[apple/icu.git] / icuSources / common / cstring.c

/*
******************************************************************************
*
*   Copyright (C) 1997-2011, International Business Machines
*   Corporation and others.  All Rights Reserved.
*
******************************************************************************
*
* File CSTRING.C
*
* @author       Helena Shih
*
* Modification History:
*
*   Date        Name        Description
*   6/18/98     hshih       Created
*   09/08/98    stephen     Added include for ctype, for Mac Port
*   11/15/99    helena      Integrated S/390 IEEE changes. 
******************************************************************************
*/


#include <stdlib.h>
#include <stdio.h>
#include "unicode/utypes.h"
#include "cmemory.h"
#include "cstring.h"
#include "uassert.h"

/*
 * We hardcode case conversion for invariant characters to match our expectation
 * and the compiler execution charset.
 * This prevents problems on systems
 * - with non-default casing behavior, like Turkish system locales where
 *   tolower('I') maps to dotless i and toupper('i') maps to dotted I
 * - where there are no lowercase Latin characters at all, or using different
 *   codes (some old EBCDIC codepages)
 *
 * This works because the compiler usually runs on a platform where the execution
 * charset includes all of the invariant characters at their expected
 * code positions, so that the char * string literals in ICU code match
 * the char literals here.
 *
 * Note that the set of lowercase Latin letters is discontiguous in EBCDIC
 * and the set of uppercase Latin letters is discontiguous as well.
 */

U_CAPI UBool U_EXPORT2
uprv_isASCIILetter(char c) {
#if U_CHARSET_FAMILY==U_EBCDIC_FAMILY
    return
        ('a'<=c && c<='i') || ('j'<=c && c<='r') || ('s'<=c && c<='z') ||
        ('A'<=c && c<='I') || ('J'<=c && c<='R') || ('S'<=c && c<='Z');
#else
    return ('a'<=c && c<='z') || ('A'<=c && c<='Z');
#endif
}

U_CAPI char U_EXPORT2
uprv_toupper(char c) {
#if U_CHARSET_FAMILY==U_EBCDIC_FAMILY
    if(('a'<=c && c<='i') || ('j'<=c && c<='r') || ('s'<=c && c<='z')) {
        c=(char)(c+('A'-'a'));
    }
#else
    if('a'<=c && c<='z') {
        c=(char)(c+('A'-'a'));
    }
#endif
    return c;
}


#if 0
/*
 * Commented out because cstring.h defines uprv_tolower() to be
 * the same as either uprv_asciitolower() or uprv_ebcdictolower()
 * to reduce the amount of code to cover with tests.
 *
 * Note that this uprv_tolower() definition is likely to work for most
 * charset families, not just ASCII and EBCDIC, because its #else branch
 * is written generically.
 */
U_CAPI char U_EXPORT2
uprv_tolower(char c) {
#if U_CHARSET_FAMILY==U_EBCDIC_FAMILY
    if(('A'<=c && c<='I') || ('J'<=c && c<='R') || ('S'<=c && c<='Z')) {
        c=(char)(c+('a'-'A'));
    }
#else
    if('A'<=c && c<='Z') {
        c=(char)(c+('a'-'A'));
    }
#endif
    return c;
}
#endif

U_CAPI char U_EXPORT2
uprv_asciitolower(char c) {
    if(0x41<=c && c<=0x5a) {
        c=(char)(c+0x20);
    }
    return c;
}

U_CAPI char U_EXPORT2
uprv_ebcdictolower(char c) {
    if( (0xc1<=(uint8_t)c && (uint8_t)c<=0xc9) ||
        (0xd1<=(uint8_t)c && (uint8_t)c<=0xd9) ||
        (0xe2<=(uint8_t)c && (uint8_t)c<=0xe9)
    ) {
        c=(char)(c-0x40);
    }
    return c;
}


U_CAPI char* U_EXPORT2
T_CString_toLowerCase(char* str)
{
    char* origPtr = str;

    if (str) {
        do
            *str = (char)uprv_tolower(*str);
        while (*(str++));
    }

    return origPtr;
}

U_CAPI char* U_EXPORT2
T_CString_toUpperCase(char* str)
{
    char* origPtr = str;

    if (str) {
        do
            *str = (char)uprv_toupper(*str);
        while (*(str++));
    }

    return origPtr;
}

/*
 * Takes a int32_t and fills in  a char* string with that number "radix"-based.
 * Does not handle negative values (makes an empty string for them).
 * Writes at most 12 chars ("-2147483647" plus NUL).
 * Returns the length of the string (not including the NUL).
 */
U_CAPI int32_t U_EXPORT2
T_CString_integerToString(char* buffer, int32_t v, int32_t radix)
{
    char      tbuf[30];
    int32_t   tbx    = sizeof(tbuf);
    uint8_t   digit;
    int32_t   length = 0;
    uint32_t  uval;
    
    U_ASSERT(radix>=2 && radix<=16);
    uval = (uint32_t) v;
    if(v<0 && radix == 10) {
        /* Only in base 10 do we conside numbers to be signed. */
        uval = (uint32_t)(-v); 
        buffer[length++] = '-';
    }
    
    tbx = sizeof(tbuf)-1;
    tbuf[tbx] = 0;   /* We are generating the digits backwards.  Null term the end. */
    do {
        digit = (uint8_t)(uval % radix);
        tbuf[--tbx] = (char)(T_CString_itosOffset(digit));
        uval  = uval / radix;
    } while (uval != 0);
    
    /* copy converted number into user buffer  */
    uprv_strcpy(buffer+length, tbuf+tbx);
    length += sizeof(tbuf) - tbx -1;
    return length;
}


/*
 * Takes a int64_t and fills in  a char* string with that number "radix"-based.
 * Writes at most 21: chars ("-9223372036854775807" plus NUL).
 * Returns the length of the string, not including the terminating NULL.
 */
U_CAPI int32_t U_EXPORT2
T_CString_int64ToString(char* buffer, int64_t v, uint32_t radix)
{
    char      tbuf[30];
    int32_t   tbx    = sizeof(tbuf);
    uint8_t   digit;
    int32_t   length = 0;
    uint64_t  uval;
    
    U_ASSERT(radix>=2 && radix<=16);
    uval = (uint64_t) v;
    if(v<0 && radix == 10) {
        /* Only in base 10 do we conside numbers to be signed. */
        uval = (uint64_t)(-v); 
        buffer[length++] = '-';
    }
    
    tbx = sizeof(tbuf)-1;
    tbuf[tbx] = 0;   /* We are generating the digits backwards.  Null term the end. */
    do {
        digit = (uint8_t)(uval % radix);
        tbuf[--tbx] = (char)(T_CString_itosOffset(digit));
        uval  = uval / radix;
    } while (uval != 0);
    
    /* copy converted number into user buffer  */
    uprv_strcpy(buffer+length, tbuf+tbx);
    length += sizeof(tbuf) - tbx -1;
    return length;
}


U_CAPI int32_t U_EXPORT2
T_CString_stringToInteger(const char *integerString, int32_t radix)
{
    char *end;
    return uprv_strtoul(integerString, &end, radix);

}

U_CAPI int U_EXPORT2
uprv_stricmp(const char *str1, const char *str2) {
    if(str1==NULL) {
        if(str2==NULL) {
            return 0;
        } else {
            return -1;
        }
    } else if(str2==NULL) {
        return 1;
    } else {
        /* compare non-NULL strings lexically with lowercase */
        int rc;
        unsigned char c1, c2;

        for(;;) {
            c1=(unsigned char)*str1;
            c2=(unsigned char)*str2;
            if(c1==0) {
                if(c2==0) {
                    return 0;
                } else {
                    return -1;
                }
            } else if(c2==0) {
                return 1;
            } else {
                /* compare non-zero characters with lowercase */
                rc=(int)(unsigned char)uprv_tolower(c1)-(int)(unsigned char)uprv_tolower(c2);
                if(rc!=0) {
                    return rc;
                }
            }
            ++str1;
            ++str2;
        }
    }
}

U_CAPI int U_EXPORT2
uprv_strnicmp(const char *str1, const char *str2, uint32_t n) {
    if(str1==NULL) {
        if(str2==NULL) {
            return 0;
        } else {
            return -1;
        }
    } else if(str2==NULL) {
        return 1;
    } else {
        /* compare non-NULL strings lexically with lowercase */
        int rc;
        unsigned char c1, c2;

        for(; n--;) {
            c1=(unsigned char)*str1;
            c2=(unsigned char)*str2;
            if(c1==0) {
                if(c2==0) {
                    return 0;
                } else {
                    return -1;
                }
            } else if(c2==0) {
                return 1;
            } else {
                /* compare non-zero characters with lowercase */
                rc=(int)(unsigned char)uprv_tolower(c1)-(int)(unsigned char)uprv_tolower(c2);
                if(rc!=0) {
                    return rc;
                }
            }
            ++str1;
            ++str2;
        }
    }

    return 0;
}

U_CAPI char* U_EXPORT2
uprv_strdup(const char *src) {
    size_t len = uprv_strlen(src) + 1;
    char *dup = (char *) uprv_malloc(len);

    if (dup) {
        uprv_memcpy(dup, src, len);
    }

    return dup;
}

U_CAPI char* U_EXPORT2
uprv_strndup(const char *src, int32_t n) {
    char *dup;

    if(n < 0) {
        dup = uprv_strdup(src);
    } else {
        dup = (char*)uprv_malloc(n+1);
        if (dup) { 
            uprv_memcpy(dup, src, n);
            dup[n] = 0;
        }
    }

    return dup;
}
Commit	Line	Data
b75a7d8f A	1	/*
	2	******************************************************************************
	3	*
4388f060	4	* Copyright (C) 1997-2011, International Business Machines
b75a7d8f A	5	* Corporation and others. All Rights Reserved.
	6	*
	7	******************************************************************************
	8	*
	9	* File CSTRING.C
	10	*
	11	* @author Helena Shih
	12	*
	13	* Modification History:
	14	*
	15	* Date Name Description
	16	* 6/18/98 hshih Created
	17	* 09/08/98 stephen Added include for ctype, for Mac Port
	18	* 11/15/99 helena Integrated S/390 IEEE changes.
	19	******************************************************************************
	20	*/
	21
	22
	23
	24	#include <stdlib.h>
374ca955	25	#include <stdio.h>
b75a7d8f A	26	#include "unicode/utypes.h"
	27	#include "cmemory.h"
	28	#include "cstring.h"
374ca955	29	#include "uassert.h"
b75a7d8f A	30
	31	/*
	32	* We hardcode case conversion for invariant characters to match our expectation
	33	* and the compiler execution charset.
	34	* This prevents problems on systems
	35	* - with non-default casing behavior, like Turkish system locales where
	36	* tolower('I') maps to dotless i and toupper('i') maps to dotted I
	37	* - where there are no lowercase Latin characters at all, or using different
	38	* codes (some old EBCDIC codepages)
	39	*
	40	* This works because the compiler usually runs on a platform where the execution
	41	* charset includes all of the invariant characters at their expected
	42	* code positions, so that the char * string literals in ICU code match
	43	* the char literals here.
	44	*
	45	* Note that the set of lowercase Latin letters is discontiguous in EBCDIC
	46	* and the set of uppercase Latin letters is discontiguous as well.
	47	*/
	48
4388f060 A	49	U_CAPI UBool U_EXPORT2
	50	uprv_isASCIILetter(char c) {
	51	#if U_CHARSET_FAMILY==U_EBCDIC_FAMILY
	52	return
	53	('a'<=c && c<='i') \|\| ('j'<=c && c<='r') \|\| ('s'<=c && c<='z') \|\|
	54	('A'<=c && c<='I') \|\| ('J'<=c && c<='R') \|\| ('S'<=c && c<='Z');
	55	#else
	56	return ('a'<=c && c<='z') \|\| ('A'<=c && c<='Z');
	57	#endif
	58	}
	59
b75a7d8f A	60	U_CAPI char U_EXPORT2
	61	uprv_toupper(char c) {
	62	#if U_CHARSET_FAMILY==U_EBCDIC_FAMILY
	63	if(('a'<=c && c<='i') \|\| ('j'<=c && c<='r') \|\| ('s'<=c && c<='z')) {
	64	c=(char)(c+('A'-'a'));
	65	}
	66	#else
	67	if('a'<=c && c<='z') {
	68	c=(char)(c+('A'-'a'));
	69	}
	70	#endif
	71	return c;
	72	}
	73
374ca955 A	74
	75	#if 0
	76	/*
	77	* Commented out because cstring.h defines uprv_tolower() to be
	78	* the same as either uprv_asciitolower() or uprv_ebcdictolower()
	79	* to reduce the amount of code to cover with tests.
	80	*
	81	* Note that this uprv_tolower() definition is likely to work for most
	82	* charset families, not just ASCII and EBCDIC, because its #else branch
	83	* is written generically.
	84	*/
b75a7d8f A	85	U_CAPI char U_EXPORT2
	86	uprv_tolower(char c) {
	87	#if U_CHARSET_FAMILY==U_EBCDIC_FAMILY
	88	if(('A'<=c && c<='I') \|\| ('J'<=c && c<='R') \|\| ('S'<=c && c<='Z')) {
	89	c=(char)(c+('a'-'A'));
	90	}
	91	#else
	92	if('A'<=c && c<='Z') {
	93	c=(char)(c+('a'-'A'));
	94	}
	95	#endif
	96	return c;
	97	}
374ca955 A	98	#endif
	99
	100	U_CAPI char U_EXPORT2
	101	uprv_asciitolower(char c) {
	102	if(0x41<=c && c<=0x5a) {
	103	c=(char)(c+0x20);
	104	}
	105	return c;
	106	}
	107
	108	U_CAPI char U_EXPORT2
	109	uprv_ebcdictolower(char c) {
	110	if( (0xc1<=(uint8_t)c && (uint8_t)c<=0xc9) \|\|
	111	(0xd1<=(uint8_t)c && (uint8_t)c<=0xd9) \|\|
	112	(0xe2<=(uint8_t)c && (uint8_t)c<=0xe9)
	113	) {
	114	c=(char)(c-0x40);
	115	}
	116	return c;
	117	}
b75a7d8f A	118
	119
	120	U_CAPI char* U_EXPORT2
	121	T_CString_toLowerCase(char* str)
	122	{
	123	char* origPtr = str;
	124
	125	if (str) {
	126	do
	127	str = (char)uprv_tolower(str);
	128	while (*(str++));
	129	}
	130
	131	return origPtr;
	132	}
	133
	134	U_CAPI char* U_EXPORT2
	135	T_CString_toUpperCase(char* str)
	136	{
	137	char* origPtr = str;
	138
	139	if (str) {
	140	do
	141	str = (char)uprv_toupper(str);
	142	while (*(str++));
	143	}
	144
	145	return origPtr;
	146	}
	147
	148	/*
	149	* Takes a int32_t and fills in a char* string with that number "radix"-based.
	150	* Does not handle negative values (makes an empty string for them).
374ca955 A	151	* Writes at most 12 chars ("-2147483647" plus NUL).
374ca955 A	152	* Returns the length of the string (not including the NUL).
b75a7d8f A	153	*/
b75a7d8f A	154	U_CAPI int32_t U_EXPORT2
374ca955	155	T_CString_integerToString(char* buffer, int32_t v, int32_t radix)
b75a7d8f	156	{
374ca955 A	157	char tbuf[30];
	158	int32_t tbx = sizeof(tbuf);
	159	uint8_t digit;
	160	int32_t length = 0;
	161	uint32_t uval;
	162
	163	U_ASSERT(radix>=2 && radix<=16);
	164	uval = (uint32_t) v;
	165	if(v<0 && radix == 10) {
	166	/* Only in base 10 do we conside numbers to be signed. */
	167	uval = (uint32_t)(-v);
	168	buffer[length++] = '-';
b75a7d8f	169	}
374ca955 A	170
	171	tbx = sizeof(tbuf)-1;
	172	tbuf[tbx] = 0; /* We are generating the digits backwards. Null term the end. */
	173	do {
	174	digit = (uint8_t)(uval % radix);
	175	tbuf[--tbx] = (char)(T_CString_itosOffset(digit));
	176	uval = uval / radix;
	177	} while (uval != 0);
	178
	179	/* copy converted number into user buffer */
	180	uprv_strcpy(buffer+length, tbuf+tbx);
	181	length += sizeof(tbuf) - tbx -1;
	182	return length;
	183	}
b75a7d8f A	184
b75a7d8f A	185
b75a7d8f	186
374ca955 A	187	/*
	188	* Takes a int64_t and fills in a char* string with that number "radix"-based.
	189	* Writes at most 21: chars ("-9223372036854775807" plus NUL).
	190	* Returns the length of the string, not including the terminating NULL.
	191	*/
	192	U_CAPI int32_t U_EXPORT2
	193	T_CString_int64ToString(char* buffer, int64_t v, uint32_t radix)
	194	{
	195	char tbuf[30];
	196	int32_t tbx = sizeof(tbuf);
	197	uint8_t digit;
	198	int32_t length = 0;
	199	uint64_t uval;
	200
	201	U_ASSERT(radix>=2 && radix<=16);
	202	uval = (uint64_t) v;
	203	if(v<0 && radix == 10) {
	204	/* Only in base 10 do we conside numbers to be signed. */
	205	uval = (uint64_t)(-v);
	206	buffer[length++] = '-';
	207	}
	208
	209	tbx = sizeof(tbuf)-1;
	210	tbuf[tbx] = 0; /* We are generating the digits backwards. Null term the end. */
	211	do {
	212	digit = (uint8_t)(uval % radix);
	213	tbuf[--tbx] = (char)(T_CString_itosOffset(digit));
	214	uval = uval / radix;
	215	} while (uval != 0);
	216
	217	/* copy converted number into user buffer */
	218	uprv_strcpy(buffer+length, tbuf+tbx);
	219	length += sizeof(tbuf) - tbx -1;
	220	return length;
b75a7d8f A	221	}
	222
	223
	224	U_CAPI int32_t U_EXPORT2
	225	T_CString_stringToInteger(const char *integerString, int32_t radix)
	226	{
	227	char *end;
374ca955	228	return uprv_strtoul(integerString, &end, radix);
b75a7d8f A	229
b75a7d8f A	230	}
4388f060	231
b75a7d8f	232	U_CAPI int U_EXPORT2
4388f060	233	uprv_stricmp(const char str1, const char str2) {
b75a7d8f A	234	if(str1==NULL) {
	235	if(str2==NULL) {
	236	return 0;
	237	} else {
	238	return -1;
	239	}
	240	} else if(str2==NULL) {
	241	return 1;
	242	} else {
	243	/* compare non-NULL strings lexically with lowercase */
	244	int rc;
	245	unsigned char c1, c2;
	246
	247	for(;;) {
	248	c1=(unsigned char)*str1;
	249	c2=(unsigned char)*str2;
	250	if(c1==0) {
	251	if(c2==0) {
	252	return 0;
	253	} else {
	254	return -1;
	255	}
	256	} else if(c2==0) {
	257	return 1;
	258	} else {
	259	/* compare non-zero characters with lowercase */
	260	rc=(int)(unsigned char)uprv_tolower(c1)-(int)(unsigned char)uprv_tolower(c2);
	261	if(rc!=0) {
	262	return rc;
	263	}
	264	}
	265	++str1;
	266	++str2;
	267	}
	268	}
	269	}
	270
	271	U_CAPI int U_EXPORT2
4388f060	272	uprv_strnicmp(const char str1, const char str2, uint32_t n) {
b75a7d8f A	273	if(str1==NULL) {
	274	if(str2==NULL) {
	275	return 0;
	276	} else {
	277	return -1;
	278	}
	279	} else if(str2==NULL) {
	280	return 1;
	281	} else {
	282	/* compare non-NULL strings lexically with lowercase */
	283	int rc;
	284	unsigned char c1, c2;
	285
	286	for(; n--;) {
	287	c1=(unsigned char)*str1;
	288	c2=(unsigned char)*str2;
	289	if(c1==0) {
	290	if(c2==0) {
	291	return 0;
	292	} else {
	293	return -1;
	294	}
	295	} else if(c2==0) {
	296	return 1;
	297	} else {
	298	/* compare non-zero characters with lowercase */
	299	rc=(int)(unsigned char)uprv_tolower(c1)-(int)(unsigned char)uprv_tolower(c2);
	300	if(rc!=0) {
	301	return rc;
	302	}
	303	}
	304	++str1;
	305	++str2;
	306	}
	307	}
	308
	309	return 0;
	310	}
	311
	312	U_CAPI char* U_EXPORT2
	313	uprv_strdup(const char *src) {
374ca955	314	size_t len = uprv_strlen(src) + 1;
b75a7d8f A	315	char dup = (char ) uprv_malloc(len);
	316
	317	if (dup) {
	318	uprv_memcpy(dup, src, len);
	319	}
	320
	321	return dup;
	322	}
374ca955 A	323
	324	U_CAPI char* U_EXPORT2
	325	uprv_strndup(const char *src, int32_t n) {
	326	char *dup;
	327
	328	if(n < 0) {
	329	dup = uprv_strdup(src);
	330	} else {
	331	dup = (char*)uprv_malloc(n+1);
	332	if (dup) {
	333	uprv_memcpy(dup, src, n);
	334	dup[n] = 0;
	335	}
	336	}
	337
	338	return dup;
	339	}