I recently needed to authenticate to a remote API using an SSL client certificate, and had a bit of trouble getting LWP::UserAgent to work with it.

The examples I found which looked like they should work involved e.g.:

use LWP::UserAgent;

my $ua = LWP::UserAgent->new(
    ssl_opts => {
        SSL_use_cert => 1,
        SSL_cert_file   => "/path/to/clientcert.crt",
        SSL_key_file    => "/path/to/privatekey.key",
    },
);

However, that didn’t work; changing the paths to the cert/key to non-existent files didn’t cause any difference, so I suspected that those options were actually being ignored.

After a fair bit of digging, the option I found that actually worked was loading Net::SSL first, to make LWP use Net::SSLeay, and setting env vars to the client cert to use:

use Net::SSL;
use LWP::UserAgent;

$ENV{HTTPS_CERT_FILE} = "/path/to/clientcert.crt";
$ENV{HTTPS_KEY_FILE}  = "/path/to/privatekey.key";
my $ua = LWP::UserAgent->new();

This, to me, is pretty icky – I’d much rather pass config to affect just that single LWP object. However, it gets it working.

Yesterday, I attended the 2011 London Perl Workshop – my first ever Perl conference.

I had a good day, met a few members of the Perl community I knew from online interactions who I’d never met in person before, saw some good talks, and partook in some free food and beer (kindly paid for by the sponsors, including my employer, UK2).

Some brief mentions of talks I attended:

Matt S Trout (mst) – First, Tak wrote the world‎

I have IO::Pipeline. I have App::FatPacker. I have IPC::Command::Multiplex. And yet I still couldn’t whip up a five line example of bolting them all together that made a compelling argument for a perl-loving sysadmin to stop using fabric.

This problem, among others, will be solved by the conclusion of this talk.

Tak sounds like something which will be very useful to me – running code on multiple other hosts via SSH, but including Perl code – with all locally-installed modules available for use at the remote end!

As mst went through explaining how it all worked, my thoughts went from “hmm, useful”, to “hmm, useful but looks over-engineered, not sure it needs to be that complex” to “whoah, that’s genius”. Fatpacking and sending code to the remote side, which then adds a coderef to @INC which requests other modules from the local end, sent over and loaded remotely, is awesomely creative.

These kind of tricks remind me of why I love Perl.

Mike Whitaker (‎Penfold‎) – ‎Perl and Unicode, the 5.14 edition‎

A very good talk on handling Unicode safely in Perl, and the gotchas to avoid. Provided major impetus for me to upgrade to 5.14, too.

Zefram – ‎why time is difficult‎

Dates, times, time intervals, clocks, calendars, and related phenomena are major contributors to hassle in programming, and the source of innumerable bugs.

Zefram’s talk, whilst barely Perl related, was very interesting, and very well delivered. I hadn’t realised quite how complex time was :)

Zefram’s amusing lightning talk on doing away with source code by simply storing bytecode and editing it by deparsing the source, editing it, then “compiling” back to bytecode was also entertaining.

Claes Jakobsson (‎claes‎) – ‎Don’t debug now, debug later

Runops::Recorder is a alternate runloop for perl that writes down what your program does to disk for playback later

It also comes with a viewer and some helper classes for you to write your own playback tools such as diffs etc.

This looks like a very useful debugging tool, recording the path of execution through your code and writing it to a file which can then later be “replayed” using a viewer – much like single-stepping or tracing through the debugger, but after the fact. The ability to leave it running and have it dump out a configurable amount of trace data when a die is encountered looks excellently useful for catching intermittent / rare problems – you should be able to leave it in place, wait until the problem occurs, then replay what happened leading up to the die to see what was going on.

Future versions should also be able to track changes to variables, etc, which will be very useful indeed.

There were a couple of workshops I’d like to have attended, but which I didn’t; partly because they conflicted with talks I wanted to see, and partly because I didn’t have a laptop with me to “work along” and didn’t think I could take much of value away from them.

Andrew Solomon – ‎[[TRAINING SESSION]] Web development for beginners using Dancer‎

As a core developer for the Dancer perl web framework I’d love to have attended Andrew Solomon’s workshop, to see what was being taught, and offer any input desired. Unfortunately, I wasn’t there, but I’ll be looking with interest for any feedback from people who were, and what they learned and what they thought of Dancer if they hadn’t encountered it before. Making Perl accessible for new users is an important thing.

Gabor Szabo (‎szabgab‎) – ‎[[TRAINING SESSION]] Testing in Perl

I’d also like to have taken part in Gabor’s workshops, but they were in two parts and conflicted with several other talks I wanted to see.

I met a few members of the Perl community who I knew from online interactions but had never met in meatspace, so it was great to meet them. Unfortunately, there were a few others I’d meant to go introduce myself to, but never got a chance to do so – including Tatsuhiko Miyagawa and Gabor Szgabo.

Overall, it was a good day, and I imagine there’s a very good chance I’ll be back next year :)

I recently wanted to produce some graphs from a web app powered by the Dancer Perl web framework, and reevaluated the various Perl graphing moduiles out there.

Modules I considered were:

• Chart::Strip
• Chart::Graph
• Google::Chart
• Chart::Clicker
• Chart::Gnuplot

Unfortunately, I didn’t have time to do a full in-depth writeup trying every module like the excellent ones Neil Bowers has been doing, but I thought I’d write up a quick post on the choice I made, with example code, in case it helps other people looking to graph potentially irregularly-spaced time-based data samples in Perl easily.

Chart::Clicker looked to be a nice choice (with a nice example of doing just what I want given as the topic answer to a question on StackOverflow), but had a huge chain of dependencies, finally failing when demanding Cairo and various X11 libraries (on my headless server).

Chart::Strip seemed to do exactly what I wanted in a simple way, but I encountered a div-by-zero bug when dealing with a certain dataset with > 89 data points.

I reported this to the author, Jeff Weisgberg in RT #72288, and he promptly released 1.08 with a fix (thanks Jeff!).

Chart::Strip made it simple to do what I wanted:

my @dataset;
while (my $row = $sth->fetchrow_hashref) {
    push @dataset, { time => $row->{timestamp}, value => $row->{value} };
}

my $chart = Chart::Strip->new( title => "My chart" );
$chart->add_data(\@dataset, { style => 'line' });

# then get the chart as an image with $chart->png

Nice and easy, just what I wanted – a way to say “here’s some timestamps and values (quite possibly irregularly spaced) – work out how to plot this sensibly for me”.

The resulting graphs look good enough to me, e.g.:





(Rendered intentionally a little smaller to fit the blog; naturally the graphs can be whatever size you want. Also, I had to use the transparent option to disable transparent backgrounds.)