#include <string.h>
#include <ctype.h>
#include <stdlib.h>
#include <curl/curl.h>
#include <stdlib.h>

#include "local_scan.h"

#define BUFSIZE 16384

typedef struct {
		char *url;
		void *next;
		int   res;
		} url_t;

static url_t *cl_url = NULL;

static char err_msg[CURL_ERROR_SIZE];
static CURL *curl;
static char buffer[BUFSIZE];
uschar *reason;

//static void cleanup(void) __attribute__((destructor));

//-------------------------------------------------------------------

static int test_domain( char *dom )
{
uschar *r;

if ((debug_selector & D_local_scan) != 0)
	debug_printf("SURBL: Testing url %s",dom);

r = expand_string(string_sprintf("${lookup dnsdb{a=%s.multi.surbl.org}{$value}{}}",dom));

if (!r)
	{
	log_write(0,LOG_MAIN,"SURBL: Expansion error: %s",expand_string_message);
	return -1;
	}

if (!*r) return 1;

reason = string_sprintf("%s listed at multi.surbl.org (%s)",dom,r);
return 0;
}

//-------------------------------------------------------------------

static char *follow_url ( char *url )
{
CURLcode ret;
static   char *nurl;

curl_easy_setopt(curl,CURLOPT_URL,url);

ret = curl_easy_perform(curl);

if (ret && (debug_selector & D_local_scan) != 0)
	{
	debug_printf("SURBL: Error navigating %s: %s",url,err_msg);
	}

curl_easy_getinfo(curl,CURLINFO_EFFECTIVE_URL,&nurl);
return nurl;
}

//-------------------------------------------------------------------

static int process_url ( char *url )
{
char  *nurl,*durl,*dom,*s;
char  *d1,*d2,*d3;
int   r;
url_t *t;

if ((debug_selector & D_local_scan) != 0)
    debug_printf("SURBL: Processing url %s",url);

for (t=cl_url; t; t=t->next)
	{
	if (!strcmp(url,t->url))
		{
		if (t->res!=-1)
			{
			if ((debug_selector & D_local_scan) != 0)
			    debug_printf("SURBL: URL already tested");
			if (!t->res)
				reason = string_sprintf("%s listed at multi.surbl.org",url);
			return t->res;
			}
		break;
		}
	}

if (!t)
	{
	t= store_get_perm(sizeof(url_t));
	if (!t) return -1;
	t->url = store_get_perm(strlen(url)+1);
	if (!t->url) return -1;
	strcpy(t->url,url);
	t->next = cl_url;
	t->res = 1;
	cl_url = t;
	}

nurl = follow_url(url);
durl = curl_unescape(nurl,0);

if ((debug_selector & D_local_scan) != 0)
    debug_printf("SURBL: Found to be %s",durl);

if (strlen(durl)<7) return 1;
	
// Get hostname

dom = durl+7;

if (*dom=='/') dom++;

for (s=dom; *s; s++)
    {
    if (*s=='@')
        dom=s+1;
    else
        if (*s=='/' || *s=='?')
            {
            *s=0;
            break;
            }
    }

// Strip www.

if (!strncmp(dom,"www.",4)) dom += 4;

if ((debug_selector & D_local_scan) != 0)
    debug_printf("SURBL: Cleaned to %s",dom);

// Get host parts

d1=d2=d3=NULL;

for (s = dom; *s; s++)
    {
    if (*s=='.')
        {
        d3=d2;
        d2=d1;
        d1=s+1;
        }
	else
		if (*s==':')
			{
			*s=0;
			break;
			}
    }

// Check it

r = -1;

if (!d1) return 1;

if (isdigit(*d1))
	{
	if (d1 && d2 && d3)
		{
		*d1=*d2=*d3=0;
		r = test_domain(CS string_sprintf("%s.%s.%s.%s",d1+1,d2+1,d3+1,dom));
		}
	}
else
	{
	if (strlen(d1)==2 && d3)
		{
		r = test_domain(d3);
		}

	if (r!=0)
		{
		if (d2)
			r = test_domain(d2);
		else
			r = test_domain(dom);
		}
	}

t->res = r;
curl_free(durl);

if (r==0)
	return 0;
else
	return 1;
}

//-------------------------------------------------------------------

static int process_data ( void )
{
char *s,*u,c;

s = buffer;

while ((s = strstr(s,"http://"))!=NULL)
	{
	c=0;

	if (s>buffer && (*(s-1)=='"' || *(s-1)=='\''))
		{
		c=*(s-1);
		}

	u=s;

	for (; *s; s++)
		{
		if (*s=='\\') 
			{
			s++;
			continue;
			}
		if (*s==c) break;
		if (c) continue;
		if (isspace(*s) || *s=='<' || *s=='>' || *s=='"') break;
		}

	*s=0;
	s++;

	if (!process_url(u)) return 0;
	}

return 1;
}

//-------------------------------------------------------------------

static int scan_file ( uschar **yield, char *file )
{
FILE *fp;

if ((fp=fopen(file,"rt"))==NULL)
	{
	*yield = string_sprintf("Unable to open %s",file);
	return ERROR;
	}

if ((debug_selector & D_local_scan) != 0)
    debug_printf("SURBL: Testing file %s",file);

while (!feof(fp))
	{
	fgets(buffer,BUFSIZE,fp);

	if (!process_data())
		{
		if ((debug_selector & D_local_scan) != 0)
    		debug_printf("SURBL: Listed");
		*yield = reason;
		fclose(fp);
		return OK;
		}
	}

fclose(fp);
*yield = string_copy(US "0");
return OK;
}

//-------------------------------------------------------------------

void cleanup ( void )
{
if (curl)
    {
    curl_easy_cleanup(curl);
    curl = NULL;
    }

log_write(0,LOG_MAIN,"SURBL: cleanup");
}

//-------------------------------------------------------------------

int surbl(uschar **yield, int argc, uschar *argv[])
{

if (!curl)
	{
	curl=curl_easy_init();

	if (!curl)
		{
		*yield = string_copy(US "Unable to init curl");
		return ERROR;
		}

	curl_easy_setopt(curl,CURLOPT_NOPROGRESS,1);
	curl_easy_setopt(curl,CURLOPT_FAILONERROR,1);
	curl_easy_setopt(curl,CURLOPT_ERRORBUFFER,err_msg);
	curl_easy_setopt(curl,CURLOPT_NOBODY,1);
	curl_easy_setopt(curl,CURLOPT_FOLLOWLOCATION,1);
	curl_easy_setopt(curl,CURLOPT_AUTOREFERER,1);
	curl_easy_setopt(curl,CURLOPT_MAXREDIRS,5);
	curl_easy_setopt(curl,CURLOPT_USERAGENT,"Mozilla/4.0 (compatible; MSIE 6.0; Win32);");
	curl_easy_setopt(curl,CURLOPT_TIMEOUT,10);

	//atexit(cleanup);
	}

return scan_file(yield, CS argv[0]);
}