Some changes related to URL-parsing for upcoming ICAP additions:

 - findTrailingHTTPVersion() now has optional 'char *end' arg for
   parsing non-terminated buffers.  If end defaults to NULL, then
   the string must be NULL terminated.
 - urlParseMethod() now has optional 'char *end' arg.
 - urlParseProtocol() now has optional 'char *end' arg.
 - urlParse() now has optional 'HttpRequest *' arg.  If non-NULL,
   urlParse() does not allocate a new HttpRequest and returns
   this passed arg.

diff --git a/src/client_side.cc b/src/client_side.cc
index 1ef357d98e1a97a22e75dd51ae43330df9dd2300..a9c160d8c7bced870adf6d77fe289e8f4ae2a86a 100644 (file)
--- a/src/client_side.cc
+++ b/src/client_side.cc
@@ -1,6 +1,6 @@
 
 /*
- * $Id: client_side.cc,v 1.694 2005/09/14 17:10:38 serassio Exp $
+ * $Id: client_side.cc,v 1.695 2005/09/14 21:09:38 wessels Exp $
  *
  * DEBUG: section 33    Client-side Routines
  * AUTHOR: Duane Wessels
@@ -153,7 +153,6 @@ static void clientUpdateSocketStats(log_type logType, size_t size);
 
 static ClientSocketContext *clientParseRequestMethod(char *inbuf, method_t * method_p, ConnStateData::Pointer & conn);
 static char *skipLeadingSpace(char *aString);
-static char *findTrailingHTTPVersion(char *uriAndHTTPVersion);
 #if UNUSED_CODE
 static void trimTrailingSpaces(char *aString, size_t len);
 #endif
@@ -1615,18 +1614,26 @@ skipLeadingSpace(char *aString)
     return result;
 }
 
-static char *
-findTrailingHTTPVersion(char *uriAndHTTPVersion)
+/*
+ * 'end' defaults to NULL for backwards compatibility
+ * remove default value if we ever get rid of NULL-terminated
+ * request buffers.
+ */
+const char *
+findTrailingHTTPVersion(const char *uriAndHTTPVersion, const char *end)
 {
-    char *token;
+    if (NULL == end) {
+        end = uriAndHTTPVersion + strcspn(uriAndHTTPVersion, "\r\n");
+        assert(end);
+    }
 
-    for (token = strchr(uriAndHTTPVersion, '\n'); token > uriAndHTTPVersion; token--) {
-        if (*token == '\n' || *token == '\r')
+    for (; end > uriAndHTTPVersion; end--) {
+        if (*end == '\n' || *end == '\r')
             continue;
 
-        if (xisspace(*token)) {
-            if (strncasecmp(token + 1, "HTTP/", 5) == 0)
-                return token + 1;
+        if (xisspace(*end)) {
+            if (strncasecmp(end + 1, "HTTP/", 5) == 0)
+                return end + 1;
             else
                 break;
         }
@@ -1849,7 +1856,7 @@ parseHttpRequest(ConnStateData::Pointer & conn, method_t * method_p,
     ClientHttpRequest *http;
     ClientSocketContext *result;
     StoreIOBuffer tempBuffer;
-    char *http_version;
+    const char *http_version;
 
     /* pre-set these values to make aborting simpler */
     *prefix_p = NULL;
@@ -1905,7 +1912,7 @@ parseHttpRequest(ConnStateData::Pointer & conn, method_t * method_p,
 
     /* Is there a legitimate first line to the headers ? */
     if ((result = clientParseHttpRequestLine(inbuf, conn, method_p, &url,
-                  http_ver, http_version))) {
+                  http_ver, (char *) http_version))) {
         /* something wrong, abort */
         xfree(inbuf);
         return result;

diff --git a/src/protos.h b/src/protos.h
index 562910c4a7331ce0d1fc670916ceb7838dedeeb4..093ea1c5f61dc6ef53208d67174dfd325f7703b7 100644 (file)
--- a/src/protos.h
+++ b/src/protos.h
@@ -1,6 +1,6 @@
 
 /*
- * $Id: protos.h,v 1.509 2005/09/12 23:28:57 wessels Exp $
+ * $Id: protos.h,v 1.510 2005/09/14 21:09:38 wessels Exp $
  *
  *
  * SQUID Web Proxy Cache          http://www.squid-cache.org/
@@ -819,10 +819,10 @@ SQUIDCEXTERN void unlinkdUnlink(const char *);
 
 SQUIDCEXTERN char *url_convert_hex(char *org_url, int allocate);
 SQUIDCEXTERN char *url_escape(const char *url);
-SQUIDCEXTERN protocol_t urlParseProtocol(const char *);
-SQUIDCEXTERN method_t urlParseMethod(const char *);
+SQUIDCEXTERN protocol_t urlParseProtocol(const char *, const char *e = NULL);
+SQUIDCEXTERN method_t urlParseMethod(const char *, const char *e = NULL);
 SQUIDCEXTERN void urlInitialize(void);
-SQUIDCEXTERN HttpRequest *urlParse(method_t, char *);
+SQUIDCEXTERN HttpRequest *urlParse(method_t, char *, HttpRequest *request = NULL);
 SQUIDCEXTERN const char *urlCanonical(HttpRequest *);
 SQUIDCEXTERN char *urlRInternal(const char *host, u_short port, const char *dir, const char *name);
 SQUIDCEXTERN char *urlInternal(const char *dir, const char *name);

diff --git a/src/url.cc b/src/url.cc
index fc8086e7c258c72aee2eecf8ee527749d61c9be2..36e5e96eaf53c6f60838036fb4e8596155be7469 100644 (file)
--- a/src/url.cc
+++ b/src/url.cc
@@ -1,6 +1,6 @@
 
 /*
- * $Id: url.cc,v 1.146 2003/08/10 11:00:45 robertc Exp $
+ * $Id: url.cc,v 1.147 2005/09/14 21:09:38 wessels Exp $
  *
  * DEBUG: section 23    URL Parsing
  * AUTHOR: Duane Wessels
@@ -189,9 +189,13 @@ method_t &operator++ (method_t &aMethod)
     return aMethod;
 }
 
-
+/*
+ * urlParseMethod() takes begin and end pointers, but for backwards
+ * compatibility, end defaults to NULL, in which case we assume begin
+ * is NULL-terminated.
+ */
 method_t
-urlParseMethod(const char *s)
+urlParseMethod(const char *b, const char *e)
 {
     method_t method = METHOD_NONE;
     /*
@@ -200,57 +204,79 @@ urlParseMethod(const char *s)
      * which have the form %EXT[0-9][0-9]
      */
 
-    if (*s == '%')
+    if (*b == '%')
         return METHOD_NONE;
 
+    /*
+     * if e is NULL, b must be NULL terminated and we
+     * make e point to the first whitespace character
+     * after b.
+     */
+    if (NULL == e)
+        e = b + strcspn(b, w_space);
+
     for (++method; method < METHOD_ENUM_END; ++method) {
-        if (0 == strcasecmp(s, RequestMethodStr[method]))
+        if (0 == strncasecmp(b, RequestMethodStr[method], e-b))
             return method;
     }
 
     return METHOD_NONE;
 }
 
-
+/*
+ * urlParseProtocol() takes begin (b) and end (e) pointers, but for
+ * backwards compatibility, e defaults to NULL, in which case we
+ * assume b is NULL-terminated.
+ */
 protocol_t
-urlParseProtocol(const char *s)
+urlParseProtocol(const char *b, const char *e)
 {
+    /*
+     * if e is NULL, b must be NULL terminated and we
+     * make e point to the first whitespace character
+     * after b.
+     */
+
+    if (NULL == e)
+        e = b + strcspn(b, ":");
+
+    int len = e - b;
+
     /* test common stuff first */
 
-    if (strcasecmp(s, "http") == 0)
+    if (strncasecmp(b, "http", len) == 0)
         return PROTO_HTTP;
 
-    if (strcasecmp(s, "ftp") == 0)
+    if (strncasecmp(b, "ftp", len) == 0)
         return PROTO_FTP;
 
-    if (strcasecmp(s, "https") == 0)
+    if (strncasecmp(b, "https", len) == 0)
         return PROTO_HTTPS;
 
-    if (strcasecmp(s, "file") == 0)
+    if (strncasecmp(b, "file", len) == 0)
         return PROTO_FTP;
 
-    if (strcasecmp(s, "gopher") == 0)
+    if (strncasecmp(b, "gopher", len) == 0)
         return PROTO_GOPHER;
 
-    if (strcasecmp(s, "wais") == 0)
+    if (strncasecmp(b, "wais", len) == 0)
         return PROTO_WAIS;
 
-    if (strcasecmp(s, "cache_object") == 0)
+    if (strncasecmp(b, "cache_object", len) == 0)
         return PROTO_CACHEOBJ;
 
-    if (strcasecmp(s, "urn") == 0)
+    if (strncasecmp(b, "urn", len) == 0)
         return PROTO_URN;
 
-    if (strcasecmp(s, "whois") == 0)
+    if (strncasecmp(b, "whois", len) == 0)
         return PROTO_WHOIS;
 
-    if (strcasecmp(s, "internal") == 0)
+    if (strncasecmp(b, "internal", len) == 0)
         return PROTO_INTERNAL;
 
     return PROTO_NONE;
 }
 
-
 int
 urlDefaultPort(protocol_t p)
 {
@@ -284,14 +310,19 @@ urlDefaultPort(protocol_t p)
     }
 }
 
+/*
+ * Parse a URI/URL.
+ *
+ * If the 'request' arg is non-NULL, put parsed values there instead
+ * of allocating a new HttpRequest.
+ */
 HttpRequest *
-urlParse(method_t method, char *url)
+urlParse(method_t method, char *url, HttpRequest *request)
 {
     LOCAL_ARRAY(char, proto, MAX_URL);
     LOCAL_ARRAY(char, login, MAX_URL);
     LOCAL_ARRAY(char, host, MAX_URL);
     LOCAL_ARRAY(char, urlpath, MAX_URL);
-    HttpRequest *request = NULL;
     char *t = NULL;
     char *q = NULL;
     int port;
@@ -426,7 +457,14 @@ urlParse(method_t method, char *url)
         }
     }
 
-    request = requestCreate(method, protocol, urlpath);
+    if (NULL == request)
+        request = requestCreate(method, protocol, urlpath);
+    else {
+        request->method = method;
+        request->protocol = protocol;
+        request->urlpath = urlpath;
+    }
+
     xstrncpy(request->host, host, SQUIDHOSTNAMELEN);
     xstrncpy(request->login, login, MAX_LOGIN_SZ);
     request->port = (u_short) port;