Add text dump to pdfium_test.

samples/pdfium_test.cc

 // Copyright (c) 2010 The Chromium Authors. All rights reserved.
 // Use of this source code is governed by a BSD-style license that can be
 // found in the LICENSE file.
 
 #include <limits.h>
 #include <stdio.h>
 #include <stdlib.h>
 #include <string.h>
 
 #include <list>
@@ skipping 24 matching lines @@
 #define snprintf _snprintf
 #endif
 
 #ifdef PDF_ENABLE_SKIA
 #include "third_party/skia/include/core/SkPictureRecorder.h"
 #include "third_party/skia/include/core/SkStream.h"
 #endif
 
 enum OutputFormat {
   OUTPUT_NONE,
+  OUTPUT_TEXT,
   OUTPUT_PPM,
   OUTPUT_PNG,
 #ifdef _WIN32
   OUTPUT_BMP,
   OUTPUT_EMF,
 #endif
 #ifdef PDF_ENABLE_SKIA
   OUTPUT_SKP,
 #endif
 };
@@ skipping 50 matching lines @@
       dest_line[(w * 3) + 1] = src_line[(w * 4) + 1];
       // B
       dest_line[(w * 3) + 2] = src_line[w * 4];
     }
   }
   fwrite(result, out_len, 1, fp);
   delete[] result;
   fclose(fp);
 }
 
+void WriteText(FPDF_PAGE page, const char* pdf_name, int num) {
+  char filename[256];
+  int chars_formatted =
+      snprintf(filename, sizeof(filename), "%s.%d.txt", pdf_name, num);
+  if (chars_formatted < 0 ||
+      static_cast<size_t>(chars_formatted) >= sizeof(filename)) {
+    fprintf(stderr, "Filename %s is too long\n", filename);
+    return;
+  }
+
+  FILE* fp = fopen(filename, "w");
+  if (!fp) {
+    fprintf(stderr, "Failed to open %s for output\n", filename);
+    return;
+  }
+
+  // Output in UTF32-LE.
+  uint32_t bom = 0x0000FEFF;
+  fwrite(&bom, sizeof(bom), 1, fp);
+
+  FPDF_TEXTPAGE textpage = FPDFText_LoadPage(page);
+  for (int i = 0; i < FPDFText_CountChars(textpage); i++) {
+    uint32_t c = FPDFText_GetUnicode(textpage, i);
+    fwrite(&c, sizeof(c), 1, fp);
+  }
+
+  FPDFText_ClosePage(textpage);
+
+  (void)fclose(fp);
+}
+
 static void WritePng(const char* pdf_name, int num, const void* buffer_void,
                      int stride, int width, int height) {
   if (!CheckDimensions(stride, width, height))
     return;
 
   std::vector<unsigned char> png_encoding;
   const unsigned char* buffer = static_cast<const unsigned char*>(buffer_void);
   if (!image_diff_png::EncodeBGRAPNG(
           buffer, width, height, stride, false, &png_encoding)) {
     fprintf(stderr, "Failed to convert bitmap to PNG\n");
@@ skipping 222 matching lines @@
         fprintf(stderr, "Duplicate or conflicting --ppm argument\n");
         return false;
       }
       options->output_format = OUTPUT_PPM;
     } else if (cur_arg == "--png") {
       if (options->output_format != OUTPUT_NONE) {
         fprintf(stderr, "Duplicate or conflicting --png argument\n");
         return false;
       }
       options->output_format = OUTPUT_PNG;
+    } else if (cur_arg == "--txt") {
+      if (options->output_format != OUTPUT_NONE) {
+        fprintf(stderr, "Duplicate or conflicting --txt argument\n");
+        return false;
+      }
+      options->output_format = OUTPUT_TEXT;
 #ifdef PDF_ENABLE_SKIA
     } else if (cur_arg == "--skp") {
       if (options->output_format != OUTPUT_NONE) {
         fprintf(stderr, "Duplicate or conflicting --skp argument\n");
         return false;
       }
       options->output_format = OUTPUT_SKP;
 #endif
     } else if (cur_arg.size() > 11 &&
                cur_arg.compare(0, 11, "--font-dir=") == 0) {
@@ skipping 154 matching lines @@
     switch (options.output_format) {
 #ifdef _WIN32
       case OUTPUT_BMP:
         WriteBmp(name.c_str(), page_index, buffer, stride, width, height);
         break;
 
       case OUTPUT_EMF:
         WriteEmf(page, name.c_str(), page_index);
         break;
 #endif
+      case OUTPUT_TEXT:
+        WriteText(page, name.c_str(), page_index);
+        break;
+
       case OUTPUT_PNG:
         WritePng(name.c_str(), page_index, buffer, stride, width, height);
         break;
 
       case OUTPUT_PPM:
         WritePpm(name.c_str(), page_index, buffer, stride, width, height);
         break;
 
 #ifdef PDF_ENABLE_SKIA
       case OUTPUT_SKP: {
@@ skipping 198 matching lines @@
   printf("%s\n", config.c_str());
 }
 
 static const char usage_string[] =
     "Usage: pdfium_test [OPTION] [FILE]...\n"
     "  --show-config     - print build options and exit\n"
     "  --send-events     - send input described by .evt file\n"
     "  --bin-dir=<path>  - override path to v8 external data\n"
     "  --font-dir=<path> - override path to external fonts\n"
     "  --scale=<number>  - scale output size by number (e.g. 0.5)\n"
+    "  --txt - write page text in UTF32-LE <pdf-name.<page-number>.txt\n"

[hal.canary, 2016/06/16 14:32:15]

<pdf-name>.<page-number>.txt

[dsinclair, 2016/06/16 14:34:37]

Done.

 #ifdef _WIN32
     "  --bmp - write page images <pdf-name>.<page-number>.bmp\n"
     "  --emf - write page meta files <pdf-name>.<page-number>.emf\n"
 #endif  // _WIN32
     "  --png - write page images <pdf-name>.<page-number>.png\n"
 #ifdef PDF_ENABLE_SKIA
     "  --skp - write page images <pdf-name>.<page-number>.skp\n"
 #endif
     "";
 
@@ skipping 77 matching lines @@
   }
 
   FPDF_DestroyLibrary();
 #ifdef PDF_ENABLE_V8
   v8::V8::ShutdownPlatform();
   delete platform;
 #endif  // PDF_ENABLE_V8
 
   return 0;
 }