HELP REGEXP_LIB              match / all / split / replace conveniences

    uses regexp;

A friendly surface over Poplog's built-in regular expression engine
(the one Ved search uses; REF * REGEXP documents it fully).  Compiled
patterns are cached, so repeating a pattern string costs one property
lookup.

PATTERN SYNTAX (Ved style, not PCRE): ordinary characters are
literal; operators take a @ prefix:

    @.          any character
    @*          zero or more of the previous item
    @[a-z@]     character class ( @[^...@] negated )
    @< @>       word boundaries
    @a @z       anchor at start / end of string
    @?          any single item (Ved's wildcard)

So digits are '@[0-9@]@[0-9@]@*', and a line of '-'s is '@a-@*@z'.

regexp_search(PATTERN, S, I) -> (START, NCHARS)
    First match at or after byte position I; both false if none.
    substring(START, NCHARS, S) is the matching text.

regexp_matches(PATTERN, S) -> BOOL

regexp_first(PATTERN, S) -> SUBSTRING or false

regexp_all(PATTERN, S) -> LIST
    Every non-overlapping match, left to right.

regexp_split(PATTERN, S) -> LIST
    S split around matches:
        regexp_split('@[,;@]', 'a,b;c') -> ['a' 'b' 'c']

regexp_replace(PATTERN, S, NEW) -> S
    Every match replaced by the literal string NEW.

Zero-length matches advance by one character, so walks always
terminate.  A malformed pattern (e.g. '@[a-z' with no closing class)
mishaps with the engine's error message.

See also: REF * REGEXP (full syntax incl. case-folding and long
expressions), LIB * STRUTILS (fixed-string operations),
tools/tests/test_regexp.p.

--- pop/help/regexp_lib