[PATCH 1 of 1 V2] histedit: abort if there are multiple roots in "--outgoing" revisions

[Matt Mackall]

On Tue, 2013-09-17 at 05:25 +0900, FUJIWARA Katsunori wrote:
> At Sat, 14 Sep 2013 15:12:13 -0500,
> Matt Mackall wrote:
> > 
> > On Thu, 2013-09-12 at 22:33 -0500, Kevin Bullock wrote:
> > > On 9 Sep 2013, at 8:51 AM, FUJIWARA Katsunori wrote:
> > > 
> > > > # HG changeset patch
> > > > # User FUJIWARA Katsunori <foozy at lares.dti.ne.jp>
> > > > # Date 1378733863 -32400
> > > > #      Mon Sep 09 22:37:43 2013 +0900
> > > > # Node ID f0edc4e70f976b20d0c886a287b472a629789329
> > > > # Parent  1d07bf106c2ad1c7ef5e257e754ca8d858bd04b0
> > > > histedit: abort if there are multiple roots in "--outgoing" revisions
> > > 
> > > LGTM except that I think we need to bikeshed the messages we give the
> > > user here. Telling them the outgoing changesets have "multiple roots"
> > > will be terrifying. I don't have any better suggestions, though (yet).
> > 
> > Can we bikeshed by follow-up patch then?
> > 
> > This appears to be a bugfix that belongs on stable, yes?
> 
> I don't think strongly that this should be fiexed on stable, because
> (1) no issue has been reported yet about "histedit --outgoing" AFAIK
> (according to BTS search result), and (2) this problem doesn't seem so
> serious. And, I dropped "STABLE" flag from this patch series.

Having a BTS issue is not a requirement for fixing a bug. If histedit
already does something vaguely reasonable in this case, then perhaps
it's not a bug. But it seems like it's doing something silly currently.

-- 
Mathematics is the supreme nostalgia of our time.