| 1 | static char sccsid[] = "@(#)lexi.c 4.1 (Berkeley) %G%"; |
| 2 | |
| 3 | /* |
| 4 | |
| 5 | Copyright (C) 1976 |
| 6 | by the |
| 7 | Board of Trustees |
| 8 | of the |
| 9 | University of Illinois |
| 10 | |
| 11 | All rights reserved |
| 12 | |
| 13 | |
| 14 | NAME: |
| 15 | lexi |
| 16 | |
| 17 | FUNCTION: |
| 18 | This is the token scanner for indent |
| 19 | |
| 20 | ALGORITHM: |
| 21 | 1) Strip off intervening blanks and/or tabs. |
| 22 | 2) If it is an alphanumeric token, move it to the token buffer "token". |
| 23 | Check if it is a special reserved word that indent will want to |
| 24 | know about. |
| 25 | 3) Non-alphanumeric tokens are handled with a big switch statement. A |
| 26 | flag is kept to remember if the last token was a "unary delimiter", |
| 27 | which forces a following operator to be unary as opposed to binary. |
| 28 | |
| 29 | PARAMETERS: |
| 30 | None |
| 31 | |
| 32 | RETURNS: |
| 33 | An integer code indicating the type of token scanned. |
| 34 | |
| 35 | GLOBALS: |
| 36 | buf_ptr = |
| 37 | had_eof |
| 38 | last_u_d = Set to true iff this token is a "unary delimiter" |
| 39 | |
| 40 | CALLS: |
| 41 | fill_buffer |
| 42 | printf (lib) |
| 43 | |
| 44 | CALLED BY: |
| 45 | main |
| 46 | |
| 47 | NOTES: |
| 48 | Start of comment is passed back so that the comment can be scanned by |
| 49 | pr_comment. |
| 50 | |
| 51 | Strings and character literals are returned just like identifiers. |
| 52 | |
| 53 | HISTORY: |
| 54 | initial coding November 1976 D A Willcox of CAC |
| 55 | 1/7/77 D A Willcox of CAC Fix to provide proper handling |
| 56 | of "int a -1;" |
| 57 | |
| 58 | */\f |
| 59 | |
| 60 | /* Here we have the token scanner for indent. It scans off one token and |
| 61 | puts it in the global variable "token". It returns a code, indicating the |
| 62 | type of token scanned. */ |
| 63 | |
| 64 | #include "indent_globs.h"; |
| 65 | #include "indent_codes.h"; |
| 66 | |
| 67 | |
| 68 | |
| 69 | #define alphanum 1 |
| 70 | #define opchar 3 |
| 71 | |
| 72 | struct templ { |
| 73 | char *rwd; |
| 74 | int rwcode; |
| 75 | }; |
| 76 | |
| 77 | struct templ specials[] = |
| 78 | { |
| 79 | "switch", 1, |
| 80 | "case", 2, |
| 81 | "struct", 3, |
| 82 | "default", 2, |
| 83 | "int", 4, |
| 84 | "char", 4, |
| 85 | "float", 4, |
| 86 | "double", 4, |
| 87 | "long", 4, |
| 88 | "short", 4, |
| 89 | "typdef", 4, |
| 90 | "unsigned", 4, |
| 91 | "register", 4, |
| 92 | "static", 4, |
| 93 | "global", 4, |
| 94 | "extern", 4, |
| 95 | "if", 5, |
| 96 | "while", 5, |
| 97 | "for", 5, |
| 98 | "else", 6, |
| 99 | "do", 6, |
| 100 | "sizeof", 0, |
| 101 | 0, 0 |
| 102 | }; |
| 103 | |
| 104 | char chartype[128] = |
| 105 | { /* this is used to facilitate the decision of what type |
| 106 | (alphanumeric, operator) each character is */ |
| 107 | 0, 0, 0, 0, 0, 0, 0, 0, |
| 108 | 0, 0, 0, 0, 0, 0, 0, 0, |
| 109 | 0, 0, 0, 0, 0, 0, 0, 0, |
| 110 | 0, 0, 0, 0, 0, 0, 0, 0, |
| 111 | 0, 3, 0, 0, 0, 3, 3, 0, |
| 112 | 0, 0, 3, 3, 0, 3, 3, 3, |
| 113 | 1, 1, 1, 1, 1, 1, 1, 1, |
| 114 | 1, 1, 0, 0, 3, 3, 3, 3, |
| 115 | 0, 1, 1, 1, 1, 1, 1, 1, |
| 116 | 1, 1, 1, 1, 1, 1, 1, 1, |
| 117 | 1, 1, 1, 1, 1, 1, 1, 1, |
| 118 | 1, 1, 1, 0, 0, 0, 3, 1, |
| 119 | 0, 1, 1, 1, 1, 1, 1, 1, |
| 120 | 1, 1, 1, 1, 1, 1, 1, 1, |
| 121 | 1, 1, 1, 1, 1, 1, 1, 1, |
| 122 | 1, 1, 1, 0, 3, 0, 3, 0 |
| 123 | }; |
| 124 | |
| 125 | int last_nl = true; |
| 126 | /* this is true if the last thing scanned was a newline */ |
| 127 | |
| 128 | |
| 129 | |
| 130 | int lexi () { |
| 131 | register char *tok; |
| 132 | /* local pointer to next char in token */ |
| 133 | register int i; |
| 134 | /* local loop counter */ |
| 135 | register char *j; |
| 136 | /* used for searching thru list of reserved words */ |
| 137 | int unary_delim; |
| 138 | /* this is set to 1 if the current token forces a following operator to be |
| 139 | unary */ |
| 140 | static int last_code; |
| 141 | /* the last token type returned */ |
| 142 | static int l_struct; |
| 143 | /* set to 1 if the last token was 'struct' */ |
| 144 | int found_it; |
| 145 | int code; /* internal code to be returned */ |
| 146 | char qchar; /* the delimiter character for a string */ |
| 147 | |
| 148 | tok = token; /* point to start of place to save token */ |
| 149 | unary_delim = false; |
| 150 | col_1 = last_nl; /* tell world that this token started in column |
| 151 | 1 iff the last thing scanned was nl */ |
| 152 | last_nl = false; |
| 153 | |
| 154 | while (*buf_ptr == ' ' || *buf_ptr == '\t') { |
| 155 | /* get rid of blanks */ |
| 156 | col_1 = false; /* leading blanks imply token is not in column 1 |
| 157 | */ |
| 158 | if (++buf_ptr >= buf_end) |
| 159 | fill_buffer (); |
| 160 | } |
| 161 | |
| 162 | /*----------------------------------------------------------*\ |
| 163 | | Scan an alphanumeric token |
| 164 | \*----------------------------------------------------------*/ |
| 165 | |
| 166 | if (chartype[*buf_ptr & 0177] == alphanum) { |
| 167 | /* we have a character or number */ |
| 168 | while (chartype[*buf_ptr & 0177] == alphanum) { |
| 169 | /* copy it over */ |
| 170 | *tok++ = *buf_ptr++; |
| 171 | if (buf_ptr >= buf_end) |
| 172 | fill_buffer (); |
| 173 | } |
| 174 | |
| 175 | *tok++ = '\0'; |
| 176 | |
| 177 | if (l_struct) { /* if last token was 'struct', then this token |
| 178 | should be treated as a declaration */ |
| 179 | l_struct = false; |
| 180 | last_code = ident; |
| 181 | last_u_d = true; |
| 182 | return (decl); |
| 183 | } |
| 184 | |
| 185 | last_u_d = false; /* operator after indentifier is binary */ |
| 186 | |
| 187 | for (i = 0; specials[i].rwd != 0; ++i) { |
| 188 | /* this loop will check if the token is a keyword. if so, a following |
| 189 | operator is unary */ |
| 190 | last_code = ident; /* remember that this is the code we will return |
| 191 | */ |
| 192 | j = specials[i].rwd; |
| 193 | /* point at ith reserved word */ |
| 194 | tok = token; /* point at scanned toekn */ |
| 195 | found_it = true; /* set to false if not found */ |
| 196 | do { |
| 197 | if (*tok++ != *j) { |
| 198 | found_it = false; |
| 199 | break; |
| 200 | } |
| 201 | } while (*j++); |
| 202 | |
| 203 | if (found_it) { /* we have a keyword */ |
| 204 | last_u_d = true; |
| 205 | switch (specials[i].rwcode) { |
| 206 | case 1: /* it is a switch */ |
| 207 | return (swstmt); |
| 208 | case 2: /* a case or default */ |
| 209 | return (casestmt); |
| 210 | |
| 211 | case 3: /* a "struct" */ |
| 212 | l_struct = true; |
| 213 | /* Next time around, we will want to know that we have had |
| 214 | a 'struct' */ |
| 215 | case 4: /* one of the declaration keywords */ |
| 216 | if(p_l_follow) break; /* inside parens: cast */ |
| 217 | last_code = decl; |
| 218 | return (decl); |
| 219 | |
| 220 | case 5: /* if, while, for */ |
| 221 | return (sp_paren); |
| 222 | |
| 223 | case 6: /* do, else */ |
| 224 | return (sp_nparen); |
| 225 | |
| 226 | default: /* all others are treated like any other |
| 227 | identifier */ |
| 228 | return (ident); |
| 229 | } /* end of switch */ |
| 230 | } /* end of if (found_it) */ |
| 231 | |
| 232 | } |
| 233 | |
| 234 | if (last_code == decl) /* if this is a declared variable, then |
| 235 | following sign is unary */ |
| 236 | last_u_d = true; /* will make "int a -1" work */ |
| 237 | last_code = ident; |
| 238 | return (ident); /* the ident is not in the list */ |
| 239 | } /* end of procesing for alpanum character */ |
| 240 | |
| 241 | |
| 242 | |
| 243 | /*----------------------------------------------------------*\ |
| 244 | | Scan a non-alphanumeric token |
| 245 | \*----------------------------------------------------------*/ |
| 246 | |
| 247 | *tok++ = *buf_ptr; /* if it is only a one-character token, it is |
| 248 | moved here */ |
| 249 | *tok = '\0'; |
| 250 | if (++buf_ptr >= buf_end) |
| 251 | fill_buffer (); |
| 252 | |
| 253 | switch (*token) { |
| 254 | case '\n': |
| 255 | unary_delim = last_u_d; |
| 256 | last_nl = true; /* remember that we just had a newline */ |
| 257 | code = (had_eof ? 0 : newline); |
| 258 | /* if data has been exausted, the newline is a dummy, and we should |
| 259 | return code to stop */ |
| 260 | break; |
| 261 | |
| 262 | case '\'': /* start of quoted character */ |
| 263 | qchar = '\''; /* remember final delimiter */ |
| 264 | goto copy_lit; /* and go to common literal code */ |
| 265 | |
| 266 | case '"': /* start of string */ |
| 267 | qchar = '"'; |
| 268 | |
| 269 | copy_lit: |
| 270 | do { /* copy the string */ |
| 271 | while (1) { /* move one character or [/<char>]<char> */ |
| 272 | if (*buf_ptr == '\n') { |
| 273 | /* check for unterminated literal */ |
| 274 | printf ("%d: Unterminated literal\n", line_no); |
| 275 | goto stop_lit; |
| 276 | /* Don't copy any more */ |
| 277 | } |
| 278 | |
| 279 | *tok = *buf_ptr++; |
| 280 | if (buf_ptr >= buf_end) |
| 281 | fill_buffer (); |
| 282 | if (had_eof || ((tok - token) > (bufsize - 2))) { |
| 283 | printf ("Unterminated literal\n"); |
| 284 | ++tok; |
| 285 | goto stop_lit; |
| 286 | /* get outof literal copying loop */ |
| 287 | } |
| 288 | |
| 289 | if (*tok == '\\') { |
| 290 | /* if escape, copy extra char */ |
| 291 | if (*buf_ptr == '\n') |
| 292 | /* check for escaped newline */ |
| 293 | ++line_no; |
| 294 | *(++tok) = *buf_ptr++; |
| 295 | ++tok; /* we must increment this again because we |
| 296 | copied two chars */ |
| 297 | if (buf_ptr >= buf_end) |
| 298 | fill_buffer (); |
| 299 | } |
| 300 | else |
| 301 | break; /* we copied one character */ |
| 302 | } /* end of while (1) */ |
| 303 | } while (*tok++ != qchar); |
| 304 | |
| 305 | stop_lit: |
| 306 | code = ident; |
| 307 | break; |
| 308 | |
| 309 | case ('('): |
| 310 | case ('['): |
| 311 | unary_delim = true; |
| 312 | code = lparen; |
| 313 | break; |
| 314 | |
| 315 | case (')'): |
| 316 | case (']'): |
| 317 | code = rparen; |
| 318 | break; |
| 319 | |
| 320 | case '#': |
| 321 | unary_delim = last_u_d; |
| 322 | code = preesc; |
| 323 | break; |
| 324 | |
| 325 | case '?': |
| 326 | unary_delim = true; |
| 327 | code = question; |
| 328 | break; |
| 329 | |
| 330 | case (':'): |
| 331 | code = colon; |
| 332 | unary_delim = true; |
| 333 | break; |
| 334 | |
| 335 | case (';'): |
| 336 | unary_delim = true; |
| 337 | code = semicolon; |
| 338 | break; |
| 339 | |
| 340 | case ('{'): |
| 341 | unary_delim = true; |
| 342 | code = lbrace; |
| 343 | break; |
| 344 | |
| 345 | case ('}'): |
| 346 | unary_delim = true; |
| 347 | code = rbrace; |
| 348 | break; |
| 349 | |
| 350 | case 014: /* a form feed */ |
| 351 | unary_delim = last_u_d; |
| 352 | last_nl = true; /* remember this so we can set 'col_1' right */ |
| 353 | code = form_feed; |
| 354 | break; |
| 355 | |
| 356 | case (','): |
| 357 | unary_delim = true; |
| 358 | code = comma; |
| 359 | break; |
| 360 | |
| 361 | case '.': |
| 362 | unary_delim = false; |
| 363 | code = period; |
| 364 | break; |
| 365 | |
| 366 | case '-': |
| 367 | case '+': /* check for -, +, --, ++ */ |
| 368 | code = (last_u_d ? unary_op : binary_op); |
| 369 | unary_delim = true; |
| 370 | |
| 371 | if (*buf_ptr == token[0]) { |
| 372 | /* check for doubled character */ |
| 373 | *tok++ = *buf_ptr++; |
| 374 | /* buffer overflow will be checked at end of loop */ |
| 375 | if (last_code == ident || last_code == rparen) { |
| 376 | code = (last_u_d ? unary_op : postop); |
| 377 | /* check for following ++ or -- */ |
| 378 | unary_delim = false; |
| 379 | } |
| 380 | } |
| 381 | else |
| 382 | if (*buf_ptr == '>' || *buf_ptr == '=') |
| 383 | /* check for operator -> or += */ |
| 384 | *tok++ = *buf_ptr++; |
| 385 | /* buffer overflow will be checked at end of switch */ |
| 386 | |
| 387 | break; |
| 388 | |
| 389 | case '=': |
| 390 | if (chartype[*buf_ptr] == opchar) { |
| 391 | /* we have two char assignment */ |
| 392 | *tok++ = *buf_ptr; |
| 393 | /* move second character */ |
| 394 | if (++buf_ptr >= buf_end) |
| 395 | fill_buffer (); |
| 396 | } |
| 397 | |
| 398 | code = binary_op; |
| 399 | unary_delim = true; |
| 400 | if (token[1] != '<' && token[1] != '>') |
| 401 | /* check for possible 3 char operator */ |
| 402 | break; |
| 403 | /* can drop thru!!! */ |
| 404 | |
| 405 | case '>': |
| 406 | case '<': |
| 407 | case '!': /* ops like <, <<, <=, !=, etc */ |
| 408 | if (*buf_ptr == '>' || *buf_ptr == '<' || *buf_ptr == '=') { |
| 409 | *tok++ = *buf_ptr; |
| 410 | if (++buf_ptr >= buf_end) |
| 411 | fill_buffer (); |
| 412 | } |
| 413 | |
| 414 | if (*buf_ptr == '=') |
| 415 | *tok++ = *buf_ptr++; |
| 416 | code = (last_u_d ? unary_op : binary_op); |
| 417 | unary_delim = true; |
| 418 | break; |
| 419 | |
| 420 | default: |
| 421 | if (token[0] == '/' && *buf_ptr == '*') { |
| 422 | /* it is start of comment */ |
| 423 | *tok++ = '*'; |
| 424 | |
| 425 | if (++buf_ptr >= buf_end) |
| 426 | fill_buffer (); |
| 427 | |
| 428 | code = comment; |
| 429 | unary_delim = last_u_d; |
| 430 | break; |
| 431 | } |
| 432 | |
| 433 | while (*(tok - 1) == *buf_ptr || *buf_ptr=='=') { |
| 434 | /* handle ||, &&, etc, and also things as in int *****i */ |
| 435 | *tok++ = *buf_ptr; |
| 436 | if (++buf_ptr >= buf_end) |
| 437 | fill_buffer (); |
| 438 | } |
| 439 | |
| 440 | |
| 441 | code = (last_u_d ? unary_op : binary_op); |
| 442 | unary_delim = true; |
| 443 | |
| 444 | |
| 445 | } /* end of switch */ |
| 446 | |
| 447 | if (code != newline) { |
| 448 | l_struct = false; |
| 449 | last_code = code; |
| 450 | } |
| 451 | |
| 452 | if (buf_ptr >= buf_end) /* check for input buffer empty */ |
| 453 | fill_buffer (); |
| 454 | last_u_d = unary_delim; |
| 455 | *tok = '\0'; /* null terminate the token */ |
| 456 | return (code); |
| 457 | }; |