On Tue, Jul 31, 2012 at 02:28:21PM +0000, Trond Myklebust wrote:

On Tue, 2012-07-31 at 22:10 +0800, Fengguang Wu wrote:

quoted

quoted

quoted

Easy enough, the patch that fixes it is the last one in the series that I sent in yesterday.

I don't understand. That is the patch series that Fengguang was testing
afaik. His tests were on 

tree:   git://git.linux-nfs.org/projects/trondmy/linux-nfs.git devel
head:   5c13c9e1c15ee2ca9ab2b953224001af53d9be09

which includes your patch series from yesterday, no?

Trond, I typically do commit-by-commit tests and will complain if
*any* point of the tree is not bisect-able. A fix at the HEAD won't
stop the email notification for a defect in the middle point..

On the other hand, I do maintain a list of non-rebaseable
tree/branches, on which the HEAD commit will be tested first, and only
if any problems are found, go back to find out the first bad commit.

If you prefer the latter behavior, I can add your tree or any branch
of it to the non-rebaseable list.

Ah... I see now... So you are saying that the end result is indeed
correct, but the bisection fails...

Sorry for the confusions!

The 'devel' tree is usually a mixture of rebaseable and non-rebaseable:
anything that is already been committed to the nfs-for-next branch is
non-rebaseable, while the rest usually is...

Got it. I'll mark 'nfs-for-next' and 'devel' as non-rebaseable, and do
thorough bisect tests on the other branches.

Thanks,
Fengguang