Intent of TestCommitMerge.test_merge_new_file? (was: [RFC] browsing history API)

[Aaron Bentley]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Robert Collins wrote:

| (Reparents are not listed in
| the versioned-file graph, ...

Actually, it appears that they are, because ie.describe_changes counts
reparentings as renames:

~        if old_entry.parent_id != new_entry.parent_id:
~            renamed = True
~        elif old_entry.name != new_entry.name:
~            renamed = True

I am working on nested tree support.  Now each new tree has its own root
id, and that breaks test_merge_new_file.  Because wtx and wty each has
their own root id, and so moving x/eks into y is a reparenting.

Is this meant to be an example of unrelated trees being grafted
together?  If so, eks should not wind up in y's root directory, by the
time I'm done.

Or is it meant to be an example of different changes to the same branch
not interfering with each other?  If so, x and y should have the same
root id.

Aaron
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iD8DBQFEdvKB0F+nu1YWqI0RAiy3AJ467eqdW3Q3qgfIprSt+tFpZBgkbgCfRCgF
NSuXTgVQwYYlbd+8C397rC4=
=vDan
-----END PGP SIGNATURE-----