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 "git-compat-util.h"
   9 #include "config.h"
  10 #include "environment.h"
  11 #include "gettext.h"
  12 #include "path.h"
  13 #include "strbuf.h"
  14 #include "utf8.h"
  15 #include "precompose_utf8.h"
  16
  17 typedef char *iconv_ibp;
  18 static const char *repo_encoding = "UTF-8";
  19 static const char *path_encoding = "UTF-8-MAC";
  20
  21 static size_t has_non_ascii(const char *s, size_t maxlen, size_t *strlen_c)
  22 {
  23         const uint8_t *ptr = (const uint8_t *)s;
  24         size_t strlen_chars = 0;
  25         size_t ret = 0;
  26
  27         if (!ptr || !*ptr)
  28                 return 0;
  29
  30         while (*ptr && maxlen) {
  31                 if (*ptr & 0x80)
  32                         ret++;
  33                 strlen_chars++;
  34                 ptr++;
  35                 maxlen--;
  36         }
  37         if (strlen_c)
  38                 *strlen_c = strlen_chars;
  39
  40         return ret;
  41 }
  42
  43
  44 void probe_utf8_pathname_composition(void)
  45 {
  46         struct strbuf path = STRBUF_INIT;
  47         static const char *auml_nfc = "\xc3\xa4";
  48         static const char *auml_nfd = "\x61\xcc\x88";
  49         int output_fd;
  50         if (precomposed_unicode != -1)
  51                 return; /* We found it defined in the global config, respect it */
  52         git_path_buf(&path, "%s", auml_nfc);
  53         output_fd = open(path.buf, O_CREAT|O_EXCL|O_RDWR, 0600);
  54         if (output_fd >= 0) {
  55                 close(output_fd);
  56                 git_path_buf(&path, "%s", auml_nfd);
  57                 precomposed_unicode = access(path.buf, R_OK) ? 0 : 1;
  58                 git_config_set("core.precomposeunicode",
  59                                precomposed_unicode ? "true" : "false");
  60                 git_path_buf(&path, "%s", auml_nfc);
  61                 if (unlink(path.buf))
  62                         die_errno(_("failed to unlink '%s'"), path.buf);
  63         }
  64         strbuf_release(&path);
  65 }
  66
  67 const char *precompose_string_if_needed(const char *in)
  68 {
  69         size_t inlen;
  70         size_t outlen;
  71         if (!in)
  72                 return NULL;
  73         if (has_non_ascii(in, (size_t)-1, &inlen)) {
  74                 iconv_t ic_prec;
  75                 char *out;
  76                 if (precomposed_unicode < 0)
  77                         git_config_get_bool("core.precomposeunicode", &precomposed_unicode);
  78                 if (precomposed_unicode != 1)
  79                         return in;
  80                 ic_prec = iconv_open(repo_encoding, path_encoding);
  81                 if (ic_prec == (iconv_t) -1)
  82                         return in;
  83
  84                 out = reencode_string_iconv(in, inlen, ic_prec, 0, &outlen);
  85                 if (out) {
  86                         if (outlen == inlen && !memcmp(in, out, outlen))
  87                                 free(out); /* no need to return indentical */
  88                         else
  89                                 in = out;
  90                 }
  91                 iconv_close(ic_prec);
  92
  93         }
  94         return in;
  95 }
  96
  97 const char *precompose_argv_prefix(int argc, const char **argv, const char *prefix)
  98 {
  99         int i = 0;
 100
 101         while (i < argc) {
 102                 argv[i] = precompose_string_if_needed(argv[i]);
 103                 i++;
 104         }
 105         return precompose_string_if_needed(prefix);
 106 }
 107
 108
 109 PREC_DIR *precompose_utf8_opendir(const char *dirname)
 110 {
 111         PREC_DIR *prec_dir = xmalloc(sizeof(PREC_DIR));
 112         prec_dir->dirent_nfc = xmalloc(sizeof(dirent_prec_psx));
 113         prec_dir->dirent_nfc->max_name_len = sizeof(prec_dir->dirent_nfc->d_name);
 114
 115         prec_dir->dirp = opendir(dirname);
 116         if (!prec_dir->dirp) {
 117                 free(prec_dir->dirent_nfc);
 118                 free(prec_dir);
 119                 return NULL;
 120         } else {
 121                 int ret_errno = errno;
 122                 prec_dir->ic_precompose = iconv_open(repo_encoding, path_encoding);
 123                 /* if iconv_open() fails, die() in readdir() if needed */
 124                 errno = ret_errno;
 125         }
 126
 127         return prec_dir;
 128 }
 129
 130 struct dirent_prec_psx *precompose_utf8_readdir(PREC_DIR *prec_dir)
 131 {
 132         struct dirent *res;
 133         res = readdir(prec_dir->dirp);
 134         if (res) {
 135                 size_t namelenz = strlen(res->d_name) + 1; /* \0 */
 136                 size_t new_maxlen = namelenz;
 137
 138                 int ret_errno = errno;
 139
 140                 if (new_maxlen > prec_dir->dirent_nfc->max_name_len) {
 141                         size_t new_len = sizeof(dirent_prec_psx) + new_maxlen -
 142                                 sizeof(prec_dir->dirent_nfc->d_name);
 143
 144                         prec_dir->dirent_nfc = xrealloc(prec_dir->dirent_nfc, new_len);
 145                         prec_dir->dirent_nfc->max_name_len = new_maxlen;
 146                 }
 147
 148                 prec_dir->dirent_nfc->d_ino  = res->d_ino;
 149                 prec_dir->dirent_nfc->d_type = res->d_type;
 150
 151                 if ((precomposed_unicode == 1) && has_non_ascii(res->d_name, (size_t)-1, NULL)) {
 152                         if (prec_dir->ic_precompose == (iconv_t)-1) {
 153                                 die("iconv_open(%s,%s) failed, but needed:\n"
 154                                                 "    precomposed unicode is not supported.\n"
 155                                                 "    If you want to use decomposed unicode, run\n"
 156                                                 "    \"git config core.precomposeunicode false\"\n",
 157                                                 repo_encoding, path_encoding);
 158                         } else {
 159                                 iconv_ibp       cp = (iconv_ibp)res->d_name;
 160                                 size_t inleft = namelenz;
 161                                 char *outpos = &prec_dir->dirent_nfc->d_name[0];
 162                                 size_t outsz = prec_dir->dirent_nfc->max_name_len;
 163                                 errno = 0;
 164                                 iconv(prec_dir->ic_precompose, &cp, &inleft, &outpos, &outsz);
 165                                 if (errno || inleft) {
 166                                         /*
 167                                          * iconv() failed and errno could be E2BIG, EILSEQ, EINVAL, EBADF
 168                                          * MacOS X avoids illegal byte sequences.
 169                                          * If they occur on a mounted drive (e.g. NFS) it is not worth to
 170                                          * die() for that, but rather let the user see the original name
 171                                         */
 172                                         namelenz = 0; /* trigger strlcpy */
 173                                 }
 174                         }
 175                 } else
 176                         namelenz = 0;
 177
 178                 if (!namelenz)
 179                         strlcpy(prec_dir->dirent_nfc->d_name, res->d_name,
 180                                                         prec_dir->dirent_nfc->max_name_len);
 181
 182                 errno = ret_errno;
 183                 return prec_dir->dirent_nfc;
 184         }
 185         return NULL;
 186 }
 187
 188
 189 int precompose_utf8_closedir(PREC_DIR *prec_dir)
 190 {
 191         int ret_value;
 192         int ret_errno;
 193         ret_value = closedir(prec_dir->dirp);
 194         ret_errno = errno;
 195         if (prec_dir->ic_precompose != (iconv_t)-1)
 196                 iconv_close(prec_dir->ic_precompose);
 197         free(prec_dir->dirent_nfc);
 198         free(prec_dir);
 199         errno = ret_errno;
 200         return ret_value;
 201 }