---------------------
PatchSet 1201 
Date: 2005/01/26 19:51:33
Author: nlewycky
Branch: prefetching
Tag: (none) 
Log:
Moved analyzer out of HttpStateData into a clientStream in
clientReplyContext.
Don't try to analyze encoded content.

Members: 
	src/HTMLAnalysisStream.cc:1.1->1.1.2.1 
	src/HTMLAnalysisStream.h:1.1->1.1.2.1 
	src/HTMLPrefetchParser.cc:1.2.2.1->1.2.2.2(DEAD) 
	src/HTMLPrefetchParser.h:1.2.2.1->1.2.2.2(DEAD) 
	src/Makefile.am:1.57.2.2->1.57.2.3 
	src/PrefetchStream.cc:1.1.2.1->1.1.2.2 
	src/client_side_reply.cc:1.57->1.57.2.1 
	src/http.cc:1.43.2.1->1.43.2.2 
	src/http.h:1.10.2.1->1.10.2.2 

--- /dev/null	Wed Feb 14 13:33:00 2007
+++ squid3/src/HTMLAnalysisStream.cc	Wed Feb 14 13:34:10 2007
@@ -0,0 +1,383 @@
+
+/*
+ * $Id: HTMLAnalysisStream.cc,v 1.1.2.1 2005/01/26 19:51:33 nlewycky Exp $
+ *
+ * DEBUG: section 93    HTML parsing and fetching
+ * AUTHOR: Nick Lewycky
+ *
+ * SQUID Web Proxy Cache          http://www.squid-cache.org/
+ * ----------------------------------------------------------
+ *
+ *  Squid is the result of efforts by numerous individuals from
+ *  the Internet community; see the CONTRIBUTORS file for full
+ *  details.   Many organizations have provided support for Squid's
+ *  development; see the SPONSORS file for full details.  Squid is
+ *  Copyrighted (C) 2001 by the Regents of the University of
+ *  California; see the COPYRIGHT file for full details.  Squid
+ *  incorporates software developed and/or copyrighted by other
+ *  sources; see the CREDITS file for full details.
+ *
+ *  This program is free software; you can redistribute it and/or modify
+ *  it under the terms of the GNU General Public License as published by
+ *  the Free Software Foundation; either version 2 of the License, or
+ *  (at your option) any later version.
+ *  
+ *  This program is distributed in the hope that it will be useful,
+ *  but WITHOUT ANY WARRANTY; without even the implied warranty of
+ *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ *  GNU General Public License for more details.
+ *  
+ *  You should have received a copy of the GNU General Public License
+ *  along with this program; if not, write to the Free Software
+ *  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111, USA.
+ *
+ *  Copyright (c) 2004, Nick Lewycky <nicholas@mxc.ca>
+ *
+ */
+
+#include <stdexcept>
+#include <string>
+
+#include <libxml/SAX2.h>
+
+#include "squid.h"
+#include "Debug.h"
+#include "HttpReply.h"
+#include "HttpRequest.h"
+#include "HTMLAnalysisStream.h"
+#include "http.h"
+#include "PrefetchStream.h"
+#include "protos.h"
+
+using namespace std;
+
+CBDATA_CLASS_INIT(HTMLAnalysisStream);
+
+/* NOTE: this does get called after detach. */
+void
+htmlBufferData(clientStreamNode *node, ClientHttpRequest *req,
+               HttpReply *reply, StoreIOBuffer buffer)
+{
+    assert(node->data.getRaw());
+    debugs(93, 1, "bufferData");
+    clientStreamCallback(node, req, reply, buffer);
+    HTMLAnalysisStream::Pointer self =
+        dynamic_cast<HTMLAnalysisStream *>(node->data.getRaw());
+    if (!self.getRaw())
+    {
+        /* detach time. but if it's not my node->data, whose is it? */
+        return;
+    }
+
+    //self->parse(buffer.data, buffer.length, 1 /*?*/);
+    self->parse(node->next()->readBuffer.data + (reply?reply->hdr_sz:0),
+                node->next()->readBuffer.length - (reply?reply->hdr_sz:0),
+                1);
+}
+
+void
+htmlStreamRead(clientStreamNode *node, ClientHttpRequest *req)
+{
+    HTMLAnalysisStream::Pointer self =
+        dynamic_cast<HTMLAnalysisStream *>(node->data.getRaw());
+
+    debugs(93, 1, "before read");
+    clientStreamRead(node, req, /*self->buffer*/ node->next()->readBuffer);
+    debugs(93, 1, "after read");
+}
+
+void
+htmlStreamDetach(clientStreamNode *node, ClientHttpRequest *req)
+{
+  debugs(93, 1, "detach");
+  clientStreamDetach(node, req);
+}
+
+clientStream_status_t
+htmlStreamStatus(clientStreamNode *node, ClientHttpRequest *req)
+{
+    debugs(93, 1, "status");
+    return clientStreamStatus(node, req);
+}
+
+HTMLAnalysisStream::HTMLAnalysisStream(const HttpRequest *req)
+        :
+        request(req),
+        parser(htmlCreatePushParserCtxt(&handler, static_cast<void *>(this),
+                                        NULL, 0, NULL, XML_CHAR_ENCODING_NONE))
+{
+    if (!parser)
+        throw runtime_error("Unable to create parser.");
+
+    xmlSubstituteEntitiesDefault(1);
+
+    relative_url = urlCanonicalClean(request);
+
+    debugs(93, 1, "analyzing " << relative_url);
+}
+
+HTMLAnalysisStream::~HTMLAnalysisStream()
+{
+}
+
+inline void
+HTMLAnalysisStream::prefetch(const string &url)
+{
+    PrefetchStream::prefetch(url, relative_url, request);
+}
+
+void
+HTMLAnalysisStream::parse(const char *document, size_t len, bool partial)
+{
+    debugs(93, 1, "chunk to parse, " << len << " long " << (partial?"partial":""));
+    debugs(93, 1, "chunk is: " << string(document, len));
+    htmlParseChunk(parser, document, len, partial ? 1 : 0);
+}
+
+void
+HTMLAnalysisStream::start_element_handler(void *userData,
+        const xmlChar * xname, const xmlChar ** xattr)
+{
+    const char  *name = reinterpret_cast<const char  *>(xname);
+    const char **attr = reinterpret_cast<const char **>(xattr);
+
+    if (!name || !attr)
+        return;
+
+    HTMLAnalysisStream *self = static_cast<HTMLAnalysisStream *>(userData);
+
+    // Example:
+    // <img src="https://squid.sourceforge.net/changesets/squid3/patches/foo.png">
+    // <frame src="https://squid.sourceforge.net/changesets/squid3/patches/foo.html">
+    // <iframe src="https://squid.sourceforge.net/changesets/squid3/patches/foo.html">
+    if (strcasecmp(name, "img") == 0 || strcasecmp(name, "frame") == 0 ||
+        strcasecmp(name, "iframe") == 0) {
+        while (*attr) {
+            if (strcasecmp(*attr, "src") == 0) {
+                self->prefetch(*(attr + 1));
+                return;
+            }
+
+            attr += 2;
+        }
+
+        return;
+    }
+
+    // Example:
+    // <link rel="stylesheet" href="https://squid.sourceforge.net/changesets/squid3/patches/foo.css">
+    // <link href="https://squid.sourceforge.net/changesets/squid3/patches/foo.html" rel="prefetch">
+    bool is_fetchable = false;
+    string href;
+    if (strcasecmp(name, "link") == 0) {
+        while (*attr) {
+            if (strcasecmp(*attr, "href") == 0) {
+                if (is_fetchable) {
+                    self->prefetch(*(attr + 1));
+                    return;
+                } else
+                    href = *(attr + 1);
+            } else if (strcasecmp(*attr, "rel") == 0) {
+                if (strcasecmp(*(attr + 1), "stylesheet") != 0 &&
+                        strcasecmp(*(attr + 1), "prefetch") != 0)
+                    return;
+
+                if (!href.empty()) {
+                    self->prefetch(*(attr + 1));
+                    return;
+                } else
+                    is_fetchable = true;
+            }
+
+            attr += 2;
+        }
+
+        return;
+    }
+
+    // Example:
+    // <meta http-equiv="link" content="<foo.html>; rel=prefetch">
+    bool is_httpequiv = false;
+
+    if (strcasecmp(name, "meta") == 0) {
+        while (*attr) {
+            if (strcasecmp(*attr, "content") == 0) {
+                // TODO: replace this with some incantation of httpHeaderParse?
+
+                /* From RFC 2068:
+                 *
+                 * Link           = "Link" ":" #("<" URI ">" *( ";" link-param )
+                 *
+                 * link-param     = ( ( "rel" "=" relationship )
+                 *                    | ( "rev" "=" relationship )
+                 *                    | ( "title" "=" quoted-string )
+                 *                    | ( "anchor" "=" <"> URI <"> )
+                 *                    | ( link-extension ) )
+                 *
+                 * link-extension = token [ "=" ( token | quoted-string ) ]
+                 *
+                 * relationship   = sgml-name
+                 *                | ( <"> sgml-name *( SP sgml-name) <"> )
+                 *
+                 * sgml-name      = ALPHA *( ALPHA | DIGIT | "." | "-" )
+                 */
+
+                const char *p1 = *(attr + 1), *p2;
+
+                if (!p1)
+                    return;     // eg. <meta content>
+
+                while (*p1 && *p1 == ' ')
+                    p1++;
+
+                if (*p1 == '<')
+                    p1++;
+                else
+                    return;     // malformed (no leading '<')
+
+                p2 = p1;
+
+                while (*p2 && *p2 != '>' && strncasecmp(p2, "&gt;", 4) != 0)
+                    p2++;
+
+                if (!*p2)
+                    return;     // malformed (end of string before '>')
+
+                href = string(p1, p2);
+
+                p2++;   // skip over closing '>'
+
+                while (*p2 == ';' && !is_fetchable) {
+                    if (!*p2 == ';')
+                        return;
+
+                    p2++;       // skip over separating ';'
+
+                    while (*p2 && *p2 == ' ')
+                        p2++;
+
+                    if (strncasecmp(p2, "rel", 3) == 0) {
+                        p2 += 3;        // skip over 'rel'
+
+                        while (*p2 == ' ')
+                            p2++;
+
+                        if (*p2 != '=')
+                            return;     // malformed (no '=' in link-param)
+
+                        p2++;   // skip over '='
+
+                        while (*p2 == ' ')
+                            p2++;
+
+                        bool is_quoted = *p2 == '"';
+
+                        if (is_quoted)
+                            p2++;
+
+                        if (strncasecmp(p2, "prefetch", 8) == 0) {
+                            p2 += 8;
+
+                            if (is_quoted && *p2 != '"')
+                                return;
+                            else {
+                                if (is_httpequiv) {
+                                    self->prefetch(href);
+                                    return;
+                                } else
+                                    is_fetchable = true;
+                            }
+                        } else
+                            return;
+                    } else {
+                        while (*p2 && *p2 != '=')
+                            p2++;
+
+                        p2++;
+
+                        bool is_quoted = *p2 == '"';
+
+                        if (is_quoted) {
+                            while (*p2 && *p2 != '"')
+                                p2++;
+
+                            while (*p2 == ' ')
+                                p2++;
+
+                            p2++;
+                        } else
+                            while (*p2 && *p2 != ' ')
+                                p2++;
+                    }
+                }
+
+                if (!is_fetchable)
+                    return;     // no [further] link-param, hence no rel=prefetch
+
+            } else if (strcasecmp(*attr, "http-equiv") == 0 &&
+                       strcasecmp(*(attr + 1), "link") == 0) {
+                if (!href.empty()) {
+                    self->prefetch(href);
+                    return;
+                } else
+                    is_httpequiv = true;
+            }
+
+            attr += 2;
+        }
+
+        return;
+    }
+
+    // Example:
+    // <base href="">
+    if (strcasecmp(name, "base") == 0) {
+        while (*attr) {
+            if (strcasecmp(*attr, "href") == 0) {
+                self->relative_url = *(attr + 1);
+                return;
+            }
+
+            attr += 2;
+        }
+
+        return;
+    }
+}
+
+htmlSAXHandler
+HTMLAnalysisStream::handler = {
+    NULL,                       // internal subset
+    NULL,                       // is standalone
+    NULL,                       // has internal subset
+    NULL,                       // has external subset
+    NULL,                       // resolve entity
+    NULL,                       // get entity
+    NULL,                       // entity declaration
+    NULL,                       // notation declaration
+    NULL,                       // attribute declaration
+    NULL,                       // element declaraion
+    NULL,                       // unparsed entity
+    NULL,                       // set document locator
+    NULL,                       // start document
+    NULL,                       // end document
+    HTMLAnalysisStream::start_element_handler,  // start element
+    NULL,                       // end element
+    NULL,                       // reference
+    NULL,                       // characters
+    NULL,                       // ignorable whitespace
+    NULL,                       // processing instruction
+    NULL,                       // comment
+    NULL,                       // warning
+    NULL,                       // error
+    NULL,                       // fatal error
+    NULL,                       // get parameter entity
+    NULL,                       // cdata block
+    NULL,                       // external subset
+    XML_SAX2_MAGIC,
+    NULL,                       // private
+    NULL,                       // start element namespace
+    NULL,                       // end element namespace
+    NULL                        // xml structured error
+};
+
--- /dev/null	Wed Feb 14 13:33:00 2007
+++ squid3/src/HTMLAnalysisStream.h	Wed Feb 14 13:34:10 2007
@@ -0,0 +1,66 @@
+#ifndef SQUID_HTMLANALYSISSTREAM_H
+#define SQUID_HTMLANALYSISSTREAM_H
+
+#include <set>
+#include <string>
+
+#include <libxml/HTMLparser.h>
+#include <libxml/SAX2.h>
+
+#include "defines.h"
+#include "clientStream.h"
+#include "client_side_request.h"
+#include "Store.h"
+#include "StoreClient.h"
+
+class HttpRequest;
+
+void htmlBufferData(clientStreamNode *, ClientHttpRequest *,
+                    HttpReply *, StoreIOBuffer);
+void
+htmlStreamRead(clientStreamNode *node, ClientHttpRequest *req);
+void
+htmlStreamDetach(clientStreamNode *node, ClientHttpRequest *req);
+clientStream_status_t
+htmlStreamStatus(clientStreamNode *node, ClientHttpRequest *req);
+
+class HTMLAnalysisStream : public RefCountable
+{
+    friend void
+    htmlBufferData(clientStreamNode *, ClientHttpRequest *,
+                        HttpReply *, StoreIOBuffer);
+    friend void
+    htmlStreamRead(clientStreamNode *node, ClientHttpRequest *req);
+    friend void
+    htmlStreamDetach(clientStreamNode *node, ClientHttpRequest *req);
+    friend clientStream_status_t
+    htmlStreamStatus(clientStreamNode *node, ClientHttpRequest *req);
+
+
+public:
+    typedef RefCount<HTMLAnalysisStream> Pointer;
+
+    explicit HTMLAnalysisStream(const HttpRequest *);
+    ~HTMLAnalysisStream();
+
+private:
+    void prefetch(const std::string &);
+
+    void parse(const char *document, size_t len, bool partial);
+
+    static void start_element_handler(void *, const xmlChar *,
+	const xmlChar **);
+
+    const HttpRequest *request;
+
+    StoreIOBuffer buffer;
+
+    htmlParserCtxtPtr parser;
+    std::string relative_url;
+
+    static htmlSAXHandler handler;
+
+    CBDATA_CLASS2(HTMLAnalysisStream);
+};
+
+#endif // header guard
--- squid3/src/HTMLPrefetchParser.cc	Wed Feb 14 13:34:10 2007
+++ /dev/null	Wed Feb 14 13:33:00 2007
@@ -1,347 +0,0 @@
-
-/*
- * $Id: HTMLPrefetchParser.cc,v 1.2.2.1 2005/01/13 03:53:04 nlewycky Exp $
- *
- * DEBUG: section 93    HTML parsing and fetching
- * AUTHOR: Nick Lewycky
- *
- * SQUID Web Proxy Cache          http://www.squid-cache.org/
- * ----------------------------------------------------------
- *
- *  Squid is the result of efforts by numerous individuals from
- *  the Internet community; see the CONTRIBUTORS file for full
- *  details.   Many organizations have provided support for Squid's
- *  development; see the SPONSORS file for full details.  Squid is
- *  Copyrighted (C) 2001 by the Regents of the University of
- *  California; see the COPYRIGHT file for full details.  Squid
- *  incorporates software developed and/or copyrighted by other
- *  sources; see the CREDITS file for full details.
- *
- *  This program is free software; you can redistribute it and/or modify
- *  it under the terms of the GNU General Public License as published by
- *  the Free Software Foundation; either version 2 of the License, or
- *  (at your option) any later version.
- *  
- *  This program is distributed in the hope that it will be useful,
- *  but WITHOUT ANY WARRANTY; without even the implied warranty of
- *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- *  GNU General Public License for more details.
- *  
- *  You should have received a copy of the GNU General Public License
- *  along with this program; if not, write to the Free Software
- *  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111, USA.
- *
- *  Copyright (c) 2004, Nick Lewycky <nicholas@mxc.ca>
- *
- */
-
-#include <stdexcept>
-#include <string>
-
-#include <libxml/SAX2.h>
-
-#include "squid.h"
-#include "Debug.h"
-#include "HttpReply.h"
-#include "HttpRequest.h"
-#include "HTMLPrefetchParser.h"
-#include "http.h"
-#include "PrefetchStream.h"
-#include "protos.h"
-
-using namespace std;
-
-HTMLPrefetchParser::HTMLPrefetchParser(const HttpStateData * state)
-        :
-        stateData(state),
-        parser(NULL)
-{}
-
-HTMLPrefetchParser::~HTMLPrefetchParser()
-{
-    if (parser)
-        htmlFreeParserCtxt(parser);
-}
-
-void
-HTMLPrefetchParser::init()
-{
-    if (parser)
-        return;
-
-    parser = htmlCreatePushParserCtxt(&handler, static_cast<void *>(this),
-                                      NULL, 0, NULL, XML_CHAR_ENCODING_NONE);
-
-    if (!parser)
-        throw runtime_error("Unable to create parser.");
-
-    xmlSubstituteEntitiesDefault(1);
-
-    relative_url = urlCanonical(stateData->orig_request);
-
-    debugs(93, 1, "analyzing " << relative_url);
-}
-
-inline void
-HTMLPrefetchParser::prefetch(const string &url)
-{
-    PrefetchStream::prefetch(url, relative_url, stateData->orig_request);
-}
-
-void
-HTMLPrefetchParser::real_parse(const char *document, size_t len, bool partial)
-{
-    debugs(93, 1, "parse to chunk, " << len << " long " << (partial?"partial":""));
-    debugs(93, 1, "chunk is: " << string(document, len));
-    htmlParseChunk(parser, document, len, partial ? 1 : 0);
-}
-
-/* This code is ugly for two reasons. One is the need to inter-operate
- * with a C library. The other is a need for high-performance leading
- * to lots of short circuiting.
- */
-void
-HTMLPrefetchParser::start_element_handler(void *userData,
-        const xmlChar * xname, const xmlChar ** xattr)
-{
-    const char  *name = reinterpret_cast<const char  *>(xname);
-    const char **attr = reinterpret_cast<const char **>(xattr);
-
-    if (!name || !attr)
-        return;
-
-    HTMLPrefetchParser *self = static_cast<HTMLPrefetchParser *>(userData);
-
-    // Example:
-    // <img src="http://www.example.com/foo/foo.png">
-    if (strcasecmp(name, "img") == 0) {
-        while (*attr) {
-            if (strcasecmp(name, "img") == 0 &&
-                strcasecmp(*attr, "src") == 0) {
-                self->prefetch(*(attr + 1));
-                return;
-            }
-
-            attr += 2;
-        }
-
-        return;
-    }
-
-    // Example:
-    // <link rel="stylesheet" href="http://www.example.com/foo/foo.css">
-    // <link href="http://www.example.com/foo/foo.html" rel="prefetch">
-    bool is_fetchable = false;
-    string href;
-    if (strcasecmp(name, "link") == 0) {
-        while (*attr) {
-            if (strcasecmp(*attr, "href") == 0) {
-                if (is_fetchable) {
-                    self->prefetch(*(attr + 1));
-                    return;
-                } else
-                    href = *(attr + 1);
-            } else if (strcasecmp(*attr, "rel") == 0) {
-                if (strcasecmp(*(attr + 1), "stylesheet") != 0 &&
-                        strcasecmp(*(attr + 1), "prefetch") != 0)
-                    return;
-
-                if (!href.empty()) {
-                    self->prefetch(*(attr + 1));
-                    return;
-                } else
-                    is_fetchable = true;
-            }
-
-            attr += 2;
-        }
-
-        return;
-    }
-
-    // Example:
-    // <meta http-equiv="link" content="<foo.html>; rel=prefetch">
-    bool is_httpequiv = false;
-
-    if (strcasecmp(name, "meta") == 0) {
-        while (*attr) {
-            if (strcasecmp(*attr, "content") == 0) {
-                // TODO: replace this with some incantation of httpHeaderParse?
-
-                /* From RFC 2068:
-                 *
-                 * Link           = "Link" ":" #("<" URI ">" *( ";" link-param )
-                 *
-                 * link-param     = ( ( "rel" "=" relationship )
-                 *                    | ( "rev" "=" relationship )
-                 *                    | ( "title" "=" quoted-string )
-                 *                    | ( "anchor" "=" <"> URI <"> )
-                 *                    | ( link-extension ) )
-                 *
-                 * link-extension = token [ "=" ( token | quoted-string ) ]
-                 *
-                 * relationship   = sgml-name
-                 *                | ( <"> sgml-name *( SP sgml-name) <"> )
-                 *
-                 * sgml-name      = ALPHA *( ALPHA | DIGIT | "." | "-" )
-                 */
-
-                const char *p1 = *(attr + 1), *p2;
-
-                if (!p1)
-                    return;     // eg. <meta content>
-
-                while (*p1 && *p1 == ' ')
-                    p1++;
-
-                if (*p1 == '<')
-                    p1++;
-                else
-                    return;     // malformed (no leading '<')
-
-                p2 = p1;
-
-                while (*p2 && *p2 != '>' && strncasecmp(p2, "&gt;", 4) != 0)
-                    p2++;
-
-                if (!*p2)
-                    return;     // malformed (end of string before '>')
-
-                href = string(p1, p2);
-
-                p2++;   // skip over closing '>'
-
-                while (*p2 == ';' && !is_fetchable) {
-                    if (!*p2 == ';')
-                        return;
-
-                    p2++;       // skip over separating ';'
-
-                    while (*p2 && *p2 == ' ')
-                        p2++;
-
-                    if (strncasecmp(p2, "rel", 3) == 0) {
-                        p2 += 3;        // skip over 'rel'
-
-                        while (*p2 == ' ')
-                            p2++;
-
-                        if (*p2 != '=')
-                            return;     // malformed (no '=' in link-param)
-
-                        p2++;   // skip over '='
-
-                        while (*p2 == ' ')
-                            p2++;
-
-                        bool is_quoted = *p2 == '"';
-
-                        if (is_quoted)
-                            p2++;
-
-                        if (strncasecmp(p2, "prefetch", 8) == 0) {
-                            p2 += 8;
-
-                            if (is_quoted && *p2 != '"')
-                                return;
-                            else {
-                                if (is_httpequiv) {
-                                    self->prefetch(href);
-                                    return;
-                                } else
-                                    is_fetchable = true;
-                            }
-                        } else
-                            return;
-                    } else {
-                        while (*p2 && *p2 != '=')
-                            p2++;
-
-                        p2++;
-
-                        bool is_quoted = *p2 == '"';
-
-                        if (is_quoted) {
-                            while (*p2 && *p2 != '"')
-                                p2++;
-
-                            while (*p2 == ' ')
-                                p2++;
-
-                            p2++;
-                        } else
-                            while (*p2 && *p2 != ' ')
-                                p2++;
-                    }
-                }
-
-                if (!is_fetchable)
-                    return;     // no [further] link-param, hence no rel=prefetch
-
-            } else if (strcasecmp(*attr, "http-equiv") == 0 &&
-                       strcasecmp(*(attr + 1), "link") == 0) {
-                if (!href.empty()) {
-                    self->prefetch(href);
-                    return;
-                } else
-                    is_httpequiv = true;
-            }
-
-            attr += 2;
-        }
-
-        return;
-    }
-
-    // Example:
-    // <base href="">
-    if (strcasecmp(name, "base") == 0) {
-        while (*attr) {
-            if (strcasecmp(*attr, "href") == 0) {
-                self->relative_url = *(attr + 1);
-                return;
-            }
-
-            attr += 2;
-        }
-
-        return;
-    }
-}
-
-htmlSAXHandler
-HTMLPrefetchParser::handler = {
-    NULL,                       // internal subset
-    NULL,                       // is standalone
-    NULL,                       // has internal subset
-    NULL,                       // has external subset
-    NULL,                       // resolve entity
-    NULL,                       // get entity
-    NULL,                       // entity declaration
-    NULL,                       // notation declaration
-    NULL,                       // attribute declaration
-    NULL,                       // element declaraion
-    NULL,                       // unparsed entity
-    NULL,                       // set document locator
-    NULL,                       // start document
-    NULL,                       // end document
-    HTMLPrefetchParser::start_element_handler,  // start element
-    NULL,                       // end element
-    NULL,                       // reference
-    NULL,                       // characters
-    NULL,                       // ignorable whitespace
-    NULL,                       // processing instruction
-    NULL,                       // comment
-    NULL,                       // warning
-    NULL,                       // error
-    NULL,                       // fatal error
-    NULL,                       // get parameter entity
-    NULL,                       // cdata block
-    NULL,                       // external subset
-    XML_SAX2_MAGIC,
-    NULL,                       // private
-    NULL,                       // start element namespace
-    NULL,                       // end element namespace
-    NULL                        // xml structured error
-};
-
--- squid3/src/HTMLPrefetchParser.h	Wed Feb 14 13:34:10 2007
+++ /dev/null	Wed Feb 14 13:33:00 2007
@@ -1,48 +0,0 @@
-#ifndef SQUID_HTMLPREFETCHPARSER_H
-#define SQUID_HTMLPREFETCHPARSER_H
-
-#include <set>
-#include <string>
-
-#include <libxml/HTMLparser.h>
-#include <libxml/SAX2.h>
-
-#include "defines.h"
-#include "clientStream.h"
-#include "Store.h"
-#include "StoreClient.h"
-
-class HttpStateData;
-
-class HTMLPrefetchParser
-{
-public:
-    explicit HTMLPrefetchParser(const HttpStateData *);
-    ~HTMLPrefetchParser();
-
-    void init();
-
-    inline void parse(const char *document, size_t len, bool partial)
-    {
-	if (!parser)
-	    return;
-	real_parse(document, len, partial);
-    }
-
-private:
-    void prefetch(const std::string &);
-
-    void real_parse(const char *document, size_t len, bool partial);
-
-    static void start_element_handler(void *, const xmlChar *,
-	const xmlChar **);
-
-    const HttpStateData *stateData;
-
-    htmlParserCtxtPtr parser;
-    std::string relative_url;
-
-    static htmlSAXHandler handler;
-};
-
-#endif // header guard
Index: squid3/src/Makefile.am
===================================================================
RCS file: /cvsroot/squid-sf//squid3/src/Makefile.am,v
retrieving revision 1.57.2.2
retrieving revision 1.57.2.3
diff -u -r1.57.2.2 -r1.57.2.3
--- squid3/src/Makefile.am	13 Jan 2005 03:53:04 -0000	1.57.2.2
+++ squid3/src/Makefile.am	26 Jan 2005 19:51:33 -0000	1.57.2.3
@@ -1,7 +1,7 @@
 #
 #  Makefile for the Squid Object Cache server
 #
-#  $Id: Makefile.am,v 1.57.2.2 2005/01/13 03:53:04 nlewycky Exp $
+#  $Id: Makefile.am,v 1.57.2.3 2005/01/26 19:51:33 nlewycky Exp $
 #
 #  Uncomment and customize the following to suit your needs:
 #
@@ -93,8 +93,8 @@
 endif
 
 HTMLPREFETCH_ALL_SOURCE = \
-	HTMLPrefetchParser.cc \
-	HTMLPrefetchParser.h \
+	HTMLAnalysisStream.cc \
+	HTMLAnalysisStream.h \
 	PrefetchStream.cc \
 	PrefetchStream.h
 if USE_HTMLPREFETCH
Index: squid3/src/PrefetchStream.cc
===================================================================
RCS file: /cvsroot/squid-sf//squid3/src/Attic/PrefetchStream.cc,v
retrieving revision 1.1.2.1
retrieving revision 1.1.2.2
diff -u -r1.1.2.1 -r1.1.2.2
--- squid3/src/PrefetchStream.cc	13 Jan 2005 03:53:05 -0000	1.1.2.1
+++ squid3/src/PrefetchStream.cc	26 Jan 2005 19:51:34 -0000	1.1.2.2
@@ -1,6 +1,6 @@
 
 /*
- * $Id: PrefetchStream.cc,v 1.1.2.1 2005/01/13 03:53:05 nlewycky Exp $
+ * $Id: PrefetchStream.cc,v 1.1.2.2 2005/01/26 19:51:34 nlewycky Exp $
  *
  * DEBUG: section 93    HTML parsing and fetching
  * AUTHOR: Nick Lewycky
@@ -146,6 +146,11 @@
     dynamic_cast<PrefetchStream *>(node->data.getRaw())->detach(node, req);
 }
 
+/* OH NO! This isn't supposed to be akin to esiStreamRead, it's
+   actually esiProcessStream! Well that explains a lot.
+   NOTE: remember to review this code and make sure I've whacked the
+   stupid out of it.
+*/
 void
 PrefetchStream::bufferData(clientStreamNode *node, ClientHttpRequest *req,
                            HttpReply *reply, StoreIOBuffer buffer)
Index: squid3/src/client_side_reply.cc
===================================================================
RCS file: /cvsroot/squid-sf//squid3/src/client_side_reply.cc,v
retrieving revision 1.57
retrieving revision 1.57.2.1
diff -u -r1.57 -r1.57.2.1
--- squid3/src/client_side_reply.cc	21 Dec 2004 16:37:17 -0000	1.57
+++ squid3/src/client_side_reply.cc	26 Jan 2005 19:51:34 -0000	1.57.2.1
@@ -1,6 +1,6 @@
 
 /*
- * $Id: client_side_reply.cc,v 1.57 2004/12/21 16:37:17 squidadm Exp $
+ * $Id: client_side_reply.cc,v 1.57.2.1 2005/01/26 19:51:34 nlewycky Exp $
  *
  * DEBUG: section 88    Client-side Reply Routines
  * AUTHOR: Robert Collins (Originally Duane Wessels in client_side.c)
@@ -52,6 +52,9 @@
 #include "DelayPools.h"
 #endif
 #include "client_side.h"
+#if USE_HTMLPREFETCH
+#include "HTMLAnalysisStream.h"
+#endif
 
 CBDATA_CLASS_INIT(clientReplyContext);
 
@@ -1920,6 +1923,26 @@
     ("clientReplyContext::sendMoreData: Appending %d bytes after %d bytes of headers\n",
      (int) body_size, rep->hdr_sz);
 
+#if USE_HTMLPREFETCH
+
+    if (Config.onoff.htmlPrefetch) {
+        const String content_type = http->storeEntry()->getReply()->content_type;
+
+        if (content_type.buf() &&
+            (content_type.caseCmp("text/html") == 0 ||
+             content_type.caseCmp("application/xhtml+xml") == 0) &&
+            !httpHeaderHas(&http->storeEntry()->getReply()->header,
+                           HDR_CONTENT_ENCODING) &&
+            !http->request->range) {
+            HTMLAnalysisStream *htmlas = new HTMLAnalysisStream(http->request);
+            clientStreamInsertHead(&http->client_stream, htmlStreamRead,
+                                   htmlBufferData, htmlStreamDetach,
+                                   htmlStreamStatus, htmlas);
+        }
+    }
+
+#endif
+
 #if ESI
 
     if (http->flags.accel && rep->sline.status != HTTP_FORBIDDEN &&
Index: squid3/src/http.cc
===================================================================
RCS file: /cvsroot/squid-sf//squid3/src/http.cc,v
retrieving revision 1.43.2.1
retrieving revision 1.43.2.2
diff -u -r1.43.2.1 -r1.43.2.2
--- squid3/src/http.cc	12 Jan 2005 21:58:29 -0000	1.43.2.1
+++ squid3/src/http.cc	26 Jan 2005 19:51:36 -0000	1.43.2.2
@@ -1,6 +1,6 @@
 
 /*
- * $Id: http.cc,v 1.43.2.1 2005/01/12 21:58:29 nlewycky Exp $
+ * $Id: http.cc,v 1.43.2.2 2005/01/26 19:51:36 nlewycky Exp $
  *
  * DEBUG: section 11    Hypertext Transfer Protocol (HTTP)
  * AUTHOR: Harvest Derived
@@ -35,7 +35,7 @@
 
 /*
  * Anonymizing patch by lutz@as-node.jena.thur.de
- * have a look into http-anon.c to get more informations.
+ * have a look into http-anon.c to get more information.
  */
 
 #include "squid.h"
@@ -50,9 +50,6 @@
 #if DELAY_POOLS
 #include "DelayPools.h"
 #endif
-#if USE_HTMLPREFETCH
-#include "HTMLPrefetchParser.h"
-#endif
 
 #include <memory>
 
@@ -1064,17 +1061,6 @@
             }
         }
 
-#if USE_HTMLPREFETCH
-        if (Config.onoff.htmlPrefetch) {
-            String content_type = entry->getReply()->content_type;
-
-            if (content_type.buf() &&
-                    (content_type.caseCmp("text/html") == 0 ||
-                     content_type.caseCmp("application/xhtml+xml") == 0))
-                htmlpp->init();
-        }
-#endif
-
         PROF_start(HttpStateData_processReplyData);
         processReplyData(buf, len);
         PROF_stop(HttpStateData_processReplyData);
@@ -1101,11 +1087,6 @@
          */
 
         if (len > end) {
-#if USE_HTMLPREFETCH
-            htmlpp->parse(buf + end, len - end,
-                          persistentConnStatus() == INCOMPLETE_MSG);
-#endif
-
             tempBuffer.data = (char *) buf + end;
             tempBuffer.length = len - end;
             tempBuffer.offset = currentOffset;
@@ -1115,10 +1096,6 @@
 
         flags.headers_pushed = 1;
     } else {
-#if USE_HTMLPREFETCH
-        htmlpp->parse(buf, len, persistentConnStatus() == INCOMPLETE_MSG);
-#endif
-
         tempBuffer.data = (char *) buf;
         tempBuffer.length = len;
         tempBuffer.offset = currentOffset;
@@ -1723,11 +1700,6 @@
     httpState->entry = fwd->entry;
     httpState->fd = fd;
 
-#if USE_HTMLPREFETCH
-    auto_ptr<HTMLPrefetchParser> htmlpp(new HTMLPrefetchParser(httpState));
-    httpState->htmlpp = htmlpp;
-#endif
-
     if (fwd->servers)
         httpState->_peer = fwd->servers->_peer;	/* might be NULL */
 
Index: squid3/src/http.h
===================================================================
RCS file: /cvsroot/squid-sf//squid3/src/http.h,v
retrieving revision 1.10.2.1
retrieving revision 1.10.2.2
diff -u -r1.10.2.1 -r1.10.2.2
--- squid3/src/http.h	12 Jan 2005 21:58:30 -0000	1.10.2.1
+++ squid3/src/http.h	26 Jan 2005 19:51:36 -0000	1.10.2.2
@@ -1,6 +1,6 @@
 
 /*
- * $Id: http.h,v 1.10.2.1 2005/01/12 21:58:30 nlewycky Exp $
+ * $Id: http.h,v 1.10.2.2 2005/01/26 19:51:36 nlewycky Exp $
  *
  *
  * SQUID Web Proxy Cache          http://www.squid-cache.org/
@@ -34,15 +34,9 @@
 #ifndef SQUID_HTTP_H
 #define SQUID_HTTP_H
 
-#include <memory>
-
 #include "StoreIOBuffer.h"
 #include "comm.h"
 
-#if USE_HTMLPREFETCH
-class HTMLPrefetchParser;
-#endif
-
 class HttpStateData
 {
 
@@ -73,10 +67,6 @@
     bool surrogateNoStore;
     void processSurrogateControl(HttpReply *);
 
-#if USE_HTMLPREFETCH
-         std::auto_ptr<HTMLPrefetchParser> htmlpp;
-#endif
-
   private:
     enum ConnectionStatus {
 	INCOMPLETE_MSG,