michael@0: /*
michael@0: *******************************************************************************
michael@0: *
michael@0: *   Copyright (C) 1999-2010, International Business Machines
michael@0: *   Corporation and others.  All Rights Reserved.
michael@0: *
michael@0: *******************************************************************************
michael@0: *   file name:  uinvchar.c
michael@0: *   encoding:   US-ASCII
michael@0: *   tab size:   8 (not used)
michael@0: *   indentation:2
michael@0: *
michael@0: *   created on: 2004sep14
michael@0: *   created by: Markus W. Scherer
michael@0: *
michael@0: *   Functions for handling invariant characters, moved here from putil.c
michael@0: *   for better modularization.
michael@0: */
michael@0: 
michael@0: #include "unicode/utypes.h"
michael@0: #include "unicode/ustring.h"
michael@0: #include "udataswp.h"
michael@0: #include "cstring.h"
michael@0: #include "cmemory.h"
michael@0: #include "uassert.h"
michael@0: #include "uinvchar.h"
michael@0: 
michael@0: /* invariant-character handling --------------------------------------------- */
michael@0: 
michael@0: /*
michael@0:  * These maps for ASCII to/from EBCDIC map invariant characters (see utypes.h)
michael@0:  * appropriately for most EBCDIC codepages.
michael@0:  *
michael@0:  * They currently also map most other ASCII graphic characters,
michael@0:  * appropriately for codepages 37 and 1047.
michael@0:  * Exceptions: The characters for []^ have different codes in 37 & 1047.
michael@0:  * Both versions are mapped to ASCII.
michael@0:  *
michael@0:  *    ASCII 37 1047
michael@0:  * [     5B BA   AD
michael@0:  * ]     5D BB   BD
michael@0:  * ^     5E B0   5F
michael@0:  *
michael@0:  * There are no mappings for variant characters from Unicode to EBCDIC.
michael@0:  *
michael@0:  * Currently, C0 control codes are also included in these maps.
michael@0:  * Exceptions: S/390 Open Edition swaps LF and NEL codes compared with other
michael@0:  * EBCDIC platforms; both codes (15 and 25) are mapped to ASCII LF (0A),
michael@0:  * but there is no mapping for ASCII LF back to EBCDIC.
michael@0:  *
michael@0:  *    ASCII EBCDIC S/390-OE
michael@0:  * LF    0A     25       15
michael@0:  * NEL   85     15       25
michael@0:  *
michael@0:  * The maps below explicitly exclude the variant
michael@0:  * control and graphical characters that are in ASCII-based
michael@0:  * codepages at 0x80 and above.
michael@0:  * "No mapping" is expressed by mapping to a 00 byte.
michael@0:  *
michael@0:  * These tables do not establish a converter or a codepage.
michael@0:  */
michael@0: 
michael@0: static const uint8_t asciiFromEbcdic[256]={
michael@0:     0x00, 0x01, 0x02, 0x03, 0x00, 0x09, 0x00, 0x7f, 0x00, 0x00, 0x00, 0x0b, 0x0c, 0x0d, 0x0e, 0x0f,
michael@0:     0x10, 0x11, 0x12, 0x13, 0x00, 0x0a, 0x08, 0x00, 0x18, 0x19, 0x00, 0x00, 0x1c, 0x1d, 0x1e, 0x1f,
michael@0:     0x00, 0x00, 0x00, 0x00, 0x00, 0x0a, 0x17, 0x1b, 0x00, 0x00, 0x00, 0x00, 0x00, 0x05, 0x06, 0x07,
michael@0:     0x00, 0x00, 0x16, 0x00, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x00, 0x14, 0x15, 0x00, 0x1a,
michael@0: 
michael@0:     0x20, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x2e, 0x3c, 0x28, 0x2b, 0x7c,
michael@0:     0x26, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x21, 0x24, 0x2a, 0x29, 0x3b, 0x5e,
michael@0:     0x2d, 0x2f, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x2c, 0x25, 0x5f, 0x3e, 0x3f,
michael@0:     0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x60, 0x3a, 0x23, 0x40, 0x27, 0x3d, 0x22,
michael@0: 
michael@0:     0x00, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, 0x68, 0x69, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
michael@0:     0x00, 0x6a, 0x6b, 0x6c, 0x6d, 0x6e, 0x6f, 0x70, 0x71, 0x72, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
michael@0:     0x00, 0x7e, 0x73, 0x74, 0x75, 0x76, 0x77, 0x78, 0x79, 0x7a, 0x00, 0x00, 0x00, 0x5b, 0x00, 0x00,
michael@0:     0x5e, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x5b, 0x5d, 0x00, 0x5d, 0x00, 0x00,
michael@0: 
michael@0:     0x7b, 0x41, 0x42, 0x43, 0x44, 0x45, 0x46, 0x47, 0x48, 0x49, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
michael@0:     0x7d, 0x4a, 0x4b, 0x4c, 0x4d, 0x4e, 0x4f, 0x50, 0x51, 0x52, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
michael@0:     0x5c, 0x00, 0x53, 0x54, 0x55, 0x56, 0x57, 0x58, 0x59, 0x5a, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
michael@0:     0x30, 0x31, 0x32, 0x33, 0x34, 0x35, 0x36, 0x37, 0x38, 0x39, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00
michael@0: };
michael@0: 
michael@0: static const uint8_t ebcdicFromAscii[256]={
michael@0:     0x00, 0x01, 0x02, 0x03, 0x37, 0x2d, 0x2e, 0x2f, 0x16, 0x05, 0x00, 0x0b, 0x0c, 0x0d, 0x0e, 0x0f,
michael@0:     0x10, 0x11, 0x12, 0x13, 0x3c, 0x3d, 0x32, 0x26, 0x18, 0x19, 0x3f, 0x27, 0x1c, 0x1d, 0x1e, 0x1f,
michael@0:     0x40, 0x00, 0x7f, 0x00, 0x00, 0x6c, 0x50, 0x7d, 0x4d, 0x5d, 0x5c, 0x4e, 0x6b, 0x60, 0x4b, 0x61,
michael@0:     0xf0, 0xf1, 0xf2, 0xf3, 0xf4, 0xf5, 0xf6, 0xf7, 0xf8, 0xf9, 0x7a, 0x5e, 0x4c, 0x7e, 0x6e, 0x6f,
michael@0: 
michael@0:     0x00, 0xc1, 0xc2, 0xc3, 0xc4, 0xc5, 0xc6, 0xc7, 0xc8, 0xc9, 0xd1, 0xd2, 0xd3, 0xd4, 0xd5, 0xd6,
michael@0:     0xd7, 0xd8, 0xd9, 0xe2, 0xe3, 0xe4, 0xe5, 0xe6, 0xe7, 0xe8, 0xe9, 0x00, 0x00, 0x00, 0x00, 0x6d,
michael@0:     0x00, 0x81, 0x82, 0x83, 0x84, 0x85, 0x86, 0x87, 0x88, 0x89, 0x91, 0x92, 0x93, 0x94, 0x95, 0x96,
michael@0:     0x97, 0x98, 0x99, 0xa2, 0xa3, 0xa4, 0xa5, 0xa6, 0xa7, 0xa8, 0xa9, 0x00, 0x00, 0x00, 0x00, 0x07,
michael@0: 
michael@0:     0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
michael@0:     0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
michael@0:     0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
michael@0:     0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
michael@0: 
michael@0:     0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
michael@0:     0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
michael@0:     0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
michael@0:     0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
michael@0: };
michael@0: 
michael@0: /* Same as asciiFromEbcdic[] except maps all letters to lowercase. */
michael@0: static const uint8_t lowercaseAsciiFromEbcdic[256]={
michael@0:     0x00, 0x01, 0x02, 0x03, 0x00, 0x09, 0x00, 0x7f, 0x00, 0x00, 0x00, 0x0b, 0x0c, 0x0d, 0x0e, 0x0f,
michael@0:     0x10, 0x11, 0x12, 0x13, 0x00, 0x0a, 0x08, 0x00, 0x18, 0x19, 0x00, 0x00, 0x1c, 0x1d, 0x1e, 0x1f,
michael@0:     0x00, 0x00, 0x00, 0x00, 0x00, 0x0a, 0x17, 0x1b, 0x00, 0x00, 0x00, 0x00, 0x00, 0x05, 0x06, 0x07,
michael@0:     0x00, 0x00, 0x16, 0x00, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x00, 0x14, 0x15, 0x00, 0x1a,
michael@0: 
michael@0:     0x20, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x2e, 0x3c, 0x28, 0x2b, 0x7c,
michael@0:     0x26, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x21, 0x24, 0x2a, 0x29, 0x3b, 0x5e,
michael@0:     0x2d, 0x2f, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x2c, 0x25, 0x5f, 0x3e, 0x3f,
michael@0:     0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x60, 0x3a, 0x23, 0x40, 0x27, 0x3d, 0x22,
michael@0: 
michael@0:     0x00, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, 0x68, 0x69, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
michael@0:     0x00, 0x6a, 0x6b, 0x6c, 0x6d, 0x6e, 0x6f, 0x70, 0x71, 0x72, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
michael@0:     0x00, 0x7e, 0x73, 0x74, 0x75, 0x76, 0x77, 0x78, 0x79, 0x7a, 0x00, 0x00, 0x00, 0x5b, 0x00, 0x00,
michael@0:     0x5e, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x5b, 0x5d, 0x00, 0x5d, 0x00, 0x00,
michael@0: 
michael@0:     0x7b, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, 0x68, 0x69, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
michael@0:     0x7d, 0x6a, 0x6b, 0x6c, 0x6d, 0x6e, 0x6f, 0x70, 0x71, 0x72, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
michael@0:     0x7c, 0x00, 0x73, 0x74, 0x75, 0x76, 0x77, 0x78, 0x79, 0x7a, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
michael@0:     0x30, 0x31, 0x32, 0x33, 0x34, 0x35, 0x36, 0x37, 0x38, 0x39, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00
michael@0: };
michael@0: 
michael@0: /*
michael@0:  * Bit sets indicating which characters of the ASCII repertoire
michael@0:  * (by ASCII/Unicode code) are "invariant".
michael@0:  * See utypes.h for more details.
michael@0:  *
michael@0:  * As invariant are considered the characters of the ASCII repertoire except
michael@0:  * for the following:
michael@0:  * 21  '!' <exclamation mark>
michael@0:  * 23  '#' <number sign>
michael@0:  * 24  '$' <dollar sign>
michael@0:  *
michael@0:  * 40  '@' <commercial at>
michael@0:  *
michael@0:  * 5b  '[' <left bracket>
michael@0:  * 5c  '\' <backslash>
michael@0:  * 5d  ']' <right bracket>
michael@0:  * 5e  '^' <circumflex>
michael@0:  *
michael@0:  * 60  '`' <grave accent>
michael@0:  *
michael@0:  * 7b  '{' <left brace>
michael@0:  * 7c  '|' <vertical line>
michael@0:  * 7d  '}' <right brace>
michael@0:  * 7e  '~' <tilde>
michael@0:  */
michael@0: static const uint32_t invariantChars[4]={
michael@0:     0xfffffbff, /* 00..1f but not 0a */
michael@0:     0xffffffe5, /* 20..3f but not 21 23 24 */
michael@0:     0x87fffffe, /* 40..5f but not 40 5b..5e */
michael@0:     0x87fffffe  /* 60..7f but not 60 7b..7e */
michael@0: };
michael@0: 
michael@0: /*
michael@0:  * test unsigned types (or values known to be non-negative) for invariant characters,
michael@0:  * tests ASCII-family character values
michael@0:  */
michael@0: #define UCHAR_IS_INVARIANT(c) (((c)<=0x7f) && (invariantChars[(c)>>5]&((uint32_t)1<<((c)&0x1f)))!=0)
michael@0: 
michael@0: /* test signed types for invariant characters, adds test for positive values */
michael@0: #define SCHAR_IS_INVARIANT(c) ((0<=(c)) && UCHAR_IS_INVARIANT(c))
michael@0: 
michael@0: #if U_CHARSET_FAMILY==U_ASCII_FAMILY
michael@0: #define CHAR_TO_UCHAR(c) c
michael@0: #define UCHAR_TO_CHAR(c) c
michael@0: #elif U_CHARSET_FAMILY==U_EBCDIC_FAMILY
michael@0: #define CHAR_TO_UCHAR(u) asciiFromEbcdic[u]
michael@0: #define UCHAR_TO_CHAR(u) ebcdicFromAscii[u]
michael@0: #else
michael@0: #   error U_CHARSET_FAMILY is not valid
michael@0: #endif
michael@0: 
michael@0: 
michael@0: U_CAPI void U_EXPORT2
michael@0: u_charsToUChars(const char *cs, UChar *us, int32_t length) {
michael@0:     UChar u;
michael@0:     uint8_t c;
michael@0: 
michael@0:     /*
michael@0:      * Allow the entire ASCII repertoire to be mapped _to_ Unicode.
michael@0:      * For EBCDIC systems, this works for characters with codes from
michael@0:      * codepages 37 and 1047 or compatible.
michael@0:      */
michael@0:     while(length>0) {
michael@0:         c=(uint8_t)(*cs++);
michael@0:         u=(UChar)CHAR_TO_UCHAR(c);
michael@0:         U_ASSERT((u!=0 || c==0)); /* only invariant chars converted? */
michael@0:         *us++=u;
michael@0:         --length;
michael@0:     }
michael@0: }
michael@0: 
michael@0: U_CAPI void U_EXPORT2
michael@0: u_UCharsToChars(const UChar *us, char *cs, int32_t length) {
michael@0:     UChar u;
michael@0: 
michael@0:     while(length>0) {
michael@0:         u=*us++;
michael@0:         if(!UCHAR_IS_INVARIANT(u)) {
michael@0:             U_ASSERT(FALSE); /* Variant characters were used. These are not portable in ICU. */
michael@0:             u=0;
michael@0:         }
michael@0:         *cs++=(char)UCHAR_TO_CHAR(u);
michael@0:         --length;
michael@0:     }
michael@0: }
michael@0: 
michael@0: U_CAPI UBool U_EXPORT2
michael@0: uprv_isInvariantString(const char *s, int32_t length) {
michael@0:     uint8_t c;
michael@0: 
michael@0:     for(;;) {
michael@0:         if(length<0) {
michael@0:             /* NUL-terminated */
michael@0:             c=(uint8_t)*s++;
michael@0:             if(c==0) {
michael@0:                 break;
michael@0:             }
michael@0:         } else {
michael@0:             /* count length */
michael@0:             if(length==0) {
michael@0:                 break;
michael@0:             }
michael@0:             --length;
michael@0:             c=(uint8_t)*s++;
michael@0:             if(c==0) {
michael@0:                 continue; /* NUL is invariant */
michael@0:             }
michael@0:         }
michael@0:         /* c!=0 now, one branch below checks c==0 for variant characters */
michael@0: 
michael@0:         /*
michael@0:          * no assertions here because these functions are legitimately called
michael@0:          * for strings with variant characters
michael@0:          */
michael@0: #if U_CHARSET_FAMILY==U_ASCII_FAMILY
michael@0:         if(!UCHAR_IS_INVARIANT(c)) {
michael@0:             return FALSE; /* found a variant char */
michael@0:         }
michael@0: #elif U_CHARSET_FAMILY==U_EBCDIC_FAMILY
michael@0:         c=CHAR_TO_UCHAR(c);
michael@0:         if(c==0 || !UCHAR_IS_INVARIANT(c)) {
michael@0:             return FALSE; /* found a variant char */
michael@0:         }
michael@0: #else
michael@0: #   error U_CHARSET_FAMILY is not valid
michael@0: #endif
michael@0:     }
michael@0:     return TRUE;
michael@0: }
michael@0: 
michael@0: U_CAPI UBool U_EXPORT2
michael@0: uprv_isInvariantUString(const UChar *s, int32_t length) {
michael@0:     UChar c;
michael@0: 
michael@0:     for(;;) {
michael@0:         if(length<0) {
michael@0:             /* NUL-terminated */
michael@0:             c=*s++;
michael@0:             if(c==0) {
michael@0:                 break;
michael@0:             }
michael@0:         } else {
michael@0:             /* count length */
michael@0:             if(length==0) {
michael@0:                 break;
michael@0:             }
michael@0:             --length;
michael@0:             c=*s++;
michael@0:         }
michael@0: 
michael@0:         /*
michael@0:          * no assertions here because these functions are legitimately called
michael@0:          * for strings with variant characters
michael@0:          */
michael@0:         if(!UCHAR_IS_INVARIANT(c)) {
michael@0:             return FALSE; /* found a variant char */
michael@0:         }
michael@0:     }
michael@0:     return TRUE;
michael@0: }
michael@0: 
michael@0: /* UDataSwapFn implementations used in udataswp.c ------- */
michael@0: 
michael@0: /* convert ASCII to EBCDIC and verify that all characters are invariant */
michael@0: U_CAPI int32_t U_EXPORT2
michael@0: uprv_ebcdicFromAscii(const UDataSwapper *ds,
michael@0:                      const void *inData, int32_t length, void *outData,
michael@0:                      UErrorCode *pErrorCode) {
michael@0:     const uint8_t *s;
michael@0:     uint8_t *t;
michael@0:     uint8_t c;
michael@0: 
michael@0:     int32_t count;
michael@0: 
michael@0:     if(pErrorCode==NULL || U_FAILURE(*pErrorCode)) {
michael@0:         return 0;
michael@0:     }
michael@0:     if(ds==NULL || inData==NULL || length<0 || (length>0 && outData==NULL)) {
michael@0:         *pErrorCode=U_ILLEGAL_ARGUMENT_ERROR;
michael@0:         return 0;
michael@0:     }
michael@0: 
michael@0:     /* setup and swapping */
michael@0:     s=(const uint8_t *)inData;
michael@0:     t=(uint8_t *)outData;
michael@0:     count=length;
michael@0:     while(count>0) {
michael@0:         c=*s++;
michael@0:         if(!UCHAR_IS_INVARIANT(c)) {
michael@0:             udata_printError(ds, "uprv_ebcdicFromAscii() string[%d] contains a variant character in position %d\n",
michael@0:                              length, length-count);
michael@0:             *pErrorCode=U_INVALID_CHAR_FOUND;
michael@0:             return 0;
michael@0:         }
michael@0:         *t++=ebcdicFromAscii[c];
michael@0:         --count;
michael@0:     }
michael@0: 
michael@0:     return length;
michael@0: }
michael@0: 
michael@0: /* this function only checks and copies ASCII strings without conversion */
michael@0: U_CFUNC int32_t
michael@0: uprv_copyAscii(const UDataSwapper *ds,
michael@0:                const void *inData, int32_t length, void *outData,
michael@0:                UErrorCode *pErrorCode) {
michael@0:     const uint8_t *s;
michael@0:     uint8_t c;
michael@0: 
michael@0:     int32_t count;
michael@0: 
michael@0:     if(pErrorCode==NULL || U_FAILURE(*pErrorCode)) {
michael@0:         return 0;
michael@0:     }
michael@0:     if(ds==NULL || inData==NULL || length<0 || (length>0 && outData==NULL)) {
michael@0:         *pErrorCode=U_ILLEGAL_ARGUMENT_ERROR;
michael@0:         return 0;
michael@0:     }
michael@0: 
michael@0:     /* setup and checking */
michael@0:     s=(const uint8_t *)inData;
michael@0:     count=length;
michael@0:     while(count>0) {
michael@0:         c=*s++;
michael@0:         if(!UCHAR_IS_INVARIANT(c)) {
michael@0:             udata_printError(ds, "uprv_copyFromAscii() string[%d] contains a variant character in position %d\n",
michael@0:                              length, length-count);
michael@0:             *pErrorCode=U_INVALID_CHAR_FOUND;
michael@0:             return 0;
michael@0:         }
michael@0:         --count;
michael@0:     }
michael@0: 
michael@0:     if(length>0 && inData!=outData) {
michael@0:         uprv_memcpy(outData, inData, length);
michael@0:     }
michael@0: 
michael@0:     return length;
michael@0: }
michael@0: 
michael@0: /* convert EBCDIC to ASCII and verify that all characters are invariant */
michael@0: U_CFUNC int32_t
michael@0: uprv_asciiFromEbcdic(const UDataSwapper *ds,
michael@0:                      const void *inData, int32_t length, void *outData,
michael@0:                      UErrorCode *pErrorCode) {
michael@0:     const uint8_t *s;
michael@0:     uint8_t *t;
michael@0:     uint8_t c;
michael@0: 
michael@0:     int32_t count;
michael@0: 
michael@0:     if(pErrorCode==NULL || U_FAILURE(*pErrorCode)) {
michael@0:         return 0;
michael@0:     }
michael@0:     if(ds==NULL || inData==NULL || length<0 ||  (length>0 && outData==NULL)) {
michael@0:         *pErrorCode=U_ILLEGAL_ARGUMENT_ERROR;
michael@0:         return 0;
michael@0:     }
michael@0: 
michael@0:     /* setup and swapping */
michael@0:     s=(const uint8_t *)inData;
michael@0:     t=(uint8_t *)outData;
michael@0:     count=length;
michael@0:     while(count>0) {
michael@0:         c=*s++;
michael@0:         if(c!=0 && ((c=asciiFromEbcdic[c])==0 || !UCHAR_IS_INVARIANT(c))) {
michael@0:             udata_printError(ds, "uprv_asciiFromEbcdic() string[%d] contains a variant character in position %d\n",
michael@0:                              length, length-count);
michael@0:             *pErrorCode=U_INVALID_CHAR_FOUND;
michael@0:             return 0;
michael@0:         }
michael@0:         *t++=c;
michael@0:         --count;
michael@0:     }
michael@0: 
michael@0:     return length;
michael@0: }
michael@0: 
michael@0: /* this function only checks and copies EBCDIC strings without conversion */
michael@0: U_CFUNC int32_t
michael@0: uprv_copyEbcdic(const UDataSwapper *ds,
michael@0:                 const void *inData, int32_t length, void *outData,
michael@0:                 UErrorCode *pErrorCode) {
michael@0:     const uint8_t *s;
michael@0:     uint8_t c;
michael@0: 
michael@0:     int32_t count;
michael@0: 
michael@0:     if(pErrorCode==NULL || U_FAILURE(*pErrorCode)) {
michael@0:         return 0;
michael@0:     }
michael@0:     if(ds==NULL || inData==NULL || length<0 || (length>0 && outData==NULL)) {
michael@0:         *pErrorCode=U_ILLEGAL_ARGUMENT_ERROR;
michael@0:         return 0;
michael@0:     }
michael@0: 
michael@0:     /* setup and checking */
michael@0:     s=(const uint8_t *)inData;
michael@0:     count=length;
michael@0:     while(count>0) {
michael@0:         c=*s++;
michael@0:         if(c!=0 && ((c=asciiFromEbcdic[c])==0 || !UCHAR_IS_INVARIANT(c))) {
michael@0:             udata_printError(ds, "uprv_copyEbcdic() string[%] contains a variant character in position %d\n",
michael@0:                              length, length-count);
michael@0:             *pErrorCode=U_INVALID_CHAR_FOUND;
michael@0:             return 0;
michael@0:         }
michael@0:         --count;
michael@0:     }
michael@0: 
michael@0:     if(length>0 && inData!=outData) {
michael@0:         uprv_memcpy(outData, inData, length);
michael@0:     }
michael@0: 
michael@0:     return length;
michael@0: }
michael@0: 
michael@0: /* compare invariant strings; variant characters compare less than others and unlike each other */
michael@0: U_CFUNC int32_t
michael@0: uprv_compareInvAscii(const UDataSwapper *ds,
michael@0:                      const char *outString, int32_t outLength,
michael@0:                      const UChar *localString, int32_t localLength) {
michael@0:     int32_t minLength;
michael@0:     UChar32 c1, c2;
michael@0:     uint8_t c;
michael@0: 
michael@0:     if(outString==NULL || outLength<-1 || localString==NULL || localLength<-1) {
michael@0:         return 0;
michael@0:     }
michael@0: 
michael@0:     if(outLength<0) {
michael@0:         outLength=(int32_t)uprv_strlen(outString);
michael@0:     }
michael@0:     if(localLength<0) {
michael@0:         localLength=u_strlen(localString);
michael@0:     }
michael@0: 
michael@0:     minLength= outLength<localLength ? outLength : localLength;
michael@0: 
michael@0:     while(minLength>0) {
michael@0:         c=(uint8_t)*outString++;
michael@0:         if(UCHAR_IS_INVARIANT(c)) {
michael@0:             c1=c;
michael@0:         } else {
michael@0:             c1=-1;
michael@0:         }
michael@0: 
michael@0:         c2=*localString++;
michael@0:         if(!UCHAR_IS_INVARIANT(c2)) {
michael@0:             c2=-2;
michael@0:         }
michael@0: 
michael@0:         if((c1-=c2)!=0) {
michael@0:             return c1;
michael@0:         }
michael@0: 
michael@0:         --minLength;
michael@0:     }
michael@0: 
michael@0:     /* strings start with same prefix, compare lengths */
michael@0:     return outLength-localLength;
michael@0: }
michael@0: 
michael@0: U_CFUNC int32_t
michael@0: uprv_compareInvEbcdic(const UDataSwapper *ds,
michael@0:                       const char *outString, int32_t outLength,
michael@0:                       const UChar *localString, int32_t localLength) {
michael@0:     int32_t minLength;
michael@0:     UChar32 c1, c2;
michael@0:     uint8_t c;
michael@0: 
michael@0:     if(outString==NULL || outLength<-1 || localString==NULL || localLength<-1) {
michael@0:         return 0;
michael@0:     }
michael@0: 
michael@0:     if(outLength<0) {
michael@0:         outLength=(int32_t)uprv_strlen(outString);
michael@0:     }
michael@0:     if(localLength<0) {
michael@0:         localLength=u_strlen(localString);
michael@0:     }
michael@0: 
michael@0:     minLength= outLength<localLength ? outLength : localLength;
michael@0: 
michael@0:     while(minLength>0) {
michael@0:         c=(uint8_t)*outString++;
michael@0:         if(c==0) {
michael@0:             c1=0;
michael@0:         } else if((c1=asciiFromEbcdic[c])!=0 && UCHAR_IS_INVARIANT(c1)) {
michael@0:             /* c1 is set */
michael@0:         } else {
michael@0:             c1=-1;
michael@0:         }
michael@0: 
michael@0:         c2=*localString++;
michael@0:         if(!UCHAR_IS_INVARIANT(c2)) {
michael@0:             c2=-2;
michael@0:         }
michael@0: 
michael@0:         if((c1-=c2)!=0) {
michael@0:             return c1;
michael@0:         }
michael@0: 
michael@0:         --minLength;
michael@0:     }
michael@0: 
michael@0:     /* strings start with same prefix, compare lengths */
michael@0:     return outLength-localLength;
michael@0: }
michael@0: 
michael@0: U_CAPI int32_t U_EXPORT2
michael@0: uprv_compareInvEbcdicAsAscii(const char *s1, const char *s2) {
michael@0:     int32_t c1, c2;
michael@0: 
michael@0:     for(;; ++s1, ++s2) {
michael@0:         c1=(uint8_t)*s1;
michael@0:         c2=(uint8_t)*s2;
michael@0:         if(c1!=c2) {
michael@0:             if(c1!=0 && ((c1=asciiFromEbcdic[c1])==0 || !UCHAR_IS_INVARIANT(c1))) {
michael@0:                 c1=-(int32_t)(uint8_t)*s1;
michael@0:             }
michael@0:             if(c2!=0 && ((c2=asciiFromEbcdic[c2])==0 || !UCHAR_IS_INVARIANT(c2))) {
michael@0:                 c2=-(int32_t)(uint8_t)*s2;
michael@0:             }
michael@0:             return c1-c2;
michael@0:         } else if(c1==0) {
michael@0:             return 0;
michael@0:         }
michael@0:     }
michael@0: }
michael@0: 
michael@0: U_CAPI char U_EXPORT2
michael@0: uprv_ebcdicToLowercaseAscii(char c) {
michael@0:     return (char)lowercaseAsciiFromEbcdic[(uint8_t)c];
michael@0: }
michael@0: 
michael@0: U_INTERNAL uint8_t* U_EXPORT2
michael@0: uprv_aestrncpy(uint8_t *dst, const uint8_t *src, int32_t n)
michael@0: {
michael@0:   uint8_t *orig_dst = dst;
michael@0: 
michael@0:   if(n==-1) { 
michael@0:     n = uprv_strlen((const char*)src)+1; /* copy NUL */
michael@0:   }
michael@0:   /* copy non-null */
michael@0:   while(*src && n>0) {
michael@0:     *(dst++) = asciiFromEbcdic[*(src++)];
michael@0:     n--;
michael@0:   }
michael@0:   /* pad */
michael@0:   while(n>0) {
michael@0:     *(dst++) = 0;
michael@0:     n--;
michael@0:   }
michael@0:   return orig_dst;
michael@0: }
michael@0: 
michael@0: U_INTERNAL uint8_t* U_EXPORT2
michael@0: uprv_eastrncpy(uint8_t *dst, const uint8_t *src, int32_t n)
michael@0: {
michael@0:   uint8_t *orig_dst = dst;
michael@0: 
michael@0:   if(n==-1) { 
michael@0:     n = uprv_strlen((const char*)src)+1; /* copy NUL */
michael@0:   }
michael@0:   /* copy non-null */
michael@0:   while(*src && n>0) {
michael@0:     char ch = ebcdicFromAscii[*(src++)];
michael@0:     if(ch == 0) {
michael@0:       ch = ebcdicFromAscii[0x3f]; /* questionmark (subchar) */
michael@0:     }
michael@0:     *(dst++) = ch;
michael@0:     n--;
michael@0:   }
michael@0:   /* pad */
michael@0:   while(n>0) {
michael@0:     *(dst++) = 0;
michael@0:     n--;
michael@0:   }
michael@0:   return orig_dst;
michael@0: }
michael@0: