Snobol4 Reference Manual: Chapter 7

Chapter 7 : PATTERNS AND PATTERN FUNCTIONS

The SNOBOL4 pattern matcher is called the "scanner." The "cursor" is the scanner's pointer into the subject string; it points between subject characters (no relation to your CRT cursor). It is initially zero when positioned to the left of the subject, and is incremented as the scanner moves to the right in the subject.

7.1 PRIMITIVE PATTERNS

These variables initially contain the primitive patterns of the same name. They may be set to other values by a program, and restored to their original value from the corresponding protected keywords.

ABORT
Causes immediate failure of the entire pattern match, without seeking alternatives.
ARB
Matches zero or more characters of the subject string. It matches the shortest possible substring.
BAL
Matches any nonnull string which is balanced with respect to parentheses. A string without parentheses is considered balanced. BAL matches the shortest string possible.
FAIL
Causes failure of this portion of the pattern match, causing the scanner to backtrack and try alternatives.
FENCE
Matches the null string and succeeds when the scanner is moving left to right in a pattern, but fails if the scanner has to back up through it, seeking alternatives.
REM
Matches zero or more characters from the current cursor position to the end of the subject string.
SUCCEED
Matches the null string and always succeeds.

Altering these primitive patterns can produce very confusing programs, unless the new value encompasses the old, like this:

    ARB = &ARB . OUTPUT

7.2 PRIMITIVE PATTERN FUNCTIONS

These functions produce a pattern based on the argument supplied. The argument data type is shown below---other data types or expressions will be converted to the required type if possible.

Pattern functions may be combined with other primitive patterns, functions, and strings using the alternation and concatenation operators to produce larger patterns.

ANY(string) Match one character from set
Matches exactly one character from the set of characters specified by the argument string.
ARBNO(pattern) Match repeated pattern
Matches zero or more consecutive occurrences of the string matched by the argument pattern. ARBNO matches the shortest string possible--initially the null string--and only tries to match pattern if other pattern components in the statement require it.
BREAK(string) Match characters not in set
Matches zero or more characters provided they are not in the set of characters in the argument string. That is, it matches up to, but not including, a character from the argument string.
LEN(integer) Match fixed length string
Matches a string of the specified length. There are no restrictions on the subject string characters. An argument of zero will match the null string.
NOTANY(string) Match one character not in set
Matches exactly one character provided it is not in the set of characters specified by the argument string.
POS(integer) Verify scanner position
Succeeds if the scanner's current cursor position in the subject string is equal to the specified integer value. This function merely verifies scanner position---it does not consume or match any subject characters. POS(0) as the first component of a pattern produces an anchored pattern match.
RPOS(integer) Verify scanner position from end
Succeeds if the scanner's current cursor position in the subject string is the specified number of characters from the end of the string. Like POS(), it verifies scanner position but does not consume any characters. RPOS(0) as the last component of a pattern forces the pattern to match to the end of the subject string.
RTAB(integer) Match through position counting from end
Matches all characters from the current cursor position up to the specified cursor position, counting from the end of the subject string. RTAB(N) matches characters up to, but not including, the final N characters of the subject.
SPAN(string) Match characters in set
Matches one or more characters from the set of characters specified by the argument string. SPAN will not match the null string; at least one character from the argument string must be found in the subject.
TAB(integer) Match through fixed position
Matches all characters from the current cursor position up to the specified cursor position. TAB(N) matches characters up to, and including, the initial N characters of the subject. TAB will match the null string if the target position and current cursor position are the same. The function fails if the current scanner position is to the right of the target position.

Previous chapter · Next

Next chapter ·

Table of Contents