1 files changed, 132 insertions, 208 deletions

diff --git a/ext/unicorn/http11/http11.c b/ext/unicorn/http11/http11.c
index d5c364a..cd7a8f7 100644
--- a/ext/unicorn/http11/http11.c
+++ b/ext/unicorn/http11/http11.c
@@ -1,4 +1,5 @@
 /**
+ * Copyright (c) 2009 Eric Wong (all bugs are Eric's fault)
  * Copyright (c) 2005 Zed A. Shaw
  * You can redistribute it and/or modify it under the same terms as Ruby.
  */
@@ -8,6 +9,16 @@
 #include <string.h>
 #include "http11_parser.h"
 
+static http_parser *data_get(VALUE self)
+{
+  http_parser *http;
+
+  Data_Get_Struct(self, http_parser, http);
+  if (!http)
+    rb_raise(rb_eArgError, "NULL found for http when shouldn't be.");
+  return http;
+}
+
 #ifndef RSTRING_PTR
 #define RSTRING_PTR(s) (RSTRING(s)->ptr)
 #endif
@@ -18,40 +29,50 @@
 static VALUE mUnicorn;
 static VALUE cHttpParser;
 static VALUE eHttpParserError;
+static VALUE sym_http_body;
 
 #define HTTP_PREFIX "HTTP_"
 #define HTTP_PREFIX_LEN (sizeof(HTTP_PREFIX) - 1)
 
+static VALUE global_rack_url_scheme;
 static VALUE global_request_method;
 static VALUE global_request_uri;
 static VALUE global_fragment;
 static VALUE global_query_string;
 static VALUE global_http_version;
-static VALUE global_content_length;
-static VALUE global_http_content_length;
 static VALUE global_request_path;
-static VALUE global_content_type;
-static VALUE global_http_content_type;
-static VALUE global_http_body;
-static VALUE global_gateway_interface;
-static VALUE global_gateway_interface_value;
+static VALUE global_path_info;
 static VALUE global_server_name;
 static VALUE global_server_port;
 static VALUE global_server_protocol;
 static VALUE global_server_protocol_value;
 static VALUE global_http_host;
+static VALUE global_http_x_forwarded_proto;
 static VALUE global_port_80;
+static VALUE global_port_443;
 static VALUE global_localhost;
+static VALUE global_http;
 
 /** Defines common length and error messages for input length validation. */
-#define DEF_MAX_LENGTH(N,length) const size_t MAX_##N##_LENGTH = length; const char *MAX_##N##_LENGTH_ERR = "HTTP element " # N  " is longer than the " # length " allowed length."
+#define DEF_MAX_LENGTH(N, length) \
+  static const size_t MAX_##N##_LENGTH = length; \
+  static const char * const MAX_##N##_LENGTH_ERR = \
+    "HTTP element " # N  " is longer than the " # length " allowed length."
 
-/** Validates the max length of given input and throws an HttpParserError exception if over. */
-#define VALIDATE_MAX_LENGTH(len, N) if(len > MAX_##N##_LENGTH) { rb_raise(eHttpParserError, MAX_##N##_LENGTH_ERR); }
+/**
+ * Validates the max length of given input and throws an HttpParserError
+ * exception if over.
+ */
+#define VALIDATE_MAX_LENGTH(len, N) do { \
+  if (len > MAX_##N##_LENGTH) \
+    rb_raise(eHttpParserError, MAX_##N##_LENGTH_ERR); \
+} while (0)
 
 /** Defines global strings in the init method. */
-#define DEF_GLOBAL(N, val)   global_##N = rb_obj_freeze(rb_str_new2(val)); rb_global_variable(&global_##N)
-
+#define DEF_GLOBAL(N, val) do { \
+  global_##N = rb_obj_freeze(rb_str_new(val, sizeof(val) - 1)); \
+  rb_global_variable(&global_##N); \
+} while (0)
 
 /* Defines the maximum allowed lengths for various input elements.*/
 DEF_MAX_LENGTH(FIELD_NAME, 256);
@@ -109,26 +130,13 @@ static struct common_field common_http_fields[] = {
         f("USER_AGENT"),
         f("VIA"),
         f("X_FORWARDED_FOR"), /* common for proxies */
+        f("X_FORWARDED_PROTO"), /* common for proxies */
         f("X_REAL_IP"), /* common for proxies */
         f("WARNING")
 # undef f
 };
 
-/*
- * qsort(3) and bsearch(3) improve average performance slightly, but may
- * not be worth it for lack of portability to certain platforms...
- */
-#if defined(HAVE_QSORT_BSEARCH)
-/* sort by length, then by name if there's a tie */
-static int common_field_cmp(const void *a, const void *b)
-{
-  struct common_field *cfa = (struct common_field *)a;
-  struct common_field *cfb = (struct common_field *)b;
-  signed long diff = cfa->len - cfb->len;
-  return diff ? diff : memcmp(cfa->name, cfb->name, cfa->len);
-}
-#endif /* HAVE_QSORT_BSEARCH */
-
+/* this function is not performance-critical */
 static void init_common_fields(void)
 {
   int i;
@@ -137,32 +145,21 @@ static void init_common_fields(void)
   memcpy(tmp, HTTP_PREFIX, HTTP_PREFIX_LEN);
 
   for(i = 0; i < ARRAY_SIZE(common_http_fields); cf++, i++) {
-    memcpy(tmp + HTTP_PREFIX_LEN, cf->name, cf->len + 1);
-    cf->value = rb_obj_freeze(rb_str_new(tmp, HTTP_PREFIX_LEN + cf->len));
+    /* Rack doesn't like certain headers prefixed with "HTTP_" */
+    if (!strcmp("CONTENT_LENGTH", cf->name) ||
+        !strcmp("CONTENT_TYPE", cf->name)) {
+      cf->value = rb_str_new(cf->name, cf->len);
+    } else {
+      memcpy(tmp + HTTP_PREFIX_LEN, cf->name, cf->len + 1);
+      cf->value = rb_str_new(tmp, HTTP_PREFIX_LEN + cf->len);
+    }
+    cf->value = rb_obj_freeze(cf->value);
     rb_global_variable(&cf->value);
   }
-
-#if defined(HAVE_QSORT_BSEARCH)
-  qsort(common_http_fields,
-        ARRAY_SIZE(common_http_fields),
-        sizeof(struct common_field),
-        common_field_cmp);
-#endif /* HAVE_QSORT_BSEARCH */
 }
 
 static VALUE find_common_field_value(const char *field, size_t flen)
 {
-#if defined(HAVE_QSORT_BSEARCH)
-  struct common_field key;
-  struct common_field *found;
-  key.name = field;
-  key.len = (signed long)flen;
-  found = (struct common_field *)bsearch(&key, common_http_fields,
-                                         ARRAY_SIZE(common_http_fields),
-                                         sizeof(struct common_field),
-                                         common_field_cmp);
-  return found ? found->value : Qnil;
-#else /* !HAVE_QSORT_BSEARCH */
   int i;
   struct common_field *cf = common_http_fields;
   for(i = 0; i < ARRAY_SIZE(common_http_fields); i++, cf++) {
@@ -170,21 +167,17 @@ static VALUE find_common_field_value(const char *field, size_t flen)
       return cf->value;
   }
   return Qnil;
-#endif /* !HAVE_QSORT_BSEARCH */
 }
 
 static void http_field(void *data, const char *field,
                        size_t flen, const char *value, size_t vlen)
 {
   VALUE req = (VALUE)data;
-  VALUE v = Qnil;
   VALUE f = Qnil;
 
   VALIDATE_MAX_LENGTH(flen, FIELD_NAME);
   VALIDATE_MAX_LENGTH(vlen, FIELD_VALUE);
 
-  v = rb_str_new(value, vlen);
-
   f = find_common_field_value(field, flen);
 
   if (f == Qnil) {
@@ -203,9 +196,11 @@ static void http_field(void *data, const char *field,
     memcpy(RSTRING_PTR(f) + HTTP_PREFIX_LEN, field, flen);
     assert(*(RSTRING_PTR(f) + RSTRING_LEN(f)) == '\0'); /* paranoia */
     /* fprintf(stderr, "UNKNOWN HEADER <%s>\n", RSTRING_PTR(f)); */
+  } else if (f == global_http_host && rb_hash_aref(req, f) != Qnil) {
+    return;
   }
 
-  rb_hash_aset(req, f, v);
+  rb_hash_aset(req, f, rb_str_new(value, vlen));
 }
 
 static void request_method(void *data, const char *at, size_t length)
@@ -217,6 +212,16 @@ static void request_method(void *data, const char *at, size_t length)
   rb_hash_aset(req, global_request_method, val);
 }
 
+static void scheme(void *data, const char *at, size_t length)
+{
+  rb_hash_aset((VALUE)data, global_rack_url_scheme, rb_str_new(at, length));
+}
+
+static void host(void *data, const char *at, size_t length)
+{
+  rb_hash_aset((VALUE)data, global_http_host, rb_str_new(at, length));
+}
+
 static void request_uri(void *data, const char *at, size_t length)
 {
   VALUE req = (VALUE)data;
@@ -226,6 +231,13 @@ static void request_uri(void *data, const char *at, size_t length)
 
   val = rb_str_new(at, length);
   rb_hash_aset(req, global_request_uri, val);
+
+  /* "OPTIONS * HTTP/1.1\r\n" is a valid request */
+  if (length == 1 && *at == '*') {
+    val = rb_str_new(NULL, 0);
+    rb_hash_aset(req, global_request_path, val);
+    rb_hash_aset(req, global_path_info, val);
+  }
 }
 
 static void fragment(void *data, const char *at, size_t length)
@@ -248,6 +260,10 @@ static void request_path(void *data, const char *at, size_t length)
 
   val = rb_str_new(at, length);
   rb_hash_aset(req, global_request_path, val);
+
+  /* rack says PATH_INFO must start with "/" or be empty */
+  if (!(length == 1 && *at == '*'))
+    rb_hash_aset(req, global_path_info, val);
 }
 
 static void query_string(void *data, const char *at, size_t length)
@@ -268,46 +284,49 @@ static void http_version(void *data, const char *at, size_t length)
   rb_hash_aset(req, global_http_version, val);
 }
 
-/** Finalizes the request header to have a bunch of stuff that's
-  needed. */
-
+/** Finalizes the request header to have a bunch of stuff that's needed. */
 static void header_done(void *data, const char *at, size_t length)
 {
   VALUE req = (VALUE)data;
-  VALUE temp = Qnil;
-  VALUE ctype = Qnil;
-  VALUE clen = Qnil;
-  char *colon = NULL;
-
-  clen = rb_hash_aref(req, global_http_content_length);
-  if(clen != Qnil) {
-    rb_hash_aset(req, global_content_length, clen);
+  VALUE server_name = global_localhost;
+  VALUE server_port = global_port_80;
+  VALUE temp;
+
+  /* rack requires QUERY_STRING */
+  if (rb_hash_aref(req, global_query_string) == Qnil)
+    rb_hash_aset(req, global_query_string, rb_str_new(NULL, 0));
+
+  /* set rack.url_scheme to "https" or "http", no others are allowed by Rack */
+  if ((temp = rb_hash_aref(req, global_rack_url_scheme)) == Qnil) {
+    if ((temp = rb_hash_aref(req, global_http_x_forwarded_proto)) != Qnil &&
+        RSTRING_LEN(temp) == 5 &&
+        !memcmp("https", RSTRING_PTR(temp), 5))
+      server_port = global_port_443;
+    else
+      temp = global_http;
+    rb_hash_aset(req, global_rack_url_scheme, temp);
+  } else if (RSTRING_LEN(temp) == 5 && !memcmp("https", RSTRING_PTR(temp), 5)) {
+    server_port = global_port_443;
   }
 
-  ctype = rb_hash_aref(req, global_http_content_type);
-  if(ctype != Qnil) {
-    rb_hash_aset(req, global_content_type, ctype);
-  }
+  /* parse and set the SERVER_NAME and SERVER_PORT variables */
+  if ((temp = rb_hash_aref(req, global_http_host)) != Qnil) {
+    char *colon = memchr(RSTRING_PTR(temp), ':', RSTRING_LEN(temp));
+    if (colon) {
+      long port_start = colon - RSTRING_PTR(temp) + 1;
 
-  rb_hash_aset(req, global_gateway_interface, global_gateway_interface_value);
-  if((temp = rb_hash_aref(req, global_http_host)) != Qnil) {
-    colon = memchr(RSTRING_PTR(temp), ':', RSTRING_LEN(temp));
-    if(colon != NULL) {
-      rb_hash_aset(req, global_server_name, rb_str_substr(temp, 0, colon - RSTRING_PTR(temp)));
-      rb_hash_aset(req, global_server_port, 
-          rb_str_substr(temp, colon - RSTRING_PTR(temp)+1, 
-            RSTRING_LEN(temp)));
+      server_name = rb_str_substr(temp, 0, colon - RSTRING_PTR(temp));
+      if ((RSTRING_LEN(temp) - port_start) > 0)
+        server_port = rb_str_substr(temp, port_start, RSTRING_LEN(temp));
     } else {
-      rb_hash_aset(req, global_server_name, temp);
-      rb_hash_aset(req, global_server_port, global_port_80);
+      server_name = temp;
     }
-  } else {
-    rb_hash_aset(req, global_server_name, global_localhost);
-    rb_hash_aset(req, global_server_port, global_port_80);
   }
+  rb_hash_aset(req, global_server_name, server_name);
+  rb_hash_aset(req, global_server_port, server_port);
 
   /* grab the initial body and stuff it into the hash */
-  rb_hash_aset(req, global_http_body, rb_str_new(at, length));
+  rb_hash_aset(req, sym_http_body, rb_str_new(at, length));
   rb_hash_aset(req, global_server_protocol, global_server_protocol_value);
 }
 
@@ -325,14 +344,6 @@ static VALUE HttpParser_alloc(VALUE klass)
   VALUE obj;
   http_parser *hp = ALLOC_N(http_parser, 1);
   TRACE();
-  hp->http_field = http_field;
-  hp->request_method = request_method;
-  hp->request_uri = request_uri;
-  hp->fragment = fragment;
-  hp->request_path = request_path;
-  hp->query_string = query_string;
-  hp->http_version = http_version;
-  hp->header_done = header_done;
   http_parser_init(hp);
 
   obj = Data_Wrap_Struct(klass, NULL, HttpParser_free, hp);
@@ -349,9 +360,7 @@ static VALUE HttpParser_alloc(VALUE klass)
  */
 static VALUE HttpParser_init(VALUE self)
 {
-  http_parser *http = NULL;
-  DATA_GET(self, http_parser, http);
-  http_parser_init(http);
+  http_parser_init(data_get(self));
 
   return self;
 }
@@ -366,9 +375,7 @@ static VALUE HttpParser_init(VALUE self)
  */
 static VALUE HttpParser_reset(VALUE self)
 {
-  http_parser *http = NULL;
-  DATA_GET(self, http_parser, http);
-  http_parser_init(http);
+  http_parser_init(data_get(self));
 
   return Qnil;
 }
@@ -376,140 +383,58 @@ static VALUE HttpParser_reset(VALUE self)
 
 /**
  * call-seq:
- *    parser.finish -> true/false
- *
- * Finishes a parser early which could put in a "good" or bad state.
- * You should call reset after finish it or bad things will happen.
- */
-static VALUE HttpParser_finish(VALUE self)
-{
-  http_parser *http = NULL;
-  DATA_GET(self, http_parser, http);
-  http_parser_finish(http);
-
-  return http_parser_is_finished(http) ? Qtrue : Qfalse;
-}
-
-
-/**
- * call-seq:
- *    parser.execute(req_hash, data, start) -> Integer
+ *    parser.execute(req_hash, data) -> true/false
  *
- * Takes a Hash and a String of data, parses the String of data filling in the Hash
- * returning an Integer to indicate how much of the data has been read.  No matter
- * what the return value, you should call HttpParser#finished? and HttpParser#error?
- * to figure out if it's done parsing or there was an error.
+ * Takes a Hash and a String of data, parses the String of data filling
+ * in the Hash returning a boolean to indicate whether or not parsing
+ * is finished.
  * 
- * This function now throws an exception when there is a parsing error.  This makes 
- * the logic for working with the parser much easier.  You can still test for an 
- * error, but now you need to wrap the parser with an exception handling block.
- *
- * The third argument allows for parsing a partial request and then continuing
- * the parsing from that position.  It needs all of the original data as well 
- * so you have to append to the data buffer as you read.
- */
-static VALUE HttpParser_execute(VALUE self, VALUE req_hash,
-                                VALUE data, VALUE start)
-{
-  http_parser *http = NULL;
-  int from = 0;
-  char *dptr = NULL;
-  long dlen = 0;
-
-  DATA_GET(self, http_parser, http);
-
-  from = FIX2INT(start);
-  dptr = RSTRING_PTR(data);
-  dlen = RSTRING_LEN(data);
-
-  if(from >= dlen) {
-    rb_raise(eHttpParserError, "Requested start is after data buffer end.");
-  } else {
-    http->data = (void *)req_hash;
-    http_parser_execute(http, dptr, dlen, from);
-
-    VALIDATE_MAX_LENGTH(http_parser_nread(http), HEADER);
-
-    if(http_parser_has_error(http)) {
-      rb_raise(eHttpParserError, "Invalid HTTP format, parsing fails.");
-    } else {
-      return INT2FIX(http_parser_nread(http));
-    }
-  }
-}
-
-
-
-/**
- * call-seq:
- *    parser.error? -> true/false
- *
- * Tells you whether the parser is in an error state.
+ * This function now throws an exception when there is a parsing error.
+ * This makes the logic for working with the parser much easier.  You
+ * will need to wrap the parser with an exception handling block.
  */
-static VALUE HttpParser_has_error(VALUE self)
-{
-  http_parser *http = NULL;
-  DATA_GET(self, http_parser, http);
-
-  return http_parser_has_error(http) ? Qtrue : Qfalse;
-}
-
 
-/**
- * call-seq:
- *    parser.finished? -> true/false
- *
- * Tells you whether the parser is finished or not and in a good state.
- */
-static VALUE HttpParser_is_finished(VALUE self)
+static VALUE HttpParser_execute(VALUE self, VALUE req_hash, VALUE data)
 {
-  http_parser *http = NULL;
-  DATA_GET(self, http_parser, http);
+  http_parser *http = data_get(self);
+  char *dptr = RSTRING_PTR(data);
+  long dlen = RSTRING_LEN(data);
 
-  return http_parser_is_finished(http) ? Qtrue : Qfalse;
-}
+  if (http->nread < dlen) {
+    http->data = (void *)req_hash;
+    http_parser_execute(http, dptr, dlen);
 
+    VALIDATE_MAX_LENGTH(http->nread, HEADER);
 
-/**
- * call-seq:
- *    parser.nread -> Integer
- *
- * Returns the amount of data processed so far during this processing cycle.  It is
- * set to 0 on initialize or reset calls and is incremented each time execute is called.
- */
-static VALUE HttpParser_nread(VALUE self)
-{
-  http_parser *http = NULL;
-  DATA_GET(self, http_parser, http);
+    if (!http_parser_has_error(http))
+      return http_parser_is_finished(http) ? Qtrue : Qfalse;
 
-  return INT2FIX(http->nread);
+    rb_raise(eHttpParserError, "Invalid HTTP format, parsing fails.");
+  }
+  rb_raise(eHttpParserError, "Requested start is after data buffer end.");
 }
 
-void Init_http11()
+void Init_http11(void)
 {
-
   mUnicorn = rb_define_module("Unicorn");
 
+  DEF_GLOBAL(rack_url_scheme, "rack.url_scheme");
   DEF_GLOBAL(request_method, "REQUEST_METHOD");
   DEF_GLOBAL(request_uri, "REQUEST_URI");
   DEF_GLOBAL(fragment, "FRAGMENT");
   DEF_GLOBAL(query_string, "QUERY_STRING");
   DEF_GLOBAL(http_version, "HTTP_VERSION");
   DEF_GLOBAL(request_path, "REQUEST_PATH");
-  DEF_GLOBAL(content_length, "CONTENT_LENGTH");
-  DEF_GLOBAL(http_content_length, "HTTP_CONTENT_LENGTH");
-  DEF_GLOBAL(http_body, "HTTP_BODY");
-  DEF_GLOBAL(content_type, "CONTENT_TYPE");
-  DEF_GLOBAL(http_content_type, "HTTP_CONTENT_TYPE");
-  DEF_GLOBAL(gateway_interface, "GATEWAY_INTERFACE");
-  DEF_GLOBAL(gateway_interface_value, "CGI/1.2");
+  DEF_GLOBAL(path_info, "PATH_INFO");
   DEF_GLOBAL(server_name, "SERVER_NAME");
   DEF_GLOBAL(server_port, "SERVER_PORT");
   DEF_GLOBAL(server_protocol, "SERVER_PROTOCOL");
   DEF_GLOBAL(server_protocol_value, "HTTP/1.1");
-  DEF_GLOBAL(http_host, "HTTP_HOST");
+  DEF_GLOBAL(http_x_forwarded_proto, "HTTP_X_FORWARDED_PROTO");
   DEF_GLOBAL(port_80, "80");
+  DEF_GLOBAL(port_443, "443");
   DEF_GLOBAL(localhost, "localhost");
+  DEF_GLOBAL(http, "http");
 
   eHttpParserError = rb_define_class_under(mUnicorn, "HttpParserError", rb_eIOError);
 
@@ -517,10 +442,9 @@ void Init_http11()
   rb_define_alloc_func(cHttpParser, HttpParser_alloc);
   rb_define_method(cHttpParser, "initialize", HttpParser_init,0);
   rb_define_method(cHttpParser, "reset", HttpParser_reset,0);
-  rb_define_method(cHttpParser, "finish", HttpParser_finish,0);
-  rb_define_method(cHttpParser, "execute", HttpParser_execute,3);
-  rb_define_method(cHttpParser, "error?", HttpParser_has_error,0);
-  rb_define_method(cHttpParser, "finished?", HttpParser_is_finished,0);
-  rb_define_method(cHttpParser, "nread", HttpParser_nread,0);
+  rb_define_method(cHttpParser, "execute", HttpParser_execute,2);
+  sym_http_body = ID2SYM(rb_intern("http_body"));
   init_common_fields();
+  global_http_host = find_common_field_value("HOST", 4);
+  assert(global_http_host != Qnil);
 }