Use concat overloading for "foo$_->$*"

This is the only discrepancy between $$_ and $_->$* that I know about.

To get ->@... interpolation to work, we have to emit a special
POSTJOIN token, which has just the right precedence to get it to apply
to the right amount of code before it, which perly.y then turns into a
regular join(...).

->$* and ->$#* were also going through that same code path, though it
turns out that simply omitting the POSTJOIN token for these dollar
tokens Just Works.  (I thought the fix would be more complicated.)

Now $_->$* within quotes becomes a direct argument to the concat ope-
rator, instead of being wrapped in a stringify(...) (what join(...)
optimises to with a single-item list).

2 files changed, 15 insertions, 2 deletions

diff --git a/t/op/postfixderef.t b/t/op/postfixderef.t
index 77988bfd3c..c3fa968d49 100644
--- a/t/op/postfixderef.t
+++ b/t/op/postfixderef.t
@@ -16,7 +16,7 @@ BEGIN {
 
 use strict qw(refs subs);
 
-plan(115);
+plan(116);
 
 {
     no strict 'refs';
@@ -352,4 +352,16 @@ is "$_->@{foo}", "foo->7 8 9", '->@{ does not interpolate without feature';
     is "@{[foo->@*]}", "7 8 9", '->@* inside "@{...}"';
     is "@{[foo->@[0,1]]}", "7 8", '->@[ inside "@{...}"';
     is "@{[foo->@{foo}]}", "oof", '->@{ inside "@{...}"';
+
+    # "foo $_->$*" should be equivalent to "foo $$_", which uses concat
+    # overloading
+    package o {
+	use overload fallback=>1,
+	    '""' => sub { $_[0][0] },
+	    '.'  => sub { bless [ "$_[$_[2]]"." plus "."$_[!$_[2]]" ] };
+    }
+    my $o = bless ["overload"], o::;
+    my $ref = \$o;
+    is "foo$ref->$*bar", "foo plus overload plus bar",
+       '"foo $s->$* bar" does concat overloading';
 }
diff --git a/toke.c b/toke.c
index b16544ba66..70f00f3406 100644
--- a/toke.c
+++ b/toke.c
@@ -1944,7 +1944,8 @@ S_postderef(pTHX_ int const funny, char const next)
 	if (PL_lex_state == LEX_INTERPNORMAL && !PL_lex_brackets) {
 	    assert('@' == funny || '$' == funny || DOLSHARP == funny);
 	    PL_lex_state = LEX_INTERPEND;
-	    force_next(POSTJOIN);
+	    if ('@' == funny)
+		force_next(POSTJOIN);
 	}
 	force_next(next);
 	PL_bufptr+=2;