Opened 12 years ago
Last modified 12 years ago
#1397 defect new
Trial reports SUCCESS!?! and ERROR for the same test
Reported by: | jknight | Owned by: | |
---|---|---|---|
Priority: | high | Milestone: | |
Component: | trial | Keywords: | waiting |
Cc: | Jonathan Lange, jknight, Jean-Paul Calderone, spiv | Branch: | |
Author: |
Description
Change History (8)
comment:2 Changed 12 years ago by
testCallLaterOrder2 is managing to call log.err(). I guess this is because of unhandled errors in callLaters. So, this can be reduced to an earlier problem -- #1384
comment:3 Changed 12 years ago by
Cc: | Jean-Paul Calderone spiv added |
---|
My last comment is mistaken.
The real problem is that log error calls are handled in cleanup, and todo has no authority there.
This creates two topics of interest:
# Should .todo cover the entire test, or just the test method? # log.err calls should be handled differently.
#1409 addresses the second point.
What about the first? (CC: opening up the discussion to other vocal trial users)
comment:4 Changed 12 years ago by
Keywords: | waiting added |
---|
comment:6 Changed 12 years ago by
I think I agree with jknight that todo should cover the entire test.
In general, it's surprising any time multiple results are associated with a single test.
comment:7 Changed 12 years ago by
So, I'm happy now to make .todo cover the entire test.
However, as stated elsewhere (citation needed) I am committed to tests being able to have multiple results. Or rather, I am convinced that trial shouldn't hide errors, unless explicitly asked to (ala todo).
comment:8 Changed 7 years ago by
Owner: | Jonathan Lange deleted |
---|