[MERGE] Refactor commit to prepare for population by tree walking

[Aaron Bentley]

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

Ian Clatworthy wrote:
> @@ -594,24 +590,64 @@
>  
>          specific_files = self.specific_files
>          mutter("Selecting files for commit with filter %s", specific_files)
> -        work_inv = self.work_tree.inventory
> -        assert work_inv.root is not None
> -        self.pb_entries_total = len(work_inv)
>  
>          # Check and warn about old CommitBuilders
> -        entries = work_inv.iter_entries()
> +        root_added_already = False
>          if not self.builder.record_root_entry:
>              symbol_versioning.warn('CommitBuilders should support recording'
>                  ' the root entry as of bzr 0.10.', DeprecationWarning, 
>                  stacklevel=1)
>              self.builder.new_inventory.add(self.basis_inv.root.copy())
> -            entries.next()
> -
> +            root_added_already = True
> +
> +        # Build the new inventory
> +        self._populate_from_inventory(specific_files, root_added_already)

This refactoring looks odd.  For one thing, the root hasn't really been
added.  For another, root_added_already is redundant with
self.builder.record_root_entry.

> +        # If specific files/directories were nominated, it is possible
> +        # that some data from outside those needs to be preserved from
> +        # the basis tree. For example, if a file x is moved from out of
> +        # directory foo into directory bar and the user requests
> +        # ``commit foo``, then information about bar/x must also be
> +        # recorded.

^^^ This implies that bar knows what its children are, which isn't
really accurate.  Strictly speaking, bar only needs to be committed if
it wasn't a directory in the basis tree-- the current implementation is
overkill.

> +            # Record an entry for this item
> +            # Note: I don't particularly want to have the existing_ie

It's confusing that you're referring to new_ie as existing_ie here.

Aaron
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFGl9lV0F+nu1YWqI0RAvxVAJ9ayrN0u/B9PIDAvLMW+LYilDQeYQCfUBEb
pHmXAIZHgZr1A7xkJJK7wZs=
=w8zG
-----END PGP SIGNATURE-----