Don't you just wish that repeat counts would repeat enclosed captures? Then you could use

m[
    ( ^ [^(:]+ )
    (?:
        (?: \( ( \d+ ) \) )?
        :
        (?: .*? \( ( [^)]+ ) \) ){1,}
        .*
    )?
    $
]x;
[download]

to grab repetative elements instead of having to do (or rather not do) things with built in limits, like this.

#! perl -slw
use strict;

while( <DATA> ) {
    chomp;
    my @bits = grep{ defined }
        m[
            ( ^ [^(:]+ )
            (?:
                (?: \( ( \d+ ) \) )?
                :
                (?:
                    .*?
                    \( ( [^)]+ ) \)
                    (?:
                        .*?
                        \( ( [^)]+ ) \)
                        (?:
                            .*?
                            \( ( [^)]+ ) \)
                            (?:
                                .*?
                                \( ( [^)]+ ) \)
                            )?
                        )?
                    )?
                )?
                .*
            )?
            $
        ]x;
    print join'/', @bits;
}

=Output
P:\test>junk
OFF
SUCCESS/abc
ERROR/1/27/11:03
WARNING/1
TEST/255/this/that/the other/and this
=cut

__DATA__
OFF
SUCCESS: (abc)
ERROR(1): disk number (27) crashed at (11:03)
WARNING(1): system is rebooting
TEST(255): (this) (that) (the other) (and this) (but not this)
[download]