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

// © 2016 and later: Unicode, Inc. and others.
// License & terms of use: http://www.unicode.org/copyright.html
/*
******************************************************************************
*
*   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
f3c0d7a5 A	1	// © 2016 and later: Unicode, Inc. and others.
f3c0d7a5 A	2	// License & terms of use: http://www.unicode.org/copyright.html
b75a7d8f A	3	/*
	4	******************************************************************************
	5	*
4388f060	6	* Copyright (C) 1997-2011, International Business Machines
b75a7d8f A	7	* Corporation and others. All Rights Reserved.
	8	*
	9	******************************************************************************
	10	*
	11	* File CSTRING.C
	12	*
	13	* @author Helena Shih
	14	*
	15	* Modification History:
	16	*
	17	* Date Name Description
	18	* 6/18/98 hshih Created
	19	* 09/08/98 stephen Added include for ctype, for Mac Port
	20	* 11/15/99 helena Integrated S/390 IEEE changes.
	21	******************************************************************************
	22	*/
	23
	24
	25
	26	#include <stdlib.h>
374ca955	27	#include <stdio.h>
b75a7d8f A	28	#include "unicode/utypes.h"
	29	#include "cmemory.h"
	30	#include "cstring.h"
374ca955	31	#include "uassert.h"
b75a7d8f A	32
	33	/*
	34	* We hardcode case conversion for invariant characters to match our expectation
	35	* and the compiler execution charset.
	36	* This prevents problems on systems
	37	* - with non-default casing behavior, like Turkish system locales where
	38	* tolower('I') maps to dotless i and toupper('i') maps to dotted I
	39	* - where there are no lowercase Latin characters at all, or using different
	40	* codes (some old EBCDIC codepages)
	41	*
	42	* This works because the compiler usually runs on a platform where the execution
	43	* charset includes all of the invariant characters at their expected
	44	* code positions, so that the char * string literals in ICU code match
	45	* the char literals here.
	46	*
	47	* Note that the set of lowercase Latin letters is discontiguous in EBCDIC
	48	* and the set of uppercase Latin letters is discontiguous as well.
	49	*/
	50
4388f060 A	51	U_CAPI UBool U_EXPORT2
	52	uprv_isASCIILetter(char c) {
	53	#if U_CHARSET_FAMILY==U_EBCDIC_FAMILY
	54	return
	55	('a'<=c && c<='i') \|\| ('j'<=c && c<='r') \|\| ('s'<=c && c<='z') \|\|
	56	('A'<=c && c<='I') \|\| ('J'<=c && c<='R') \|\| ('S'<=c && c<='Z');
	57	#else
	58	return ('a'<=c && c<='z') \|\| ('A'<=c && c<='Z');
	59	#endif
	60	}
	61
b75a7d8f A	62	U_CAPI char U_EXPORT2
	63	uprv_toupper(char c) {
	64	#if U_CHARSET_FAMILY==U_EBCDIC_FAMILY
	65	if(('a'<=c && c<='i') \|\| ('j'<=c && c<='r') \|\| ('s'<=c && c<='z')) {
	66	c=(char)(c+('A'-'a'));
	67	}
	68	#else
	69	if('a'<=c && c<='z') {
	70	c=(char)(c+('A'-'a'));
	71	}
	72	#endif
	73	return c;
	74	}
	75
374ca955 A	76
	77	#if 0
	78	/*
	79	* Commented out because cstring.h defines uprv_tolower() to be
	80	* the same as either uprv_asciitolower() or uprv_ebcdictolower()
	81	* to reduce the amount of code to cover with tests.
	82	*
	83	* Note that this uprv_tolower() definition is likely to work for most
	84	* charset families, not just ASCII and EBCDIC, because its #else branch
	85	* is written generically.
	86	*/
b75a7d8f A	87	U_CAPI char U_EXPORT2
	88	uprv_tolower(char c) {
	89	#if U_CHARSET_FAMILY==U_EBCDIC_FAMILY
	90	if(('A'<=c && c<='I') \|\| ('J'<=c && c<='R') \|\| ('S'<=c && c<='Z')) {
	91	c=(char)(c+('a'-'A'));
	92	}
	93	#else
	94	if('A'<=c && c<='Z') {
	95	c=(char)(c+('a'-'A'));
	96	}
	97	#endif
	98	return c;
	99	}
374ca955 A	100	#endif
	101
	102	U_CAPI char U_EXPORT2
	103	uprv_asciitolower(char c) {
	104	if(0x41<=c && c<=0x5a) {
	105	c=(char)(c+0x20);
	106	}
	107	return c;
	108	}
	109
	110	U_CAPI char U_EXPORT2
	111	uprv_ebcdictolower(char c) {
	112	if( (0xc1<=(uint8_t)c && (uint8_t)c<=0xc9) \|\|
	113	(0xd1<=(uint8_t)c && (uint8_t)c<=0xd9) \|\|
	114	(0xe2<=(uint8_t)c && (uint8_t)c<=0xe9)
	115	) {
	116	c=(char)(c-0x40);
	117	}
	118	return c;
	119	}
b75a7d8f A	120
	121
	122	U_CAPI char* U_EXPORT2
	123	T_CString_toLowerCase(char* str)
	124	{
	125	char* origPtr = str;
	126
	127	if (str) {
	128	do
	129	str = (char)uprv_tolower(str);
	130	while (*(str++));
	131	}
	132
	133	return origPtr;
	134	}
	135
	136	U_CAPI char* U_EXPORT2
	137	T_CString_toUpperCase(char* str)
	138	{
	139	char* origPtr = str;
	140
	141	if (str) {
	142	do
	143	str = (char)uprv_toupper(str);
	144	while (*(str++));
	145	}
	146
	147	return origPtr;
	148	}
	149
	150	/*
	151	* Takes a int32_t and fills in a char* string with that number "radix"-based.
	152	* Does not handle negative values (makes an empty string for them).
374ca955 A	153	* Writes at most 12 chars ("-2147483647" plus NUL).
374ca955 A	154	* Returns the length of the string (not including the NUL).
b75a7d8f A	155	*/
b75a7d8f A	156	U_CAPI int32_t U_EXPORT2
374ca955	157	T_CString_integerToString(char* buffer, int32_t v, int32_t radix)
b75a7d8f	158	{
374ca955 A	159	char tbuf[30];
	160	int32_t tbx = sizeof(tbuf);
	161	uint8_t digit;
	162	int32_t length = 0;
	163	uint32_t uval;
	164
	165	U_ASSERT(radix>=2 && radix<=16);
	166	uval = (uint32_t) v;
	167	if(v<0 && radix == 10) {
	168	/* Only in base 10 do we conside numbers to be signed. */
	169	uval = (uint32_t)(-v);
	170	buffer[length++] = '-';
b75a7d8f	171	}
374ca955 A	172
	173	tbx = sizeof(tbuf)-1;
	174	tbuf[tbx] = 0; /* We are generating the digits backwards. Null term the end. */
	175	do {
	176	digit = (uint8_t)(uval % radix);
	177	tbuf[--tbx] = (char)(T_CString_itosOffset(digit));
	178	uval = uval / radix;
	179	} while (uval != 0);
	180
	181	/* copy converted number into user buffer */
	182	uprv_strcpy(buffer+length, tbuf+tbx);
	183	length += sizeof(tbuf) - tbx -1;
	184	return length;
	185	}
b75a7d8f A	186
b75a7d8f A	187
b75a7d8f	188
374ca955 A	189	/*
	190	* Takes a int64_t and fills in a char* string with that number "radix"-based.
	191	* Writes at most 21: chars ("-9223372036854775807" plus NUL).
	192	* Returns the length of the string, not including the terminating NULL.
	193	*/
	194	U_CAPI int32_t U_EXPORT2
	195	T_CString_int64ToString(char* buffer, int64_t v, uint32_t radix)
	196	{
	197	char tbuf[30];
	198	int32_t tbx = sizeof(tbuf);
	199	uint8_t digit;
	200	int32_t length = 0;
	201	uint64_t uval;
	202
	203	U_ASSERT(radix>=2 && radix<=16);
	204	uval = (uint64_t) v;
	205	if(v<0 && radix == 10) {
	206	/* Only in base 10 do we conside numbers to be signed. */
	207	uval = (uint64_t)(-v);
	208	buffer[length++] = '-';
	209	}
	210
	211	tbx = sizeof(tbuf)-1;
	212	tbuf[tbx] = 0; /* We are generating the digits backwards. Null term the end. */
	213	do {
	214	digit = (uint8_t)(uval % radix);
	215	tbuf[--tbx] = (char)(T_CString_itosOffset(digit));
	216	uval = uval / radix;
	217	} while (uval != 0);
	218
	219	/* copy converted number into user buffer */
	220	uprv_strcpy(buffer+length, tbuf+tbx);
	221	length += sizeof(tbuf) - tbx -1;
	222	return length;
b75a7d8f A	223	}
	224
	225
	226	U_CAPI int32_t U_EXPORT2
	227	T_CString_stringToInteger(const char *integerString, int32_t radix)
	228	{
	229	char *end;
374ca955	230	return uprv_strtoul(integerString, &end, radix);
b75a7d8f A	231
b75a7d8f A	232	}
4388f060	233
b75a7d8f	234	U_CAPI int U_EXPORT2
4388f060	235	uprv_stricmp(const char str1, const char str2) {
b75a7d8f A	236	if(str1==NULL) {
	237	if(str2==NULL) {
	238	return 0;
	239	} else {
	240	return -1;
	241	}
	242	} else if(str2==NULL) {
	243	return 1;
	244	} else {
	245	/* compare non-NULL strings lexically with lowercase */
	246	int rc;
	247	unsigned char c1, c2;
	248
	249	for(;;) {
	250	c1=(unsigned char)*str1;
	251	c2=(unsigned char)*str2;
	252	if(c1==0) {
	253	if(c2==0) {
	254	return 0;
	255	} else {
	256	return -1;
	257	}
	258	} else if(c2==0) {
	259	return 1;
	260	} else {
	261	/* compare non-zero characters with lowercase */
	262	rc=(int)(unsigned char)uprv_tolower(c1)-(int)(unsigned char)uprv_tolower(c2);
	263	if(rc!=0) {
	264	return rc;
	265	}
	266	}
	267	++str1;
	268	++str2;
	269	}
	270	}
	271	}
	272
	273	U_CAPI int U_EXPORT2
4388f060	274	uprv_strnicmp(const char str1, const char str2, uint32_t n) {
b75a7d8f A	275	if(str1==NULL) {
	276	if(str2==NULL) {
	277	return 0;
	278	} else {
	279	return -1;
	280	}
	281	} else if(str2==NULL) {
	282	return 1;
	283	} else {
	284	/* compare non-NULL strings lexically with lowercase */
	285	int rc;
	286	unsigned char c1, c2;
	287
	288	for(; n--;) {
	289	c1=(unsigned char)*str1;
	290	c2=(unsigned char)*str2;
	291	if(c1==0) {
	292	if(c2==0) {
	293	return 0;
	294	} else {
	295	return -1;
	296	}
	297	} else if(c2==0) {
	298	return 1;
	299	} else {
	300	/* compare non-zero characters with lowercase */
	301	rc=(int)(unsigned char)uprv_tolower(c1)-(int)(unsigned char)uprv_tolower(c2);
	302	if(rc!=0) {
	303	return rc;
	304	}
	305	}
	306	++str1;
	307	++str2;
	308	}
	309	}
	310
	311	return 0;
	312	}
	313
	314	U_CAPI char* U_EXPORT2
	315	uprv_strdup(const char *src) {
374ca955	316	size_t len = uprv_strlen(src) + 1;
b75a7d8f A	317	char dup = (char ) uprv_malloc(len);
	318
	319	if (dup) {
	320	uprv_memcpy(dup, src, len);
	321	}
	322
	323	return dup;
	324	}
374ca955 A	325
	326	U_CAPI char* U_EXPORT2
	327	uprv_strndup(const char *src, int32_t n) {
	328	char *dup;
	329
	330	if(n < 0) {
	331	dup = uprv_strdup(src);
	332	} else {
	333	dup = (char*)uprv_malloc(n+1);
	334	if (dup) {
	335	uprv_memcpy(dup, src, n);
	336	dup[n] = 0;
	337	}
	338	}
	339
	340	return dup;
	341	}