1/* $NetBSD: pi.c,v 1.24 2023/08/26 15:18:27 rillig Exp $ */ 2 3/* 4 * Copyright (c) 1980, 1993 5 * The Regents of the University of California. All rights reserved. 6 * 7 * Redistribution and use in source and binary forms, with or without 8 * modification, are permitted provided that the following conditions 9 * are met: 10 * 1. Redistributions of source code must retain the above copyright 11 * notice, this list of conditions and the following disclaimer. 12 * 2. Redistributions in binary form must reproduce the above copyright 13 * notice, this list of conditions and the following disclaimer in the 14 * documentation and/or other materials provided with the distribution. 15 * 3. Neither the name of the University nor the names of its contributors 16 * may be used to endorse or promote products derived from this software 17 * without specific prior written permission. 18 * 19 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 20 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 21 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 22 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 23 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 24 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 25 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 26 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 27 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 28 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 29 * SUCH DAMAGE. 30 */ 31 32#include <sys/cdefs.h> 33#ifndef lint 34#if 0 35static char sccsid[] = "@(#)pi.c 8.1 (Berkeley) 6/6/93"; 36#endif 37__RCSID("$NetBSD: pi.c,v 1.24 2023/08/26 15:18:27 rillig Exp $"); 38#endif /* not lint */ 39 40#include <stdio.h> 41#include <ctype.h> 42#include <string.h> 43#include <stdlib.h> 44#include "error.h" 45 46#if 0 /* not const-correct */ 47static char *unk_hdr[] = {"In", "program", "???"}; 48#else 49DECL_STRINGS_3(static, unk_hdr, "In", "program", "???"); 50#endif 51 52static char *c_linenumber; 53static char **c_header = &unk_hdr[0]; 54 55static bool alldigits(const char *); 56static bool isdateformat(int, char **); 57static bool instringset(const char *, const char **); 58static bool piptr(const char *); 59 60 61/* 62 * Attempt to handle error messages produced by pi (and by pc) 63 * 64 * problem #1: There is no file name available when a file does not 65 * use a #include; this will have to be given to error 66 * in the command line. 67 * problem #2: pi doesn't always tell you what line number 68 * a error refers to; for example during the tree 69 * walk phase of code generation and error detection, 70 * an error can refer to "variable foo in procedure bletch" 71 * without giving a line number 72 * problem #3: line numbers, when available, are attached to 73 * the source line, along with the source line itself 74 * These line numbers must be extracted, and 75 * the source line thrown away. 76 * problem #4: Some error messages produce more than one line number 77 * on the same message. 78 * There are only two (I think): 79 * %s undefined on line%s 80 * %s improperly used on line%s 81 * here, the %s makes line plural or singular. 82 * 83 * Here are the error strings used in pi version 1.2 that can refer 84 * to a file name or line number: 85 * 86 * Multiply defined label in case, lines %d and %d 87 * Goto %s from line %d is into a structured statement 88 * End matched %s on line %d 89 * Inserted keyword end matching %s on line %d 90 * 91 * Here are the general pi patterns recognized: 92 * define piptr == -.*^-.* 93 * define msg = .* 94 * define digit = [0-9] 95 * definename = .* 96 * define date_format letter*3 letter*3 (digit | (digit digit)) 97 * (digit | (digit digit)):digit*2 digit*4 98 * 99 * {e,E} (piptr) (msg) Encounter an error during textual scan 100 * E {digit}* - (msg) Have an error message that refers to a new line 101 * E - msg Have an error message that refers to current 102 * function, program or procedure 103 * (date_format) (name): When switch compilation files 104 * ... (msg) When refer to the previous line 105 * 'In' ('procedure'|'function'|'program') (name): 106 * pi is now complaining about 2nd pass errors. 107 * 108 * Here is the output from a compilation 109 * 110 * 111 * 2 var i:integer; 112 * e --------------^--- Inserted ';' 113 * E 2 - All variables must be declared in one var part 114 * E 5 - Include filename must end in .i 115 * Mon Apr 21 15:56 1980 test.h: 116 * 2 begin 117 * e ------^--- Inserted ';' 118 * Mon Apr 21 16:06 1980 test.p: 119 * E 2 - Function type must be specified 120 * 6 procedure foo(var x:real); 121 * e ------^--- Inserted ';' 122 * In function bletch: 123 * E - No assignment to the function variable 124 * w - variable x is never used 125 * E 6 - foo is already defined in this block 126 * In procedure foo: 127 * w - variable x is neither used nor set 128 * 9 z : = 23; 129 * E --------------^--- Undefined variable 130 * 10 y = [1]; 131 * e ----------------^--- Inserted ':' 132 * 13 z := 345.; 133 * e -----------------------^--- Digits required after decimal point 134 * E 10 - Constant set involved in non set context 135 * E 11 - Type clash: real is incompatible with integer 136 * ... Type of expression clashed with type of variable in assignment 137 * E 12 - Parameter type not identical to type of var parameter x of foo 138 * In program mung: 139 * w - variable y is never used 140 * w - type foo is never used 141 * w - function bletch is never used 142 * E - z undefined on lines 9 13 143 */ 144static const char *Months[] = { 145 "Jan", "Feb", "Mar", "Apr", "May", "Jun", 146 "Jul", "Aug", "Sep", "Oct", "Nov", "Dec", 147 NULL 148}; 149static const char *Days[] = { 150 "Sun", "Mon", "Tue", "Wed", "Thu", "Fri", "Sat", NULL 151}; 152static const char *Piroutines[] = { 153 "program", "function", "procedure", NULL 154}; 155 156 157static bool structured, multiple; 158 159#if 0 /* not const-correct */ 160static char *pi_Endmatched[] = {"End", "matched"}; 161static char *pi_Inserted[] = {"Inserted", "keyword", "end", "matching"}; 162 163static char *pi_multiple[] = {"Multiply", "defined", "label", "in", "case,", "line"}; 164static char *pi_structured[] = {"is", "into", "a", "structured", "statement"}; 165 166static char *pi_und1[] = {"undefined", "on", "line"}; 167static char *pi_und2[] = {"undefined", "on", "lines"}; 168static char *pi_imp1[] = {"improperly", "used", "on", "line"}; 169static char *pi_imp2[] = {"improperly", "used", "on", "lines"}; 170 171#else 172DECL_STRINGS_2(static, pi_Endmatched, "End", "matched"); 173DECL_STRINGS_4(static, pi_Inserted, "Inserted", "keyword", "end", "matching"); 174 175DECL_STRINGS_6(static, pi_multiple, 176 "Multiply", "defined", "label", "in", "case,", "line"); 177DECL_STRINGS_5(static, pi_structured, 178 "is", "into", "a", "structured", "statement"); 179 180DECL_STRINGS_3(static, pi_und1, "undefined", "on", "line"); 181DECL_STRINGS_3(static, pi_und2, "undefined", "on", "lines"); 182DECL_STRINGS_4(static, pi_imp1, "improperly", "used", "on", "line"); 183DECL_STRINGS_4(static, pi_imp2, "improperly", "used", "on", "lines"); 184 185#endif 186 187static bool 188alldigits(const char *string) 189{ 190 for (; *string != '\0' && isdigit((unsigned char)*string); string++) 191 continue; 192 return *string == '\0'; 193} 194 195static bool 196instringset(const char *member, const char **set) 197{ 198 for (; *set != NULL; set++) 199 if (strcmp(*set, member) == 0) 200 return true; 201 return false; 202} 203 204static bool 205isdateformat(int wordc, char **wordv) 206{ 207 return wordc == 5 208 && instringset(wordv[0], Days) 209 && instringset(wordv[1], Months) 210 && alldigits(wordv[2]) 211 && alldigits(wordv[4]); 212} 213 214static bool 215piptr(const char *string) 216{ 217 if (*string != '-') 218 return false; 219 while (*string == '-') 220 string++; 221 if (*string != '^') 222 return false; 223 string++; 224 while (*string == '-') 225 string++; 226 return *string == '\0'; 227} 228 229Errorclass 230pi(void) 231{ 232 char **nwordv; 233 234 nwordv = NULL; 235 if (cur_wordc < 2) 236 return C_UNKNOWN; 237 if (strlen(cur_wordv[1]) == 1 238 && ( cur_wordv[1][0] == 'e' || cur_wordv[1][0] == 'E') 239 && piptr(cur_wordv[2]) 240 ) { 241 bool longpiptr = false; 242 243 /* 244 * We have recognized a first pass error of the form: 245 * letter ------^---- message 246 * 247 * turn into an error message of the form: 248 * 249 * file line 'pascal errortype' letter \n |---- message 250 * or of the form: 251 * file line letter |---- message 252 * when there are strlen("(*[pi]") or more 253 * preceding '-' on the error pointer. 254 * 255 * Where the | is intended to be a down arrow, so that 256 * the pi error messages can be inserted above the 257 * line in error, instead of below. (All of the other 258 * languages put their messages before the source line, 259 * instead of after it as does pi.) 260 * 261 * where the pointer to the error has been truncated 262 * by 6 characters to account for the fact that 263 * the pointer points into a tab preceded input line. 264 */ 265 language = INPI; 266 (void)substitute(cur_wordv[2], '^', '|'); 267 longpiptr = position(cur_wordv[2],'|') > (6+8); 268 nwordv = wordvsplice(longpiptr ? 2 : 4, cur_wordc, cur_wordv+1); 269 nwordv[0] = strdup(currentfilename); 270 nwordv[1] = strdup(c_linenumber); 271 if (!longpiptr) { 272 nwordv[2] = Strdup("pascal errortype"); /* XXX leaked */ 273 nwordv[3] = cur_wordv[1]; 274 nwordv[4] = strdup("%%%\n"); 275 if (strlen(nwordv[5]) > (8-2)) /* this is the pointer */ 276 nwordv[5] += (8-2); /* bump over 6 characters */ 277 } 278 cur_wordv = nwordv - 1; /* convert to 1 based */ 279 cur_wordc += longpiptr ? 2 : 4; 280 return C_TRUE; 281 } 282 if (cur_wordc >= 4 283 && strlen(cur_wordv[1]) == 1 284 && (*cur_wordv[1] == 'E' || *cur_wordv[1] == 'w' || *cur_wordv[1] == 'e') 285 && alldigits(cur_wordv[2]) 286 && strlen(cur_wordv[3]) == 1 287 && cur_wordv[3][0] == '-' 288 ) { 289 /* 290 * Message of the form: letter linenumber - message 291 * Turn into form: filename linenumber letter - message 292 */ 293 language = INPI; 294 nwordv = wordvsplice(1, cur_wordc, cur_wordv + 1); 295 nwordv[0] = strdup(currentfilename); 296 nwordv[1] = cur_wordv[2]; 297 nwordv[2] = cur_wordv[1]; 298 c_linenumber = cur_wordv[2]; 299 cur_wordc += 1; 300 cur_wordv = nwordv - 1; 301 return C_TRUE; 302 } 303 if (cur_wordc >= 3 304 && strlen(cur_wordv[1]) == 1 305 && (*cur_wordv[1] == 'E' || *cur_wordv[1] == 'w' || *cur_wordv[1] == 'e') 306 && strlen(cur_wordv[2]) == 1 307 && cur_wordv[2][0] == '-' 308 ) { 309 /* 310 * Message of the form: letter - message 311 * 312 * This happens only when we are traversing the tree 313 * during the second pass of pi, and discover semantic 314 * errors. 315 * 316 * We have already (presumably) saved the header message 317 * and can now construct a nulled error message for the 318 * current file. 319 * 320 * Turns into a message of the form: 321 * filename (header) letter - message 322 * 323 * First, see if it is a message referring to more than 324 * one line number. Only of the form: 325 * %s undefined on line%s 326 * %s improperly used on line%s 327 */ 328 bool undefined = false; 329 int wordindex; 330 331 language = INPI; 332 if ((undefined = wordv_eq(cur_wordv+2, 3, pi_und1)) 333 || (undefined = wordv_eq(cur_wordv+2, 3, pi_und2)) 334 || wordv_eq(cur_wordv+2, 4, pi_imp1) 335 || wordv_eq(cur_wordv+2, 4, pi_imp2) 336 ) { 337 for (wordindex = undefined ? 5 : 6; 338 wordindex <= cur_wordc; 339 wordindex++) { 340 if (nwordv != NULL) { 341 free(nwordv[0]); 342 free(nwordv); 343 } 344 nwordv = wordvsplice(2, undefined ? 2 : 3, cur_wordv+1); 345 nwordv[0] = strdup(currentfilename); 346 nwordv[1] = cur_wordv[wordindex]; 347 if (wordindex != cur_wordc) 348 erroradd(undefined ? 4 : 5, nwordv, 349 C_TRUE, C_UNKNOWN); 350 } 351 cur_wordc = undefined ? 4 : 5; 352 cur_wordv = nwordv - 1; 353 return C_TRUE; 354 } 355 356 nwordv = wordvsplice(1+3, cur_wordc, cur_wordv+1); 357 nwordv[0] = strdup(currentfilename); 358 nwordv[1] = strdup(c_header[0]); 359 nwordv[2] = strdup(c_header[1]); 360 nwordv[3] = strdup(c_header[2]); 361 cur_wordv = nwordv - 1; 362 cur_wordc += 1 + 3; 363 return C_THISFILE; 364 } 365 if (strcmp(cur_wordv[1], "...") == 0 && c_linenumber != NULL && 366 currentfilename != default_currentfilename) { 367 /* 368 * have a continuation error message 369 * of the form: ... message 370 * Turn into form : filename linenumber message 371 */ 372 language = INPI; 373 nwordv = wordvsplice(1, cur_wordc, cur_wordv+1); 374 nwordv[0] = strdup(currentfilename); 375 nwordv[1] = strdup(c_linenumber); 376 cur_wordv = nwordv - 1; 377 cur_wordc += 1; 378 return C_TRUE; 379 } 380 if (cur_wordc == 6 381 && lastchar(cur_wordv[6]) == ':' 382 && isdateformat(5, cur_wordv + 1) 383 ) { 384 /* 385 * Have message that tells us we have changed files 386 */ 387 language = INPI; 388 currentfilename = strdup(cur_wordv[6]); 389 clob_last(currentfilename, '\0'); 390 return C_SYNC; 391 } 392 if (cur_wordc == 3 393 && strcmp(cur_wordv[1], "In") == 0 394 && lastchar(cur_wordv[3]) == ':' 395 && instringset(cur_wordv[2], Piroutines) 396 ) { 397 language = INPI; 398 c_header = wordvsplice(0, cur_wordc, cur_wordv+1); 399 return C_SYNC; 400 } 401 402 /* 403 * now, check for just the line number followed by the text 404 */ 405 if (alldigits(cur_wordv[1])) { 406 language = INPI; 407 c_linenumber = cur_wordv[1]; 408 return C_IGNORE; 409 } 410 411 /* 412 * Attempt to match messages refering to a line number 413 * 414 * Multiply defined label in case, lines %d and %d 415 * Goto %s from line %d is into a structured statement 416 * End matched %s on line %d 417 * Inserted keyword end matching %s on line %d 418 */ 419 structured = false; 420 multiple = false; 421 if ( 422 (cur_wordc == 6 && wordv_eq(cur_wordv+1, 2, pi_Endmatched)) 423 || (cur_wordc == 8 && wordv_eq(cur_wordv+1, 4, pi_Inserted)) 424 || (multiple = (cur_wordc == 9 && wordv_eq(cur_wordv+1,6, pi_multiple))) 425 || (structured = (cur_wordc == 10 && wordv_eq(cur_wordv+6,5, pi_structured))) 426 ) { 427 language = INPI; 428 nwordv = wordvsplice(2, cur_wordc, cur_wordv+1); 429 nwordv[0] = strdup(currentfilename); 430 nwordv[1] = structured ? cur_wordv [5] : cur_wordv[cur_wordc]; 431 cur_wordc += 2; 432 cur_wordv = nwordv - 1; 433 if (!multiple) 434 return C_TRUE; 435 erroradd(cur_wordc, nwordv, C_TRUE, C_UNKNOWN); 436 nwordv = wordvsplice(0, cur_wordc, nwordv); 437 nwordv[1] = cur_wordv[cur_wordc - 2]; 438 return C_TRUE; 439 } 440 return C_UNKNOWN; 441} 442