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 "utf8.h"
  11 #include "precompose_utf8.h"
  12
  13 typedef char *iconv_ibp;
  14 static const char *repo_encoding = "UTF-8";
  15 static const char *path_encoding = "UTF-8-MAC";
  16
  17 static size_t has_non_ascii(const char *s, size_t maxlen, size_t *strlen_c)
  18 {
  19         const uint8_t *ptr = (const uint8_t *)s;
  20         size_t strlen_chars = 0;
  21         size_t ret = 0;
  22
  23         if (!ptr || !*ptr)
  24                 return 0;
  25
  26         while (*ptr && maxlen) {
  27                 if (*ptr & 0x80)
  28                         ret++;
  29                 strlen_chars++;
  30                 ptr++;
  31                 maxlen--;
  32         }
  33         if (strlen_c)
  34                 *strlen_c = strlen_chars;
  35
  36         return ret;
  37 }
  38
  39
  40 void probe_utf8_pathname_composition(void)
  41 {
  42         struct strbuf path = STRBUF_INIT;
  43         static const char *auml_nfc = "\xc3\xa4";
  44         static const char *auml_nfd = "\x61\xcc\x88";
  45         int output_fd;
  46         if (precomposed_unicode != -1)
  47                 return; /* We found it defined in the global config, respect it */
  48         git_path_buf(&path, "%s", auml_nfc);
  49         output_fd = open(path.buf, O_CREAT|O_EXCL|O_RDWR, 0600);
  50         if (output_fd >= 0) {
  51                 close(output_fd);
  52                 git_path_buf(&path, "%s", auml_nfd);
  53                 precomposed_unicode = access(path.buf, R_OK) ? 0 : 1;
  54                 git_config_set("core.precomposeunicode",
  55                                precomposed_unicode ? "true" : "false");
  56                 git_path_buf(&path, "%s", auml_nfc);
  57                 if (unlink(path.buf))
  58                         die_errno(_("failed to unlink '%s'"), path.buf);
  59         }
  60         strbuf_release(&path);
  61 }
  62
  63
  64 void precompose_argv(int argc, const char **argv)
  65 {
  66         int i = 0;
  67         const char *oldarg;
  68         char *newarg;
  69         iconv_t ic_precompose;
  70
  71         if (precomposed_unicode != 1)
  72                 return;
  73
  74         ic_precompose = iconv_open(repo_encoding, path_encoding);
  75         if (ic_precompose == (iconv_t) -1)
  76                 return;
  77
  78         while (i < argc) {
  79                 size_t namelen;
  80                 oldarg = argv[i];
  81                 if (has_non_ascii(oldarg, (size_t)-1, &namelen)) {
  82                         newarg = reencode_string_iconv(oldarg, namelen, ic_precompose, 0, NULL);
  83                         if (newarg)
  84                                 argv[i] = newarg;
  85                 }
  86                 i++;
  87         }
  88         iconv_close(ic_precompose);
  89 }
  90
  91
  92 PREC_DIR *precompose_utf8_opendir(const char *dirname)
  93 {
  94         PREC_DIR *prec_dir = xmalloc(sizeof(PREC_DIR));
  95         prec_dir->dirent_nfc = xmalloc(sizeof(dirent_prec_psx));
  96         prec_dir->dirent_nfc->max_name_len = sizeof(prec_dir->dirent_nfc->d_name);
  97
  98         prec_dir->dirp = opendir(dirname);
  99         if (!prec_dir->dirp) {
 100                 free(prec_dir->dirent_nfc);
 101                 free(prec_dir);
 102                 return NULL;
 103         } else {
 104                 int ret_errno = errno;
 105                 prec_dir->ic_precompose = iconv_open(repo_encoding, path_encoding);
 106                 /* if iconv_open() fails, die() in readdir() if needed */
 107                 errno = ret_errno;
 108         }
 109
 110         return prec_dir;
 111 }
 112
 113 struct dirent_prec_psx *precompose_utf8_readdir(PREC_DIR *prec_dir)
 114 {
 115         struct dirent *res;
 116         res = readdir(prec_dir->dirp);
 117         if (res) {
 118                 size_t namelenz = strlen(res->d_name) + 1; /* \0 */
 119                 size_t new_maxlen = namelenz;
 120
 121                 int ret_errno = errno;
 122
 123                 if (new_maxlen > prec_dir->dirent_nfc->max_name_len) {
 124                         size_t new_len = sizeof(dirent_prec_psx) + new_maxlen -
 125                                 sizeof(prec_dir->dirent_nfc->d_name);
 126
 127                         prec_dir->dirent_nfc = xrealloc(prec_dir->dirent_nfc, new_len);
 128                         prec_dir->dirent_nfc->max_name_len = new_maxlen;
 129                 }
 130
 131                 prec_dir->dirent_nfc->d_ino  = res->d_ino;
 132                 prec_dir->dirent_nfc->d_type = res->d_type;
 133
 134                 if ((precomposed_unicode == 1) && has_non_ascii(res->d_name, (size_t)-1, NULL)) {
 135                         if (prec_dir->ic_precompose == (iconv_t)-1) {
 136                                 die("iconv_open(%s,%s) failed, but needed:\n"
 137                                                 "    precomposed unicode is not supported.\n"
 138                                                 "    If you want to use decomposed unicode, run\n"
 139                                                 "    \"git config core.precomposeunicode false\"\n",
 140                                                 repo_encoding, path_encoding);
 141                         } else {
 142                                 iconv_ibp       cp = (iconv_ibp)res->d_name;
 143                                 size_t inleft = namelenz;
 144                                 char *outpos = &prec_dir->dirent_nfc->d_name[0];
 145                                 size_t outsz = prec_dir->dirent_nfc->max_name_len;
 146                                 errno = 0;
 147                                 iconv(prec_dir->ic_precompose, &cp, &inleft, &outpos, &outsz);
 148                                 if (errno || inleft) {
 149                                         /*
 150                                          * iconv() failed and errno could be E2BIG, EILSEQ, EINVAL, EBADF
 151                                          * MacOS X avoids illegal byte sequences.
 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
 154                                         */
 155                                         namelenz = 0; /* trigger strlcpy */
 156                                 }
 157                         }
 158                 } else
 159                         namelenz = 0;
 160
 161                 if (!namelenz)
 162                         strlcpy(prec_dir->dirent_nfc->d_name, res->d_name,
 163                                                         prec_dir->dirent_nfc->max_name_len);
 164
 165                 errno = ret_errno;
 166                 return prec_dir->dirent_nfc;
 167         }
 168         return NULL;
 169 }
 170
 171
 172 int precompose_utf8_closedir(PREC_DIR *prec_dir)
 173 {
 174         int ret_value;
 175         int ret_errno;
 176         ret_value = closedir(prec_dir->dirp);
 177         ret_errno = errno;
 178         if (prec_dir->ic_precompose != (iconv_t)-1)
 179                 iconv_close(prec_dir->ic_precompose);
 180         free(prec_dir->dirent_nfc);
 181         free(prec_dir);
 182         errno = ret_errno;
 183         return ret_value;
 184 }