Mercurial and Agile

[Matt Mackall]

On Thu, 2009-07-09 at 09:48 +0200, Gilles Moris wrote:
> I tried this workflow, but it does not work.
> The first update (reverse merge) merges not only s but also everything
> between a -> p. This makes the merge of update non symmetric:

Yes, you've found a bug. Here's my simplified test case:

rm -rf a
hg init a
cd a
echo a > a
hg ci -Aqm0
echo b >> a
hg ci -Am1
hg up 0 -q
echo change > a
echo a >> a
hg up 1 -q
hg up 0 --debug

which outputs:

resolving manifests
 overwrite False partial False
 ancestor 21b2e120c769 local 21b2e120c769+ remote 456cf78ff65e
 a: versions differ -> m
preserving a for resolve of a
picked tool 'kdiff3' for a (binary False symlink False)
merging a
my a at 21b2e120c769+ other a at 456cf78ff65e ancestor a at 456cf78ff65e
 premerge successful
0 files updated, 1 files merged, 0 files removed, 0 files unresolved

Notice how on the third line, ancestor is p1, but on line 8, ancestor is
the target revision again. This looks like another instance of

http://mercurial.selenic.com/bts/issue1327

Guess I'll have to fix that.

-- 
http://selenic.com : development and support for Mercurial and Linux