[PATCH 2 of 3 RFC] flags: introduce a command for manipulating file flags (issue2020)

[timeless]

Matt Harbison wrote:
> Interesting idea.  I wish it were less verbose, but I don't have a better
> idea.

the only thing that could make it less verbose is anchoring it to
another existing command -- I won't have resources to check this week.

> The octal idea is interesting too, but since Mercurial really only tracks a
> single x-bit and then turns them on for each r-bit (see posix.setflags()),
> there will be a bunch of combinations that make little or no sense.  Should
> they be rejected?  Just the u+x bit tested?

I'm not saying we should support it today, just noting that this
proposal could support it eventually.