icuSources/tools/toolutil/uparse.c

   1 /*
   2 *******************************************************************************
   3 *
   4 *   Copyright (C) 2000-2011, International Business Machines
   5 *   Corporation and others.  All Rights Reserved.
   6 *
   7 *******************************************************************************
   8 *   file name:  uparse.c
   9 *   encoding:   US-ASCII
  10 *   tab size:   8 (not used)
  11 *   indentation:4
  12 *
  13 *   created on: 2000apr18
  14 *   created by: Markus W. Scherer
  15 *
  16 *   This file provides a parser for files that are delimited by one single
  17 *   character like ';' or TAB. Example: the Unicode Character Properties files
  18 *   like UnicodeData.txt are semicolon-delimited.
  19 */
  20
  21 #include "unicode/utypes.h"
  22 #include "unicode/uchar.h"
  23 #include "unicode/ustring.h"
  24 #include "unicode/utf16.h"
  25 #include "cstring.h"
  26 #include "filestrm.h"
  27 #include "uparse.h"
  28 #include "ustr_imp.h"
  29
  30 #include <stdio.h>
  31
  32 U_CAPI const char * U_EXPORT2
  33 u_skipWhitespace(const char *s) {
  34     while(U_IS_INV_WHITESPACE(*s)) {
  35         ++s;
  36     }
  37     return s;
  38 }
  39
  40 U_CAPI char * U_EXPORT2
  41 u_rtrim(char *s) {
  42     char *end=uprv_strchr(s, 0);
  43     while(s<end && U_IS_INV_WHITESPACE(*(end-1))) {
  44         *--end = 0;
  45     }
  46     return end;
  47 }
  48
  49 /*
  50  * If the string starts with # @missing: then return the pointer to the
  51  * following non-whitespace character.
  52  * Otherwise return the original pointer.
  53  * Unicode 5.0 adds such lines in some data files to document
  54  * default property values.
  55  * Poor man's regex for variable amounts of white space.
  56  */
  57 static const char *
  58 getMissingLimit(const char *s) {
  59     const char *s0=s;
  60     if(
  61         *(s=u_skipWhitespace(s))=='#' &&
  62         *(s=u_skipWhitespace(s+1))=='@' &&
  63         0==strncmp((s=u_skipWhitespace(s+1)), "missing", 7) &&
  64         *(s=u_skipWhitespace(s+7))==':'
  65     ) {
  66         return u_skipWhitespace(s+1);
  67     } else {
  68         return s0;
  69     }
  70 }
  71
  72 U_CAPI void U_EXPORT2
  73 u_parseDelimitedFile(const char *filename, char delimiter,
  74                      char *fields[][2], int32_t fieldCount,
  75                      UParseLineFn *lineFn, void *context,
  76                      UErrorCode *pErrorCode) {
  77     FileStream *file;
  78     char line[300];
  79     char *start, *limit;
  80     int32_t i, length;
  81
  82     if(U_FAILURE(*pErrorCode)) {
  83         return;
  84     }
  85
  86     if(fields==NULL || lineFn==NULL || fieldCount<=0) {
  87         *pErrorCode=U_ILLEGAL_ARGUMENT_ERROR;
  88         return;
  89     }
  90
  91     if(filename==NULL || *filename==0 || (*filename=='-' && filename[1]==0)) {
  92         filename=NULL;
  93         file=T_FileStream_stdin();
  94     } else {
  95         file=T_FileStream_open(filename, "r");
  96     }
  97     if(file==NULL) {
  98         *pErrorCode=U_FILE_ACCESS_ERROR;
  99         return;
 100     }
 101
 102     while(T_FileStream_readLine(file, line, sizeof(line))!=NULL) {
 103         /* remove trailing newline characters */
 104         length=(int32_t)(u_rtrim(line)-line);
 105
 106         /*
 107          * detect a line with # @missing:
 108          * start parsing after that, or else from the beginning of the line
 109          * set the default warning for @missing lines
 110          */
 111         start=(char *)getMissingLimit(line);
 112         if(start==line) {
 113             *pErrorCode=U_ZERO_ERROR;
 114         } else {
 115             *pErrorCode=U_USING_DEFAULT_WARNING;
 116         }
 117
 118         /* skip this line if it is empty or a comment */
 119         if(*start==0 || *start=='#') {
 120             continue;
 121         }
 122
 123         /* remove in-line comments */
 124         limit=uprv_strchr(start, '#');
 125         if(limit!=NULL) {
 126             /* get white space before the pound sign */
 127             while(limit>start && U_IS_INV_WHITESPACE(*(limit-1))) {
 128                 --limit;
 129             }
 130
 131             /* truncate the line */
 132             *limit=0;
 133         }
 134
 135         /* skip lines with only whitespace */
 136         if(u_skipWhitespace(start)[0]==0) {
 137             continue;
 138         }
 139
 140         /* for each field, call the corresponding field function */
 141         for(i=0; i<fieldCount; ++i) {
 142             /* set the limit pointer of this field */
 143             limit=start;
 144             while(*limit!=delimiter && *limit!=0) {
 145                 ++limit;
 146             }
 147
 148             /* set the field start and limit in the fields array */
 149             fields[i][0]=start;
 150             fields[i][1]=limit;
 151
 152             /* set start to the beginning of the next field, if any */
 153             start=limit;
 154             if(*start!=0) {
 155                 ++start;
 156             } else if(i+1<fieldCount) {
 157                 *pErrorCode=U_PARSE_ERROR;
 158                 limit=line+length;
 159                 i=fieldCount;
 160                 break;
 161             }
 162         }
 163
 164         /* error in a field function? */
 165         if(U_FAILURE(*pErrorCode)) {
 166             break;
 167         }
 168
 169         /* call the field function */
 170         lineFn(context, fields, fieldCount, pErrorCode);
 171         if(U_FAILURE(*pErrorCode)) {
 172             break;
 173         }
 174     }
 175
 176     if(filename!=NULL) {
 177         T_FileStream_close(file);
 178     }
 179 }
 180
 181 /*
 182  * parse a list of code points
 183  * store them as a UTF-32 string in dest[destCapacity]
 184  * return the number of code points
 185  */
 186 U_CAPI int32_t U_EXPORT2
 187 u_parseCodePoints(const char *s,
 188                   uint32_t *dest, int32_t destCapacity,
 189                   UErrorCode *pErrorCode) {
 190     char *end;
 191     uint32_t value;
 192     int32_t count;
 193
 194     if(U_FAILURE(*pErrorCode)) {
 195         return 0;
 196     }
 197     if(s==NULL || destCapacity<0 || (destCapacity>0 && dest==NULL)) {
 198         *pErrorCode=U_ILLEGAL_ARGUMENT_ERROR;
 199         return 0;
 200     }
 201
 202     count=0;
 203     for(;;) {
 204         s=u_skipWhitespace(s);
 205         if(*s==';' || *s==0) {
 206             return count;
 207         }
 208
 209         /* read one code point */
 210         value=(uint32_t)uprv_strtoul(s, &end, 16);
 211         if(end<=s || (!U_IS_INV_WHITESPACE(*end) && *end!=';' && *end!=0) || value>=0x110000) {
 212             *pErrorCode=U_PARSE_ERROR;
 213             return 0;
 214         }
 215
 216         /* append it to the destination array */
 217         if(count<destCapacity) {
 218             dest[count++]=value;
 219         } else {
 220             *pErrorCode=U_BUFFER_OVERFLOW_ERROR;
 221         }
 222
 223         /* go to the following characters */
 224         s=end;
 225     }
 226 }
 227
 228 /*
 229  * parse a list of code points
 230  * store them as a string in dest[destCapacity]
 231  * set the first code point in *pFirst
 232  * @return The length of the string in numbers of UChars.
 233  */
 234 U_CAPI int32_t U_EXPORT2
 235 u_parseString(const char *s,
 236               UChar *dest, int32_t destCapacity,
 237               uint32_t *pFirst,
 238               UErrorCode *pErrorCode) {
 239     char *end;
 240     uint32_t value;
 241     int32_t destLength;
 242
 243     if(U_FAILURE(*pErrorCode)) {
 244         return 0;
 245     }
 246     if(s==NULL || destCapacity<0 || (destCapacity>0 && dest==NULL)) {
 247         *pErrorCode=U_ILLEGAL_ARGUMENT_ERROR;
 248     }
 249
 250     if(pFirst!=NULL) {
 251         *pFirst=0xffffffff;
 252     }
 253
 254     destLength=0;
 255     for(;;) {
 256         s=u_skipWhitespace(s);
 257         if(*s==';' || *s==0) {
 258             if(destLength<destCapacity) {
 259                 dest[destLength]=0;
 260             } else if(destLength==destCapacity) {
 261                 *pErrorCode=U_STRING_NOT_TERMINATED_WARNING;
 262             } else {
 263                 *pErrorCode=U_BUFFER_OVERFLOW_ERROR;
 264             }
 265             return destLength;
 266         }
 267
 268         /* read one code point */
 269         value=(uint32_t)uprv_strtoul(s, &end, 16);
 270         if(end<=s || (!U_IS_INV_WHITESPACE(*end) && *end!=';' && *end!=0) || value>=0x110000) {
 271             *pErrorCode=U_PARSE_ERROR;
 272             return 0;
 273         }
 274
 275         /* store the first code point */
 276         if(pFirst!=NULL) {
 277             *pFirst=value;
 278             pFirst=NULL;
 279         }
 280
 281         /* append it to the destination array */
 282         if((destLength+U16_LENGTH(value))<=destCapacity) {
 283             U16_APPEND_UNSAFE(dest, destLength, value);
 284         } else {
 285             destLength+=U16_LENGTH(value);
 286         }
 287
 288         /* go to the following characters */
 289         s=end;
 290     }
 291 }
 292
 293 /* read a range like start or start..end */
 294 U_CAPI int32_t U_EXPORT2
 295 u_parseCodePointRangeAnyTerminator(const char *s,
 296                                    uint32_t *pStart, uint32_t *pEnd,
 297                                    const char **terminator,
 298                                    UErrorCode *pErrorCode) {
 299     char *end;
 300     uint32_t value;
 301
 302     if(U_FAILURE(*pErrorCode)) {
 303         return 0;
 304     }
 305     if(s==NULL || pStart==NULL || pEnd==NULL) {
 306         *pErrorCode=U_ILLEGAL_ARGUMENT_ERROR;
 307         return 0;
 308     }
 309
 310     /* read the start code point */
 311     s=u_skipWhitespace(s);
 312     value=(uint32_t)uprv_strtoul(s, &end, 16);
 313     if(end<=s || value>=0x110000) {
 314         *pErrorCode=U_PARSE_ERROR;
 315         return 0;
 316     }
 317     *pStart=*pEnd=value;
 318
 319     /* is there a "..end"? */
 320     s=u_skipWhitespace(end);
 321     if(*s!='.' || s[1]!='.') {
 322         *terminator=end;
 323         return 1;
 324     }
 325     s=u_skipWhitespace(s+2);
 326
 327     /* read the end code point */
 328     value=(uint32_t)uprv_strtoul(s, &end, 16);
 329     if(end<=s || value>=0x110000) {
 330         *pErrorCode=U_PARSE_ERROR;
 331         return 0;
 332     }
 333     *pEnd=value;
 334
 335     /* is this a valid range? */
 336     if(value<*pStart) {
 337         *pErrorCode=U_PARSE_ERROR;
 338         return 0;
 339     }
 340
 341     *terminator=end;
 342     return value-*pStart+1;
 343 }
 344
 345 U_CAPI int32_t U_EXPORT2
 346 u_parseCodePointRange(const char *s,
 347                       uint32_t *pStart, uint32_t *pEnd,
 348                       UErrorCode *pErrorCode) {
 349     const char *terminator;
 350     int32_t rangeLength=
 351         u_parseCodePointRangeAnyTerminator(s, pStart, pEnd, &terminator, pErrorCode);
 352     if(U_SUCCESS(*pErrorCode)) {
 353         terminator=u_skipWhitespace(terminator);
 354         if(*terminator!=';' && *terminator!=0) {
 355             *pErrorCode=U_PARSE_ERROR;
 356             return 0;
 357         }
 358     }
 359     return rangeLength;
 360 }
 361
 362 U_CAPI int32_t U_EXPORT2
 363 u_parseUTF8(const char *source, int32_t sLen, char *dest, int32_t destCapacity, UErrorCode *status) {
 364     const char *read = source;
 365     int32_t i = 0;
 366     unsigned int value = 0;
 367     if(sLen == -1) {
 368         sLen = (int32_t)strlen(source);
 369     }
 370
 371     while(read < source+sLen) {
 372         sscanf(read, "%2x", &value);
 373         if(i < destCapacity) {
 374             dest[i] = (char)value;
 375         }
 376         i++;
 377         read += 2;
 378     }
 379     return u_terminateChars(dest, destCapacity, i, status);
 380 }