|
Packit |
aea12f |
/* Decomposition of Unicode characters.
|
|
Packit Service |
991b93 |
Copyright (C) 2009-2020 Free Software Foundation, Inc.
|
|
Packit |
aea12f |
Written by Bruno Haible <bruno@clisp.org>, 2009.
|
|
Packit |
aea12f |
|
|
Packit |
aea12f |
This program is free software: you can redistribute it and/or
|
|
Packit |
aea12f |
modify it under the terms of either:
|
|
Packit |
aea12f |
|
|
Packit |
aea12f |
* the GNU Lesser General Public License as published by the Free
|
|
Packit |
aea12f |
Software Foundation; either version 3 of the License, or (at your
|
|
Packit |
aea12f |
option) any later version.
|
|
Packit |
aea12f |
|
|
Packit |
aea12f |
or
|
|
Packit |
aea12f |
|
|
Packit |
aea12f |
* the GNU General Public License as published by the Free
|
|
Packit |
aea12f |
Software Foundation; either version 2 of the License, or (at your
|
|
Packit |
aea12f |
option) any later version.
|
|
Packit |
aea12f |
|
|
Packit |
aea12f |
or both in parallel, as here.
|
|
Packit |
aea12f |
This program is distributed in the hope that it will be useful,
|
|
Packit |
aea12f |
but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
Packit |
aea12f |
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
|
|
Packit |
aea12f |
Lesser General Public License for more details.
|
|
Packit |
aea12f |
|
|
Packit |
aea12f |
You should have received a copy of the GNU Lesser General Public License
|
|
Packit |
aea12f |
along with this program. If not, see <https://www.gnu.org/licenses/>. */
|
|
Packit |
aea12f |
|
|
Packit |
aea12f |
#include <config.h>
|
|
Packit |
aea12f |
|
|
Packit |
aea12f |
/* Specification. */
|
|
Packit |
aea12f |
#include "uninorm.h"
|
|
Packit |
aea12f |
|
|
Packit |
aea12f |
#include "uninorm/decomposition-table.h"
|
|
Packit |
aea12f |
|
|
Packit |
aea12f |
int
|
|
Packit |
aea12f |
uc_decomposition (ucs4_t uc, int *decomp_tag, ucs4_t *decomposition)
|
|
Packit |
aea12f |
{
|
|
Packit |
aea12f |
if (uc >= 0xAC00 && uc < 0xD7A4)
|
|
Packit |
aea12f |
{
|
|
Packit |
aea12f |
/* Hangul syllable. See Unicode standard, chapter 3, section
|
|
Packit |
aea12f |
"Hangul Syllable Decomposition", See also the clarification at
|
|
Packit Service |
991b93 |
<https://www.unicode.org/versions/Unicode5.1.0/>, section
|
|
Packit |
aea12f |
"Clarification of Hangul Jamo Handling". */
|
|
Packit |
aea12f |
unsigned int t;
|
|
Packit |
aea12f |
|
|
Packit |
aea12f |
uc -= 0xAC00;
|
|
Packit |
aea12f |
t = uc % 28;
|
|
Packit |
aea12f |
|
|
Packit |
aea12f |
*decomp_tag = UC_DECOMP_CANONICAL;
|
|
Packit |
aea12f |
if (t == 0)
|
|
Packit |
aea12f |
{
|
|
Packit |
aea12f |
unsigned int v, l;
|
|
Packit |
aea12f |
|
|
Packit |
aea12f |
uc = uc / 28;
|
|
Packit |
aea12f |
v = uc % 21;
|
|
Packit |
aea12f |
l = uc / 21;
|
|
Packit |
aea12f |
|
|
Packit |
aea12f |
decomposition[0] = 0x1100 + l;
|
|
Packit |
aea12f |
decomposition[1] = 0x1161 + v;
|
|
Packit |
aea12f |
return 2;
|
|
Packit |
aea12f |
}
|
|
Packit |
aea12f |
else
|
|
Packit |
aea12f |
{
|
|
Packit |
aea12f |
#if 1 /* Return the pairwise decomposition, not the full decomposition. */
|
|
Packit |
aea12f |
decomposition[0] = 0xAC00 + uc - t; /* = 0xAC00 + (l * 21 + v) * 28; */
|
|
Packit |
aea12f |
decomposition[1] = 0x11A7 + t;
|
|
Packit |
aea12f |
return 2;
|
|
Packit |
aea12f |
#else
|
|
Packit |
aea12f |
unsigned int v, l;
|
|
Packit |
aea12f |
|
|
Packit |
aea12f |
uc = uc / 28;
|
|
Packit |
aea12f |
v = uc % 21;
|
|
Packit |
aea12f |
l = uc / 21;
|
|
Packit |
aea12f |
|
|
Packit |
aea12f |
decomposition[0] = 0x1100 + l;
|
|
Packit |
aea12f |
decomposition[1] = 0x1161 + v;
|
|
Packit |
aea12f |
decomposition[2] = 0x11A7 + t;
|
|
Packit |
aea12f |
return 3;
|
|
Packit |
aea12f |
#endif
|
|
Packit |
aea12f |
}
|
|
Packit |
aea12f |
}
|
|
Packit |
aea12f |
else if (uc < 0x110000)
|
|
Packit |
aea12f |
{
|
|
Packit |
aea12f |
unsigned short entry = decomp_index (uc);
|
|
Packit |
aea12f |
if (entry != (unsigned short)(-1))
|
|
Packit |
aea12f |
{
|
|
Packit |
aea12f |
const unsigned char *p;
|
|
Packit |
aea12f |
unsigned int element;
|
|
Packit |
aea12f |
unsigned int length;
|
|
Packit |
aea12f |
|
|
Packit |
aea12f |
p = &gl_uninorm_decomp_chars_table[3 * (entry & 0x7FFF)];
|
|
Packit |
aea12f |
element = (p[0] << 16) | (p[1] << 8) | p[2];
|
|
Packit |
aea12f |
/* The first element has 5 bits for the decomposition type. */
|
|
Packit |
aea12f |
*decomp_tag = (element >> 18) & 0x1f;
|
|
Packit |
aea12f |
length = 1;
|
|
Packit |
aea12f |
for (;;)
|
|
Packit |
aea12f |
{
|
|
Packit |
aea12f |
/* Every element has an 18 bits wide Unicode code point. */
|
|
Packit |
aea12f |
*decomposition = element & 0x3ffff;
|
|
Packit |
aea12f |
/* Bit 23 tells whether there are more elements, */
|
|
Packit |
aea12f |
if ((element & (1 << 23)) == 0)
|
|
Packit |
aea12f |
break;
|
|
Packit |
aea12f |
p += 3;
|
|
Packit |
aea12f |
element = (p[0] << 16) | (p[1] << 8) | p[2];
|
|
Packit |
aea12f |
decomposition++;
|
|
Packit |
aea12f |
length++;
|
|
Packit |
aea12f |
}
|
|
Packit |
aea12f |
return length;
|
|
Packit |
aea12f |
}
|
|
Packit |
aea12f |
}
|
|
Packit |
aea12f |
return -1;
|
|
Packit |
aea12f |
}
|