2
* Converts filenames from decomposed unicode into precomposed unicode.
6
#define PRECOMPOSE_UNICODE_C
10
#include "precompose_utf8.h"
12
typedef char *iconv_ibp;
13
static const char *repo_encoding = "UTF-8";
14
static const char *path_encoding = "UTF-8-MAC";
16
static size_t has_non_ascii(const char *s, size_t maxlen, size_t *strlen_c)
18
const uint8_t *ptr = (const uint8_t *)s;
19
size_t strlen_chars = 0;
25
while (*ptr && maxlen) {
33
*strlen_c = strlen_chars;
39
void probe_utf8_pathname_composition(char *path, int len)
41
static const char *auml_nfc = "\xc3\xa4";
42
static const char *auml_nfd = "\x61\xcc\x88";
44
if (precomposed_unicode != -1)
45
return; /* We found it defined in the global config, respect it */
46
strcpy(path + len, auml_nfc);
47
output_fd = open(path, O_CREAT|O_EXCL|O_RDWR, 0600);
50
strcpy(path + len, auml_nfd);
51
/* Indicate to the user, that we can configure it to true */
52
if (!access(path, R_OK))
53
git_config_set("core.precomposeunicode", "false");
54
/* To be backward compatible, set precomposed_unicode to 0 */
55
precomposed_unicode = 0;
56
strcpy(path + len, auml_nfc);
58
die_errno(_("failed to unlink '%s'"), path);
63
void precompose_argv(int argc, const char **argv)
68
iconv_t ic_precompose;
70
if (precomposed_unicode != 1)
73
ic_precompose = iconv_open(repo_encoding, path_encoding);
74
if (ic_precompose == (iconv_t) -1)
80
if (has_non_ascii(oldarg, (size_t)-1, &namelen)) {
81
newarg = reencode_string_iconv(oldarg, namelen, ic_precompose, NULL);
87
iconv_close(ic_precompose);
91
PREC_DIR *precompose_utf8_opendir(const char *dirname)
93
PREC_DIR *prec_dir = xmalloc(sizeof(PREC_DIR));
94
prec_dir->dirent_nfc = xmalloc(sizeof(dirent_prec_psx));
95
prec_dir->dirent_nfc->max_name_len = sizeof(prec_dir->dirent_nfc->d_name);
97
prec_dir->dirp = opendir(dirname);
98
if (!prec_dir->dirp) {
99
free(prec_dir->dirent_nfc);
103
int ret_errno = errno;
104
prec_dir->ic_precompose = iconv_open(repo_encoding, path_encoding);
105
/* if iconv_open() fails, die() in readdir() if needed */
112
struct dirent_prec_psx *precompose_utf8_readdir(PREC_DIR *prec_dir)
115
res = readdir(prec_dir->dirp);
117
size_t namelenz = strlen(res->d_name) + 1; /* \0 */
118
size_t new_maxlen = namelenz;
120
int ret_errno = errno;
122
if (new_maxlen > prec_dir->dirent_nfc->max_name_len) {
123
size_t new_len = sizeof(dirent_prec_psx) + new_maxlen -
124
sizeof(prec_dir->dirent_nfc->d_name);
126
prec_dir->dirent_nfc = xrealloc(prec_dir->dirent_nfc, new_len);
127
prec_dir->dirent_nfc->max_name_len = new_maxlen;
130
prec_dir->dirent_nfc->d_ino = res->d_ino;
131
prec_dir->dirent_nfc->d_type = res->d_type;
133
if ((precomposed_unicode == 1) && has_non_ascii(res->d_name, (size_t)-1, NULL)) {
134
if (prec_dir->ic_precompose == (iconv_t)-1) {
135
die("iconv_open(%s,%s) failed, but needed:\n"
136
" precomposed unicode is not supported.\n"
137
" If you want to use decomposed unicode, run\n"
138
" \"git config core.precomposeunicode false\"\n",
139
repo_encoding, path_encoding);
141
iconv_ibp cp = (iconv_ibp)res->d_name;
142
size_t inleft = namelenz;
143
char *outpos = &prec_dir->dirent_nfc->d_name[0];
144
size_t outsz = prec_dir->dirent_nfc->max_name_len;
147
cnt = iconv(prec_dir->ic_precompose, &cp, &inleft, &outpos, &outsz);
148
if (errno || inleft) {
150
* iconv() failed and errno could be E2BIG, EILSEQ, EINVAL, EBADF
151
* MacOS X avoids illegal byte sequemces.
152
* If they occur on a mounted drive (e.g. NFS) it is not worth to
153
* die() for that, but rather let the user see the original name
155
namelenz = 0; /* trigger strlcpy */
162
strlcpy(prec_dir->dirent_nfc->d_name, res->d_name,
163
prec_dir->dirent_nfc->max_name_len);
166
return prec_dir->dirent_nfc;
172
int precompose_utf8_closedir(PREC_DIR *prec_dir)
176
ret_value = closedir(prec_dir->dirp);
178
if (prec_dir->ic_precompose != (iconv_t)-1)
179
iconv_close(prec_dir->ic_precompose);
180
free(prec_dir->dirent_nfc);