+const char *
+uc_name (ucs4_t uc, char buffer[16])
+{
+ if (uc >= 0x20 && uc < 0x7f)
+ snprintf (buffer, 16, "`%c'", uc);
+ else
+ snprintf (buffer, 16, "U+%04X", uc);
+ return buffer;
+}
+\f
+/* UTF-8 functions that deal with uppercase/lowercase distinctions. */
+
+/* Returns a hash value for the N bytes of UTF-8 encoded data starting at S,
+ with lowercase and uppercase letters treated as equal, starting from
+ BASIS. */
+unsigned int
+utf8_hash_case_bytes (const char *s, size_t n, unsigned int basis)
+{
+ uint8_t folded_buf[2048];
+ size_t folded_len = sizeof folded_buf;
+ uint8_t *folded_s;
+ unsigned int hash;
+
+ folded_s = u8_casefold (CHAR_CAST (const uint8_t *, s), n,
+ NULL, UNINORM_NFKD, folded_buf, &folded_len);
+ if (folded_s != NULL)
+ {
+ hash = hash_bytes (folded_s, folded_len, basis);
+ if (folded_s != folded_buf)
+ free (folded_s);
+ }
+ else
+ {
+ if (errno == ENOMEM)
+ xalloc_die ();
+ hash = hash_bytes (s, n, basis);
+ }
+
+ return hash;
+}
+
+/* Returns a hash value for null-terminated UTF-8 string S, with lowercase and
+ uppercase letters treated as equal, starting from BASIS. */
+unsigned int
+utf8_hash_case_string (const char *s, unsigned int basis)
+{
+ return utf8_hash_case_bytes (s, strlen (s), basis);
+}
+
+/* Compares UTF-8 strings A and B case-insensitively.
+ Returns a negative value if A < B, zero if A == B, positive if A > B. */
+int
+utf8_strcasecmp (const char *a, const char *b)
+{
+ return utf8_strncasecmp (a, strlen (a), b, strlen (b));
+}
+
+/* Compares UTF-8 strings A (with length AN) and B (with length BN)
+ case-insensitively.
+ Returns a negative value if A < B, zero if A == B, positive if A > B. */
+int
+utf8_strncasecmp (const char *a, size_t an, const char *b, size_t bn)
+{
+ int result;
+
+ if (u8_casecmp (CHAR_CAST (const uint8_t *, a), an,
+ CHAR_CAST (const uint8_t *, b), bn,
+ NULL, UNINORM_NFKD, &result))
+ {
+ if (errno == ENOMEM)
+ xalloc_die ();
+
+ result = memcmp (a, b, MIN (an, bn));
+ if (result == 0)
+ result = an < bn ? -1 : an > bn;
+ }
+
+ return result;
+}
+
+static char *
+utf8_casemap (const char *s,
+ uint8_t *(*f) (const uint8_t *, size_t, const char *, uninorm_t,
+ uint8_t *, size_t *))
+{
+ char *result;
+ size_t size;
+
+ result = CHAR_CAST (char *,
+ f (CHAR_CAST (const uint8_t *, s), strlen (s) + 1,
+ NULL, NULL, NULL, &size));
+ if (result == NULL)
+ {
+ if (errno == ENOMEM)
+ xalloc_die ();
+
+ result = xstrdup (s);
+ }
+ return result;
+}
+
+char *
+utf8_to_upper (const char *s)
+{
+ return utf8_casemap (s, u8_toupper);
+}
+
+char *
+utf8_to_lower (const char *s)
+{
+ return utf8_casemap (s, u8_tolower);
+}
+\f
+bool
+get_encoding_info (struct encoding_info *e, const char *name)
+{
+ const struct substring in = SS_LITERAL_INITIALIZER (
+ "\t\n\v\f\r "
+ "!\"#$%&'()*+,-./0123456789:;<=>?@"
+ "ABCDEFGHIJKLMNOPQRSTUVWXYZ[\\]^_`"
+ "abcdefghijklmnopqrstuvwxyz{|}~");
+
+ struct substring out, cr, lf, space;
+ bool ok;
+
+ memset (e, 0, sizeof *e);
+
+ cr = recode_substring_pool (name, "UTF-8", ss_cstr ("\r"), NULL);
+ lf = recode_substring_pool (name, "UTF-8", ss_cstr ("\n"), NULL);
+ space = recode_substring_pool (name, "UTF-8", ss_cstr (" "), NULL);
+ ok = (cr.length >= 1
+ && cr.length <= MAX_UNIT
+ && cr.length == lf.length
+ && cr.length == space.length);
+ if (!ok)
+ {
+ fprintf (stderr, "warning: encoding `%s' is not supported.\n", name);
+ ss_dealloc (&cr);
+ ss_dealloc (&lf);
+ ss_dealloc (&space);
+ ss_alloc_substring (&cr, ss_cstr ("\r"));
+ ss_alloc_substring (&lf, ss_cstr ("\n"));
+ ss_alloc_substring (&space, ss_cstr (" "));
+ }
+
+ e->unit = cr.length;
+ memcpy (e->cr, cr.string, e->unit);
+ memcpy (e->lf, lf.string, e->unit);
+ memcpy (e->space, space.string, e->unit);
+
+ ss_dealloc (&cr);
+ ss_dealloc (&lf);
+ ss_dealloc (&space);
+
+ out = recode_substring_pool ("UTF-8", name, in, NULL);
+ e->is_ascii_compatible = ss_equals (in, out);
+ ss_dealloc (&out);
+
+ if (!e->is_ascii_compatible && e->unit == 1)
+ {
+ out = recode_substring_pool ("UTF-8", name, ss_cstr ("A"), NULL);
+ e->is_ebcdic_compatible = (out.length == 1
+ && (uint8_t) out.string[0] == 0xc1);
+ ss_dealloc (&out);
+ }
+ else
+ e->is_ebcdic_compatible = false;
+
+ return ok;
+}
+
+bool
+is_encoding_ascii_compatible (const char *encoding)
+{
+ struct encoding_info e;
+
+ get_encoding_info (&e, encoding);
+ return e.is_ascii_compatible;
+}
+
+bool
+is_encoding_ebcdic_compatible (const char *encoding)
+{
+ struct encoding_info e;
+
+ get_encoding_info (&e, encoding);
+ return e.is_ebcdic_compatible;
+}
+
+/* Returns true if iconv can convert ENCODING to and from UTF-8,
+ otherwise false. */
+bool
+is_encoding_supported (const char *encoding)
+{
+ return (create_iconv__ ("UTF-8", encoding)->conv != (iconv_t) -1
+ && create_iconv__ (encoding, "UTF-8")->conv != (iconv_t) -1);
+}
+
+/* Returns true if E is the name of a UTF-8 encoding.
+
+ XXX Possibly we should test not E as a string but its properties via
+ iconv. */
+bool
+is_encoding_utf8 (const char *e)
+{
+ return ((e[0] == 'u' || e[0] == 'U')
+ && (e[1] == 't' || e[1] == 'T')
+ && (e[2] == 'f' || e[2] == 'F')
+ && ((e[3] == '8' && e[4] == '\0')
+ || (e[3] == '-' && e[4] == '8' && e[5] == '\0')));
+}
+\f
+static struct encoding_category *categories;
+static int n_categories;
+
+static void SENTINEL (0)
+add_category (size_t *allocated_categories, const char *category, ...)
+{
+ struct encoding_category *c;
+ const char *encodings[16];
+ va_list args;
+ int i, n;
+
+ /* Count encoding arguments. */
+ va_start (args, category);
+ n = 0;
+ while ((encodings[n] = va_arg (args, const char *)) != NULL)
+ {
+ const char *encoding = encodings[n];
+ if (!strcmp (encoding, "Auto") || is_encoding_supported (encoding))
+ n++;
+ }
+ assert (n < sizeof encodings / sizeof *encodings);
+ va_end (args);
+
+ if (n == 0)
+ return;
+
+ if (n_categories >= *allocated_categories)
+ categories = x2nrealloc (categories,
+ allocated_categories, sizeof *categories);
+
+ c = &categories[n_categories++];
+ c->category = category;
+ c->encodings = xmalloc (n * sizeof *c->encodings);
+ for (i = 0; i < n; i++)
+ c->encodings[i] = encodings[i];
+ c->n_encodings = n;
+}
+
+static void
+init_encoding_categories (void)
+{
+ static bool inited;
+ size_t alloc;
+
+ if (inited)
+ return;
+ inited = true;
+
+ alloc = 0;
+ add_category (&alloc, "Unicode", "UTF-8", "UTF-16", "UTF-16BE", "UTF-16LE",
+ "UTF-32", "UTF-32BE", "UTF-32LE", NULL_SENTINEL);
+ add_category (&alloc, _("Arabic"), "IBM864", "ISO-8859-6", "Windows-1256",
+ NULL_SENTINEL);
+ add_category (&alloc, _("Armenian"), "ARMSCII-8", NULL_SENTINEL);
+ add_category (&alloc, _("Baltic"), "ISO-8859-13", "ISO-8859-4",
+ "Windows-1257", NULL_SENTINEL);
+ add_category (&alloc, _("Celtic"), "ISO-8859-14", NULL_SENTINEL);
+ add_category (&alloc, _("Central European"), "IBM852", "ISO-8859-2",
+ "Mac-CentralEurope", "Windows-1250", NULL_SENTINEL);
+ add_category (&alloc, _("Chinese Simplified"), "GB18030", "GB2312", "GBK",
+ "HZ-GB-2312", "ISO-2022-CN", NULL_SENTINEL);
+ add_category (&alloc, _("Chinese Traditional"), "Big5", "Big5-HKSCS",
+ "EUC-TW", NULL_SENTINEL);
+ add_category (&alloc, _("Croatian"), "MacCroatian", NULL_SENTINEL);
+ add_category (&alloc, _("Cyrillic"), "IBM855", "ISO-8859-5", "ISO-IR-111",
+ "KOI8-R", "MacCyrillic", NULL_SENTINEL);
+ add_category (&alloc, _("Cyrillic/Russian"), "IBM866", NULL_SENTINEL);
+ add_category (&alloc, _("Cyrillic/Ukrainian"), "KOI8-U", "MacUkrainian",
+ NULL_SENTINEL);
+ add_category (&alloc, _("Georgian"), "GEOSTD8", NULL_SENTINEL);
+ add_category (&alloc, _("Greek"), "ISO-8859-7", "MacGreek", NULL_SENTINEL);
+ add_category (&alloc, _("Gujarati"), "MacGujarati", NULL_SENTINEL);
+ add_category (&alloc, _("Gurmukhi"), "MacGurmukhi", NULL_SENTINEL);
+ add_category (&alloc, _("Hebrew"), "IBM862", "ISO-8859-8-I", "Windows-1255",
+ NULL_SENTINEL);
+ add_category (&alloc, _("Hebrew Visual"), "ISO-8859-8", NULL_SENTINEL);
+ add_category (&alloc, _("Hindi"), "MacDevangari", NULL_SENTINEL);
+ add_category (&alloc, _("Icelandic"), "MacIcelandic", NULL_SENTINEL);
+ add_category (&alloc, _("Japanese"), "EUC-JP", "ISO-2022-JP", "Shift_JIS",
+ NULL_SENTINEL);
+ add_category (&alloc, _("Korean"), "EUC-KR", "ISO-2022-KR", "JOHAB", "UHC",
+ NULL_SENTINEL);
+ add_category (&alloc, _("Nordic"), "ISO-8859-10", NULL_SENTINEL);
+ add_category (&alloc, _("Romanian"), "ISO-8859-16", "MacRomanian",
+ NULL_SENTINEL);
+ add_category (&alloc, _("South European"), "ISO-8859-3", NULL_SENTINEL);
+ add_category (&alloc, _("Thai"), "ISO-8859-11", "TIS-620", "Windows-874",
+ NULL_SENTINEL);
+ add_category (&alloc, _("Turkish"), "IBM857", "ISO-8859-9", "Windows-1254",
+ NULL_SENTINEL);
+ add_category (&alloc, _("Vietnamese"), "TVCN", "VISCII", "VPS",
+ "Windows-1258", NULL_SENTINEL);
+ add_category (&alloc, _("Western European"), "ISO-8859-1", "ISO-8859-15",
+ "Windows-1252", "IBM850", "MacRoman", NULL_SENTINEL);
+}
+
+/* Returns an array of "struct encoding_category" that contains only the
+ categories and encodings that the system supports. */
+struct encoding_category *
+get_encoding_categories (void)
+{
+ init_encoding_categories ();
+ return categories;
+}
+
+/* Returns the number of elements in the array returned by
+ get_encoding_categories(). */
+size_t
+get_n_encoding_categories (void)
+{
+ init_encoding_categories ();
+ return n_categories;
+}