1 #include <stdio.h> 2 #include <stdlib.h> 3 #include <string.h> 4 #include <ctype.h> 5 #if (defined(_WIN32) || defined(_MSDOS) || defined(__IBMC__)) 6 #include <io.h> 7 #else 8 #include <unistd.h> 9 #endif 10 #include "cpp.h" 11 12 13 static char wbuf[4 * OBS]; 14 static char *wbp = wbuf; 15 static int EBCDIC_ExternTokenDetected = 0; 16 static int EBCDIC_StartTokenDetected = 0; 17 18 unsigned char toLatin1[256] = 19 { 20 0x00, 0x01, 0x02, 0x03, 0x9c, 0x09, 0x86, 0x7f, 0x97, 0x8d, 21 0x8e, 0x0b, 0x0c, 0x0d, 0x0e, 0x0f, 0x10, 0x11, 0x12, 0x13, 22 0x9d, 0x0a, 0x08, 0x87, 0x18, 0x19, 0x92, 0x8f, 0x1c, 0x1d, 23 0x1e, 0x1f, 0x80, 0x81, 0x82, 0x83, 0x84, 0x85, 0x17, 0x1b, 24 0x88, 0x89, 0x8a, 0x8b, 0x8c, 0x05, 0x06, 0x07, 0x90, 0x91, 25 0x16, 0x93, 0x94, 0x95, 0x96, 0x04, 0x98, 0x99, 0x9a, 0x9b, 26 0x14, 0x15, 0x9e, 0x1a, 0x20, 0xa0, 0xe2, 0xe4, 0xe0, 0xe1, 27 0xe3, 0xe5, 0xe7, 0xf1, 0xa2, 0x2e, 0x3c, 0x28, 0x2b, 0x7c, 28 0x26, 0xe9, 0xea, 0xeb, 0xe8, 0xed, 0xee, 0xef, 0xec, 0xdf, 29 0x21, 0x24, 0x2a, 0x29, 0x3b, 0x5e, 0x2d, 0x2f, 0xc2, 0xc4, 30 0xc0, 0xc1, 0xc3, 0xc5, 0xc7, 0xd1, 0xa6, 0x2c, 0x25, 0x5f, 31 0x3e, 0x3f, 0xf8, 0xc9, 0xca, 0xcb, 0xc8, 0xcd, 0xce, 0xcf, 32 0xcc, 0x60, 0x3a, 0x23, 0x40, 0x27, 0x3d, 0x22, 33 0xd8, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, 0x68, 0x69, 34 0xab, 0xbb, 0xf0, 0xfd, 0xfe, 0xb1, 0xb0, 0x6a, 0x6b, 0x6c, 35 0x6d, 0x6e, 0x6f, 0x70, 0x71, 0x72, 0xaa, 0xba, 0xe6, 0xb8, 36 0xc6, 0xa4, 0xb5, 0x7e, 0x73, 0x74, 0x75, 0x76, 0x77, 0x78, 37 0x79, 0x7a, 0xa1, 0xbf, 0xd0, 0x5b, 0xde, 0xae, 0xac, 0xa3, 38 0xa5, 0xb7, 0xa9, 0xa7, 0xb6, 0xbc, 0xbd, 0xbe, 0xdd, 0xa8, 39 0xaf, 0x5d, 0xb4, 0xd7, 0x7b, 0x41, 0x42, 0x43, 0x44, 0x45, 40 0x46, 0x47, 0x48, 0x49, 0xad, 0xf4, 0xf6, 0xf2, 0xf3, 0xf5, 41 0x7d, 0x4a, 0x4b, 0x4c, 0x4d, 0x4e, 0x4f, 0x50, 0x51, 0x52, 42 0xb9, 0xfb, 0xfc, 0xf9, 0xfa, 0xff, 0x5c, 0xf7, 0x53, 0x54, 43 0x55, 0x56, 0x57, 0x58, 0x59, 0x5a, 0xb2, 0xd4, 0xd6, 0xd2, 44 0xd3, 0xd5, 0x30, 0x31, 0x32, 0x33, 0x34, 0x35, 0x36, 0x37, 45 0x38, 0x39, 0xb3, 0xdb, 0xdc, 0xd9, 0xda, 0x9f 46 }; 47 48 #define MASK "\\x%x" 49 50 int 51 memcpy_EBCDIC( char * pwbuf, uchar *p, int len ) 52 { 53 int currpos = 0; 54 int processedchars = 0; 55 56 if( len == 0 ) 57 return 0; 58 59 if( len == 1 ) 60 { 61 *pwbuf = *p; 62 return 1; 63 } 64 65 /* copy spaces until " or ' */ 66 while( (p[ processedchars ] != '\"') && (p[ processedchars ] != '\'') ) 67 pwbuf[ currpos++ ] = p[ processedchars++ ]; 68 69 /* copy first " or ' */ 70 pwbuf[ currpos++ ] = p[ processedchars++ ]; 71 72 /* convert all characters until " or ' */ 73 while( processedchars < (len - 1) ) 74 { 75 if( p[ processedchars ] == '\\' ) 76 { 77 switch( p[ ++processedchars ] ) 78 { 79 case 'n': 80 currpos += sprintf( &pwbuf[ currpos ], MASK, toLatin1['\n'] ); 81 processedchars++; 82 break; 83 84 case 't': 85 currpos += sprintf( &pwbuf[ currpos ], MASK, toLatin1['\t'] ); 86 processedchars++; 87 break; 88 89 case 'v': 90 currpos += sprintf( &pwbuf[ currpos ], MASK, toLatin1['\v'] ); 91 processedchars++; 92 break; 93 94 case 'b': 95 currpos += sprintf( &pwbuf[ currpos ], MASK, toLatin1['\b'] ); 96 processedchars++; 97 break; 98 99 case 'r': 100 currpos += sprintf( &pwbuf[ currpos ], MASK, toLatin1['\r'] ); 101 processedchars++; 102 break; 103 104 case 'f': 105 currpos += sprintf( &pwbuf[ currpos ], MASK, toLatin1['\f'] ); 106 processedchars++; 107 break; 108 109 case 'a': 110 currpos += sprintf( &pwbuf[ currpos ], MASK, toLatin1['\a'] ); 111 processedchars++; 112 break; 113 114 case '\\': 115 currpos += sprintf( &pwbuf[ currpos ], MASK, toLatin1['\\'] ); 116 processedchars++; 117 break; 118 119 case '?': 120 currpos += sprintf( &pwbuf[ currpos ], MASK, toLatin1['\?'] ); 121 processedchars++; 122 break; 123 124 case '\'': 125 currpos += sprintf( &pwbuf[ currpos ], MASK, toLatin1['\''] ); 126 processedchars++; 127 break; 128 129 case '"': 130 currpos += sprintf( &pwbuf[ currpos ], MASK, toLatin1['\"'] ); 131 processedchars++; 132 break; 133 134 /* octal coded character? -> copy */ 135 case '0': 136 case '1': 137 case '2': 138 case '3': 139 case '4': 140 case '5': 141 case '6': 142 case '7': 143 { 144 int startpos = currpos; 145 146 pwbuf[ currpos++ ] = '\\'; 147 148 while( p[ processedchars ] >= '0' && p[ processedchars ] <= '7' && (currpos < startpos + 4) ) 149 pwbuf[ currpos++ ] = (unsigned char)p[ processedchars++ ]; 150 break; 151 } 152 153 /* hex coded character? -> copy */ 154 case 'x': 155 case 'X': 156 { 157 int startpos = currpos; 158 159 pwbuf[ currpos++ ] = '\\'; 160 pwbuf[ currpos++ ] = 'x'; 161 processedchars++; 162 163 while( isxdigit( p[ processedchars ] ) && (currpos < startpos + 4) ) 164 pwbuf[ currpos++ ] = (unsigned char)p[ processedchars++ ]; 165 break; 166 } 167 168 } 169 } 170 else 171 currpos += sprintf( &pwbuf[ currpos ], MASK, toLatin1[p[ processedchars++ ]] ); 172 173 } 174 175 /* copy last " or ' */ 176 pwbuf[ currpos++ ] = p[ processedchars ]; 177 178 return currpos; 179 } 180 181 void 182 maketokenrow(int size, Tokenrow * trp) 183 { 184 trp->max = size; 185 if (size > 0) 186 trp->bp = (Token *) domalloc(size * sizeof(Token)); 187 else 188 trp->bp = NULL; 189 trp->tp = trp->bp; 190 trp->lp = trp->bp; 191 } 192 193 Token * 194 growtokenrow(Tokenrow * trp) 195 { 196 int ncur = trp->tp - trp->bp; 197 int nlast = trp->lp - trp->bp; 198 199 trp->max = 3 * trp->max / 2 + 1; 200 trp->bp = (Token *) realloc(trp->bp, trp->max * sizeof(Token)); 201 trp->lp = &trp->bp[nlast]; 202 trp->tp = &trp->bp[ncur]; 203 return trp->lp; 204 } 205 206 /* 207 * Compare a row of tokens, ignoring the content of WS; return !=0 if different 208 */ 209 int 210 comparetokens(Tokenrow * tr1, Tokenrow * tr2) 211 { 212 Token *tp1, *tp2; 213 214 tp1 = tr1->tp; 215 tp2 = tr2->tp; 216 if (tr1->lp - tp1 != tr2->lp - tp2) 217 return 1; 218 for (; tp1 < tr1->lp; tp1++, tp2++) 219 { 220 if (tp1->type != tp2->type 221 || (tp1->wslen == 0) != (tp2->wslen == 0) 222 || tp1->len != tp2->len 223 || strncmp((char *) tp1->t, (char *) tp2->t, tp1->len) != 0) 224 return 1; 225 } 226 return 0; 227 } 228 229 /* 230 * replace ntok tokens starting at dtr->tp with the contents of str. 231 * tp ends up pointing just beyond the replacement. 232 * Canonical whitespace is assured on each side. 233 */ 234 void 235 insertrow(Tokenrow * dtr, int ntok, Tokenrow * str) 236 { 237 int nrtok = rowlen(str); 238 239 dtr->tp += ntok; 240 adjustrow(dtr, nrtok - ntok); 241 dtr->tp -= ntok; 242 movetokenrow(dtr, str); 243 dtr->tp += nrtok; 244 } 245 246 /* 247 * make sure there is WS before trp->tp, if tokens might merge in the output 248 */ 249 void 250 makespace(Tokenrow * trp, Token * ntp) 251 { 252 uchar *tt; 253 Token *tp = trp->tp; 254 255 if (tp >= trp->lp) 256 return; 257 258 if (ntp->wslen) 259 { 260 tt = newstring(tp->t, tp->len, ntp->wslen); 261 strncpy((char *)tt, (char *)ntp->t - ntp->wslen, ntp->wslen); 262 tp->t = tt + ntp->wslen; 263 tp->wslen = ntp->wslen; 264 tp->flag |= XPWS; 265 } 266 } 267 268 /* 269 * Copy an entire tokenrow into another, at tp. 270 * It is assumed that there is enough space. 271 * Not strictly conforming. 272 */ 273 void 274 movetokenrow(Tokenrow * dtr, Tokenrow * str) 275 { 276 int nby; 277 278 /* nby = sizeof(Token) * (str->lp - str->bp); */ 279 nby = (char *) str->lp - (char *) str->bp; 280 memmove(dtr->tp, str->bp, nby); 281 } 282 283 /* 284 * Move the tokens in a row, starting at tr->tp, rightward by nt tokens; 285 * nt may be negative (left move). 286 * The row may need to be grown. 287 * Non-strictly conforming because of the (char *), but easily fixed 288 */ 289 void 290 adjustrow(Tokenrow * trp, int nt) 291 { 292 int nby, size; 293 294 if (nt == 0) 295 return; 296 size = (trp->lp - trp->bp) + nt; 297 while (size > trp->max) 298 growtokenrow(trp); 299 /* nby = sizeof(Token) * (trp->lp - trp->tp); */ 300 nby = (char *) trp->lp - (char *) trp->tp; 301 if (nby) 302 memmove(trp->tp + nt, trp->tp, nby); 303 trp->lp += nt; 304 } 305 306 /* 307 * Copy a row of tokens into the destination holder, allocating 308 * the space for the contents. Return the destination. 309 */ 310 Tokenrow * 311 copytokenrow(Tokenrow * dtr, Tokenrow * str) 312 { 313 int len = rowlen(str); 314 315 maketokenrow(len, dtr); 316 movetokenrow(dtr, str); 317 dtr->lp += len; 318 return dtr; 319 } 320 321 /* 322 * Produce a copy of a row of tokens. Start at trp->tp. 323 * The value strings are copied as well. The first token 324 * has WS available. 325 */ 326 Tokenrow * 327 normtokenrow(Tokenrow * trp) 328 { 329 Token *tp; 330 Tokenrow *ntrp = new(Tokenrow); 331 int len; 332 333 len = trp->lp - trp->tp; 334 if (len <= 0) 335 len = 1; 336 maketokenrow(len, ntrp); 337 for (tp = trp->tp; tp < trp->lp; tp++) 338 { 339 *ntrp->lp = *tp; 340 if (tp->len) 341 { 342 ntrp->lp->t = newstring(tp->t, tp->len, 1); 343 *ntrp->lp->t++ = ' '; 344 if (tp->wslen) 345 ntrp->lp->wslen = 1; 346 } 347 ntrp->lp++; 348 } 349 if (ntrp->lp > ntrp->bp) 350 ntrp->bp->wslen = 0; 351 return ntrp; 352 } 353 354 /* 355 * Debugging 356 */ 357 void 358 peektokens(Tokenrow * trp, char *str) 359 { 360 Token *tp; 361 362 tp = trp->tp; 363 flushout(); 364 if (str) 365 fprintf(stderr, "%s ", str); 366 if (tp < trp->bp || tp > trp->lp) 367 fprintf(stderr, "(tp offset %ld) ", (long int) (tp - trp->bp)); 368 for (tp = trp->bp; tp < trp->lp && tp < trp->bp + 32; tp++) 369 { 370 if (tp->type != NL) 371 { 372 int c = tp->t[tp->len]; 373 374 tp->t[tp->len] = 0; 375 fprintf(stderr, "%s", tp->t); 376 tp->t[tp->len] = (uchar) c; 377 } 378 fprintf(stderr, tp == trp->tp ? "{%x*} " : "{%x} ", tp->type); 379 } 380 fprintf(stderr, "\n"); 381 fflush(stderr); 382 } 383 384 void 385 puttokens(Tokenrow * trp) 386 { 387 Token *tp; 388 int len; 389 uchar *p; 390 391 if (Vflag) 392 peektokens(trp, ""); 393 tp = trp->bp; 394 for (; tp < trp->lp; tp++) 395 { 396 if (tp->type != NL) 397 { 398 len = tp->len + tp->wslen; 399 p = tp->t - tp->wslen; 400 401 /* add parameter check to delete operator? */ 402 if( Dflag ) 403 { 404 if( (tp->type == NAME) && (strncmp( (char*)p, "delete", len ) == 0) ) 405 { 406 Token* ntp = tp; 407 ntp++; 408 409 if( ntp->type == NAME ) 410 { 411 uchar* np = ntp->t - ntp->wslen; 412 int nlen = ntp->len + ntp->wslen; 413 414 memcpy(wbp, "if(", 3 ); 415 wbp += 4; 416 memcpy(wbp, np, nlen ); 417 wbp += nlen; 418 memcpy(wbp, ")", 1 ); 419 wbp++; 420 421 memcpy(wbp, p, len); 422 } 423 } 424 } 425 426 /* EBCDIC to ANSI conversion requested? */ 427 if( Aflag ) 428 { 429 /* keyword __ToLatin1__ found? -> do conversion! */ 430 if( EBCDIC_StartTokenDetected ) 431 { 432 /* previous token was 'extern'? -> don't convert current token! */ 433 if( EBCDIC_ExternTokenDetected ) 434 { 435 EBCDIC_ExternTokenDetected = 0; 436 memcpy(wbp, p, len); 437 } 438 else 439 { 440 /* current token is keyword 'extern'? -> don't convert following token! */ 441 if( (tp->wslen == 0) && (strncmp( (char*)p, "extern", len ) == 0) ) 442 { 443 EBCDIC_ExternTokenDetected = 1; 444 memcpy(wbp, p, len); 445 } 446 else 447 { 448 /* token is string or char? -> process EBCDIC to ANSI conversion */ 449 if ((tp->type == STRING) || (tp->type == CCON)) 450 len = memcpy_EBCDIC(wbp, p, len); 451 else 452 memcpy(wbp, p, len); 453 } 454 } 455 } 456 else 457 /* keyword __ToLatin1__ found? -> don't copy keyword and start conversion */ 458 if( (tp->type == NAME) && (strncmp( (char*)p, "__ToLatin1__", len) == 0) ) 459 { 460 EBCDIC_StartTokenDetected = 1; 461 len = 0; 462 } 463 else 464 memcpy(wbp, p, len); 465 } 466 else 467 memcpy(wbp, p, len); 468 469 wbp += len; 470 } 471 else 472 *wbp++ = '\n'; 473 474 if (wbp >= &wbuf[OBS]) 475 { 476 if ( write(1, wbuf, OBS) != -1 ) { 477 if (wbp > &wbuf[OBS]) 478 memcpy(wbuf, wbuf + OBS, wbp - &wbuf[OBS]); 479 wbp -= OBS; 480 } 481 else exit(1); 482 } 483 } 484 trp->tp = tp; 485 if (cursource->fd == 0) 486 flushout(); 487 } 488 489 void 490 flushout(void) 491 { 492 if (wbp > wbuf) 493 { 494 if ( write(1, wbuf, wbp - wbuf) != -1) 495 wbp = wbuf; 496 else 497 exit(1); 498 } 499 } 500 501 /* 502 * turn a row into just a newline 503 */ 504 void 505 setempty(Tokenrow * trp) 506 { 507 trp->tp = trp->bp; 508 trp->lp = trp->bp + 1; 509 *trp->bp = nltoken; 510 } 511 512 /* 513 * generate a number 514 */ 515 char * 516 outnum(char *p, int n) 517 { 518 if (n >= 10) 519 p = outnum(p, n / 10); 520 *p++ = (char) (n % 10 + '0'); 521 return p; 522 } 523 524 /* 525 * allocate and initialize a new string from s, of length l, at offset o 526 * Null terminated. 527 */ 528 uchar * 529 newstring(uchar * s, int l, int o) 530 { 531 uchar *ns = (uchar *) domalloc(l + o + 1); 532 533 ns[l + o] = '\0'; 534 return (uchar *) strncpy((char *) ns + o, (char *) s, l) - o; 535 } 536