read-cache: optionally disallow HFS+ .git variants

The point of disallowing ".git" in the index is that we
would never want to accidentally overwrite files in the
repository directory. But this means we need to respect the
filesystem's idea of when two paths are equal. The prior
commit added a helper to make such a comparison for HFS+;
let's use it in verify_path.

We make this check optional for two reasons:

  1. It restricts the set of allowable filenames, which is
     unnecessary for people who are not on HFS+. In practice
     this probably doesn't matter, though, as the restricted
     names are rather obscure and almost certainly would
     never come up in practice.

  2. It has a minor performance penalty for every path we
     insert into the index.

This patch ties the check to the core.protectHFS config
option. Though this is expected to be most useful on OS X,
we allow it to be set everywhere, as HFS+ may be mounted on
other platforms. The variable does default to on for OS X,
though.

Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>

Documentation/config.txt

@@ -234,6 +234,11 @@ core.precomposeunicode::
 	When false, file names are handled fully transparent by Git,
 	which is backward compatible with older versions of Git.
 
+core.protectHFS::
+	If set to true, do not allow checkout of paths that would
+	be considered equivalent to `.git` on an HFS+ filesystem.
+	Defaults to `true` on Mac OS, and `false` elsewhere.
+
 core.trustctime::
 	If false, the ctime differences between the index and the
 	working tree are ignored; useful when the inode change time

cache.h

@@ -584,6 +584,7 @@ extern int fsync_object_files;
 extern int core_preload_index;
 extern int core_apply_sparse_checkout;
 extern int precomposed_unicode;
+extern int protect_hfs;
 
 /*
  * The character that begins a commented line in user-editable file

config.c

@@ -881,6 +881,11 @@ static int git_default_core_config(const char *var, const char *value)
 		return 0;
 	}
 
+	if (!strcmp(var, "core.protecthfs")) {
+		protect_hfs = git_config_bool(var, value);
+		return 0;
+	}
+
 	/* Add other config variables here and to Documentation/config.txt. */
 	return 0;
 }

config.mak.uname

@@ -97,6 +97,7 @@ ifeq ($(uname_S),Darwin)
 	HAVE_DEV_TTY = YesPlease
 	COMPAT_OBJS += compat/precompose_utf8.o
 	BASIC_CFLAGS += -DPRECOMPOSE_UNICODE
+	BASIC_CFLAGS += -DPROTECT_HFS_DEFAULT=1
 endif
 ifeq ($(uname_S),SunOS)
 	NEEDS_SOCKET = YesPlease

environment.c

@@ -63,6 +63,11 @@ int precomposed_unicode = -1; /* see probe_utf8_pathname_composition() */
 struct startup_info *startup_info;
 unsigned long pack_size_limit_cfg;
 
+#ifndef PROTECT_HFS_DEFAULT
+#define PROTECT_HFS_DEFAULT 0
+#endif
+int protect_hfs = PROTECT_HFS_DEFAULT;
+
 /*
  * The character that begins a commented line in user-editable file
  * that is subject to stripspace.

read-cache.c

@@ -14,6 +14,7 @@
 #include "resolve-undo.h"
 #include "strbuf.h"
 #include "varint.h"
+#include "utf8.h"
 
 static struct cache_entry *refresh_cache_entry(struct cache_entry *ce, int really);
 
@@ -786,6 +787,8 @@ int verify_path(const char *path)
 			return 1;
 		if (is_dir_sep(c)) {
 inside:
+			if (protect_hfs && is_hfs_dotgit(path))
+				return 0;
 			c = *path++;
 			if ((c == '.' && !verify_dotfile(path)) ||
 			    is_dir_sep(c) || c == '\0')

t/t1014-read-tree-confusing.sh

@@ -11,23 +11,39 @@ test_expect_success 'create base tree' '
 	tree=$(git rev-parse HEAD^{tree})
 '
 
-while read path; do
-	test_expect_success "reject $path at end of path" '
+test_expect_success 'enable core.protectHFS for rejection tests' '
+	git config core.protectHFS true
+'
+
+while read path pretty; do
+	: ${pretty:=$path}
+	test_expect_success "reject $pretty at end of path" '
 		printf "100644 blob %s\t%s" "$blob" "$path" >tree &&
 		bogus=$(git mktree <tree) &&
 		test_must_fail git read-tree $bogus
 	'
 
-	test_expect_success "reject $path as subtree" '
+	test_expect_success "reject $pretty as subtree" '
 		printf "040000 tree %s\t%s" "$tree" "$path" >tree &&
 		bogus=$(git mktree <tree) &&
 		test_must_fail git read-tree $bogus
 	'
-done <<-\EOF
+done <<-EOF
 .
 ..
 .git
 .GIT
+${u200c}.Git {u200c}.Git
+.gI${u200c}T .gI{u200c}T
+.GiT${u200c} .GiT{u200c}
 EOF
 
+test_expect_success 'utf-8 paths allowed with core.protectHFS off' '
+	test_when_finished "git read-tree HEAD" &&
+	test_config core.protectHFS false &&
+	printf "100644 blob %s\t%s" "$blob" ".gi${u200c}t" >tree &&
+	ok=$(git mktree <tree) &&
+	git read-tree $ok
+'
+
 test_done

t/test-lib.sh

@@ -154,7 +154,11 @@ _z40=0000000000000000000000000000000000000000
 LF='
 '
 
-export _x05 _x40 _z40 LF
+# UTF-8 ZERO WIDTH NON-JOINER, which HFS+ ignores
+# when case-folding filenames
+u200c=$(printf '\342\200\214')
+
+export _x05 _x40 _z40 LF u200c
 
 # Each test should start with something like this, after copyright notices:
 #