Previous Previous chapter · Next Next chapter · Contents Table of Contents

Chapter 3 : STATEMENTS

Each line of input to SNOBOL4 consists of a sequence of ASCII characters, terminated by a carriage return.

Comment and control statements are always one line long. However, a program statement may occupy several lines if necessary. A continuation mark (plus sign or period) is placed in the first column of the additional lines.


An asterisk (*) in character position one denotes a comment card. All text through the end-of-line is copied to the listing file, but is otherwise ignored by SNOBOL4.


Control statements provide instructions to the SNOBOL4 compiler. They begin with a minus (-) in character position one. Controls may be specified in upper- or lower-case, regardless of the current state of case-folding. Unrecognized controls are ignored. SNOBOL4 defaults to -LIST LEFT and -CASE 1.


If a line is not a control or comment statement, it is considered SNOBOL4 program text. A SNOBOL4 statement may have up to five components. The general form of a statement is:
Statement elements are separated by blank or tab.

Ignoring the LABEL and GOTO fields for a moment, the remaining elements may appear in various combinations to create different types of statements:

Evaluate expression: SUBJECT
The expression comprising the subject is evaluated. It may invoke primitive and program-defined functions.

Assignment statement: SUBJECT = REPLACEMENT
The value on the right is assigned to the variable on the left. If failure occurs when evaluating the subject or replacement components, the assignment does not occur.

Pattern match: SUBJECT PATTERN
The subject and pattern expressions are evaluated, and the specified pattern is applied to the subject string, producing success or failure.

Pattern match with replacement: SUBJECT PATTERN = REPLACEMENT
If the pattern match succeeds, the replacement expression is evaluated and replaces the portion of the subject matched. Only the matched portion is replaced; characters adjacent to the matching substring are not disturbed.

If the equal sign (=) is present but the replacement field is absent, the null string is assumed as the value of the replacement field.

The GOTO field provides two-way branching to test the success or failure of the preceding statement elements.

3.3.1 Label Field

If a label is present, it must begin with the first character of the line. Labels provide a name for the statement, and serve as the target for transfer of control from the GOTO field of any statement. Labels must begin with a letter or digit, optionally followed by an arbitrary string of characters. The label field is terminated by the character blank, tab, or semicolon. If the first character of a line is blank or tab, the label field is absent.

If case-folding is in effect, lower-case letters are converted to upper-case before defining the label.

3.3.2 Subject Field

The subject field specifies the string which will be the subject of pattern matching. It also specifies the left side of a simple assignment statement if pattern matching is absent.

In an assignment statement, the subject must be a variable name, an unprotected keyword, or a field-reference function from a program-defined data type. If a string is produced by evaluating an expression, the indirect ($) operator must be used to reference the underlying variable.

If the subject appears in pattern matching without replacement, the subject must evaluate to a string. The string is scanned left to right during the pattern match. If the subject evaluates to an integer, it is automatically converted to a string. If replacement is present, the same subject restrictions of assignment statements apply. Thus, a literal string is a valid subject only if replacement is absent.

If the expression comprising the subject contains the concatenation operator, the subject must be surrounded by parenthesis. This allows SNOBOL4 to distinguish concatenation blanks within the subject from the blank between subject and pattern.

3.3.3 Pattern Field

The pattern may be a simple string, or a complex expression involving primitive pattern functions. The pattern specifies one or more strings which are systematically searched for in the subject. The pattern match succeeds if a match is found, and fails otherwise. The &FULLSCAN keyword determines whether the search is exhaustive, or if heuristics will be applied to prevent futile match attempts.

The pattern may assign various matching components to variables with the binary assignment operators dot and dollar sign (., $).

3.3.4 Replacement Field

In an assignment statement, there are very few restrictions on the replacement field. If the subject is an unprotected keyword, the replacement field must evaluate to an integer value. If the subject is a variable, the replacement field is assigned directly to it, without type conversion.

If there is pattern matching on the left side of the statement, the replacement field must evaluate to a string, so that it may be inserted into the matched portion of the subject string.

Replacement occurs only if evaluation of the subject, pattern, and replacement succeed. Primitive functions which return success or failure may be used in the replacement field as predicate functions. Since they return the null string, they do not alter the replacement value. However, their failure can prevent replacement from occurring, and can be tested in the GOTO field.

3.3.5 GOTO Field

Statement execution normally proceeds sequentially from one statement to the next. The GOTO field allows this flow to be altered by directing the SNOBOL4 system to continue execution elsewhere. The GOTO field is set off from the preceding statement elements by blank or tab, and colon (:). It may assume three forms: unconditional, conditional, and direct.

The "unconditional GOTO" causes control to be transferred to the specified labeled statement. The label is enclosed in parenthesis, and may be a name, or the result of evaluating an expression and applying the indirect operator ($). Transfer is made to the labeled statement regardless of the success or failure outcome of the earlier parts of the statement.

The "conditional GOTO" similarly specifies control transfer to a labeled statement, but it depends on the success or failure of the statement. The letter S precedes the parenthesized label where control goes next if the statement succeeds. The letter F specifies the branch to be taken if the statement fails. For example:

    :S(LOOP)         Branches to label LOOP if the statement succeeds.

    :F(ERROR)        Branches to label ERROR if the statement fails.

    :S(OK) F(NOGO)   Branches to label OK on success, to NOGO on failure.

    :(AGAIN)         Unconditionally transfers control to label AGAIN.

    :($('VAR' N))    Branches to the label obtained by concatenating
                     the string 'VAR' with the value of variable N.
The "direct GOTO" is used to branch to a block of code compiled with the CODE function. If the code contains labels, a regular GOTO could branch to the label and begin execution in the code block. The direct GOTO will branch to the start of the code block, labeled or not. A direct GOTO is specified by placing in angle brackets the name of the variable which points to the code block.

Direct GOTOs may be made conditional by preceding them with S or F. They may also appear with regular GOTOs:

    VAR = CODE(string)         :S<VAR> F(COMPILE_ERROR)
The lower-case letters "s" and "f" may be used interchangeably with "S" and "F", regardless of case-folding.

The GOTO field may appear on a line without any subject, pattern, and replacement. The absent SNOBOL4 statement is assumed to have succeeded.


A SNOBOL4 statement may be divided across several lines by placing a plus (+) or period (.) in character position one of the successive lines. There is no limit to the number of continuation statements allowed. The statement must be divided at a point where a blank or tab could appear as an operator or separator; it cannot be split in the middle of a name or quoted string.

Very long strings may be entered on multiple lines, using the implicit blank between lines as a concatenation operator:

            LONG_STRING = "This is an example of a very long "
    +   "string that wends its way across multiple continua"
    +   "tion statements.  There is an implicit blank at the "
    +   "beginning of each line that provides the concatenation"
    +   " operator between segments."


The semicolon character may be used to place several statements on one line. Each semicolon terminates the current statement and behaves like a new "column one" for the statement which follows. Only program statements are permitted after the semicolon; control and continuation statements are not allowed. Here are some examples:
    I = 1;     J = 2;      S PAT = 'HENRI'       :S(YES)
    I = 1;OUT  OUTPUT = A<I>  :F(END);  I = I + 1 :(OUT)
Because of its poor readability, placing labels in the middle of a statement is strongly discouraged.

As a language extension, Vanilla SNOBOL4 permits a comment statement after the semicolon. This provides a simple device for end-of-line comments:

    PARA    NEXT = GETNEXT() :F(FRETURN) ;* Return if EOF
            IDENT(NEXT)      :S(RETURN)  ;* Return on empty line
            PARA = PARA NEXT :(PARA)     ;* Splice line


The last statement in a program must be an END statement. The word END appears in the label field, beginning in column one. Normally, it is the only word on the line:
            . . .
            OUTPUT = 'All done'
After reading the END statement, compilation ends, and execution begins immediately with the very first program statement. When the program is done, it should flow into the END statement, or use a GOTO to transfer to it.

Occasionally, we would like to begin execution at other than the first statement. If we place a statement label in the subject field of the END statement, execution will begin there. For example, this statement will cause execution to begin at the statement labeled START:

    END     START

Previous Previous chapter · Next Next chapter · Contents Table of Contents