cmpilato@tigris.org writes:
> --- branches/fs-atomic-renames/subversion/libsvn_fs_base/fs.h (original)
> +++ branches/fs-atomic-renames/subversion/libsvn_fs_base/fs.h Wed Sep 14 15:28:13 2005
> @@ -276,7 +276,9 @@
> typedef enum
> {
> copy_kind_real = 1, /* real copy */
> - copy_kind_soft /* soft copy */
> + copy_kind_soft, /* soft copy */
> + copy_kind_move, /* real move */
> + copy_kind_soft_move /* soft move */
>
> } copy_kind_t;
Thank goodness for libsvn_fs_base/notes/fs-history's explanation of
soft copies and, by implication, soft moves :-).
> @@ -328,6 +330,18 @@
> const char *lock_token;
>
> } lock_node_t;
> +
> +
> +/*** Directory Entry ***/
> +typedef struct
> +{
> + /* entry node revision id */
> + const svn_fs_id_t *id;
> +
> + /* (optional) move id */
> + const char *move_id;
> +
> +} directory_entry_t;
It might be useful to have a doc string explaining why this structure
is necessary, i.e., why it is not the same as svn_fs_dirent_t. Would
you approve this language?:
/*** Directory Entry ***/
/* This serves a different purpose from the public svn_fs_dirent_t.
Whereas svn_fs_dirent_t is designed to convey externally
interesting information to the outside world, this structure is
designed to carry around private information (e.g., the move_id)
inside the filesystem. The 'svn_fs_id_t *id' field tells you
when they represent the same directory entry, of course. */
typedef struct
{
...
} directory_entry_t;
-Karl
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Thu Sep 15 18:25:40 2005