joshclark has asked for the wisdom of the Perl Monks concerning the following question:

Hello monks,

I'm writing a routine that needs to deal with bytes even when the string contains multibyte characters. "use bytes" of course does the trick but would break on pre-5.6 perl (alas, I need to be able to have this run on 5.004+). So, I need to ask the compiler to use the bytes pragma only if it exists...

I *think* this will do the trick but wanted to confirm with those who are wiser than I that I'm not missing anything important here:

BEGIN { eval {require bytes; import bytes; 1} }

Thanks for any advice!

Replies are listed 'Best First'.
Re: use bytes without breaking perl 5.005 or 5.004?
by lestrrat (Deacon) on Sep 26, 2003 at 17:12 UTC

    Matts showed me a clever way to do it the other day. I thought it was beautiful:

    $INC{ '' }++ if $] < 5.006

    Once that's in %INC, calls to "use bytes" do nothing, apparently


      I had previously convinced myself that because BEGIN requires a block, I couldn't fake out 'use warnings' for people who don't have it but then I realized this works fine:

      BEGIN { if( eval { require warnings } ) { warnings->import(); } else { $^W= 1; } }
      because the pragmas already expect to be inside of a (invisible) block and so apply themselves to one block further out.

      I like this method because it allows the user to install a back-ported or or whatever and use it if they want to and I don't have to trust my knowledge of when the module was added. (:

      Anyway, it is another way to do it.

                      - tye
        Thanks Tye! Applying this to "use bytes," am I correct that your solution is basically the same approach as the one I originally suggested? In other words, isn't this...
        BEGIN { if( eval { require bytes } ) { bytes->import(); } }
        ...pretty much the same as this:
        BEGIN { eval {require bytes; import bytes; 1} }
        I'm a bit new to trying to import modules and pragmas outside of the regular "use" function, so I'm just trying to be sure that I understand the difference, if any.

        Thanks! Josh
      Hey, interesting! So, the idea would be to do it like this?
      BEGIN { $INC{ '' }++ if $] < 5.006 use bytes; }
      This will cause 5.005 and below to ignore "use bytes" and cause 5.6+ to use the bytes pragma as usual? Nifty!

      Thanks for the help, Josh
Re: use bytes without breaking perl 5.005 or 5.004?
by bart (Canon) on Sep 27, 2003 at 00:57 UTC
    You can make a copy of the string that is marked as a byte string, even if the original string is UTF8, by doing this:
    $bytes = pack "C0a*", $bytes_or_utf8;
    That should work on pre-5.6 perls too — even though it won't do much.