Provides classes for converting Palm database data to/from a PalmDocument object, which can be used by the framework.

This package provides classes that handle the writing of data to an OutputStream object for the {@link org.openoffice.xmerge.DocumentSerializer DocumentSerializer} interface for; as well as the reading of data from an InputStream object for the framework's {@link org.openoffice.xmerge.DocumentDeserializer DocumentDeserializer} interface. Both these framework interfaces are simply converters from server-side documents to device specific documents and vice-versa. Since all Palm databases have a general record oriented format, a Palm database converter specific I/O stream format is specified for the Palm sync client application to handle the byte stream in a generic way. This also means that Palm database converters should read and/or write using this I/O stream format as specified in the next section.

Note that the format of the byte stream is not exactly that of a PDB file encoding. It does not need to contain the PDB header information nor record indices section. Instead, it contains the following ...

    set header
       4 bytes - creator id
       4 bytes - type id
       2 bytes - PDB header version
       2 bytes - PDB header attribute
       unsigned 2 bytes - number of PDB data to follow

    for each PDB,
       32 bytes - name of PDB i
       unsigned 2 bytes - number of records in PDB i

       for each record contained in PDB i,
          1 byte - record attributes
          unsigned 2 bytes - size of record j in PDB i
          x bytes - data

Note that each PDB section is appended by another if there is more than one.

Since the PalmDocument class takes care of the writing and reading of this format through its write and read methods, respectively, this format shall also be referred to as the PalmDocument stream format.

Usage of the classes for the specified I/O stream

When converting from a server document to device document(s), the framework requires writing the device document(s) to an OutputStream object via the DocumentSerializer interface. Note that a single server document may be converted into multiple PDB's on the Palm device. Each worksheet in the document is converted into a PalmDocument . Thus, if there is more than one worksheet in the document, more than one PalmDocument will be produced by the DocumentSerializer.

A DocumentSerializer creates a ConvertData object, which contains all of the PalmDocuments. The {@link org.openoffice.xmerge.converter.palm.PalmDocument#write write} method to write to the given OutputStream. The PalmDocument object will take care of writing the data in the specified format.

A DocumentDeserializer can use the PalmDocument object's {@link org.openoffice.xmerge.converter.palm.PalmDocument#read read} method to fill in all the PalmDocument object's data.

PDB file encoding/decoding

The PalmDocument object's read and write functions are provided by the PdbDecoder and PdbEncoder objects. The PdbEncoder class provides the functionality of encoding a PalmDB object into an InputStream, while the PdbDecoder class provides the functionality of decoding a PDB file into an OutputStream.

Refer to the class description of each for usage.

Important Note

Methods in these classes are not thread safe for performance reasons. Users of these classes will have to make sure that the usage of these classes are done in a proper manner. Possibly more on this later.

TODO list

1. Merge the PalmDB, PdbDecoder and PdbEncoder classes into the PalmDocument class.
2. After reading more on the palm file format spec, I realized that there are certain optional fields that may need to be addressed still, like the appInfo block and sortInfo block.
3. The current PdbDecoder only returns a PalmDB object. There are other information that we may want to expose from the PDB decoding process.
4. Investigate on different language encoding on the Palm and how that affects the PDB name.