Stringified objects are utf-8 encoded

[jixam]

Stringified objects are utf-8 encoded

[rurban]

Well, if you encode it twice (TO_JSON plus string overload). I'll check if the overloading callback can be omitted for the TO_JSON case, but stacking both is the really the users fault. And I don't know if skipping the overload callback is ok then.

[jixam]

The problem is with S which is not stacked. The TO_JSON works as expected, even with an overload present.

Here is a (currently failing) testcase that I managed to create:

--- t/54_stringify.t
+++ t/54_stringify.t
@@ -34,6 +34,12 @@ my $enc = $json->encode( { obj => $object } );
 is( $enc, '{"obj":"Foo <foo>"}', "mg object stringified" )
   or diag($enc);

+$object = bless ["\x{1f603}"], 'Foo';
+$enc = $json->encode( { obj => $object } );
+
+is( $enc, "{\"obj\":\"Foo <\x{1f603}>\"}", "mg object stringified Unicode" )
+  or diag($enc);
+
 $enc = $json->encode( { time => $time } );
 isa_ok($time, "Time::Piece");

(Edit: updated to the right version of the diff)

[rurban]

Ok, I'll check if I get my head around it. Double encoding doesn't look too good then

[rurban]

Fix coming today. Just need to fix 5.8 also

[jixam]

Thank you for the quick turnaround!