Reputation
Top tag
Next privilege 150 Rep.
Create new tags
Badges
4
Newest
 Supporter
Impact
~404 people reached

  • 0 posts edited
  • 0 helpful flags
  • 1 vote cast
May
30
awarded  Supporter
May
30
comment How should I test the output of image renderers thoroughly, efficiently and in a branchable way
(enter is 'save', whoops) At the moment speed of image comparison is not the problem, but disk storage space. I had thought hashes would solve this, but then we have problems seeing 'what changed'. I clearly need to add some way of comparing one build output against another. Also, we don't use any eyeball comparison - everything is done automatically till failure. On failure, we can recognise our own artifacts, or we have analysts for 'judgement calls'.
May
30
comment How should I test the output of image renderers thoroughly, efficiently and in a branchable way
Thanks for your suggestion of the all pairs approach. Combinatorics was (/is) a new field to me, so plenty of interesting stuff to learn.
May
26
awarded  Editor
May
26
revised How should I test the output of image renderers thoroughly, efficiently and in a branchable way
Clarified that image comparison between expected/actual is automated.
May
26
comment How should I test the output of image renderers thoroughly, efficiently and in a branchable way
Sorry, I over edited and lost clarity! The initial comparison (in the copy and pasted test code) is automated. It's done by an internal image comparison library that does pixel based diffs. Only if the automated comparison fails do we do a human inspection to checkthe changes are acceptable.
May
26
awarded  Student
May
26
asked How should I test the output of image renderers thoroughly, efficiently and in a branchable way
May
11
awarded  Autobiographer