feat: changes to image proxy, proxy favicons

src/Cache/Cache.c

@@ -11,15 +11,20 @@
 static char cache_dir[BUFFER_SIZE_MEDIUM] = {0};
 static int cache_ttl_search_val = DEFAULT_CACHE_TTL_SEARCH;
 static int cache_ttl_infobox_val = DEFAULT_CACHE_TTL_INFOBOX;
+static int cache_ttl_image_val = DEFAULT_CACHE_TTL_IMAGE;
 
 void set_cache_ttl_search(int ttl) { cache_ttl_search_val = ttl; }
 
 void set_cache_ttl_infobox(int ttl) { cache_ttl_infobox_val = ttl; }
 
+void set_cache_ttl_image(int ttl) { cache_ttl_image_val = ttl; }
+
 int get_cache_ttl_search(void) { return cache_ttl_search_val; }
 
 int get_cache_ttl_infobox(void) { return cache_ttl_infobox_val; }
 
+int get_cache_ttl_image(void) { return cache_ttl_image_val; }
+
 static void md5_hash(const char *str, char *output) {
   unsigned char hash[EVP_MAX_MD_SIZE];
   unsigned int hash_len;

src/Cache/Cache.h

@@ -17,7 +17,9 @@ char *cache_compute_key(const char *query, int page, const char *engine_name);
 
 void set_cache_ttl_search(int ttl);
 void set_cache_ttl_infobox(int ttl);
+void set_cache_ttl_image(int ttl);
 int get_cache_ttl_search(void);
 int get_cache_ttl_infobox(void);
+int get_cache_ttl_image(void);
 
 #endif

src/Config.c

@@ -92,6 +92,8 @@ int load_config(const char *filename, Config *config) {
           config->cache_ttl_search = atoi(value);
         } else if (strcmp(key, "ttl_infobox") == 0) {
           config->cache_ttl_infobox = atoi(value);
+        } else if (strcmp(key, "ttl_image") == 0) {
+          config->cache_ttl_image = atoi(value);
         }
       } else if (strcmp(section, "engines") == 0) {
         if (strcmp(key, "engines") == 0) {

src/Config.h

@@ -6,6 +6,7 @@
 #define DEFAULT_CACHE_DIR "/tmp/omnisearch_cache"
 #define DEFAULT_CACHE_TTL_SEARCH 3600
 #define DEFAULT_CACHE_TTL_INFOBOX 86400
+#define DEFAULT_CACHE_TTL_IMAGE 604800
 #define DEFAULT_MAX_PROXY_RETRIES 3
 
 #define BUFFER_SIZE_SMALL 256
@@ -42,6 +43,7 @@ typedef struct {
   char cache_dir[512];
   int cache_ttl_search;
   int cache_ttl_infobox;
+  int cache_ttl_image;
   char engines[512];
 } Config;
 

src/Main.c

@@ -52,6 +52,7 @@ int main() {
                 .cache_dir = DEFAULT_CACHE_DIR,
                 .cache_ttl_search = DEFAULT_CACHE_TTL_SEARCH,
                 .cache_ttl_infobox = DEFAULT_CACHE_TTL_INFOBOX,
+                .cache_ttl_image = DEFAULT_CACHE_TTL_IMAGE,
                 .engines = ""};
 
   if (load_config("config.ini", &cfg) != 0) {
@@ -72,6 +73,7 @@ int main() {
 
   set_cache_ttl_search(cfg.cache_ttl_search);
   set_cache_ttl_infobox(cfg.cache_ttl_infobox);
+  set_cache_ttl_image(cfg.cache_ttl_image);
 
   if (cfg.proxy_list_file[0] != '\0') {
     if (load_proxy_list(cfg.proxy_list_file) < 0) {

src/Routes/ImageProxy.c

@@ -1,9 +1,16 @@
 #include "ImageProxy.h"
+#include "../Cache/Cache.h"
 #include "../Proxy/Proxy.h"
+#include <arpa/inet.h>
 #include <curl/curl.h>
+#include <curl/urlapi.h>
+#include <netdb.h>
+#include <netinet/in.h>
+#include <openssl/evp.h>
 #include <stdio.h>
 #include <stdlib.h>
 #include <string.h>
+#include <sys/socket.h>
 
 #define MAX_IMAGE_SIZE (10 * 1024 * 1024)
 
@@ -13,7 +20,97 @@ typedef struct {
   size_t capacity;
 } MemoryBuffer;
 
+static int is_private_ip(const char *ip_str) {
+  struct in_addr addr;
+  if (inet_pton(AF_INET, ip_str, &addr) != 1) {
+    return 0;
+  }
+
+  uint32_t ip = ntohl(addr.s_addr);
+
+  // 10.0.0.0/8
+  if ((ip >> 24) == 10) {
+    return 1;
+  }
+
+  // 172.16.0.0/12
+  if ((ip >> 20) == 0xAC) {
+    uint8_t second = (ip >> 16) & 0xFF;
+    if (second >= 16 && second <= 31) {
+      return 1;
+    }
+  }
+
+  // 192.168.0.0/16
+  if ((ip >> 16) == 0xC0A8) {
+    return 1;
+  }
+
+  // 127.0.0.0/8
+  if ((ip >> 24) == 127) {
+    return 1;
+  }
+
+  // 169.254.0.0/16
+  if ((ip >> 16) == 0xA9FE) {
+    return 1;
+  }
+
+  return 0;
+}
+
+static int is_private_hostname(const char *hostname) {
+  struct addrinfo hints, *res, *p;
+  memset(&hints, 0, sizeof(hints));
+  hints.ai_family = AF_INET;
+  hints.ai_socktype = SOCK_STREAM;
+
+  int err = getaddrinfo(hostname, NULL, &hints, &res);
+  if (err != 0) {
+    return 0;
+  }
+
+  for (p = res; p != NULL; p = p->ai_next) {
+    if (p->ai_family == AF_INET) {
+      struct sockaddr_in *ipv4 = (struct sockaddr_in *)p->ai_addr;
+      char ip_str[INET_ADDRSTRLEN];
+      inet_ntop(AF_INET, &(ipv4->sin_addr), ip_str, INET_ADDRSTRLEN);
+
+      if (is_private_ip(ip_str)) {
+        freeaddrinfo(res);
+        return 1;
+      }
+    }
+  }
+
+  freeaddrinfo(res);
+  return 0;
+}
+
 static int is_allowed_domain(const char *url) {
+  CURLU *h = curl_url();
+  if (!h) {
+    return -1;
+  }
+
+  curl_url_set(h, CURLUPART_URL, url, 0);
+
+  char *scheme = NULL;
+  curl_url_get(h, CURLUPART_SCHEME, &scheme, 0);
+
+  int valid_scheme = 0;
+  if (scheme && (strcasecmp(scheme, "http") == 0 || strcasecmp(scheme, "https") == 0)) {
+    valid_scheme = 1;
+  }
+
+  if (scheme)
+    curl_free(scheme);
+
+  if (!valid_scheme) {
+    curl_url_cleanup(h);
+    return -1;
+  }
+
   const char *protocol = strstr(url, "://");
   if (!protocol) {
     protocol = url;
@@ -30,23 +127,20 @@ static int is_allowed_domain(const char *url) {
   }
   strncpy(host, protocol, host_len);
 
-  const char *allowed_domains[] = {"mm.bing.net", "th.bing.com", NULL};
+  char *colon = strchr(host, ':');
-
+  if (colon) {
-  for (int i = 0; allowed_domains[i] != NULL; i++) {
+    *colon = '\0';
-    size_t domain_len = strlen(allowed_domains[i]);
-    size_t host_str_len = strlen(host);
-
-    if (host_str_len >= domain_len) {
-      const char *suffix = host + host_str_len - domain_len;
-      if (strcmp(suffix, allowed_domains[i]) == 0) {
-        return 1;
-      }
-    }
   }
 
+  if (is_private_hostname(host)) {
+    curl_url_cleanup(h);
     return 0;
   }
 
+  curl_url_cleanup(h);
+  return 1;
+}
+
 static size_t write_callback(void *contents, size_t size, size_t nmemb,
                              void *userp) {
   size_t realsize = size * nmemb;
@@ -73,6 +167,31 @@ static size_t write_callback(void *contents, size_t size, size_t nmemb,
   return realsize;
 }
 
+static char *url_encode_key(const char *url) {
+  char *hash = malloc(33);
+  if (!hash)
+    return NULL;
+
+  unsigned char md5hash[16];
+  EVP_MD_CTX *ctx = EVP_MD_CTX_new();
+  if (!ctx) {
+    free(hash);
+    return NULL;
+  }
+
+  EVP_DigestInit_ex(ctx, EVP_md5(), NULL);
+  EVP_DigestUpdate(ctx, url, strlen(url));
+  EVP_DigestFinal_ex(ctx, md5hash, NULL);
+  EVP_MD_CTX_free(ctx);
+
+  for (int i = 0; i < 16; i++) {
+    sprintf(hash + (i * 2), "%02x", md5hash[i]);
+  }
+  hash[32] = '\0';
+
+  return hash;
+}
+
 int image_proxy_handler(UrlParams *params) {
   const char *url = NULL;
   for (int i = 0; i < params->count; i++) {
@@ -87,13 +206,59 @@ int image_proxy_handler(UrlParams *params) {
     return 0;
   }
 
-  if (!is_allowed_domain(url)) {
+  int domain_check = is_allowed_domain(url);
-    send_response("Domain not allowed");
+  if (domain_check == -1) {
+    send_response("Invalid URL scheme");
+    return 0;
+  }
+  if (domain_check == 0) {
+    send_response("Private addresses are not allowed");
+    return 0;
+  }
+
+  char *cache_key = url_encode_key(url);
+  if (!cache_key) {
+    send_response("Failed to generate cache key");
+    return 0;
+  }
+
+  char *cached_data = NULL;
+  size_t cached_size = 0;
+  int cache_ttl = get_cache_ttl_image();
+
+  if (cache_get(cache_key, cache_ttl, &cached_data, &cached_size) == 0) {
+    char content_type[64] = {0};
+
+    const char *ext = strrchr(url, '.');
+    if (ext) {
+      if (strcasecmp(ext, ".png") == 0) {
+        strncpy(content_type, "image/png", sizeof(content_type) - 1);
+      } else if (strcasecmp(ext, ".gif") == 0) {
+        strncpy(content_type, "image/gif", sizeof(content_type) - 1);
+      } else if (strcasecmp(ext, ".webp") == 0) {
+        strncpy(content_type, "image/webp", sizeof(content_type) - 1);
+      } else if (strcasecmp(ext, ".svg") == 0) {
+        strncpy(content_type, "image/svg+xml", sizeof(content_type) - 1);
+      } else if (strcasecmp(ext, ".ico") == 0) {
+        strncpy(content_type, "image/x-icon", sizeof(content_type) - 1);
+      } else if (strcasecmp(ext, ".bmp") == 0) {
+        strncpy(content_type, "image/bmp", sizeof(content_type) - 1);
+      }
+    }
+
+    if (strlen(content_type) == 0) {
+      strncpy(content_type, "image/jpeg", sizeof(content_type) - 1);
+    }
+
+    serve_data(cached_data, cached_size, content_type);
+    free(cached_data);
+    free(cache_key);
     return 0;
   }
 
   CURL *curl = curl_easy_init();
   if (!curl) {
+    free(cache_key);
     send_response("Failed to initialize curl");
     return 0;
   }
@@ -102,6 +267,7 @@ int image_proxy_handler(UrlParams *params) {
 
   if (!buf.data) {
     curl_easy_cleanup(curl);
+    free(cache_key);
     send_response("Memory allocation failed");
     return 0;
   }
@@ -111,6 +277,10 @@ int image_proxy_handler(UrlParams *params) {
   curl_easy_setopt(curl, CURLOPT_WRITEDATA, &buf);
   curl_easy_setopt(curl, CURLOPT_FOLLOWLOCATION, 1L);
   curl_easy_setopt(curl, CURLOPT_TIMEOUT, 10L);
+  curl_easy_setopt(curl, CURLOPT_USERAGENT,
+                   "Mozilla/5.0 (Windows NT 10.0; Win64; x64) "
+                   "AppleWebKit/537.36 (KHTML, like Gecko) "
+                   "Chrome/120.0.0.0 Safari/537.36");
   apply_proxy_settings(curl);
 
   CURLcode res = curl_easy_perform(curl);
@@ -130,14 +300,27 @@ int image_proxy_handler(UrlParams *params) {
 
   if (res != CURLE_OK || response_code != 200) {
     free(buf.data);
+    free(cache_key);
     send_response("Failed to fetch image");
     return 0;
   }
 
+  if (strlen(content_type) == 0 ||
+      strncmp(content_type, "image/", 6) != 0) {
+    free(buf.data);
+    free(cache_key);
+    send_response("Invalid content type");
+    return 0;
+  }
+
   const char *mime_type =
       strlen(content_type) > 0 ? content_type : "image/jpeg";
+
+  cache_set(cache_key, buf.data, buf.size);
+
   serve_data(buf.data, buf.size, mime_type);
 
   free(buf.data);
+  free(cache_key);
   return 0;
 }

templates/results.html

@@ -68,7 +68,7 @@
                 <div class="result">
                     <div class="result-header">
                         <div class="result-favicon"
-                            style="background-image: url('https://{{result[4]}}/favicon.ico'), url('https://{{result[4]}}/favicon.png');">
+                            style="background-image: url('/proxy?url=https://{{result[4]}}/favicon.ico'), url('/proxy?url=https://{{result[4]}}/favicon.png');">
                         </div>
                         <span class="url">
                             {{result[1]}}