====== Testing Plans ======

Started these plans regarding testing during the Evergreen Hack-A-Way 2012.

===== Easy wins =====

  * Source code analysis (could all be tied to git hooks) [Dan Scott]
    * POD
    * template compilation
    * perlcritic
  * pgtap [Galen Charlton]
    * eg_db_config.pl could invoke some pgtap tests
    * go for a method that works well with buildbot
  * srfsh testing [Mike Rylander]
    * broad coverage relatively soon?
    * we get regression tests from this
  * initial Test::More type stuff for perl utility functions (the ones that don't require a big mock environment)

===== Harder wins =====

  * refactoring code so more of it can be covered by unit tests
  * automated UI tests. selenium? etc?

==== Other benefits ====

  * Documentation of intent can come from testing

===== Other action items =====

  * Expanding test data set [Bill Erickson] [[https://bugs.launchpad.net/evergreen/+bug/1066888|Bug 1066888]]
  * Whitespace/style fixing commit soon (post 2.3.1?) [Dan Wells]
    * Schedule recurring ones of these for post-major-release periods
  * Human testing
    * Submit kernel of testing workflows/steps [Chris Sharp / Rogan Hamby]
    * Other community members will generalize these as necessary
    * Schedule a period for lots of community members dev and end-user to actually run through these scripts and produce bug reports
    * Find coordinator(s) for human testing [Mike Rylander]
  * Launchpad Management
    * The "chief nagger" will follow up and get attention for outstanding bug reports [Ben Shum]
    * Send out weekly/bi-weekly digest of totally unresponded-to Launchpad bugs