 # define s6_addr16 __u6_addr.__u6_addr16 #endif

int main(int argc, char **argv) { struct ifreq *ifr; struct ifconf ifc; char buf[32768]; int s= socket(AF_INET, SOCK_DGRAM, 0); if (-1 == s) return -1; ifc.ifc_len= sizeof(buf); ifc.ifc_buf= buf; if (ioctl(s, SIOCGIFCONF, &ifc) < 0) { perror("SIOCGIFCONF"); goto die; } for (ifr= ifc.ifc_req; (char *)ifr < (char *)ifc.ifc_req + ifc.ifc_len;) { switch (ifr->ifr_addr.sa_family) { case AF_INET: { struct sockaddr_in *sin= (struct sockaddr_in *)&ifr->ifr_addr; printf("%s\t%s\n", ifr->ifr_name, inet_ntoa(sin->sin_addr)); break; } case AF_INET6: { struct sockaddr_in6 *sin6= (struct sockaddr_in6 *)&ifr->ifr_addr; struct in6_addr *in6= &sin6->sin6_addr; printf("%s\t%x:%x:%x:%x:%x:%x:%x:%x\n", ifr->ifr_name, in6->s6_addr16[0], in6->s6_addr16[1], in6->s6_addr16[2], in6-

s6_addr16[3],

in6->s6_addr16[4], in6->s6_addr16[5], in6->s6_addr16[6], in6-

s6_addr16[7]);

break; } default: break; } # if defined(linux) ++ifr; # else ifr= (struct ifreq *)((char *)ifr + sizeof(ifr->ifr_name) + ifr-

ifr_addr.sa_len);

# endif } die: close(s); return 0; }

yes, but OS-X is not Linux...

Also I'm not sure it's correct http://bugs.squeak.org/view.php?id=7392

On 26-Aug-09, at 4:30 PM, Bert Freudenberg wrote:

On 27.08.2009, at 01:10, Ian Piumarta wrote:

...

On Aug 26, 2009, at 3:15 PM, Andreas Raab wrote:

...

David T. Lewis wrote:

...

#!/bin/sh

Wow, without sed? I'm impressed ;-) And you know what, that's just what I'll do. Thanks for the snippet.

If you change your mind, your primitive could evolve trivially from the following.

So the extended SocketPlugin cannot list all ip addresses under Linux? On John's VM it does I thought ...

• Bert -

-- = = = ======================================================================== John M. McIntosh johnmci@smalltalkconsulting.com Twitter: squeaker68882 Corporate Smalltalk Consulting Ltd. http://www.smalltalkconsulting.com = = = ========================================================================