conv.c

关系型数据库 Postgresql 6.5.2

/*
 * conversion between client encoding and server internal encoding
 * (currently mule internal code (mic) is used)
 * Tatsuo Ishii
 * WIN1250 client encoding support contributed by Pavel Behal
 *
 * $Id: conv.c,v 1.12 1999/07/11 22:47:20 ishii Exp $
 *
 *
 */

#include <stdio.h>
#include <string.h>

#include "mb/pg_wchar.h"

/*
 * convert bogus chars that cannot be represented in the current
encoding
 * system.
 */
static void
printBogusChar(unsigned char **mic, unsigned char **p)
{
	char		strbuf[16];
	int			l = pg_mic_mblen(*mic);

	*(*p)++ = '(';
	while (l--)
	{
		sprintf(strbuf, "%02x", *(*mic)++);
		*(*p)++ = strbuf[0];
		*(*p)++ = strbuf[1];
	}
	*(*p)++ = ')';
}

/*
 * SJIS ---> MIC
 */
static void
sjis2mic(unsigned char *sjis, unsigned char *p, int len)
{
	int			c1,
				c2;

	while (len > 0 && (c1 = *sjis++))
	{
		if (c1 >= 0xa1 && c1 <= 0xdf)
		{						/* 1 byte kana? */
			len--;
			*p++ = LC_JISX0201K;
			*p++ = c1;
		}
		else if (c1 > 0x7f)
		{						/* kanji? */
			c2 = *sjis++;
			len -= 2;
			*p++ = LC_JISX0208;
			*p++ = ((c1 & 0x3f) << 1) + 0x9f + (c2 > 0x9e);
			*p++ = c2 + ((c2 > 0x9e) ? 2 : 0x60) + (c2 < 0x80);
		}
		else
		{						/* should be ASCII */
			len--;
			*p++ = c1;
		}
	}
	*p = '\0';
}

/*
 * MIC ---> SJIS
 */
static void
mic2sjis(unsigned char *mic, unsigned char *p, int len)
{
	int			c1,
				c2;

	while (len > 0 && (c1 = *mic))
	{
		len -= pg_mic_mblen(mic++);

		if (c1 == LC_JISX0201K)
			*p++ = *mic++;
		else if (c1 == LC_JISX0208)
		{
			c1 = *mic++;
			c2 = *mic++;
			*p++ = ((c1 - 0xa1) >> 1) + ((c1 < 0xdf) ? 0x81 : 0xc1);
			*p++ = c2 - ((c1 & 1) ? ((c2 < 0xe0) ? 0x61 : 0x60) : 2);
		}
		else if (c1 > 0x7f)
		{						/* cannot convert to SJIS! */
			mic--;
			printBogusChar(&mic, &p);
		}
		else
		{						/* should be ASCII */
			*p++ = c1;
		}
	}
	*p = '\0';
}

/*
 * EUC_JP ---> MIC
 */
static void
euc_jp2mic(unsigned char *euc, unsigned char *p, int len)
{
	int			c1;

	while (len > 0 && (c1 = *euc++))
	{
		if (c1 == SS2)
		{						/* 1 byte kana? */
			len -= 2;
			*p++ = LC_JISX0201K;
			*p++ = *euc++;
		}
		else if (c1 == SS3)
		{						/* JIS X0212 kanji? */
			len -= 3;
			*p++ = LC_JISX0212;
			*p++ = *euc++;
			*p++ = *euc++;
		}
		else if (c1 & 0x80)
		{						/* kanji? */
			len -= 2;
			*p++ = LC_JISX0208;
			*p++ = c1;
			*p++ = *euc++;
		}
		else
		{						/* should be ASCII */
			len--;
			*p++ = c1;
		}
	}
	*p = '\0';
}

/*
 * MIC ---> EUC_JP
 */
static void
mic2euc_jp(unsigned char *mic, unsigned char *p, int len)
{
	int			c1;

	while (len > 0 && (c1 = *mic))
	{
		len -= pg_mic_mblen(mic++);

		if (c1 == LC_JISX0201K)
		{
			*p++ = SS2;
			*p++ = *mic++;
		}
		else if (c1 == LC_JISX0212)
		{
			*p++ = SS3;
			*p++ = *mic++;
			*p++ = *mic++;
		}
		else if (c1 == LC_JISX0208)
		{
			*p++ = *mic++;
			*p++ = *mic++;
		}
		else if (c1 > 0x7f)
		{						/* cannot convert to EUC_JP! */
			mic--;
			printBogusChar(&mic, &p);
		}
		else
		{						/* should be ASCII */
			*p++ = c1;
		}
	}
	*p = '\0';
}

/*
 * EUC_KR ---> MIC
 */
static void
euc_kr2mic(unsigned char *euc, unsigned char *p, int len)
{
	int			c1;

	while (len > 0 && (c1 = *euc++))
	{
		if (c1 & 0x80)
		{
			len -= 2;
			*p++ = LC_KS5601;
			*p++ = c1;
			*p++ = *euc++;
		}
		else
		{						/* should be ASCII */
			len--;
			*p++ = c1;
		}
	}
	*p = '\0';
}

/*
 * MIC ---> EUC_KR
 */
static void
mic2euc_kr(unsigned char *mic, unsigned char *p, int len)
{
	int			c1;

	while (len > 0 && (c1 = *mic))
	{
		len -= pg_mic_mblen(mic++);

		if (c1 == LC_KS5601)
		{
			*p++ = *mic++;
			*p++ = *mic++;
		}
		else if (c1 > 0x7f)
		{						/* cannot convert to EUC_KR! */
			mic--;
			printBogusChar(&mic, &p);
		}
		else
		{						/* should be ASCII */
			*p++ = c1;
		}
	}
	*p = '\0';
}

/*
 * EUC_CN ---> MIC
 */
static void
euc_cn2mic(unsigned char *euc, unsigned char *p, int len)
{
	int			c1;

	while (len > 0 && (c1 = *euc++))
	{
		if (c1 & 0x80)
		{
			len -= 2;
			*p++ = LC_GB2312_80;
			*p++ = c1;
			*p++ = *euc++;
		}
		else
		{						/* should be ASCII */
			len--;
			*p++ = c1;
		}
	}
	*p = '\0';
}

/*
 * MIC ---> EUC_CN
 */
static void
mic2euc_cn(unsigned char *mic, unsigned char *p, int len)
{
	int			c1;

	while (len > 0 && (c1 = *mic))
	{
		len -= pg_mic_mblen(mic++);

		if (c1 == LC_GB2312_80)
		{
			*p++ = *mic++;
			*p++ = *mic++;
		}
		else if (c1 > 0x7f)
		{						/* cannot convert to EUC_CN! */
			mic--;
			printBogusChar(&mic, &p);
		}
		else
		{						/* should be ASCII */
			*p++ = c1;
		}
	}
	*p = '\0';
}

/*
 * EUC_TW ---> MIC
 */
static void
euc_tw2mic(unsigned char *euc, unsigned char *p, int len)
{
	int			c1;

	while (len > 0 && (c1 = *euc++))
	{
		if (c1 == SS2)
		{
			len -= 4;
			c1 = *euc++;		/* plane No. */
			if (c1 == 0xa1)
				*p++ = LC_CNS11643_1;
			else if (c1 == 0xa2)
				*p++ = LC_CNS11643_2;
			else
			{
				*p++ = 0x9d;	/* LCPRV2 */
				*p++ = 0xa3 - c1 + LC_CNS11643_3;
			}
			*p++ = *euc++;
			*p++ = *euc++;
		}
		else if (c1 & 0x80)
		{						/* CNS11643-1 */
			len -= 2;
			*p++ = LC_CNS11643_1;
			*p++ = c1;
			*p++ = *euc++;
		}
		else
		{						/* should be ASCII */
			len--;
			*p++ = c1;
		}
	}
	*p = '\0';
}

/*
 * MIC ---> EUC_TW
 */
static void
mic2euc_tw(unsigned char *mic, unsigned char *p, int len)
{
	int			c1;

	while (len > 0 && (c1 = *mic))
	{
		len -= pg_mic_mblen(mic++);

		if (c1 == LC_CNS11643_1 || c1 == LC_CNS11643_2)
		{
			*p++ = *mic++;
			*p++ = *mic++;
		}
		else if (c1 == 0x9d)
		{						/* LCPRV2? */
			*p++ = SS2;
			*p++ = c1 - LC_CNS11643_3 + 0xa3;
			*p++ = *mic++;
			*p++ = *mic++;
		}
		else if (c1 > 0x7f)
		{						/* cannot convert to EUC_TW! */
			mic--;
			printBogusChar(&mic, &p);
		}
		else
		{						/* should be ASCII */
			*p++ = c1;
		}
	}
	*p = '\0';
}

/*
 * Big5 ---> MIC
 */
static void
big52mic(unsigned char *big5, unsigned char *p, int len)
{
	unsigned short c1;
	unsigned short big5buf,
				cnsBuf;
	unsigned char lc;
	char		bogusBuf[2];
	int			i;

	while (len > 0 && (c1 = *big5++))
	{
		if (c1 <= 0x7fU)
		{						/* ASCII */
			len--;
			*p++ = c1;
		}
		else
		{
			len -= 2;
			big5buf = c1 << 8;
			c1 = *big5++;
			big5buf |= c1;
			cnsBuf = BIG5toCNS(big5buf, &lc);
			if (lc != 0)
			{
				if (lc == LC_CNS11643_3 || lc == LC_CNS11643_4)
				{
					*p++ = 0x9d;/* LCPRV2 */
				}
				*p++ = lc;		/* Plane No. */
				*p++ = (cnsBuf >> 8) & 0x00ff;
				*p++ = cnsBuf & 0x00ff;
			}
			else
			{					/* cannot convert */
				big5 -= 2;
				*p++ = '(';
				for (i = 0; i < 2; i++)
				{
					sprintf(bogusBuf, "%02x", *big5++);
					*p++ = bogusBuf[0];
					*p++ = bogusBuf[1];
				}
				*p++ = ')';
			}
		}
	}
	*p = '\0';
}

/*
 * MIC ---> Big5
 */
static void
mic2big5(unsigned char *mic, unsigned char *p, int len)
{
	int			l;
	unsigned short c1;
	unsigned short big5buf,
				cnsBuf;

	while (len > 0 && (c1 = *mic))
	{
		l = pg_mic_mblen(mic++);
		len -= l;

		/* 0x9d means LCPRV2 */
		if (c1 == LC_CNS11643_1 || c1 == LC_CNS11643_2 || c1 == 0x9d)
		{
			if (c1 == 0x9d)
			{
				c1 = *mic++;	/* get plane no. */
			}
			cnsBuf = (*mic++) << 8;
			cnsBuf |= (*mic++) & 0x00ff;
			big5buf = CNStoBIG5(cnsBuf, c1);
			if (big5buf == 0)
			{					/* cannot convert to Big5! */
				mic -= l;
				printBogusChar(&mic, &p);
			}
			else
			{
				*p++ = (big5buf >> 8) & 0x00ff;
				*p++ = big5buf & 0x00ff;
			}
		}
		else if (c1 <= 0x7f)	/* ASCII */
			*p++ = c1;
		else
		{						/* cannot convert to Big5! */
			mic--;
			printBogusChar(&mic, &p);
		}
	}
	*p = '\0';
}

/*
 * LATINn ---> MIC
 */