[KoCo-CVS] [Commit] KoreanCodecs/src hangul.c Setup.in
Brought to you by:
perky
Menu
▾
▴
[KoCo-CVS] [Commit] KoreanCodecs/src hangul.c Setup.in
|
From: Chang <pe...@us...> - 2002-04-24 14:17:01
|
perky 02/04/24 07:16:56
Modified: src Setup.in
Added: src hangul.c
Log:
- Add ROUGH implementation of korean.c.hangul module
Revision Changes Path
1.2 +1 -0 KoreanCodecs/src/Setup.in
Index: Setup.in
===================================================================
RCS file: /cvsroot/koco/KoreanCodecs/src/Setup.in,v
retrieving revision 1.1
retrieving revision 1.2
diff -u -r1.1 -r1.2
--- Setup.in 14 Mar 2002 21:10:53 -0000 1.1
+++ Setup.in 24 Apr 2002 14:16:56 -0000 1.2
@@ -1,2 +1,3 @@
*shared*
_koco _koco.c
+hangul hangul.c
1.1 KoreanCodecs/src/hangul.c
Index: hangul.c
===================================================================
/*
* hangul.c
*
* KoreanCodecs Hangul Module C Implementation
*
* Author : Hye-Shik Chang <pe...@fa...>
* Date : $Date: 2002/04/24 14:16:56 $
* Created : 25 April 2002
*
* $Revision: 1.1 $
*/
static char *version =
"$Id: hangul.c,v 1.1 2002/04/24 14:16:56 perky Exp $";
#include "Python.h"
enum { /* Jaeum Codes on U+3100 */
G = 0x3131, GG, GS, N, NJ, NH, D, DD, L, LG, LM, LB,
LS, LT, LP, LH, M, B, BB, BS, S, SS, NG, J, JJ, C, K, T, P, H
};
enum { /* Moeum Codes on U+3100 */
A = 0x314f, AE, YA, YAE, EO, E, YEO, YE, O, WA, WAE, OE, YO,
U, WEO, WE, WI, YU, EU, YI, I
};
#define NCHOSUNG 19
#define NJUNGSUNG 21
#define NJONGSUNG 28
#define NJAEUM 30
#define NMOEUM 21
#define JAEUM_BOTTOM G
#define JAEUM_TOP H
#define MOEUM_BOTTOM A
#define MOEUM_TOP I
#define HANGUL_BOTTOM 0xac00
#define HANGUL_TOP 0xd7a3
#define JBASE_CHOSUNG 0x1100
#define JBASE_JUNGSUNG 0x1161
#define JBASE_JONGSUNG 0x11A8
#define CHOSUNG_FILLER 0x115f
#define JUNGSUNG_FILLER 0x1160
#define F_JAEUM 0x01
#define F_MOEUM 0x02
#define F_CHOSUNG 0x04
#define F_JUNGSUNG 0x08
#define F_JONGSUNG 0x10
#define MAX_MULTIJAMO 3
typedef struct _jamotype {
char *name;
Py_UNICODE code;
int multi[MAX_MULTIJAMO];
int flags;
} jamotype;
#define CODE(c) #c,c
#define NOMULTI {0,0,0}
#define JC (F_JAEUM | F_CHOSUNG)
#define JJ (F_JAEUM | F_JONGSUNG)
#define JCJ (F_JAEUM | F_CHOSUNG | F_JONGSUNG)
#define MJ (F_MOEUM | F_JUNGSUNG)
jamotype jamos[] = {
/* JAEUM */
{ CODE(G), NOMULTI, JCJ }, { CODE(GG), {G, G,}, JCJ }, { CODE(GS), {G, S,}, JJ },
{ CODE(N), NOMULTI, JCJ }, { CODE(NJ), {N, J,}, JJ }, { CODE(NH), {N, H,}, JJ },
{ CODE(D), NOMULTI, JCJ }, { CODE(DD), {D, D,}, JC }, { CODE(L), NOMULTI, JCJ },
{ CODE(LG), {L, G,}, JJ }, { CODE(LM), {L, M,}, JJ }, { CODE(LB), {L, B,}, JJ },
{ CODE(LS), {L, S,}, JJ }, { CODE(LT), {L, T,}, JJ }, { CODE(LP), {L, P,}, JJ },
{ CODE(LH), {L, H,}, JJ }, { CODE(M), NOMULTI, JCJ }, { CODE(B), NOMULTI, JCJ },
{ CODE(BB), {B, B,}, JC }, { CODE(BS), {B, S,}, JJ }, { CODE(S), NOMULTI, JCJ },
{ CODE(SS), {S, S,}, JCJ }, { CODE(NG), NOMULTI, JCJ }, { CODE(J), NOMULTI, JCJ },
{ CODE(JJ), {J, J,}, JC }, { CODE(C), NOMULTI, JCJ }, { CODE(K), NOMULTI, JCJ },
{ CODE(T), NOMULTI, JCJ }, { CODE(P), NOMULTI, JCJ }, { CODE(H), NOMULTI, JCJ },
/* MOEUM */
{ CODE(A), NOMULTI, MJ }, { CODE(AE), {A, I,}, MJ }, { CODE(YA), NOMULTI, MJ },
{ CODE(YAE), {YA,I}, MJ }, { CODE(EO), NOMULTI, MJ }, { CODE(E), NOMULTI, MJ },
{ CODE(YEO), NOMULTI, MJ }, { CODE(YE), {YEO,I}, MJ }, { CODE(O), NOMULTI, MJ },
{ CODE(WA), {O, A}, MJ }, { CODE(WAE), {O,A,I}, MJ }, { CODE(OE), {O, I}, MJ },
{ CODE(YO), NOMULTI, MJ }, { CODE(U), NOMULTI, MJ }, { CODE(WEO), {U, EO}, MJ },
{ CODE(WE), {U, E}, MJ }, { CODE(WI), {U, I}, MJ }, { CODE(YU), NOMULTI, MJ },
{ CODE(EU), NOMULTI, MJ }, { CODE(YI), {EU, I}, MJ }, { CODE(I), NOMULTI, MJ },
/* END MARKER */
{ 0, 0, NOMULTI, 0 },
};
#undef JC, JJ, JCJ, MJ, NOMULTI, CODE
#define isJaeum(c) (JAEUM_BOTTOM <= (c) && (c) <= JAEUM_TOP)
#define isMoeum(c) (MOEUM_BOTTOM <= (c) && (c) <= MOEUM_TOP)
static char Py_isJaeum__doc__[] = "isJaeum(code): Verify whether the code is Jaeum.";
static PyObject *
Py_isJaeum(PyObject *self, PyObject *args)
{
Py_UNICODE *code;
int codelen;
if (!PyArg_ParseTuple(args, "u#:isJaeum", &code, &codelen))
return NULL;
if (codelen < 1) {
PyErr_Format(PyExc_ValueError, "need not null unicode string");
return NULL;
}
if (isJaeum(*code)) {
Py_INCREF(Py_True);
return Py_True;
} else {
Py_INCREF(Py_False);
return Py_False;
}
}
static char Py_isMoeum__doc__[] = "isMoeum(code): Verify whether the code is Moeum.";
static PyObject *
Py_isMoeum(PyObject *self, PyObject *args)
{
Py_UNICODE *code;
int codelen;
if (!PyArg_ParseTuple(args, "u#:isMoeum", &code, &codelen))
return NULL;
if (codelen < 1) {
PyErr_Format(PyExc_ValueError, "need not null unicode string");
return NULL;
}
if (isMoeum(*code)) {
Py_INCREF(Py_True);
return Py_True;
} else {
Py_INCREF(Py_False);
return Py_False;
}
}
#if 0
static char cp949_encode__doc__[] = "CP949 encoder";
static PyObject *
cp949_encode(PyObject *self, PyObject *args)
{
Py_UNICODE *argptr, *srccur, *srcend;
int arglen, errtype = error_strict;
char *errors = NULL;
unsigned char *destptr, *destcur, *decbuf;
PyObject *r;
if (!PyArg_ParseTuple(args, "u#|z:cp949_encode", &argptr, &arglen, &errors))
return NULL;
errtype = error_type(errors);
if (errtype == error_undef)
return NULL;
destcur = destptr = PyMem_New(unsigned char, arglen*2+1);
for (srccur = argptr, srcend = argptr + arglen; srccur < srcend; srccur++) {
if (*srccur <= 0x7F)
*(destcur++) = *srccur;
else {
decbuf = _ksc5601_encode(*srccur);
if (!decbuf)
decbuf = _uhc_encode(*srccur);
if(decbuf == 0) {
switch (errtype) {
case error_strict:
PyMem_Del(destptr);
PyErr_Format(PyExc_UnicodeError,
"CP949 encoding error: invalid character \\u%04x",
*srccur);
return NULL;
break;
case error_replace:
*(destcur++) = 0xa1;
*(destcur++) = 0xa1;
break;
/* case error_ignore: break; */
}
} else {
*(destcur++) = decbuf[0];
*(destcur++) = decbuf[1];
}
}
}
r = codec_tuple(PyString_FromStringAndSize((char*)destptr, destcur - destptr), arglen);
PyMem_Del(destptr);
return r;
}
#endif
/* List of methods defined in the module */
#define meth(name, func, doc) {name, (PyCFunction)func, METH_VARARGS, doc}
static struct PyMethodDef hangul_methods[] = {
meth("isJaeum", Py_isJaeum, Py_isJaeum__doc__),
meth("isMoeum", Py_isMoeum, Py_isMoeum__doc__),
{NULL, NULL},
};
#define SET_INTCONSTANT(dict, value) \
PyDict_SetItemString(dict, #value, PyInt_FromLong((long) value))
#define SET_STRCONSTANT(dict, value) \
PyDict_SetItemString(dict, #value, PyString_FromString(value))
#define SET_CHARCONSTANT(dict, value) \
PyDict_SetItemString(dict, #value, PyString_FromFormat("%c", value))
/* Initialization function for the module */
void
inithangul(void)
{
PyObject *m, *d, *tmp;
Py_UNICODE tuni[2];
int i;
/* Create the module and add the functions */
m = Py_InitModule("hangul", hangul_methods);
/* Add some symbolic constants to the module */
d = PyModule_GetDict(m);
SET_INTCONSTANT(d, NCHOSUNG);
SET_INTCONSTANT(d, NJUNGSUNG);
SET_INTCONSTANT(d, NJONGSUNG);
{
PyObject *Chosung, *Jungsung, *Jongsung;
PyObject *Jaeum, *Moeum;
PyObject *JaeumDict, *MoeumDict;
PyObject *JaeumCodes, *MoeumCodes;
PyObject *JaeumMulti, *MoeumMulti;
int cur_cho, cur_jung, cur_jong;
int cur_jaeum, cur_moeum;
jamotype *jamo;
/* Bind Chosung, Jungsung, Jongsung lists */
cur_cho = cur_jung = cur_jong = 0;
Chosung = PyList_New(NCHOSUNG);
Jungsung = PyList_New(NJUNGSUNG);
Jongsung = PyList_New(NJONGSUNG);
PyDict_SetItemString(d, "Chosung", Chosung);
PyDict_SetItemString(d, "Jungsung", Jungsung);
PyDict_SetItemString(d, "Jongsung", Jongsung);
PyList_SET_ITEM(Jongsung, cur_jong++, PyUnicode_FromUnicode(NULL, 0));
/* Create Jaeum and Moeum meta class */
JaeumDict = PyDict_New();
MoeumDict = PyDict_New();
tmp = PyString_FromString("Jaeum");
Jaeum = PyClass_New(NULL, JaeumDict, tmp);
Py_DECREF(tmp);
tmp = PyString_FromString("Moeum");
Moeum = PyClass_New(NULL, MoeumDict, tmp);
Py_DECREF(tmp);
/* Bind meta class members */
PyDict_SetItemString(d, "Jaeum", Jaeum);
PyDict_SetItemString(d, "Moeum", Moeum);
PyDict_SetItemString(JaeumDict, "Chosung", Chosung);
PyDict_SetItemString(MoeumDict, "Jungsung", Jungsung);
PyDict_SetItemString(JaeumDict, "Jongsung", Jongsung);
/* Create Jaeum and Moeum Members */
JaeumCodes = PyTuple_New(NJAEUM);
MoeumCodes = PyTuple_New(NMOEUM);
JaeumMulti = PyDict_New();
MoeumMulti = PyDict_New();
cur_jaeum = cur_moeum = 0;
PyDict_SetItemString(JaeumDict, "Codes", JaeumCodes);
PyDict_SetItemString(MoeumDict, "Codes", MoeumCodes);
PyDict_SetItemString(JaeumDict, "Width", PyInt_FromLong(NJAEUM));
PyDict_SetItemString(MoeumDict, "Width", PyInt_FromLong(NMOEUM));
PyDict_SetItemString(JaeumDict, "MultiElement", JaeumMulti);
PyDict_SetItemString(MoeumDict, "MultiElement", MoeumMulti);
for (jamo = jamos; jamo->name; jamo++) {
PyObject *unijamo, *multicls;
int tuplen;
tuni[0] = jamo->code;
unijamo = PyUnicode_FromUnicode(tuni, 1);
PyDict_SetItemString(d, jamo->name, unijamo);
Py_INCREF(unijamo); /* PuTyple_SET_ITEM steals reference */
if (jamo->flags & F_JAEUM) {
PyTuple_SET_ITEM(JaeumCodes, cur_jaeum++, unijamo);
if (jamo->flags & F_CHOSUNG) {
PyList_SET_ITEM(Chosung, cur_cho++, unijamo);
PyDict_SetItemString(JaeumDict, jamo->name, unijamo);
}
if (jamo->flags & F_JONGSUNG) {
PyList_SET_ITEM(Jongsung, cur_jong++, unijamo);
PyDict_SetItemString(JaeumDict, jamo->name, unijamo);
}
multicls = JaeumMulti;
} else { /* Moeum */
PyTuple_SET_ITEM(MoeumCodes, cur_moeum++, unijamo);
if (jamo->flags & F_JUNGSUNG) {
PyList_SET_ITEM(Jungsung, cur_jung++, unijamo);
PyDict_SetItemString(MoeumDict, jamo->name, unijamo);
}
multicls = MoeumMulti;
}
if (jamo->multi[0]) {
tuplen = jamo->multi[2] ? 3 : 2;
tmp = PyTuple_New(tuplen);
for (i = 0; i < tuplen; i++) {
tuni[0] = jamo->multi[i];
PyTuple_SET_ITEM(tmp, i, PyUnicode_FromUnicode(tuni, 1));
}
PyDict_SetItem(multicls, unijamo, tmp);
Py_DECREF(tmp);
}
}
Py_DECREF(JaeumDict);
Py_DECREF(MoeumDict);
}
tmp = PyTuple_New(2);
tuni[0] = HANGUL_BOTTOM;
PyTuple_SET_ITEM(tmp, 0, PyUnicode_FromUnicode(tuni, 1));
tuni[0] = HANGUL_TOP;
PyTuple_SET_ITEM(tmp, 1, PyUnicode_FromUnicode(tuni, 1));
PyDict_SetItemString(d, "ZONE", tmp);
Py_DECREF(tmp);
tuni[0] = JBASE_CHOSUNG;
PyDict_SetItemString(d, "JBASE_CHOSUNG", PyUnicode_FromUnicode(tuni, 1));
tuni[0] = JBASE_JUNGSUNG;
PyDict_SetItemString(d, "JBASE_JUNGSUNG", PyUnicode_FromUnicode(tuni, 1));
tuni[0] = JBASE_JONGSUNG;
PyDict_SetItemString(d, "JBASE_JONGSUNG", PyUnicode_FromUnicode(tuni, 1));
tuni[0] = CHOSUNG_FILLER;
PyDict_SetItemString(d, "CHOSUNG_FILLER", PyUnicode_FromUnicode(tuni, 1));
tuni[0] = JUNGSUNG_FILLER;
PyDict_SetItemString(d, "JUNGSUNG_FILLER", PyUnicode_FromUnicode(tuni, 1));
PyDict_SetItemString(d, "version", PyString_FromString(version));
/* Check for errors */
if (PyErr_Occurred())
Py_FatalError("can't initialize the hangul module");
}
|
