[Cyan]<Tioga2>Base>T2FileOps.mesa!2

T2FileOps.Mesa

Last Edited by: Paxton, August 19, 1983 11:20 am

File representation for Tioga2

File representation

file = text control properties trailer

text = characters of text nodes separated by CR's and padded with 0 at end if necessary

control = control information for parsing the file; starts with 0 to indicate end of plain text part of file

properties = {key value} -- both keys and values stored as ropes

trailer = -- 14 bytes at end of file

1 byte controlTrailerId, 1 byte (DataPaddedFlag | DataUnpaddedFlag), 1 byte version number, 1 byte of flags, 2 bytes for length of properties, 4 bytes for length of text, 4 bytes for length of file

Branch node

branch =

OP [if OP.class then className] [if OP.format then formatName] {prop}

(BRANCHCONTENTSPECS contentsSpecs | {item}) -- the contents of the branch

(NOCHILD -- branch has no children -- |

CHILDSPECS childSpecs -- class op will create children from the specs -- |

STARTC {child} ENDC -- the children will be parsed left to right when parse the parent -- |

ENDITEMS {child} ENDC {ctrlLen dataLen} numChildren -- can create unparsed children --)

The OP also includes a comment bit.

The ctrlLen dataLen pairs are in reverse order for the children. Thus when read from end, get info for first child first.

Need opcode after items since branches don't have a special starting code, and therefore we wouldn't be able to tell the end of the items from the beginning of the children if the first child didn't have comment or properties.

The STARTC&ENDC version is called "whole tree mode" and is used for branches inside list or box nodes and for branches nested more than 2 levels deep in the document tree. In this mode, the child branches are parsed at the same time the parent is parsed, so there is no need for control and data lengths. If the parent is stored in whole tree mode, the children must all be stored that way also. No child can begin with ENDC, so the parsing of the children can simply go on until it encounters an ENDC.

We put an ENDC after the children even in the case where we give the lengths and number of children info so that we can be sure to be able to parse a branch even if we don't know the length of the control information ahead of time.

Basic, Box, or List node

basic, box, or list = OP [if OP.class then className] [if OP.format then formatName] {prop}

(STARTBR {branch} | ITEMCONTENTSPECS contentsSpecs | {item | basic}) (NODATA | CLASSDATA data)

The OP includes bits saying whether or not there is a class name and a format name.

The OP also includes a bit saying whether or not this is a comment node.

If there is no class name, the class defaults to "Basic", "Box", or "List" respectively.

If there is no format name, the format defaults to "Default".

There are no contents for a basic node.

If the contents are branches, they are represented using whole tree mode as defined above. Notice that the contents are either all branches or all non-branches.

No branch, item, or basic can begin with NODATA or CLASSDATA, so the parsing of the contents can go on until it encounters one of these ops.

Text node

text = OP [if OP.class then className] [if OP.format then formatName]

[if OP.runs then runs] [if OP.comment then text] [if OP.props then {prop} ENDPROPS]

The OP has bits for presence of class name, format name, runs, and properties.

There is also a comment bit saying whether or not this is a comment node. If it is, the text appears in the control info.

Miscellaneous

item = text | box | list

prop = PROPERTY propertyName propertyValueSpecs | APPLICATION | FORMAL

runs = numberOfRuns {run}

run = (lookSeq | NOLOOKS) runLength

lookSeq = {addLooks} lastLooks

addlooks = BYTE [addLooksFirst..addLooksLast]

lastLooks = BYTE [lastLooksFirst..lastLooksLast]

T2FileOps: CEDAR DEFINITIONS = BEGIN

General definitions

classBit: CARDINAL = 1; -- non-default class name

formatBit: CARDINAL = 2; -- non-default format name

commentBit: CARDINAL = 4; -- comment property set true

runsBit: CARDINAL = 8; -- text node has looks runs

propsBit: CARDINAL = 16; -- text node has properties

Op: TYPE = CHAR;

General ops

ENDOFFILE: Op = 0C;

NOCHILD: Op = ENDOFFILE+1;

BRANCHCONTENTSPECS: Op = NOCHILD+1;

CHILDSPECS: Op = BRANCHCONTENTSPECS+1;

ITEMCONTENTSPECS: Op = CHILDSPECS+1;

STARTC: Op = ITEMCONTENTSPECS+1;

ENDC: Op = STARTC+1;

STARTBR: Op = ENDC+1;

ENDITEMS: Op = STARTBR+1;

NODATA: Op = ENDITEMS+1;

CLASSDATA: Op = NODATA+1;

ENDPROPS: Op = CLASSDATA+1;

PROPERTY: Op = ENDPROPS+1;

APPLICATION: Op = PROPERTY+1;

FORMAL: Op = APPLICATION+1;

Looks related ops

numLooks: CARDINAL = 48; -- reserve extra 16 bits for future use such as OS character set

NOLOOKS: Op = FORMAL+1;

addLooksFirst: Op = NOLOOKS+1;

addLooksLast: Op = addLooksFirst+numLooks-1;

lastLooksFirst: Op = addLooksLast+1;

lastLooksLast: Op = lastLooksFirst+numLooks-1;

Branch nodes

branch: Op = lastLooksLast+1;

lastBranch: Op = branch+commentBit*2-1;

Text nodes

text: Op = lastBranch+1;

lastText: Op = text+propsBit*2-1;

List nodes

list: Op = lastText+1;

lastList: Op = list+commentBit*2-1;

Box nodes

box: Op = lastList+1;

lastBox: Op = box+commentBit*2-1;

Basic nodes

basic: Op = lastBox+1;

lastBasic: Op = basic+commentBit*2-1;

Lengths

lengths are given in as few bytes as possible according to following scheme

first bit of each byte tells if other bytes follow (up to 4 bytes max)

concat other 7 bits of bytes with leading 0's to form length

thus can represent lengths up to 2**28-1 if use 4 bytes

lengths of 127 or less fit in 1 byte

lengths in [128..4095] fit in 2 bytes

IntBytes: TYPE = MACHINE DEPENDENT RECORD [

first (0:9..15): [0..127] ← 0, -- first byte written on file

second (0:2..8): [0..127] ← 0, -- second byte written

thirdBottom (0:0..1): [0..3] ← 0, -- bottom bits of third byte written

thirdTop (1:11..15): [0..31] ← 0, -- top bits of third byte written

fourth (1:4..10): [0..127] ← 0, -- fourth byte written

unused (1:0..3): [0..15] ← 0 -- these are always 0

];

LengthByte: TYPE = MACHINE DEPENDENT RECORD [

unused (0:0..7): [0..255] ← 0,

others (0:8..8): BOOLEAN ← FALSE, -- tells if other bytes are present

data (0:9..15): [0..127] ← 0];

ThirdByte: TYPE = MACHINE DEPENDENT RECORD [

unused (0:0..7): [0..255] ← 0,

others (0:8..8): BOOLEAN ← FALSE,

dataTop (0:9..13): [0..31] ← 0, -- top 5 bits of data

dataBottom (0:14..15): [0..3] ← 0 -- bottom 2 bits of data

];

Trailer info

controlTrailerId: CHAR = 232C;

DataPaddedFlag: CHAR = 205C;

DataUnpaddedFlag: CHAR = 250C;

versionNumber: CHAR = 2C;

Trailer flags

isBranchFlagBit: CARDINAL = 1;

END.