> M. Nooning wrote: > >You don't need "use encoding ..." (or "use utf8") for that. > >It's only needed when your script itself uses UTF8 characters > >(say, in literal strings). > > > I had the same problem at one time. I was advised to put > use Encode::Unicode;
That was probably me :) > as the first line "use" in my program even though my program > does not use it. I did, and the problem went away. The reason is that deep down in the guts of Tk.pm (or some Tk/* module used by it) there's (on Win32 only) require Encoding::Unicode; except that it's not written that way, but more like my $foo = ..lengthy decision tree... require "Encoding::$foo"; which is not detectable by static analysis, i.e. Module::ScanDeps. But I thought it's rules had since been amended... Cheers, Roderich
