show details for failing examples tagged with aggregate_failures: true

[skaestle]

This will fix Issue #56 and was heavily inspired by #74 and all credit for the specs go to @jasoncodes

It outputs the exception.message in the case of a RSpec::Expectations::MultipleExpectationsNotMetError. This emulates what Core::Formatters::JsonFormatter does: https://github.com/rspec/rspec-core/blob/main/lib/rspec/core/formatters/json_formatter.rb#L43

Before

<?xml version="1.0" encoding="UTF-8"?>
<testsuite name="rspec" tests="1" skipped="0" failures="1" errors="0" time="0.014878" timestamp="2024-09-10T10:30:24+02:00" hostname="GREM-C02C90V3MD6V.local">
     <properties>
          <property name="seed" value="12345"/>
     </properties>
     <testcase classname="spec.example_spec" name="some example specs should support multiple failures" file="./spec/example_spec.rb" time="0.013631"><failure message="RSpec::Expectations::MultipleExpectationsNotMetError" type="RSpec::Expectations::MultipleExpectationsNotMetError">
     </failure></testcase>
</testsuite>

After

<?xml version="1.0" encoding="UTF-8"?>
<testsuite errors="0" failures="1" hostname="GREM-C02C90V3MD6V.local" name="rspec" skipped="0" tests="1" time="0.010382" timestamp="2024-09-10T10:31:22+02:00">
     <properties>
          <property name="seed" value="12345"/>
     </properties>
     <testcase classname="spec.example_spec" file="./spec/example_spec.rb" name="some example specs should support multiple failures" time="0.009463">
          <failure message="RSpec::Expectations::MultipleExpectationsNotMetError" type="RSpec::Expectations::MultipleExpectationsNotMetError">Got 2 failures from failure aggregation block:

  1) expected: 1
          got: &quot;foo&quot;

     (compared using ==)

     ./spec/example_spec.rb:53:in `block (2 levels) in &lt;top (required)&gt;'

  2) expected: 2
          got: &quot;bar&quot;

     (compared using ==)

     ./spec/example_spec.rb:54:in `block (2 levels) in &lt;top (required)&gt;'</failure>
     </testcase>
</testsuite>

[skaestle]

hi there @sj26. any chance you could have a look at this PR? we have this code running on our CI (gitlab) for awhile and we have yet to see any adverse side-effects

Let me know if there is anything to help get this PR into the main