2

Sometimes after autotest runs specs and encounters an failure, it will immediately re-run all the specs in the project, which (a) I have no need for and (b) clears the terminal so I can't read the results of the failed spec.

Why is this happening?

John Bachir
  • 22,495
  • 29
  • 154
  • 227
  • I don't think I've ever seen that behavior. Is one of your tests writing/touching a file in the directory, causing autotest to mistakenly think a source file changed? – the Tin Man Jan 31 '11 at 22:10
  • When you say "encounters an error" do you mean that you had a failing test, or that there was actually a syntax error in your code, etc. (resulting in a big backtrace dumped into your output)? – Dylan Markow Jan 31 '11 at 22:12
  • @tinman: hmm, interesting idea @dmarkow: whoops, i meant, a failure. I'll fix the question... – John Bachir Jan 31 '11 at 23:07

1 Answers1

1

This can happen for a number of reasons:

  1. A file was touched that is not in the ignore list or has a mapping.
  2. Autotest could not parse your test output for fail state and assumed that walking back up the failure tree is reasonable.
  3. Your tests sent an interrupt to autotest.
raggi
  • 1,290
  • 9
  • 12
  • FWIW, I'm still seeing behavior where the above 3 don't seem to explain it. Is there any way I can have autotest tell me why it chose to re-run everything? – John Bachir Feb 20 '11 at 02:49
  • If it's re-running all specs, then it's assuming that all your specs passed. This suggests that the most likely cause is that it's not parsing your test output correctly. More details are required to answer the original question. No there is no immediately configurable way to output the reason, but the output is uniquely indicative in each of the above cases. – raggi Mar 06 '11 at 09:04