A Short History of Perl Development

Perl has gone through several eras of development, each with its own story. Knowing what’s come before can help you appreciate where Perl is now.

In the late 1980s, Larry Wall created the Perl language (although that was not the first name he tried to use). He mostly worked on his own with some feedback from the Usenet community. Kids today probably have never seen a newsgroup, but it was the social media of the time, and that’s where he first released Perl in 1987.

Eventually Perl was interesting enough to have books about it (this book in particular, and Programming perl, with pink covers before our publisher moved Perl books to blue covers). Perl was promoted to version 4. That’s also around the time Perl’s popularity exploded, and when many people learned (or stopped learning) Perl. Quite frankly, it’s this era that set most of the world’s expectations of Perl, and more disappointedly, Perl programmers. But that’s spilt milk.

But Perl 4 didn’t have object-oriented features, good ways to make complex data structures, or lexical scoping. Around 1993, Larry started working on Perl 5, the current major version and the one we write about in this book.

To move from Perl 4 to Perl 5, a gang of Perlers created the Perl 5 Porters to ensure that Perl 5 was ported to hundreds of different platforms. Today the group is still there, although the people have changed. You can read more about their process in perlpolicy.

The perlhist documentation lists each Perl release, including its date and maintainer. After Larry released Perl 5.0 in 1994, other people took responsibility for some releases. After a new version release, other people typically came in to maintain the old version. It was a bit haphazard and ad hoc, but it worked for a while.

The Porters made big changes in Perl 5.6, and again in Perl 5.8. Perl was going through some growing pains, including the switch to handling Unicode. From Perl 5.004 to 5.005 was a little over a year, but from 5.005 to 5.6 was almost two years. The lag time from Perl 5.6 to 5.8 was over two years. However, the time between releases was increasing.

After Perl 5.8, people knew the code needed drastic changes for continued development. Chip Salzenberg tried rewriting Perl in C++ in a secret project he had dubbed “Topaz.” It didn’t work out, although he learned interesting lessons in the process. Around the same time, Larry and a few others had the idea to start Perl 6, a complete rewrite of the code base to allow easier development and modern features.

Installing a Recent Perl

Before you think about installing a new Perl, check if the one that you have is good enough. The -v command-line switch tells you which version you have:

$ perl -v

This is perl 5, version 34, subversion 0 (v5.34.0)

If you have a recent enough version, you don’t have to do any more work. What you do next depends on how much work you’d like to do.

If you are stuck on a system without a compiler, you can try precompiled versions, including Strawberry Perl (Windows) or ActivePerl Community Edition (macOS, Windows, Linux, and others).

You can compile your own perl. We think that everyone should try compiling it themselves at least once in their lives. Part of being a programmer is understanding how actual computers work, and compiling source, managing libraries, and such are part of that. You can download the perl source from CPAN. We tend to have them all installed so we can play with any version we like.

Once you unpack the source, you can configure the installation to tell it where you want to install it. You don’t need special privileges to do this since you can install it into any directory you control:

$ ./Configure -des -Dprefix=/path/where/you/want/perl

We like to install several perls, so we create version-specific directories for each one:

$ ./Configure -des -Dprefix=/usr/local/perls/perl-5.34.0

From there, tell make to install it—this might take a bit:

$ make install

You might want to test the result before you install it. If the test step fails, make will not run the install step:

$ make test install

Once installed, we can use the new perl by specifying it in the shebang line:

#!/usr/local/perls/perl-5.34.0/bin/perl

You can also use the perlbrew application to install and manage several Perls. It’s doing the same thing you did in the previous step, but automated. See http://perlbrew.pl for details.

Experimental Features

Let’s get down to the actual features and how you use them. We aren’t going to list every feature or fully explain the features we highlight. We want to show you how to use any new feature, not particular new features.

You can enable experimental features in several ways. The first is the -E command-line switch, introduced in v5.10. Like the -e switch, it specifies the program as an argument, but -E also enables all new features:

$ perl -E "say q(Hello World)"

Inside a program, you can enable new features with use and the version number in any format:

use v5.34;
use 5.34.0;
use 5.034;

Remember that since v5.12, specifying the version with use also implicitly turns on strict.

You can also specify the minimum version without loading new features; do it with require:

require v5.34;

The feature module allows you to load features when you want them. In the use example, Perl was doing this for you by implicitly calling it to load the tag associated with that version:

use feature qw(:5.10);

Instead of loading all the new features for a particular version, you can load them individually. In Chapter 4 we showed you state (a stable feature that showed up in v5.10) and signatures (an experimental feature introduced in v5.20):

use feature qw(state signatures);

If you want to turn off all new features, perhaps because you have an old script that doesn’t work with the newer Perls, you can disable them:

no feature qw(:all);

Of course, you need a version of Perl that has the feature module for this to work. That probably means that you are running an old program on a new perl, or perhaps you’re still getting used to the new features and don’t want to accidentally use them.

use v5.20;
use feature qw(signatures);

use v5.20;
use feature qw(signatures);

sub division ( $m, $n ) {
  eval { $m / $n }
}

The signatures feature is experimental at features.pl line 4.

no warnings qw(experimental::signatures);

no warnings qw(experimental);

use experimental qw(current_sub);

require v5.10;
sub say {
  print "Someone said "@_"
";
}

say( "Hello Fred!" );

{ # use the built-in say in here
use feature qw(say);
say "Hello Barney!";
}

say( "Hello Dino!" );

Someone said "Hello Fred!"
Hello Barney!
Someone said "Hello Dino!"