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 "path.h"
  13 #include "utf8.h"
  14 #include "precompose_utf8.h"
  15
  16 typedef char *iconv_ibp;
  17 static const char *repo_encoding = "UTF-8";
  18 static const char *path_encoding = "UTF-8-MAC";
  19
  20 static size_t has_non_ascii(const char *s, size_t maxlen, size_t *strlen_c)
  21 {
  22         const uint8_t *ptr = (const uint8_t *)s;
  23         size_t strlen_chars = 0;
  24         size_t ret = 0;
  25
  26         if (!ptr || !*ptr)
  27                 return 0;
  28
  29         while (*ptr && maxlen) {
  30                 if (*ptr & 0x80)
  31                         ret++;
  32                 strlen_chars++;
  33                 ptr++;
  34                 maxlen--;
  35         }
  36         if (strlen_c)
  37                 *strlen_c = strlen_chars;
  38
  39         return ret;
  40 }
  41
  42
  43 void probe_utf8_pathname_composition(void)
  44 {
  45         struct strbuf path = STRBUF_INIT;
  46         static const char *auml_nfc = "\xc3\xa4";
  47         static const char *auml_nfd = "\x61\xcc\x88";
  48         int output_fd;
  49         if (precomposed_unicode != -1)
  50                 return; /* We found it defined in the global config, respect it */
  51         git_path_buf(&path, "%s", auml_nfc);
  52         output_fd = open(path.buf, O_CREAT|O_EXCL|O_RDWR, 0600);
  53         if (output_fd >= 0) {
  54                 close(output_fd);
  55                 git_path_buf(&path, "%s", auml_nfd);
  56                 precomposed_unicode = access(path.buf, R_OK) ? 0 : 1;
  57                 git_config_set("core.precomposeunicode",
  58                                precomposed_unicode ? "true" : "false");
  59                 git_path_buf(&path, "%s", auml_nfc);
  60                 if (unlink(path.buf))
  61                         die_errno(_("failed to unlink '%s'"), path.buf);
  62         }
  63         strbuf_release(&path);
  64 }
  65
  66 const char *precompose_string_if_needed(const char *in)
  67 {
  68         size_t inlen;
  69         size_t outlen;
  70         if (!in)
  71                 return NULL;
  72         if (has_non_ascii(in, (size_t)-1, &inlen)) {
  73                 iconv_t ic_prec;
  74                 char *out;
  75                 if (precomposed_unicode < 0)
  76                         git_config_get_bool("core.precomposeunicode", &precomposed_unicode);
  77                 if (precomposed_unicode != 1)
  78                         return in;
  79                 ic_prec = iconv_open(repo_encoding, path_encoding);
  80                 if (ic_prec == (iconv_t) -1)
  81                         return in;
  82
  83                 out = reencode_string_iconv(in, inlen, ic_prec, 0, &outlen);
  84                 if (out) {
  85                         if (outlen == inlen && !memcmp(in, out, outlen))
  86                                 free(out); /* no need to return indentical */
  87                         else
  88                                 in = out;
  89                 }
  90                 iconv_close(ic_prec);
  91
  92         }
  93         return in;
  94 }
  95
  96 const char *precompose_argv_prefix(int argc, const char **argv, const char *prefix)
  97 {
  98         int i = 0;
  99
 100         while (i < argc) {
 101                 argv[i] = precompose_string_if_needed(argv[i]);
 102                 i++;
 103         }
 104         return precompose_string_if_needed(prefix);
 105 }
 106
 107
 108 PREC_DIR *precompose_utf8_opendir(const char *dirname)
 109 {
 110         PREC_DIR *prec_dir = xmalloc(sizeof(PREC_DIR));
 111         prec_dir->dirent_nfc = xmalloc(sizeof(dirent_prec_psx));
 112         prec_dir->dirent_nfc->max_name_len = sizeof(prec_dir->dirent_nfc->d_name);
 113
 114         prec_dir->dirp = opendir(dirname);
 115         if (!prec_dir->dirp) {
 116                 free(prec_dir->dirent_nfc);
 117                 free(prec_dir);
 118                 return NULL;
 119         } else {
 120                 int ret_errno = errno;
 121                 prec_dir->ic_precompose = iconv_open(repo_encoding, path_encoding);
 122                 /* if iconv_open() fails, die() in readdir() if needed */
 123                 errno = ret_errno;
 124         }
 125
 126         return prec_dir;
 127 }
 128
 129 struct dirent_prec_psx *precompose_utf8_readdir(PREC_DIR *prec_dir)
 130 {
 131         struct dirent *res;
 132         res = readdir(prec_dir->dirp);
 133         if (res) {
 134                 size_t namelenz = strlen(res->d_name) + 1; /* \0 */
 135                 size_t new_maxlen = namelenz;
 136
 137                 int ret_errno = errno;
 138
 139                 if (new_maxlen > prec_dir->dirent_nfc->max_name_len) {
 140                         size_t new_len = sizeof(dirent_prec_psx) + new_maxlen -
 141                                 sizeof(prec_dir->dirent_nfc->d_name);
 142
 143                         prec_dir->dirent_nfc = xrealloc(prec_dir->dirent_nfc, new_len);
 144                         prec_dir->dirent_nfc->max_name_len = new_maxlen;
 145                 }
 146
 147                 prec_dir->dirent_nfc->d_ino  = res->d_ino;
 148                 prec_dir->dirent_nfc->d_type = res->d_type;
 149
 150                 if ((precomposed_unicode == 1) && has_non_ascii(res->d_name, (size_t)-1, NULL)) {
 151                         if (prec_dir->ic_precompose == (iconv_t)-1) {
 152                                 die("iconv_open(%s,%s) failed, but needed:\n"
 153                                                 "    precomposed unicode is not supported.\n"
 154                                                 "    If you want to use decomposed unicode, run\n"
 155                                                 "    \"git config core.precomposeunicode false\"\n",
 156                                                 repo_encoding, path_encoding);
 157                         } else {
 158                                 iconv_ibp       cp = (iconv_ibp)res->d_name;
 159                                 size_t inleft = namelenz;
 160                                 char *outpos = &prec_dir->dirent_nfc->d_name[0];
 161                                 size_t outsz = prec_dir->dirent_nfc->max_name_len;
 162                                 errno = 0;
 163                                 iconv(prec_dir->ic_precompose, &cp, &inleft, &outpos, &outsz);
 164                                 if (errno || inleft) {
 165                                         /*
 166                                          * iconv() failed and errno could be E2BIG, EILSEQ, EINVAL, EBADF
 167                                          * MacOS X avoids illegal byte sequences.
 168                                          * If they occur on a mounted drive (e.g. NFS) it is not worth to
 169                                          * die() for that, but rather let the user see the original name
 170                                         */
 171                                         namelenz = 0; /* trigger strlcpy */
 172                                 }
 173                         }
 174                 } else
 175                         namelenz = 0;
 176
 177                 if (!namelenz)
 178                         strlcpy(prec_dir->dirent_nfc->d_name, res->d_name,
 179                                                         prec_dir->dirent_nfc->max_name_len);
 180
 181                 errno = ret_errno;
 182                 return prec_dir->dirent_nfc;
 183         }
 184         return NULL;
 185 }
 186
 187
 188 int precompose_utf8_closedir(PREC_DIR *prec_dir)
 189 {
 190         int ret_value;
 191         int ret_errno;
 192         ret_value = closedir(prec_dir->dirp);
 193         ret_errno = errno;
 194         if (prec_dir->ic_precompose != (iconv_t)-1)
 195                 iconv_close(prec_dir->ic_precompose);
 196         free(prec_dir->dirent_nfc);
 197         free(prec_dir);
 198         errno = ret_errno;
 199         return ret_value;
 200 }