Examples of com.puppetlabs.geppetto.junitresult.util.JunitresultAggregator.aggregate()

• com.puppetlabs.geppetto.junitresult.util.JunitresultAggregator.aggregate()
  Creates an aggregated JunitResult from all JUnit XML files discovered in a location under the given reportDir. The resulting JunitResult will have a name that is based on the path from the given rootDir to the discovered file (including the discovered file's name). The following rules are applied:

  • If the name of the file starts with "SPEC-" this part is dropped.
  • The suffix ".xml" is not included.
  • If a loaded testsuide contains testcases and these have different classname attributes, the testsuite is repackaged as one suite per classname. e.g. if ./mystack/allroottests.xml contains testcases from ./spec/x_spec.rb, and ./y_spec.rb (and each has testcases a and b) the resulting strucure is:

     mystack : suite  allroottests : suite spec/x_spec : suite a : tc b : tc y_spec : suite a : tc b : tc 

  • If a loaded testsuite has the same name as a single embedded testcase, the structure is flattented into a single testcase. This handles the output from ci_reporter. As an example, if tests are named the same way in two different ..._spec.rb files Game#score in a_spec.rb and b_spec.rb the result is:

     mystack : suite Game : suite Game.0 : suite Game-score : suite Game#score : tc Game-score.0 :suite 

    Without the flattening, each tc would be wrapped in a testsuite named the same way as the tc. Note that ci_reporter solves non unique naming by appending a ".n" to the filenames, but this is not applied inside the resulting xml. (The aggregator's method of using the name of the file as the name of the suite solves this issue - even if it is impossible to figure out where it comes from, it is at least unique. This is really a ci_reporter issue).
  • If the given reportDir, and rootDir are the same the last segment of the rootDir is used as the name of the root testsuite.
  • If a loaded testsuite's timestamp is missing, it will be derived from the file ctime it is loaded from.

  When scanning, all ".xml" files in the structure are assumed to be JUnit reports. Failure to load a file will insert a synthetic testsuite named after the file, and with a synthetic failed testcase.

  @param reportDir @param rootDir @return