On 05/08/2015 08:34 PM, Justin Clift wrote:
On 8 May 2015, at 13:16, Jeff Darcy <jdarcy@xxxxxxxxxx> wrote:
<snip>
Perhaps the change that's needed
is to make the fixing of likely-spurious test failures a higher
priority than adding new features.
YES! A million times Yes.
We need to move this project to operating with _0 regression
failures_ as the normal state of things for master and
release branches.
Regression failures for CR's in development... sure, that's a
normal part of development.
But any time a regression failure happens in _master_ or a
release branch should be a case of _get this fixed pronto_.
Agree completely.
Regression test failures need to be dealt with the highest priority. We
have not been dealing with this well right now but let us ensure that we
do not reach the same state anytime in the future. Let all of us be
vigilant to observe regression test failures and engage the right people
to have those fixed.
Thanks,
Vijay
_______________________________________________
Gluster-devel mailing list
Gluster-devel@xxxxxxxxxxx
http://www.gluster.org/mailman/listinfo/gluster-devel
