compat/precompose_utf8.c

   1 /*
   2  * Converts filenames from decomposed unicode into precomposed unicode.
   3  * Used on MacOS X.
   4  */
   5
   6 #define PRECOMPOSE_UNICODE_C
   7
   8 #include "cache.h"
   9 #include "config.h"
  10 #include "environment.h"
  11 #include "gettext.h"
  12 #include "utf8.h"
  13 #include "precompose_utf8.h"
  14
  15 typedef char *iconv_ibp;
  16 static const char *repo_encoding = "UTF-8";
  17 static const char *path_encoding = "UTF-8-MAC";
  18
  19 static size_t has_non_ascii(const char *s, size_t maxlen, size_t *strlen_c)
  20 {
  21         const uint8_t *ptr = (const uint8_t *)s;
  22         size_t strlen_chars = 0;
  23         size_t ret = 0;
  24
  25         if (!ptr || !*ptr)
  26                 return 0;
  27
  28         while (*ptr && maxlen) {
  29                 if (*ptr & 0x80)
  30                         ret++;
  31                 strlen_chars++;
  32                 ptr++;
  33                 maxlen--;
  34         }
  35         if (strlen_c)
  36                 *strlen_c = strlen_chars;
  37
  38         return ret;
  39 }
  40
  41
  42 void probe_utf8_pathname_composition(void)
  43 {
  44         struct strbuf path = STRBUF_INIT;
  45         static const char *auml_nfc = "\xc3\xa4";
  46         static const char *auml_nfd = "\x61\xcc\x88";
  47         int output_fd;
  48         if (precomposed_unicode != -1)
  49                 return; /* We found it defined in the global config, respect it */
  50         git_path_buf(&path, "%s", auml_nfc);
  51         output_fd = open(path.buf, O_CREAT|O_EXCL|O_RDWR, 0600);
  52         if (output_fd >= 0) {
  53                 close(output_fd);
  54                 git_path_buf(&path, "%s", auml_nfd);
  55                 precomposed_unicode = access(path.buf, R_OK) ? 0 : 1;
  56                 git_config_set("core.precomposeunicode",
  57                                precomposed_unicode ? "true" : "false");
  58                 git_path_buf(&path, "%s", auml_nfc);
  59                 if (unlink(path.buf))
  60                         die_errno(_("failed to unlink '%s'"), path.buf);
  61         }
  62         strbuf_release(&path);
  63 }
  64
  65 const char *precompose_string_if_needed(const char *in)
  66 {
  67         size_t inlen;
  68         size_t outlen;
  69         if (!in)
  70                 return NULL;
  71         if (has_non_ascii(in, (size_t)-1, &inlen)) {
  72                 iconv_t ic_prec;
  73                 char *out;
  74                 if (precomposed_unicode < 0)
  75                         git_config_get_bool("core.precomposeunicode", &precomposed_unicode);
  76                 if (precomposed_unicode != 1)
  77                         return in;
  78                 ic_prec = iconv_open(repo_encoding, path_encoding);
  79                 if (ic_prec == (iconv_t) -1)
  80                         return in;
  81
  82                 out = reencode_string_iconv(in, inlen, ic_prec, 0, &outlen);
  83                 if (out) {
  84                         if (outlen == inlen && !memcmp(in, out, outlen))
  85                                 free(out); /* no need to return indentical */
  86                         else
  87                                 in = out;
  88                 }
  89                 iconv_close(ic_prec);
  90
  91         }
  92         return in;
  93 }
  94
  95 const char *precompose_argv_prefix(int argc, const char **argv, const char *prefix)
  96 {
  97         int i = 0;
  98
  99         while (i < argc) {
 100                 argv[i] = precompose_string_if_needed(argv[i]);
 101                 i++;
 102         }
 103         return precompose_string_if_needed(prefix);
 104 }
 105
 106
 107 PREC_DIR *precompose_utf8_opendir(const char *dirname)
 108 {
 109         PREC_DIR *prec_dir = xmalloc(sizeof(PREC_DIR));
 110         prec_dir->dirent_nfc = xmalloc(sizeof(dirent_prec_psx));
 111         prec_dir->dirent_nfc->max_name_len = sizeof(prec_dir->dirent_nfc->d_name);
 112
 113         prec_dir->dirp = opendir(dirname);
 114         if (!prec_dir->dirp) {
 115                 free(prec_dir->dirent_nfc);
 116                 free(prec_dir);
 117                 return NULL;
 118         } else {
 119                 int ret_errno = errno;
 120                 prec_dir->ic_precompose = iconv_open(repo_encoding, path_encoding);
 121                 /* if iconv_open() fails, die() in readdir() if needed */
 122                 errno = ret_errno;
 123         }
 124
 125         return prec_dir;
 126 }
 127
 128 struct dirent_prec_psx *precompose_utf8_readdir(PREC_DIR *prec_dir)
 129 {
 130         struct dirent *res;
 131         res = readdir(prec_dir->dirp);
 132         if (res) {
 133                 size_t namelenz = strlen(res->d_name) + 1; /* \0 */
 134                 size_t new_maxlen = namelenz;
 135
 136                 int ret_errno = errno;
 137
 138                 if (new_maxlen > prec_dir->dirent_nfc->max_name_len) {
 139                         size_t new_len = sizeof(dirent_prec_psx) + new_maxlen -
 140                                 sizeof(prec_dir->dirent_nfc->d_name);
 141
 142                         prec_dir->dirent_nfc = xrealloc(prec_dir->dirent_nfc, new_len);
 143                         prec_dir->dirent_nfc->max_name_len = new_maxlen;
 144                 }
 145
 146                 prec_dir->dirent_nfc->d_ino  = res->d_ino;
 147                 prec_dir->dirent_nfc->d_type = res->d_type;
 148
 149                 if ((precomposed_unicode == 1) && has_non_ascii(res->d_name, (size_t)-1, NULL)) {
 150                         if (prec_dir->ic_precompose == (iconv_t)-1) {
 151                                 die("iconv_open(%s,%s) failed, but needed:\n"
 152                                                 "    precomposed unicode is not supported.\n"
 153                                                 "    If you want to use decomposed unicode, run\n"
 154                                                 "    \"git config core.precomposeunicode false\"\n",
 155                                                 repo_encoding, path_encoding);
 156                         } else {
 157                                 iconv_ibp       cp = (iconv_ibp)res->d_name;
 158                                 size_t inleft = namelenz;
 159                                 char *outpos = &prec_dir->dirent_nfc->d_name[0];
 160                                 size_t outsz = prec_dir->dirent_nfc->max_name_len;
 161                                 errno = 0;
 162                                 iconv(prec_dir->ic_precompose, &cp, &inleft, &outpos, &outsz);
 163                                 if (errno || inleft) {
 164                                         /*
 165                                          * iconv() failed and errno could be E2BIG, EILSEQ, EINVAL, EBADF
 166                                          * MacOS X avoids illegal byte sequences.
 167                                          * If they occur on a mounted drive (e.g. NFS) it is not worth to
 168                                          * die() for that, but rather let the user see the original name
 169                                         */
 170                                         namelenz = 0; /* trigger strlcpy */
 171                                 }
 172                         }
 173                 } else
 174                         namelenz = 0;
 175
 176                 if (!namelenz)
 177                         strlcpy(prec_dir->dirent_nfc->d_name, res->d_name,
 178                                                         prec_dir->dirent_nfc->max_name_len);
 179
 180                 errno = ret_errno;
 181                 return prec_dir->dirent_nfc;
 182         }
 183         return NULL;
 184 }
 185
 186
 187 int precompose_utf8_closedir(PREC_DIR *prec_dir)
 188 {
 189         int ret_value;
 190         int ret_errno;
 191         ret_value = closedir(prec_dir->dirp);
 192         ret_errno = errno;
 193         if (prec_dir->ic_precompose != (iconv_t)-1)
 194                 iconv_close(prec_dir->ic_precompose);
 195         free(prec_dir->dirent_nfc);
 196         free(prec_dir);
 197         errno = ret_errno;
 198         return ret_value;
 199 }