bring it in sync to HEAD changes

configure.ac

@@ -255,6 +255,20 @@ elif test x$enable_gtk_test = xtry; then
 fi
 AM_CONDITIONAL(BUILD_GTK_TEST, test x$enable_gtk_test = xyes)
 
+AC_ARG_ENABLE(abiword-output,
+              AC_HELP_STRING([--disable-abiword-output],
+                             [Don't build the abiword backend.]),
+              enable_abiword_output=$enableval,
+              enable_abiword_output="try")
+if test x$enable_abiword_output = xyes; then
+  PKG_CHECK_MODULES(ABIWORD, libxml-2.0)
+elif test x$enable_abiword_output = xtry; then
+  PKG_CHECK_MODULES(ABIWORD, libxml-2.0,
+                    [enable_abiword_output="yes"],
+                    [enable_abiword_output="no"])
+fi
+
+AM_CONDITIONAL(BUILD_ABIWORD_OUTPUT, test x$enable_abiword_output = xyes)
 
 AC_ARG_ENABLE(utils,
               AC_HELP_STRING([--disable-utils],
@@ -321,6 +335,7 @@ echo ""
 echo "Building poppler with support for:"
 echo "  splash output:      $enable_splash_output"
 echo "  cairo output:       $enable_cairo_output"
+echo "  abiword output:     $enable_abiword_output"
 echo "  qt wrapper:         $enable_poppler_qt"
 echo "  qt4 wrapper:        $enable_poppler_qt4"
 echo "  glib wrapper:       $enable_poppler_glib"

poppler/ABWOutputDev.h

+//========================================================================
+//
+// ABWOutputDev.h
+//
+// Copyright 2006 Jauco Noordzij
+//
+//========================================================================
+
+#ifndef ABWOUTPUTDEV_H
+#define ABWOUTPUTDEV_H
+
+#ifdef __GNUC__
+#pragma interface
+#endif
+
+#include <stdio.h>
+#include "goo/gtypes.h"
+#include "goo/GooList.h"
+#include "GfxFont.h"
+#include "OutputDev.h"
+#include "Catalog.h"
+#include "UnicodeMap.h"
+#include "PDFDoc.h"
+
+#include <libxml/parser.h>
+#include <libxml/tree.h>
+#include <libxml/xpath.h>
+#include <libxml/xpathInternals.h>
+
+#ifdef WIN32
+#  define SLASH '\\'
+#else
+#  define SLASH '/'
+#endif
+
+#define xoutRound(x) ((int)(x + 0.5))
+
+class GfxState;
+class GooString;
+
+//------------------------------------------------------------------------
+// ABWOutputDev
+//------------------------------------------------------------------------
+
+class ABWOutputDev: public OutputDev {
+public:
+
+  // Open a text output file.  If <fileName> is NULL, no file is written
+  // (this is useful, e.g., for searching text).  If <useASCII7> is true,
+  // text is converted to 7-bit ASCII; otherwise, text is converted to
+  // 8-bit ISO Latin-1.  <useASCII7> should also be set for Japanese
+  // (EUC-JP) text.  If <rawOrder> is true, the text is kept in content
+  // stream order.
+  ABWOutputDev(xmlDocPtr ext_doc);
+
+  // Destructor.
+  virtual ~ABWOutputDev();
+
+  // Check if file was successfully created.
+  virtual GBool isOk() { return gTrue; }
+
+  //---- get info about output device
+
+  // Does this device use upside-down coordinates?
+  // (Upside-down means (0,0) is the top left corner of the page.)
+  virtual GBool upsideDown() { return gTrue; }
+
+  // Does this device use drawChar() or drawString()?
+  virtual GBool useDrawChar() { return gTrue; }
+
+  // Does this device use beginType3Char/endType3Char?  Otherwise,
+  // text in Type 3 fonts will be drawn with drawChar/drawString.
+  virtual GBool interpretType3Chars() { return gFalse; }
+
+  // Does this device need non-text content?
+  virtual GBool needNonText() { return gFalse; }
+
+  //----- initialization and control
+
+  // Start a page.
+  virtual void startPage(int pageNum, GfxState *state);
+
+  // End a page.
+  virtual void endPage();
+
+  //----- update text state
+  virtual void updateFont(GfxState *state);
+
+  //----- text drawing
+  //new feature    
+  virtual int DevType() {return 1234;}
+
+  int getPageWidth() { return maxPageWidth; }
+  int getPageHeight() { return maxPageHeight; }
+  float getBiggestSeperator(xmlNodePtr N_set, unsigned int direction, float * C1, float * C2);
+  void recursiveXYC(xmlNodePtr nodeset);
+  void splitNodes(float splitValue, unsigned int direction, xmlNodePtr N_parent, double extravalue);
+  virtual void beginString(GfxState *state, GooString *s);
+  virtual void endString(GfxState *state);
+  virtual void drawChar(GfxState *state, double x, double y,
+			double dx, double dy,
+			double originX, double originY,
+			CharCode code, int nBytes, Unicode *u, int uLen);
+  void beginWord(GfxState *state, double x, double y);
+  void endWord();
+  void beginTextBlock(GfxState *state, double x, double y);
+  void endTextBlock();
+  void interpretXYTree();
+  void ATP_recursive(xmlNodePtr N_cur);
+  void cleanUpNode(xmlNodePtr N_parent, bool aggregateInfo);
+  void transformPage(xmlNodePtr N_parent);
+  void generateParagraphs();
+  void addAlignment(xmlNodePtr N_parent);
+  void setPDFDoc(PDFDoc *priv_pdfdoc);
+  void createABW();
+
+private:
+  int maxPageWidth;
+  int maxPageHeight;
+  int G_pageNum;
+  int Style, maxStyle;
+  //A lot of values are nice to have around. I think that declaring some 
+  //global variables that contain these values is faster & easier than reading
+  //them from the xml tree every time.
+  double height;
+  double wordSpace, charSpace;
+  double X1,X2,Y1,Y2,horDist, verDist, curDx, curDy;
+  bool mightBreak;
+  xmlDocPtr doc;
+  /* node pointers */
+  xmlNodePtr N_root, N_content, N_page, N_style, N_text, N_styleset, N_Block, N_word, N_column, N_colset;
+  xmlNodePtr outputDoc;
+  xmlXPathContextPtr xpathCtx;
+  static const unsigned int HORIZONTAL = 0;
+  static const unsigned int VERTICAL = 1;
+  UnicodeMap *uMap;
+  PDFDoc *pdfdoc;
+};
+#endif

poppler/Makefile.am

@@ -73,12 +73,27 @@ zlib_libs = 					\
 
 endif
 
+if BUILD_ABIWORD_OUTPUT
+
+abiword_sources =				\
+	ABWOutputDev.h				\
+	ABWOutputDev.cc
+
+abiword_includes =				\
+	$(ABIWORD_CFLAGS)
+
+abiword_libs =					\
+	$(ABIWORD_LIBS)
+
+endif
+
 INCLUDES =					\
 	-I$(top_srcdir)				\
 	-I$(top_srcdir)/goo			\
 	$(splash_includes)			\
 	$(cairo_includes)			\
 	$(arthur_includes)			\
+	$(abiword_includes)			\
 	$(FREETYPE_CFLAGS)			\
 	$(FONTCONFIG_CFLAGS)
 
@@ -94,6 +109,7 @@ libpoppler_la_LIBADD =				\
 	$(cairo_libs)				\
 	$(libjpeg_libs)				\
 	$(zlib_libs)				\
+	$(abiword_libs)				\
 	$(FREETYPE_LIBS)			\
 	$(FONTCONFIG_LIBS)
 
@@ -169,6 +185,7 @@ libpoppler_la_SOURCES =		\
 	$(arthur_sources)	\
 	$(libjpeg_sources)	\
 	$(zlib_sources)		\
+	$(abiword_sources)	\
 	Annot.cc		\
 	Array.cc 		\
 	BuiltinFont.cc		\

poppler/PSOutputDev.cc

@@ -1073,6 +1073,11 @@ void PSOutputDev::init(PSOutputFunc outputFuncA, void *outputStreamA,
     imgURX = paperWidth;
     imgURY = paperHeight;
   }
+  if (imgLLX == 0 && imgURX == 0 && imgLLY == 0 && imgURY == 0) {
+    imgLLX = imgLLY = 0;
+    imgURX = paperWidth;
+    imgURY = paperHeight;
+  }
   manualCtrl = manualCtrlA;
   if (mode == psModeForm) {
     lastPage = firstPage;

qt4/src/poppler-embeddedfile.cc

@@ -64,7 +64,6 @@ EmbeddedFile::EmbeddedFile(const EmbeddedFile &ef)
 
 EmbeddedFile::~EmbeddedFile()
 {
-	m_embeddedFile->m_streamObject.free();
 	delete m_embeddedFile;
 }
 

utils/Makefile.am

@@ -10,12 +10,23 @@ pdftoppm_manpage = pdftoppm.1
 
 endif
 
+if BUILD_ABIWORD_OUTPUT
+
+pdftoabw_SOURCES =				\
+	pdftoabw.cc				\
+	$(common)
+
+pdftoabw_binary = pdftoabw
+
+endif
+
 INCLUDES =					\
 	-I$(top_srcdir)				\
 	-I$(top_srcdir)/utils			\
 	-I$(top_srcdir)/poppler			\
 	$(UTILS_CFLAGS)				\
-	$(FONTCONFIG_CFLAGS)
+	$(FONTCONFIG_CFLAGS)			\
+	$(ABIWORD_CFLAGS)
 
 LDADD =						\
 	$(top_builddir)/poppler/libpoppler.la	\
@@ -29,7 +40,8 @@ bin_PROGRAMS =					\
 	pdftops					\
 	pdftotext				\
 	pdftohtml				\
-	$(pdftoppm_binary)
+	$(pdftoppm_binary)			\
+	$(pdftoabw_binary)
 
 dist_man1_MANS =				\
 	pdffonts.1				\

utils/pdftoabw.cc

+//========================================================================
+//
+// pdftohtml.cc
+//
+//
+// Copyright 1999-2000 G. Ovtcharov
+//========================================================================
+
+#include "config.h"
+#include <poppler-config.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <stddef.h>
+#include <string.h>
+#include <dirent.h>
+#include <time.h>
+#include "parseargs.h"
+#include "goo/GooString.h"
+#include "goo/gmem.h"
+#include "Object.h"
+#include "Stream.h"
+#include "Array.h"
+#include "Dict.h"
+#include "XRef.h"
+#include "Catalog.h"
+#include "Page.h"
+#include "PDFDoc.h"
+#include "ABWOutputDev.h"
+#include "PSOutputDev.h"
+#include "GlobalParams.h"
+#include "Error.h"
+#include "goo/gfile.h"
+#include <libxml/parser.h>
+#include <libxml/tree.h>
+
+static int firstPage = 1;
+static int lastPage = 0;
+static GBool printHelp = gFalse;
+GBool stout = gFalse;
+static char ownerPassword[33] = "";
+static char userPassword[33] = "";
+
+static char textEncName[128] = "";
+
+static ArgDesc argDesc[] = {
+  {"-f",      argInt,      &firstPage,     0,
+   "first page to convert"},
+  {"-l",      argInt,      &lastPage,      0,
+   "last page to convert"},
+  {"-h",      argFlag,     &printHelp,     0,
+   "print usage information"},
+  {"--help",   argFlag,     &printHelp,     0,
+   "print usage information"},
+  {"--stdout"  ,argFlag,    &stout,         0,
+   "use standard output"},
+  {"--opw",    argString,   ownerPassword,  sizeof(ownerPassword),
+   "owner password (for encrypted files)"},
+  {"--upw",    argString,   userPassword,   sizeof(userPassword),
+   "user password (for encrypted files)"},
+  {NULL}
+};
+
+int main(int argc, char *argv[]) {
+  PDFDoc *doc = NULL;
+  GooString *fileName = NULL;
+  GooString *abwFileName = NULL;
+  ABWOutputDev *abwOut = NULL;
+  GBool ok;
+  GooString *ownerPW, *userPW;
+  Object info;
+
+  int result = 1;
+  
+  char * outpName;
+  xmlDocPtr XMLdoc = NULL;
+
+  // parse args
+  parseArgs(argDesc, &argc, argv);
+  globalParams = new GlobalParams();
+
+  fileName = new GooString(argv[1]);
+  if (stout || (argc < 2)){
+    outpName = "-";
+  }
+  else {
+    outpName = argv[2];
+  }
+
+  if (ownerPassword[0]) {
+    ownerPW = new GooString(ownerPassword);
+  } else {
+    ownerPW = NULL;
+  }
+  if (userPassword[0]) {
+    userPW = new GooString(userPassword);
+  } else {
+    userPW = NULL;
+  }
+
+  doc = new PDFDoc(fileName, ownerPW, userPW);
+
+  if (userPW) {
+    delete userPW;
+  }
+  if (ownerPW) {
+    delete ownerPW;
+  }
+
+  if (!doc || !doc->isOk())
+    {
+      fprintf (stderr, "Error opening PDF %s\n", fileName->getCString());
+      goto error;
+    }
+
+  // check for copy permission
+  if (!doc->okToCopy()) {
+    fprintf(stderr, "Copying of text from this document is not allowed.\n");
+    goto error;
+  }
+
+  XMLdoc = xmlNewDoc(BAD_CAST "1.0");
+  abwOut = new ABWOutputDev(XMLdoc);
+  abwOut->setPDFDoc(doc);
+
+  if (lastPage == 0 || lastPage > doc->getNumPages ()) lastPage = doc->getNumPages();
+  if (firstPage < 1) firstPage = 1;
+
+  if (abwOut->isOk())
+  {
+    doc->displayPages(abwOut, firstPage, lastPage, 72, 72, 0, gTrue, gFalse, gFalse);
+    abwOut->createABW();
+  }
+
+  if (xmlSaveFormatFileEnc(outpName, XMLdoc, "UTF-8", 1) == -1)
+    {
+      fprintf (stderr, "Error saving to %s\n", outpName);
+      goto error;
+    }
+
+  result = 0;
+
+ error:
+  // clean up
+  if(globalParams) delete globalParams;
+  if(doc) delete doc;
+  if(XMLdoc) xmlFreeDoc(XMLdoc);
+  if(abwOut) delete abwOut;
+  
+  // check for memory leaks
+  Object::memCheck(stderr);
+  gMemReport(stderr);
+
+  return result;
+}