The purpose of
libpandoc is to make the Haskell library
Pandoc available for use from C and other non-Haskell
environments that support C FFI. Pandoc and
text conversion between HTML, Markdown, LaTeX, OpenDocument and other
The latest version is available at GitHub.
is licensed under GPL version 2 or later, which is also Pandoc's
Building follows standard Haskell conventions and requires the Haskell Platform:
$ cabal configure [--global] $ cabal install --only-dependencies $ cabal build
A successful build creates the shared library file in
./dist/build/libpandoc.so/libpandoc.so. Installation and use of
the library is platform-dependent.
Note: On Windows, you should edit the
libpandoc.cabal file to rename
the extension of the shared object to dll.
that take care of installing the header and library in
For your convenience, an
./install.bat is provided that installs the
shared library under
The C interface is defined in
#include <pandoc.h> pandoc_init(); char* error = pandoc(1024 /* buffer size */, "markdown" /* input format */, "html" /* output format */, NULL /* JSON settings */, reader /* the reader function */, writer /* the writer function */, user_data /* private user data */); pandoc_exit();
gcc [my-file.c] -lpandoc
Haskell runtime has to be started and stopped explicitly via the
The reader function is in the following form:
int reader(char *buf, void *user_data);
buf is the buffer to be filled. The size of this buffer is the same
as provided as the first argument to the
the same pointer passed as the last argument of the
pandoc function. The
reader function must fill the buffer with the input to be converted by Pandoc.
The return value is the number of characters read. The reader is no longer called
when this value is zero.
The writer function is in the following form:
void writer(const char *buf, int len, void *user_data);
buf is the buffer to be written,
len is the number of elements in
the buffer and
user_data is the last argument of the
user_data of the reader. The writer function must write the
contents of the buffer as the output of the conversion by Pandoc.
Input and Output Formats
Input and output formats depend on Pandoc version the library is built
against. They are passed as strings. Reader and writer names are the same
as understood by Pandoc, for example
The settings parameter allows to customize the text transformation by passing Pandoc settings as an JSON-encoded string. The JSON format is derived automatically from Pandoc data type declarations by generic programming.
defaults.json is a printout of the default settings. NOTE: it may be outdated
with respect to the current library version. The custom settings
passed by the user are merged with the default settings, so only the
fields that have non-default values have to be provided.
- A .NET/C# interface is available as libpandoc-dotnet (currently outdated).
- Switched to JSON format for settings instead of XML
- Completed support for all Pandoc readers and writers
- Updated to Pandoc version 1.16 and higher
- 0.7 - Updated to Pandoc version 1.13 and higher
- 0.6 - Updated to Pandoc version 1.10 and higher
- 0.5 - Implemented XML generics to support all config settings.
Original author is Anton Tayanovskyy firstname.lastname@example.org.
Shahbaz Youssefi email@example.com is the current maintainer. Bug reports and feature requests are welcome.
- Katherine Whitlock