Move utils/JSInfo.cc utils/JSInfo.h to core poppler

and add new JSInfo::scanJS() variant that returns
immediately after finding JS. This variant is used
by newly added PDFDoc::hasJavascript() method which
is in turn used by newly added poppler-glib function
poppler_document_has_javascript()

4 files changed, 343 insertions, 0 deletions

diff --git a/poppler/JSInfo.cc b/poppler/JSInfo.cc
new file mode 100644
index 00000000..4d3f5919
--- /dev/null
+++ b/poppler/JSInfo.cc
@@ -0,0 +1,268 @@
+//========================================================================
+//
+// JSInfo.cc
+//
+// This file is licensed under the GPLv2 or later
+//
+// Copyright (C) 2013 Adrian Johnson <ajohnson@redneon.com>
+// Copyright (C) 2017, 2020 Albert Astals Cid <aacid@kde.org>
+// Copyright (C) 2018 Klarälvdalens Datakonsult AB, a KDAB Group company, <info@kdab.com>. Work sponsored by the LiMux project of the city of Munich
+// Copyright (C) 2020 Oliver Sander <oliver.sander@tu-dresden.de>
+//
+// 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
+//
+//========================================================================
+
+
+#include "config.h"
+#include <cstdio>
+#include "Object.h"
+#include "Dict.h"
+#include "Annot.h"
+#include "PDFDoc.h"
+#include "JSInfo.h"
+#include "Link.h"
+#include "Form.h"
+#include "UnicodeMap.h"
+#include "UTF.h"
+// #include "Win32Console.h"
+
+JSInfo::JSInfo(PDFDoc *docA, int firstPage) {
+  doc = docA;
+  currentPage = firstPage + 1;
+}
+
+JSInfo::~JSInfo() {
+}
+
+void JSInfo::printJS(const GooString *js) {
+  Unicode *u =  nullptr;
+  char buf[8];
+  int i, n, len;
+
+  if (!js || !js->c_str())
+    return;
+
+  len = TextStringToUCS4(js, &u);
+  for (i = 0; i < len; i++) {
+    n = uniMap->mapUnicode(u[i], buf, sizeof(buf));
+    fwrite(buf, 1, n, file);
+  }
+  gfree(u);
+}
+
+void JSInfo::scanLinkAction(LinkAction *link, const char *action) {
+  if (!link)
+    return;
+
+  if (link->getKind() == actionJavaScript) {
+    hasJS = true;
+    if (print) {
+      LinkJavaScript *linkjs = static_cast<LinkJavaScript *>(link);
+      if (linkjs->isOk()) {
+	const std::string& s = linkjs->getScript();
+	fprintf(file, "%s:\n", action);
+	GooString gooS = GooString(s);
+	printJS(&gooS);
+	fputs("\n\n", file);
+      }
+    }
+  }
+
+  if (link->getKind() == actionRendition) {
+    LinkRendition *linkr = static_cast<LinkRendition *>(link);
+    if (!linkr->getScript().empty()) {
+      hasJS = true;
+      if (print) {
+        fprintf(file, "%s (Rendition):\n", action);
+        const GooString s(linkr->getScript());
+        printJS(&s);
+        fputs("\n\n", file);
+      }
+    }
+  }
+}
+
+void JSInfo::scanJS(int nPages) {
+  print = false;
+  file = nullptr;
+  onlyFirstJS = false;
+  scan(nPages);
+}
+
+void JSInfo::scanJS(int nPages, FILE *fout, const UnicodeMap *uMap) {
+  print = true;
+  file = fout;
+  uniMap = uMap;
+  onlyFirstJS = false;
+  scan(nPages);
+}
+
+void JSInfo::scanJS(int nPages, bool stopOnFirstJS) {
+  print = false;
+  file = nullptr;
+  onlyFirstJS = stopOnFirstJS;
+  scan(nPages);
+}
+
+void JSInfo::scan(int nPages) {
+  Page *page;
+  Annots *annots;
+  int lastPage;
+
+  hasJS = false;
+
+  // Names
+  int numNames = doc->getCatalog()->numJS();
+  if (numNames > 0) {
+    hasJS = true;
+    if (onlyFirstJS) {
+      return;
+    }
+    if (print) {
+      for (int i = 0; i < numNames; i++) {
+	fprintf(file, "Name Dictionary \"%s\":\n", doc->getCatalog()->getJSName(i)->c_str());
+	GooString *js = doc->getCatalog()->getJS(i);
+	printJS(js);
+	delete js;
+	fputs("\n\n", file);
+      }
+    }
+  }
+
+  // document actions
+  scanLinkAction(doc->getCatalog()->getAdditionalAction(Catalog::actionCloseDocument).get(),
+                 "Before Close Document");
+  scanLinkAction(doc->getCatalog()->getAdditionalAction(Catalog::actionSaveDocumentStart).get(),
+                 "Before Save Document");
+  scanLinkAction(doc->getCatalog()->getAdditionalAction(Catalog::actionSaveDocumentFinish).get(),
+                 "After Save Document");
+  scanLinkAction(doc->getCatalog()->getAdditionalAction(Catalog::actionPrintDocumentStart).get(),
+                 "Before Print Document");
+  scanLinkAction(doc->getCatalog()->getAdditionalAction(Catalog::actionPrintDocumentFinish).get(),
+                 "After Print Document");
+
+  if (onlyFirstJS && hasJS) {
+    return;
+  }
+  // form field actions
+  if (doc->getCatalog()->getFormType() == Catalog::AcroForm) {
+    Form *form = doc->getCatalog()->getForm();
+    for (int i = 0; i < form->getNumFields(); i++) {
+      FormField *field = form->getRootField(i);
+      for (int j = 0; j < field->getNumWidgets(); j++) {
+	FormWidget *widget = field->getWidget(j);
+	scanLinkAction(widget->getActivationAction(),
+                       "Field Activated");
+	scanLinkAction(widget->getAdditionalAction(Annot::actionFieldModified).get(),
+                       "Field Modified");
+	scanLinkAction(widget->getAdditionalAction(Annot::actionFormatField).get(),
+                       "Format Field");
+	scanLinkAction(widget->getAdditionalAction(Annot::actionValidateField).get(),
+                       "Validate Field");
+	scanLinkAction(widget->getAdditionalAction(Annot::actionCalculateField).get(),
+                       "Calculate Field");
+        if (onlyFirstJS && hasJS) {
+          return;
+        }
+      }
+    }
+  }
+
+  // scan pages
+
+  if (currentPage > doc->getNumPages()) {
+    return;
+  }
+
+  lastPage = currentPage + nPages;
+  if (lastPage > doc->getNumPages() + 1) {
+    lastPage = doc->getNumPages() + 1;
+  }
+
+  for (int pg = currentPage; pg < lastPage; ++pg) {
+    page = doc->getPage(pg);
+    if (!page) continue;
+
+    // page actions (open, close)
+    scanLinkAction(page->getAdditionalAction(Page::actionOpenPage).get(), "Page Open");
+    scanLinkAction(page->getAdditionalAction(Page::actionClosePage).get(), "Page Close");
+
+    if (onlyFirstJS && hasJS) {
+          return;
+    }
+    // annotation actions (links, screen, widget)
+    annots = page->getAnnots();
+    for (int i = 0; i < annots->getNumAnnots(); ++i) {
+      if (annots->getAnnot(i)->getType() == Annot::typeLink) {
+	AnnotLink *annot = static_cast<AnnotLink *>(annots->getAnnot(i));
+	scanLinkAction(annot->getAction(), "Link Annotation Activated");
+	if (onlyFirstJS && hasJS) {
+	  return;
+	}
+      } else if (annots->getAnnot(i)->getType() == Annot::typeScreen) {
+	AnnotScreen *annot = static_cast<AnnotScreen *>(annots->getAnnot(i));
+	scanLinkAction(annot->getAction(),
+                       "Screen Annotation Activated");
+	scanLinkAction(annot->getAdditionalAction(Annot::actionCursorEntering).get(),
+                       "Screen Annotation Cursor Enter");
+	scanLinkAction(annot->getAdditionalAction(Annot::actionCursorLeaving).get(),
+                       "Screen Annotation Cursor Leave");
+	scanLinkAction(annot->getAdditionalAction(Annot::actionMousePressed).get(),
+                       "Screen Annotation Mouse Pressed");
+	scanLinkAction(annot->getAdditionalAction(Annot::actionMouseReleased).get(),
+                       "Screen Annotation Mouse Released");
+	scanLinkAction(annot->getAdditionalAction(Annot::actionFocusIn).get(),
+                       "Screen Annotation Focus In");
+	scanLinkAction(annot->getAdditionalAction(Annot::actionFocusOut).get(),
+                       "Screen Annotation Focus Out");
+	scanLinkAction(annot->getAdditionalAction(Annot::actionPageOpening).get(),
+                       "Screen Annotation Page Open");
+	scanLinkAction(annot->getAdditionalAction(Annot::actionPageClosing).get(),
+                       "Screen Annotation Page Close");
+	scanLinkAction(annot->getAdditionalAction(Annot::actionPageVisible).get(),
+                       "Screen Annotation Page Visible");
+	scanLinkAction(annot->getAdditionalAction(Annot::actionPageInvisible).get(),
+                       "Screen Annotation Page Invisible");
+
+	if (onlyFirstJS && hasJS) {
+	  return;
+	}
+      } else if (annots->getAnnot(i)->getType() == Annot::typeWidget) {
+	AnnotWidget *annot = static_cast<AnnotWidget *>(annots->getAnnot(i));
+	scanLinkAction(annot->getAction(),
+                       "Widget Annotation Activated");
+	scanLinkAction(annot->getAdditionalAction(Annot::actionCursorEntering).get(),
+                       "Widget Annotation Cursor Enter");
+	scanLinkAction(annot->getAdditionalAction(Annot::actionCursorLeaving).get(),
+                       "Widget Annotation Cursor Leave");
+	scanLinkAction(annot->getAdditionalAction(Annot::actionMousePressed).get(),
+                       "Widget Annotation Mouse Pressed");
+	scanLinkAction(annot->getAdditionalAction(Annot::actionMouseReleased).get(),
+                       "Widget Annotation Mouse Released");
+	scanLinkAction(annot->getAdditionalAction(Annot::actionFocusIn).get(),
+                       "Widget Annotation Focus In");
+	scanLinkAction(annot->getAdditionalAction(Annot::actionFocusOut).get(),
+                       "Widget Annotation Focus Out");
+	scanLinkAction(annot->getAdditionalAction(Annot::actionPageOpening).get(),
+                       "Widget Annotation Page Open");
+	scanLinkAction(annot->getAdditionalAction(Annot::actionPageClosing).get(),
+                       "Widget Annotation Page Close");
+	scanLinkAction(annot->getAdditionalAction(Annot::actionPageVisible).get(),
+                       "Widget Annotation Page Visible");
+	scanLinkAction(annot->getAdditionalAction(Annot::actionPageInvisible).get(),
+                       "Widget Annotation Page Invisible");
+	if (onlyFirstJS && hasJS) {
+	  return;
+	}
+      }
+    }
+  }
+
+  currentPage = lastPage;
+}
+
+bool JSInfo::containsJS() {
+  return hasJS;
+}
diff --git a/poppler/JSInfo.h b/poppler/JSInfo.h
new file mode 100644
index 00000000..89e0a3d0
--- /dev/null
+++ b/poppler/JSInfo.h
@@ -0,0 +1,66 @@
+//========================================================================
+//
+// JSInfo.h
+//
+// This file is licensed under the GPLv2 or later
+//
+// Copyright (C) 2013 Adrian Johnson <ajohnson@redneon.com>
+// Copyright (C) 2020 Albert Astals Cid <aacid@kde.org>
+// Copyright (C) 2018 Klarälvdalens Datakonsult AB, a KDAB Group company, <info@kdab.com>. Work sponsored by the LiMux project of the city of Munich
+// Copyright (C) 2020 Oliver Sander <oliver.sander@tu-dresden.de>
+//
+// 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
+//
+//========================================================================
+
+#ifndef JS_INFO_H
+#define JS_INFO_H
+
+#include <cstdio>
+#include "Object.h"
+#include "PDFDoc.h"
+
+#include "Link.h"
+#include "UnicodeMap.h"
+
+class PDFDoc;
+
+class JSInfo {
+public:
+
+  // Constructor.
+  JSInfo(PDFDoc *doc, int firstPage = 0);
+
+  // Destructor.
+  ~JSInfo();
+
+  // scan for JS in the PDF
+  void scanJS(int nPages);
+
+  // scan and print JS in the PDF
+  void scanJS(int nPages, FILE *fout, const UnicodeMap *uMap);
+
+  // scan but exit after finding first JS in the PDF
+  void scanJS(int nPages, bool stopOnFirstJS);
+
+  // return true if PDF contains JavaScript
+  bool containsJS();
+
+private:
+
+  PDFDoc *doc;
+  int currentPage;
+  bool hasJS;
+  bool print;
+  FILE *file;
+  const UnicodeMap *uniMap;
+  bool onlyFirstJS; /* stop scanning after finding first JS */
+
+  void scan(int nPages);
+  void scanLinkAction(LinkAction *link, const char *action);
+  void printJS(const GooString *js);
+
+};
+
+#endif
diff --git a/poppler/PDFDoc.cc b/poppler/PDFDoc.cc
index 4b5621fc..6ad884cf 100644
--- a/poppler/PDFDoc.cc
+++ b/poppler/PDFDoc.cc
@@ -86,6 +86,7 @@
 #include "PDFDoc.h"
 #include "Hints.h"
 #include "UTF.h"
+#include "JSInfo.h"
 
 //------------------------------------------------------------------------
 
@@ -2152,3 +2153,9 @@ Page *PDFDoc::getPage(int page)
 
   return catalog->getPage(page);
 }
+
+bool PDFDoc::hasJavascript() {
+  JSInfo jsInfo(this);
+  jsInfo.scanJS (getNumPages(), true);
+  return jsInfo.containsJS();
+}
diff --git a/poppler/PDFDoc.h b/poppler/PDFDoc.h
index 7f4fb167..ca8bf6b6 100644
--- a/poppler/PDFDoc.h
+++ b/poppler/PDFDoc.h
@@ -349,6 +349,8 @@ public:
                                      Goffset uxrefOffset, OutStream* outStr, XRef *xRef);
   static void writeXRefStreamTrailer (Object &&trailerDict, XRef *uxref, Ref *uxrefStreamRef,
                                       Goffset uxrefOffset, OutStream* outStr, XRef *xRef);
+  // scans the PDF and returns whether it contains any javascript
+  bool hasJavascript();
 
 private:
   // insert referenced objects in XRef