todo

Sun公司Dream项目

Fix bug where peek 4, get 1 loop will cause data to continually come
from resid buffer (if resid buffer will fit back into next buffer, copy
it there.)

Add code to turn CRC checks on and off in Mpeg2Pat and Mpeg2Pmt, then
if Crc's are on call pipeCrcClear calls in right places.

Verify all free's and pipe cleanup.

Move free routines up by new routines.

Make parser object in MpegDump.c

Return relative position in info for parsers where it makes sense
	(e.g. ES, PSI's)

Add bit-level put routines to Pipe.

Make sure all stuff bytes are now skipped with pipeSkipStuffBytes()

Put source in more hierarchical structure:
	mmp

	mpeg

	mpeg1    mpeg2

Make includes follow this path, too

Variable Naming Guidelines
	Map field names by mapping all upper case letters to lower case,
	eliminating _, and capitalizing letter after _

	Fields which have a defined bit pattern are named, b0000,
	where 0000 represents the binary bit pattern msb-first.

	Duplicate field names are resolved by appending an _1,
	_2, etc.  (For markerBit fields, b000 fields, etc).

	Fields that span multiple bytes are broken into segments
	that are contained within a single byte, each subfield is
	named by appending a _ separated first bit, last bit
	specification (e.g. programClockReference23_16)

	Types are named by either mpeg parser descriptor name
	prefixed by module name (e.g. MpegTs), or if no mpeg
	parser given, by name of controlling flag.  Names are
	constructed like field names except initial letter is
	capitalized.

	Function macros are named by prefix with all caps associated
	type name with words separated by _ followed by field name
	on which they operate.

	Enum elements are named by mapping type name to all caps,
	separating words by _,  then appending a descriptive identifier
	follow the same conventions.

	Value macros are named by mapping the type name to which they
	apply to all caps, separating words by _ and then followed by
	the field name to which they apply, finally followed by a
	descriptive identifier.

	Types which are not complete definitions of the corresponding
	Mpeg type (because of variable length fields, payloads, etc)
	should have a comment indicating the missing data.

	In all cases where duplicate words show up in names, the
	duplicate words may be eliminated