On Sun, Apr 07, 2019 at 03:45:36PM +0300, Nikolay Borisov wrote: > > > On 7.04.19 г. 14:54 ч., Anand Jain wrote: > > On 6/4/19 8:02 pm, Eryu Guan wrote: > >> On Fri, Apr 05, 2019 at 04:21:10PM +0300, Nikolay Borisov wrote: > >>> > >>> > >>> On 3.04.19 г. 20:04 ч., Anand Jain wrote: > >>>> Add more property validation cases which are fixed by the patches [1] > >>>> [1] > >>>> btrfs: fix vanished compression property after failed set > >>>> btrfs: fix zstd compression parameter > >>>> > >>>> Signed-off-by: Anand Jain <anand.jain@xxxxxxxxxx> > >>> > >>> Reviewed-by: Nikolay Borisov <nborisov@xxxxxxxx> > >> > >> Thanks for the test and the review! > >> > >> But this looks like a targeted regression test that may fail an existing > >> test. It's better to write a new test for this. > > > > Regression is only when fstests is upgraded. This > > test case mentions the prerequisite kernel patches [1]. > > So that should suffice the concern? > > I agree with Anand here, this is an extension to an existing test, which > covers specific feature. IMO it's not good to always introduce new tests > because every invocation of a test comes with an overhead of spawning > processes and whatnot. THis is not a problem for 10 tests, but currently > for btrfs we execute around 600 tests each one "wasting" some cycles to > spawn bash processes to execute the actual test. Fair enough. We're having more tests now, and we do consider "merging" some tests into one case. Thanks, Eryu