You are viewing the version of this documentation from Perl 5.8.0. View the latest version



Test::Harness::Straps - detailed analysis of test results


use Test::Harness::Straps;

my $strap = Test::Harness::Straps->new;

# Various ways to interpret a test
my %results = $strap->analyze($name, \@test_output);
my %results = $strap->analyze_fh($name, $test_filehandle);
my %results = $strap->analyze_file($test_file);

my %total = $strap->total_results;

# Altering the behavior of the strap  UNIMPLEMENTED
my $verbose_output = $strap->dump_verbose();


THIS IS ALPHA SOFTWARE in that the interface is subject to change in incompatible ways. It is otherwise stable.

Test::Harness is limited to printing out its results. This makes analysis of the test results difficult for anything but a human. To make it easier for programs to work with test results, we provide Test::Harness::Straps. Instead of printing the results, straps provide them as raw data. You can also configure how the tests are to be run.

The interface is currently incomplete. Please contact the author if you'd like a feature added or something change or just have comments.


my $strap = Test::Harness::Straps->new;

Initialize a new strap.


my %results = $strap->analyze($name, \@test_output);

Analyzes the output of a single test, assigning it the given $name for use in the total report. Returns the %results of the test. See Results.

@test_output should be the raw output from the test, including newlines.

my %results = $strap->analyze_fh($name, $test_filehandle);

Like analyze, but it reads from the given filehandle.

my %results = $strap->analyze_file($test_file);

Like analyze, but it runs the given $test_file and parses it's results. It will also use that name for the total report.


The %results returned from analyze() contain the following information:

passing           true if the whole test is considered a pass 
                  (or skipped), false if its a failure

exit              the exit code of the test run, if from a file
wait              the wait code of the test run, if from a file

max               total tests which should have been run
seen              total tests actually seen
skip_all          if the whole test was skipped, this will 
                    contain the reason.

ok                number of tests which passed 
                    (including todo and skips)

todo              number of todo tests seen
bonus             number of todo tests which 
                    unexpectedly passed

skip              number of tests skipped

So a successful test should have max == seen == ok.

There is one final item, the details.

details           an array ref reporting the result of 
                  each test looks like this:

  $results{details}[$test_num - 1] = 
          { ok        => is the test considered ok?
            actual_ok => did it literally say 'ok'?
            name      => name of the test (if any)
            type      => 'skip' or 'todo' (if any)
            reason    => reason for the above (if any)

Element 0 of the details is test #1. I tried it with element 1 being #1 and 0 being empty, this is less awkward.


See examples/mini_harness.plx for an example of use.


Michael G Schwern <>