1 /*
2 *******************************************************************************
3 *
4 *   Copyright (C) 1998-2014, International Business Machines
5 *   Corporation and others.  All Rights Reserved.
6 *
7 *******************************************************************************
8 *
9 * File ucbuf.c
10 *
11 * Modification History:
12 *
13 *   Date        Name        Description
14 *   05/10/01    Ram         Creation.
15 *******************************************************************************
16 */
17 
18 #include "unicode/utypes.h"
19 #include "unicode/putil.h"
20 #include "unicode/uchar.h"
21 #include "unicode/ucnv.h"
22 #include "unicode/ucnv_err.h"
23 #include "unicode/ustring.h"
24 #include "unicode/utf16.h"
25 #include "filestrm.h"
26 #include "cstring.h"
27 #include "cmemory.h"
28 #include "ustrfmt.h"
29 #include "ucbuf.h"
30 #include <stdio.h>
31 
32 #if !UCONFIG_NO_CONVERSION
33 
34 
35 #define MAX_IN_BUF 1000
36 #define MAX_U_BUF 1500
37 #define CONTEXT_LEN 20
38 
39 struct UCHARBUF {
40     UChar* buffer;
41     UChar* currentPos;
42     UChar* bufLimit;
43     int32_t bufCapacity;
44     int32_t remaining;
45     int32_t signatureLength;
46     FileStream* in;
47     UConverter* conv;
48     UBool showWarning; /* makes this API not produce any errors */
49     UBool isBuffered;
50 };
51 
52 U_CAPI UBool U_EXPORT2
ucbuf_autodetect_fs(FileStream * in,const char ** cp,UConverter ** conv,int32_t * signatureLength,UErrorCode * error)53 ucbuf_autodetect_fs(FileStream* in, const char** cp, UConverter** conv, int32_t* signatureLength, UErrorCode* error){
54     char start[8];
55     int32_t numRead;
56 
57     UChar target[1]={ 0 };
58     UChar* pTarget;
59     const char* pStart;
60 
61     /* read a few bytes */
62     numRead=T_FileStream_read(in, start, sizeof(start));
63 
64     *cp = ucnv_detectUnicodeSignature(start, numRead, signatureLength, error);
65 
66     /* unread the bytes beyond what was consumed for U+FEFF */
67     T_FileStream_rewind(in);
68     if (*signatureLength > 0) {
69         T_FileStream_read(in, start, *signatureLength);
70     }
71 
72     if(*cp==NULL){
73         *conv =NULL;
74         return FALSE;
75     }
76 
77     /* open the converter for the detected Unicode charset */
78     *conv = ucnv_open(*cp,error);
79 
80     /* convert and ignore initial U+FEFF, and the buffer overflow */
81     pTarget = target;
82     pStart = start;
83     ucnv_toUnicode(*conv, &pTarget, target+1, &pStart, start+*signatureLength, NULL, FALSE, error);
84     *signatureLength = (int32_t)(pStart - start);
85     if(*error==U_BUFFER_OVERFLOW_ERROR) {
86         *error=U_ZERO_ERROR;
87     }
88 
89     /* verify that we successfully read exactly U+FEFF */
90     if(U_SUCCESS(*error) && (pTarget!=(target+1) || target[0]!=0xfeff)) {
91         *error=U_INTERNAL_PROGRAM_ERROR;
92     }
93 
94 
95     return TRUE;
96 }
ucbuf_isCPKnown(const char * cp)97 static UBool ucbuf_isCPKnown(const char* cp){
98     if(ucnv_compareNames("UTF-8",cp)==0){
99         return TRUE;
100     }
101     if(ucnv_compareNames("UTF-16BE",cp)==0){
102         return TRUE;
103     }
104     if(ucnv_compareNames("UTF-16LE",cp)==0){
105         return TRUE;
106     }
107     if(ucnv_compareNames("UTF-16",cp)==0){
108         return TRUE;
109     }
110     if(ucnv_compareNames("UTF-32",cp)==0){
111         return TRUE;
112     }
113     if(ucnv_compareNames("UTF-32BE",cp)==0){
114         return TRUE;
115     }
116     if(ucnv_compareNames("UTF-32LE",cp)==0){
117         return TRUE;
118     }
119     if(ucnv_compareNames("SCSU",cp)==0){
120         return TRUE;
121     }
122     if(ucnv_compareNames("BOCU-1",cp)==0){
123         return TRUE;
124     }
125     if(ucnv_compareNames("UTF-7",cp)==0){
126         return TRUE;
127     }
128     return FALSE;
129 }
130 
131 U_CAPI FileStream * U_EXPORT2
ucbuf_autodetect(const char * fileName,const char ** cp,UConverter ** conv,int32_t * signatureLength,UErrorCode * error)132 ucbuf_autodetect(const char* fileName, const char** cp,UConverter** conv, int32_t* signatureLength,UErrorCode* error){
133     FileStream* in=NULL;
134     if(error==NULL || U_FAILURE(*error)){
135         return NULL;
136     }
137     if(conv==NULL || cp==NULL || fileName==NULL){
138         *error = U_ILLEGAL_ARGUMENT_ERROR;
139         return NULL;
140     }
141     /* open the file */
142     in= T_FileStream_open(fileName,"rb");
143 
144     if(in == NULL){
145         *error=U_FILE_ACCESS_ERROR;
146         return NULL;
147     }
148 
149     if(ucbuf_autodetect_fs(in,cp,conv,signatureLength,error)) {
150         return in;
151     } else {
152         ucnv_close(*conv);
153         *conv=NULL;
154         T_FileStream_close(in);
155         return NULL;
156     }
157 }
158 
159 /* fill the uchar buffer */
160 static UCHARBUF*
ucbuf_fillucbuf(UCHARBUF * buf,UErrorCode * error)161 ucbuf_fillucbuf( UCHARBUF* buf,UErrorCode* error){
162     UChar* pTarget=NULL;
163     UChar* target=NULL;
164     const char* source=NULL;
165     char  carr[MAX_IN_BUF] = {'\0'};
166     char* cbuf =  carr;
167     int32_t inputRead=0;
168     int32_t outputWritten=0;
169     int32_t offset=0;
170     const char* sourceLimit =NULL;
171     int32_t cbufSize=0;
172     pTarget = buf->buffer;
173     /* check if we arrived here without exhausting the buffer*/
174     if(buf->currentPos<buf->bufLimit){
175         offset = (int32_t)(buf->bufLimit-buf->currentPos);
176         memmove(buf->buffer,buf->currentPos,offset* sizeof(UChar));
177     }
178 
179 #if UCBUF_DEBUG
180     memset(pTarget+offset,0xff,sizeof(UChar)*(MAX_IN_BUF-offset));
181 #endif
182     if(buf->isBuffered){
183         cbufSize = MAX_IN_BUF;
184         /* read the file */
185         inputRead=T_FileStream_read(buf->in,cbuf,cbufSize-offset);
186         buf->remaining-=inputRead;
187 
188     }else{
189         cbufSize = T_FileStream_size(buf->in);
190         cbuf = (char*)uprv_malloc(cbufSize);
191         if (cbuf == NULL) {
192         	*error = U_MEMORY_ALLOCATION_ERROR;
193         	return NULL;
194         }
195         inputRead= T_FileStream_read(buf->in,cbuf,cbufSize);
196         buf->remaining-=inputRead;
197     }
198 
199     /* just to be sure...*/
200     if ( 0 == inputRead )
201        buf->remaining = 0;
202 
203     target=pTarget;
204     /* convert the bytes */
205     if(buf->conv){
206         /* set the callback to stop */
207         UConverterToUCallback toUOldAction ;
208         void* toUOldContext;
209         void* toUNewContext=NULL;
210         ucnv_setToUCallBack(buf->conv,
211            UCNV_TO_U_CALLBACK_STOP,
212            toUNewContext,
213            &toUOldAction,
214            (const void**)&toUOldContext,
215            error);
216         /* since state is saved in the converter we add offset to source*/
217         target = pTarget+offset;
218         source = cbuf;
219         sourceLimit = source + inputRead;
220         ucnv_toUnicode(buf->conv,&target,target+(buf->bufCapacity-offset),
221                         &source,sourceLimit,NULL,
222                         (UBool)(buf->remaining==0),error);
223 
224         if(U_FAILURE(*error)){
225             char context[CONTEXT_LEN+1];
226             char preContext[CONTEXT_LEN+1];
227             char postContext[CONTEXT_LEN+1];
228             int8_t len = CONTEXT_LEN;
229             int32_t start=0;
230             int32_t stop =0;
231             int32_t pos =0;
232             /* use erro1 to preserve the error code */
233             UErrorCode error1 =U_ZERO_ERROR;
234 
235             if( buf->showWarning==TRUE){
236                 fprintf(stderr,"\n###WARNING: Encountered abnormal bytes while"
237                                " converting input stream to target encoding: %s\n",
238                                u_errorName(*error));
239             }
240 
241 
242             /* now get the context chars */
243             ucnv_getInvalidChars(buf->conv,context,&len,&error1);
244             context[len]= 0 ; /* null terminate the buffer */
245 
246             pos = (int32_t)(source - cbuf - len);
247 
248             /* for pre-context */
249             start = (pos <=CONTEXT_LEN)? 0 : (pos - (CONTEXT_LEN-1));
250             stop  = pos-len;
251 
252             memcpy(preContext,cbuf+start,stop-start);
253             /* null terminate the buffer */
254             preContext[stop-start] = 0;
255 
256             /* for post-context */
257             start = pos+len;
258             stop  = (int32_t)(((pos+CONTEXT_LEN)<= (sourceLimit-cbuf) )? (pos+(CONTEXT_LEN-1)) : (sourceLimit-cbuf));
259 
260             memcpy(postContext,source,stop-start);
261             /* null terminate the buffer */
262             postContext[stop-start] = 0;
263 
264             if(buf->showWarning ==TRUE){
265                 /* print out the context */
266                 fprintf(stderr,"\tPre-context: %s\n",preContext);
267                 fprintf(stderr,"\tContext: %s\n",context);
268                 fprintf(stderr,"\tPost-context: %s\n", postContext);
269             }
270 
271             /* reset the converter */
272             ucnv_reset(buf->conv);
273 
274             /* set the call back to substitute
275              * and restart conversion
276              */
277             ucnv_setToUCallBack(buf->conv,
278                UCNV_TO_U_CALLBACK_SUBSTITUTE,
279                toUNewContext,
280                &toUOldAction,
281                (const void**)&toUOldContext,
282                &error1);
283 
284             /* reset source and target start positions */
285             target = pTarget+offset;
286             source = cbuf;
287 
288             /* re convert */
289             ucnv_toUnicode(buf->conv,&target,target+(buf->bufCapacity-offset),
290                             &source,sourceLimit,NULL,
291                             (UBool)(buf->remaining==0),&error1);
292 
293         }
294         outputWritten = (int32_t)(target - pTarget);
295 
296 #if UCBUF_DEBUG
297         {
298             int i;
299             target = pTarget;
300             for(i=0;i<numRead;i++){
301               /*  printf("%c", (char)(*target++));*/
302             }
303         }
304 #endif
305 
306     }else{
307         u_charsToUChars(cbuf,target+offset,inputRead);
308         outputWritten=((buf->remaining>cbufSize)? cbufSize:inputRead+offset);
309     }
310     buf->currentPos = pTarget;
311     buf->bufLimit=pTarget+outputWritten;
312     *buf->bufLimit=0; /*NUL terminate*/
313     if(cbuf!=carr){
314         uprv_free(cbuf);
315     }
316     return buf;
317 }
318 
319 
320 
321 /* get a UChar from the stream*/
322 U_CAPI int32_t U_EXPORT2
ucbuf_getc(UCHARBUF * buf,UErrorCode * error)323 ucbuf_getc(UCHARBUF* buf,UErrorCode* error){
324     if(error==NULL || U_FAILURE(*error)){
325         return FALSE;
326     }
327     if(buf->currentPos>=buf->bufLimit){
328         if(buf->remaining==0){
329             return U_EOF;
330         }
331         buf=ucbuf_fillucbuf(buf,error);
332         if(U_FAILURE(*error)){
333             return U_EOF;
334         }
335     }
336 
337     return *(buf->currentPos++);
338 }
339 
340 /* get a UChar32 from the stream*/
341 U_CAPI int32_t U_EXPORT2
ucbuf_getc32(UCHARBUF * buf,UErrorCode * error)342 ucbuf_getc32(UCHARBUF* buf,UErrorCode* error){
343     int32_t retVal = (int32_t)U_EOF;
344     if(error==NULL || U_FAILURE(*error)){
345         return FALSE;
346     }
347     if(buf->currentPos+1>=buf->bufLimit){
348         if(buf->remaining==0){
349             return U_EOF;
350         }
351         buf=ucbuf_fillucbuf(buf,error);
352         if(U_FAILURE(*error)){
353             return U_EOF;
354         }
355     }
356     if(U16_IS_LEAD(*(buf->currentPos))){
357         retVal=U16_GET_SUPPLEMENTARY(buf->currentPos[0],buf->currentPos[1]);
358         buf->currentPos+=2;
359     }else{
360         retVal = *(buf->currentPos++);
361     }
362     return retVal;
363 }
364 
365 /* u_unescapeAt() callback to return a UChar*/
366 static UChar U_CALLCONV
_charAt(int32_t offset,void * context)367 _charAt(int32_t offset, void *context) {
368     return ((UCHARBUF*) context)->currentPos[offset];
369 }
370 
371 /* getc and escape it */
372 U_CAPI int32_t U_EXPORT2
ucbuf_getcx32(UCHARBUF * buf,UErrorCode * error)373 ucbuf_getcx32(UCHARBUF* buf,UErrorCode* error) {
374     int32_t length;
375     int32_t offset;
376     UChar32 c32,c1,c2;
377     if(error==NULL || U_FAILURE(*error)){
378         return FALSE;
379     }
380     /* Fill the buffer if it is empty */
381     if (buf->currentPos >=buf->bufLimit-2) {
382         ucbuf_fillucbuf(buf,error);
383     }
384 
385     /* Get the next character in the buffer */
386     if (buf->currentPos < buf->bufLimit) {
387         c1 = *(buf->currentPos)++;
388     } else {
389         c1 = U_EOF;
390     }
391 
392     c2 = *(buf->currentPos);
393 
394     /* If it isn't a backslash, return it */
395     if (c1 != 0x005C) {
396         return c1;
397     }
398 
399     /* Determine the amount of data in the buffer */
400     length = (int32_t)(buf->bufLimit - buf->currentPos);
401 
402     /* The longest escape sequence is \Uhhhhhhhh; make sure
403        we have at least that many characters */
404     if (length < 10) {
405 
406         /* fill the buffer */
407         ucbuf_fillucbuf(buf,error);
408         length = (int32_t)(buf->bufLimit - buf->buffer);
409     }
410 
411     /* Process the escape */
412     offset = 0;
413     c32 = u_unescapeAt(_charAt, &offset, length, (void*)buf);
414 
415     /* check if u_unescapeAt unescaped and converted
416      * to c32 or not
417      */
418     if(c32==0xFFFFFFFF){
419         if(buf->showWarning) {
420             char context[CONTEXT_LEN+1];
421             int32_t len = CONTEXT_LEN;
422             if(length < len) {
423                 len = length;
424             }
425             context[len]= 0 ; /* null terminate the buffer */
426             u_UCharsToChars( buf->currentPos, context, len);
427             fprintf(stderr,"Bad escape: [%c%s]...\n", (int)c1, context);
428         }
429         *error= U_ILLEGAL_ESCAPE_SEQUENCE;
430         return c1;
431     }else if(c32!=c2 || (c32==0x0075 && c2==0x0075 && c1==0x005C) /* for \u0075 c2=0x0075 and c32==0x0075*/){
432         /* Update the current buffer position */
433         buf->currentPos += offset;
434     }else{
435         /* unescaping failed so we just return
436          * c1 and not consume the buffer
437          * this is useful for rules with escapes
438          * in resouce bundles
439          * eg: \' \\ \"
440          */
441         return c1;
442     }
443 
444     return c32;
445 }
446 
447 U_CAPI UCHARBUF* U_EXPORT2
ucbuf_open(const char * fileName,const char ** cp,UBool showWarning,UBool buffered,UErrorCode * error)448 ucbuf_open(const char* fileName,const char** cp,UBool showWarning, UBool buffered, UErrorCode* error){
449 
450     FileStream* in = NULL;
451     int32_t fileSize=0;
452     const char* knownCp;
453     if(error==NULL || U_FAILURE(*error)){
454         return NULL;
455     }
456     if(cp==NULL || fileName==NULL){
457         *error = U_ILLEGAL_ARGUMENT_ERROR;
458         return FALSE;
459     }
460     if (!uprv_strcmp(fileName, "-")) {
461         in = T_FileStream_stdin();
462     }else{
463         in = T_FileStream_open(fileName, "rb");
464     }
465 
466     if(in!=NULL){
467         UCHARBUF* buf =(UCHARBUF*) uprv_malloc(sizeof(UCHARBUF));
468         fileSize = T_FileStream_size(in);
469         if(buf == NULL){
470             *error = U_MEMORY_ALLOCATION_ERROR;
471             T_FileStream_close(in);
472             return NULL;
473         }
474         buf->in=in;
475         buf->conv=NULL;
476         buf->showWarning = showWarning;
477         buf->isBuffered = buffered;
478         buf->signatureLength=0;
479         if(*cp==NULL || **cp=='\0'){
480             /* don't have code page name... try to autodetect */
481             ucbuf_autodetect_fs(in,cp,&buf->conv,&buf->signatureLength,error);
482         }else if(ucbuf_isCPKnown(*cp)){
483             /* discard BOM */
484             ucbuf_autodetect_fs(in,&knownCp,&buf->conv,&buf->signatureLength,error);
485         }
486         if(U_SUCCESS(*error) && buf->conv==NULL) {
487             buf->conv=ucnv_open(*cp,error);
488         }
489         if(U_FAILURE(*error)){
490             ucnv_close(buf->conv);
491             uprv_free(buf);
492             T_FileStream_close(in);
493             return NULL;
494         }
495 
496         if((buf->conv==NULL) && (buf->showWarning==TRUE)){
497             fprintf(stderr,"###WARNING: No converter defined. Using codepage of system.\n");
498         }
499         buf->remaining=fileSize-buf->signatureLength;
500         if(buf->isBuffered){
501             buf->bufCapacity=MAX_U_BUF;
502         }else{
503             buf->bufCapacity=buf->remaining+buf->signatureLength+1/*for terminating nul*/;
504         }
505         buf->buffer=(UChar*) uprv_malloc(U_SIZEOF_UCHAR * buf->bufCapacity );
506         if (buf->buffer == NULL) {
507             *error = U_MEMORY_ALLOCATION_ERROR;
508             ucbuf_close(buf);
509             return NULL;
510         }
511         buf->currentPos=buf->buffer;
512         buf->bufLimit=buf->buffer;
513         if(U_FAILURE(*error)){
514             fprintf(stderr, "Could not open codepage [%s]: %s\n", *cp, u_errorName(*error));
515             ucbuf_close(buf);
516             return NULL;
517         }
518         ucbuf_fillucbuf(buf,error);
519         if(U_FAILURE(*error)){
520             ucbuf_close(buf);
521             return NULL;
522         }
523         return buf;
524     }
525     *error =U_FILE_ACCESS_ERROR;
526     return NULL;
527 }
528 
529 
530 
531 /* TODO: this method will fail if at the
532  * begining of buffer and the uchar to unget
533  * is from the previous buffer. Need to implement
534  * system to take care of that situation.
535  */
536 U_CAPI void U_EXPORT2
ucbuf_ungetc(int32_t c,UCHARBUF * buf)537 ucbuf_ungetc(int32_t c,UCHARBUF* buf){
538     /* decrement currentPos pointer
539      * if not at the begining of buffer
540      */
541     if(buf->currentPos!=buf->buffer){
542         if(*(buf->currentPos-1)==c){
543             buf->currentPos--;
544         } else {
545             /* ungetc failed - did not match. */
546         }
547     } else {
548        /* ungetc failed - beginning of buffer. */
549     }
550 }
551 
552 /* frees the resources of UChar* buffer */
553 static void
ucbuf_closebuf(UCHARBUF * buf)554 ucbuf_closebuf(UCHARBUF* buf){
555     uprv_free(buf->buffer);
556     buf->buffer = NULL;
557 }
558 
559 /* close the buf and release resources*/
560 U_CAPI void U_EXPORT2
ucbuf_close(UCHARBUF * buf)561 ucbuf_close(UCHARBUF* buf){
562     if(buf!=NULL){
563         if(buf->conv){
564             ucnv_close(buf->conv);
565         }
566         T_FileStream_close(buf->in);
567         ucbuf_closebuf(buf);
568         uprv_free(buf);
569     }
570 }
571 
572 /* rewind the buf and file stream */
573 U_CAPI void U_EXPORT2
ucbuf_rewind(UCHARBUF * buf,UErrorCode * error)574 ucbuf_rewind(UCHARBUF* buf,UErrorCode* error){
575     if(error==NULL || U_FAILURE(*error)){
576         return;
577     }
578     if(buf){
579         buf->currentPos=buf->buffer;
580         buf->bufLimit=buf->buffer;
581         T_FileStream_rewind(buf->in);
582         buf->remaining=T_FileStream_size(buf->in)-buf->signatureLength;
583 
584         ucnv_resetToUnicode(buf->conv);
585         if(buf->signatureLength>0) {
586             UChar target[1]={ 0 };
587             UChar* pTarget;
588             char start[8];
589             const char* pStart;
590             int32_t numRead;
591 
592             /* read the signature bytes */
593             numRead=T_FileStream_read(buf->in, start, buf->signatureLength);
594 
595             /* convert and ignore initial U+FEFF, and the buffer overflow */
596             pTarget = target;
597             pStart = start;
598             ucnv_toUnicode(buf->conv, &pTarget, target+1, &pStart, start+numRead, NULL, FALSE, error);
599             if(*error==U_BUFFER_OVERFLOW_ERROR) {
600                 *error=U_ZERO_ERROR;
601             }
602 
603             /* verify that we successfully read exactly U+FEFF */
604             if(U_SUCCESS(*error) && (numRead!=buf->signatureLength || pTarget!=(target+1) || target[0]!=0xfeff)) {
605                 *error=U_INTERNAL_PROGRAM_ERROR;
606             }
607         }
608     }
609 }
610 
611 
612 U_CAPI int32_t U_EXPORT2
ucbuf_size(UCHARBUF * buf)613 ucbuf_size(UCHARBUF* buf){
614     if(buf){
615         if(buf->isBuffered){
616             return (T_FileStream_size(buf->in)-buf->signatureLength)/ucnv_getMinCharSize(buf->conv);
617         }else{
618             return (int32_t)(buf->bufLimit - buf->buffer);
619         }
620     }
621     return 0;
622 }
623 
624 U_CAPI const UChar* U_EXPORT2
ucbuf_getBuffer(UCHARBUF * buf,int32_t * len,UErrorCode * error)625 ucbuf_getBuffer(UCHARBUF* buf,int32_t* len,UErrorCode* error){
626     if(error==NULL || U_FAILURE(*error)){
627         return NULL;
628     }
629     if(buf==NULL || len==NULL){
630         *error = U_ILLEGAL_ARGUMENT_ERROR;
631         return NULL;
632     }
633     *len = (int32_t)(buf->bufLimit - buf->buffer);
634     return buf->buffer;
635 }
636 
637 U_CAPI const char* U_EXPORT2
ucbuf_resolveFileName(const char * inputDir,const char * fileName,char * target,int32_t * len,UErrorCode * status)638 ucbuf_resolveFileName(const char* inputDir, const char* fileName, char* target, int32_t* len, UErrorCode* status){
639     int32_t requiredLen = 0;
640     int32_t dirlen =  0;
641     int32_t filelen = 0;
642     if(status==NULL || U_FAILURE(*status)){
643         return NULL;
644     }
645 
646     if(inputDir == NULL || fileName == NULL || len==NULL || (target==NULL && *len>0)){
647         *status = U_ILLEGAL_ARGUMENT_ERROR;
648         return NULL;
649     }
650 
651 
652     dirlen  = (int32_t)uprv_strlen(inputDir);
653     filelen = (int32_t)uprv_strlen(fileName);
654     if(inputDir[dirlen-1] != U_FILE_SEP_CHAR) {
655         requiredLen = dirlen + filelen + 2;
656         if((*len < requiredLen) || target==NULL){
657             *len = requiredLen;
658             *status = U_BUFFER_OVERFLOW_ERROR;
659             return NULL;
660         }
661 
662         target[0] = '\0';
663         /*
664          * append the input dir to openFileName if the first char in
665          * filename is not file seperation char and the last char input directory is  not '.'.
666          * This is to support :
667          * genrb -s. /home/icu/data
668          * genrb -s. icu/data
669          * The user cannot mix notations like
670          * genrb -s. /icu/data --- the absolute path specified. -s redundant
671          * user should use
672          * genrb -s. icu/data  --- start from CWD and look in icu/data dir
673          */
674         if( (fileName[0] != U_FILE_SEP_CHAR) && (inputDir[dirlen-1] !='.')){
675             uprv_strcpy(target, inputDir);
676             target[dirlen]     = U_FILE_SEP_CHAR;
677         }
678         target[dirlen + 1] = '\0';
679     } else {
680         requiredLen = dirlen + filelen + 1;
681         if((*len < requiredLen) || target==NULL){
682             *len = requiredLen;
683             *status = U_BUFFER_OVERFLOW_ERROR;
684             return NULL;
685         }
686 
687         uprv_strcpy(target, inputDir);
688     }
689 
690     uprv_strcat(target, fileName);
691     return target;
692 }
693 /*
694  * Unicode TR 13 says any of the below chars is
695  * a new line char in a readline function in addition
696  * to CR+LF combination which needs to be
697  * handled seperately
698  */
ucbuf_isCharNewLine(UChar c)699 static UBool ucbuf_isCharNewLine(UChar c){
700     switch(c){
701     case 0x000A: /* LF  */
702     case 0x000D: /* CR  */
703     case 0x000C: /* FF  */
704     case 0x0085: /* NEL */
705     case 0x2028: /* LS  */
706     case 0x2029: /* PS  */
707         return TRUE;
708     default:
709         return FALSE;
710     }
711 }
712 
713 U_CAPI const UChar* U_EXPORT2
ucbuf_readline(UCHARBUF * buf,int32_t * len,UErrorCode * err)714 ucbuf_readline(UCHARBUF* buf,int32_t* len,UErrorCode* err){
715     UChar* temp = buf->currentPos;
716     UChar* savePos =NULL;
717     UChar c=0x0000;
718     if(buf->isBuffered){
719         /* The input is buffered we have to do more
720         * for returning a pointer U_TRUNCATED_CHAR_FOUND
721         */
722         for(;;){
723             c = *temp++;
724             if(buf->remaining==0){
725                 return NULL; /* end of file is reached return NULL */
726             }
727             if(temp>=buf->bufLimit && buf->currentPos == buf->buffer){
728                 *err= U_TRUNCATED_CHAR_FOUND;
729                 return NULL;
730             }else{
731                 ucbuf_fillucbuf(buf,err);
732                 if(U_FAILURE(*err)){
733                     return NULL;
734                 }
735             }
736             /*
737              * Accoding to TR 13 readLine functions must interpret
738              * CR, CR+LF, LF, NEL, PS, LS or FF as line seperators
739              */
740             /* Windows CR LF */
741             if(c ==0x0d && temp <= buf->bufLimit && *temp == 0x0a ){
742                 *len = (int32_t)(temp++ - buf->currentPos);
743                 savePos = buf->currentPos;
744                 buf->currentPos = temp;
745                 return savePos;
746             }
747             /* else */
748 
749             if (temp>=buf->bufLimit|| ucbuf_isCharNewLine(c)){  /* Unipad inserts 2028 line separators! */
750                 *len = (int32_t)(temp - buf->currentPos);
751                 savePos = buf->currentPos;
752                 buf->currentPos = temp;
753                 return savePos;
754             }
755         }
756     }else{
757     /* we know that all input is read into the internal
758     * buffer so we can safely return pointers
759         */
760         for(;;){
761             c = *temp++;
762 
763             if(buf->currentPos==buf->bufLimit){
764                 return NULL; /* end of file is reached return NULL */
765             }
766             /* Windows CR LF */
767             if(c ==0x0d && temp <= buf->bufLimit && *temp == 0x0a ){
768                 *len = (int32_t)(temp++ - buf->currentPos);
769                 savePos = buf->currentPos;
770                 buf->currentPos = temp;
771                 return savePos;
772             }
773             /* else */
774             if (temp>=buf->bufLimit|| ucbuf_isCharNewLine(c)) {  /* Unipad inserts 2028 line separators! */
775                 *len = (int32_t)(temp - buf->currentPos);
776                 savePos = buf->currentPos;
777                 buf->currentPos = temp;
778                 return savePos;
779             }
780         }
781     }
782     /* not reached */
783     /* A compiler warning will appear if all paths don't contain a return statement. */
784 /*    return NULL;*/
785 }
786 #endif
787