On 05/12/2011 09:39 AM, Ingo Weinhold wrote:
I use Mercurial all the time, but I don't go too far in depth typically so I can't say for sure, but I doubt it. I was thinking more along the lines of adding a tag to each revision with the output of the date command. It could offer sufficient precision that including the hash would be unnecessary. I could just be completely out in left field on this one, but it was just a thought.On Thu, 12 May 2011 07:26:17 -0400 Jon Yoder<darkwyrm@xxxxxxxxx> wrote:On 05/12/2011 06:26 AM, Ingo Weinhold wrote:I'm still fan of the comparable revision number idea. Maybe the datepartcan be extended by another component to make it unique? E.g. "2011-05-11-13" would denote the 13th changeset that day. Would that be possible?I am, too. Another possibility would be to include the time down to the nanosecond level -- something like `date -u "+%Y-%m-%d-%k:%M.%N"` could get the job in a way that is quite readable and may not even need to include the hash itself. Just my $0.02.Do the tools even store times with that precision? At least a simple "log" only produces full seconds with either. git's --date=raw option (which I'd expect to show what is actually stored) produces what looks like a number of seconds since some point in time.