November 2004 - Darwin-dev - lists.apple.com

Re: My IP Address
by Garth Cummings 30 Nov '04

30 Nov '04

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Hi Jay, On Nov 30, 2004, at 1:34 PM, Paul Forgey wrote: Paul makes an excellent suggestion here. --gc __________________________________________________________________ Garth Cummings Apple Developer Technical Support gcummings(a)apple.com _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… This email sent to site_archiver(a)lists.apple.com You could also advertise your service in Rendezvous and let clients find you. http://developer.apple.com/technicalsupport smime.p7s

1 0

Re: My IP Address
by Paul Forgey 30 Nov '04

30 Nov '04

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com On Nov 30, 2004, at 12:58 PM, Mark Gilbert wrote: _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… Also, I am creating (in C) an HTML web page with a 'refresh' line, to make the page automatically reload every 2 seconds (to generate a live update of status). In order to do that I have to code my own IP address into the HTML which I return to the client. In this case I would use getsockname() as another person responding suggested. Since you already have a connection with the web client, you now know which interface is being used with it. If you do this per connection, then you will work rather reliably in most cases. Even if there are multiple interfaces or your assigned address changes. So, first of all, I would like to have a status window in my app's GUI which tells a local user which address they need to enter remotely to control the app (without having to look elsewhere in the computer to find it). As has been stated before there is no 100% reliable or sensible way to do that. You may well have multiple interfaces to separate networks which don't route. If you are trying to emulate what the sharing panel does, it uses the fqdn if the system has one, otherwise I believe (maybe wrongly) it uses the IP address of the interface with highest "priority" as specified in the network panel. You could do a best guess by using gethostname() and then resolving the result with gethostbyname(), but that assumes a network where host names have been assigned and are resolvable in DNS. This really is one of the main reasons DNS exists in the first place :). You could also advertise your service in Rendezvous and let clients find you. This email sent to site_archiver(a)lists.apple.com

1 0

Re: [APPL:DarwinDev] Re: My IP Address
by Jay A. Kreibich 30 Nov '04

30 Nov '04

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com User-agent: Mutt/1.4.2.1i On Tue, Nov 30, 2004 at 08:58:42PM +0000, Mark Gilbert scratched on the wall: > So, first of all, I would like to have a status window in my app's > GUI which tells a local user which address they need to enter > remotely to control the app (without having to look elsewhere in the > computer to find it). That's a really good reason, although I might try to probe for the hostname, rather than just the IP address. If a machine has multiple interfaces (== multiple IP addresses) things can get complex. > Also, I am creating (in C) an HTML web page with a 'refresh' line, to > make the page automatically reload every 2 seconds (to generate a > live update of status). In order to do that I have to code my own IP > address into the HTML which I return to the client. If you just want the system to reload the same page, you don't need any URL. Just say "Refresh: 2" or whatever, and it will reload the same URL. If you must use a different URL, you can usually grab the hostname from the "Host:" header in the HTTP 1.1 request. That's kind of cheating, but it would likely work well. Failing that, use the hostname. Two seconds is a bit quick, BTW, especially on a browser like Safari that delays the page rendering on downloads. -j -- Jay A. Kreibich | Comm. Technologies, R&D jak(a)uiuc.edu | Campus IT & Edu. Svcs. <http://www.uiuc.edu/~jak> | University of Illinois at U/C _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… This email sent to site_archiver(a)lists.apple.com

1 0

Re: My IP Address
by Mark Gilbert 30 Nov '04

30 Nov '04

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com The problem with the question is: if the client managed to connect to the web server, why does it need to ask for an address? A cgi-bin to figure out a local address to send back to the client that's already connected to it is somewhat pointless. It isn't pointless because you're assuming those addresses are the same, and that's a very bad assumption because of NAT. In fact, there are several web servers that offer the opposite service-- you connect, and they say "you came from THIS address." If the client is behind a NAT box, the reported address may or may not be different than what your local machine thinks its own address is. In fact, if you need to register a "client" with a dynamic DNS service or something like that, an external reverse probe like this is essential to getting the correct globally-routed address. I'm not sure what the original poster is trying to do, but I assume they know what they are doing. Well, I may not know completely what I am doing, but I am having a go..:-) Thanks for all the replies, I think I have plenty to go on with. Cheers Mark Gilbert. -- mark(a)gallery.co.uk Tel: +44 208 340 5677 fax: +44 870 055 7790 http://www.gallery.co.uk New Product ! - Zen MetaCorder Location Sound Field Recorder http://www.metacorder.info _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… > Basically, I am adding a web based control interface to my application. The idea is that someone could control my app remotely from another CPU via a web browser. This is done by a very, very basic web-type server built inside my app (its just an open socket which spits out simple HTML as a reply to a client's connection). All the HTML is generated dynamically in C. So, first of all, I would like to have a status window in my app's GUI which tells a local user which address they need to enter remotely to control the app (without having to look elsewhere in the computer to find it). Also, I am creating (in C) an HTML web page with a 'refresh' line, to make the page automatically reload every 2 seconds (to generate a live update of status). In order to do that I have to code my own IP address into the HTML which I return to the client. So, there are 2 'pre-emptive' reasons why I need to know my own IP address, before starting any communications with the outside world. Hope this clears things up. I am sure there are simpler ways to do this with clever HTML or CGI's, but I am just a humble C programmer who likes to stick to what I know best, and it all needs to be wrapped up self contained and auto configuring inside my app. This email sent to site_archiver(a)lists.apple.com

1 0

Re: How to get my IP Address ?
by Noah Williamsson 30 Nov '04

30 Nov '04

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com User-agent: Mozilla Thunderbird 0.9 (Windows/20041103) Hi Mark! Folks. Assuming you're using C and the BSD API you could use something like this. #include <stdio.h> #include <unistd.h> #include <string.h> #include <sys/types.h> #include <sys/socket.h> #include <netinet/in.h> #include <arpa/inet.h> #include <errno.h> int main(int argc, char **argv) { int s; struct sockaddr_in sin; int slen; if(argc == 1) { printf("Usage: %s <ip>\n", argv[0]); return -1; } // Connect to port 80 somewhere s = socket(PF_INET, SOCK_STREAM, 0); memset(&sin, 0, sizeof(sin)); sin.sin_family = PF_INET; sin.sin_port = htons(80); sin.sin_addr.s_addr = inet_addr(argv[1]); if(connect(s, (struct sockaddr *)&sin, sizeof(sin)) < 0) { perror("connect()"); return -1; } // Get socket's local name memset(&sin, 0, sizeof(sin)); slen = sizeof(sin); if(getsockname(s, (struct sockaddr *)&sin, &slen) < 0) { perror("getsockname()"); close(s); return -1; } printf("Local IP is %s\n", inet_ntoa(sin.sin_addr)); // The local port is available in ntohs(sin.sin_port) close(s); return 0; } Here's a sample run: insight$ ./1 127.0.0.1 # Use the loopback interface Local IP is 127.0.0.1 insight$ ./1 216.239.59.99 # Google IP, will use external interface Local IP is 194.18.12.187 insight$ As you can see, this code will show you the IP of the outgoing interface, which may or may not be the IP the peer sees due to NAT, as pointed out in other posts. getsockname(3) will get you information about the local peer and getpeername(3) will do the same, but for the remote peer. -- noah _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… Mark Gilbert wrote: I have setup a very simple socket from one machine to another. I want to send back very very simple HTML. within the HTML I need to include my own IP address, but I am not sure how to get it. I tried gethostbyname("localhost"), but this returns 127.0.0.1 which works on the local machine, but is not my address on the network. This email sent to site_archiver(a)lists.apple.com

1 0

Re: [APPL:DarwinDev] Re: How to get my IP Address ?
by Jay A. Kreibich 30 Nov '04

30 Nov '04

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com User-agent: Mutt/1.4.2.1i On Mon, Nov 29, 2004 at 05:30:10PM -0800, Jan Brittenson scratched on the wall: > > The problem with the question is: if the client managed to connect to > the web server, why does it need to ask for an address? A cgi-bin > to figure out a local address to send back to the client that's > already connected to it is somewhat pointless. It isn't pointless because you're assuming those addresses are the same, and that's a very bad assumption because of NAT. In fact, there are several web servers that offer the opposite service-- you connect, and they say "you came from THIS address." If the client is behind a NAT box, the reported address may or may not be different than what your local machine thinks its own address is. In fact, if you need to register a "client" with a dynamic DNS service or something like that, an external reverse probe like this is essential to getting the correct globally-routed address. I'm not sure what the original poster is trying to do, but I assume they know what they are doing. I'll make the standard soap-box speech that network protocol design is very very difficult-- even at the application layer-- and you should really know what you're doing if you are trying to roll your own; that includes the standard warning that unless you are specifically trying to do some type of NAT discovery system, encoding end-point IP addresses in an application protocol stream is usually a Bad Idea, not the least of which is because it is easy for the system to lie. That said, there are still reasons why you might want to do this, especially in a NAT environment. It's just too bad they can't use a standard server like Apache. This could be solved with a dozen line script that simply returns the CGI env var "SERVER_ADDR". It's also worth pointing out that this _is_ a standard CGI env var, meaning somebody somewhere thought it was important enough to provide. -j -- Jay A. Kreibich | Comm. Technologies, R&D jak(a)uiuc.edu | Campus IT & Edu. Svcs. <http://www.uiuc.edu/~jak> | University of Illinois at U/C _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… This email sent to site_archiver(a)lists.apple.com

1 0

Re: How to get my IP Address ?
by Jan Brittenson 30 Nov '04

30 Nov '04

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com User-agent: Mozilla Thunderbird 0.9 (Macintosh/20041103) Jerry Pendergraft wrote: Not sure I completely understand your situtation, but you probably don't need to know your own ip address. If you send out a packet to another machine, that packet will contain a source address which is your address. Such address is put in by the hardware and handles the case of your machine having multiple interfaces. The software on the other machine will then use the address in the packet to reply. If you look in that reply packet the source will be the peer and the destination will be you. The problem with the question is: if the client managed to connect to the web server, why does it need to ask for an address? A cgi-bin to figure out a local address to send back to the client that's already connected to it is somewhat pointless. Now, the web server may direct traffic to a different host. But no amount of looking up local interfaces is going to help with this, nor do I think this is what the OP wondered about. But in general, what I would do is return a name where the service can be found; the name is resolved through name services by the client into a list of addresses. The addresses are tried in order to reach the server. This permits DNS round-robin, reverse NAT and port forwarding, multihoming, redirection, etc, to all just magically work. A server in a DHCP environment may need dynamic DNS or equivalent. For more symmetric peer LAN applications there's tons of discovery protocols out there (including Rendezvous), which is what I think Jerry is reaching for above. But this leaves out the role of the web service: SMIL? Generic XML data exchange vehicle? Session management? Load balancing? Edge redirection? Ticketing and admission? Or just a starting point? IMO more info is needed, and if there isn't more info it's time to go back to the drawing board and rethink. :) _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… This email sent to site_archiver(a)lists.apple.com

1 0

Re: How to get my IP Address ?
by Jerry Pendergraft 29 Nov '04

29 Nov '04

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Not sure I completely understand your situtation, but you probably don't need to know your own ip address. If you send out a packet to another machine, that packet will contain a source address which is your address. Such address is put in by the hardware and handles the case of your machine having multiple interfaces. The software on the other machine will then use the address in the packet to reply. If you look in that reply packet the source will be the peer and the destination will be you. -- Jerry Pendergraft jerry.pendergraft(a)endocardial.com Endocardial Solutions voice: 651-523-6935 1350 Energy Lane, Suite 110 fax: 651-644-7897 St Paul, MN 55108-5254 mobile: 651-491-0163 On Sat, 27 Nov 2004, Mark Gilbert wrote: > Folks. > > I have setup a very simple socket from one machine to another. I > want to send back very very simple HTML. > > within the HTML I need to include my own IP address, but I am not > sure how to get it. > > I tried gethostbyname("localhost"), but this returns 127.0.0.1 which > works on the local machine, but is not my address on the network. > > For example, I happen to know that my machine is 192.169.1.104, and I > can reach my machine remotely using this. How does my app learn that > this is my address ? > > I am sure its a really simple trick, but I am obviously missing it. > > Cheers > > Mark Gilbert. > _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… This email sent to site_archiver(a)lists.apple.com

1 0

Re: How to get my IP Address ?
by Marc Majka 29 Nov '04

29 Nov '04

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… This email sent to site_archiver(a)lists.apple.com #include <stdio.h> #include <stdlib.h> #include <sys/types.h> #include <sys/socket.h> #include <netinet/in.h> #include <arpa/inet.h> #include <net/if.h> #include <ifaddrs.h> #include <errno.h> int main(int argc, char *argv[]) { struct ifaddrs *myaddrs, *ifa; struct sockaddr_in *s4; struct sockaddr_in6 *s6; int status; /* buf must be big enough for an IPv6 address (e.g. 3ffe:2fa0:1010:ca22:020a:95ff:fe8a:1cf8) */ char buf[64]; status = getifaddrs(&myaddrs); if (status != 0) { perror("getifaddrs"); exit(1); } for (ifa = myaddrs; ifa != NULL; ifa = ifa->ifa_next) { if (ifa->ifa_addr == NULL) continue; if ((ifa->ifa_flags & IFF_UP) == 0) continue; if (ifa->ifa_addr->sa_family == AF_INET) { s4 = (struct sockaddr_in *)(ifa->ifa_addr); if (inet_ntop(ifa->ifa_addr->sa_family, (void *)&(s4->sin_addr), buf, sizeof(buf)) == NULL) { printf("%s: inet_ntop failed!\n", ifa->ifa_name); } else { printf("%s: %s\n", ifa->ifa_name, buf); } } else if (ifa->ifa_addr->sa_family == AF_INET6) { s6 = (struct sockaddr_in6 *)(ifa->ifa_addr); if (inet_ntop(ifa->ifa_addr->sa_family, (void *)&(s6->sin6_addr), buf, sizeof(buf)) == NULL) { printf("%s: inet_ntop failed!\n", ifa->ifa_name); } else { printf("%s: %s\n", ifa->ifa_name, buf); } } } freeifaddrs(myaddrs); exit(0); }

1 0

Re[4]: Preprocesor ?
by Igor Shmukler 28 Nov '04

28 Nov '04

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com > The C++ standard, ISO/IEC 14882 (2003), does not say anything about the > names of source files. This is left up to implementations. > > The GCC compiler accepts a number of extensions, including .cc, .cpp, > and .cxx. There's no reason, other than personal preference or your > organization's coding standards, to prefer one over another. (I do > recommend against using .C, though. It's fine as far as GCC is > concerned, but it interacts poorly with case-insensitive file systems.) My bad, you are correct. Thank you for noticing this. I pointed out that I am not positive this was a part of the standard. The .cpp is however a common extension agreed upon by almost everyone. It's what both Mozilla requires (http://www.mozilla.org/hacking/portable-cpp.html#file_extension_is_cpp) and what would Microsoft IDE make for you. If fact every place I worked for last few years had .cpp as a standard extension. The Apple Darwin sources are all .cpp just the same. _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… This email sent to site_archiver(a)lists.apple.com

1 0