src: support IPv6 Zone ID as per RFC 6874

IPv6 scoped address contains Zone ID, which is appended to IPv6
address after '%' separator.  RFC 6874 says that Zone ID after '%'
must consist of 1*(unreserved or pct-encoded).  This commit adds this
IPv6 Zone ID support.

PR-URL: #253
Reviewed-By: Fedor Indutny <[email protected]>

http_parser.c

@@ -400,6 +400,8 @@ enum http_host_state
   , s_http_host
   , s_http_host_v6
   , s_http_host_v6_end
+  , s_http_host_v6_zone_start
+  , s_http_host_v6_zone
   , s_http_host_port_start
   , s_http_host_port
 };
@@ -2211,6 +2213,23 @@ http_parse_host_char(enum http_host_state s, const char ch) {
         return s_http_host_v6;
       }
 
+      if (s == s_http_host_v6 && ch == '%') {
+        return s_http_host_v6_zone_start;
+      }
+      break;
+
+    case s_http_host_v6_zone:
+      if (ch == ']') {
+        return s_http_host_v6_end;
+      }
+
+    /* FALLTHROUGH */
+    case s_http_host_v6_zone_start:
+      /* RFC 6874 Zone ID consists of 1*( unreserved / pct-encoded) */
+      if (IS_ALPHANUM(ch) || ch == '%' || ch == '.' || ch == '-' || ch == '_' ||
+          ch == '~') {
+        return s_http_host_v6_zone;
+      }
       break;
 
     case s_http_host_port:
@@ -2261,6 +2280,11 @@ http_parse_host(const char * buf, struct http_parser_url *u, int found_at) {
         u->field_data[UF_HOST].len++;
         break;
 
+      case s_http_host_v6_zone_start:
+      case s_http_host_v6_zone:
+        u->field_data[UF_HOST].len++;
+        break;
+
       case s_http_host_port:
         if (s != s_http_host_port) {
           u->field_data[UF_PORT].off = p - buf;
@@ -2290,6 +2314,8 @@ http_parse_host(const char * buf, struct http_parser_url *u, int found_at) {
     case s_http_host_start:
     case s_http_host_v6_start:
     case s_http_host_v6:
+    case s_http_host_v6_zone_start:
+    case s_http_host_v6_zone:
     case s_http_host_port_start:
     case s_http_userinfo:
     case s_http_userinfo_start:

test.c

@@ -2918,6 +2918,59 @@ const struct url_test url_tests[] =
   ,.rv=1 /* s_dead */
   }
 
+, {.name="ipv6 address with Zone ID"
+  ,.url="http://[fe80::a%25eth0]/"
+  ,.is_connect=0
+  ,.u=
+    {.field_set= (1<<UF_SCHEMA) | (1<<UF_HOST) | (1<<UF_PATH)
+    ,.port=0
+    ,.field_data=
+      {{  0,  4 } /* UF_SCHEMA */
+      ,{  8, 14 } /* UF_HOST */
+      ,{  0,  0 } /* UF_PORT */
+      ,{ 23,  1 } /* UF_PATH */
+      ,{  0,  0 } /* UF_QUERY */
+      ,{  0,  0 } /* UF_FRAGMENT */
+      ,{  0,  0 } /* UF_USERINFO */
+      }
+    }
+  ,.rv=0
+  }
+
+, {.name="ipv6 address with Zone ID, but '%' is not percent-encoded"
+  ,.url="http://[fe80::a%eth0]/"
+  ,.is_connect=0
+  ,.u=
+    {.field_set= (1<<UF_SCHEMA) | (1<<UF_HOST) | (1<<UF_PATH)
+    ,.port=0
+    ,.field_data=
+      {{  0,  4 } /* UF_SCHEMA */
+      ,{  8, 12 } /* UF_HOST */
+      ,{  0,  0 } /* UF_PORT */
+      ,{ 21,  1 } /* UF_PATH */
+      ,{  0,  0 } /* UF_QUERY */
+      ,{  0,  0 } /* UF_FRAGMENT */
+      ,{  0,  0 } /* UF_USERINFO */
+      }
+    }
+  ,.rv=0
+  }
+
+, {.name="ipv6 address ending with '%'"
+  ,.url="http://[fe80::a%]/"
+  ,.rv=1 /* s_dead */
+  }
+
+, {.name="ipv6 address with Zone ID including bad character"
+  ,.url="http://[fe80::a%$HOME]/"
+  ,.rv=1 /* s_dead */
+  }
+
+, {.name="just ipv6 Zone ID"
+  ,.url="http://[%eth0]/"
+  ,.rv=1 /* s_dead */
+  }
+
 #if HTTP_PARSER_STRICT
 
 , {.name="tab in URL"