[andy/viking.git] / doc / dev / File-Format-Outline

WARNING: OLD file format. Will be discarded. Read FF2
----------
Magic Number:
6c45e131

lE..

at beginning of file. I love magic numbers.

For Serialization, we need to separate realizing a layer and adding it to a aggregate. An aggregate 
layer, when realized, should realize all of its children, and add all of its children to the treeview. 
An aggregate should realize a layer and add it to the treeview in add_layer only if it is already 
realized (we cannot assume that it is.) This will pave the road for serialization.

Tree
  Child
    Child


START
(TopLayer)
layer type ID
layer name len + layer name
Layer Data. Call Layer serialize-in func.

(if type == aggregate, len = # of layers)

Example:

Top
+ Aggy
  + Aggy2
    - TRW
    - MAP
  - COORD
+ Aggy1
  - TRW
  - TRW
  - TRW

File:

0	(Aggregate)
4
Aggy
2	(2 Sublayers)

0	(Aggregate)
5
Aggy2
2	(2 Subs)

1	(TRW)
3
TRW
(TRW DATA)

2	(MAP)
3
MAP
(MAP DATA)

3	(COORD)
5
COORD
(COORD DATA)

1
5
Aggy1
3	(3 subs)

1
3
TRW
100
(TRW DATA)

1
3
TRW
100
(TRW DATA)

1
3
TRW
100
(TRW DATA)

(END)

Parsing
Call agg parse function (top layer) with file format.
agg parse function parses, calls appropriate function.
repeats add infinitum.

----

Functions:

Aggregate Read:
for loop thru # of layers:
  read a byte, get id, create a layer of such type via the "serialize_read" func which 
  returns a layer I can then read.
  repeat, enjoy.

Ahhh, the joys of Recursion.

Aggregate Write:
write length of layers array.
for each layer:
  write ID.
  run serialize_write.


MAX # of layers: 2^16
Commit	Line	Data
50a14534 EB	1	WARNING: OLD file format. Will be discarded. Read FF2
	2	----------
	3	Magic Number:
	4	6c45e131
	5
	6	lE..
	7
	8	at beginning of file. I love magic numbers.
	9
	10	For Serialization, we need to separate realizing a layer and adding it to a aggregate. An aggregate
	11	layer, when realized, should realize all of its children, and add all of its children to the treeview.
	12	An aggregate should realize a layer and add it to the treeview in add_layer only if it is already
	13	realized (we cannot assume that it is.) This will pave the road for serialization.
	14
	15	Tree
	16	Child
	17	Child
	18
	19
	20	START
	21	(TopLayer)
	22	layer type ID
	23	layer name len + layer name
	24	Layer Data. Call Layer serialize-in func.
	25
	26	(if type == aggregate, len = # of layers)
	27
	28	Example:
	29
	30	Top
	31	+ Aggy
	32	+ Aggy2
	33	- TRW
	34	- MAP
	35	- COORD
	36	+ Aggy1
	37	- TRW
	38	- TRW
	39	- TRW
	40
	41	File:
	42
	43	0 (Aggregate)
	44	4
	45	Aggy
	46	2 (2 Sublayers)
	47
	48	0 (Aggregate)
	49	5
	50	Aggy2
	51	2 (2 Subs)
	52
	53	1 (TRW)
	54	3
	55	TRW
	56	(TRW DATA)
	57
	58	2 (MAP)
	59	3
	60	MAP
	61	(MAP DATA)
	62
	63	3 (COORD)
	64	5
65	COORD
66	(COORD DATA)
67
68	1
69	5
70	Aggy1
71	3 (3 subs)
72
73	1
74	3
75	TRW
76	100
77	(TRW DATA)
78
79	1
80	3
81	TRW
82	100
83	(TRW DATA)
84
85	1
86	3
87	TRW
88	100
89	(TRW DATA)
90
91	(END)
92
93	Parsing
94	Call agg parse function (top layer) with file format.
95	agg parse function parses, calls appropriate function.
96	repeats add infinitum.
97
98	----
99
100	Functions:
101
102	Aggregate Read:
103	for loop thru # of layers:
104	read a byte, get id, create a layer of such type via the "serialize_read" func which
105	returns a layer I can then read.
106	repeat, enjoy.
107
108	Ahhh, the joys of Recursion.
109
110	Aggregate Write:
111	write length of layers array.
112	for each layer:
113	write ID.
114	run serialize_write.
115
116
117	MAX # of layers: 2^16
118