noderb-http 0.0.1 → 0.0.3

data/ext/noderb_http_extension/http_parser.h

@@ -28,7 +28,7 @@ extern "C" {
 #define HTTP_PARSER_VERSION_MINOR 0
 
 #include <sys/types.h>
-#if defined(_WIN32) && !defined(__MINGW32__)
+#if defined(_WIN32) && !defined(__MINGW32__) && !defined(_MSC_VER)
 typedef __int8 int8_t;
 typedef unsigned __int8 uint8_t;
 typedef __int16 int16_t;
@@ -51,6 +51,13 @@ typedef int ssize_t;
 # define HTTP_PARSER_STRICT 1
 #endif
 
+/* Compile with -DHTTP_PARSER_DEBUG=1 to add extra debugging information to
+ * the error reporting facility.
+ */
+#ifndef HTTP_PARSER_DEBUG
+# define HTTP_PARSER_DEBUG 0
+#endif
+
 
 /* Maximium header size allowed */
 #define HTTP_MAX_HEADER_SIZE (80*1024)
@@ -58,6 +65,7 @@ typedef int ssize_t;
 
 typedef struct http_parser http_parser;
 typedef struct http_parser_settings http_parser_settings;
+typedef struct http_parser_result http_parser_result;
 
 
 /* Callbacks should return non-zero to indicate an error. The parser will
@@ -125,6 +133,72 @@ enum flags
   };
 
 
+/* Map for errno-related constants
+ * 
+ * The provided argument should be a macro that takes 2 arguments.
+ */
+#define HTTP_ERRNO_MAP(XX)                                           \
+  /* No error */                                                     \
+  XX(OK, "success")                                                  \
+                                                                     \
+  /* Callback-related errors */                                      \
+  XX(CB_message_begin, "the on_message_begin callback failed")       \
+  XX(CB_path, "the on_path callback failed")                         \
+  XX(CB_query_string, "the on_query_string callback failed")         \
+  XX(CB_url, "the on_url callback failed")                           \
+  XX(CB_fragment, "the on_fragment callback failed")                 \
+  XX(CB_header_field, "the on_header_field callback failed")         \
+  XX(CB_header_value, "the on_header_value callback failed")         \
+  XX(CB_headers_complete, "the on_headers_complete callback failed") \
+  XX(CB_body, "the on_body callback failed")                         \
+  XX(CB_message_complete, "the on_message_complete callback failed") \
+                                                                     \
+  /* Parsing-related errors */                                       \
+  XX(INVALID_EOF_STATE, "stream ended at an unexpected time")        \
+  XX(HEADER_OVERFLOW,                                                \
+     "too many header bytes seen; overflow detected")                \
+  XX(CLOSED_CONNECTION,                                              \
+     "data received after completed connection: close message")      \
+  XX(INVALID_VERSION, "invalid HTTP version")                        \
+  XX(INVALID_STATUS, "invalid HTTP status code")                     \
+  XX(INVALID_METHOD, "invalid HTTP method")                          \
+  XX(INVALID_URL, "invalid URL")                                     \
+  XX(INVALID_HOST, "invalid host")                                   \
+  XX(INVALID_PORT, "invalid port")                                   \
+  XX(INVALID_PATH, "invalid path")                                   \
+  XX(INVALID_QUERY_STRING, "invalid query string")                   \
+  XX(INVALID_FRAGMENT, "invalid fragment")                           \
+  XX(LF_EXPECTED, "LF character expected")                           \
+  XX(INVALID_HEADER_TOKEN, "invalid character in header")            \
+  XX(INVALID_CONTENT_LENGTH,                                         \
+     "invalid character in content-length header")                   \
+  XX(INVALID_CHUNK_SIZE,                                             \
+     "invalid character in chunk size header")                       \
+  XX(INVALID_CONSTANT, "invalid constant string")                    \
+  XX(INVALID_INTERNAL_STATE, "encountered unexpected internal state")\
+  XX(STRICT, "strict mode assertion failed")                         \
+  XX(UNKNOWN, "an unknown error occurred")
+
+
+/* Define HPE_* values for each errno value above */
+#define HTTP_ERRNO_GEN(n, s) HPE_##n,
+enum http_errno {
+  HTTP_ERRNO_MAP(HTTP_ERRNO_GEN)
+};
+#undef HTTP_ERRNO_GEN
+
+
+/* Get an http_errno value from an http_parser */
+#define HTTP_PARSER_ERRNO(p)            ((enum http_errno) (p)->http_errno)
+
+/* Get the line number that generated the current error */
+#if HTTP_PARSER_DEBUG
+#define HTTP_PARSER_ERRNO_LINE(p)       ((p)->error_lineno)
+#else
+#define HTTP_PARSER_ERRNO_LINE(p)       0
+#endif
+
+
 struct http_parser {
   /** PRIVATE **/
   unsigned char type : 2;
@@ -141,13 +215,18 @@ struct http_parser {
   unsigned short http_minor;
   unsigned short status_code; /* responses only */
   unsigned char method;    /* requests only */
+  unsigned char http_errno : 7;
 
   /* 1 = Upgrade header was present and the parser has exited because of that.
    * 0 = No upgrade header present.
    * Should be checked when http_parser_execute() returns in addition to
    * error checking.
    */
-  char upgrade;
+  char upgrade : 1;
+
+#if HTTP_PARSER_DEBUG
+  uint32_t error_lineno;
+#endif
 
   /** PUBLIC **/
   void *data; /* A pointer to get hook to the "connection" or "socket" object */
@@ -156,10 +235,7 @@ struct http_parser {
 
 struct http_parser_settings {
   http_cb      on_message_begin;
-  http_data_cb on_path;
-  http_data_cb on_query_string;
   http_data_cb on_url;
-  http_data_cb on_fragment;
   http_data_cb on_header_field;
   http_data_cb on_header_value;
   http_cb      on_headers_complete;
@@ -186,7 +262,13 @@ size_t http_parser_execute(http_parser *parser,
 int http_should_keep_alive(http_parser *parser);
 
 /* Returns a string version of the HTTP method. */
-const char *http_method_str(enum http_method);
+const char *http_method_str(enum http_method m);
+
+/* Return a string name of the given error */
+const char *http_errno_name(enum http_errno err);
+
+/* Return a string description of the given error */
+const char *http_errno_description(enum http_errno err);
 
 #ifdef __cplusplus
 }

data/ext/noderb_http_extension/noderb_http.c

@@ -2,9 +2,11 @@
 
 typedef struct {
     long parser;
+    int pre_headers;
 } nodeRb_http;
 
 VALUE nodeRbHttpParser;
+VALUE nodeRbHttpPointer;
 
 VALUE nodeRb_get_object_from_id(long id) {
     return rb_funcall(rb_const_get(rb_cObject, rb_intern("ObjectSpace")), rb_intern("_id2ref"), 1, rb_int2inum(id));
@@ -17,101 +19,89 @@ int nodeRb_http_on_message_begin(http_parser* parser) {
     return 0;
 }
 
-int nodeRb_http_on_message_complete(http_parser* parser) {
-    nodeRb_http* client = parser->data;
-    VALUE self = nodeRb_get_object_from_id(client->parser);
-    if (http_should_keep_alive(parser)) {
-        rb_funcall(self, rb_intern("on_close_keep_alive"), 0);
-    };
-    rb_funcall(self, rb_intern("on_message_complete"), 0);
-    return 0;
-}
-
-int nodeRb_http_on_headers_complete(http_parser* parser) {
+int nodeRb_http_on_url(http_parser* parser, const char *buf, size_t len) {
     nodeRb_http* client = parser->data;
     VALUE self = nodeRb_get_object_from_id(client->parser);
-    rb_funcall(self, rb_intern("on_headers_complete_internal"), 0);
+    if(parser->type == HTTP_REQUEST){
+        rb_funcall(self, rb_intern("on_method"), 1, rb_str_new2(http_method_str(parser->method)));
+    }
+    rb_funcall(self, rb_intern("on_url"), 1, rb_str_new(buf, len));
     return 0;
 }
 
 int nodeRb_http_on_header_field(http_parser* parser, const char *buf, size_t len) {
     nodeRb_http* client = parser->data;
     VALUE self = nodeRb_get_object_from_id(client->parser);
-    rb_funcall(self, rb_intern("on_header_field_internal"), 1, rb_str_new(buf, len));
+    rb_funcall(self, rb_intern("on_header_field"), 1, rb_str_new(buf, len));
     return 0;
 }
 
 int nodeRb_http_on_header_value(http_parser* parser, const char *buf, size_t len) {
     nodeRb_http* client = parser->data;
     VALUE self = nodeRb_get_object_from_id(client->parser);
-    rb_funcall(self, rb_intern("on_header_value_internal"), 1, rb_str_new(buf, len));
+    rb_funcall(self, rb_intern("on_header_value"), 1, rb_str_new(buf, len));
     return 0;
 }
 
-int nodeRb_http_on_path(http_parser* parser, const char *buf, size_t len) {
-    nodeRb_http* client = parser->data;
-    VALUE self = nodeRb_get_object_from_id(client->parser);
-    rb_funcall(self, rb_intern("on_method"), 1, rb_str_new2(http_method_str(parser->method)));
-    rb_funcall(self, rb_intern("on_path"), 1, rb_str_new(buf, len));
-    return 0;
-}
-
-int nodeRb_http_on_query_string(http_parser* parser, const char *buf, size_t len) {
-    nodeRb_http* client = parser->data;
-    VALUE self = nodeRb_get_object_from_id(client->parser);
-    rb_funcall(self, rb_intern("on_query_string"), 1, rb_str_new(buf, len));
-    return 0;
-}
-
-int nodeRb_http_on_url(http_parser* parser, const char *buf, size_t len) {
+int nodeRb_http_on_headers_complete(http_parser* parser) {
     nodeRb_http* client = parser->data;
     VALUE self = nodeRb_get_object_from_id(client->parser);
-    rb_funcall(self, rb_intern("on_url"), 1, rb_str_new(buf, len));
+    rb_funcall(self, rb_intern("on_version"), 2, rb_int2inum(parser->http_major), rb_int2inum(parser->http_minor));
+    if(parser->type == HTTP_RESPONSE){
+        rb_funcall(self, rb_intern("on_status_code"), 1, rb_int2inum(parser->status_code));
+    }
+    if (http_should_keep_alive(parser)) {       
+        rb_funcall(self, rb_intern("on_keep_alive"), 1, Qtrue);
+    }
+    rb_funcall(self, rb_intern("on_headers_complete"), 0);
     return 0;
 }
 
-int nodeRb_http_on_fragment(http_parser* parser, const char *buf, size_t len) {
+int nodeRb_http_on_body(http_parser* parser, const char *buf, size_t len) {
     nodeRb_http* client = parser->data;
     VALUE self = nodeRb_get_object_from_id(client->parser);
-    rb_funcall(self, rb_intern("on_fragment"), 1, rb_str_new(buf, len));
+    rb_funcall(self, rb_intern("on_body"), 1, rb_str_new(buf, len));
     return 0;
 }
 
-int nodeRb_http_on_body(http_parser* parser, const char *buf, size_t len) {
+int nodeRb_http_on_message_complete(http_parser* parser) {
     nodeRb_http* client = parser->data;
     VALUE self = nodeRb_get_object_from_id(client->parser);
-    rb_funcall(self, rb_intern("on_body"), 1, rb_str_new(buf, len));
+    rb_funcall(self, rb_intern("on_message_complete"), 0);
     return 0;
 }
 
-VALUE nodeRb_http_setup(VALUE self) {
+VALUE nodeRb_http_setup(VALUE self, VALUE type) {
     http_parser_settings* settings = malloc(sizeof (http_parser_settings));
 
     settings->on_message_begin = nodeRb_http_on_message_begin;
-    settings->on_message_complete = nodeRb_http_on_message_complete;
-    settings->on_headers_complete = nodeRb_http_on_headers_complete;
-
-    settings->on_path = nodeRb_http_on_path;
-    settings->on_query_string = nodeRb_http_on_query_string;
+    
     settings->on_url = nodeRb_http_on_url;
 
     settings->on_header_field = nodeRb_http_on_header_field;
     settings->on_header_value = nodeRb_http_on_header_value;
 
-    settings->on_fragment = nodeRb_http_on_fragment;
+    settings->on_headers_complete = nodeRb_http_on_headers_complete;
+        
     settings->on_body = nodeRb_http_on_body;
 
+    settings->on_message_complete = nodeRb_http_on_message_complete;
 
     http_parser* parser = malloc(sizeof (http_parser));
-    http_parser_init(parser, HTTP_REQUEST);
+    
+    if(ID2SYM(rb_intern("response")) == type){
+        http_parser_init(parser, HTTP_RESPONSE);
+    }else{
+        http_parser_init(parser, HTTP_REQUEST);
+    }   
 
     nodeRb_http* client = malloc(sizeof (nodeRb_http));
     parser->data = client;
 
     client->parser = rb_num2long(rb_obj_id(self));
 
-    rb_iv_set(self, "@settings", Data_Wrap_Struct(nodeRbHttpParser, 0, NULL, settings));
-    rb_iv_set(self, "@parser", Data_Wrap_Struct(nodeRbHttpParser, 0, NULL, parser));
+    rb_iv_set(self, "@settings", Data_Wrap_Struct(nodeRbHttpPointer, 0, NULL, settings));
+    rb_iv_set(self, "@parser", Data_Wrap_Struct(nodeRbHttpPointer, 0, NULL, parser));
 };
 
 VALUE nodeRb_http_parse(VALUE self, VALUE data) {
@@ -132,8 +122,8 @@ VALUE nodeRb_http_parse(VALUE self, VALUE data) {
 
     if (parser->upgrade) {
         rb_funcall(self, rb_intern("on_upgrade"), 0);
-    } else if (parsed != RSTRING_LEN(data)) {
-        rb_funcall(self, rb_intern("on_error"), 0);
+    } else if (parsed != (unsigned) RSTRING_LEN(data)) {
+        rb_funcall(self, rb_intern("on_error"), 2, rb_str_new2(http_errno_name(parser->http_errno)), rb_str_new2(http_errno_description(parser->http_errno)));
     };
 };
 
@@ -149,7 +139,7 @@ VALUE nodeRb_http_dispose(VALUE self) {
     free(parser);
 };
 
-void Init_noderb_http() {
+void Init_noderb_http_extension() {
     // Define module
     VALUE nodeRb = rb_define_module("NodeRb");
     // Modules
@@ -157,9 +147,10 @@ void Init_noderb_http() {
     // Http
     VALUE nodeRbHttp = rb_define_module_under(nodeRbModules, "Http");
     // Http parser
-    nodeRbHttpParser = rb_define_class_under(nodeRbHttp, "Parser", rb_cObject);
+    nodeRbHttpPointer = rb_define_class_under(nodeRbHttp, "Pointer", rb_cObject);
+    nodeRbHttpParser = rb_define_module_under(nodeRbHttp, "Parser");
     // Methods
-    rb_define_method(nodeRbHttpParser, "setup", nodeRb_http_setup, 0);
+    rb_define_method(nodeRbHttpParser, "setup", nodeRb_http_setup, 1);
     rb_define_method(nodeRbHttpParser, "parse", nodeRb_http_parse, 1);
     rb_define_method(nodeRbHttpParser, "dispose", nodeRb_http_dispose, 0);
 }

data/lib/noderb/modules/http/parser.rb

@@ -0,0 +1,112 @@
+module NodeRb
+  module Modules
+    module Http
+
+      module Parser
+
+        def on_message_begin
+          @_header_name = ""
+          @_header_value = ""
+          @headers = {}
+          @body = ""
+          @method = ""
+          @url = ""
+          @status_code = 0
+          @upgrade = false
+          @keep_alive = false
+          @_active = true
+        end
+        
+        def on_method method
+          @method = method
+        end
+
+        def on_url url
+          @url = url
+        end
+
+        def on_header_field name
+          if @_header_state == :value
+            on_header(@_header_name, @_header_value)
+            @_header_name = ""
+            @_header_value = ""
+          end
+          @_header_state = :field
+          @_header_name << name
+        end
+
+        def on_header_value value
+          @_header_state = :value
+          @_header_value ||= ""
+          @_header_value << value
+        end
+
+        def on_header name, value
+          @headers[name] = value if @_active
+        end
+        
+        def on_version major, minor
+          @version_major = major if @_active
+          @version_minor = minor if @_active
+        end
+        
+        def on_status_code status_code
+          @status_code = status_code if @_active
+        end
+        
+        def on_keep_alive keep_alive
+          @keep_alive = keep_alive if @_active
+        end
+
+        def on_headers_complete
+          on_header(@_header_name, @_header_value)
+          @_header_name = ""
+          @_header_value = ""
+          on_message_header
+        end
+
+        def on_upgrade
+          @_active = false
+          @upgrade = true
+        end
+
+        def on_body body
+          @body << body if @_active
+        end
+        
+        def on_message_complete
+          on_message_body
+        end
+        
+        def on_error name, description
+          @_active = true
+          on_message_error(name, description)
+        end
+        
+        def on_message_header
+          # User overrides this to get whole message header
+          # ------
+          # @headers => Hash
+          # @url => String
+          # @version_major => Integer
+          # @version_minor => Integer
+          # @status_code (response) => Integer
+          # @method (request) => String
+          # @upgrade => Boolean
+          # @keep_alive => Boolean
+        end
+        
+        def on_message_body
+          # User overrides this to get whole message body
+          # @body => String
+        end
+
+        def on_message_error name, description
+          # User overrides this to be notified of errors
+        end
+
+      end
+
+    end
+  end
+end

data/lib/noderb/modules/http/version.rb

@@ -0,0 +1,9 @@
+module NodeRb
+  module Modules
+    module Http
+      
+      VERSION = "0.0.3" unless const_defined?(:VERSION)
+    
+    end
+  end
+end

data/lib/noderb/modules/http.rb

@@ -1,99 +1,2 @@
-module NodeRb
-  module Modules
-    module Http
-
-      module Rack
-
-        def initialize
-          @env = {}
-        end
-
-        def on_header name, value
-        end
-
-      end
-
-      class Parser
-
-        def on_message_begin
-        end
-
-        def on_method method
-        end
-
-        def on_path path
-        end
-
-        def on_query_string query_string
-        end
-
-        def on_url url
-        end
-
-        def on_fragment fragment
-        end
-
-        def on_header_field_internal name
-          if @_header_state == :value
-            on_header_value(@_header_value)
-            on_header(@_header_name, @_header_value)
-            @_header_name = nil
-            @_header_value = nil
-          end
-          @_header_state = :field
-          @_header_name ||= ""
-          @_header_name << name
-        end
-
-        def on_header_field name
-          
-        end
-
-        def on_header_value_internal value
-          if @_header_state == :field
-            on_header_field(@_header_name)
-          end
-          @_header_state = :value
-          @_header_value ||= ""
-          @_header_value << value
-        end
-
-        def on_header_value value
-
-        end
-
-        def on_header name, value
-        end
-
-        def on_headers_complete_internal
-          on_header_value(@_header_value)
-          on_header(@_header_name, @_header_value)
-          @_header_name = nil
-          @_header_value = nil
-          on_headers_complete
-        end
-
-        def on_headers_complete
-          
-        end
-
-        def on_close_keep_alive
-        end
-
-        def on_upgrade
-        end
-
-        def on_body body
-        end
-
-        def on_message_complete
-        end
-
-        def on_error
-        end
-
-      end
-
-    end
-  end
-end
+require "noderb/modules/http/version"
+require "noderb/modules/http/parser"

data/lib/noderb-http.rb

@@ -1,2 +1,2 @@
-require 'noderb_http_extension/noderb_http'
-require "noderb/modules/http"
+require 'noderb_http_extension'
+require 'noderb/modules/http'

data/test/cases/curl_get.rb

@@ -0,0 +1,25 @@
+CASES << {
+    :name => "curl_get",
+    :type => :request,
+    :message_complete_on_eof => false,
+    :data => [
+      "GET /test HTTP/1.1\n",
+      "User-Agent: curl/7.18.0 (i486-pc-linux-gnu) libcurl/7.18.0 OpenSSL/0.9.8g zlib/1.2.3.3 libidn/1.1\n",
+      "Host: 0.0.0.0=5000\n",
+      "Accept: */*\n",
+      "\n"
+    ],    
+    :should_keep_alive => true,
+    :http_major => 1,
+    :http_minor => 1,
+    :method => "GET",
+    :request_url => "/test",
+    :num_headers => 3,
+    :headers => { 
+        "User-Agent" => "curl/7.18.0 (i486-pc-linux-gnu) libcurl/7.18.0 OpenSSL/0.9.8g zlib/1.2.3.3 libidn/1.1",
+        "Host" => "0.0.0.0=5000",
+        "Accept" => "*/*",
+    },
+    :body => "",
+    :upgrade => false
+}

data/test/cases/patch_request.rb

@@ -0,0 +1,29 @@
+CASES << {
+    :name => "patch_request",
+    :type => :request,
+    :message_complete_on_eof => false,
+    :data => [
+      "PATCH /file.txt HTTP/1.1\n",
+      "Host: www.example.com\n",
+      "Content-Type: application/example\n",
+      "If-Match: \"e0023aa4e\"\n",
+      "Content-Length: 10\n",
+      "\n",
+      "abcde",
+      "efghi"
+    ],
+    :should_keep_alive => true,
+    :http_major => 1,
+    :http_minor => 1,
+    :method => "PATCH",
+    :request_url => "/file.txt",
+    :num_headers => 4,
+    :headers => { 
+        "Host" => "www.example.com",
+        "Content-Length" => "10",
+        "If-Match" => "\"e0023aa4e\"",
+        "Content-Type" => "application/example"
+    },
+    :body => "abcdeefghi",
+    :upgrade => false
+}

data/test/cases/response.rb

@@ -0,0 +1,38 @@
+CASES << {
+    :name => "response",
+    :type => :response,
+    :message_complete_on_eof => false,
+    :data => [
+      "HTTP/1.0 301 Moved Permanently\n",
+      "Date: Thu, 03 Jun 2010 09:56:32 GMT\n",
+       "Server: Apache/2.2.3 (Red Hat)\n",
+      "Cache-Control: public\n",
+      "Pragma: \n",
+      "Location: http://www.bonjourmadame.fr/\n",
+      "Vary: Accept-Encoding\n",
+      "Content-Length: 0\n",
+      "Content-Type: text/html; charset=UTF-8\n",
+      "Connection: keep-alive\n",
+      "\n"
+    ],    
+    :should_keep_alive => true,
+    :http_major => 1,
+    :http_minor => 0,
+    :method => "",
+    :request_url => "",
+    :num_headers => 9,
+    :status_code => 301,
+    :headers => { 
+      "Date" => "Thu, 03 Jun 2010 09:56:32 GMT",
+      "Server" => "Apache/2.2.3 (Red Hat)",
+      "Cache-Control" => "public",
+      "Pragma" => "",
+      "Location" => "http://www.bonjourmadame.fr/",
+      "Vary" => "Accept-Encoding",
+      "Content-Length" => "0",
+      "Content-Type" => "text/html; charset=UTF-8",
+      "Connection" => "keep-alive"
+    },
+    :upgrade => false,
+    :body => ""
+}