why 'merging'?

[Robert Figura]

Mads Kiilerich <mads at kiilerich.com> wrote:
> On 09/02/2009 03:27 PM, Neal Becker wrote:
> >   hg pull ssh://nbecker6/xxx
> > pulling from ssh://nbecker6/xxx
> > searching for changes
> > adding changesets
> > adding manifests
> > adding file changes
> > added 1 changesets with 1 changes to 1 files
> > (run 'hg update' to get a working copy)
> >
> > hg update
> > merging test/test_ldpc_pipelined.py
> > 0 files updated, 1 files merged, 0 files removed, 0 files unresolved
> >
> > The use of the word 'merge' confused me.  After the so-called merge, I find
> > that the remote file and the local file are binary-equal.
> >    
> 
> That is probably because you locally had an uncommitted change exactly 
> like what had been committed to the remote repo. The merge process 
> concluded that the same change had been made both places, and the result 
> was ... what it was.

It could be unrelated or it might be just me seeing ghosts but with the
following scheme related to mq i often run into empty merges:

branch devel: has file "filea"
branch subproduct: has files "filea" and "fileb"
file a is supposed to be changed only in devel branch

hg update subproduct
vi filea
hg qnew -f apatch
vi fileb
hg qnew -f sub:bpatch
hg qpop -a
hg update devel
hg qpush apatch
hg qfinish -a
hg update subproduct
hg qpush sub:bpatch
hg qfinish -a
hg merge -r devel

Before the merge happens filea should be unmodified relative to
subproduct due to updating to subproduct. But still it *sometimes*
seems to do empty merges.

In the past some confusions had happened and changes for a were
errorneously commited to the subproduct branch. Maybe this is relevant.

Or Maybe it happens when i reorder patches before qfinish-ing.

And well it might be that i'm just confused or am doing something else
wrong since i didn't run elaborate tests yet.

Regards
  - Robert Figura

-- 
http://digital-gecko.net/