Christoph Hellwig wrote:
> On Tue, Apr 21, 2009 at 03:13:30PM -0500, Eric Sandeen wrote:
>> I was experimenting w/ default lazy-count=1, and it broke a few tests.
>> Seems simplest to enforce lazy-count=0 in the common.log routines
>> for now, and fix up the output? (vs. updating all the expected
>> output for these tests...)
>
> Didn't you just send a patch to make lazy-count=1 the default? :)
>
> I'd rather see either a filter dealing with both variants or make
> sure we always test the newer and more important case (lazy-count=1)
> here.
I don't think that a filter will be possible, it pretty fundamentally
changes log traffic doesn't it? I could make 2 expected outputs,
perhaps, one for each... I'll double check though.
I just made it forced to 0 because I don't think these tests are really
testing anything relevant to whether or not lazy-count is enabled...
-Eric
|