Fabric: A framework for code generation for resource-constraint devices based on XML Schema/WSDL.
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Failed to load latest commit information.


Fabric is a framework for code generation based on several scientific
publications [1,2,3]. It is comprised of components for
  - parsing XML Schema and WSDL files into a tree, for

  - a module API to walk the tree for codegeneration, and for

  - code generation (C, C++, Java) using simple APIs instead of
    writing to stdout.

In addition, a number of modules are contained that generate code
  - microFibre: An XML (de-)compression scheme usable for
    ressource-constraint devices [2,3] that generates code for
    Java and C++.

  - macroFibre: An XML (de-)serialization scheme for
    ressource-constraint devices [3] that generates code for
    Java and C++.

  - XML: XML (de-)serialization for easy conversion from/to

  - protobuf: Converts XML Schema files to Google protobuf
    (http://code.google.com/p/protobuf/) files.

In 2011/12 the Fabric framework was extended by two more modules:
  - typegen: A generator to create JavaBean-like container
    classes from XML schema documents for both Java and C++.

  - exi: An implementation of the Efficient XML Interchange
    specification (http://www.w3.org/TR/exi) for XML (de-)

Fabric works as follows:
  1.) Input files are XML Schema and WSDL files which are parsed
    into a tree. This tree contains most information from the schema
    but abstracts away from all the nasty details of XML Schema
    and WSDL. It automatically resolves all imports, references,
    etc. and the resulting tree can directly be used for code
    generation based on tree-walking.

  2.) A so-called workspace is created that can be used for code
    generation. A workspace represents a set of files of a certain
    language (e.g., Java files) in memory. Users of the workspace use
    language-specific APIs to add components to the files (e.g., a
    Java class). To such elements, further stuff can be added (e.g.,
    methods). The benefit over directly writing code to stdout is
    that code can be added at any time in any order. Once the
    generation is done, the workspace can be serialized to the disk.

  3.) A set of modules is invoked to handle the parsed tree. Modules are
    basically tree-walkers that react to events that are the result
    from walking the tree (e.g., handling a root element of a schema).

  4.) The workspace is written to the disk.

Any feedback will be greatly appreciated, at the GitHub project page
(https://github.com/pfisterer/fabric) or by contacting
Dennis Pfisterer (fabric__A_T__farberg__de).

For feedback on the type generator or EXI module, please visit the
Fabric fork at https://www.github.com/nepa/fabric and contact one
of its maintainers.

No installation is required. To build fabric, you need Java 6 or higher
 and Maven 2 or higher (http://maven.apache.org/).

Before cloning this repository, be sure to enable automatic
conversion of CRLF/LF on your machine using "git config
--global core.autocrlf input". For more information, please
refer to http://help.github.com/dealing-with-lineendings/

Clone the repository using
"git clone git://github.com/pfisterer/fabric.git"

To build, run "mvn install", this will build fabric and place the
generated jar file in target/ and in your local Maven repository.


  java -jar core/target/fabric.core-1.0.one-jar.jar --help

Fabric is licenced under a BSD license. For details, see

If you use Fabric for scientific publications, I appreciate
if the papers below are referenced.

[1] Dennis Pfisterer, Horst Hellbrueck, and Stefan Fischer. Fabric:
  Towards data type-centric middleware synthesis. In Proceedings
  of the Euro-American Workshop on Middleware for Sensor Networks
  in conjunction with the International Conference on Distributed
  Computing (DCOSS 06), 2006.

[2] Dennis Pfisterer, Marco Wegner, Horst Hellbrueck, Christian Werner,
  and Stefan Fischer. Energy-optimized data serialization for
  heterogeneous WSNs using middleware synthesis. In Proceedings of
  The Sixth Annual Mediterranean Ad Hoc Networking Workshop
  (Med-Hoc-Net 2007), pages 180-187, June 2007.

[3] Dennis Pfisterer: Comprehensive Development Support for Wireless
  Sensor Networks, PhD Thesis, Institute of Telematics, University
  of Luebeck, Germany, 2007.