Removing BASE-DELAYWRITE in make test

Jeremy Allison jra at
Thu May 21 02:36:09 GMT 2009

On Wed, May 20, 2009 at 06:43:11PM -0700, Tim Prouty wrote:
> You have probably noticed that most of the build machines flip-flop
> between passing and failing BASE-DELAYWRITE.  This flip-flopping
> causes many of the build machines to send a build break mail for
> nearly every commit that is pushed.  This makes it time-consuming and
> error-prone to differentiate real build breaks from BASE-DELAYWRITE
> failures.
> Ideally we would fix BASE-DELAYWRITE, but since this has been
> happening for at least the last few months and no one has looked into
> it, I propose we remove it from the source3 'make test' temporarily.
> It really seems to be doing more harm than good in its current form.
> When someone has time to give it the love it needs, then it can easily
> be added back.

The problem is when it passes sometimes, there's little
incentive to look closely :-).

If it's bugging you, can you give me a week or so to
look into it and see if I can tweak any of the timing
parameters to see what might be the problem ?

Of course, it always passes on a local 'make test', that's
the problem :-).


More information about the samba-technical mailing list