moved to pub

git-svn-id: svn://10.0.0.236/trunk@165 18797224-902f-48f8-a5cc-f745e15eee43

mozilla/parser/htmlparser/robot/Makefile

@@ -0,0 +1,54 @@
+#!gmake
+#
+# The contents of this file are subject to the Netscape Public License
+# Version 1.0 (the "NPL"); you may not use this file except in
+# compliance with the NPL.  You may obtain a copy of the NPL at
+# http://www.mozilla.org/NPL/
+#
+# Software distributed under the NPL is distributed on an "AS IS" basis,
+# WITHOUT WARRANTY OF ANY KIND, either express or implied. See the NPL
+# for the specific language governing rights and limitations under the
+# NPL.
+#
+# The Initial Developer of this code under the NPL is Netscape
+# Communications Corporation.  Portions created by Netscape are
+# Copyright (C) 1998 Netscape Communications Corporation.  All Rights
+# Reserved.
+
+DEPTH = ../../..
+
+include $(DEPTH)/config/config.mk
+
+CPPSRCS =			\
+	RobotMain.cpp		\
+	nsRobotSink.cpp		\
+	$(NULL)
+
+REQUIRES=xpcom raptor
+
+OBJS	= $(CPPSRCS:.cpp=.o)
+
+EX_LIBS = \
+	$(DIST)/lib/libraptorbase.a \
+	$(DIST)/lib/libraptorhtmlpars.a \
+	$(DIST)/lib/libnetlib.a \
+	$(DIST)/lib/libpref.a	\
+	$(DIST)/lib/libxp.a	\
+	$(DIST)/lib/libjs.a	\
+	$(DIST)/lib/libxpcom.a  \
+	$(DIST)/lib/libplc21.a	\
+	$(DIST)/lib/libplds21.a	\
+	$(DIST)/lib/libnspr21.a	\
+	-lXm			\
+	-lm			\
+	$(NULL)
+
+PROG = htmlrobot
+
+$(PROG):$(OBJS)
+	@$(MAKE_OBJDIR)
+	$(CCC) -o $@ $(OBJS) $(LDFLAGS) $(EX_LIBS) $(OS_LIBS)
+
+TARGETS = $(PROG)
+
+include $(DEPTH)/config/rules.mk

mozilla/parser/htmlparser/robot/RobotMain.cpp

@@ -0,0 +1,111 @@
+/* -*- Mode: C++; tab-width: 2; indent-tabs-mode: nil; c-basic-offset: 2 -*- */
+/*
+ * The contents of this file are subject to the Netscape Public License
+ * Version 1.0 (the "NPL"); you may not use this file except in
+ * compliance with the NPL.  You may obtain a copy of the NPL at
+ * http://www.mozilla.org/NPL/
+ *
+ * Software distributed under the NPL is distributed on an "AS IS" basis,
+ * WITHOUT WARRANTY OF ANY KIND, either express or implied. See the NPL
+ * for the specific language governing rights and limitations under the
+ * NPL.
+ *
+ * The Initial Developer of this code under the NPL is Netscape
+ * Communications Corporation.  Portions created by Netscape are
+ * Copyright (C) 1998 Netscape Communications Corporation.  All Rights
+ * Reserved.
+ */
+#include "nsIRobotSink.h"
+#include "nsIRobotSinkObserver.h"
+#include "nsIParser.h"
+#include "nsVoidArray.h"
+#include "nsString.h"
+#include "nsIURL.h"
+
+static nsVoidArray* gWorkList;
+
+static NS_DEFINE_IID(kIRobotSinkObserverIID, NS_IROBOTSINKOBSERVER_IID);
+
+class RobotSinkObserver : public nsIRobotSinkObserver {
+public:
+  RobotSinkObserver() {
+    NS_INIT_REFCNT();
+  }
+
+  ~RobotSinkObserver() {
+  }
+
+  NS_DECL_ISUPPORTS
+
+  NS_IMETHOD ProcessLink(const nsString& aURLSpec);
+};
+
+NS_IMPL_ISUPPORTS(RobotSinkObserver, kIRobotSinkObserverIID);
+
+NS_IMETHODIMP RobotSinkObserver::ProcessLink(const nsString& aURLSpec)
+{
+  fputs(aURLSpec, stdout);
+  printf("\n");
+  return NS_OK;
+}
+
+//----------------------------------------------------------------------
+
+int main(int argc, char **argv)
+{
+  gWorkList = new nsVoidArray();
+
+  int i;
+  for (i = 1; i < argc; i++) {
+    gWorkList->AppendElement(new nsString(argv[i]));
+  }
+
+  RobotSinkObserver* myObserver = new RobotSinkObserver();
+  NS_ADDREF(myObserver);
+
+  for (;;) {
+    PRInt32 n = gWorkList->Count();
+    if (0 == n) {
+      break;
+    }
+    nsString* urlName = (nsString*) gWorkList->ElementAt(n - 1);
+    gWorkList->RemoveElementAt(n - 1);
+
+    // Create url
+    nsIURL* url;
+    nsresult rv = NS_NewURL(&url, *urlName);
+    if (NS_OK != rv) {
+      printf("invalid URL: '");
+      fputs(*urlName, stdout);
+      printf("'\n");
+      return -1;
+    }
+    delete urlName;
+
+    nsIParser* parser;
+    rv = NS_NewHTMLParser(&parser);
+    if (NS_OK != rv) {
+      printf("can't make parser\n");
+      return -1;
+    }
+
+    nsIRobotSink* sink;
+    rv = NS_NewRobotSink(&sink);
+    if (NS_OK != rv) {
+      printf("can't make parser\n");
+      return -1;
+    }
+    sink->Init(url);
+    sink->AddObserver(myObserver);
+
+    parser->SetContentSink(sink);
+    parser->Parse(url);
+    NS_RELEASE(sink);
+    NS_RELEASE(parser);
+    NS_RELEASE(url);
+  }
+
+  NS_RELEASE(myObserver);
+
+  return 0;
+}

mozilla/parser/htmlparser/robot/makefile.win

@@ -0,0 +1,62 @@
+#!nmake
+#
+# The contents of this file are subject to the Netscape Public License
+# Version 1.0 (the "NPL"); you may not use this file except in
+# compliance with the NPL.  You may obtain a copy of the NPL at
+# http://www.mozilla.org/NPL/
+#
+# Software distributed under the NPL is distributed on an "AS IS" basis,
+# WITHOUT WARRANTY OF ANY KIND, either express or implied. See the NPL
+# for the specific language governing rights and limitations under the
+# NPL.
+#
+# The Initial Developer of this code under the NPL is Netscape
+# Communications Corporation.  Portions created by Netscape are
+# Copyright (C) 1998 Netscape Communications Corporation.  All Rights
+# Reserved.
+
+DEPTH=..\..\.. 
+IGNORE_MANIFEST=1
+
+MAKE_OBJ_TYPE   = EXE
+PROGRAM = .\$(OBJDIR)\htmlrobot.exe
+
+MISCDEP=			\
+ $(DIST)\lib\raptorhtmlpars.lib	\
+ $(DIST)\lib\xpcom32.lib	\
+ $(LIBNSPR)			\
+ $(DIST)\lib\libplc21.lib	\
+ $(NULL)
+
+MYLIBS=				\
+ $(DIST)\lib\raptorhtmlpars.lib	\
+ $(DIST)\lib\raptorbase.lib	\
+ $(DIST)\lib\xpcom32.lib	\
+ $(LIBNSPR)			\
+ $(DIST)\lib\libplc21.lib	\
+ $(NULL)
+
+LLIBS= $(MYLIBS)                \
+ shell32.lib                    \
+ -SUBSYSTEM:CONSOLE
+
+DEFINES=-D_IMPL_NS_HTMLPARS 
+
+CPPSRCS=			\
+ nsRobotSink.cpp		\
+ RobotMain.cpp			\
+ $(NULL)
+
+CPP_OBJS=			\
+ .\$(OBJDIR)\nsRobotSink.obj	\
+ .\$(OBJDIR)\RobotMain.obj
+
+LINCS=-I$(XPDIST)\public\xpcom -I$(XPDIST)\public\raptor  -I..\src
+
+include <$(DEPTH)\config\rules.mak>
+
+install:: $(PROGRAM)
+	$(MAKE_INSTALL) $(PROGRAM) $(DIST)\bin
+
+clobber::
+	rm -f $(DIST)\bin\htmlrobot.exe

mozilla/parser/htmlparser/robot/nsIRobotSink.h

@@ -0,0 +1,39 @@
+/* -*- Mode: C++; tab-width: 2; indent-tabs-mode: nil; c-basic-offset: 2 -*- */
+/*
+ * The contents of this file are subject to the Netscape Public License
+ * Version 1.0 (the "NPL"); you may not use this file except in
+ * compliance with the NPL.  You may obtain a copy of the NPL at
+ * http://www.mozilla.org/NPL/
+ *
+ * Software distributed under the NPL is distributed on an "AS IS" basis,
+ * WITHOUT WARRANTY OF ANY KIND, either express or implied. See the NPL
+ * for the specific language governing rights and limitations under the
+ * NPL.
+ *
+ * The Initial Developer of this code under the NPL is Netscape
+ * Communications Corporation.  Portions created by Netscape are
+ * Copyright (C) 1998 Netscape Communications Corporation.  All Rights
+ * Reserved.
+ */
+#ifndef nsIRobotSink_h___
+#define nsIRobotSink_h___
+
+#include "nsIHTMLContentSink.h"
+class nsIURL;
+class nsIRobotSinkObserver;
+
+/* 61256800-cfd8-11d1-9328-00805f8add32 */
+#define NS_IROBOTSINK_IID     \
+{ 0x61256800, 0xcfd8, 0x11d1, \
+  {0x93, 0x28, 0x00, 0x80, 0x5f, 0x8a, 0xdd, 0x32} }
+
+class nsIRobotSink : public nsIHTMLContentSink {
+public:
+  NS_IMETHOD Init(nsIURL* aDocumentURL) = 0;
+  NS_IMETHOD AddObserver(nsIRobotSinkObserver* aObserver) = 0;
+  NS_IMETHOD RemoveObserver(nsIRobotSinkObserver* aObserver) = 0;
+};
+
+extern nsresult NS_NewRobotSink(nsIRobotSink** aInstancePtrResult);
+
+#endif /* nsIRobotSink_h___ */

mozilla/parser/htmlparser/robot/nsIRobotSinkObserver.h

@@ -0,0 +1,34 @@
+/* -*- Mode: C++; tab-width: 2; indent-tabs-mode: nil; c-basic-offset: 2 -*- */
+/*
+ * The contents of this file are subject to the Netscape Public License
+ * Version 1.0 (the "NPL"); you may not use this file except in
+ * compliance with the NPL.  You may obtain a copy of the NPL at
+ * http://www.mozilla.org/NPL/
+ *
+ * Software distributed under the NPL is distributed on an "AS IS" basis,
+ * WITHOUT WARRANTY OF ANY KIND, either express or implied. See the NPL
+ * for the specific language governing rights and limitations under the
+ * NPL.
+ *
+ * The Initial Developer of this code under the NPL is Netscape
+ * Communications Corporation.  Portions created by Netscape are
+ * Copyright (C) 1998 Netscape Communications Corporation.  All Rights
+ * Reserved.
+ */
+#ifndef nsIRobotSinkObserver_h___
+#define nsIRobotSinkObserver_h___
+
+#include "nsISupports.h"
+class nsString;
+
+/* fab1d970-cfda-11d1-9328-00805f8add32 */
+#define NS_IROBOTSINKOBSERVER_IID \
+{ 0xfab1d970, 0xcfda, 0x11d1,	  \
+  {0x93, 0x28, 0x00, 0x80, 0x5f, 0x8a, 0xdd, 0x32} }
+
+class nsIRobotSinkObserver : public nsISupports {
+public:
+  NS_IMETHOD ProcessLink(const nsString& aURLSpec) = 0;
+};
+
+#endif /* nsIRobotSinkObserver_h___ */

mozilla/parser/htmlparser/robot/nsRobotSink.cpp

@@ -0,0 +1,285 @@
+/* -*- Mode: C++; tab-width: 2; indent-tabs-mode: nil; c-basic-offset: 2 -*- */
+/*
+ * The contents of this file are subject to the Netscape Public License
+ * Version 1.0 (the "NPL"); you may not use this file except in
+ * compliance with the NPL.  You may obtain a copy of the NPL at
+ * http://www.mozilla.org/NPL/
+ *
+ * Software distributed under the NPL is distributed on an "AS IS" basis,
+ * WITHOUT WARRANTY OF ANY KIND, either express or implied. See the NPL
+ * for the specific language governing rights and limitations under the
+ * NPL.
+ *
+ * The Initial Developer of this code under the NPL is Netscape
+ * Communications Corporation.  Portions created by Netscape are
+ * Copyright (C) 1998 Netscape Communications Corporation.  All Rights
+ * Reserved.
+ */
+#include "nsIRobotSink.h"
+#include "nsIRobotSinkObserver.h"
+#include "nsIParserNode.h"
+#include "nsString.h"
+#include "nsIURL.h"
+#include "nsCRT.h"
+#include "nsVoidArray.h"
+class nsIDocument;
+
+// TODO
+// - add in base tag support
+// - get links from other sources:
+//      - LINK tag
+//      - STYLE SRC
+//      - IMG SRC
+//      - LAYER SRC
+
+static NS_DEFINE_IID(kISupportsIID, NS_ISUPPORTS_IID);
+static NS_DEFINE_IID(kIHTMLContentSinkIID, NS_IHTMLCONTENTSINK_IID);
+static NS_DEFINE_IID(kIRobotSinkIID, NS_IROBOTSINK_IID);
+
+class RobotSink : public nsIRobotSink {
+public:
+  RobotSink();
+  ~RobotSink();
+
+  void* operator new(size_t size) {
+    void* rv = ::operator new(size);
+    nsCRT::zero(rv, size);
+    return (void*) rv;
+  }
+
+  // nsISupports
+  NS_DECL_ISUPPORTS
+
+  // nsIHTMLContentSink
+  virtual PRBool SetTitle(const nsString& aValue);
+  virtual PRBool OpenHTML(const nsIParserNode& aNode);
+  virtual PRBool CloseHTML(const nsIParserNode& aNode);
+  virtual PRBool OpenHead(const nsIParserNode& aNode);
+  virtual PRBool CloseHead(const nsIParserNode& aNode);
+  virtual PRBool OpenBody(const nsIParserNode& aNode);
+  virtual PRBool CloseBody(const nsIParserNode& aNode);
+  virtual PRBool OpenForm(const nsIParserNode& aNode);
+  virtual PRBool CloseForm(const nsIParserNode& aNode);
+  virtual PRBool OpenFrameset(const nsIParserNode& aNode);
+  virtual PRBool CloseFrameset(const nsIParserNode& aNode);
+  virtual PRBool OpenContainer(const nsIParserNode& aNode);
+  virtual PRBool CloseContainer(const nsIParserNode& aNode);
+  virtual PRBool CloseTopmostContainer();
+  virtual PRBool AddLeaf(const nsIParserNode& aNode);
+
+  // nsIRobotSink
+  NS_IMETHOD Init(nsIURL* aDocumentURL);
+  NS_IMETHOD AddObserver(nsIRobotSinkObserver* aObserver);
+  NS_IMETHOD RemoveObserver(nsIRobotSinkObserver* aObserver);
+
+  void ProcessLink(const nsString& aLink);
+
+protected:
+  nsIURL* mDocumentURL;
+  nsVoidArray mObservers;
+};
+
+nsresult NS_NewRobotSink(nsIRobotSink** aInstancePtrResult)
+{
+  RobotSink* it = new RobotSink();
+  return it->QueryInterface(kIRobotSinkIID, (void**) aInstancePtrResult);
+}
+
+RobotSink::RobotSink()
+{
+}
+
+RobotSink::~RobotSink()
+{
+  NS_IF_RELEASE(mDocumentURL);
+  PRInt32 i, n = mObservers.Count();
+  for (i = 0; i < n; i++) {
+    nsIRobotSinkObserver* cop = (nsIRobotSinkObserver*)mObservers.ElementAt(i);
+    NS_RELEASE(cop);
+  }
+}  
+
+NS_IMPL_ADDREF(RobotSink);
+
+NS_IMPL_RELEASE(RobotSink);
+
+NS_IMETHODIMP RobotSink::QueryInterface(REFNSIID aIID, void** aInstancePtr)
+{
+  if (NULL == aInstancePtr) {
+    return NS_ERROR_NULL_POINTER;
+  }
+  if (aIID.Equals(kIRobotSinkIID)) {
+    *aInstancePtr = (void*) this;
+    AddRef();
+    return NS_OK;
+  }
+  if (aIID.Equals(kIHTMLContentSinkIID)) {
+    *aInstancePtr = (void*) this;
+    AddRef();
+    return NS_OK;
+  }
+  if (aIID.Equals(kISupportsIID)) {
+    *aInstancePtr = (void*) ((nsISupports*)this);
+    AddRef();
+    return NS_OK;
+  }
+  return NS_NOINTERFACE;
+}
+
+PRBool RobotSink::SetTitle(const nsString& aValue)
+{
+  return PR_TRUE;
+}
+
+PRBool RobotSink::OpenHTML(const nsIParserNode& aNode)
+{
+  return PR_TRUE;
+}
+
+PRBool RobotSink::CloseHTML(const nsIParserNode& aNode)
+{
+  return PR_TRUE;
+}
+
+PRBool RobotSink::OpenHead(const nsIParserNode& aNode)
+{
+  return PR_TRUE;
+}
+
+PRBool RobotSink::CloseHead(const nsIParserNode& aNode)
+{
+  return PR_TRUE;
+}
+
+PRBool RobotSink::OpenBody(const nsIParserNode& aNode)
+{
+  return PR_TRUE;
+}
+
+PRBool RobotSink::CloseBody(const nsIParserNode& aNode)
+{
+  return PR_TRUE;
+}
+
+PRBool RobotSink::OpenForm(const nsIParserNode& aNode)
+{
+  return PR_TRUE;
+}
+
+PRBool RobotSink::CloseForm(const nsIParserNode& aNode)
+{
+  return PR_TRUE;
+}
+
+PRBool RobotSink::OpenFrameset(const nsIParserNode& aNode)
+{
+  return PR_TRUE;
+}
+
+PRBool RobotSink::CloseFrameset(const nsIParserNode& aNode)
+{
+  return PR_TRUE;
+}
+
+PRBool RobotSink::OpenContainer(const nsIParserNode& aNode)
+{
+  nsAutoString tmp(aNode.GetText());
+  tmp.ToUpperCase();
+  if (tmp.Equals("A")) {
+    nsAutoString k, v;
+    PRInt32 ac = aNode.GetAttributeCount();
+    for (PRInt32 i = 0; i < ac; i++) {
+      // Get upper-cased key
+      const nsString& key = aNode.GetKeyAt(i);
+      k.Truncate();
+      k.Append(key);
+      k.ToUpperCase();
+      if (k.Equals("HREF")) {
+        // Get value and remove mandatory quotes
+        v.Truncate();
+        v.Append(aNode.GetValueAt(i));
+        PRUnichar first = v.First();
+        if ((first == '"') || (first == '\'')) {
+          if (v.Last() == first) {
+            v.Cut(0, 1);
+            PRInt32 pos = v.Length() - 1;
+            if (pos >= 0) {
+              v.Cut(pos, 1);
+            }
+          } else {
+            // Mismatched quotes - leave them in
+          }
+        }
+        ProcessLink(v);
+      }
+    }
+  }
+  return PR_TRUE;
+}
+
+PRBool RobotSink::CloseContainer(const nsIParserNode& aNode)
+{
+  return PR_TRUE;
+}
+
+PRBool RobotSink::CloseTopmostContainer()
+{
+  return PR_TRUE;
+}
+
+PRBool RobotSink::AddLeaf(const nsIParserNode& aNode)
+{
+  return PR_TRUE;
+}
+
+NS_IMETHODIMP RobotSink::Init(nsIURL* aDocumentURL)
+{
+  NS_IF_RELEASE(mDocumentURL);
+  mDocumentURL = aDocumentURL;
+  NS_IF_ADDREF(aDocumentURL);
+  return NS_OK;
+}
+
+NS_IMETHODIMP RobotSink::AddObserver(nsIRobotSinkObserver* aObserver)
+{
+  if (mObservers.AppendElement(aObserver)) {
+    NS_ADDREF(aObserver);
+    return NS_OK;
+  }
+  return NS_ERROR_OUT_OF_MEMORY;
+}
+
+NS_IMETHODIMP RobotSink::RemoveObserver(nsIRobotSinkObserver* aObserver)
+{
+  if (mObservers.RemoveElement(aObserver)) {
+    NS_RELEASE(aObserver);
+    return NS_OK;
+  }
+  //XXX return NS_ERROR_NOT_FOUND;
+  return NS_OK;
+}
+
+void RobotSink::ProcessLink(const nsString& aLink)
+{
+  nsAutoString absURLSpec(aLink);
+
+  // Make link absolute
+  // XXX base tag handling
+  nsIURL* docURL = mDocumentURL;
+  if (nsnull != docURL) {
+    nsIURL* absurl;
+    nsresult rv = NS_NewURL(&absurl, docURL, aLink);
+    if (NS_OK == rv) {
+      absURLSpec.Truncate();
+      absurl->ToString(absURLSpec);
+      NS_RELEASE(absurl);
+    }
+  }
+
+  // Now give link to robot observers
+  PRInt32 i, n = mObservers.Count();
+  for (i = 0; i < n; i++) {
+    nsIRobotSinkObserver* cop = (nsIRobotSinkObserver*)mObservers.ElementAt(i);
+    cop->ProcessLink(absURLSpec);
+  }
+}