pi.c revision 1.18 1 /* $NetBSD: pi.c,v 1.18 2011/08/17 13:11:22 christos 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
35 static char sccsid[] = "@(#)pi.c 8.1 (Berkeley) 6/6/93";
36 #endif
37 __RCSID("$NetBSD: pi.c,v 1.18 2011/08/17 13:11:22 christos 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 */
47 static char *unk_hdr[] = {"In", "program", "???"};
48 #else
49 DECL_STRINGS_3(static, unk_hdr, "In", "program", "???");
50 #endif
51
52 static char *c_linenumber;
53 static char **c_header = &unk_hdr[0];
54
55 static boolean alldigits(const char *);
56 static boolean isdateformat(int, char **);
57 static boolean instringset(const char *, const char **);
58 static boolean 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 */
144 static const char *Months[] = {
145 "Jan", "Feb", "Mar", "Apr", "May", "Jun",
146 "Jul", "Aug", "Sep", "Oct", "Nov", "Dec",
147 0
148 };
149 static const char *Days[] = {
150 "Sun", "Mon", "Tue", "Wed", "Thu", "Fri", "Sat", 0
151 };
152 static const char *Piroutines[] = {
153 "program", "function", "procedure", 0
154 };
155
156
157 static boolean structured, multiple;
158
159 #if 0 /* not const-correct */
160 static char *pi_Endmatched[] = {"End", "matched"};
161 static char *pi_Inserted[] = {"Inserted", "keyword", "end", "matching"};
162
163 static char *pi_multiple[] = {"Mutiply", "defined", "label", "in", "case,", "line"};
164 static char *pi_structured[] = {"is", "into", "a", "structured", "statement"};
165
166 static char *pi_und1[] = {"undefined", "on", "line"};
167 static char *pi_und2[] = {"undefined", "on", "lines"};
168 static char *pi_imp1[] = {"improperly", "used", "on", "line"};
169 static char *pi_imp2[] = {"improperly", "used", "on", "lines"};
170
171 #else
172 DECL_STRINGS_2(static, pi_Endmatched, "End", "matched");
173 DECL_STRINGS_4(static, pi_Inserted, "Inserted", "keyword", "end", "matching");
174
175 DECL_STRINGS_6(static, pi_multiple,
176 "Mutiply", "defined", "label", "in", "case,", "line");
177 DECL_STRINGS_5(static, pi_structured,
178 "is", "into", "a", "structured", "statement");
179
180 DECL_STRINGS_3(static, pi_und1, "undefined", "on", "line");
181 DECL_STRINGS_3(static, pi_und2, "undefined", "on", "lines");
182 DECL_STRINGS_4(static, pi_imp1, "improperly", "used", "on", "line");
183 DECL_STRINGS_4(static, pi_imp2, "improperly", "used", "on", "lines");
184
185 #endif
186
187 static boolean
188 alldigits(const char *string)
189 {
190 for (; *string && isdigit((unsigned char)*string); string++)
191 continue;
192 return (*string == '\0');
193 }
194
195 static boolean
196 instringset(const char *member, const char **set)
197 {
198 for (; *set; set++) {
199 if (strcmp(*set, member) == 0)
200 return true;
201 }
202 return false;
203 }
204
205 static boolean
206 isdateformat(int wordc, char **wordv)
207 {
208 return (
209 (wordc == 5)
210 && (instringset(wordv[0], Days))
211 && (instringset(wordv[1], Months))
212 && (alldigits(wordv[2]))
213 && (alldigits(wordv[4])));
214 }
215
216 static boolean
217 piptr(const char *string)
218 {
219 if (*string != '-')
220 return false;
221 while (*string && *string == '-')
222 string++;
223 if (*string != '^')
224 return false;
225 string++;
226 while (*string && *string == '-')
227 string++;
228 return (*string == '\0');
229 }
230
231 Errorclass
232 pi(void)
233 {
234 char **nwordv;
235
236 nwordv = NULL;
237 if (cur_wordc < 2)
238 return (C_UNKNOWN);
239 if (strlen(cur_wordv[1]) == 1
240 && ( cur_wordv[1][0] == 'e' || cur_wordv[1][0] == 'E')
241 && piptr(cur_wordv[2])
242 ) {
243 boolean longpiptr = 0;
244
245 /*
246 * We have recognized a first pass error of the form:
247 * letter ------^---- message
248 *
249 * turn into an error message of the form:
250 *
251 * file line 'pascal errortype' letter \n |---- message
252 * or of the form:
253 * file line letter |---- message
254 * when there are strlen("(*[pi]") or more
255 * preceding '-' on the error pointer.
256 *
257 * Where the | is intended to be a down arrow, so that
258 * the pi error messages can be inserted above the
259 * line in error, instead of below. (All of the other
260 * languages put their messages before the source line,
261 * instead of after it as does pi.)
262 *
263 * where the pointer to the error has been truncated
264 * by 6 characters to account for the fact that
265 * the pointer points into a tab preceded input line.
266 */
267 language = INPI;
268 (void)substitute(cur_wordv[2], '^', '|');
269 longpiptr = position(cur_wordv[2],'|') > (6+8);
270 nwordv = wordvsplice(longpiptr ? 2 : 4, cur_wordc, cur_wordv+1);
271 nwordv[0] = strdup(currentfilename);
272 nwordv[1] = strdup(c_linenumber);
273 if (!longpiptr) {
274 nwordv[2] = Strdup("pascal errortype"); /* XXX leaked */
275 nwordv[3] = cur_wordv[1];
276 nwordv[4] = strdup("%%%\n");
277 if (strlen(nwordv[5]) > (8-2)) /* this is the pointer */
278 nwordv[5] += (8-2); /* bump over 6 characters */
279 }
280 cur_wordv = nwordv - 1; /* convert to 1 based */
281 cur_wordc += longpiptr ? 2 : 4;
282 return (C_TRUE);
283 }
284 if (cur_wordc >= 4
285 && strlen(cur_wordv[1]) == 1
286 && (*cur_wordv[1] == 'E' || *cur_wordv[1] == 'w' || *cur_wordv[1] == 'e')
287 && alldigits(cur_wordv[2])
288 && strlen(cur_wordv[3]) == 1
289 && cur_wordv[3][0] == '-'
290 ) {
291 /*
292 * Message of the form: letter linenumber - message
293 * Turn into form: filename linenumber letter - message
294 */
295 language = INPI;
296 nwordv = wordvsplice(1, cur_wordc, cur_wordv + 1);
297 nwordv[0] = strdup(currentfilename);
298 nwordv[1] = cur_wordv[2];
299 nwordv[2] = cur_wordv[1];
300 c_linenumber = cur_wordv[2];
301 cur_wordc += 1;
302 cur_wordv = nwordv - 1;
303 return (C_TRUE);
304 }
305 if (cur_wordc >= 3
306 && strlen(cur_wordv[1]) == 1
307 && (*cur_wordv[1] == 'E' || *cur_wordv[1] == 'w' || *cur_wordv[1] == 'e')
308 && strlen(cur_wordv[2]) == 1
309 && cur_wordv[2][0] == '-'
310 ) {
311 /*
312 * Message of the form: letter - message
313 *
314 * This happens only when we are traversing the tree
315 * during the second pass of pi, and discover semantic
316 * errors.
317 *
318 * We have already (presumably) saved the header message
319 * and can now construct a nulled error message for the
320 * current file.
321 *
322 * Turns into a message of the form:
323 * filename (header) letter - message
324 *
325 * First, see if it is a message referring to more than
326 * one line number. Only of the form:
327 * %s undefined on line%s
328 * %s improperly used on line%s
329 */
330 boolean undefined = 0;
331 int wordindex;
332
333 language = INPI;
334 if ((undefined = (wordvcmp(cur_wordv+2, 3, pi_und1) == 0))
335 || (undefined = (wordvcmp(cur_wordv+2, 3, pi_und2) == 0))
336 || wordvcmp(cur_wordv+2, 4, pi_imp1) == 0
337 || wordvcmp(cur_wordv+2, 4, pi_imp2) == 0
338 ) {
339 for (wordindex = undefined ? 5 : 6;
340 wordindex <= cur_wordc;
341 wordindex++) {
342 if (nwordv) {
343 free(nwordv[0]);
344 free(nwordv);
345 }
346 nwordv = wordvsplice(2, undefined ? 2 : 3, cur_wordv+1);
347 nwordv[0] = strdup(currentfilename);
348 nwordv[1] = cur_wordv[wordindex];
349 if (wordindex != cur_wordc)
350 erroradd(undefined ? 4 : 5, nwordv,
351 C_TRUE, C_UNKNOWN);
352 }
353 cur_wordc = undefined ? 4 : 5;
354 cur_wordv = nwordv - 1;
355 return (C_TRUE);
356 }
357
358 nwordv = wordvsplice(1+3, cur_wordc, cur_wordv+1);
359 nwordv[0] = strdup(currentfilename);
360 nwordv[1] = strdup(c_header[0]);
361 nwordv[2] = strdup(c_header[1]);
362 nwordv[3] = strdup(c_header[2]);
363 cur_wordv = nwordv - 1;
364 cur_wordc += 1 + 3;
365 return (C_THISFILE);
366 }
367 if (strcmp(cur_wordv[1], "...") == 0 && c_linenumber &&
368 currentfilename != default_currentfilename) {
369 /*
370 * have a continuation error message
371 * of the form: ... message
372 * Turn into form : filename linenumber message
373 */
374 language = INPI;
375 nwordv = wordvsplice(1, cur_wordc, cur_wordv+1);
376 nwordv[0] = strdup(currentfilename);
377 nwordv[1] = strdup(c_linenumber);
378 cur_wordv = nwordv - 1;
379 cur_wordc += 1;
380 return (C_TRUE);
381 }
382 if (cur_wordc == 6
383 && lastchar(cur_wordv[6]) == ':'
384 && isdateformat(5, cur_wordv + 1)
385 ) {
386 /*
387 * Have message that tells us we have changed files
388 */
389 language = INPI;
390 currentfilename = strdup(cur_wordv[6]);
391 clob_last(currentfilename, '\0');
392 return (C_SYNC);
393 }
394 if (cur_wordc == 3
395 && strcmp(cur_wordv[1], "In") == 0
396 && lastchar(cur_wordv[3]) == ':'
397 && instringset(cur_wordv[2], Piroutines)
398 ) {
399 language = INPI;
400 c_header = wordvsplice(0, cur_wordc, cur_wordv+1);
401 return (C_SYNC);
402 }
403
404 /*
405 * now, check for just the line number followed by the text
406 */
407 if (alldigits(cur_wordv[1])) {
408 language = INPI;
409 c_linenumber = cur_wordv[1];
410 return (C_IGNORE);
411 }
412
413 /*
414 * Attempt to match messages refering to a line number
415 *
416 * Multiply defined label in case, lines %d and %d
417 * Goto %s from line %d is into a structured statement
418 * End matched %s on line %d
419 * Inserted keyword end matching %s on line %d
420 */
421 multiple = structured = 0;
422 if (
423 (cur_wordc == 6 && wordvcmp(cur_wordv+1, 2, pi_Endmatched) == 0)
424 || (cur_wordc == 8 && wordvcmp(cur_wordv+1, 4, pi_Inserted) == 0)
425 || (multiple = (cur_wordc == 9 && wordvcmp(cur_wordv+1,6, pi_multiple) == 0))
426 || (structured = (cur_wordc == 10 && wordvcmp(cur_wordv+6,5, pi_structured) == 0))
427 ) {
428 language = INPI;
429 nwordv = wordvsplice(2, cur_wordc, cur_wordv+1);
430 nwordv[0] = strdup(currentfilename);
431 nwordv[1] = structured ? cur_wordv [5] : cur_wordv[cur_wordc];
432 cur_wordc += 2;
433 cur_wordv = nwordv - 1;
434 if (!multiple)
435 return (C_TRUE);
436 erroradd(cur_wordc, nwordv, C_TRUE, C_UNKNOWN);
437 nwordv = wordvsplice(0, cur_wordc, nwordv);
438 nwordv[1] = cur_wordv[cur_wordc - 2];
439 return (C_TRUE);
440 }
441 return (C_UNKNOWN);
442 }
443