More Boom In The Room

Batch-checking that a set of Perl modules compile

2008-07-10T11:44:00.000-04:00

cd module-dir; find . -name '*.pm' -exec perl -c {} \; 2>&1 | fgrep -v 'syntax OK'

HOWTO fix Perl IO::Uncompress::RawInflate errors in CPAN shell

2008-06-16T17:57:00.005-04:00

I've run into this problem repeatedly on my work PC. Unfortunately the solutions I've seen seem to be specific to Linux distributions, and I'm using Cygwin on Windows XP. So here's my (hopefully cross-platform) solution: manually download the IO::Compress::Zlib tarball from CPAN, then build and install it manually as per the README file. The CPAN shell should now be able to unpack things again.

Under-appreciated Perl functions: map, grep and exists

2008-04-11T15:55:00.002-04:00

I've recently taken more direct ownership of a big Perl application at work, and in the process of adding a new feature noticed a lot of this type of looping code:

# loop 1
my %foo_types = ();
foreach my $entry_data_ref (@entry_data) {
    $foo_types{ $entry_data_ref->[19] } = $entry_data_ref->[20];
}
my $foo_types_regex = '(' . join('|', keys %foo_types) . ')';

# loop 2
foreach my $data_ref (@data) {
    my $id = $data_ref->[0];
    push(@ids_to_disable, $id) if $type_cd =~ m/^$foo_types_regex$/;
    my $type_cd = $data_ref->[21];
}

I managed to simplify it down to this:

# loop 1
my %foo_types = map { $_->[19] => $_->[20] } @entry_data;

# loop 2
my @ids_to_disable = map $_->[0], grep exists $foo_types{ $_->[21] }, @data;

See the difference? Both loops are concerned with generating one list from another -- the first transforms a list into a hash, the second uses the keys of that hash to determine the contents of another list. This is what map (and its cousin, grep) are built for. The foreach loop in the original hides the intent of the code, and it's likely slower as well since map knows how many elements it needs to process and can preallocate space on the list it returns.

The original loop 1 also populated a string with hash keys to use as a regular expression later in order to find out what entries to get out of @data in loop 2. But again, we have a much more appropriate function for this: exists. Checking for a hash hit is near-instantaneous compared to working through a long regular expression, and again it makes the code smaller and easier to understand and debug.

map and grep come from the more functional-programming side of Perl, and they seem to be very underused and under-appreciated by people used to more traditional procedural languages like C and Java. It's unfortunate, because it leads to much more bloated code that takes longer to debug, and doesn't use the language's full capabilities.

I'm the new Perl XML::Parser MacPorts maintainer

2007-12-13T10:59:00.000-05:00

Not really a big deal, but I like MacPorts and I use this module a lot. Mainly this will be a matter of monitoring CPAN for new releases.

Visual Basic for the Web

2006-11-02T13:49:00.001-05:00

According to Tim Bray, Andi Gutmans said at the recent Zend/PHP Conference that "the right way to think about PHP is as Visual Basic for the Web."

My new job is at an avowed "PHP shop", even though the position was advertised on a Perl-oriented job board. A lot of the technical portion of my interview was spent pointing out that I had skills in both -- one of my future co-workers even admitted to me that he was biased towards me as a "Perl guy" instead of just as a programmer.

Coverity Study Ranks LAMP Code Quality

2006-09-18T08:39:00.000-04:00

Interesting excerpt to me, since Perl is my weapon of choice:

"Chelf was surprised, however, by the performance of the Perl language.
"Of the LAMP stack, Perl had the best defect density well passed standard deviation and better than the average, Chelf said.
Perl had a defect density of only 0.186. In comparison Python had a defect density of 0.372 and PHP was actually above both the baseline and LAMP averages at 0.474." ("Coverity Study Ranks LAMP Code Quality", Sean Michael Kerner, internetnews.com)

Tags:

$Philadelphia->isa("YAPC::NA") ?

2006-08-02T17:17:00.000-04:00

Geoff Avery of phl.pm (the Philadelphia Perl mongers) has submitted our fair city to be the site for next year's Yet Another Perl Conference. We should find out by September 1 whether Philly (OK, Camden) made it.

I was at the first Perl Conference (now OSCON) in San Jose, and haven't been to many other professional conferences after that. It'd be nice to have one so close.

Tags:

The other "P"s in OPAL

2006-06-27T11:02:00.002-04:00

So Oracle has recently been embracing the non-database parts of the LAMP web development stack and given it the name OPAL. They've got a PHP Developer Center, a n integrated Zend PHP engine, even a blog and discussion forum on PHP.

So where's the love for the other, more general-purpose languages that are part of the web development stack? Why is DBD::Oracle so freakin' hard to set up? Why has DCOracle2 languished umaintained for four years?

I have nothing against PHP, and have even done some work in it (including this site). But Perl is my weapon of choice, and I'm sure other developers have theirs. Oracle would do well not to alienate us by making us struggle with their software. They're losing market share to MySQL when it comes to web and rapid application development -- my bread-and-butter.

I like Oracle in most other respects. Their Express Edition database is a godsend for doing quickie work in preparation for moving it to enterprise-sized big iron. Application Express is drop-dead simple for the majority of CRUD-like web apps. I just wish they would go all in and show better support for us non-PHP developers out there.

Back in the saddle

2006-06-14T14:18:00.000-04:00

For the past week I've been working on a project that finally gets me back into real web development (as opposed to the faux-web dev I've done for the past several months with Oracle APEX and doctoring up a hairball Windows Perl script.) New weapons in my arsenal this time around are CGI::Application, DBIx::Class, and Template Toolkit, which together give me a fairly good set of loosely-coupled MVC-oriented tools.

I tried Jifty, but the lack of Oracle support was a deal-breaker for me. (Sent in a bug fix to help that effort along, though.) Also tried Catalyst, but couldn't get my head wrapped around it. Maybe I'm just too stupid for these full-stack kits, and will need a couple go-arounds with the more lightweight CGI::App before making another attempt.

The smallness of CGI::App is definitely a draw, though. The full-stack MVC toolkits impose a pretty strict application structure -- so much so that they use helper scripts to set it up for you. (Ruby on Rails obviously popularized this trend.) CGI::App lets me put together a quickie module hidden behind a small instance script, and then grow it as needed. Heck, I could see myself rewriting this site (dude, the band's been dead for two years, move on) with a thin CGI::App shell for the blog, and then adding bits as the mood (and free time) strikes.

So, yeah. After another several month drought, this is my second post in as many days, and it's another geekfest. Oh well.

Yay, my first Perl Monks post

2006-03-27T18:24:00.000-05:00

Ever wanted a quickie example on Microsoft WMI and Win32::OLE? Yeah, me neither. :-/

CGI dump script

2006-03-09T15:27:00.000-05:00

For my own reference, a Perl CGI script to dump out received parameters, taint mode status, and Perl's internal configuration. In order to run taint scripts on Microsoft IIS (version 5.1 on Windows XP Pro SP2) with ActivePerl I had to create a mapping (Default Web Site Properties → Home Directory → Configuration... → Mappings) for files with a .tpl extension to execute with C:\Perl\bin\perl.exe -T "%s" %s; otherwise I'd get an error about "-T" is on the #! line, it must also be used on the command line.

Here's the script:

#!/usr/bin/perl -wT

use strict;
use warnings;

use CGI::Pretty qw(:standard fatalsToBrowser);
use Config;

print header, start_html('CGI test page');

print p('Taint mode ON!') if is_tainted($ENV{PATH});

print h1('Received CGI parameters'), Dump if param;

print h1('Environment');
my $env_list;
foreach (keys %ENV) { $env_list .= dt($_) . dd($ENV{$_}) }
print dl($env_list);

print h1('Perl configuration');
print pre(Config::myconfig);

print end_html;

sub is_tainted {
 return ! eval { eval("#" . substr(join("", @_), 0, 0)); 1 };
}