mu/lib/mu-str-normalize.c

/*
** Copyright (C) 2010 Dirk-Jan C. Binnema <djcb@djcbsoftware.nl>
**
** This program is free software; you can redistribute it and/or modify
** it under the terms of the GNU General Public License as published by
** the Free Software Foundation; either version 3 of the License, or
** (at your option) any later version.
**
** This program is distributed in the hope that it will be useful,
** but WITHOUT ANY WARRANTY; without even the implied warranty of
** MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
** GNU General Public License for more details.
**
** You should have received a copy of the GNU General Public License
** along with this program; if not, write to the Free Software Foundation,
** Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
**
*/

#if HAVE_CONFIG_H
#include "config.h"
#endif /*HAVE_CONFIG_H*/


#include <glib.h>
#include <string.h>
#include <ctype.h>

#include "mu-str.h"


char*
mu_str_normalize (const char *str, gboolean downcase, GStringChunk *strchunk)
{
	char *mystr;

	g_return_val_if_fail (str, NULL);

	if (strchunk)
		mystr = g_string_chunk_insert (strchunk, str);
	else
		mystr = g_strdup (str);

	return mu_str_normalize_in_place_try (mystr, downcase, strchunk);
}


/* this implementation should work for _all_ locales. */
static char*
mu_str_normalize_in_place_generic (char *str, gboolean downcase, GStringChunk *strchunk)
{

	char *norm;
	size_t len;

	/* FIXME: add accent-folding etc. */
	if (!downcase)
		return str; /* nothing to do */

	len  = strlen (str);
	norm = g_utf8_strdown (str, len);


	if (strlen (norm) > len) {
		/* this case is rare, but does happen */
		char *copy;
		if (!strchunk)
			return norm;
		copy = g_string_chunk_insert (strchunk, norm);
		g_free (norm);
		return copy;
	}

	memcpy (str, norm, len);
	return str;
}


/*
 * this implementation works for accented chars in Unicode Blocks
 * 'Latin-1 Supplement' and 'Latin Extended-A'. An alternative (slower
 * but much simpler) implementation would be to use g_utf8_normalize
 * to decompose characters in the accent part and the character part,
 * and then get rid of the former. That would be slower than what we
 * do here, but also more *complete*.  It's unclear whether it would
 * be slower *in practice* => needs checking
 */

/* we can normalize in-place, as the normalized string will never be
 * longer than the original.  even for replacements that are 2 chars
 * wide (e.g. German ß => ss), the replacement is 2 bytes, like the
 * original 0xc3 0x9f
 *
 * note-to-self: http://www.geertvanderploeg.com/unicode-gen/
 */
char*
mu_str_normalize_in_place_try (char *str, gboolean downcase, GStringChunk *strchunk)
{
	const guchar *cur;
	int i;

	g_return_val_if_fail (str, NULL);

	if (*str == '\0')
		return str;

	for (i = 0, cur = (const guchar*)str; *cur; ++cur) {

		/* special case for plain-old ascii */
		if ((*cur < 0x80)) {
			str[i++] = downcase ? tolower (*cur) : *cur;
			continue;
		}

		if (*cur == 0xc3) { /* latin-1 supplement */
			++cur;
			switch (*cur) {

			case 0x80:
			case 0x81:
			case 0x82:
			case 0x83:
			case 0x84:
			case 0x85: str[i++] = downcase ? 'a' : 'A' ; break;

			case 0x86:
				str[i++] = downcase ? 'a' : 'A' ;
				str[i++] = 'e';
				break;

			case 0x87: str[i++] = downcase ? 'c' : 'C'; break;

			case 0x88:
			case 0x89:
			case 0x8a:
			case 0x8b:
				str[i++] = downcase ? 'e' : 'E';
				break;

			case 0x8c:
			case 0x8d:
			case 0x8e:
			case 0x8f: str[i++] = downcase ? 'i': 'I'; break;

			case 0x90: str[i++] = downcase ? 'd' : 'D'; break;
			case 0x91: str[i++] = downcase ? 'n' : 'N'; break;

			case 0x92:
			case 0x93:
			case 0x94:
			case 0x95:
			case 0x96: str[i++] = downcase ? 'o' : 'O'; break;

			case 0x99:
			case 0x9a:
			case 0x9b:
			case 0x9c: str[i++] = downcase ? 'u' : 'U'; break;

			case 0x9d: str[i++] = downcase ? 'y' : 'Y'; break;

			case 0x9e:
				str[i++] = downcase ? 't' : 'T';
				str[i++] = 'h';
				break;

			case 0x9f: str[i++] = 's'; str[i++] = 's'; break;

			case 0xa0:
			case 0xa1:
			case 0xa2:
			case 0xa3:
			case 0xa4:
			case 0xa5: str[i++] = 'a'; break;

			case 0xa6: str[i++] = 'a'; str[i++] = 'e'; break;
			case 0xa7: str[i++] = 'c'; break;

			case 0xa8:
			case 0xa9:
			case 0xaa:
			case 0xab: str[i++] = 'e'; break;

			case 0xac:
			case 0xad:
			case 0xae:
			case 0xaf: str[i++] = 'i'; break;

			case 0xb0: str[i++] = 'd'; break;
			case 0xb1: str[i++] = 'n'; break;

			case 0xb2:
			case 0xb3:
			case 0xb4:
			case 0xb5:
			case 0xb6: str[i++] = 'o'; break;

			case 0xb9:
			case 0xba:
			case 0xbb:
			case 0xbc: str[i++] = 'u'; break;

			case 0xbd: str[i++] = 'y'; break;
			case 0xbe: str[i++] = 't'; str[i++] = 'h'; break;
			case 0xbf: str[i++] = 'y'; break;

			default:
				str[i++] = *cur;
			}

		} else if (*cur == 0xc4) {  /* Latin Extended-A (0x04) */
			++cur;
			switch (*cur) {
			case 0x80:
			case 0x82:
			case 0x84: str[i++] = downcase ? 'a' : 'A'; break;

			case 0x86:
			case 0x88:
			case 0x8a:
			case 0x8c: str[i++] = downcase ? 'c' : 'C'; break;

			case 0x8e:
			case 0x90: str[i++] = downcase ? 'd' : 'D'; break;

			case 0x92:
			case 0x94:
			case 0x96:
			case 0x98:
			case 0x9a: str[i++] = downcase ? 'e' : 'E'; break;

			case 0x9c:
			case 0x9e:
			case 0xa0:
			case 0xa2: str[i++] = downcase ? 'g' : 'G'; break;

			case 0xa4:
			case 0xa6: str[i++] = downcase ? 'h' : 'H'; break;

			case 0xa8:
			case 0xaa:
			case 0xac:
			case 0xae:
			case 0xb0: str[i++] = downcase ? 'i' : 'I'; break;

			case 0xb2:
				str[i++] = downcase ? 'i' : 'I';
				str[i++] = downcase ? 'j' : 'J';
				break;


			case 0xb4: str[i++] = downcase ? 'j' : 'J'; break;

			case 0xb6: str[i++] = downcase ? 'k' : 'K'; break;

			case 0xb9:
			case 0xbb:
			case 0xbd:
			case 0xbf: str[i++] = downcase ? 'l': 'L'; break;

			case 0x81:
			case 0x83:
			case 0x85: str[i++] = 'a'; break;

			case 0x87:
			case 0x89:
			case 0x8b:
			case 0x8d: str[i++] = 'c'; break;

			case 0x8f:
			case 0x91: str[i++] = 'd'; break;

			case 0x93:
			case 0x95:
			case 0x97:
			case 0x99:
			case 0x9b: str[i++] = 'e'; break;

			case 0x9d:
			case 0x9f:
			case 0xa1:
			case 0xa: str[i++] = 'g'; break;

			case 0xa5:
			case 0xa7: str[i++] = 'h'; break;

			case 0xa9:
			case 0xab:
			case 0xad:
			case 0xaf:
			case 0xb1: str[i++] = 'i'; break;

			case 0xb3: str[i++] = 'i'; str[i++] = 'j'; break;

			case 0xb5: str[i++] = 'j'; break;

			case 0xb7:
			case 0xb8: str[i++] = 'k'; break;

			case 0xba:
			case 0xbc:
			case 0xbe: str[i++] = 'l'; break;

			default:   str[i++] = *cur; break;

			}

		} else if (*cur == 0xc5) { /* Latin Extended-A (0xc5) */
			++cur;
			switch (*cur) {
			case 0x81: str[i++] = downcase ? 'l': 'L'; break;

			case 0x83:
			case 0x85:
			case 0x87: str[i++] = downcase ? 'n': 'N'; break;

			case 0x8c:
			case 0x8e:
			case 0x90: str[i++] = downcase ? 'o': 'O'; break;

			case 0x92:
				str[i++] = downcase ? 'o':  'O';
				str[i++] = 'e';
				break;

			case 0x94:
			case 0x96:
			case 0x98: str[i++] = downcase ? 'r': 'R'; break;

			case 0x9a:
			case 0x9c:
			case 0x9e:
			case 0xa0: str[i++] = downcase ? 's': 'S'; break;

			case 0xa2:
			case 0xa4:
			case 0xa6: str[i++] = downcase ? 't': 'T'; break;

			case 0xa8:
			case 0xaa:
			case 0xac:
			case 0xae:
			case 0xb0:
			case 0xb2: str[i++] = downcase ? 'u': 'U'; break;
			case 0xb4: str[i++] = downcase ? 'w': 'W'; break;

			case 0xb6:
			case 0xb8: str[i++] = downcase ? 'y': 'Y'; break;

			case 0xb9:
			case 0xbb:
			case 0xbd: str[i++] = downcase ? 'z': 'Z'; break;

			case 0x80:
			case 0x82: str[i++] = 'l'; break;

			case 0x84:
			case 0x86:
			case 0x88:
			case 0x89:
			case 0x8a:
			case 0x8b: str[i++] = 'n'; break;

			case 0x8d:
			case 0x8f:
			case 0x91: str[i++] = 'o'; break;

			case 0x93: str[i++] = 'o'; str[i++] = 'e'; break;

			case 0x95:
			case 0x97:
			case 0x99: str[i++] = 'r'; break;

			case 0x9b:
			case 0x9d:
			case 0x9f:
			case 0xa1: str[i++] = 's'; break;

			case 0xa3:
			case 0xa5:
			case 0xa7: str[i++] = 't'; break;

			case 0xa9:
			case 0xab:
			case 0xad:
			case 0xaf:
			case 0xb1:
			case 0xb3: str[i++] = 'u'; break;

			case 0xb5: str[i++] = 'w'; break;

			case 0xb7: str[i++] = 'y'; break;

			case 0xba:
			case 0xbc:
			case 0xbe: str[i++] = 'z'; break;

			case 0xbf: str[i++] = 's'; break;

			default:   str[i++] = *cur; break;
			}
		} else {
			/* our fast-path for latin-utf8 does not work -- bummer!
			 * use something more generic (but a bit slower)
			 */
			return mu_str_normalize_in_place_generic (str, downcase, strchunk);
		}
	}

	str[i] = '\0';

	return str;
}