in reply to •Re: What is the big deal about PAR?
in thread What is the big deal about PAR?

It's the perfect replacement for commercial bundler solutions (like PerlApp and perl2exe).

I vaguely remember reading that Parrot would create executables. Is this true? Would Parrot make any features of PAR obsolete? Where's a Perl 6 internals architect when you need one?

  • Comment on Re: •Re: What is the big deal about PAR(rot)?

Replies are listed 'Best First'.
Re^3: What is the big deal about PAR(rot)?
by kelan (Deacon) on Mar 17, 2003 at 15:31 UTC

    Parrot will produce bytecode, which is really an intermediary form. It will not be directly executable, you will still need an installed version of the Parrot VM to execute it, much like you need a Java VM to execute Java bytecode. This is not to say that something with functionality like PAR will not emerge, but from what I've read so far, it doesn't look like Parrot will include that functionality itself.

    kelan

    Perl6 Grammar Student

[reply]
      While you'll need a parrot install as well as the bytecode file, that's still a darned sight better than what we have now with perl, where it's "Install perl. Install these X modules from CPAN, install this whole dir tree of stuff for my app."

      End deployment should be as simple as possible. A single file deployment is best, but barring that a persistent environment (like the JVM or parrot) that doesn't need to change with a single file for the app's not bad either.

      And I am taking a shot at being able to build a single-file combined interpreter and bytecode app, though it might be large... :)

[reply]
Re^3: What is the big deal about PAR(rot)?
by Elian (Parson) on Mar 17, 2003 at 15:47 UTC
    Parrot will potentially make all of PAR obsolete, but that's OK.

    If things work out right you'll be able to generate single-file standalone executables as well, but that's a separate thing.

[reply]
      Parrot will potentially make all of PAR obsolete

      Quick, someone tell Autrijus Tang!

      Seriously, I wonder of many active Perl projects Parrot will make less relevant. It would really stink to be working your heart out on something (like PAR) and then some other software comes out that makes all your hardwork go down the drain...

[reply]
        I don't think he'd mind. Could be wrong, in which case I'll talk with him this summer if he makes it to OSCON.

        The point is a dicey one. When moving to perl 6, do we leave alone those modules that exist for perl 5, or do we reimplement them because we need the functionality, and can potentially make it core stuff? (Real core stuff, not just shipped modules)

        This affects not only PAR, but libnet, DBI (though I hope to get Tim to do a straight port), CGI, LWP, the XML modules... While the perl5->perl6 translator or the perl 5 mode may well be just good enough, at least some of the modules will need work (like DBI, because of the XS code).

        I don't like the thought of stepping on anyone's toes, but I don't know that we should skip base functionality because someone's already done it for perl 5.

[reply]

In Section Meditations