precompose-utf8: fix spelling of "want" in error message
[git/git.git] / compat / precompose_utf8.c
CommitLineData
76759c7d
TB
1/*
2 * Converts filenames from decomposed unicode into precomposed unicode.
3 * Used on MacOS X.
5f8580ad 4 */
76759c7d
TB
5
6#define PRECOMPOSE_UNICODE_C
7
8#include "cache.h"
9#include "utf8.h"
10#include "precompose_utf8.h"
11
12typedef char *iconv_ibp;
5f8580ad
RR
13static const char *repo_encoding = "UTF-8";
14static const char *path_encoding = "UTF-8-MAC";
76759c7d 15
9a27f967 16static size_t has_non_ascii(const char *s, size_t maxlen, size_t *strlen_c)
76759c7d 17{
9a27f967 18 const uint8_t *ptr = (const uint8_t *)s;
76759c7d
TB
19 size_t strlen_chars = 0;
20 size_t ret = 0;
21
9a27f967 22 if (!ptr || !*ptr)
76759c7d 23 return 0;
76759c7d 24
9a27f967
JH
25 while (*ptr && maxlen) {
26 if (*ptr & 0x80)
76759c7d
TB
27 ret++;
28 strlen_chars++;
9a27f967 29 ptr++;
76759c7d
TB
30 maxlen--;
31 }
32 if (strlen_c)
33 *strlen_c = strlen_chars;
34
35 return ret;
36}
37
38
39void probe_utf8_pathname_composition(char *path, int len)
40{
5f8580ad
RR
41 static const char *auml_nfc = "\xc3\xa4";
42 static const char *auml_nfd = "\x61\xcc\x88";
76759c7d
TB
43 int output_fd;
44 if (precomposed_unicode != -1)
45 return; /* We found it defined in the global config, respect it */
76759c7d
TB
46 strcpy(path + len, auml_nfc);
47 output_fd = open(path, O_CREAT|O_EXCL|O_RDWR, 0600);
5f8580ad 48 if (output_fd >= 0) {
76759c7d 49 close(output_fd);
76759c7d
TB
50 strcpy(path + len, auml_nfd);
51 /* Indicate to the user, that we can configure it to true */
5f8580ad 52 if (!access(path, R_OK))
76759c7d 53 git_config_set("core.precomposeunicode", "false");
5f8580ad 54 /* To be backward compatible, set precomposed_unicode to 0 */
76759c7d 55 precomposed_unicode = 0;
76759c7d 56 strcpy(path + len, auml_nfc);
5f8580ad
RR
57 if (unlink(path))
58 die_errno(_("failed to unlink '%s'"), path);
76759c7d
TB
59 }
60}
61
62
63void precompose_argv(int argc, const char **argv)
64{
65 int i = 0;
66 const char *oldarg;
67 char *newarg;
68 iconv_t ic_precompose;
69
70 if (precomposed_unicode != 1)
71 return;
72
73 ic_precompose = iconv_open(repo_encoding, path_encoding);
74 if (ic_precompose == (iconv_t) -1)
75 return;
76
77 while (i < argc) {
78 size_t namelen;
79 oldarg = argv[i];
9a27f967 80 if (has_non_ascii(oldarg, (size_t)-1, &namelen)) {
76759c7d
TB
81 newarg = reencode_string_iconv(oldarg, namelen, ic_precompose);
82 if (newarg)
83 argv[i] = newarg;
84 }
85 i++;
86 }
87 iconv_close(ic_precompose);
88}
89
90
91PREC_DIR *precompose_utf8_opendir(const char *dirname)
92{
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);
96
97 prec_dir->dirp = opendir(dirname);
98 if (!prec_dir->dirp) {
99 free(prec_dir->dirent_nfc);
100 free(prec_dir);
101 return NULL;
102 } else {
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 */
106 errno = ret_errno;
107 }
108
109 return prec_dir;
110}
111
112struct dirent_prec_psx *precompose_utf8_readdir(PREC_DIR *prec_dir)
113{
114 struct dirent *res;
115 res = readdir(prec_dir->dirp);
116 if (res) {
117 size_t namelenz = strlen(res->d_name) + 1; /* \0 */
118 size_t new_maxlen = namelenz;
119
120 int ret_errno = errno;
121
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);
125
126 prec_dir->dirent_nfc = xrealloc(prec_dir->dirent_nfc, new_len);
127 prec_dir->dirent_nfc->max_name_len = new_maxlen;
128 }
129
130 prec_dir->dirent_nfc->d_ino = res->d_ino;
131 prec_dir->dirent_nfc->d_type = res->d_type;
132
9a27f967 133 if ((precomposed_unicode == 1) && has_non_ascii(res->d_name, (size_t)-1, NULL)) {
76759c7d
TB
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"
0f7b4c2e 137 " If you want to use decomposed unicode, run\n"
76759c7d
TB
138 " \"git config core.precomposeunicode false\"\n",
139 repo_encoding, path_encoding);
140 } else {
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;
145 size_t cnt;
146 errno = 0;
147 cnt = 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 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
154 */
155 namelenz = 0; /* trigger strlcpy */
156 }
157 }
5f8580ad 158 } else
76759c7d
TB
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
172int 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}