mirror of
https://github.com/mozilla/gecko-dev.git
synced 2024-10-31 22:25:30 +00:00
344 lines
10 KiB
C
344 lines
10 KiB
C
/* -*- Mode: C; tab-width: 4; indent-tabs-mode: nil; c-basic-offset: 2 -*-
|
|
*
|
|
* The contents of this file are subject to the Netscape Public License
|
|
* Version 1.0 (the "NPL"); you may not use this file except in
|
|
* compliance with the NPL. You may obtain a copy of the NPL at
|
|
* http://www.mozilla.org/NPL/
|
|
*
|
|
* Software distributed under the NPL is distributed on an "AS IS" basis,
|
|
* WITHOUT WARRANTY OF ANY KIND, either express or implied. See the NPL
|
|
* for the specific language governing rights and limitations under the
|
|
* NPL.
|
|
*
|
|
* The Initial Developer of this code under the NPL is Netscape
|
|
* Communications Corporation. Portions created by Netscape are
|
|
* Copyright (C) 1998 Netscape Communications Corporation. All Rights
|
|
* Reserved.
|
|
*/
|
|
/* dblower.c */
|
|
/*
|
|
This function contain the doublebyte to lower table which need to implement function in
|
|
intlcomp.c.
|
|
None of the function in this file should be made public
|
|
The interface function is INTL_GetDoubleByteToLowerMap() and it should only be called
|
|
inside libi18n
|
|
*/
|
|
#include "intlpriv.h"
|
|
#include "pintlcmp.h"
|
|
|
|
PRIVATE DoubleByteToLowerMap lower_lookup_none[]=
|
|
{
|
|
{ 0x00, 0x00, 0x00, 0x00, 0x00 } /* Terminator */
|
|
};
|
|
|
|
#ifndef XP_UNIX
|
|
PRIVATE DoubleByteToLowerMap lower_lookup_sjis[]=
|
|
{
|
|
/* Full-width Latin */
|
|
{ 0x82, 0x60, 0x7a, 0x82, 0x81 },
|
|
|
|
/* Map Full-width Hiragana to Full-width Katakana */
|
|
{ 0x82, 0x9f, 0x9f, 0x83, 0x41 }, /* A */
|
|
{ 0x82, 0xa0, 0xa0, 0x83, 0x41 },
|
|
|
|
{ 0x82, 0xa1, 0xa1, 0x83, 0x43 }, /* I */
|
|
{ 0x82, 0xa2, 0xa2, 0x83, 0x43 },
|
|
|
|
{ 0x82, 0xa3, 0xa3, 0x83, 0x45 }, /* U */
|
|
{ 0x82, 0xa4, 0xa4, 0x83, 0x45 },
|
|
|
|
{ 0x82, 0xa5, 0xa5, 0x83, 0x47 }, /* E */
|
|
{ 0x82, 0xa6, 0xa6, 0x83, 0x47 },
|
|
|
|
{ 0x82, 0xa7, 0xa7, 0x83, 0x49 }, /* O */
|
|
{ 0x82, 0xa8, 0xa8, 0x83, 0x49 },
|
|
|
|
{ 0x82, 0xa9, 0xc0, 0x83, 0x4a }, /* KA - DI */
|
|
|
|
{ 0x82, 0xc1, 0xc1, 0x83, 0x63 }, /* TU */
|
|
{ 0x82, 0xc2, 0xc2, 0x83, 0x63 },
|
|
|
|
{ 0x82, 0xc3, 0xdd, 0x83, 0x64 }, /* DU - MI */
|
|
{ 0x82, 0xde, 0xe0, 0x83, 0x80 }, /* MU - MO */
|
|
|
|
{ 0x82, 0xe1, 0xe1, 0x83, 0x84 }, /* YA */
|
|
{ 0x82, 0xe2, 0xe2, 0x83, 0x84 },
|
|
|
|
{ 0x82, 0xe3, 0xe3, 0x83, 0x86 }, /* YU */
|
|
{ 0x82, 0xe4, 0xe4, 0x83, 0x86 },
|
|
|
|
{ 0x82, 0xe5, 0xe5, 0x83, 0x88 }, /* YO */
|
|
{ 0x82, 0xe6, 0xe6, 0x83, 0x88 },
|
|
|
|
{ 0x82, 0xe7, 0xeb, 0x83, 0x89 }, /* RA - RO */
|
|
|
|
{ 0x82, 0xec, 0xec, 0x83, 0x8f }, /* WA */
|
|
{ 0x82, 0xed, 0xed, 0x83, 0x8f },
|
|
|
|
{ 0x82, 0xee, 0xf1, 0x83, 0x90 }, /* WI - N */
|
|
|
|
/* Map SMALL Full-width Katakana to Full-width Katakana */
|
|
{ 0x83, 0x40, 0x40, 0x83, 0x41 }, /* A */
|
|
{ 0x83, 0x42, 0x42, 0x83, 0x43 }, /* I */
|
|
{ 0x83, 0x43, 0x44, 0x83, 0x45 }, /* U */
|
|
{ 0x83, 0x46, 0x46, 0x83, 0x47 }, /* E */
|
|
{ 0x83, 0x48, 0x47, 0x83, 0x49 }, /* O */
|
|
{ 0x83, 0x62, 0x62, 0x83, 0x63 }, /* TU */
|
|
{ 0x83, 0x83, 0x83, 0x83, 0x84 }, /* YA */
|
|
{ 0x83, 0x85, 0x85, 0x83, 0x86 }, /* YU */
|
|
{ 0x83, 0x87, 0x87, 0x83, 0x88 }, /* YO */
|
|
{ 0x83, 0x8e, 0x8e, 0x83, 0x8f }, /* WA */
|
|
|
|
/* Full-width Greek */
|
|
{ 0x83, 0x9f, 0xb6, 0x83, 0xbf },
|
|
|
|
/* Full-width Cyrillic*/
|
|
{ 0x84, 0x40, 0x4e, 0x84, 0x70 }, /* Part 1 */
|
|
{ 0x84, 0x4f, 0x61, 0x84, 0x80 }, /* Part 2 */
|
|
|
|
{ 0x00, 0x00, 0x00, 0x00, 0x00 } /* Terminator */
|
|
};
|
|
#endif
|
|
/* Don't #ifdef XP_UNIX for lower_lookup_eucjp. It is also used by GB2312 */
|
|
PRIVATE DoubleByteToLowerMap lower_lookup_eucjp[]=
|
|
{
|
|
/* Half-width Katakana */
|
|
{ SS2, 0xa6, 0xa6, 0xa5, 0xf2 }, /* WO */
|
|
|
|
{ SS2, 0xa7, 0xa7, 0xa5, 0xa2 }, /* A */
|
|
{ SS2, 0xa8, 0xa8, 0xa5, 0xa4 }, /* I */
|
|
{ SS2, 0xa9, 0xa9, 0xa5, 0xa6 }, /* U */
|
|
{ SS2, 0xaa, 0xaa, 0xa5, 0xa8 }, /* E */
|
|
{ SS2, 0xab, 0xab, 0xa5, 0xaa }, /* O */
|
|
|
|
{ SS2, 0xac, 0xac, 0xa5, 0xe4 }, /* YA */
|
|
{ SS2, 0xad, 0xad, 0xa5, 0xe6 }, /* YU */
|
|
{ SS2, 0xae, 0xae, 0xa5, 0xe8 }, /* YO */
|
|
{ SS2, 0xaf, 0xaf, 0xa5, 0xc4 }, /* TU */
|
|
|
|
{ SS2, 0xb1, 0xb1, 0xa5, 0xa2 }, /* A */
|
|
{ SS2, 0xb2, 0xb2, 0xa5, 0xa4 }, /* I */
|
|
{ SS2, 0xb3, 0xb3, 0xa5, 0xa6 }, /* U */
|
|
{ SS2, 0xb4, 0xb4, 0xa5, 0xa8 }, /* E */
|
|
{ SS2, 0xb5, 0xb5, 0xa5, 0xaa }, /* O */
|
|
|
|
{ SS2, 0xb6, 0xb6, 0xa5, 0xab }, /* KA */
|
|
{ SS2, 0xb7, 0xb7, 0xa5, 0xad }, /* KI */
|
|
{ SS2, 0xb8, 0xb8, 0xa5, 0xaf }, /* KU */
|
|
{ SS2, 0xb9, 0xb9, 0xa5, 0xb1 }, /* KE */
|
|
{ SS2, 0xba, 0xba, 0xa5, 0xb3 }, /* KO */
|
|
|
|
{ SS2, 0xbb, 0xbb, 0xa5, 0xb5 }, /* SA */
|
|
{ SS2, 0xbc, 0xbc, 0xa5, 0xb7 }, /* SI */
|
|
{ SS2, 0xbd, 0xbd, 0xa5, 0xb9 }, /* SU */
|
|
{ SS2, 0xbe, 0xbe, 0xa5, 0xbb }, /* SE */
|
|
{ SS2, 0xbf, 0xbf, 0xa5, 0xbd }, /* SO */
|
|
|
|
{ SS2, 0xc0, 0xc0, 0xa5, 0xbf }, /* TA */
|
|
{ SS2, 0xc1, 0xc1, 0xa5, 0xc1 }, /* TI */
|
|
{ SS2, 0xc2, 0xc2, 0xa5, 0xc4 }, /* TU */
|
|
{ SS2, 0xc3, 0xc3, 0xa5, 0xc6 }, /* TE */
|
|
{ SS2, 0xc4, 0xc4, 0xa5, 0xc8 }, /* TO */
|
|
|
|
{ SS2, 0xc5, 0xc9, 0xa5, 0xca }, /* NA - NO */
|
|
|
|
{ SS2, 0xca, 0xca, 0xa5, 0xcf }, /* HA */
|
|
{ SS2, 0xcb, 0xcb, 0xa5, 0xd2 }, /* HI */
|
|
{ SS2, 0xcc, 0xcc, 0xa5, 0xd5 }, /* HU */
|
|
{ SS2, 0xcd, 0xcd, 0xa5, 0xd8 }, /* HE */
|
|
{ SS2, 0xce, 0xce, 0xa5, 0xdb }, /* HO */
|
|
|
|
{ SS2, 0xcf, 0xd3, 0xa5, 0xde }, /* MA - MO */
|
|
|
|
{ SS2, 0xd4, 0xd4, 0xa5, 0xe4 }, /* YA */
|
|
{ SS2, 0xd5, 0xd5, 0xa5, 0xe6 }, /* YU */
|
|
{ SS2, 0xd6, 0xd6, 0xa5, 0xe8 }, /* YO */
|
|
|
|
{ SS2, 0xd7, 0xdb, 0xa5, 0xe9 }, /* RA - RO */
|
|
|
|
{ SS2, 0xdc, 0xdc, 0xa5, 0xef }, /* WA */
|
|
|
|
{ SS2, 0xdd, 0xdd, 0xa5, 0xf3 }, /* N */
|
|
|
|
/* Full-width Latin */
|
|
{ 0xa3, 0xc1, 0xda, 0xa3, 0xe1 },
|
|
|
|
/* Map Full-width Hiragana to Full-width Katakana */
|
|
{ 0xa4, 0xa1, 0xa1, 0xa5, 0xa2 }, /* A */
|
|
{ 0xa4, 0xa2, 0xa2, 0xa5, 0xa2 },
|
|
|
|
{ 0xa4, 0xa3, 0xa3, 0xa5, 0xa4 }, /* I */
|
|
{ 0xa4, 0xa4, 0xa4, 0xa5, 0xa4 },
|
|
|
|
{ 0xa4, 0xa5, 0xa5, 0xa5, 0xa6 }, /* U */
|
|
{ 0xa4, 0xa6, 0xa6, 0xa5, 0xa6 },
|
|
|
|
{ 0xa4, 0xa7, 0xa7, 0xa5, 0xa8 }, /* E */
|
|
{ 0xa4, 0xa8, 0xa8, 0xa5, 0xa8 },
|
|
|
|
{ 0xa4, 0xa9, 0xa9, 0xa5, 0xaa }, /* O */
|
|
{ 0xa4, 0xaa, 0xaa, 0xa5, 0xaa },
|
|
|
|
{ 0xa4, 0xab, 0xc2, 0xa5, 0xab}, /* KA - DI */
|
|
|
|
{ 0xa4, 0xc3, 0xc3, 0xa5, 0xc4 }, /* TU */
|
|
{ 0xa4, 0xc4, 0xc4, 0xa5, 0xc4 },
|
|
|
|
{ 0xa4, 0xc5, 0xe2, 0xa5, 0xc5 }, /* DU - MO */
|
|
|
|
{ 0xa4, 0xe3, 0xe3, 0xa5, 0xe4 }, /* YA */
|
|
{ 0xa4, 0xe4, 0xe4, 0xa5, 0xe4 },
|
|
|
|
{ 0xa4, 0xe5, 0xe5, 0xa5, 0xe6 }, /* YU */
|
|
{ 0xa4, 0xe6, 0xe6, 0xa5, 0xe6 },
|
|
|
|
{ 0xa4, 0xe7, 0xe7, 0xa5, 0xe8 }, /* YO */
|
|
{ 0xa4, 0xe8, 0xe8, 0xa5, 0xe8 },
|
|
|
|
{ 0xa4, 0xe9, 0xed, 0xa5, 0xe9}, /* RA - RO */
|
|
|
|
{ 0xa4, 0xee, 0xee, 0xa5, 0xef }, /* WA */
|
|
{ 0xa4, 0xef, 0xef, 0xa5, 0xef },
|
|
|
|
{ 0xa4, 0xf0, 0xf3, 0xa5, 0xf0 }, /* WI - N */
|
|
|
|
/* Map SMALL Full-width Katakana to Full-width Katakana */
|
|
{ 0xa5, 0xa1, 0xa1, 0xa5, 0xa2 }, /* A */
|
|
{ 0xa5, 0xa3, 0xa3, 0xa5, 0xa4 }, /* I */
|
|
{ 0xa5, 0xa5, 0xa5, 0xa5, 0xa6 }, /* U */
|
|
{ 0xa5, 0xa7, 0xa7, 0xa5, 0xa8 }, /* E */
|
|
{ 0xa5, 0xa9, 0xa9, 0xa5, 0xaa }, /* O */
|
|
{ 0xa5, 0xc3, 0xc3, 0xa5, 0xc4 }, /* TU */
|
|
{ 0xa5, 0xe3, 0xe3, 0xa5, 0xe4 }, /* YA */
|
|
{ 0xa5, 0xe5, 0xe5, 0xa5, 0xe6 }, /* YU */
|
|
{ 0xa5, 0xe7, 0xe7, 0xa5, 0xe8 }, /* YO */
|
|
{ 0xa5, 0xee, 0xee, 0xa5, 0xef }, /* WA */
|
|
|
|
/* Full-width Greek */
|
|
{ 0xa6, 0xa1, 0xb8, 0xa6, 0xc1 },
|
|
/* Full-width Cyrillic*/
|
|
{ 0xa7, 0xa1, 0xc1, 0xa7, 0xd1 },
|
|
|
|
{ 0x00, 0x00, 0x00, 0x00, 0x00 } /* Terminator */
|
|
};
|
|
PRIVATE DoubleByteToLowerMap lower_lookup_big5[]=
|
|
{
|
|
/* Full-width Latin */
|
|
{ 0xa2, 0xcf, 0xe4, 0xa2, 0xe9 }, /* Part 1 A-V */
|
|
{ 0xa2, 0xe5, 0xe8, 0xa3, 0x40 }, /* Part 2 W-Z */
|
|
|
|
/* Full-width Greek */
|
|
{ 0xa3, 0x44, 0x5b, 0xa3, 0x5c },
|
|
|
|
{ 0x00, 0x00, 0x00, 0x00, 0x00 } /* Terminator */
|
|
};
|
|
PRIVATE DoubleByteToLowerMap lower_lookup_cns11643_1[]=
|
|
{
|
|
/* Roman Number */
|
|
{ 0xa4, 0xab, 0xb4, 0xa6, 0xb5 },
|
|
|
|
/* Full-width Latin */
|
|
{ 0xa4, 0xc1, 0xda, 0xa4, 0xdb },
|
|
|
|
/* Full-width Greek */
|
|
{ 0xa4, 0xf5, 0xfe, 0xa5, 0xaf }, /* Part 1 Alpha - kappa */
|
|
{ 0xa5, 0xa1, 0xae, 0xa5, 0xb9 }, /* Part 2 Lamda - Omega */
|
|
|
|
{ 0x00, 0x00, 0x00, 0x00, 0x00 } /* Terminator */
|
|
};
|
|
PRIVATE DoubleByteToLowerMap lower_lookup_ksc5601[]=
|
|
{
|
|
/* Full-width Latin */
|
|
{ 0xa3, 0xc1, 0xda, 0xa3, 0xe1 },
|
|
|
|
/* Full-width Roman Number */
|
|
{ 0xa5, 0xb0, 0xb9, 0xa5, 0xa1 },
|
|
|
|
/* Full-width Greek */
|
|
{ 0xa5, 0xc1, 0xd8, 0xa5, 0xe1 },
|
|
|
|
/* Map Full-width Hiragana to Full-width Katakana */
|
|
{ 0xaa, 0xa1, 0xa1, 0xab, 0xa2 }, /* A */
|
|
{ 0xaa, 0xa2, 0xa2, 0xab, 0xa2 },
|
|
|
|
{ 0xaa, 0xa3, 0xa3, 0xab, 0xa4 }, /* I */
|
|
{ 0xaa, 0xa4, 0xa4, 0xab, 0xa4 },
|
|
|
|
{ 0xaa, 0xa5, 0xa5, 0xab, 0xa6 }, /* U */
|
|
{ 0xaa, 0xa6, 0xa6, 0xab, 0xa6 },
|
|
|
|
{ 0xaa, 0xa7, 0xa7, 0xab, 0xa8 }, /* E */
|
|
{ 0xaa, 0xa8, 0xa8, 0xab, 0xa8 },
|
|
|
|
{ 0xaa, 0xa9, 0xa9, 0xab, 0xaa }, /* O */
|
|
{ 0xaa, 0xaa, 0xaa, 0xab, 0xaa },
|
|
|
|
{ 0xaa, 0xab, 0xc2, 0xab, 0xab}, /* KA - DI */
|
|
|
|
{ 0xaa, 0xc3, 0xc3, 0xab, 0xc4 }, /* TU */
|
|
{ 0xaa, 0xc4, 0xc4, 0xab, 0xc4 },
|
|
|
|
{ 0xaa, 0xc5, 0xe2, 0xab, 0xc5 }, /* DU - MO */
|
|
|
|
{ 0xaa, 0xe3, 0xe3, 0xab, 0xe4 }, /* YA */
|
|
{ 0xaa, 0xe4, 0xe4, 0xab, 0xe4 },
|
|
|
|
{ 0xaa, 0xe5, 0xe5, 0xab, 0xe6 }, /* YU */
|
|
{ 0xaa, 0xe6, 0xe6, 0xab, 0xe6 },
|
|
|
|
{ 0xaa, 0xe7, 0xe7, 0xab, 0xe8 }, /* YO */
|
|
{ 0xaa, 0xe8, 0xe8, 0xab, 0xe8 },
|
|
|
|
{ 0xaa, 0xe9, 0xed, 0xab, 0xe9}, /* RA - RO */
|
|
|
|
{ 0xaa, 0xee, 0xee, 0xab, 0xef }, /* WA */
|
|
{ 0xaa, 0xef, 0xef, 0xab, 0xef },
|
|
|
|
{ 0xaa, 0xf0, 0xf3, 0xab, 0xf0 }, /* WI - N */
|
|
|
|
/* Map SMALL Full-width Katakana to Full-width Katakana */
|
|
{ 0xab, 0xa1, 0xa1, 0xab, 0xa2 }, /* A */
|
|
{ 0xab, 0xa3, 0xa3, 0xab, 0xa4 }, /* I */
|
|
{ 0xab, 0xa5, 0xa5, 0xab, 0xa6 }, /* U */
|
|
{ 0xab, 0xa7, 0xa7, 0xab, 0xa8 }, /* E */
|
|
{ 0xab, 0xa9, 0xa9, 0xab, 0xaa }, /* O */
|
|
{ 0xab, 0xc3, 0xc3, 0xab, 0xc4 }, /* TU */
|
|
{ 0xab, 0xe3, 0xe3, 0xab, 0xe4 }, /* YA */
|
|
{ 0xab, 0xe5, 0xe5, 0xab, 0xe6 }, /* YU */
|
|
{ 0xab, 0xe7, 0xe7, 0xab, 0xe8 }, /* YO */
|
|
{ 0xab, 0xee, 0xee, 0xab, 0xef }, /* WA */
|
|
|
|
/* Full-width Cyrillic*/
|
|
{ 0xac, 0xa1, 0xc1, 0xac, 0xd1 },
|
|
|
|
{ 0x00, 0x00, 0x00, 0x00, 0x00 } /* Terminator */
|
|
};
|
|
|
|
MODULE_PRIVATE DoubleByteToLowerMap *INTL_GetDoubleByteToLowerMap(int16 csid)
|
|
{
|
|
switch(csid)
|
|
{
|
|
#ifndef XP_UNIX
|
|
case CS_SJIS:
|
|
return lower_lookup_sjis;
|
|
#else
|
|
case CS_EUCJP:
|
|
return lower_lookup_eucjp;
|
|
#endif
|
|
case CS_BIG5:
|
|
return lower_lookup_big5;
|
|
case CS_CNS_8BIT:
|
|
return lower_lookup_cns11643_1;
|
|
case CS_GB_8BIT:
|
|
return lower_lookup_eucjp; /* The to_lower mapping for GB 2312 and JIS0208 are exactly the same */
|
|
/* We just use the same table here. */
|
|
case CS_KSC_8BIT:
|
|
return lower_lookup_ksc5601;
|
|
default:
|
|
return lower_lookup_none;
|
|
}
|
|
}
|
|
|
|
|
|
|