getting the end of an web address in c? - c

Question

say I pass an argument www.bbc.co.uk/news/world-us-canada-11893886
I need to separate www.bbc.co.uk from /news/world-us-canada-11893886 for a HTTP GET
I have tried using strtok and strcat but I come across weird splits at runtime.
I can get www.bbc.co.uk just fine using strtok( host, "/");
I have tried using a combination of strtok and strcat to try and get all the rest of the string from the first "/" but i get an output like this...
request: da-11893886
tempString: news/world!
host: www.bbc.co.uk
Path: news/world!da-11893886
If you look at this output, the strangest part is that it always cuts out the middle section.
In this case, the "-us-cana"
the section of the code is attached below
// testing purposes
printf("argv[1]: %s\n", argv[1] );
host = malloc(sizeof(argv[1]));
strcpy(host, argv[1]);
host = strtok(host, "/");
// get the request
request = malloc(sizeof(argv[1]) + sizeof(char)*6);
char *tok, *tempString;
tempString = malloc(sizeof(argv[1]));
tok = strtok( NULL, "\0");
while( tok ) {
strcpy(tempString, tok);
printf("request: %s\n", request);
request = strcat(tempString, request);
tok = strtok(NULL, "\0");
}
printf("host: %s\n", host);
printf("Path: %s\n", request);
Thanks for looking over this.
Any direction or even a link to a site where I can figure out how to do this would be much appreciated.

Answer 1

Here's some code that does more than you want. Note that this modifies the original string - you may want to make copies instead:
void split_request(char *request, char **protocol, char **addr, char **path)
{
char *ptr = strstr(request, "://");
if(NULL == ptr)
{
*protocol = NULL;
*addr = request;
}
else
{
*protocol = request;
*addr = ptr + 3;
*ptr = '\0';
}
ptr = strchr(*addr, '/');
if(NULL == ptr)
{
*path = NULL;
}
else
{
*path = ptr + 1;
*ptr = '\0';
}
}
Please excuse any typos/obvious errors. I'm typing this in a hurry as I have work to do :P
It should get you started though.

Answer 2

I have modified your code to work the way you are expecting
main(int argc, char *argv[])
{
char *request,*host,*req;
char *tok, *tempString;
printf("argv[1]: %s\n", argv[1] );
host = malloc(strlen(argv[1]));
strcpy(host, argv[1]);
host = strtok(host, "/");
tempString = malloc(strlen(argv[1]));
tok = strtok( NULL, "\0");
printf("sizeof(tok) %d\n",strlen(tok));
strncpy(tempString, tok,strlen(tok));
while( tok ) {
tok = strtok(NULL, "\0");
if (tok != NULL) {
strncat(tempString, tok,strlen(tok));
}else {
break;
}
}
request = tempString;
printf("host: %s\n", host);
printf("Path: %s\n", request);
}
~
Output
./tmp www.bbc.co.uk/news/world-us-canada-11893886/tmp.htmlargv[1]: www.bbc.co.uk/news/world-us-canada-11893886/tmp.html
sizeof(tok) 38
host: www.bbc.co.uk
Path: news/world-us-canada-11893886/tmp.html
bash-2.03$
~

Answer 3

Use strrchr() to find the last occurrence of '/' from the rear. You will then have a pointer to the start of 'the end of the web address' if you add one to that returned pointer.
Update
Assuming your URL does not start with http://, this aught to work
#include <stdio.h>
#include <string.h>
int main(void)
{
char url[] = "www.bbc.co.uk/news/world-us-canada-11893886";
int cnt;
char host[100];
char path[100];
char request[100];
strcpy(request, strrchr(url, '/'));
strcpy(host, url);
host[cnt = strcspn(url, "/")] = '\0';
strcpy(path, &url[cnt]);
printf("host: %s\npath: %s\nrequest: %s\n", host, path, request);
return 0;
}
Output
$ ./a.out
host: www.bbc.co.uk
path: /news/world-us-canada-11893886
request: /world-us-canada-11893886

Answer 4

strrchr() returns the LAST instance of the character. He wants the FIRST instance after any http:// string.
The answer is simple:
char *address_start = strchr(in_string+8, '/');
If it's non NULl then there you are at the first / of the path.
Why +8? Because "https://" is 8 characters long and even if there is no "http://" at the beginning, no IP or web address is less than 8 characters. Even "a.b.c.d" is 7 characters long and I don't believe an IPv4 dotted numerical notation has any legal public address with all single digits. I might be wrong though. Might be worth validating the string to check it's long enough first.
Anyway, you can always pre-validate the string to see if it begins with "http" or not to determine the offset to start searching at.