1.1 --- /dev/null Thu Jan 01 00:00:00 1970 +0000
1.2 +++ b/intl/icu/source/common/udataswp.c Wed Dec 31 06:09:35 2014 +0100
1.3 @@ -0,0 +1,415 @@
1.4 +/*
1.5 +*******************************************************************************
1.6 +*
1.7 +* Copyright (C) 2003, International Business Machines
1.8 +* Corporation and others. All Rights Reserved.
1.9 +*
1.10 +*******************************************************************************
1.11 +* file name: udataswp.c
1.12 +* encoding: US-ASCII
1.13 +* tab size: 8 (not used)
1.14 +* indentation:4
1.15 +*
1.16 +* created on: 2003jun05
1.17 +* created by: Markus W. Scherer
1.18 +*
1.19 +* Definitions for ICU data transformations for different platforms,
1.20 +* changing between big- and little-endian data and/or between
1.21 +* charset families (ASCII<->EBCDIC).
1.22 +*/
1.23 +
1.24 +#include <stdarg.h>
1.25 +#include "unicode/utypes.h"
1.26 +#include "unicode/udata.h" /* UDataInfo */
1.27 +#include "ucmndata.h" /* DataHeader */
1.28 +#include "cmemory.h"
1.29 +#include "udataswp.h"
1.30 +
1.31 +/* swapping primitives ------------------------------------------------------ */
1.32 +
1.33 +static int32_t U_CALLCONV
1.34 +uprv_swapArray16(const UDataSwapper *ds,
1.35 + const void *inData, int32_t length, void *outData,
1.36 + UErrorCode *pErrorCode) {
1.37 + const uint16_t *p;
1.38 + uint16_t *q;
1.39 + int32_t count;
1.40 + uint16_t x;
1.41 +
1.42 + if(pErrorCode==NULL || U_FAILURE(*pErrorCode)) {
1.43 + return 0;
1.44 + }
1.45 + if(ds==NULL || inData==NULL || length<0 || (length&1)!=0 || outData==NULL) {
1.46 + *pErrorCode=U_ILLEGAL_ARGUMENT_ERROR;
1.47 + return 0;
1.48 + }
1.49 +
1.50 + /* setup and swapping */
1.51 + p=(const uint16_t *)inData;
1.52 + q=(uint16_t *)outData;
1.53 + count=length/2;
1.54 + while(count>0) {
1.55 + x=*p++;
1.56 + *q++=(uint16_t)((x<<8)|(x>>8));
1.57 + --count;
1.58 + }
1.59 +
1.60 + return length;
1.61 +}
1.62 +
1.63 +static int32_t U_CALLCONV
1.64 +uprv_copyArray16(const UDataSwapper *ds,
1.65 + const void *inData, int32_t length, void *outData,
1.66 + UErrorCode *pErrorCode) {
1.67 + if(pErrorCode==NULL || U_FAILURE(*pErrorCode)) {
1.68 + return 0;
1.69 + }
1.70 + if(ds==NULL || inData==NULL || length<0 || (length&1)!=0 || outData==NULL) {
1.71 + *pErrorCode=U_ILLEGAL_ARGUMENT_ERROR;
1.72 + return 0;
1.73 + }
1.74 +
1.75 + if(length>0 && inData!=outData) {
1.76 + uprv_memcpy(outData, inData, length);
1.77 + }
1.78 + return length;
1.79 +}
1.80 +
1.81 +static int32_t U_CALLCONV
1.82 +uprv_swapArray32(const UDataSwapper *ds,
1.83 + const void *inData, int32_t length, void *outData,
1.84 + UErrorCode *pErrorCode) {
1.85 + const uint32_t *p;
1.86 + uint32_t *q;
1.87 + int32_t count;
1.88 + uint32_t x;
1.89 +
1.90 + if(pErrorCode==NULL || U_FAILURE(*pErrorCode)) {
1.91 + return 0;
1.92 + }
1.93 + if(ds==NULL || inData==NULL || length<0 || (length&3)!=0 || outData==NULL) {
1.94 + *pErrorCode=U_ILLEGAL_ARGUMENT_ERROR;
1.95 + return 0;
1.96 + }
1.97 +
1.98 + /* setup and swapping */
1.99 + p=(const uint32_t *)inData;
1.100 + q=(uint32_t *)outData;
1.101 + count=length/4;
1.102 + while(count>0) {
1.103 + x=*p++;
1.104 + *q++=(uint32_t)((x<<24)|((x<<8)&0xff0000)|((x>>8)&0xff00)|(x>>24));
1.105 + --count;
1.106 + }
1.107 +
1.108 + return length;
1.109 +}
1.110 +
1.111 +static int32_t U_CALLCONV
1.112 +uprv_copyArray32(const UDataSwapper *ds,
1.113 + const void *inData, int32_t length, void *outData,
1.114 + UErrorCode *pErrorCode) {
1.115 + if(pErrorCode==NULL || U_FAILURE(*pErrorCode)) {
1.116 + return 0;
1.117 + }
1.118 + if(ds==NULL || inData==NULL || length<0 || (length&3)!=0 || outData==NULL) {
1.119 + *pErrorCode=U_ILLEGAL_ARGUMENT_ERROR;
1.120 + return 0;
1.121 + }
1.122 +
1.123 + if(length>0 && inData!=outData) {
1.124 + uprv_memcpy(outData, inData, length);
1.125 + }
1.126 + return length;
1.127 +}
1.128 +
1.129 +static uint16_t U_CALLCONV
1.130 +uprv_readSwapUInt16(uint16_t x) {
1.131 + return (uint16_t)((x<<8)|(x>>8));
1.132 +}
1.133 +
1.134 +static uint16_t U_CALLCONV
1.135 +uprv_readDirectUInt16(uint16_t x) {
1.136 + return x;
1.137 +}
1.138 +
1.139 +static uint32_t U_CALLCONV
1.140 +uprv_readSwapUInt32(uint32_t x) {
1.141 + return (uint32_t)((x<<24)|((x<<8)&0xff0000)|((x>>8)&0xff00)|(x>>24));
1.142 +}
1.143 +
1.144 +static uint32_t U_CALLCONV
1.145 +uprv_readDirectUInt32(uint32_t x) {
1.146 + return x;
1.147 +}
1.148 +
1.149 +static void U_CALLCONV
1.150 +uprv_writeSwapUInt16(uint16_t *p, uint16_t x) {
1.151 + *p=(uint16_t)((x<<8)|(x>>8));
1.152 +}
1.153 +
1.154 +static void U_CALLCONV
1.155 +uprv_writeDirectUInt16(uint16_t *p, uint16_t x) {
1.156 + *p=x;
1.157 +}
1.158 +
1.159 +static void U_CALLCONV
1.160 +uprv_writeSwapUInt32(uint32_t *p, uint32_t x) {
1.161 + *p=(uint32_t)((x<<24)|((x<<8)&0xff0000)|((x>>8)&0xff00)|(x>>24));
1.162 +}
1.163 +
1.164 +static void U_CALLCONV
1.165 +uprv_writeDirectUInt32(uint32_t *p, uint32_t x) {
1.166 + *p=x;
1.167 +}
1.168 +
1.169 +U_CAPI int16_t U_EXPORT2
1.170 +udata_readInt16(const UDataSwapper *ds, int16_t x) {
1.171 + return (int16_t)ds->readUInt16((uint16_t)x);
1.172 +}
1.173 +
1.174 +U_CAPI int32_t U_EXPORT2
1.175 +udata_readInt32(const UDataSwapper *ds, int32_t x) {
1.176 + return (int32_t)ds->readUInt32((uint32_t)x);
1.177 +}
1.178 +
1.179 +/**
1.180 + * Swap a block of invariant, NUL-terminated strings, but not padding
1.181 + * bytes after the last string.
1.182 + * @internal
1.183 + */
1.184 +U_CAPI int32_t U_EXPORT2
1.185 +udata_swapInvStringBlock(const UDataSwapper *ds,
1.186 + const void *inData, int32_t length, void *outData,
1.187 + UErrorCode *pErrorCode) {
1.188 + const char *inChars;
1.189 + int32_t stringsLength;
1.190 +
1.191 + if(pErrorCode==NULL || U_FAILURE(*pErrorCode)) {
1.192 + return 0;
1.193 + }
1.194 + if(ds==NULL || inData==NULL || length<0 || (length>0 && outData==NULL)) {
1.195 + *pErrorCode=U_ILLEGAL_ARGUMENT_ERROR;
1.196 + return 0;
1.197 + }
1.198 +
1.199 + /* reduce the strings length to not include bytes after the last NUL */
1.200 + inChars=(const char *)inData;
1.201 + stringsLength=length;
1.202 + while(stringsLength>0 && inChars[stringsLength-1]!=0) {
1.203 + --stringsLength;
1.204 + }
1.205 +
1.206 + /* swap up to the last NUL */
1.207 + ds->swapInvChars(ds, inData, stringsLength, outData, pErrorCode);
1.208 +
1.209 + /* copy the bytes after the last NUL */
1.210 + if(inData!=outData && length>stringsLength) {
1.211 + uprv_memcpy((char *)outData+stringsLength, inChars+stringsLength, length-stringsLength);
1.212 + }
1.213 +
1.214 + /* return the length including padding bytes */
1.215 + if(U_SUCCESS(*pErrorCode)) {
1.216 + return length;
1.217 + } else {
1.218 + return 0;
1.219 + }
1.220 +}
1.221 +
1.222 +U_CAPI void U_EXPORT2
1.223 +udata_printError(const UDataSwapper *ds,
1.224 + const char *fmt,
1.225 + ...) {
1.226 + va_list args;
1.227 +
1.228 + if(ds->printError!=NULL) {
1.229 + va_start(args, fmt);
1.230 + ds->printError(ds->printErrorContext, fmt, args);
1.231 + va_end(args);
1.232 + }
1.233 +}
1.234 +
1.235 +/* swap a data header ------------------------------------------------------- */
1.236 +
1.237 +U_CAPI int32_t U_EXPORT2
1.238 +udata_swapDataHeader(const UDataSwapper *ds,
1.239 + const void *inData, int32_t length, void *outData,
1.240 + UErrorCode *pErrorCode) {
1.241 + const DataHeader *pHeader;
1.242 + uint16_t headerSize, infoSize;
1.243 +
1.244 + /* argument checking */
1.245 + if(pErrorCode==NULL || U_FAILURE(*pErrorCode)) {
1.246 + return 0;
1.247 + }
1.248 + if(ds==NULL || inData==NULL || length<-1 || (length>0 && outData==NULL)) {
1.249 + *pErrorCode=U_ILLEGAL_ARGUMENT_ERROR;
1.250 + return 0;
1.251 + }
1.252 +
1.253 + /* check minimum length and magic bytes */
1.254 + pHeader=(const DataHeader *)inData;
1.255 + if( (length>=0 && length<sizeof(DataHeader)) ||
1.256 + pHeader->dataHeader.magic1!=0xda ||
1.257 + pHeader->dataHeader.magic2!=0x27 ||
1.258 + pHeader->info.sizeofUChar!=2
1.259 + ) {
1.260 + udata_printError(ds, "udata_swapDataHeader(): initial bytes do not look like ICU data\n");
1.261 + *pErrorCode=U_UNSUPPORTED_ERROR;
1.262 + return 0;
1.263 + }
1.264 +
1.265 + headerSize=ds->readUInt16(pHeader->dataHeader.headerSize);
1.266 + infoSize=ds->readUInt16(pHeader->info.size);
1.267 +
1.268 + if( headerSize<sizeof(DataHeader) ||
1.269 + infoSize<sizeof(UDataInfo) ||
1.270 + headerSize<(sizeof(pHeader->dataHeader)+infoSize) ||
1.271 + (length>=0 && length<headerSize)
1.272 + ) {
1.273 + udata_printError(ds, "udata_swapDataHeader(): header size mismatch - headerSize %d infoSize %d length %d\n",
1.274 + headerSize, infoSize, length);
1.275 + *pErrorCode=U_INDEX_OUTOFBOUNDS_ERROR;
1.276 + return 0;
1.277 + }
1.278 +
1.279 + if(length>0) {
1.280 + DataHeader *outHeader;
1.281 + const char *s;
1.282 + int32_t maxLength;
1.283 +
1.284 + /* Most of the fields are just bytes and need no swapping. */
1.285 + if(inData!=outData) {
1.286 + uprv_memcpy(outData, inData, headerSize);
1.287 + }
1.288 + outHeader=(DataHeader *)outData;
1.289 +
1.290 + outHeader->info.isBigEndian = ds->outIsBigEndian;
1.291 + outHeader->info.charsetFamily = ds->outCharset;
1.292 +
1.293 + /* swap headerSize */
1.294 + ds->swapArray16(ds, &pHeader->dataHeader.headerSize, 2, &outHeader->dataHeader.headerSize, pErrorCode);
1.295 +
1.296 + /* swap UDataInfo size and reservedWord */
1.297 + ds->swapArray16(ds, &pHeader->info.size, 4, &outHeader->info.size, pErrorCode);
1.298 +
1.299 + /* swap copyright statement after the UDataInfo */
1.300 + infoSize+=sizeof(pHeader->dataHeader);
1.301 + s=(const char *)inData+infoSize;
1.302 + maxLength=headerSize-infoSize;
1.303 + /* get the length of the string */
1.304 + for(length=0; length<maxLength && s[length]!=0; ++length) {}
1.305 + /* swap the string contents */
1.306 + ds->swapInvChars(ds, s, length, (char *)outData+infoSize, pErrorCode);
1.307 + }
1.308 +
1.309 + return headerSize;
1.310 +}
1.311 +
1.312 +/* API functions ------------------------------------------------------------ */
1.313 +
1.314 +U_CAPI UDataSwapper * U_EXPORT2
1.315 +udata_openSwapper(UBool inIsBigEndian, uint8_t inCharset,
1.316 + UBool outIsBigEndian, uint8_t outCharset,
1.317 + UErrorCode *pErrorCode) {
1.318 + UDataSwapper *swapper;
1.319 +
1.320 + if(pErrorCode==NULL || U_FAILURE(*pErrorCode)) {
1.321 + return NULL;
1.322 + }
1.323 + if(inCharset>U_EBCDIC_FAMILY || outCharset>U_EBCDIC_FAMILY) {
1.324 + *pErrorCode=U_ILLEGAL_ARGUMENT_ERROR;
1.325 + return NULL;
1.326 + }
1.327 +
1.328 + /* allocate the swapper */
1.329 + swapper=uprv_malloc(sizeof(UDataSwapper));
1.330 + if(swapper==NULL) {
1.331 + *pErrorCode=U_MEMORY_ALLOCATION_ERROR;
1.332 + return NULL;
1.333 + }
1.334 + uprv_memset(swapper, 0, sizeof(UDataSwapper));
1.335 +
1.336 + /* set values and functions pointers according to in/out parameters */
1.337 + swapper->inIsBigEndian=inIsBigEndian;
1.338 + swapper->inCharset=inCharset;
1.339 + swapper->outIsBigEndian=outIsBigEndian;
1.340 + swapper->outCharset=outCharset;
1.341 +
1.342 + swapper->readUInt16= inIsBigEndian==U_IS_BIG_ENDIAN ? uprv_readDirectUInt16 : uprv_readSwapUInt16;
1.343 + swapper->readUInt32= inIsBigEndian==U_IS_BIG_ENDIAN ? uprv_readDirectUInt32 : uprv_readSwapUInt32;
1.344 +
1.345 + swapper->writeUInt16= outIsBigEndian==U_IS_BIG_ENDIAN ? uprv_writeDirectUInt16 : uprv_writeSwapUInt16;
1.346 + swapper->writeUInt32= outIsBigEndian==U_IS_BIG_ENDIAN ? uprv_writeDirectUInt32 : uprv_writeSwapUInt32;
1.347 +
1.348 + swapper->compareInvChars= outCharset==U_ASCII_FAMILY ? uprv_compareInvAscii : uprv_compareInvEbcdic;
1.349 +
1.350 + swapper->swapArray16= inIsBigEndian==outIsBigEndian ? uprv_copyArray16 : uprv_swapArray16;
1.351 + swapper->swapArray32= inIsBigEndian==outIsBigEndian ? uprv_copyArray32 : uprv_swapArray32;
1.352 +
1.353 + if(inCharset==U_ASCII_FAMILY) {
1.354 + swapper->swapInvChars= outCharset==U_ASCII_FAMILY ? uprv_copyAscii : uprv_ebcdicFromAscii;
1.355 + } else /* U_EBCDIC_FAMILY */ {
1.356 + swapper->swapInvChars= outCharset==U_EBCDIC_FAMILY ? uprv_copyEbcdic : uprv_asciiFromEbcdic;
1.357 + }
1.358 +
1.359 + return swapper;
1.360 +}
1.361 +
1.362 +U_CAPI UDataSwapper * U_EXPORT2
1.363 +udata_openSwapperForInputData(const void *data, int32_t length,
1.364 + UBool outIsBigEndian, uint8_t outCharset,
1.365 + UErrorCode *pErrorCode) {
1.366 + const DataHeader *pHeader;
1.367 + uint16_t headerSize, infoSize;
1.368 + UBool inIsBigEndian;
1.369 + int8_t inCharset;
1.370 +
1.371 + if(pErrorCode==NULL || U_FAILURE(*pErrorCode)) {
1.372 + return NULL;
1.373 + }
1.374 + if( data==NULL ||
1.375 + (length>=0 && length<sizeof(DataHeader)) ||
1.376 + outCharset>U_EBCDIC_FAMILY
1.377 + ) {
1.378 + *pErrorCode=U_ILLEGAL_ARGUMENT_ERROR;
1.379 + return NULL;
1.380 + }
1.381 +
1.382 + pHeader=(const DataHeader *)data;
1.383 + if( (length>=0 && length<sizeof(DataHeader)) ||
1.384 + pHeader->dataHeader.magic1!=0xda ||
1.385 + pHeader->dataHeader.magic2!=0x27 ||
1.386 + pHeader->info.sizeofUChar!=2
1.387 + ) {
1.388 + *pErrorCode=U_UNSUPPORTED_ERROR;
1.389 + return 0;
1.390 + }
1.391 +
1.392 + inIsBigEndian=(UBool)pHeader->info.isBigEndian;
1.393 + inCharset=pHeader->info.charsetFamily;
1.394 +
1.395 + if(inIsBigEndian==U_IS_BIG_ENDIAN) {
1.396 + headerSize=pHeader->dataHeader.headerSize;
1.397 + infoSize=pHeader->info.size;
1.398 + } else {
1.399 + headerSize=uprv_readSwapUInt16(pHeader->dataHeader.headerSize);
1.400 + infoSize=uprv_readSwapUInt16(pHeader->info.size);
1.401 + }
1.402 +
1.403 + if( headerSize<sizeof(DataHeader) ||
1.404 + infoSize<sizeof(UDataInfo) ||
1.405 + headerSize<(sizeof(pHeader->dataHeader)+infoSize) ||
1.406 + (length>=0 && length<headerSize)
1.407 + ) {
1.408 + *pErrorCode=U_UNSUPPORTED_ERROR;
1.409 + return 0;
1.410 + }
1.411 +
1.412 + return udata_openSwapper(inIsBigEndian, inCharset, outIsBigEndian, outCharset, pErrorCode);
1.413 +}
1.414 +
1.415 +U_CAPI void U_EXPORT2
1.416 +udata_closeSwapper(UDataSwapper *ds) {
1.417 + uprv_free(ds);
1.418 +}
