@matheusbsbs / matheus-bagliano / commits / 850c3a220e

entry: report more colliding paths

In b878579ae7 (clone: report duplicate entries on case-insensitive filesystems, 2018-08-17) code was added to warn about index entries that resolve to the same file system entity (usually the cause is a case-insensitive filesystem). In Git for Windows, where inodes are not trusted (because of a performance trade-off, inodes are equal to 0 by default), that check does not compare inode numbers but the verbatim path. This logic works well when index entries' paths differ only in case. However, for file/directory conflicts only the file's path was reported, leaving the user puzzled with what that path collides. Let's try ot catch colliding paths even if one path is the prefix of the other. We do this also in setups where the file system is case-sensitive because the inode check would not be able to catch those collisions. While not a complete solution (for example, on macOS, Unicode normalization could also lead to file/directory conflicts but be missed by this logic), it is at least another defensive layer on top of what the previous commits added. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>

Johannes Schindelin committed Mar 28, 2024 at 09:44 UTC 850c3a220e7a0b1bf740fba9ac8f3f2b0486a1af

3 files changed +20 -1

dir.c

+12

@@ -88,6 +88,18 @@ int fspathncmp(const char *a, const char *b, size_t count)

88 return ignore_case ? strncasecmp(a, b, count) : strncmp(a, b, count);

89 }

91 +int paths_collide(const char *a, const char *b)

92 +{

93 + size_t len_a = strlen(a), len_b = strlen(b);

94 +

95 + if (len_a == len_b)

96 + return fspatheq(a, b);

97 +

98 + if (len_a < len_b)

99 + return is_dir_sep(b[len_a]) && !fspathncmp(a, b, len_a);

100 + return is_dir_sep(a[len_b]) && !fspathncmp(a, b, len_b);

101 +}

102 +

103 unsigned int fspathhash(const char *str)

104 {

105 return ignore_case ? strihash(str) : strhash(str);

dir.h

@@ -519,6 +519,13 @@ int fspatheq(const char *a, const char *b);

519 int fspathncmp(const char *a, const char *b, size_t count);

520 unsigned int fspathhash(const char *str);

521

522 +/*

523 + * Reports whether paths collide. This may be because the paths differ only in

524 + * case on a case-sensitive filesystem, or that one path refers to a symlink

525 + * that collides with one of the parent directories of the other.

526 + */

527 +int paths_collide(const char *a, const char *b);

528 +

529 /*

530 * The prefix part of pattern must not contains wildcards.

531 */

entry.c

+1 -1

@@ -454,7 +454,7 @@ static void mark_colliding_entries(const struct checkout *state,

454 continue;

455

456 if ((trust_ino && !match_stat_data(&dup->ce_stat_data, st)) ||

457 - (!trust_ino && !fspathcmp(ce->name, dup->name))) {

457 + paths_collide(ce->name, dup->name)) {

458 dup->ce_flags |= CE_MATCHED;

459 break;

460 }