User:Rcampbell/Toronto Summit Notes: Difference between revisions

From MozillaWiki
Jump to navigation Jump to search
No edit summary
 
m (GPHemsley moved page Robcee:Toronto Summit Notes to User:Rcampbell/Toronto Summit Notes without leaving a redirect: User page)
 
(3 intermediate revisions by one other user not shown)
Line 2: Line 2:


== automated testing ==
== automated testing ==
=== buildbot functions ===
=== Buildbot functions ===
* pull nightly
* pull nightly
* initiate test runs
* initiate test runs
* build given patch
* build given patch
** ben hearsum doing work with build team on this
** '''Ben Hearsum doing work with Build team on this'''
* select platforms
* select platforms
* reporting handled by performance testing framework
* reporting handled by performance testing framework
Line 41: Line 41:
* store some set period
* store some set period
* retain important historical milestones / benchmarks
* retain important historical milestones / benchmarks
== JS Tests ==
* specialized builds
* buildbot vs tinderbox canned builds
** requires js shell, xpcshell, can't use 'off-the-shelf' versions.
== Metrics ==
* bug density
** per component
** map bugzilla modules
** checkins churns
* lines of code
* performance thresholds
** links on wiki
* top-crashers

Latest revision as of 00:53, 11 May 2015

These are robcee's bulletized notes from the Toronto QA Development Summit on January 25th-26th, 2007. They are rough and there are large pieces missing. Feel free to fill in the gaps.

automated testing

Buildbot functions

  • pull nightly
  • initiate test runs
  • build given patch
    • Ben Hearsum doing work with Build team on this
  • select platforms
  • reporting handled by performance testing framework
  • litmus reporting
  • clean up build directory/create cleaner mirror
    • get rid of kruft
  • drive performance testing

Reporting Tools

Survey of reporting tools

  • litmus
  • buildbot
  • tinderbox

Report aggregation

  • RSS
  • Atom
  • Microformats
  • Security
    • passworded version
    • email
    • irc
  • report drill-down
    • history
    • test details
    • logs
    • store links in database for detailed analysis

Rollups

  • graphserver
  • buildbot?

History

  • store some set period
  • retain important historical milestones / benchmarks

JS Tests

  • specialized builds
  • buildbot vs tinderbox canned builds
    • requires js shell, xpcshell, can't use 'off-the-shelf' versions.

Metrics

  • bug density
    • per component
    • map bugzilla modules
    • checkins churns
  • lines of code
  • performance thresholds
    • links on wiki
  • top-crashers