Dates in diff output

[Bryan O'Sullivan]

On Fri, 2005-09-23 at 14:38 +1000, Michael Smith wrote:

> In the context of a service on the internet I believe UTC is most
> appropriate.

Well...  The nice thing about getting the committer's time is that it
makes some kind of human sense (this is what time it was where the
commit occurred).

The only value of UTC is that it lets you visually sort by time; it has
no human connection.

>  However the web interface may also be used within intranets
> where all viewers will be on the same time zone.

This is where that human bit comes in.  This makes a good case for
displaying committer-relative times (with timezone offsets) everywhere.
That way, you always get reasonably sensible behaviour, no configuration
knobs required.  You lose visual sorting, but that's of little real
value, anyway.

	<b