Defining Word Boundaries and Line Breaks
- Checking and Modifying Script Manager Global and Local Variables
- Accepting Implicit Script Codes
- Verbs for GetScript and SetScript
- Verbs for GetEnvirons and SetEnvirons
- Setting and Checking System Variables
- Setting the Active Keyboard Script
- Obtaining Script Information
- Manipulating Text
- Date and Time Utilities
- Number Utilities
and associated tables to determine word boundaries and line breaks.
boundaries of the word that includes a specified character in the block.
selection and words for word wrap (line breaking).
positions between characters-for example, an offset of 1 in Roman
text is between the first and second characters (or on the trailing
edge of the first and on the leading edge of the second).
NFindWord uses a state machine to determine word boundaries. The state
machine must start at a point at or before the beginning of the word that
includes the specified character. This can be accomplished in two ways. First,
if the specified character is sufficiently close to the beginning of the text
buffer, the state machine simply starts from the beginning of the buffer. This
is determined by the doBackupMin parameter in the tables: if the offset
parameter is less than doBackupMin, the state machine starts from the
specified character, moving backward as necessary until it encounters a word
break.
Once determined, this starting point is saved as an initial word break location.
ForwardTable until it encounters another word break. If that word break is
still before the specified character, its location is saved as the starting point,
and the state machine is restarted from that location. This process repeats
until the state machine finds a word break that is after the specified character.
current word break location as the offset pair defining the word boundaries.
The state machine operates in a similar manner whether moving backward or
forward; any differences in behavior are determined by the tables. The
machine begins in the start state (state 1). It then cycles one character at a
time until it finds a word break and exits. Each cycle proceeds as follows: the
current character is mapped to a class number. The character class and the
current state are used as indices into a two-dimensional array of byte- length
action codes. Each action code specifies the following:
• whether to mark the current offset
• the next state, which may be the exit state (state 0)
When the state machine exits, it has encountered a word break. The location of
the word break is the last character offset that was marked. In general, the
state machine marks a character offset when it determines that the word that
began at its starting point extends at least to the marked offset.
The Figure below gives two examples of the forward operation of the state
machine for word selection. In each case, the state at a given offset and the class
of the character following the offset determine (1) whether to mark that
offset, (2) whether to exit at that point, and, if not exiting, (3) what the state
at the next offset will be. When the state machine exits, the first and last
marked offsets are returned as the word boundaries.
Forward operation of the state machine for word selection
Mapping characters to classes is simple for single-byte characters, but the
process gets a little more involved for double-byte character codes. The byte
value at the current character offset is used as an offset into the ClassTable
array, an array of 256 signed bytes. If the value in ClassTable is positive, it
signifies that the byte at the current character offset is a single-byte
character, and the value in ClassTable is the class number for the character. If
the value in ClassTable is negative, it signifies that the byte at the current
character offset is the first byte of a double-byte character code, and
AuxClassTable must be used to determine the character class.
The AuxClassTable begins with a variable-length word array. The first word
contains a default class number for double-byte character codes. The following
words are offsets to RowTables, which have the same format as ClassTable, but
character code to a class number. If the value in ClassTable was -1 (or any
odd, negative number), the double-byte character code is assigned the default
class from the first word of AuxClassTable. For other double-byte characters,
value to provide an offset from the beginning of AuxClassTable to the
appropriate RowTable offset. The RowTable table specified by this offset in this
way is used to map the second byte of the character to a class number.
Note: There is a maximum of 128 classes and 128 states
(including the start and exit states).
The Figure below provides a de scription of the new break table. Note that the
high bit of the first word is set to indicate that this table is in the new format;
break table begins with an 8-word header, followed by the class and state
tables.
NFindWord header and class tables
words containing byte offsets from the beginning of the state table to the rows
of the state table; this is followed by a C-by-S byte array, where C is the
number of classes and S is the number of states. The bytes in this array are
stored with the column index varying most rapidly; that is, the bytes for the
State 1 row precede the bytes for the State 2 row. Each byte in this array is an
action code whose format is defined in the figure below.
NFindWord state table
The Figure below shows the format of an action code.
