From a38c4bc72903073727119c29331bb3ed7abf6ce5 Mon Sep 17 00:00:00 2001
From: Julian Viereck <julian.viereck@gmail.com>
Date: Tue, 11 Sep 2012 15:10:34 -0700
Subject: [PATCH] Make getTextContent return offset array and improve the
 algorithm. Make parts in viewer.js work again.

---
 src/evaluator.js | 67 ++++++++++++++++++++++++++++++++++++++++++++----
 web/viewer.html  |  2 +-
 web/viewer.js    |  2 +-
 3 files changed, 64 insertions(+), 7 deletions(-)

diff --git a/src/evaluator.js b/src/evaluator.js
index f0e775cdb..1ac32f781 100644
--- a/src/evaluator.js
+++ b/src/evaluator.js
@@ -505,7 +505,13 @@ var PartialEvaluator = (function PartialEvaluatorClosure() {
       return queue;
     },
 
-    getTextContent: function partialEvaluatorGetIRQueue(stream, resources) {
+    getTextContent: function partialEvaluatorGetIRQueue(stream, resources, state) {
+      if (!state) {
+        state = {
+          text: '',
+          mapping: []
+        };
+      }
 
       var self = this;
       var xref = this.xref;
@@ -515,18 +521,22 @@ var PartialEvaluator = (function PartialEvaluatorClosure() {
       }
 
       resources = xref.fetchIfRef(resources) || new Dict();
+      // The xobj is parsed iff it's needed, e.g. if there is a `DO` cmd.
+      var xobjs = null;
 
       var parser = new Parser(new Lexer(stream), false);
       var res = resources;
       var args = [], obj;
 
-      var text = '';
+      var text = state.text;
       var chunk = '';
+      var commandOffset = state.mapping;
       var font = null;
       while (!isEOF(obj = parser.getObj())) {
         if (isCmd(obj)) {
           var cmd = obj.cmd;
           switch (cmd) {
+            // TODO: Add support for SAVE/RESTORE and XFORM here.
             case 'Tf':
               font = handleSetFont(args[0].name).translated;
               break;
@@ -536,9 +546,12 @@ var PartialEvaluator = (function PartialEvaluatorClosure() {
                 if (typeof items[j] === 'string') {
                   chunk += fontCharsToUnicode(items[j], font);
                 } else if (items[j] < 0) {
-                  // making all negative offsets a space - better to have
-                  // a space in incorrect place than not have them at all
                   chunk += ' ';
+                } else if (items[j] < 0 && font.spacedWidth > 0) {
+                  var numFakeSpaces = Math.round(-e / font.spacedWidth);
+                  if (numFakeSpaces > 0) {
+                    chunk += ' ';
+                  }
                 }
               }
               break;
@@ -551,8 +564,49 @@ var PartialEvaluator = (function PartialEvaluatorClosure() {
             case '"':
               chunk += fontCharsToUnicode(args[2], font) + ' ';
               break;
+            case 'Do':
+              // Set the chunk such that the following if won't add something
+              // to the state.
+              chunk = '';
+
+              if (args[0].code) {
+                break;
+              }
+
+              if (!xobjs) {
+                xobjs = resources.get('XObject') || new Dict();
+              }
+
+              var name = args[0].name;
+              var xobj = xobjs.get(name);
+              if (!xobj)
+                break;
+              assertWellFormed(isStream(xobj), 'XObject should be a stream');
+
+              var type = xobj.dict.get('Subtype');
+              assertWellFormed(
+                isName(type),
+                'XObject should have a Name subtype'
+              );
+
+              if ('Form' !== type.name)
+                break;
+
+              // Add some spacing between the text here and the text of the
+              // xForm.
+              text = text + ' ';
+
+              state.text = text;
+              state = this.getTextContent(
+                xobj,
+                xobj.dict.get('Resources') || resources,
+                state
+              );
+              text = state.text;
+              break;
           } // switch
           if (chunk !== '') {
+            commandOffset.push(text.length);
             text += chunk;
             chunk = '';
           }
@@ -564,7 +618,10 @@ var PartialEvaluator = (function PartialEvaluatorClosure() {
         }
       }
 
-      return text;
+      return {
+        text: text,
+        mapping: commandOffset
+      };
     },
 
     extractDataStructures: function
diff --git a/web/viewer.html b/web/viewer.html
index 5a2f4f28c..813484c6c 100644
--- a/web/viewer.html
+++ b/web/viewer.html
@@ -88,7 +88,7 @@ limitations under the License.
           <button id="viewOutline" class="toolbarButton group" title="Show Document Outline" tabindex="2" data-l10n-id="outline">
              <span data-l10n-id="outline_label">Document Outline</span>
           </button>
-          <button id="viewSearch" class="toolbarButton group hidden" title="Search Document" tabindex="3" data-l10n-id="search_panel">
+          <button id="viewSearch" class="toolbarButton group" title="Search Document" tabindex="3" data-l10n-id="search_panel">
              <span data-l10n-id="search_panel_label">Search Document</span>
           </button>
         </div>
diff --git a/web/viewer.js b/web/viewer.js
index 5f8ded0ec..be512736f 100644
--- a/web/viewer.js
+++ b/web/viewer.js
@@ -1043,7 +1043,7 @@ var PDFView = {
     function extractPageText(pageIndex) {
       self.pages[pageIndex].pdfPage.getTextContent().then(
         function textContentResolved(textContent) {
-          self.pageText[pageIndex] = textContent;
+          self.pageText[pageIndex] = textContent.text;
           self.search();
           if ((pageIndex + 1) < self.pages.length)
             extractPageText(pageIndex + 1);