Using the Library

Working with XpdfText

The XpdfText library uses an opaque handle (type PDFHandle) to represent a PDF file. Multiple PDF files can be open simultaneously (each with its own handle).

Any program that uses the library must include the XpdfText header file, as well as the XpdfInfo header file if you use any of the info extraction functions:

#include "XpdfInfo.h" #include "XpdfText.h"

Using XpdfText, you can load PDF files and convert them to text. Typical code looks like this:

PDFHandle pdf; int err, n; err = pdfLoadFile(&pdf, "c:/test/file.pdf"); if (err != pdfOk) { /* handle the error */ } n = pdfGetNumPages(pdf); if (!pdfConvertToTextFile(pdf, 1, n, "file.txt")) { /* handle the error */ }

Using XpdfText in a multithreaded application

In a multithreaded application, the pdfInitLibrary function must be called before any other functions are called. Unlike in single-thread applications where this is optional, the pdfInitLibrary call is required in multithreaded applications. Each PDF handle must be used by only one thread. Given that constraint, all XpdfText functions (other than pdfInitLibrary) are thread-safe.

Compiling & linking on Windows

The XpdfText library is supplied as a DLL (XpdfText.dll) and an import library (XpdfText.lib).

The following instructions are for Microsoft Visual C++ 6. Similar steps should work for other development environments.

Add the include file directory: in the "Project Settings" dialog, under the "C/C++" tab, in the "Preprocessor" category, add the library include file directory (....\XpdfText\include).
Add the import library: in the "Project Settings" dialog, under the "Link" tab, in the "General" category, add the library (....\XpdfText\lib\XpdfText.lib).
Either add the library directory (....\XpdfText\lib) to your executable search path, or copy XpdfText.dll into the same directory as your application's executable.

Compiling & linking on Linux

The XpdfText library is supplied as a shared library (libXpdfText.so).

When compiling C or C++ code that uses the XpdfText library, you'll need to supply a "-I" flag pointing to the directory containing the XpdfText includes. When linking, you'll need to supply a "-L" flag pointing to the directory containing the XpdfText library, and a "-lXpdfText" flag to link with the library.

gcc -c -I/usr/local/XpdfText/include application.c gcc -o application application.o \ -L/usr/local/XpdfText/lib -lXpdfText

Look at the Makefile in the example code for a complete demonstration.

Before running the application, make sure that the XpdfText library directory is on the library search path. This this can be done either by setting the LD_LIBRARY_PATH environment variable or by editing the system-wide /etc/ld.so.conf configuration file.

Compiling & linking on Mac OS X

Using XpdfText on OS X is very similar to using it on Linux. The shared library has a different extension (libXpdfText.dylib), and you'll need to set the DYLD_LIBRARY_PATH environment variable.

Example code

The XpdfText library distribution includes four sample programs, located in the examples directory:

cvtPDFToText: convert a PDF file to a text file
cvtPDFToString: similar to cvtPDFToText, but does the conversion to a memory buffer
words: prints a word list, in a pseudo-XML format, for the first page of a PDF file
rectangle: extracts text from several rectangles on a PDF page

To build on Linux, edit the included Makefile and set the XPDFLIBDIR, XPDFINCDIR, and LIB variables according to the instructions inside the Makefile. Then run "make".

To build on Windows, create a Visual C++ project, as described above.