Nice Perl Things

[nice-perl-things-blog]

Stack traces with Carp::Always

Your program just died. Great, now what?

If the exception string ended with a \n, it didn't give you a line number. But even if it did, that's often not enough to debug the problem.

Carp::Always to the rescue.

use Carp::Always;

Put this anywhere and all exceptions (and warnings) will have a full stack trace.

You probably don't want to turn this on permanently, because the stack trace will be added to exceptions even inside an 'eval BLOCK', and that's probably not what you wanted.

So since it's a temporary debugging tool, this is a pretty bad thing to forget in the code base since it changes global behaviour.

Do you know that nothing will go weird in the entire rest of your app now that exceptions will look different? No you do not. So put a reminder comment you can 'grep' for on the same line:

use Carp::Always; ### TODO: MyInitials: Remove before committing

Yes, it needs to have your initials on it, it's your personal responsibility to remove this.

If that's too scary, you can also load it on the command line with

perl -MCarp::Always script.pl prove -MCarp::Always -r t

just for that script invocation.

Remember: Carp::Always, but only some of the time.

[nice-perl-things-blog]

Which Utils Again?

There's List::Util and there's List::MoreUtils. Or was it List::Utils?

Ok, but which module contains uniq? How about first? part?

Alleviate the need to remember all of that by using List::AllUtils instead.

use List::AllUtils "uniq";

It contains all of the functions in the two previous modules.

The only annyance now is that it's not sensibly named in the singular, like its peer Scalar::Util.

[nice-perl-things-blog]

Show the SQL for a DBIC ResultSet

This is a simplistic method for dumping the query a DBIC result set will execute.

It’s simplisitc in that it doesn’t care about quoting values, or dealing with ? that happens to not be placeholders. But whatevs.

Put this in your base ResultSet class for the schema:

=head2 as_sql() : $sql Return $sql string with the query and its parameters. =cut sub as_sql { my $self = shift; my $query = $self->as_query; my ($sql, @args) = @$$query; my @arg_values = map { "'$_->[1]'" } @args; my $match_rex = qr/ \?([ ,])/ms; while ($sql =~ $match_rex && @arg_values) { my $value = shift @arg_values; $sql =~ s/$match_rex/ $value$1/ms; } @arg_values and warn( "as_sql for ($query) didn't make a use of all query args", ); return $sql; }

In your app:

my $book_rs = $schema->resultset("Book")->search( ... ); warn "SQL is:\n" . $book_rs->as_sql . "\n";

Now you can copy the query and paste it into your db console, maybe for getting a query plan.