Implementation of a caching mechanism for compressed data. This cache uses simple advisory locking found on most modern unix file systems. Compressed files are decompressed and stored in a cache where they can be used over and over until removed from the cache. Several processes can share the cache with each reading from files. At the same time, new files can be added and the cache can be purged, without disrupting the existing read operations
Marshaller that knows how to marshal/serialize dap data objects to a C++ iostream using DAP4's receiver-makes-right scheme. This code adds checksums to the stream and uses the xdr library to encode real values if the underlying representation is not IEEE 754. It also supports computing the checksum only
A C++ stream class for chunked data. This class uses the chunked_outbuf class to provide for chunked binary serialization of data as specified by DAP4. Information to be serialized is broken into 'chunks' that are no more than 2^24 bytes in length. Each chunk is prefixed by a 4 byte header that indicates the type of chunk and size (number of bytes in the chunk body). There are three types of chunk: Data; End; and Error. In normal operation, a DAP4 data document/response is serialized as a sequence of DATA chunks followed by one END chunk (which may be zero bytes in length). If, during serialization, an error is detected, the currently buffered (but not sent) data are discarded and an ERROR chunk is sent with an error message
Output buffer for a chunked stream This performs buffered output encoding the data in the stream using the simple chunking protocol defined for DAP4's binary data transmission. Each block of data is prefixed by four bytes: A CHUNK TYPE byte followed by three bytes that are the CHUNK SIZE. There are three CHUNK TYPES: data, end and error, indicated by the code values 0x00, 0x01 and 0x02. The size of a chunk is limited to 2^24 data bytes + 4 bytes for the chunk header