next up previous contents index
Next: Polymorphism Up: Graph serialisation Previous: Roll your own   Contents   Index

Synopsis of pack_t

struct pack_t
{
  char *data;
  size_t size;
  size_t pos;
  Ptr_flag ptr_flag;
  std::vector<PtrStoreRef> alloced; //allocated data used for cleaning up 
  pack_t(size_t sz=0);
  pack_t(const char* filename, const char* mode); //pack to file
  pack_t& reseti();
  pack_t& reseto();
  pack_t& seeki(int offs);
  pack_t& seeko(int offs);
  void packraw(char *x, int sz); 
  void unpackraw(char *x, int sz);
};

data points to the beginning of the buffer maintained by pack_t. size refers to the current position of the input stream (ie the size of current valid data). pos refers to the current position of the output stream. It is an error to assign values directly to pos. It is OK to assign a value to size when setting up a pack_t variable for unpacking. Do not update size whilst packing. It is OK to assign a pointer value to data for unpacking only, however one should note that delete is called on the pointer during destruction, so in general you should reset data to NULL before the pack_t variable goes out of scope, if you don't want the object deleted (for instance if you've set it to the address of a static array).

size and pos can be reset to 0 using the reseti() and reseto() routines respectively. seeki() and seeko() allows arbitrary positioning of the streams -- the seek offset in this case is relative to the current position.

The constructor takes an integer argument which specifies the size of an initial buffer. For example:

pack_t b(N); b.size=N;
fread(b,N,1,f);
b>>foo;
is a common idiom for reading some data in from a file.

packraw and unpackraw allow arbitrary byte data to be pushed onto the buffer and taken off. This involves an extra copy operation, but is the safest way of manipulating the buffer directly.


next up previous contents index
Next: Polymorphism Up: Graph serialisation Previous: Roll your own   Contents   Index
Russell Standish 2016-09-02