On Tue, May 24, 2022 at 12:52 PM Dave Chinner <david@xxxxxxxxxxxxx> wrote: > > On Tue, May 24, 2022 at 12:04:36PM +0300, Amir Goldstein wrote: > > On Tue, May 24, 2022 at 11:32 AM Dave Chinner <david@xxxxxxxxxxxxx> wrote: > > > diff --git a/tests/xfs/026 b/tests/xfs/026 > > > index 18529003..0daa7c88 100755 > > > --- a/tests/xfs/026 > > > +++ b/tests/xfs/026 > > > @@ -9,17 +9,8 @@ > > > . ./common/preamble > > > _begin_fstest dump ioctl auto quick > > > > > > -status=0 # success is the default! > > > - > > > > All those tests that you change from success is the default > > to failure is the default, that makes sense, but > > 1. Should be documented in the commit message? > > I cleaned up a bunch of random stuff like this as I went along. If I > try to document every little change that gets done, I'll be spending > more time on describing the changes than actually doing them. I > don't have the time to rigourously document this stuff - I've got > the time to modify the code, write an overview of the changes and > that's about it. > > What does the fstests community want: better test infrastructure or > perfect commits? The former :) > > > 2. Why were those tests written this way? Do you know? > > That's just the pattern that was used by the team that wrote these > dump tests 20-odd years ago. At the time nobody cared that much > about little inconsistencies like this as long as the test failed > when something went wrong. i.e. test works, merge it, move on to the > next thing. > > Now that we have ~1700 tests to maintain, little inconsistencies are > a big deal. That's why I cleaned up the simple ones as I saw them... > That's super. Thank you for that. The reason I am asking is that with kdevops, we observed some inconsistencies between xunit reported failures and overall run success/failure. I was wondering if tests that initialize status=0 could be the source of these inconsistencies. Thanks, Amir.
