5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrewsstatic char *rcsid = "$Id: unicode.c,v 1.1 2003/06/04 00:26:16 marka Exp $";
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews * Copyright (c) 2000,2001,2002 Japan Network Information Center.
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews * All rights reserved.
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews * By using this file, you agree to the terms and conditions set forth bellow.
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews * LICENSE TERMS AND CONDITIONS
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews * The following License Terms and Conditions apply, unless a different
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews * license is obtained from Japan Network Information Center ("JPNIC"),
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews * a Japanese association, Kokusai-Kougyou-Kanda Bldg 6F, 2-3-4 Uchi-Kanda,
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews * Chiyoda-ku, Tokyo 101-0047, Japan.
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews * 1. Use, Modification and Redistribution (including distribution of any
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews * modified or derived work) in source and/or binary forms is permitted
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews * under this License Terms and Conditions.
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews * 2. Redistribution of source code must retain the copyright notices as they
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews * appear in each source code file, this License Terms and Conditions.
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews * 3. Redistribution in binary form must reproduce the Copyright Notice,
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews * this License Terms and Conditions, in the documentation and/or other
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews * materials provided with the distribution. For the purposes of binary
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews * distribution the "Copyright Notice" refers to the following language:
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews * "Copyright (c) 2000-2002 Japan Network Information Center. All rights reserved."
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews * 4. The name of JPNIC may not be used to endorse or promote products
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews * derived from this Software without specific prior written approval of
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews * 5. Disclaimer/Limitation of Liability: THIS SOFTWARE IS PROVIDED BY JPNIC
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews * PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL JPNIC BE LIABLE
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews * BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY,
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews * WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews * OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews * ADVISED OF THE POSSIBILITY OF SUCH DAMAGES.
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews * Some constants for Hangul decomposition/composition.
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews#define SLast (SBase + LCount * VCount * TCount)
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews * Symbol composition macro.
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrewstypedef int (*unicode_canonclassproc)(unsigned long v);
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrewstypedef int (*unicode_decomposeproc)(unsigned long c,
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews const unsigned long **seqp);
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrewstypedef int (*unicode_composeproc)(unsigned long c,
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews TRACE(("idn__unicode_create(version=%-.50s)\n",
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews for (v = unicode_versions; v->version != NULL; v++) {
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrewsidn__unicode_destroy(idn__unicode_version_t version) {
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews /* Nothing to do */
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrewsidn__unicode_canonicalclass(idn__unicode_version_t version, unsigned long c) {
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrewsidn__unicode_decompose(idn__unicode_version_t version,
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews unsigned long c, int *decomp_lenp) {
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews assert(v != NULL && vlen >= 0 && decomp_lenp != NULL);
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews * First, check for Hangul.
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews if ((t_offset == 0 && vlen < 2) || (t_offset > 0 && vlen < 3))
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews * Look up decomposition table. If no decomposition is defined
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews * or if it is a compatibility decomosition when canonical
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews * decomposition requested, return 'idn_notfound'.
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews if (seqidx == 0 || (compat == 0 && (seqidx & DECOMP_COMPAT) != 0))
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews * Copy the decomposed sequence. The end of the sequence are
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews * marked with END_BIT.
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews unsigned long c;
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews /* Decompose recursively. */
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews r = idn__unicode_decompose(version, compat, v, vlen, c, &dlen);
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews } else if (r == idn_notfound) {
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrewsidn__unicode_iscompositecandidate(idn__unicode_version_t version,
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews unsigned long c) {
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews /* Check for Hangul */
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews if ((LBase <= c && c < LBase + LCount) || (SBase <= c && c < SLast))
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews * Look up composition table. If there are no composition
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews * that begins with the given character, it is not a
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews * composition candidate.
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrewsidn__unicode_compose(idn__unicode_version_t version, unsigned long c1,
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews * Check for Hangul.
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews * Hangul L and V.
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews ((c1 - LBase) * VCount + (c2 - VBase)) * TCount;
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews * Hangul LV and T.
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews * Look up composition table. If the result is 0, no composition
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews * is defined. Otherwise, upper 16bits of the result contains
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews * the number of composition that begins with 'c1', and the lower
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews * 16bits is the offset in 'compose_seq'.
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews if ((n = (*version->compose_proc)(c1, &cseq)) == 0)
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews * The composite sequences are sorted by the 2nd character 'c2'.
5c526acb82c882e41b655c31f5fa4425c87b671cMark Andrews * So we can use binary search.