Getting Started with PDFium

This guide walks through some examples of using the PDFium library. For an example of using PDFium see the Chromium PDF Plugin.

Prerequisites

You will need the PDFium library on your computer. You can see the README for instructions on getting and installing PDFium.

*** note You must compile PDFium without both V8 and XFA support for the examples here to work. V8 can be compiled out by providing GYP_DEFINES="pdf_enable_v8=0 pdf_enable_xfa=0" build/gyp_pdfium.

See the V8 Getting Started guide for how to initialize PDFium when V8 is compiled into the binary.

---

PDFium Headers

PDFium's API has been broken up over several headers. You only need to include the headers for functionality you use in your application. The full set of headers can be found in the public/ folder of the repository.

In all cases you'll need to include fpdfview.h as it defines the needed methods for initialization and destruction of the library.

Initializing PDFium

The first step to using PDFium is to initialize the library. Having done so, you'll need to destroy the library when you're finished. When initializing the library you provide the FPDF_LIBRARY_CONFIG parameters to FPDF_InitLibraryWithConfig.

#include <fpdfview.h>

int main() {
  FPDF_LIBRARY_CONFIG config;
  config.version = 2;
  config.m_pUserFontPaths = NULL;
  config.m_pIsolate = NULL;
  config.m_v8EmbedderSlot = 0;

  FPDF_InitLibraryWithConfig(&config);

  FPDF_DestroyLibrary();
  return 0;
}

Currently the config.version must be set to 2. m_pUserFontPaths can be used to override the font paths searched by PDFium. If you wish to use your own font paths pass a NULL terminated list of const char* paths to use.

m_pIsolate and m_v8EmbedderSlot are both used to configure the V8 javascript engine. In the first case, you can provide an isolate through m_pIsolate for PDFium to use to store per-isolate data. Passing NULL will case PDFium to allocate a new isolate. m_v8EmbedderSlot is the embedder data slot to use in the v8::Isolate to store PDFium data. The value must be between 0 and v8::Internals::kNumIsolateDataSlots. Typically, 0 is a good choice.

For more information on using Javascript see the V8 Getting Started guide.

*** aside PDFium is built as a set of static libraries. You'll need to specify them all on the link line in order to compile. My build line was:

PDF_LIBS="-lpdfium -lfpdfapi -lfxge -lfpdfdoc -lfxcrt -lfx_agg \
-lfxcodec -lfx_lpng -lfx_libopenjpeg -lfx_lcms2 -lfx_freetype -ljpeg \
-lfx_zlib -lfdrm -lpdfwindow -lbigint -lformfiller -ljavascript \
-lfxedit"
PDF_DIR=<path/to/pdfium>

clang -I $PDF_DIR/public -o init init.c -L $PDF_DIR/out/Debug -lstdc++ -framework AppKit $PDF_LIBS

The -framework AppKit as needed as I'm building on a Mac. Internally PDFium uses C++, which is why -lstdc++ is required on the link line.

---

Loading a Document

One of the main objects in PDFium is the FPDF_DOCUMENT. The object will allow access to information from PDFs. There are four ways to to create a FPDF_DOCUMENT. FPDF_CreateNewDocument will create an empty object which can be used to create PDFs. For more information see the PDF Editing Guide.

Loading an existing document is done in one of three ways: loading from file, loading from memory, or loading via a custom loader. In all three cases you'll provide a FPDF_BYTESTRING which is the password needed to unlock the PDF, if encrypted. If the file is not encrypted the password can be NULL.

The two simplest methods are loading from file and loading from memory. To load from file, you'll provide the name of the file to open, including extension. For loading from memory you'll provide a data buffer containing the PDF and its length.

FPDF_STRING test_doc = "test_doc.pdf";
FPDF_DOCUMENT doc = FPDF_LoadDocument(test_doc, NULL);
if (!doc) {
  return 1;
}

FPDF_CloseDocument(doc);

In all three cases, FPDF_LoadDocument, FPDF_LoadMemDocument, FPDF_LoadCustomDocument a return of NULL indicates an error opening the document or that the file was not found.

You can use FPDF_GetLastError to determine what went wrong.

#include <fpdfview.h>
#include <unistd.h>
#include <stdio.h>

int main() {
  FPDF_LIBRARY_CONFIG config;
  config.version = 2;
  config.m_pUserFontPaths = NULL;
  config.m_pIsolate = NULL;
  config.m_v8EmbedderSlot = 0;

  FPDF_InitLibraryWithConfig(&config);

  FPDF_DOCUMENT doc = FPDF_LoadDocument(test_doc, NULL);
  if (!doc) {
    unsigned long err = FPDF_GetLastError();
    fprintf(stderr, "Load pdf docs unsuccessful: ");
    switch (err) {
      case FPDF_ERR_SUCCESS:
        fprintf(stderr, "Success");
        break;
      case FPDF_ERR_UNKNOWN:
        fprintf(stderr, "Unknown error");
        break;
      case FPDF_ERR_FILE:
        fprintf(stderr, "File not found or could not be opened");
        break;
      case FPDF_ERR_FORMAT:
        fprintf(stderr, "File not in PDF format or corrupted");
        break;
      case FPDF_ERR_PASSWORD:
        fprintf(stderr, "Password required or incorrect password");
        break;
      case FPDF_ERR_SECURITY:
        fprintf(stderr, "Unsupported security scheme");
        break;
      case FPDF_ERR_PAGE:
        fprintf(stderr, "Page not found or content error");
        break;
      default:
        fprintf(stderr, "Unknown error %ld", err);
    }
    fprintf(stderr, ".\n");
    goto EXIT;
  }

  FPDF_CloseDocument(doc);
EXIT:
  FPDF_DestroyLibrary();
  return 0;

While the above are simple, the preferable technique is to use a custom loader. This makes it possible to load pieces of the document only as needed. This is useful for loading documents over the network.