[PATCH] tdb: recycle dead records always at chain top

Jeremy Allison jra at samba.org
Tue Apr 8 17:52:01 MDT 2014


On Wed, Apr 09, 2014 at 01:45:12AM +0200, Michael Adam wrote:
> 
> I.e. the subtelty is that before my change, the i = 0 case, which
> is the case of the current candidate record's chain, is a special
> case treated outside of the for loop.

And that is the magic that I missed :-).

> My change removes the special treatment of this
> case, and hence moves this into the loop, but reverses
> the order of the two steps inside of the loop, so that
> the order of execution remains the same as before.
> 
> So I don't quite see how I can untangle this reasonably
> into multiple commits. (You know I am in favour of many
> small commits...) Each such commit I can imagine, would
> create a bigger change in behaviour.
> 
> If you can give me a hint, I will gladly follow up with
> an updated patch(set), but maybe the change just needs
> a better commit message.

No, I think it's good now - thanks *very* much
for the clarification.

Please feel free to push with my 'Reviewed-by:'

Thanks !

Jeremy.


More information about the samba-technical mailing list