Question

Given a regular expression with named captures, is it possible to determine which named captures are present in the re?

Note: I only have the regular expression – I don’t have a string which matches the re. I want to know if there is a way to look into the structure of the re and find all of the named captures used in the re.

Answer 1

That’s not possible in the general case, because someone could have a (??{....}) code insert that turns into something that uses a named capture of a hitherto unseen name during match execution time.

In a specific case you might be able to dig into the underlying structure from the C API. Devel::Peek suggests some places you might want to poke around:

% perl -MDevel::Peek -le 'Dump qr/(?<foo>\w+) (?&foo)/'
SV = IV(0x8033ac) at 0x8033b0
  REFCNT = 1
  FLAGS = (TEMP,ROK)
  RV = 0x803d40
  SV = REGEXP(0x8173d8) at 0x803d40
    REFCNT = 1
    FLAGS = (OBJECT,POK,FAKE,pPOK)
    PV = 0x216fc0 "(?^:(?<foo>\\w+) (?&foo))"
    CUR = 24
    LEN = 0
    STASH = 0x803b50    "Regexp"
    EXTFLAGS = 0x600000 (USE_INTUIT_NOML,USE_INTUIT_ML)
    INTFLAGS = 0x1
    NPARENS = 1
    LASTPAREN = 0
    LASTCLOSEPAREN = 0
    MINLEN = 3
    MINLENRET = 3
    GOFS = 0
    PRE_PREFIX = 4
    SEEN_EVALS = 0
    SUBLEN = 0
    SUBBEG = 0x0
    ENGINE = 0x162b60
    MOTHER_RE = 0x80fab0
    PAREN_NAMES = 0x80fa40
    SUBSTRS = 0x2044b0
    PPRIVATE = 0x206360
    OFFS = 0x204280

For example, PAREN_NAMES looks promising.

But that seems like a lot of work. What do you want to do this for?