Better SQL tracing with DBIx::Class

I've been trying to track down some SQL issues with Tau Station and to be honest, I've never been terribly happy with the output from the DBIx::Class DBIC_TRACE or the DBI DBI_TRACE. So I have something better.

Using the Perl debugger with DBIx::Class

I wrote about using the Perl debugger with Moose. In that post, I showed how to use DB::Skip to make it easier to use the Perl debugger with Moose and other modules whose guts you don't want to wade through.

Today's debugger hack will make using the debugger with DBIx::Class much easier.

Saving your test suite history

If you saw my FOSDEM talk about Building a Universe in Perl, I show some examples of the code we use to model behaviors in our universe. I start talking about out test suite at one point and that's probably the most exciting part. You see, we've been fixing our leaderboard system in part because I can do this:

That's a truncated version of our test failure report for tests failing on master. Leaderboard tests show up twice in there.

Ever have an failing test and trying to remember if it failed before? Is it fragile code? Is it fragile tests? Has it ever failed on your master branch? What percentage of your tests fail? Well, now you can find out.

Metric Time in Tau Station

If you've been following our progress with Tau Station, you know we're creating a science fiction universe in Perl for people to enjoy. As of this writing, the time in Tau Station is 194.10/51:647 GCT.

"GCT" stands for "Galactic Coordinated Time" and that's a variant of metric time. As a software developer, I wish we had that in the real world, but alas, we don't.

The GCT time listed above is roughly 194 years and 10 days after the "Catastrophe" (an apocalyptic event that effectively serves as our "epoch"). There are 100 days in a year, 100 "segments" in a day (14.4 minutes each) and 1000 units in a segment (.864 seconds each).

I love the fact that figuring out the display time for GCT is this simple:

my $days = sprintf "%9.5f" => $seconds_since_catastrophe / $second_in_a_day;
$days =~ m{^(?<year>\d+)(?<day>\d\d)\.(?<segment>\d\d)(?<unit>\d\d\d)}a;
my $gct = "$+{year}.$+{day}/$+{segment}:$+{unit} GCT";

Due to imprecision in normal dates, we don't get an exact round-trip conversion between regular DateTime objects and GCT, but so far we've not found them more than a second off.

Figuring out durations (D0.00/12.500) is similarly simple:

my $days = sprintf "%9.5f" => $duration_in_seconds / 86400;
$days =~ m{^(?<years>\d+)(?<days>\d\d)\.(?<segments>\d\d)(?<units>\d\d\d)}a;
my $duration => "D$+{years}.$+{days}/$+{segments}:$+{units}";

Of course, since that means we often need to know the total number of seconds, we have this nasty bit of code to figure that out:

sub period (%args) {
    my $seconds = delete $args{seconds} // 0;
    $seconds += ( delete $args{minutes}  // 0 ) * 60;
    $seconds += ( delete $args{hours}    // 0 ) * 3600;
    $seconds += ( delete $args{days}     // 0 ) * 86400;

    # solar year
    $seconds += ( delete $args{years}    // 0 ) * 31_556_925.97474;
    $seconds += ( delete $args{units}    // 0 ) * .864;
    $seconds += ( delete $args{segments} // 0 ) * 864;
    if ( keys %args ) {
        my $unknown = join ', ' => sort keys %args;
        croak("Unknown keys to Veure::Util::Time::period: $unknown");
    }
    return round($seconds);
}

Metric time is lovely and easy. Regular time sucks.

I really wanted to write a DBIx::Class inflator to use GCT objects instead of DateTime objects, but found too many assumptions about the use of DateTime in the DBIx::Class code, so we scrapped that bit. Darn shame.

I hope to see you at FOSDEM in Brussels

I've been rather quiet lately because between building Tau Station and working with some clients, I'm running around faster than a long-tailed cat in a room full of rocking chairs.

I'll be in Brussels for the February 4/5 2017, FOSDEM, talking about Building the Tau Station Universe in Perl. I was planning on giving a talk about testing, but I was specifically asked if I'd talk about Tau Station. While I love the project, I tried to think of a way it wouldn't sound like a 40 minute infomercial at an open source conference.

Woman relaxing in a high-tech hotel room in a space station.
Relaxing in a high-tech hotel room in a space station.

I'm pretty sure I've succeeded and I have to say, I think people are going to be really impressed with some of the code examples. In fact, it's gotten me to the point where I'm having more serious doubts about how object-oriented programming is currently taught. For example, what does the single responsibility principle mean for a combat suit that can pass the Turing test? It serves as armor, and might serve as a weapon, and might serve as a medkit, and even give the soldier a pep-talk.

Single-responsibility my ass.

The Tau Station talk will show a very simple way out of that conundrum.

We have an awesome team working on Tau Station. We've used the same hiring strategy that we use hiring for our clients at All Around the World and it's paid off in spades. You'll be impressed with their work. Fortunately, since this is FOSDEM, even if you can't make it to Brussels, the video will later be online for free and I'll post a link for you.

Hope to see you there!

Update: Once again I have had to close comments due to massive amounts of spam I have to clean out every day.

About Ovid

user-pic Freelance Perl/Testing/Agile consultant and trainer. See http://www.allaroundtheworld.fr/ for our services. If you have a problem with Perl, we will solve it for you. And don't forget to buy my book! http://www.amazon.com/Beginning-Perl-Curtis-Poe/dp/1118013840/