directory_exists_in_index_icase() dangerously assumed that it could
access one character beyond the end of its directory argument, and that
that character would unconditionally be '/'. 2eac2a4c (ls-files -k: a
directory only can be killed if the index has a non-directory,
2013-08-15) added a caller which did not respect this undocumented
assumption, and 680be044 (dir.c::test_one_path(): work around
directory_exists_in_index_icase() breakage, 2013-08-23) added a
work-around which temporarily appends a '/' before invoking
directory_exists_in_index_icase().
Since the dangerous behavior of directory_exists_in_index_icase() has
been eliminated, the work-around is now redundant, so retire it (but not
the tests added by the same commit).
Signed-off-by: Eric Sunshine <sunshine@xxxxxxxxxxxxxx>
---
dir.c | 18 +++---------------
1 file changed, 3 insertions(+), 15 deletions(-)
diff --git a/dir.c b/dir.c
index fccd479..23b6de4 100644
--- a/dir.c
+++ b/dir.c
@@ -1160,21 +1160,9 @@ static enum path_treatment treat_one_path(struct dir_struct *dir,
*/
if ((dir->flags & DIR_COLLECT_KILLED_ONLY) &&
(dtype == DT_DIR) &&
- !has_path_in_index) {
- /*
- * NEEDSWORK: directory_exists_in_index_icase()
- * assumes that one byte past the given path is
- * readable and has '/', which needs to be fixed, but
- * until then, work it around in the caller.
- */
- strbuf_addch(path, '/');
- if (directory_exists_in_index(path->buf, path->len - 1) ==
- index_nonexistent) {
- strbuf_setlen(path, path->len - 1);
- return path_none;
- }
- strbuf_setlen(path, path->len - 1);
- }
+ !has_path_in_index &&
+ (directory_exists_in_index(path->buf, path->len) == index_nonexistent))
+ return path_none;
exclude = is_excluded(dir, path->buf, &dtype);
--
1.8.4.535.g7b94f8e
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html