1.1 --- /dev/null Thu Jan 01 00:00:00 1970 +0000
1.2 +++ b/intl/icu/source/tools/toolutil/ucbuf.c Wed Dec 31 06:09:35 2014 +0100
1.3 @@ -0,0 +1,787 @@
1.4 +/*
1.5 +*******************************************************************************
1.6 +*
1.7 +* Copyright (C) 1998-2011, International Business Machines
1.8 +* Corporation and others. All Rights Reserved.
1.9 +*
1.10 +*******************************************************************************
1.11 +*
1.12 +* File ucbuf.c
1.13 +*
1.14 +* Modification History:
1.15 +*
1.16 +* Date Name Description
1.17 +* 05/10/01 Ram Creation.
1.18 +*******************************************************************************
1.19 +*/
1.20 +
1.21 +#include "unicode/utypes.h"
1.22 +#include "unicode/putil.h"
1.23 +#include "unicode/uchar.h"
1.24 +#include "unicode/ucnv.h"
1.25 +#include "unicode/ucnv_err.h"
1.26 +#include "unicode/ustring.h"
1.27 +#include "unicode/utf16.h"
1.28 +#include "filestrm.h"
1.29 +#include "cstring.h"
1.30 +#include "cmemory.h"
1.31 +#include "ustrfmt.h"
1.32 +#include "ucbuf.h"
1.33 +#include <stdio.h>
1.34 +
1.35 +#if !UCONFIG_NO_CONVERSION
1.36 +
1.37 +
1.38 +#define MAX_IN_BUF 1000
1.39 +#define MAX_U_BUF 1500
1.40 +#define CONTEXT_LEN 20
1.41 +
1.42 +struct UCHARBUF {
1.43 + UChar* buffer;
1.44 + UChar* currentPos;
1.45 + UChar* bufLimit;
1.46 + int32_t bufCapacity;
1.47 + int32_t remaining;
1.48 + int32_t signatureLength;
1.49 + FileStream* in;
1.50 + UConverter* conv;
1.51 + UBool showWarning; /* makes this API not produce any errors */
1.52 + UBool isBuffered;
1.53 +};
1.54 +
1.55 +U_CAPI UBool U_EXPORT2
1.56 +ucbuf_autodetect_fs(FileStream* in, const char** cp, UConverter** conv, int32_t* signatureLength, UErrorCode* error){
1.57 + char start[8];
1.58 + int32_t numRead;
1.59 +
1.60 + UChar target[1]={ 0 };
1.61 + UChar* pTarget;
1.62 + const char* pStart;
1.63 +
1.64 + /* read a few bytes */
1.65 + numRead=T_FileStream_read(in, start, sizeof(start));
1.66 +
1.67 + *cp = ucnv_detectUnicodeSignature(start, numRead, signatureLength, error);
1.68 +
1.69 + /* unread the bytes beyond what was consumed for U+FEFF */
1.70 + T_FileStream_rewind(in);
1.71 + if (*signatureLength > 0) {
1.72 + T_FileStream_read(in, start, *signatureLength);
1.73 + }
1.74 +
1.75 + if(*cp==NULL){
1.76 + *conv =NULL;
1.77 + return FALSE;
1.78 + }
1.79 +
1.80 + /* open the converter for the detected Unicode charset */
1.81 + *conv = ucnv_open(*cp,error);
1.82 +
1.83 + /* convert and ignore initial U+FEFF, and the buffer overflow */
1.84 + pTarget = target;
1.85 + pStart = start;
1.86 + ucnv_toUnicode(*conv, &pTarget, target+1, &pStart, start+*signatureLength, NULL, FALSE, error);
1.87 + *signatureLength = (int32_t)(pStart - start);
1.88 + if(*error==U_BUFFER_OVERFLOW_ERROR) {
1.89 + *error=U_ZERO_ERROR;
1.90 + }
1.91 +
1.92 + /* verify that we successfully read exactly U+FEFF */
1.93 + if(U_SUCCESS(*error) && (pTarget!=(target+1) || target[0]!=0xfeff)) {
1.94 + *error=U_INTERNAL_PROGRAM_ERROR;
1.95 + }
1.96 +
1.97 +
1.98 + return TRUE;
1.99 +}
1.100 +static UBool ucbuf_isCPKnown(const char* cp){
1.101 + if(ucnv_compareNames("UTF-8",cp)==0){
1.102 + return TRUE;
1.103 + }
1.104 + if(ucnv_compareNames("UTF-16BE",cp)==0){
1.105 + return TRUE;
1.106 + }
1.107 + if(ucnv_compareNames("UTF-16LE",cp)==0){
1.108 + return TRUE;
1.109 + }
1.110 + if(ucnv_compareNames("UTF-16",cp)==0){
1.111 + return TRUE;
1.112 + }
1.113 + if(ucnv_compareNames("UTF-32",cp)==0){
1.114 + return TRUE;
1.115 + }
1.116 + if(ucnv_compareNames("UTF-32BE",cp)==0){
1.117 + return TRUE;
1.118 + }
1.119 + if(ucnv_compareNames("UTF-32LE",cp)==0){
1.120 + return TRUE;
1.121 + }
1.122 + if(ucnv_compareNames("SCSU",cp)==0){
1.123 + return TRUE;
1.124 + }
1.125 + if(ucnv_compareNames("BOCU-1",cp)==0){
1.126 + return TRUE;
1.127 + }
1.128 + if(ucnv_compareNames("UTF-7",cp)==0){
1.129 + return TRUE;
1.130 + }
1.131 + return FALSE;
1.132 +}
1.133 +
1.134 +U_CAPI FileStream * U_EXPORT2
1.135 +ucbuf_autodetect(const char* fileName, const char** cp,UConverter** conv, int32_t* signatureLength,UErrorCode* error){
1.136 + FileStream* in=NULL;
1.137 + if(error==NULL || U_FAILURE(*error)){
1.138 + return NULL;
1.139 + }
1.140 + if(conv==NULL || cp==NULL || fileName==NULL){
1.141 + *error = U_ILLEGAL_ARGUMENT_ERROR;
1.142 + return NULL;
1.143 + }
1.144 + /* open the file */
1.145 + in= T_FileStream_open(fileName,"rb");
1.146 +
1.147 + if(in == NULL){
1.148 + *error=U_FILE_ACCESS_ERROR;
1.149 + return NULL;
1.150 + }
1.151 +
1.152 + if(ucbuf_autodetect_fs(in,cp,conv,signatureLength,error)) {
1.153 + return in;
1.154 + } else {
1.155 + ucnv_close(*conv);
1.156 + *conv=NULL;
1.157 + T_FileStream_close(in);
1.158 + return NULL;
1.159 + }
1.160 +}
1.161 +
1.162 +/* fill the uchar buffer */
1.163 +static UCHARBUF*
1.164 +ucbuf_fillucbuf( UCHARBUF* buf,UErrorCode* error){
1.165 + UChar* pTarget=NULL;
1.166 + UChar* target=NULL;
1.167 + const char* source=NULL;
1.168 + char carr[MAX_IN_BUF] = {'\0'};
1.169 + char* cbuf = carr;
1.170 + int32_t inputRead=0;
1.171 + int32_t outputWritten=0;
1.172 + int32_t offset=0;
1.173 + const char* sourceLimit =NULL;
1.174 + int32_t cbufSize=0;
1.175 + pTarget = buf->buffer;
1.176 + /* check if we arrived here without exhausting the buffer*/
1.177 + if(buf->currentPos<buf->bufLimit){
1.178 + offset = (int32_t)(buf->bufLimit-buf->currentPos);
1.179 + memmove(buf->buffer,buf->currentPos,offset* sizeof(UChar));
1.180 + }
1.181 +
1.182 +#if DEBUG
1.183 + memset(pTarget+offset,0xff,sizeof(UChar)*(MAX_IN_BUF-offset));
1.184 +#endif
1.185 + if(buf->isBuffered){
1.186 + cbufSize = MAX_IN_BUF;
1.187 + /* read the file */
1.188 + inputRead=T_FileStream_read(buf->in,cbuf,cbufSize-offset);
1.189 + buf->remaining-=inputRead;
1.190 +
1.191 + }else{
1.192 + cbufSize = T_FileStream_size(buf->in);
1.193 + cbuf = (char*)uprv_malloc(cbufSize);
1.194 + if (cbuf == NULL) {
1.195 + *error = U_MEMORY_ALLOCATION_ERROR;
1.196 + return NULL;
1.197 + }
1.198 + inputRead= T_FileStream_read(buf->in,cbuf,cbufSize);
1.199 + buf->remaining-=inputRead;
1.200 + }
1.201 +
1.202 + /* just to be sure...*/
1.203 + if ( 0 == inputRead )
1.204 + buf->remaining = 0;
1.205 +
1.206 + target=pTarget;
1.207 + /* convert the bytes */
1.208 + if(buf->conv){
1.209 + /* set the callback to stop */
1.210 + UConverterToUCallback toUOldAction ;
1.211 + void* toUOldContext;
1.212 + void* toUNewContext=NULL;
1.213 + ucnv_setToUCallBack(buf->conv,
1.214 + UCNV_TO_U_CALLBACK_STOP,
1.215 + toUNewContext,
1.216 + &toUOldAction,
1.217 + (const void**)&toUOldContext,
1.218 + error);
1.219 + /* since state is saved in the converter we add offset to source*/
1.220 + target = pTarget+offset;
1.221 + source = cbuf;
1.222 + sourceLimit = source + inputRead;
1.223 + ucnv_toUnicode(buf->conv,&target,target+(buf->bufCapacity-offset),
1.224 + &source,sourceLimit,NULL,
1.225 + (UBool)(buf->remaining==0),error);
1.226 +
1.227 + if(U_FAILURE(*error)){
1.228 + char context[CONTEXT_LEN+1];
1.229 + char preContext[CONTEXT_LEN+1];
1.230 + char postContext[CONTEXT_LEN+1];
1.231 + int8_t len = CONTEXT_LEN;
1.232 + int32_t start=0;
1.233 + int32_t stop =0;
1.234 + int32_t pos =0;
1.235 + /* use erro1 to preserve the error code */
1.236 + UErrorCode error1 =U_ZERO_ERROR;
1.237 +
1.238 + if( buf->showWarning==TRUE){
1.239 + fprintf(stderr,"\n###WARNING: Encountered abnormal bytes while"
1.240 + " converting input stream to target encoding: %s\n",
1.241 + u_errorName(*error));
1.242 + }
1.243 +
1.244 +
1.245 + /* now get the context chars */
1.246 + ucnv_getInvalidChars(buf->conv,context,&len,&error1);
1.247 + context[len]= 0 ; /* null terminate the buffer */
1.248 +
1.249 + pos = (int32_t)(source - cbuf - len);
1.250 +
1.251 + /* for pre-context */
1.252 + start = (pos <=CONTEXT_LEN)? 0 : (pos - (CONTEXT_LEN-1));
1.253 + stop = pos-len;
1.254 +
1.255 + memcpy(preContext,cbuf+start,stop-start);
1.256 + /* null terminate the buffer */
1.257 + preContext[stop-start] = 0;
1.258 +
1.259 + /* for post-context */
1.260 + start = pos+len;
1.261 + stop = (int32_t)(((pos+CONTEXT_LEN)<= (sourceLimit-cbuf) )? (pos+(CONTEXT_LEN-1)) : (sourceLimit-cbuf));
1.262 +
1.263 + memcpy(postContext,source,stop-start);
1.264 + /* null terminate the buffer */
1.265 + postContext[stop-start] = 0;
1.266 +
1.267 + if(buf->showWarning ==TRUE){
1.268 + /* print out the context */
1.269 + fprintf(stderr,"\tPre-context: %s\n",preContext);
1.270 + fprintf(stderr,"\tContext: %s\n",context);
1.271 + fprintf(stderr,"\tPost-context: %s\n", postContext);
1.272 + }
1.273 +
1.274 + /* reset the converter */
1.275 + ucnv_reset(buf->conv);
1.276 +
1.277 + /* set the call back to substitute
1.278 + * and restart conversion
1.279 + */
1.280 + ucnv_setToUCallBack(buf->conv,
1.281 + UCNV_TO_U_CALLBACK_SUBSTITUTE,
1.282 + toUNewContext,
1.283 + &toUOldAction,
1.284 + (const void**)&toUOldContext,
1.285 + &error1);
1.286 +
1.287 + /* reset source and target start positions */
1.288 + target = pTarget+offset;
1.289 + source = cbuf;
1.290 +
1.291 + /* re convert */
1.292 + ucnv_toUnicode(buf->conv,&target,target+(buf->bufCapacity-offset),
1.293 + &source,sourceLimit,NULL,
1.294 + (UBool)(buf->remaining==0),&error1);
1.295 +
1.296 + }
1.297 + outputWritten = (int32_t)(target - pTarget);
1.298 +
1.299 +
1.300 +#if DEBUG
1.301 + {
1.302 + int i;
1.303 + target = pTarget;
1.304 + for(i=0;i<numRead;i++){
1.305 + /* printf("%c", (char)(*target++));*/
1.306 + }
1.307 + }
1.308 +#endif
1.309 +
1.310 + }else{
1.311 + u_charsToUChars(cbuf,target+offset,inputRead);
1.312 + outputWritten=((buf->remaining>cbufSize)? cbufSize:inputRead+offset);
1.313 + }
1.314 + buf->currentPos = pTarget;
1.315 + buf->bufLimit=pTarget+outputWritten;
1.316 + *buf->bufLimit=0; /*NUL terminate*/
1.317 + if(cbuf!=carr){
1.318 + uprv_free(cbuf);
1.319 + }
1.320 + return buf;
1.321 +}
1.322 +
1.323 +
1.324 +
1.325 +/* get a UChar from the stream*/
1.326 +U_CAPI int32_t U_EXPORT2
1.327 +ucbuf_getc(UCHARBUF* buf,UErrorCode* error){
1.328 + if(error==NULL || U_FAILURE(*error)){
1.329 + return FALSE;
1.330 + }
1.331 + if(buf->currentPos>=buf->bufLimit){
1.332 + if(buf->remaining==0){
1.333 + return U_EOF;
1.334 + }
1.335 + buf=ucbuf_fillucbuf(buf,error);
1.336 + if(U_FAILURE(*error)){
1.337 + return U_EOF;
1.338 + }
1.339 + }
1.340 +
1.341 + return *(buf->currentPos++);
1.342 +}
1.343 +
1.344 +/* get a UChar32 from the stream*/
1.345 +U_CAPI int32_t U_EXPORT2
1.346 +ucbuf_getc32(UCHARBUF* buf,UErrorCode* error){
1.347 + int32_t retVal = (int32_t)U_EOF;
1.348 + if(error==NULL || U_FAILURE(*error)){
1.349 + return FALSE;
1.350 + }
1.351 + if(buf->currentPos+1>=buf->bufLimit){
1.352 + if(buf->remaining==0){
1.353 + return U_EOF;
1.354 + }
1.355 + buf=ucbuf_fillucbuf(buf,error);
1.356 + if(U_FAILURE(*error)){
1.357 + return U_EOF;
1.358 + }
1.359 + }
1.360 + if(U16_IS_LEAD(*(buf->currentPos))){
1.361 + retVal=U16_GET_SUPPLEMENTARY(buf->currentPos[0],buf->currentPos[1]);
1.362 + buf->currentPos+=2;
1.363 + }else{
1.364 + retVal = *(buf->currentPos++);
1.365 + }
1.366 + return retVal;
1.367 +}
1.368 +
1.369 +/* u_unescapeAt() callback to return a UChar*/
1.370 +static UChar U_CALLCONV
1.371 +_charAt(int32_t offset, void *context) {
1.372 + return ((UCHARBUF*) context)->currentPos[offset];
1.373 +}
1.374 +
1.375 +/* getc and escape it */
1.376 +U_CAPI int32_t U_EXPORT2
1.377 +ucbuf_getcx32(UCHARBUF* buf,UErrorCode* error) {
1.378 + int32_t length;
1.379 + int32_t offset;
1.380 + UChar32 c32,c1,c2;
1.381 + if(error==NULL || U_FAILURE(*error)){
1.382 + return FALSE;
1.383 + }
1.384 + /* Fill the buffer if it is empty */
1.385 + if (buf->currentPos >=buf->bufLimit-2) {
1.386 + ucbuf_fillucbuf(buf,error);
1.387 + }
1.388 +
1.389 + /* Get the next character in the buffer */
1.390 + if (buf->currentPos < buf->bufLimit) {
1.391 + c1 = *(buf->currentPos)++;
1.392 + } else {
1.393 + c1 = U_EOF;
1.394 + }
1.395 +
1.396 + c2 = *(buf->currentPos);
1.397 +
1.398 + /* If it isn't a backslash, return it */
1.399 + if (c1 != 0x005C) {
1.400 + return c1;
1.401 + }
1.402 +
1.403 + /* Determine the amount of data in the buffer */
1.404 + length = (int32_t)(buf->bufLimit - buf->currentPos);
1.405 +
1.406 + /* The longest escape sequence is \Uhhhhhhhh; make sure
1.407 + we have at least that many characters */
1.408 + if (length < 10) {
1.409 +
1.410 + /* fill the buffer */
1.411 + ucbuf_fillucbuf(buf,error);
1.412 + length = (int32_t)(buf->bufLimit - buf->buffer);
1.413 + }
1.414 +
1.415 + /* Process the escape */
1.416 + offset = 0;
1.417 + c32 = u_unescapeAt(_charAt, &offset, length, (void*)buf);
1.418 +
1.419 + /* check if u_unescapeAt unescaped and converted
1.420 + * to c32 or not
1.421 + */
1.422 + if(c32==0xFFFFFFFF){
1.423 + if(buf->showWarning) {
1.424 + char context[CONTEXT_LEN+1];
1.425 + int32_t len = CONTEXT_LEN;
1.426 + if(length < len) {
1.427 + len = length;
1.428 + }
1.429 + context[len]= 0 ; /* null terminate the buffer */
1.430 + u_UCharsToChars( buf->currentPos, context, len);
1.431 + fprintf(stderr,"Bad escape: [%c%s]...\n", (int)c1, context);
1.432 + }
1.433 + *error= U_ILLEGAL_ESCAPE_SEQUENCE;
1.434 + return c1;
1.435 + }else if(c32!=c2 || (c32==0x0075 && c2==0x0075 && c1==0x005C) /* for \u0075 c2=0x0075 and c32==0x0075*/){
1.436 + /* Update the current buffer position */
1.437 + buf->currentPos += offset;
1.438 + }else{
1.439 + /* unescaping failed so we just return
1.440 + * c1 and not consume the buffer
1.441 + * this is useful for rules with escapes
1.442 + * in resouce bundles
1.443 + * eg: \' \\ \"
1.444 + */
1.445 + return c1;
1.446 + }
1.447 +
1.448 + return c32;
1.449 +}
1.450 +
1.451 +U_CAPI UCHARBUF* U_EXPORT2
1.452 +ucbuf_open(const char* fileName,const char** cp,UBool showWarning, UBool buffered, UErrorCode* error){
1.453 +
1.454 + FileStream* in = NULL;
1.455 + int32_t fileSize=0;
1.456 + const char* knownCp;
1.457 + if(error==NULL || U_FAILURE(*error)){
1.458 + return NULL;
1.459 + }
1.460 + if(cp==NULL || fileName==NULL){
1.461 + *error = U_ILLEGAL_ARGUMENT_ERROR;
1.462 + return FALSE;
1.463 + }
1.464 + if (!uprv_strcmp(fileName, "-")) {
1.465 + in = T_FileStream_stdin();
1.466 + }else{
1.467 + in = T_FileStream_open(fileName, "rb");
1.468 + }
1.469 +
1.470 + if(in!=NULL){
1.471 + UCHARBUF* buf =(UCHARBUF*) uprv_malloc(sizeof(UCHARBUF));
1.472 + fileSize = T_FileStream_size(in);
1.473 + if(buf == NULL){
1.474 + *error = U_MEMORY_ALLOCATION_ERROR;
1.475 + T_FileStream_close(in);
1.476 + return NULL;
1.477 + }
1.478 + buf->in=in;
1.479 + buf->conv=NULL;
1.480 + buf->showWarning = showWarning;
1.481 + buf->isBuffered = buffered;
1.482 + buf->signatureLength=0;
1.483 + if(*cp==NULL || **cp=='\0'){
1.484 + /* don't have code page name... try to autodetect */
1.485 + ucbuf_autodetect_fs(in,cp,&buf->conv,&buf->signatureLength,error);
1.486 + }else if(ucbuf_isCPKnown(*cp)){
1.487 + /* discard BOM */
1.488 + ucbuf_autodetect_fs(in,&knownCp,&buf->conv,&buf->signatureLength,error);
1.489 + }
1.490 + if(U_SUCCESS(*error) && buf->conv==NULL) {
1.491 + buf->conv=ucnv_open(*cp,error);
1.492 + }
1.493 + if(U_FAILURE(*error)){
1.494 + ucnv_close(buf->conv);
1.495 + uprv_free(buf);
1.496 + T_FileStream_close(in);
1.497 + return NULL;
1.498 + }
1.499 +
1.500 + if((buf->conv==NULL) && (buf->showWarning==TRUE)){
1.501 + fprintf(stderr,"###WARNING: No converter defined. Using codepage of system.\n");
1.502 + }
1.503 + buf->remaining=fileSize-buf->signatureLength;
1.504 + if(buf->isBuffered){
1.505 + buf->bufCapacity=MAX_U_BUF;
1.506 + }else{
1.507 + buf->bufCapacity=buf->remaining+buf->signatureLength+1/*for terminating nul*/;
1.508 + }
1.509 + buf->buffer=(UChar*) uprv_malloc(U_SIZEOF_UCHAR * buf->bufCapacity );
1.510 + if (buf->buffer == NULL) {
1.511 + *error = U_MEMORY_ALLOCATION_ERROR;
1.512 + ucbuf_close(buf);
1.513 + return NULL;
1.514 + }
1.515 + buf->currentPos=buf->buffer;
1.516 + buf->bufLimit=buf->buffer;
1.517 + if(U_FAILURE(*error)){
1.518 + fprintf(stderr, "Could not open codepage [%s]: %s\n", *cp, u_errorName(*error));
1.519 + ucbuf_close(buf);
1.520 + return NULL;
1.521 + }
1.522 + ucbuf_fillucbuf(buf,error);
1.523 + if(U_FAILURE(*error)){
1.524 + ucbuf_close(buf);
1.525 + return NULL;
1.526 + }
1.527 + return buf;
1.528 + }
1.529 + *error =U_FILE_ACCESS_ERROR;
1.530 + return NULL;
1.531 +}
1.532 +
1.533 +
1.534 +
1.535 +/* TODO: this method will fail if at the
1.536 + * begining of buffer and the uchar to unget
1.537 + * is from the previous buffer. Need to implement
1.538 + * system to take care of that situation.
1.539 + */
1.540 +U_CAPI void U_EXPORT2
1.541 +ucbuf_ungetc(int32_t c,UCHARBUF* buf){
1.542 + /* decrement currentPos pointer
1.543 + * if not at the begining of buffer
1.544 + */
1.545 + if(buf->currentPos!=buf->buffer){
1.546 + if(*(buf->currentPos-1)==c){
1.547 + buf->currentPos--;
1.548 + } else {
1.549 + /* ungetc failed - did not match. */
1.550 + }
1.551 + } else {
1.552 + /* ungetc failed - beginning of buffer. */
1.553 + }
1.554 +}
1.555 +
1.556 +/* frees the resources of UChar* buffer */
1.557 +static void
1.558 +ucbuf_closebuf(UCHARBUF* buf){
1.559 + uprv_free(buf->buffer);
1.560 + buf->buffer = NULL;
1.561 +}
1.562 +
1.563 +/* close the buf and release resources*/
1.564 +U_CAPI void U_EXPORT2
1.565 +ucbuf_close(UCHARBUF* buf){
1.566 + if(buf!=NULL){
1.567 + if(buf->conv){
1.568 + ucnv_close(buf->conv);
1.569 + }
1.570 + T_FileStream_close(buf->in);
1.571 + ucbuf_closebuf(buf);
1.572 + uprv_free(buf);
1.573 + }
1.574 +}
1.575 +
1.576 +/* rewind the buf and file stream */
1.577 +U_CAPI void U_EXPORT2
1.578 +ucbuf_rewind(UCHARBUF* buf,UErrorCode* error){
1.579 + if(error==NULL || U_FAILURE(*error)){
1.580 + return;
1.581 + }
1.582 + if(buf){
1.583 + buf->currentPos=buf->buffer;
1.584 + buf->bufLimit=buf->buffer;
1.585 + T_FileStream_rewind(buf->in);
1.586 + buf->remaining=T_FileStream_size(buf->in)-buf->signatureLength;
1.587 +
1.588 + ucnv_resetToUnicode(buf->conv);
1.589 + if(buf->signatureLength>0) {
1.590 + UChar target[1]={ 0 };
1.591 + UChar* pTarget;
1.592 + char start[8];
1.593 + const char* pStart;
1.594 + int32_t numRead;
1.595 +
1.596 + /* read the signature bytes */
1.597 + numRead=T_FileStream_read(buf->in, start, buf->signatureLength);
1.598 +
1.599 + /* convert and ignore initial U+FEFF, and the buffer overflow */
1.600 + pTarget = target;
1.601 + pStart = start;
1.602 + ucnv_toUnicode(buf->conv, &pTarget, target+1, &pStart, start+numRead, NULL, FALSE, error);
1.603 + if(*error==U_BUFFER_OVERFLOW_ERROR) {
1.604 + *error=U_ZERO_ERROR;
1.605 + }
1.606 +
1.607 + /* verify that we successfully read exactly U+FEFF */
1.608 + if(U_SUCCESS(*error) && (numRead!=buf->signatureLength || pTarget!=(target+1) || target[0]!=0xfeff)) {
1.609 + *error=U_INTERNAL_PROGRAM_ERROR;
1.610 + }
1.611 + }
1.612 + }
1.613 +}
1.614 +
1.615 +
1.616 +U_CAPI int32_t U_EXPORT2
1.617 +ucbuf_size(UCHARBUF* buf){
1.618 + if(buf){
1.619 + if(buf->isBuffered){
1.620 + return (T_FileStream_size(buf->in)-buf->signatureLength)/ucnv_getMinCharSize(buf->conv);
1.621 + }else{
1.622 + return (int32_t)(buf->bufLimit - buf->buffer);
1.623 + }
1.624 + }
1.625 + return 0;
1.626 +}
1.627 +
1.628 +U_CAPI const UChar* U_EXPORT2
1.629 +ucbuf_getBuffer(UCHARBUF* buf,int32_t* len,UErrorCode* error){
1.630 + if(error==NULL || U_FAILURE(*error)){
1.631 + return NULL;
1.632 + }
1.633 + if(buf==NULL || len==NULL){
1.634 + *error = U_ILLEGAL_ARGUMENT_ERROR;
1.635 + return NULL;
1.636 + }
1.637 + *len = (int32_t)(buf->bufLimit - buf->buffer);
1.638 + return buf->buffer;
1.639 +}
1.640 +
1.641 +U_CAPI const char* U_EXPORT2
1.642 +ucbuf_resolveFileName(const char* inputDir, const char* fileName, char* target, int32_t* len, UErrorCode* status){
1.643 + int32_t requiredLen = 0;
1.644 + int32_t dirlen = 0;
1.645 + int32_t filelen = 0;
1.646 + if(status==NULL || U_FAILURE(*status)){
1.647 + return NULL;
1.648 + }
1.649 +
1.650 + if(inputDir == NULL || fileName == NULL || len==NULL || (target==NULL && *len>0)){
1.651 + *status = U_ILLEGAL_ARGUMENT_ERROR;
1.652 + return NULL;
1.653 + }
1.654 +
1.655 +
1.656 + dirlen = (int32_t)uprv_strlen(inputDir);
1.657 + filelen = (int32_t)uprv_strlen(fileName);
1.658 + if(inputDir[dirlen-1] != U_FILE_SEP_CHAR) {
1.659 + requiredLen = dirlen + filelen + 2;
1.660 + if((*len < requiredLen) || target==NULL){
1.661 + *len = requiredLen;
1.662 + *status = U_BUFFER_OVERFLOW_ERROR;
1.663 + return NULL;
1.664 + }
1.665 +
1.666 + target[0] = '\0';
1.667 + /*
1.668 + * append the input dir to openFileName if the first char in
1.669 + * filename is not file seperation char and the last char input directory is not '.'.
1.670 + * This is to support :
1.671 + * genrb -s. /home/icu/data
1.672 + * genrb -s. icu/data
1.673 + * The user cannot mix notations like
1.674 + * genrb -s. /icu/data --- the absolute path specified. -s redundant
1.675 + * user should use
1.676 + * genrb -s. icu/data --- start from CWD and look in icu/data dir
1.677 + */
1.678 + if( (fileName[0] != U_FILE_SEP_CHAR) && (inputDir[dirlen-1] !='.')){
1.679 + uprv_strcpy(target, inputDir);
1.680 + target[dirlen] = U_FILE_SEP_CHAR;
1.681 + }
1.682 + target[dirlen + 1] = '\0';
1.683 + } else {
1.684 + requiredLen = dirlen + filelen + 1;
1.685 + if((*len < requiredLen) || target==NULL){
1.686 + *len = requiredLen;
1.687 + *status = U_BUFFER_OVERFLOW_ERROR;
1.688 + return NULL;
1.689 + }
1.690 +
1.691 + uprv_strcpy(target, inputDir);
1.692 + }
1.693 +
1.694 + uprv_strcat(target, fileName);
1.695 + return target;
1.696 +}
1.697 +/*
1.698 + * Unicode TR 13 says any of the below chars is
1.699 + * a new line char in a readline function in addition
1.700 + * to CR+LF combination which needs to be
1.701 + * handled seperately
1.702 + */
1.703 +static UBool ucbuf_isCharNewLine(UChar c){
1.704 + switch(c){
1.705 + case 0x000A: /* LF */
1.706 + case 0x000D: /* CR */
1.707 + case 0x000C: /* FF */
1.708 + case 0x0085: /* NEL */
1.709 + case 0x2028: /* LS */
1.710 + case 0x2029: /* PS */
1.711 + return TRUE;
1.712 + default:
1.713 + return FALSE;
1.714 + }
1.715 +}
1.716 +
1.717 +U_CAPI const UChar* U_EXPORT2
1.718 +ucbuf_readline(UCHARBUF* buf,int32_t* len,UErrorCode* err){
1.719 + UChar* temp = buf->currentPos;
1.720 + UChar* savePos =NULL;
1.721 + UChar c=0x0000;
1.722 + if(buf->isBuffered){
1.723 + /* The input is buffered we have to do more
1.724 + * for returning a pointer U_TRUNCATED_CHAR_FOUND
1.725 + */
1.726 + for(;;){
1.727 + c = *temp++;
1.728 + if(buf->remaining==0){
1.729 + return NULL; /* end of file is reached return NULL */
1.730 + }
1.731 + if(temp>=buf->bufLimit && buf->currentPos == buf->buffer){
1.732 + *err= U_TRUNCATED_CHAR_FOUND;
1.733 + return NULL;
1.734 + }else{
1.735 + ucbuf_fillucbuf(buf,err);
1.736 + if(U_FAILURE(*err)){
1.737 + return NULL;
1.738 + }
1.739 + }
1.740 + /*
1.741 + * Accoding to TR 13 readLine functions must interpret
1.742 + * CR, CR+LF, LF, NEL, PS, LS or FF as line seperators
1.743 + */
1.744 + /* Windows CR LF */
1.745 + if(c ==0x0d && temp+1<=buf->bufLimit && *(temp+1) == 0x0a ){
1.746 + *len = (int32_t)(temp++ - buf->currentPos);
1.747 + savePos = buf->currentPos;
1.748 + buf->currentPos = temp;
1.749 + return savePos;
1.750 + }
1.751 + /* else */
1.752 +
1.753 + if (temp>=buf->bufLimit|| ucbuf_isCharNewLine(c)){ /* Unipad inserts 2028 line separators! */
1.754 + *len = (int32_t)(temp - buf->currentPos);
1.755 + savePos = buf->currentPos;
1.756 + buf->currentPos = temp;
1.757 + return savePos;
1.758 + }
1.759 + }
1.760 + }else{
1.761 + /* we know that all input is read into the internal
1.762 + * buffer so we can safely return pointers
1.763 + */
1.764 + for(;;){
1.765 + c = *temp++;
1.766 +
1.767 + if(buf->currentPos==buf->bufLimit){
1.768 + return NULL; /* end of file is reached return NULL */
1.769 + }
1.770 + /* Windows CR LF */
1.771 + if(c ==0x0d && temp+1<=buf->bufLimit && *(temp+1) == 0x0a ){
1.772 + *len = (int32_t)(temp++ - buf->currentPos);
1.773 + savePos = buf->currentPos;
1.774 + buf->currentPos = temp;
1.775 + return savePos;
1.776 + }
1.777 + /* else */
1.778 + if (temp>=buf->bufLimit|| ucbuf_isCharNewLine(c)) { /* Unipad inserts 2028 line separators! */
1.779 + *len = (int32_t)(temp - buf->currentPos);
1.780 + savePos = buf->currentPos;
1.781 + buf->currentPos = temp;
1.782 + return savePos;
1.783 + }
1.784 + }
1.785 + }
1.786 + /* not reached */
1.787 + /* A compiler warning will appear if all paths don't contain a return statement. */
1.788 +/* return NULL;*/
1.789 +}
1.790 +#endif
