Modern Perl for the Unfrozen Paleolithic Perl Programmer

Modern Perl for the Unfrozen
Paleolithic Perl Programmer
John SJ Anderson ⭐ Inﬁnity Interactive ⭐ @genehack
OpenWest 2016 ⭐ 14 July 2016 ⭐ Sandy UT

Disclaimer
I'mma talk pretty fast. Feel free to ask questions particularly if I'm moving too quickly

Hi Hi I'm John
• VP Tech, Inﬁnity Interactive
• Ex-biologist
• Linux
• Perl tribe
• Polyglot coder
• Just this guy, you know?
meet me!

I'M JUST A CAVEMAN
YOUR MODERN PERL FRIGHTENS AND
CONFUSES ME
I was inspired to write this talk by a few online friends who happen to be Perl programmers -- but they're not engaged with the community, they're just using Perl to get a job done, and there's a widening gulf
between what I think is reasonable and what they're doing -- so I wanted to put together sort of an info dump to bring people up to speed. this talk is going to a wide-ranging survey of what's happened in
Perl in the last 5-10 years.

Welcome to
2016!
So, hi, unfrozen Paleolithic Perl programmer! Welcome to 2016!

Quite a few things
have changed...
Many things have changed since you fell into that glacier in 2001,

There's a Clinton
running for President!

One thing is different:
Perl6 was released!
but this talk is about perl5, so that's the last i'm going to talk about perl6

We're still using
Perl5!
Most of us are still happily coding in Perl5, however

2001
Perl 5.6
When you slipped into that big hole in the ice, perl 5.6 was the latest and greatest

2016
Perl 5.24
But we're all the way up to 5.24 now. So let's talk about that 15 year gap...

We had some
bad years
there was a rough patch

5.6.0 - 22 Mar 2000
5.8.0 - 18 Jul 2002
5.10.0 - 18 Dec 2007
5.12.0 - 12 April 2010
here's the release dates for several perl versions, starting with 5.6 (latest and greatest in 2001, remember)

5.6.0 - 22 Mar 2000
5.8.0 - 18 Jul 2002
5.10.0 - 18 Dec 2007
5.12.0 - 12 April 2010
see that ﬁve year gap there? yeah, those were not great times.

5.6.0 - 22 Mar 2000
5.8.0 - 18 Jul 2002
5.10.0 - 18 Dec 2007
5.12.0 - 12 April 2010
but, starting with 5.10, which sort of brought perl 5 back to life, and then particularly, since 5.12, we've had this great thing! ...

Regular
release
cycle
a regular, predictable release cycle!

5.12.0 - 12 April 2010
5.14.0 - 14 May 2011
5.16.0 - 20 May 2012
5.18.0 - 18 May 2013
5.20.0 - 27 May 2014
5.22.0 - 01 Jun 2015
5.24.0 - 09 May 2016
here are all the releases since 5.12. notice anything?

5.12.0 - 12 April 2010
5.14.0 - 14 May 2011
5.16.0 - 20 May 2012
5.18.0 - 18 May 2013
5.20.0 - 27 May 2014
5.22.0 - 01 Jun 2015
5.24.0 - 09 May 2016
see those nice regular predictable releases, happening every year like clockwork? that's pretty cool.

underlying this is a regular series of dev releases (those are in odd numbered versions, e.g. 5.23 is the series of dev releases between 5.22 and 5.24), which happen on a
monthly cycle

Perl5
pumpking
credit for starting and more importantly continuing this goes to a couple of p5 pumpkings, which is what we call the person who is in charge of the Perl5 eﬀort -- our
Benevolent Dictator for the Moment, if you will

5.12.0 - 12 April 2010
5.14.0 - 14 May 2011
5.16.0 - 20 May 2012
5.18.0 - 18 May 2013
5.20.0 - 27 May 2014
5.22.0 - 01 Jun 2015
5.24.0 - 09 May 2016
Jesse
Rik
there were two pumpkings involved in this, Jesse Vincent for 5.12 and 5.14, and then Rik Signes for 5.16 thru 5.24 (which just came out a few months ago)

Regular
release
cycle
getting back onto a regular predictable release cycle has arguably been the single greatest factor keeping Perl5 relevant and catalyzing the Perl "Renaissance" you may
be hearing about

❤ Jesse
❤ Rik
much respect to Jesse and Rik for the work involved in this, it was *not* a trivial undertaking to start or to keep going.

❤ Sawyer
this is also a good time to mention that Rik retired from the pumpking role after releasing 5.24.0, and our new pumpking is Sawyer. Sawyer is great and as you can see
from this picture, he *really* loves hugs, so be sure to give him one when you see him at a conference.

Modern
Perl

modernperlbooks.com
Another catalyst for the Perl Renaissance is the book "Modern Perl", written by chromatic. He has made this freely available online; you can also purchase a paper copy
if that's how you roll.

❤
chromatic

modernperlbooks.com
Writing this book was a tremendous eﬀort and it's an absolutely essential read if you want to understand what we call "Modern Perl", or how we do Perl in the 21st
century.

Many
language
improvements
One of the main beneﬁts of the regular release cycle has been a number of language improvements. I'm only going to talk about a few today, but you can always read the
'perldeltas', or change summaries that come out with each new release.

Unicode
support!
In 2001, the world of Perl 5.6, Unicode existed but wasn't very well understood or very widely used. Today, of course, we live in a Unicode world, and it's something all
programmers and languages have to deal with. It's kinda too complicated to get into here, so I'm just going to say that now we can get _really_ expressive in our Perl
code

$code = "😀";
So it's much easier to tell if we're happy...

$code = "💩";
... or when things aren't going quite as well.

This should be a familiar way of communication for you!

(okay, this is my last caveman joke...)

Perl unicode support got _ok_ in 5.12, and got good in 5.14.

false values: 0, '', ( ), undef
everything else is true
undefined values: undef
everything else is defined
Quick Perl background
Before I can explain this next one, I need to do a brief refresher on a couple of aspects of Perl. Perl has the usual notion of true/false. These four values are false;
everything else is true. 'undef' is a special value; it's what Perl gives to a newly declared but uninitialized variable. That means there's this additional dimension of
defined/undefined we have to deal with

Boolean operators (or, and)
look at truthiness.
Many times, only invalid value for something is
undef. So can't look at truthiness; must look
at definedness
Quick Perl background
Now, Perl's Boolean operations *only* look at true/false. They don't care about defined/undefined at all, except for how undef is false.

But many times, you don't end up caring about true/false, you care about def/undef

if ( defined($this) ) {
$value = $this
}
else {
$value = $that
}
So that leads to writing code that looks like this... *explain code*

$value = defined($this) ? $this : $that
The more idiomatic way of writing that code is to use the ternary operator -- this just does the same thing as the previous slide in a more compact way

$value = $this // $that
defined-or
Now that I've given you that background, now I can explain this great new feature we got in 5.10, called 'defined-or'. It let's you write that code like so.

Which is pretty awesome. Defined-or was my favorite new perl feature of the 2000s

$this //= 'default'
# much like this, but…
$this ||= 'default'
deﬁned-or
You can also combine this with assignment, in much the same way you can say ||= for 'or equals'

my $copy = $orig;
$copy =~ s/swap/stuff/;
Regular expressions are a big part of Perl programming, including substitution style regexps. If you want to apply a substitution to a variable but _not_ change your
original value, you need to do something like this - make a copy, then do the sub on the copy

(my $copy = $orig) =~ s/swap/stuff/;
and this is just a more idiomatic way to do the same thing.

this makes the copy (and the parens are *required*)...

and then this applies the sub to the result of the left hand side, which means $copy gets changed and $orig is left unchanged

my $copy = $orig =~ s/swap/stuff/r;
non-destructive s///
new in 5.14, we have a new way do this. don't need the parens.

the key here is ...

this 'r' modiﬁer to the regexp, which makes

this part return the result of applying the sub to $orig (hence, 'r' - for return)

and then we just assign that to $copy.

Again, new as of 5.15

dereferencing
@{ $arrayref }
so, if you've ever worked with references in perl, you've probably done this. a reference is how perl does data abstraction, it's kinda like a pointer (only way safer). this is
how you turn a reference to an array back into the actual array

postfix
dereferencing
$arrayref->@*
New syntax for this -- not really *less* punctuation, but different punctuation. I'm not a huge fan, but some people (including former pumpking Rik Signes) really like this.

We added postfix deref in 5.20; it became non-experimental in 5.24

double
diamond
<<>>
We also got something called the 'double diamond' operator in 5.22

diamond
<>
This is related to the classic 'diamond' operator, which is used when processing ﬁles that are passed as arguments on the command line

# loop over files from command line
while (<>) {
# do something with each line
}
# call like 'script.pl file1 file2 file3'
using the diamond operator typically looks like this, you'll have a script with this while loop, and then call it with a bunch of files on the command line. this will loop over
the lines in file1, then file2, then file3, running the loop body for each one

problem!
<> uses 2-arg
open semantics!
the problem with this is <> uses 2 argument open(). why is that bad?

2-arg open
open(IN, '< file');
this is what we call a 2 arg open call. we're opening a ﬁle, for reading, and attaching it to a ﬁle handle called IN

2-arg open
open(IN, '< file');
this thing right here, in front of the ﬁle name, is how you want to open the ﬁle. this is open for reading.

2-arg open
open(IN, '> file');
where as this is open for writing. (and if you wrote this code and called the ﬁle handle IN, we can't be friends.)

2 arg is bad, because if 'ﬁle' is actually in a variable, and somehow that variable gets a '>' or '<' or '|' in it, unexpected things happen

3-arg open
open($in, '<', 'file');
this is how we write that code these days, the three argument form. here the ﬁle name is just used literally, and if it gets one of those punctuation chars in it, you'll get an
error rather than a bad surprise

problem!
<> uses 2-arg
open semantics!
so, the problem is <> uses the 2-arg form...

Means that this:
script.pl '> file1'
is gonna be a bad time
So doing something like this will open 'ﬁle1' for writing, clobbering whatever is there.

<<>> ﬁxes this.
again, available as of 5.22

subroutine
signatures
Finally, and I think this is maybe _the_ biggest feature this decade, we've also recently gotten support for subroutine signatures in 5.20

sub add {
my ($one, $two) = @_;
return $one + $two;
}
this is how you might write a simple perl function to add together two numbers. the @_ is a special variable that's used to pass in the arguments for the function.

sub add {
my ($one, $two) = @_;
return $one + $two;
}
This bit in the ﬁrst line is called "unpacking @_", and code like this is at the top of literally almost every Perl function *EVER*

sub add ($one, $two) {
return $one + $two;
}
this code does the exact same thing!

sub add ($one, $two) {
return $one + $two;
}
so we don't even have to unpack @_ ourselves! you can even provide default values, validate arguments in the function signature, and all sorts of crazy stuﬀ

Who has started using subroutine signatures? They're still experimental, but hopefully will be non-experimental in 5.26

new tools
It's not just language level features, we also have a host of new tools to make it easier to work with Perl

"system" perl:
These days, we make a distinction between the Perl that comes with your OS

"system" perl:
just say no
and encourage people not to use it for their development projects

/usr/bin/perl
Yes, that means good old /usr/bin/perl...

/usr/bin/perl
is no more. Instead, we use tools to install new Perls, outside the control of the OS

perlbrew
The ﬁrst tool that was developed for this purpose is called Perlbrew

perlbrew.pl
which you can get at perlbrew.pl

plenv
there's also a newer option, plenv

https://github.com/tokuhirom/plenv
its website isn't quite as snazzy as perlbrew's

advantages
• Solves "vendor Perl lockin" issues
• Install multiple Perls in your home directory
• ... or elsewhere
• Trivially switch from Perl version to Perl version
• Able to install modules without special permissions
• Easy to stay up to date with Perl development
either one of these tools will give you numerous advantages over using the system perl

plenv > perlbrew
• Less magic messing around with $PATH
• Can "pin" Perl different ways: globally, per-shell, or per-directory
personally, i'm using plenv these days -- the ability to easily specify a Perl version for a particular project is particularly useful to me

perlbrew > plenv
• Kicka** website

speaking of installing...
we also have a few new tools that make installing and managing modules easier

local::lib
one of them is a module called local::lib

local::lib
• Install your own copies of modules
• In your $HOME (so no special permissions needed)
• Can also install per-project modules
• Integrates well with other tools
• Not going to go into how to make it work, but the docs are great!

❤ mst
originally developed by matt trout, so thanks matt

speaking of installation tools…

cpanminus
a/k/a
cpanm
We've got a new CPAN client these days, cpanminus

% cpan Git::Wrapper
CPAN: Storable loaded ok (v2.54)
Reading '/Users/genehack/.cpan/Metadata'
Database was generated on Sat, 10 Oct 2015 01:17:02 GMT
Running install for module 'Git::Wrapper'
CPAN: LWP::UserAgent loaded ok (v6.13)
Fetching with LWP:
http://cpan.schatt.com/authors/id/G/GE/GENEHACK/Git-Wrapper-0.045.tar.gz
CPAN: YAML loaded ok (v1.15)
CPAN: Digest::SHA loaded ok (v5.95)
Fetching with LWP:
http://cpan.schatt.com/authors/id/G/GE/GENEHACK/CHECKSUMS
CPAN: Compress::Zlib loaded ok (v2.068)
Checksum for /Users/genehack/.cpan/sources/authors/id/G/GE/GENEHACK/Git-Wrapper-0.045.tar.gz ok
tmp-47326 for tmp-47326: No such file or directory at /opt/plenv/versions/5.23.2/lib/perl5/5.23.2/CPAN/Distribution
.pm line 468.
CPAN: File::Temp loaded ok (v0.2304)
CPAN: CPAN::Meta::Requirements loaded ok (v2.133)
CPAN: Parse::CPAN::Meta loaded ok (v1.4417)
CPAN: CPAN::Meta loaded ok (v2.150005)
CPAN: Module::CoreList loaded ok (v5.20150820)
Configuring G/GE/GENEHACK/Git-Wrapper-0.045.tar.gz with Makefile.PL
Locating bin:git... found at /opt/git/bin/git.
Checking if your kit is complete...
Looks good
Generating a Unix-style Makefile
Writing Makefile for Git::Wrapper
Writing MYMETA.yml and MYMETA.json
this is the output from using the default 'cpan' client to install something.

GENEHACK/Git-Wrapper-0.045.tar.gz [17/1516]
/opt/plenv/versions/5.23.2/bin/perl5.23.2 Makefile.PL -- OK
Running make for G/GE/GENEHACK/Git-Wrapper-0.045.tar.gz
cp lib/Git/Wrapper/File/RawModification.pm blib/lib/Git/Wrapper/File/RawModification.pm
cp lib/Git/Wrapper.pm blib/lib/Git/Wrapper.pm
cp lib/Git/Wrapper/Statuses.pm blib/lib/Git/Wrapper/Statuses.pm
cp lib/Git/Wrapper/Exception.pm blib/lib/Git/Wrapper/Exception.pm
cp lib/Git/Wrapper/Log.pm blib/lib/Git/Wrapper/Log.pm
cp lib/Git/Wrapper/Status.pm blib/lib/Git/Wrapper/Status.pm
Manifying 6 pod documents
GENEHACK/Git-Wrapper-0.045.tar.gz
/usr/bin/make -- OK
Running make test
PERL_DL_NONLAZY=1 "/opt/plenv/versions/5.23.2/bin/perl5.23.2" "-MExtUtils::Command::MM" "-MTest::Harness" "-e" "und
ef *Test::Harness::Switches; test_harness(0, 'blib/lib', 'blib/arch')" t/*.t
t/00-load.t ............... 1/6 # Testing Git::Wrapper 0.045
t/00-load.t ............... ok
t/author-err.t ............ skipped: these tests are for testing by the author
t/basic.t ................. # Testing git version: 2.5.2
t/basic.t ................. ok
t/git_binary.t ............ ok
t/parse_args.t ............ ok
t/path_class.t ............ # Testing git version: 2.5.2
t/path_class.t ............ ok
t/release-pod-coverage.t .. skipped: these tests are for release candidate testing
t/release-pod-syntax.t .... skipped: these tests are for release candidate testing
All tests successful.
Files=8, Tests=67, 1 wallclock secs ( 0.04 usr 0.02 sys + 0.39 cusr 0.31 csys = 0.76 CPU)
Result: PASS
GENEHACK/Git-Wrapper-0.045.tar.gz
and this is more of the output...

(still not done
but I got tired
of pasting.)

% cpanm Git::Wrapper
--> Working on Git::Wrapper
Fetching http://www.cpan.org/authors/id/G/GE/GENEHACK/Git-Wrapper-0.045.tar.gz ... OK
Configuring Git-Wrapper-0.045 ... OK
Building and testing Git-Wrapper-0.045 ... OK
Successfully installed Git-Wrapper-0.045
this is the output from cpanm installing the same thing

carton
Another tool, carton, helps manage your module dependencies in a project. Particularly useful for large projects with multiple devs

Like bundler
but for Perl
it's like ruby's bundler, but for perl modules

Freeze deps,so you
always install same
versions across
dev/staging/prod

❤ Miyagawa
both cpanm and carton were developed by Miyagawa; many thanks to him!

speaking of stuﬀ on CPAN…
so, CPAN...

metacpan.org
we have a whole new website for interacting with CPAN

search.cpan.org
search.cpan.org is still around...

but metacpan integrates and visualizes a bunch of information in a really useful way

things like a syntax-highlighted source view, linking to home pages and code repos, showing test results, and the amount of activity in a project

https://github.com/CPAN-API/metacpan-web
it's also open source, so if you can think of a way to make it better, you can

❤ MetaCPAN team
thanks to the whole metacpan team for all their hard work

Duck Duck Go
we also have a new search engine here in 2016. it's cool, and it's partially written in Perl

Duck Duck Go
!cpanm
The most useful feature, though, is ability to use 'bang searches' to restrict your search to a particular site - this is how you search metacpan

speaking of modules...
if you haven't been playing close attention to perl community goings on, there are a few new modules you may have missed

JSON::MaybeXS
anybody doing web development these days needs to interact with JSON - using JSON::MaybeXS will make sure that you have a JSON library available, picking the best
one from a number of alternatives

Cpanel::JSON::XS
JSON::XS
JSON::PP
here's how the fallback works -- and if you install JSON::MaybeXS and don't have any of these installed, it will require Cpanel::JSON::XS (as long as you have a compiler
to build the XS)

Moose
Moo
To get a handle on how we do OOP in Perl these days, you should look into Moose -- and then when you're ready to write some code, you'll probably be able to get
away with dropping down to Moo

CGI.pm
is
gone
I do have some bad news for you -- CGI.pm has been pulled out of core

(not really)
don't worry, you can still ﬁnd it on CPAN

Plack
But the current standard for web development in Perl is Plack/PSGI. Oﬀers a number of advantages over CGI, and is the basis for all modern Perl web frameworks

speaking of Perl websites…
we have quite a few new websites these days, which make it easier to keep up with the current state of things

http://cpanratings.perl.org/
one of the problems with cpan is there's just _so_ _much_ _stuﬀ_ there. it can be hard to decide which one of a dozen diﬀerent modules to use. cpanratings helps with
this problem

http://cpanratings.perl.org/
you can see individual reviews, which version they're reviewing, and so on. one downside - because ratings are about a particular version, may not reﬂect current module
state

https://metacpan.org/
MetaCPAN links to reviews as well as showing an average review score

http://cpants.cpanauthors.org/
we also have CPANTS - automated testing of some best practices around module development

http://cpants.cpanauthors.org/
here's what that looks like for a particular module. super useful if you're not sure you're doing things the "right" way

http://cpants.cpanauthors.org/author/GENEHACK
can also see modules by author, which can be useful when trying to decide whether to use somebody's code

https://metacpan.org/
MetaCPAN links to CPANTS too

http://prepan.org/
We also have PrePAN, which is a place to get feedback on module ideas you haven't even written yet

Perl 5 Porters mailing list summaries!
In a recent development, sawyer has revived the p5p weekly email summary -- excellent if you want to keep up with what's going on with perl5 development but don't
have time to follow the email list yourself

http://blogs.perl.org/users/sawyer_x/
and in a recent development, sawyer has revived the p5p weekly email summary -- excellent if you want to keep up with what's going on with perl5 development but
don't have time to follow the email list yourself

speaking of Perl community…
ﬁnally, you really should consider becoming more actively involved with the perl community, if you're working with (or even just playing with) perl

irc.perl.org
irc.perl.org had a *bit* of a bad reputation </understatement> but things have been cleaned up a bunch. there's a documented standards of conduct, abuse is not
tolerated, and it's a good place to get help from experienced perl folk

Conferences
& Workshops

& Mongers

(oh my.)
big conferences: YAPC, once a year (or so), in North America, Europe, Asia. Sometimes Brazil, Russia too

workshops: smaller, regional. Pittsburgh, DC-Baltimore, Orlando

perl mongers groups: like MasterCard, we *errywhere* you want to be

http://www.pm.org/groups/north_america.html
ﬁnally, you really should consider becoming more actively involved with the perl community, if you're working with (or even just playing with) perl, there's probably a local
perl mongers group near you

http://saltlake.pm.org/
here's the SLC one!

http://perlweekly.com/
ﬁnally, let me give a big shout out to perl weekly, which is a once a week email newsletter and website aggregating perl related news from all over the web

if you're only going to pay attention to one perl thing, perl weekly is your best choice! sign up at perlweekly.com

thanks!
thanks to the organizers for accepting my talk, thanks to all of you for attending and participating ...

thanks to my employer for giving me the time to write this talk and sending me here to deliver it

questions?
https://joind.in/event/openwest-2016/modern-perl-for-the-unfrozen-paleolithic-perl-programmer
http://tinyurl.com/unfrozen-perl
any questions? (mention feedback link)

Modern Perl for the Unfrozen Paleolithic Perl Programmer

More Related Content

What's hot (20)

Viewers also liked (6)

Similar to Modern Perl for the Unfrozen Paleolithic Perl Programmer (20)

More from John Anderson (20)

Recently uploaded (20)

Modern Perl for the Unfrozen Paleolithic Perl Programmer