---------------------
PatchSet 8572 
Date: 2006/08/31 04:52:25
Author: adri
Branch: parserwork
Tag: (none) 
Log:
Flesh out the rest of the stuff required for HTTP/0.9 parsing

Members: 
	src/HttpMsg.c:1.8.8.5->1.8.8.6 
	src/client_side.c:1.143.2.12->1.143.2.13 

Index: squid/src/HttpMsg.c
===================================================================
RCS file: /cvsroot/squid-sf//squid/src/HttpMsg.c,v
retrieving revision 1.8.8.5
retrieving revision 1.8.8.6
diff -u -r1.8.8.5 -r1.8.8.6
--- squid/src/HttpMsg.c	27 Aug 2006 02:16:37 -0000	1.8.8.5
+++ squid/src/HttpMsg.c	31 Aug 2006 04:52:25 -0000	1.8.8.6
@@ -1,6 +1,6 @@
 
 /*
- * $Id: HttpMsg.c,v 1.8.8.5 2006/08/27 02:16:37 adri Exp $
+ * $Id: HttpMsg.c,v 1.8.8.6 2006/08/31 04:52:25 adri Exp $
  *
  * DEBUG: section 74    HTTP Message
  * AUTHOR: Alex Rousskov
@@ -70,6 +70,7 @@
 	int i = 0;
 	int retcode = 0;
 	int maj = -1, min = -1;
+	int last_whitespace = -1, line_end = -1;
 
 	/* Find \r\n - end of URL+Version (and the request) */
 	for (i = 0; i < hmsg->size; i++) {
@@ -77,6 +78,7 @@
 			break;
 		}
 		if (i < hmsg->size - 1 && hmsg->buf[i - 1] == '\r' && hmsg->buf[i] == '\n') {
+			i++;
 			break;
 		}
 	}
@@ -113,78 +115,104 @@
 	}
 	hmsg->u_start = i;
 
-	/* Find whitespace; end of URL */
-	for (; i < hmsg->req_end && (! isspace(hmsg->buf[i])); i++);
-	if (i >= hmsg->req_end) {
-		retcode = 0;
-		goto finish;
-	}
-	hmsg->u_end = i - 1;
-
-	/* XXX yes, this doesn't support HTTP/0.9 requests just yet .. */
-
-	/* Find non-whitespace, version */
-	for (; i < hmsg->req_end && (isspace(hmsg->buf[i])); i++);
-	if (i >= hmsg->req_end) {
-		retcode = 0;
-		goto finish;
+	/* Find \r\n or \n - thats the end of the line. Keep track of the last whitespace! */
+	for (; i <= hmsg->req_end; i++) {
+		/* If \n - its end of line */
+		if (hmsg->buf[i] == '\n') {
+			line_end = i;
+			break;
+		}
+		/* XXX could be off-by-one wrong! */
+		if (hmsg->buf[i] == '\r' && (i + 1) <= hmsg->req_end && hmsg->buf[i+1] == '\n') {
+			line_end = i;
+			break;
+		}
+		/* If its a whitespace, note it as it'll delimit our version */
+		if (hmsg->buf[i] == ' ' || hmsg->buf[i] == '\t') {
+			last_whitespace = i;
+		}
 	}
-	hmsg->v_start = i;
-
-	/* next five characters should be HTTP/ */
-	if (i + 5 >= hmsg->req_end) {
+	debug(1, 1) ("foo: %d-%d, whitespace:%d\n", hmsg->u_start, line_end, last_whitespace);
+	if (i > hmsg->req_end) {
 		retcode = 0;
 		goto finish;
 	}
-	if (strncasecmp(&hmsg->buf[i], "HTTP/", 5) != 0) {
-		retcode = -1;
-		goto finish;
-	}
-	i += 5;
 
-	/* next should be 1 or more digits */
-	maj = 0;
-	for (; i < hmsg->req_end && (isdigit(hmsg->buf[i])); i++) {
-		maj = maj * 10;
-		maj = maj + (hmsg->buf[i]) - '0';
+	/* At this point we don't need the 'i' value; so we'll recycle it for version parsing */
 
-	}
-	if (i >= hmsg->req_end) {
-		retcode = 0;
-		goto finish;
-	}
+	/* 
+	 * At this point: line_end points to the first eol char (\r or \n);
+	 * last_whitespace points to the last whitespace char in the URL.
+	 * We know we have a full buffer here!
+	 */
+	if (last_whitespace == -1) {
+		maj = 0; min = 9;
+		hmsg->u_end = line_end - 1;
+		assert(hmsg->u_end >= hmsg->u_start);
+	} else {
+		/* Find the first non-whitespace after last_whitespace */
+		/* XXX why <= vs < ? I do need to really re-audit all of this ..*/
+		for (i = last_whitespace; i <= hmsg->req_end && isspace(hmsg->buf[i]); i++);
+		if (i > hmsg->req_end) {
+			retcode = 0;
+			goto finish;
+		}
 
-	/* next should be . */
-	if (hmsg->buf[i] != '.') {
-		retcode = -1;
-		goto finish;
-	}
-	if (i + 1 >= hmsg->req_end) {
-		retcode = 0;
-		goto finish;
-	}
+		/* is it http/ ? if so, we try parsing. If not, the URL is the whole line; version is 0.9 */
+		if (i + 5 >= hmsg->req_end || (strncasecmp(&hmsg->buf[i], "HTTP/", 5) != 0)) {
+			maj = 0; min = 9;
+			hmsg->u_end = line_end - 1;
+			assert(hmsg->u_end >= hmsg->u_start);
+		} else {
+			/* Ok, lets try parsing! Yes, this needs refactoring! */
+			hmsg->v_start = i;
+			i += 5;
+
+			/* next should be 1 or more digits */
+			maj = 0;
+			for (; i < hmsg->req_end && (isdigit(hmsg->buf[i])); i++) {
+				maj = maj * 10;
+				maj = maj + (hmsg->buf[i]) - '0';
+			}
+			if (i >= hmsg->req_end) {
+				retcode = 0;
+				goto finish;
+			}
+
+			/* next should be .; we -have- to have this as we have a whole line.. */
+			if (hmsg->buf[i] != '.') {
+				retcode = 0;
+				goto finish;
+			}
+			if (i + 1 >= hmsg->req_end) {
+				retcode = 0;
+				goto finish;
+			}
 	
-	/* next should be one or more digits */
-	i++;
-	min = 0;
-	for (; i < hmsg->req_end && (isdigit(hmsg->buf[i])); i++) {
-		min = min * 10;
-		min = min + (hmsg->buf[i]) - '0';
-
+			/* next should be one or more digits */
+			i++;
+			min = 0;
+			for (; i < hmsg->req_end && (isdigit(hmsg->buf[i])); i++) {
+				min = min * 10;
+				min = min + (hmsg->buf[i]) - '0';
+			}
+
+			/* Find whitespace, end of version */
+			hmsg->v_end = i;
+			hmsg->u_end = last_whitespace - 1;
+		}
 	}
 
-	/* Find whitespace, end of version */
-	hmsg->v_end = i;
-
-	hmsg->v_maj = maj;
-	hmsg->v_min = min;
-
 	/* 
 	 * Rightio - we have all the schtuff. Return true; we've got enough.
 	 */
 	retcode = 1;
 
 finish:
+	hmsg->v_maj = maj;
+	hmsg->v_min = min;
+	assert(maj != -1);
+	assert(min != -1);
 	debug(1, 2) ("Parser: retval %d: from %d->%d: method %d->%d; url %d->%d; version %d->%d (%d/%d)\n",
 	    retcode, hmsg->req_start, hmsg->req_end,
 	    hmsg->m_start, hmsg->m_end,
Index: squid/src/client_side.c
===================================================================
RCS file: /cvsroot/squid-sf//squid/src/client_side.c,v
retrieving revision 1.143.2.12
retrieving revision 1.143.2.13
diff -u -r1.143.2.12 -r1.143.2.13
--- squid/src/client_side.c	30 Aug 2006 07:47:42 -0000	1.143.2.12
+++ squid/src/client_side.c	31 Aug 2006 04:52:25 -0000	1.143.2.13
@@ -1,6 +1,6 @@
 
 /*
- * $Id: client_side.c,v 1.143.2.12 2006/08/30 07:47:42 adri Exp $
+ * $Id: client_side.c,v 1.143.2.13 2006/08/31 04:52:25 adri Exp $
  *
  * DEBUG: section 33    Client-side Routines
  * AUTHOR: Duane Wessels
@@ -3597,12 +3597,19 @@
     /*
      * Process headers after request line
      */
-    /* for now, assume there's going to be headers. Worry about HTTP/0.9 later */
-    assert(hmsg->headers_start > -1);
-    assert(hmsg->headers_end > -1);
+    /* XXX Are these sizes off by one? */
+    if (hmsg->v_maj > 0) {
+	    /* http/1.0 and above */
+            assert(hmsg->headers_start > -1);
+            assert(hmsg->headers_end > -1);
+            header_sz = hmsg->headers_end - hmsg->headers_start;
+            prefix_sz = hmsg->headers_end - hmsg->req_start;
+    } else {
+	    /* http/0.9 - no headers */
+	    header_sz = 0;
+	    prefix_sz = hmsg->req_end - hmsg->req_start;
+    }
 
-    header_sz = hmsg->headers_end - hmsg->headers_start;
-    prefix_sz = hmsg->headers_end - hmsg->req_start;
     //debug(33, 3) ("parseHttpRequest: req_hdr = {%s}\n", req_hdr);
     //debug(33, 3) ("parseHttpRequest: end = {%s}\n", end);