Vadim O. Ustiansky <[EMAIL PROTECTED]> writes: > >So my question is whether it is possible to completely hide >C internals from perl level.
More or less. If you associate the C pointer with the object by using MAGIC then it is very hard for perl code to get at it. (The B:: stuff can probably still mess with it...) If you then implement C struct accessors as methods you have my $get = $obj->key; $obj->key($set); You need a custom typemap for this (obviously). Tk and DBI both use this scheme. See PERL_MAGIC_ext in perlguts.pod
