On Mon, Jul 30, 2012 at 1:19 PM, C. Michael Pilato <cmpil...@collab.net> wrote: > On 07/30/2012 12:47 PM, Philip Martin wrote: >> I'm considering changing the commit code so that hashes are written in a >> stable order and the revision files are repeatable. Does anyone think >> this would be a bad idea? > > I can't think of any reason not to do this. The performance overhead will > be minimal, I should think, and you've already outlined the benefits.
As long as the performance overhead is negligible, I think having deterministic ordering is very desirable. There's just something in me that longs to have a One True Ordering in dumpfiles and revfiles and output and everything else. I can't really say why, but it feels like by having a canonical ordering, we open ourselves up for more opportunities to cache work moving forward. (How's that for nebulous and hand wavy?) -Hyrum
