[poppler] 2 commits - poppler/PDFDoc.cc poppler/PDFDoc.h poppler/XRef.cc poppler/XRef.h utils/pdfunite.cc

Albert Astals Cid aacid at kemper.freedesktop.org
Thu Apr 26 16:03:56 PDT 2012


 poppler/PDFDoc.cc |   96 ++++++++++++++++++++++++++++++++++++++----------------
 poppler/PDFDoc.h  |   14 ++++++-
 poppler/XRef.cc   |   76 ++++++++++++++++++++++++++++++++++++------
 poppler/XRef.h    |   36 +++++++++++++++++++-
 utils/pdfunite.cc |   10 +++--
 5 files changed, 185 insertions(+), 47 deletions(-)

New commits:
commit b1d5c6c0a20a4a24b42de66db23e0f63d10ca52d
Author: Fabio D'Urso <fabiodurso at hotmail.it>
Date:   Tue Apr 24 21:00:11 2012 +0200

    Output XRef stream when incrementally updating if there's already a XRef stream

diff --git a/poppler/PDFDoc.cc b/poppler/PDFDoc.cc
index 9c977a6..2dc1ac8 100644
--- a/poppler/PDFDoc.cc
+++ b/poppler/PDFDoc.cc
@@ -864,8 +864,30 @@ void PDFDoc::saveIncrementalUpdate (OutStream* outStr)
   }
 
   Guint uxrefOffset = outStr->getPos();
-  writeXRefTableTrailer(uxrefOffset, uxref, gFalse /* do not write unnecessary entries */,
-                        xref->getNumObjects(), outStr, gTrue /* incremental update */);
+  int numobjects = xref->getNumObjects();
+  const char *fileNameA = fileName ? fileName->getCString() : NULL;
+  Ref rootRef, uxrefStreamRef;
+  rootRef.num = getXRef()->getRootNum();
+  rootRef.gen = getXRef()->getRootGen();
+
+  // Output a xref stream if there is a xref stream already
+  GBool xRefStream = xref->isXRefStream();
+
+  if (xRefStream) {
+    // Append an entry for the xref stream itself
+    uxrefStreamRef.num = numobjects++;
+    uxrefStreamRef.gen = 0;
+    uxref->add(uxrefStreamRef.num, uxrefStreamRef.gen, uxrefOffset, gTrue);
+  }
+
+  Dict *trailerDict = createTrailerDict(numobjects, gTrue, getStartXRef(), &rootRef, getXRef(), fileNameA, uxrefOffset);
+  if (xRefStream) {
+    writeXRefStreamTrailer(trailerDict, uxref, &uxrefStreamRef, uxrefOffset, outStr, getXRef());
+  } else {
+    writeXRefTableTrailer(trailerDict, uxref, gFalse, uxrefOffset, outStr, getXRef());
+  }
+
+  delete trailerDict;
   delete uxref;
 }
 
@@ -1203,6 +1225,25 @@ void PDFDoc::writeXRefTableTrailer(Dict *trailerDict, XRef *uxref, GBool writeAl
   outStr->printf( "%%%%EOF\r\n");
 }
 
+void PDFDoc::writeXRefStreamTrailer (Dict *trailerDict, XRef *uxref, Ref *uxrefStreamRef, Guint uxrefOffset, OutStream* outStr, XRef *xRef)
+{
+  GooString stmData;
+
+  // Fill stmData and some trailerDict fields
+  uxref->writeStreamToBuffer(&stmData, trailerDict, xRef);
+
+  // Create XRef stream object and write it
+  Object obj1;
+  MemStream *mStream = new MemStream( stmData.getCString(), 0,
+                                      stmData.getLength(), obj1.initDict(trailerDict) );
+  writeObject(obj1.initStream(mStream), uxrefStreamRef, outStr, xRef, 0);
+  obj1.free();
+
+  outStr->printf( "startxref\r\n");
+  outStr->printf( "%i\r\n", uxrefOffset);
+  outStr->printf( "%%%%EOF\r\n");
+}
+
 void PDFDoc::writeXRefTableTrailer(Guint uxrefOffset, XRef *uxref, GBool writeAllEntries,
                                    int uxrefSize, OutStream* outStr, GBool incrUpdate)
 {
diff --git a/poppler/XRef.cc b/poppler/XRef.cc
index 3b0d7f8..1574371 100644
--- a/poppler/XRef.cc
+++ b/poppler/XRef.cc
@@ -1240,10 +1240,10 @@ void XRef::removeIndirectObject(Ref r) {
   e->updated = true;
 }
 
-void XRef::writeTableToFile(OutStream* outStr, GBool writeAllEntries) {
+void XRef::writeXRef(XRef::XRefWriter *writer, GBool writeAllEntries) {
   //create free entries linked-list
   if (getEntry(0)->gen != 65535) {
-    error(errInternal, -1, "XRef::writeToFile, entry 0 of the XRef is invalid (gen != 65535)\n");
+    error(errInternal, -1, "XRef::writeXRef, entry 0 of the XRef is invalid (gen != 65535)\n");
   }
   int lastFreeEntry = 0;
   for (int i=0; i<size; i++) {
@@ -1254,18 +1254,13 @@ void XRef::writeTableToFile(OutStream* outStr, GBool writeAllEntries) {
   }
 
   if (writeAllEntries) {
-    //write the new xref
-    outStr->printf("xref\r\n");
-    outStr->printf("%i %i\r\n", 0, size);
+    writer->startSection(0, size);
     for (int i=0; i<size; i++) {
       XRefEntry *e = getEntry(i);
-
       if(e->gen > 65535) e->gen = 65535; //cap generation number to 65535 (required by PDFReference)
-      outStr->printf("%010i %05i %c\r\n", e->offset, e->gen, (e->type==xrefEntryFree)?'f':'n');
+      writer->writeEntry(e->offset, e->gen, e->type);
     }
   } else {
-    //write the new xref
-    outStr->printf("xref\r\n");
     int i = 0;
     while (i < size) {
       int j;
@@ -1275,11 +1270,11 @@ void XRef::writeTableToFile(OutStream* outStr, GBool writeAllEntries) {
       }
       if (j-i != 0)
       {
-        outStr->printf("%i %i\r\n", i, j-i);
+        writer->startSection(i, j-i);
         for (int k=i; k<j; k++) {
           XRefEntry *e = getEntry(k);
           if(e->gen > 65535) e->gen = 65535; //cap generation number to 65535 (required by PDFReference)
-          outStr->printf("%010i %05i %c\r\n", e->offset, e->gen, (e->type==xrefEntryFree)?'f':'n');
+          writer->writeEntry(e->offset, e->gen, e->type);
         }
         i = j;
       }
@@ -1288,6 +1283,65 @@ void XRef::writeTableToFile(OutStream* outStr, GBool writeAllEntries) {
   }
 }
 
+XRef::XRefTableWriter::XRefTableWriter(OutStream* outStrA) {
+  outStr = outStrA;
+}
+
+void XRef::XRefTableWriter::startSection(int first, int count) {
+  outStr->printf("%i %i\r\n", first, count);
+}
+
+void XRef::XRefTableWriter::writeEntry(Guint offset, int gen, XRefEntryType type) {
+  outStr->printf("%010i %05i %c\r\n", offset, gen, (type==xrefEntryFree)?'f':'n');
+}
+
+void XRef::writeTableToFile(OutStream* outStr, GBool writeAllEntries) {
+  XRefTableWriter writer(outStr);
+  outStr->printf("xref\r\n");
+  writeXRef(&writer, writeAllEntries);
+}
+
+XRef::XRefStreamWriter::XRefStreamWriter(Object *indexA, GooString *stmBufA) {
+  index = indexA;
+  stmBuf = stmBufA;
+}
+
+void XRef::XRefStreamWriter::startSection(int first, int count) {
+  Object obj;
+  index->arrayAdd( obj.initInt(first) );
+  index->arrayAdd( obj.initInt(count) );
+}
+
+void XRef::XRefStreamWriter::writeEntry(Guint offset, int gen, XRefEntryType type) {
+  char data[7];
+  data[0] = (type==xrefEntryFree) ? 0 : 1;
+  data[1] = (offset >> 24) & 0xff;
+  data[2] = (offset >> 16) & 0xff;
+  data[3] = (offset >> 8) & 0xff;
+  data[4] = offset & 0xff;
+  data[5] = (gen >> 8) & 0xff;
+  data[6] = gen & 0xff;
+  stmBuf->append(data, 7);
+}
+
+void XRef::writeStreamToBuffer(GooString *stmBuf, Dict *xrefDict, XRef *xref) {
+  Object index;
+  index.initArray(xref);
+  stmBuf->clear();
+
+  XRefStreamWriter writer(&index, stmBuf);
+  writeXRef(&writer, gFalse);
+
+  Object obj1, obj2;
+  xrefDict->set("Type", obj1.initName("XRef"));
+  xrefDict->set("Index", &index);
+  obj2.initArray(xref);
+  obj2.arrayAdd( obj1.initInt(1) );
+  obj2.arrayAdd( obj1.initInt(4) );
+  obj2.arrayAdd( obj1.initInt(2) );
+  xrefDict->set("W", &obj2);
+}
+
 GBool XRef::parseEntry(Guint offset, XRefEntry *entry)
 {
   GBool r;
diff --git a/poppler/XRef.h b/poppler/XRef.h
index ab8047c..148a5ce 100644
--- a/poppler/XRef.h
+++ b/poppler/XRef.h
@@ -79,6 +79,9 @@ public:
   // Is xref table valid?
   GBool isOk() { return ok; }
 
+  // Is the last XRef section a stream or a table?
+  GBool isXRefStream() { return xRefStream; }
+
   // Get the error code (if isOk() returns false).
   int getErrorCode() { return errCode; }
 
@@ -135,7 +138,11 @@ public:
   Ref addIndirectObject (Object* o);
   void removeIndirectObject(Ref r);
   void add(int num, int gen,  Guint offs, GBool used);
+
+  // Output XRef table to stream
   void writeTableToFile(OutStream* outStr, GBool writeAllEntries);
+  // Output XRef stream contents to GooString and fill trailerDict fields accordingly
+  void writeStreamToBuffer(GooString *stmBuf, Dict *xrefDict, XRef *xref);
 
 private:
 
@@ -176,6 +183,33 @@ private:
   GBool constructXRef(GBool *wasReconstructed);
   GBool parseEntry(Guint offset, XRefEntry *entry);
 
+  class XRefWriter {
+  public:
+    virtual void startSection(int first, int count) = 0;
+    virtual void writeEntry(Guint offset, int gen, XRefEntryType type) = 0;
+    virtual ~XRefWriter() {};
+  };
+
+  class XRefTableWriter: public XRefWriter {
+  public:
+    XRefTableWriter(OutStream* outStrA);
+    void startSection(int first, int count);
+    void writeEntry(Guint offset, int gen, XRefEntryType type);
+  private:
+    OutStream* outStr;
+  };
+
+  class XRefStreamWriter: public XRefWriter {
+  public:
+    XRefStreamWriter(Object *index, GooString *stmBuf);
+    void startSection(int first, int count);
+    void writeEntry(Guint offset, int gen, XRefEntryType type);
+  private:
+    Object *index;
+    GooString *stmBuf;
+  };
+
+  void writeXRef(XRefWriter *writer, GBool writeAllEntries);
 };
 
 #endif
commit 2ecf3b2e49a4c35e995d25016b810592260edfeb
Author: Fabio D'Urso <fabiodurso at hotmail.it>
Date:   Tue Apr 24 18:10:15 2012 +0200

    Refactoring of XRef table write support (in preparation for XRef stream write support)
    
     - Trailer dictionary creation now lives in its own function "createTrailerDict"
       (that will be used by XRef stream creation too)
     - writeXRefTableTrailer (WAS writeTrailer) now takes care of writing the XRef
       table too (previously it was demanded to the caller)

diff --git a/poppler/PDFDoc.cc b/poppler/PDFDoc.cc
index 898bcbb..9c977a6 100644
--- a/poppler/PDFDoc.cc
+++ b/poppler/PDFDoc.cc
@@ -716,12 +716,14 @@ int PDFDoc::savePageAs(GooString *name, int pageNo)
   page.free();
 
   Guint uxrefOffset = outStr->getPos();
-  yRef->writeToFile(outStr, gFalse /* do not write unnecessary entries */);
-
   Ref ref;
   ref.num = rootNum;
   ref.gen = 0;
-  writeTrailer(uxrefOffset, objectsCount, outStr, gFalse, 0, &ref, getXRef(), name->getCString(), outStr->getPos());
+  Dict *trailerDict = createTrailerDict(objectsCount, gFalse, 0, &ref, getXRef(),
+                                        name->getCString(), uxrefOffset);
+  writeXRefTableTrailer(trailerDict, yRef, gFalse /* do not write unnecessary entries */,
+                        uxrefOffset, outStr, getXRef());
+  delete trailerDict;
 
   outStr->close();
   fclose(f);
@@ -862,10 +864,8 @@ void PDFDoc::saveIncrementalUpdate (OutStream* outStr)
   }
 
   Guint uxrefOffset = outStr->getPos();
-  uxref->writeToFile(outStr, gFalse /* do not write unnecessary entries */);
-
-  writeTrailer(uxrefOffset, xref->getNumObjects(), outStr, gTrue);
-
+  writeXRefTableTrailer(uxrefOffset, uxref, gFalse /* do not write unnecessary entries */,
+                        xref->getNumObjects(), outStr, gTrue /* incremental update */);
   delete uxref;
 }
 
@@ -902,13 +902,9 @@ void PDFDoc::saveCompleteRewrite (OutStream* outStr)
     }
   }
   Guint uxrefOffset = outStr->getPos();
-  uxref->writeToFile(outStr, gTrue /* write all entries */);
-
-  writeTrailer(uxrefOffset, uxref->getNumObjects(), outStr, gFalse);
-
-
+  writeXRefTableTrailer(uxrefOffset, uxref, gTrue /* write all entries */,
+                        uxref->getNumObjects(), outStr, gFalse /* complete rewrite */);
   delete uxref;
-
 }
 
 void PDFDoc::writeDictionnary (Dict* dict, OutStream* outStr, XRef *xRef, Guint numOffset)
@@ -1109,10 +1105,8 @@ Guint PDFDoc::writeObject (Object* obj, Ref* ref, OutStream* outStr, XRef *xRef,
   return offset;
 }
 
-void PDFDoc::writeTrailer(Guint uxrefOffset, int uxrefSize, 
-			  OutStream* outStr, GBool incrUpdate,
-			  Guint startxRef, Ref *root, XRef *xRef, const char *fileName,
-			  Guint fileSize)
+Dict *PDFDoc::createTrailerDict(int uxrefSize, GBool incrUpdate, Guint startxRef,
+                                Ref *root, XRef *xRef, const char *fileName, Guint fileSize)
 {
   Dict *trailerDict = new Dict(xRef);
   Object obj1;
@@ -1120,7 +1114,6 @@ void PDFDoc::writeTrailer(Guint uxrefOffset, int uxrefSize,
   trailerDict->set("Size", &obj1);
   obj1.free();
 
-
   //build a new ID, as recommended in the reference, uses:
   // - current time
   // - file name
@@ -1130,7 +1123,9 @@ void PDFDoc::writeTrailer(Guint uxrefOffset, int uxrefSize,
   char buffer[256];
   sprintf(buffer, "%i", (int)time(NULL));
   message.append(buffer);
-  message.append(fileName);
+
+  if (fileName)
+    message.append(fileName);
 
   sprintf(buffer, "%i", fileSize);
   message.append(buffer);
@@ -1162,7 +1157,7 @@ void PDFDoc::writeTrailer(Guint uxrefOffset, int uxrefSize,
     //only update the second part of the array
     xRef->getTrailerDict()->getDict()->lookup("ID", &obj4);
     if (!obj4.isArray()) {
-      error(errSyntaxWarning, -1, "PDFDoc::writeTrailer original file's ID entry isn't an array. Trying to continue");
+      error(errSyntaxWarning, -1, "PDFDoc::createTrailerDict original file's ID entry isn't an array. Trying to continue");
     } else {
       //Get the first part of the ID
       obj4.arrayGet(0,&obj3); 
@@ -1194,24 +1189,25 @@ void PDFDoc::writeTrailer(Guint uxrefOffset, int uxrefSize,
       trailerDict->set("Info", &obj5);
     }
   }
-  
+
+  return trailerDict;
+}
+
+void PDFDoc::writeXRefTableTrailer(Dict *trailerDict, XRef *uxref, GBool writeAllEntries, Guint uxrefOffset, OutStream* outStr, XRef *xRef)
+{
+  uxref->writeTableToFile( outStr, writeAllEntries );
   outStr->printf( "trailer\r\n");
   writeDictionnary(trailerDict, outStr, xRef, 0);
   outStr->printf( "\r\nstartxref\r\n");
   outStr->printf( "%i\r\n", uxrefOffset);
   outStr->printf( "%%%%EOF\r\n");
-
-  delete trailerDict;
 }
 
-void PDFDoc::writeTrailer(Guint uxrefOffset, int uxrefSize, OutStream* outStr, GBool incrUpdate)
+void PDFDoc::writeXRefTableTrailer(Guint uxrefOffset, XRef *uxref, GBool writeAllEntries,
+                                   int uxrefSize, OutStream* outStr, GBool incrUpdate)
 {
-  const char *fileNameA;
-  if (fileName)
-    fileNameA = fileName->getCString();
-  else
-    fileNameA = "streamwithoutfilename.pdf";
-  // file size
+  const char *fileNameA = fileName ? fileName->getCString() : NULL;
+  // file size (doesn't include the trailer)
   unsigned int fileSize = 0;
   int c;
   str->reset();
@@ -1222,7 +1218,10 @@ void PDFDoc::writeTrailer(Guint uxrefOffset, int uxrefSize, OutStream* outStr, G
   Ref ref;
   ref.num = getXRef()->getRootNum();
   ref.gen = getXRef()->getRootGen();
-  writeTrailer(uxrefOffset, uxrefSize, outStr, incrUpdate, getStartXRef(), &ref, getXRef(), fileNameA, fileSize);
+  Dict * trailerDict = createTrailerDict(uxrefSize, incrUpdate, getStartXRef(), &ref,
+                                         getXRef(), fileNameA, fileSize);
+  writeXRefTableTrailer(trailerDict, uxref, writeAllEntries, uxrefOffset, outStr, getXRef());
+  delete trailerDict;
 }
 
 void PDFDoc::writeHeader(OutStream *outStr, int major, int minor)
diff --git a/poppler/PDFDoc.h b/poppler/PDFDoc.h
index ccb1b22..468f698 100644
--- a/poppler/PDFDoc.h
+++ b/poppler/PDFDoc.h
@@ -23,6 +23,7 @@
 // Copyright (C) 2010 Hib Eris <hib at hiberis.nl>
 // Copyright (C) 2010 Srinivas Adicherla <srinivas.adicherla at geodesic.com>
 // Copyright (C) 2011 Thomas Freitag <Thomas.Freitag at alfa.de>
+// Copyright (C) 2012 Fabio D'Urso <fabiodurso at hotmail.it>
 //
 // To see a description of the changes please see the Changelog file that
 // came with your tarball or type make ChangeLog if you are building from git
@@ -244,8 +245,14 @@ public:
   Guint writePageObjects(OutStream *outStr, XRef *xRef, Guint numOffset);
   static Guint writeObject (Object *obj, Ref *ref, OutStream* outStr, XRef *xref, Guint numOffset);
   static void writeHeader(OutStream *outStr, int major, int minor);
-  static void writeTrailer (Guint uxrefOffset, int uxrefSize, OutStream* outStr, GBool incrUpdate,
-                            Guint startxRef, Ref *root, XRef *xRef, const char *fileName, Guint fileSize);
+
+  // Ownership goes to the caller
+  static Dict *createTrailerDict (int uxrefSize, GBool incrUpdate, Guint startxRef,
+                                  Ref *root, XRef *xRef, const char *fileName, Guint fileSize);
+  static void writeXRefTableTrailer (Dict *trailerDict, XRef *uxref, GBool writeAllEntries,
+                                     Guint uxrefOffset, OutStream* outStr, XRef *xRef);
+  static void writeXRefStreamTrailer (Dict *trailerDict, XRef *uxref, Ref *uxrefStreamRef,
+                                      Guint uxrefOffset, OutStream* outStr, XRef *xRef);
 
 private:
   // insert referenced objects in XRef
@@ -260,7 +267,8 @@ private:
   { writeDictionnary(dict, outStr, getXRef(), 0); }
   static void writeStream (Stream* str, OutStream* outStr);
   static void writeRawStream (Stream* str, OutStream* outStr);
-  void writeTrailer (Guint uxrefOffset, int uxrefSize, OutStream* outStr, GBool incrUpdate);
+  void writeXRefTableTrailer (Guint uxrefOffset, XRef *uxref, GBool writeAllEntries,
+                              int uxrefSize, OutStream* outStr, GBool incrUpdate);
   static void writeString (GooString* s, OutStream* outStr);
   void saveIncrementalUpdate (OutStream* outStr);
   void saveCompleteRewrite (OutStream* outStr);
diff --git a/poppler/XRef.cc b/poppler/XRef.cc
index dfc6d73..3b0d7f8 100644
--- a/poppler/XRef.cc
+++ b/poppler/XRef.cc
@@ -1240,7 +1240,7 @@ void XRef::removeIndirectObject(Ref r) {
   e->updated = true;
 }
 
-void XRef::writeToFile(OutStream* outStr, GBool writeAllEntries) {
+void XRef::writeTableToFile(OutStream* outStr, GBool writeAllEntries) {
   //create free entries linked-list
   if (getEntry(0)->gen != 65535) {
     error(errInternal, -1, "XRef::writeToFile, entry 0 of the XRef is invalid (gen != 65535)\n");
diff --git a/poppler/XRef.h b/poppler/XRef.h
index 4cba474..ab8047c 100644
--- a/poppler/XRef.h
+++ b/poppler/XRef.h
@@ -135,7 +135,7 @@ public:
   Ref addIndirectObject (Object* o);
   void removeIndirectObject(Ref r);
   void add(int num, int gen,  Guint offs, GBool used);
-  void writeToFile(OutStream* outStr, GBool writeAllEntries);
+  void writeTableToFile(OutStream* outStr, GBool writeAllEntries);
 
 private:
 
diff --git a/utils/pdfunite.cc b/utils/pdfunite.cc
index 93850a4..212f89b 100644
--- a/utils/pdfunite.cc
+++ b/utils/pdfunite.cc
@@ -6,6 +6,7 @@
 //
 // Copyright (C) 2011 Thomas Freitag <Thomas.Freitag at alfa.de>
 // Copyright (C) 2012 Arseny Solokha <asolokha at gmx.com>
+// Copyright (C) 2012 Fabio D'Urso <fabiodurso at hotmail.it>
 //
 //========================================================================
 #include <PDFDoc.h>
@@ -161,13 +162,14 @@ int main (int argc, char *argv[])
     objectsCount++;
   }
   Guint uxrefOffset = outStr->getPos();
-  yRef->writeToFile(outStr, gFalse /* do not write unnecessary entries */ );
-
   Ref ref;
   ref.num = rootNum;
   ref.gen = 0;
-  PDFDoc::writeTrailer(uxrefOffset, objectsCount, outStr, (GBool) gFalse, 0,
-	&ref, yRef, fileName, outStr->getPos());
+  Dict *trailerDict = PDFDoc::createTrailerDict(objectsCount, gFalse, 0, &ref, yRef,
+                                                fileName, outStr->getPos());
+  PDFDoc::writeXRefTableTrailer(trailerDict, yRef, gFalse /* do not write unnecessary entries */,
+                                uxrefOffset, outStr, yRef);
+  delete trailerDict;
 
   outStr->close();
   fclose(f);


More information about the poppler mailing list