---------------------
PatchSet 1401 
Date: 2001/01/27 02:38:58
Author: rbcollins
Branch: rbcollins_filters
Tag: (none) 
Log:
sample filter working

Members: 
	src/client_side.c:1.1.1.3.4.1.4.15.2.2->1.1.1.3.4.1.4.15.2.3 

Index: squid/src/client_side.c
===================================================================
RCS file: /cvsroot/squid-sf//squid/src/client_side.c,v
retrieving revision 1.1.1.3.4.1.4.15.2.2
retrieving revision 1.1.1.3.4.1.4.15.2.3
diff -u -r1.1.1.3.4.1.4.15.2.2 -r1.1.1.3.4.1.4.15.2.3
--- squid/src/client_side.c	26 Jan 2001 13:18:32 -0000	1.1.1.3.4.1.4.15.2.2
+++ squid/src/client_side.c	27 Jan 2001 02:38:58 -0000	1.1.1.3.4.1.4.15.2.3
@@ -1,6 +1,6 @@
 
 /*
- * $Id: client_side.c,v 1.1.1.3.4.1.4.15.2.2 2001/01/26 13:18:32 rbcollins Exp $
+ * $Id: client_side.c,v 1.1.1.3.4.1.4.15.2.3 2001/01/27 02:38:58 rbcollins Exp $
  *
  * DEBUG: section 33    Client-side Routines
  * AUTHOR: Duane Wessels
@@ -1991,6 +1991,103 @@
 //    comm_write(http->conn->fd, buf, len, clientWriteBodyComplete, http, NULL);
 }
 
+#define TESTFILTER 1
+#if TESTFILTER
+
+char *
+strnchr(const char *str, char chr, size_t len) {
+    size_t pos=0;
+    while (pos<len) {
+        if (tolower(*str)==tolower(chr))
+	    return (char *) str;
+	pos++;
+	str++;
+    }
+    return NULL;
+}
+
+static void
+clientFilterOnUnload(const char *buf, size_t len, dlink_list *filter_list, FILTER_list *filters, unsigned int flags, void *data) {
+#define pattern "onunload.*\".*\""
+    clientHttpRequest *http = data;
+    FILTER_list *temp_filter;
+    char *pos, *chr,*startpos, searchstring[]="onunload";
+    regex_t comp;
+
+    /* this looks for browser DOM handlers of the type onunload="..." and
+     * removes them from the data passed to the client
+     * caveats and problems: 
+     * 1. theres probably a set of library functions for matching patterns at arbitrary
+     * places in a datastream (ie handles buffer edge issues, optimum search patterns
+     * and the like. Well this is a proof of concept filter - live with it :-]
+     * 2. Some code may not quote the handler properly. I haven't tried to build
+     * a full blown parser here - see 1.
+     * 3. this will match patterns in jpegs etc as well at this point. It needs a 
+     * acl or content type check in clientSendMoreData before adding the filter to the
+     * list.
+     *
+     * 4. I'm making use of the properties of the fixed search string to simplify
+     * the search. IT IS NOT GENERAL PURPOSE 
+
+
+     * The compiled expression isn't cached. This _is_ a performance issue.
+     * !!! !!!!!!!! !!!!!!!!!!
+     */
+
+
+    temp_filter=filters->node.next->data;
+
+    chr=searchstring;
+    pos=0;
+    startpos=buf;
+    
+    while (startpos) {
+    startpos=strnchr(startpos, *chr, len-(startpos-buf));
+
+    if (startpos) {
+	/* possibly present in this buffer */
+        chr++;
+	pos=startpos+1;
+	while (pos) {
+	    if (tolower(*pos)==tolower(*chr)) {
+		chr++;
+		pos++;
+	    }
+	    if (pos-buf>=len){
+		/* end of data */
+		temp_filter->filter(buf,len,filter_list, temp_filter, flags, temp_filter->data);
+		return;
+	    }
+	    if (*chr=='\0') {
+		/* end of search string - we've got a match */
+		/* send the first set of data */
+		if (startpos-buf)
+		    temp_filter->filter(buf,startpos-buf,filter_list, temp_filter, flags, temp_filter->data);
+		/* send the filter set of data */
+		temp_filter->filter("nonsense",8,filter_list, temp_filter, flags, temp_filter->data);
+		/* and the rest */
+		temp_filter->filter(startpos+8,len-(startpos-buf)-8,filter_list, temp_filter, flags,temp_filter->data);
+		return;
+	    }
+	    if(tolower(*pos)!=tolower(*chr)) {
+		/* reset the search criteria to after the corrent position */
+		startpos=pos;
+		pos=NULL;
+		chr=searchstring;
+	    }
+	}
+    }
+    else {
+        temp_filter->filter(buf,len,filter_list, temp_filter, flags, temp_filter->data);
+        return;
+    }
+
+    }
+    temp_filter->filter(buf,len,filter_list, temp_filter, flags, temp_filter->data);
+    
+}
+#endif
+
 /*
  * accepts chunk of a http message in buf, parses prefix, filters headers and
  * such, writes processed message to the client's socket
@@ -2097,6 +2194,15 @@
 	temp_filter->data=http;
 	/* cbDataLock(http); ? */
 	dlinkAdd(temp_filter, &temp_filter->node, &http->repfilters);
+#if TESTFILTER
+	temp_filter=xmalloc(sizeof(FILTER_list));
+	temp_filter->filter=clientFilterOnUnload;
+	temp_filter->data=http;
+	/* cbDataLock(http); ? */
+	dlinkAddTail(temp_filter, &temp_filter->node, &http->repfilters);
+#endif
+
+
 #if 0
 	/* this can't be done until the process for locking/freeing buffers
  	 * is looked into
@@ -2334,7 +2440,6 @@
          * write it.
          */
         debug(33,8)("clientWriteComplete: Pushing queued data\n");
-        http->out.offset += http->mb.size;
         http->flags.reply_write_in_progress=1;
         comm_write_mbuf(http->conn->fd, http->mb, clientWriteComplete, http);
         /* zero the membuf - the data is untouched */