Add a helper function to compare file contents
In the next commit, Git will learn to disallow hooks during `git clone` operations _except_ when those hooks come from the templates (which are inherently supposed to be trusted). To that end, we add a function to compare the contents of two files. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Johannes Schindelin committed
Mar 30, 2024 at 15:59 UTC
584de0b4c235209fa60ca4a733678472263bdce0
4 files changed
+123
cache.h
+14
@@ -1785,6 +1785,20 @@ int copy_fd(int ifd, int ofd);
1785
int copy_file(const char *dst, const char *src, int mode);
1786
int copy_file_with_time(const char *dst, const char *src, int mode);
1787
1788
+/*
1789
+ * Compare the file mode and contents of two given files.
1790
+ *
1791
+ * If both files are actually symbolic links, the function returns 1 if the link
1792
+ * targets are identical or 0 if they are not.
1793
+ *
1794
+ * If any of the two files cannot be accessed or in case of read failures, this
1795
+ * function returns 0.
1796
+ *
1797
+ * If the file modes and contents are identical, the function returns 1,
1798
+ * otherwise it returns 0.
1799
+ */
1800
+int do_files_match(const char *path1, const char *path2);
1801
+
1802
void write_or_die(int fd, const void *buf, size_t count);
1803
void fsync_or_die(int fd, const char *);
1804
int fsync_component(enum fsync_component component, int fd);
copy.c
+58
@@ -65,3 +65,61 @@ int copy_file_with_time(const char *dst, const char *src, int mode)
65
return copy_times(dst, src);
66
return status;
67
}
68
+
69
+static int do_symlinks_match(const char *path1, const char *path2)
70
+{
71
+ struct strbuf buf1 = STRBUF_INIT, buf2 = STRBUF_INIT;
72
+ int ret = 0;
73
+
74
+ if (!strbuf_readlink(&buf1, path1, 0) &&
75
+ !strbuf_readlink(&buf2, path2, 0))
76
+ ret = !strcmp(buf1.buf, buf2.buf);
77
+
78
+ strbuf_release(&buf1);
79
+ strbuf_release(&buf2);
80
+ return ret;
81
+}
82
+
83
+int do_files_match(const char *path1, const char *path2)
84
+{
85
+ struct stat st1, st2;
86
+ int fd1 = -1, fd2 = -1, ret = 1;
87
+ char buf1[8192], buf2[8192];
88
+
89
+ if ((fd1 = open_nofollow(path1, O_RDONLY)) < 0 ||
90
+ fstat(fd1, &st1) || !S_ISREG(st1.st_mode)) {
91
+ if (fd1 < 0 && errno == ELOOP)
92
+ /* maybe this is a symbolic link? */
93
+ return do_symlinks_match(path1, path2);
94
+ ret = 0;
95
+ } else if ((fd2 = open_nofollow(path2, O_RDONLY)) < 0 ||
96
+ fstat(fd2, &st2) || !S_ISREG(st2.st_mode)) {
97
+ ret = 0;
98
+ }
99
+
100
+ if (ret)
101
+ /* to match, neither must be executable, or both */
102
+ ret = !(st1.st_mode & 0111) == !(st2.st_mode & 0111);
103
+
104
+ if (ret)
105
+ ret = st1.st_size == st2.st_size;
106
+
107
+ while (ret) {
108
+ ssize_t len1 = read_in_full(fd1, buf1, sizeof(buf1));
109
+ ssize_t len2 = read_in_full(fd2, buf2, sizeof(buf2));
110
+
111
+ if (len1 < 0 || len2 < 0 || len1 != len2)
112
+ ret = 0; /* read error or different file size */
113
+ else if (!len1) /* len2 is also 0; hit EOF on both */
114
+ break; /* ret is still true */
115
+ else
116
+ ret = !memcmp(buf1, buf2, len1);
117
+ }
118
+
119
+ if (fd1 >= 0)
120
+ close(fd1);
121
+ if (fd2 >= 0)
122
+ close(fd2);
123
+
124
+ return ret;
125
+}
t/helper/test-path-utils.c
+10
@@ -495,6 +495,16 @@ int cmd__path_utils(int argc, const char **argv)
495
return !!res;
496
}
497
498
+ if (argc == 4 && !strcmp(argv[1], "do_files_match")) {
499
+ int ret = do_files_match(argv[2], argv[3]);
500
+
501
+ if (ret)
502
+ printf("equal\n");
503
+ else
504
+ printf("different\n");
505
+ return !ret;
506
+ }
507
+
508
fprintf(stderr, "%s: unknown function name: %s\n", argv[0],
509
argv[1] ? argv[1] : "(there was none)");
510
return 1;
t/t0060-path-utils.sh
+41
@@ -560,4 +560,45 @@ test_expect_success !VALGRIND,RUNTIME_PREFIX,CAN_EXEC_IN_PWD '%(prefix)/ works'
560
test_cmp expect actual
561
'
562
563
+test_expect_success 'do_files_match()' '
564
+ test_seq 0 10 >0-10.txt &&
565
+ test_seq -1 10 >-1-10.txt &&
566
+ test_seq 1 10 >1-10.txt &&
567
+ test_seq 1 9 >1-9.txt &&
568
+ test_seq 0 8 >0-8.txt &&
569
+
570
+ test-tool path-utils do_files_match 0-10.txt 0-10.txt >out &&
571
+
572
+ assert_fails() {
573
+ test_must_fail \
574
+ test-tool path-utils do_files_match "$1" "$2" >out &&
575
+ grep different out
576
+ } &&
577
+
578
+ assert_fails 0-8.txt 1-9.txt &&
579
+ assert_fails -1-10.txt 0-10.txt &&
580
+ assert_fails 1-10.txt 1-9.txt &&
581
+ assert_fails 1-10.txt .git &&
582
+ assert_fails does-not-exist 1-10.txt &&
583
+
584
+ if test_have_prereq FILEMODE
585
+ then
586
+ cp 0-10.txt 0-10.x &&
587
+ chmod a+x 0-10.x &&
588
+ assert_fails 0-10.txt 0-10.x
589
+ fi &&
590
+
591
+ if test_have_prereq SYMLINKS
592
+ then
593
+ ln -sf 0-10.txt symlink &&
594
+ ln -s 0-10.txt another-symlink &&
595
+ ln -s over-the-ocean yet-another-symlink &&
596
+ ln -s "$PWD/0-10.txt" absolute-symlink &&
597
+ assert_fails 0-10.txt symlink &&
598
+ test-tool path-utils do_files_match symlink another-symlink &&
599
+ assert_fails symlink yet-another-symlink &&
600
+ assert_fails symlink absolute-symlink
601
+ fi
602
+'
603
+
604
test_done