Skip to content

Commit 0f4256b

Browse files
ethomsonethomson
committed
repository: improve hashfile for absolute paths
When `git_repository_hashfile` is handed an absolute path, it determines whether the path is within the repository's working directory or not. This is necessary when there is no `as_path` specified. If the path is within the working directory, then the given path should be used for attribute lookups (it is the effective `as_path`). If it is not within the working directory, then it is _not_ eligible. Importantly, now we will _never_ pass an absolute path down to attribute lookup functions.
1 parent d6c7ca3 commit 0f4256b

File tree

3 files changed

+108
-17
lines changed

3 files changed

+108
-17
lines changed

‎include/git2/repository.h‎

Lines changed: 7 additions & 5 deletions
Original file line numberDiff line numberDiff line change
@@ -762,13 +762,15 @@ GIT_EXTERN(int) git_repository_mergehead_foreach(
762762
*
763763
* @param out Output value of calculated SHA
764764
* @param repo Repository pointer
765-
* @param path Path to file on disk whose contents should be hashed. If the
766-
* repository is not NULL, this can be a relative path.
765+
* @param path Path to file on disk whose contents should be hashed. This
766+
* may be an absolute path or a relative path, in which case it
767+
* will be treated as a path within the working directory.
767768
* @param type The object type to hash as (e.g. GIT_OBJECT_BLOB)
768769
* @param as_path The path to use to look up filtering rules. If this is
769-
* NULL, then the `path` parameter will be used instead. If
770-
* this is passed as the empty string, then no filters will be
771-
* applied when calculating the hash.
770+
* an empty string then no filters will be applied when
771+
* calculating the hash. If this is `NULL` and the `path`
772+
* parameter is a file within the repository's working
773+
* directory, then the `path` will be used.
772774
* @return 0 on success, or an error code
773775
*/
774776
GIT_EXTERN(int) git_repository_hashfile(

‎src/repository.c‎

Lines changed: 13 additions & 11 deletions
Original file line numberDiff line numberDiff line change
@@ -2840,34 +2840,36 @@ int git_repository_hashfile(
28402840
git_file fd = -1;
28412841
uint64_t len;
28422842
git_buf full_path = GIT_BUF_INIT;
2843+
const char *workdir = git_repository_workdir(repo);
28432844

28442845
/* as_path can be NULL */
28452846
GIT_ASSERT_ARG(out);
28462847
GIT_ASSERT_ARG(path);
28472848
GIT_ASSERT_ARG(repo);
28482849

2849-
/* At some point, it would be nice if repo could be NULL to just
2850-
* apply filter rules defined in system and global files, but for
2851-
* now that is not possible because git_filters_load() needs it.
2852-
*/
2853-
2854-
if ((error = git_path_join_unrooted(
2855-
&full_path, path, git_repository_workdir(repo), NULL)) < 0 ||
2850+
if ((error = git_path_join_unrooted(&full_path, path, workdir, NULL)) < 0 ||
28562851
(error = git_path_validate_workdir_buf(repo, &full_path)) < 0)
28572852
return error;
28582853

2859-
if (!as_path)
2860-
as_path = path;
2854+
/*
2855+
* NULL as_path means that we should derive it from the
2856+
* given path.
2857+
*/
2858+
if (!as_path) {
2859+
if (workdir && !git__prefixcmp(full_path.ptr, workdir))
2860+
as_path = full_path.ptr + strlen(workdir);
2861+
else
2862+
as_path = "";
2863+
}
28612864

28622865
/* passing empty string for "as_path" indicated --no-filters */
28632866
if (strlen(as_path) > 0) {
28642867
error = git_filter_list_load(
28652868
&fl, repo, NULL, as_path,
28662869
GIT_FILTER_TO_ODB, GIT_FILTER_DEFAULT);
2870+
28672871
if (error < 0)
28682872
return error;
2869-
} else {
2870-
error = 0;
28712873
}
28722874

28732875
/* at this point, error is a count of the number of loaded filters */

‎tests/repo/hashfile.c‎

Lines changed: 88 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -10,6 +10,7 @@ void test_repo_hashfile__initialize(void)
1010

1111
void test_repo_hashfile__cleanup(void)
1212
{
13+
cl_fixture_cleanup("absolute");
1314
cl_git_sandbox_cleanup();
1415
_repo = NULL;
1516
}
@@ -38,10 +39,18 @@ void test_repo_hashfile__simple(void)
3839
git_buf_dispose(&full);
3940
}
4041

41-
void test_repo_hashfile__filtered(void)
42+
void test_repo_hashfile__filtered_in_workdir(void)
4243
{
44+
git_buf root = GIT_BUF_INIT, txt = GIT_BUF_INIT, bin = GIT_BUF_INIT;
45+
char cwd[GIT_PATH_MAX];
4346
git_oid a, b;
4447

48+
cl_must_pass(p_getcwd(cwd, GIT_PATH_MAX));
49+
cl_must_pass(p_mkdir("absolute", 0777));
50+
cl_git_pass(git_buf_joinpath(&root, cwd, "status"));
51+
cl_git_pass(git_buf_joinpath(&txt, root.ptr, "testfile.txt"));
52+
cl_git_pass(git_buf_joinpath(&bin, root.ptr, "testfile.bin"));
53+
4554
cl_repo_set_bool(_repo, "core.autocrlf", true);
4655

4756
cl_git_append2file("status/.gitattributes", "*.txt text\n*.bin binary\n\n");
@@ -55,21 +64,41 @@ void test_repo_hashfile__filtered(void)
5564
cl_git_pass(git_repository_hashfile(&b, _repo, "testfile.txt", GIT_OBJECT_BLOB, NULL));
5665
cl_assert(git_oid_cmp(&a, &b));
5766

67+
/* not equal hashes because of filtering when specified by absolute path */
68+
cl_git_pass(git_odb_hashfile(&a, "status/testfile.txt", GIT_OBJECT_BLOB));
69+
cl_git_pass(git_repository_hashfile(&b, _repo, txt.ptr, GIT_OBJECT_BLOB, NULL));
70+
cl_assert(git_oid_cmp(&a, &b));
71+
5872
/* equal hashes because filter is binary */
5973
cl_git_pass(git_odb_hashfile(&a, "status/testfile.bin", GIT_OBJECT_BLOB));
6074
cl_git_pass(git_repository_hashfile(&b, _repo, "testfile.bin", GIT_OBJECT_BLOB, NULL));
6175
cl_assert_equal_oid(&a, &b);
6276

77+
/* equal hashes because filter is binary when specified by absolute path */
78+
cl_git_pass(git_odb_hashfile(&a, "status/testfile.bin", GIT_OBJECT_BLOB));
79+
cl_git_pass(git_repository_hashfile(&b, _repo, bin.ptr, GIT_OBJECT_BLOB, NULL));
80+
cl_assert_equal_oid(&a, &b);
81+
6382
/* equal hashes when 'as_file' points to binary filtering */
6483
cl_git_pass(git_odb_hashfile(&a, "status/testfile.txt", GIT_OBJECT_BLOB));
6584
cl_git_pass(git_repository_hashfile(&b, _repo, "testfile.txt", GIT_OBJECT_BLOB, "foo.bin"));
6685
cl_assert_equal_oid(&a, &b);
6786

87+
/* equal hashes when 'as_file' points to binary filtering (absolute path) */
88+
cl_git_pass(git_odb_hashfile(&a, "status/testfile.txt", GIT_OBJECT_BLOB));
89+
cl_git_pass(git_repository_hashfile(&b, _repo, txt.ptr, GIT_OBJECT_BLOB, "foo.bin"));
90+
cl_assert_equal_oid(&a, &b);
91+
6892
/* not equal hashes when 'as_file' points to text filtering */
6993
cl_git_pass(git_odb_hashfile(&a, "status/testfile.bin", GIT_OBJECT_BLOB));
7094
cl_git_pass(git_repository_hashfile(&b, _repo, "testfile.bin", GIT_OBJECT_BLOB, "foo.txt"));
7195
cl_assert(git_oid_cmp(&a, &b));
7296

97+
/* not equal hashes when 'as_file' points to text filtering */
98+
cl_git_pass(git_odb_hashfile(&a, "status/testfile.bin", GIT_OBJECT_BLOB));
99+
cl_git_pass(git_repository_hashfile(&b, _repo, bin.ptr, GIT_OBJECT_BLOB, "foo.txt"));
100+
cl_assert(git_oid_cmp(&a, &b));
101+
73102
/* equal hashes when 'as_file' is empty and turns off filtering */
74103
cl_git_pass(git_odb_hashfile(&a, "status/testfile.txt", GIT_OBJECT_BLOB));
75104
cl_git_pass(git_repository_hashfile(&b, _repo, "testfile.txt", GIT_OBJECT_BLOB, ""));
@@ -79,7 +108,65 @@ void test_repo_hashfile__filtered(void)
79108
cl_git_pass(git_repository_hashfile(&b, _repo, "testfile.bin", GIT_OBJECT_BLOB, ""));
80109
cl_assert_equal_oid(&a, &b);
81110

111+
cl_git_pass(git_odb_hashfile(&a, "status/testfile.txt", GIT_OBJECT_BLOB));
112+
cl_git_pass(git_repository_hashfile(&b, _repo, txt.ptr, GIT_OBJECT_BLOB, ""));
113+
cl_assert_equal_oid(&a, &b);
114+
115+
cl_git_pass(git_odb_hashfile(&a, "status/testfile.bin", GIT_OBJECT_BLOB));
116+
cl_git_pass(git_repository_hashfile(&b, _repo, bin.ptr, GIT_OBJECT_BLOB, ""));
117+
cl_assert_equal_oid(&a, &b);
118+
82119
/* some hash type failures */
83120
cl_git_fail(git_odb_hashfile(&a, "status/testfile.txt", 0));
84121
cl_git_fail(git_repository_hashfile(&b, _repo, "testfile.txt", GIT_OBJECT_ANY, NULL));
122+
123+
git_buf_dispose(&txt);
124+
git_buf_dispose(&bin);
125+
git_buf_dispose(&root);
126+
}
127+
128+
void test_repo_hashfile__filtered_outside_workdir(void)
129+
{
130+
git_buf root = GIT_BUF_INIT, txt = GIT_BUF_INIT, bin = GIT_BUF_INIT;
131+
char cwd[GIT_PATH_MAX];
132+
git_oid a, b;
133+
134+
cl_must_pass(p_getcwd(cwd, GIT_PATH_MAX));
135+
cl_must_pass(p_mkdir("absolute", 0777));
136+
cl_git_pass(git_buf_joinpath(&root, cwd, "absolute"));
137+
cl_git_pass(git_buf_joinpath(&txt, root.ptr, "testfile.txt"));
138+
cl_git_pass(git_buf_joinpath(&bin, root.ptr, "testfile.bin"));
139+
140+
cl_repo_set_bool(_repo, "core.autocrlf", true);
141+
cl_git_append2file("status/.gitattributes", "*.txt text\n*.bin binary\n\n");
142+
143+
/* create some sample content with CRLF in it */
144+
cl_git_mkfile("absolute/testfile.txt", "content\r\n");
145+
cl_git_mkfile("absolute/testfile.bin", "other\r\nstuff\r\n");
146+
147+
/* not equal hashes because of filtering */
148+
cl_git_pass(git_odb_hashfile(&a, "absolute/testfile.txt", GIT_OBJECT_BLOB));
149+
cl_git_pass(git_repository_hashfile(&b, _repo, txt.ptr, GIT_OBJECT_BLOB, "testfile.txt"));
150+
cl_assert(git_oid_cmp(&a, &b));
151+
152+
/* equal hashes because filter is binary */
153+
cl_git_pass(git_odb_hashfile(&a, "absolute/testfile.bin", GIT_OBJECT_BLOB));
154+
cl_git_pass(git_repository_hashfile(&b, _repo, bin.ptr, GIT_OBJECT_BLOB, "testfile.bin"));
155+
cl_assert_equal_oid(&a, &b);
156+
157+
/*
158+
* equal hashes because no filtering occurs for absolute paths outside the working
159+
* directory unless as_path is specified
160+
*/
161+
cl_git_pass(git_odb_hashfile(&a, "absolute/testfile.txt", GIT_OBJECT_BLOB));
162+
cl_git_pass(git_repository_hashfile(&b, _repo, txt.ptr, GIT_OBJECT_BLOB, NULL));
163+
cl_assert_equal_oid(&a, &b);
164+
165+
cl_git_pass(git_odb_hashfile(&a, "absolute/testfile.bin", GIT_OBJECT_BLOB));
166+
cl_git_pass(git_repository_hashfile(&b, _repo, bin.ptr, GIT_OBJECT_BLOB, NULL));
167+
cl_assert_equal_oid(&a, &b);
168+
169+
git_buf_dispose(&txt);
170+
git_buf_dispose(&bin);
171+
git_buf_dispose(&root);
85172
}

0 commit comments

Comments
 (0)