Annotation of src/usr.bin/checknr/checknr.c, Revision 1.11
1.11 ! deraadt 1: /* $OpenBSD: checknr.c,v 1.10 2003/06/10 22:20:45 deraadt Exp $ */
1.1 deraadt 2: /* $NetBSD: checknr.c,v 1.4 1995/03/26 04:10:19 glass Exp $ */
3:
4: /*
5: * Copyright (c) 1980, 1993
6: * The Regents of the University of California. All rights reserved.
7: *
8: * Redistribution and use in source and binary forms, with or without
9: * modification, are permitted provided that the following conditions
10: * are met:
11: * 1. Redistributions of source code must retain the above copyright
12: * notice, this list of conditions and the following disclaimer.
13: * 2. Redistributions in binary form must reproduce the above copyright
14: * notice, this list of conditions and the following disclaimer in the
15: * documentation and/or other materials provided with the distribution.
1.9 millert 16: * 3. Neither the name of the University nor the names of its contributors
1.1 deraadt 17: * may be used to endorse or promote products derived from this software
18: * without specific prior written permission.
19: *
20: * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
21: * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
22: * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
23: * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
24: * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
25: * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
26: * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
27: * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
28: * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
29: * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
30: * SUCH DAMAGE.
31: */
32:
33: #ifndef lint
34: static char copyright[] =
35: "@(#) Copyright (c) 1980, 1993\n\
36: The Regents of the University of California. All rights reserved.\n";
37: #endif /* not lint */
38:
39: #ifndef lint
40: #if 0
41: static char sccsid[] = "@(#)checknr.c 8.1 (Berkeley) 6/6/93";
42: #else
1.11 ! deraadt 43: static char rcsid[] = "$OpenBSD: checknr.c,v 1.10 2003/06/10 22:20:45 deraadt Exp $";
1.1 deraadt 44: #endif
45: #endif /* not lint */
46:
47: /*
48: * checknr: check an nroff/troff input file for matching macro calls.
49: * we also attempt to match size and font changes, but only the embedded
50: * kind. These must end in \s0 and \fP resp. Maybe more sophistication
51: * later but for now think of these restrictions as contributions to
52: * structured typesetting.
53: */
54: #include <stdio.h>
55: #include <string.h>
1.11 ! deraadt 56: #include <stdlib.h>
! 57: #include <unistd.h>
1.1 deraadt 58: #include <ctype.h>
59:
60: #define MAXSTK 100 /* Stack size */
61: #define MAXBR 100 /* Max number of bracket pairs known */
62: #define MAXCMDS 500 /* Max number of commands known */
63:
64: /*
65: * The stack on which we remember what we've seen so far.
66: */
67: struct stkstr {
68: int opno; /* number of opening bracket */
69: int pl; /* '+', '-', ' ' for \s, 1 for \f, 0 for .ft */
70: int parm; /* parm to size, font, etc */
71: int lno; /* line number the thing came in in */
72: } stk[MAXSTK];
73: int stktop;
74:
1.5 deraadt 75: void usage(void);
76: void addmac(char *);
77: void process(FILE *);
78: void pe(int);
79: int eq(char *, char *);
80: void complain(int);
81: void prop(int);
82: void chkcmd(char *, char *);
83: void addcmd(char *);
84: void nomatch(char *);
85: void checkknown(char *);
86: int binsrch(char *);
87:
1.1 deraadt 88: /*
89: * The kinds of opening and closing brackets.
90: */
91: struct brstr {
92: char *opbr;
93: char *clbr;
94: } br[MAXBR] = {
95: /* A few bare bones troff commands */
96: #define SZ 0
1.5 deraadt 97: { "sz", "sz" }, /* also \s */
1.1 deraadt 98: #define FT 1
1.5 deraadt 99: { "ft", "ft" }, /* also \f */
1.1 deraadt 100: /* the -mm package */
1.5 deraadt 101: { "AL", "LE" },
102: { "AS", "AE" },
103: { "BL", "LE" },
104: { "BS", "BE" },
105: { "DF", "DE" },
106: { "DL", "LE" },
107: { "DS", "DE" },
108: { "FS", "FE" },
109: { "ML", "LE" },
110: { "NS", "NE" },
111: { "RL", "LE" },
112: { "VL", "LE" },
1.1 deraadt 113: /* the -ms package */
1.5 deraadt 114: { "AB", "AE" },
115: { "BD", "DE" },
116: { "CD", "DE" },
117: { "DS", "DE" },
118: { "FS", "FE" },
119: { "ID", "DE" },
120: { "KF", "KE" },
121: { "KS", "KE" },
122: { "LD", "DE" },
123: { "LG", "NL" },
124: { "QS", "QE" },
125: { "RS", "RE" },
126: { "SM", "NL" },
127: { "XA", "XE" },
128: { "XS", "XE" },
1.1 deraadt 129: /* The -me package */
1.5 deraadt 130: { "(b", ")b" },
131: { "(c", ")c" },
132: { "(d", ")d" },
133: { "(f", ")f" },
134: { "(l", ")l" },
135: { "(q", ")q" },
136: { "(x", ")x" },
137: { "(z", ")z" },
1.1 deraadt 138: /* Things needed by preprocessors */
1.5 deraadt 139: { "EQ", "EN" },
140: { "TS", "TE" },
1.1 deraadt 141: /* Refer */
1.5 deraadt 142: { "[", "]" },
143: { 0, },
1.1 deraadt 144: };
145:
146: /*
147: * All commands known to nroff, plus macro packages.
148: * Used so we can complain about unrecognized commands.
149: */
150: char *knowncmds[MAXCMDS] = {
151: "$c", "$f", "$h", "$p", "$s", "(b", "(c", "(d", "(f", "(l", "(q", "(t",
152: "(x", "(z", ")b", ")c", ")d", ")f", ")l", ")q", ")t", ")x", ")z", "++",
153: "+c", "1C", "1c", "2C", "2c", "@(", "@)", "@C", "@D", "@F", "@I", "@M",
154: "@c", "@e", "@f", "@h", "@m", "@n", "@o", "@p", "@r", "@t", "@z", "AB",
155: "AE", "AF", "AI", "AL", "AM", "AS", "AT", "AU", "AX", "B", "B1", "B2",
156: "BD", "BE", "BG", "BL", "BS", "BT", "BX", "C1", "C2", "CD", "CM", "CT",
157: "D", "DA", "DE", "DF", "DL", "DS", "DT", "EC", "EF", "EG", "EH", "EM",
158: "EN", "EQ", "EX", "FA", "FD", "FE", "FG", "FJ", "FK", "FL", "FN", "FO",
159: "FQ", "FS", "FV", "FX", "H", "HC", "HD", "HM", "HO", "HU", "I", "ID",
160: "IE", "IH", "IM", "IP", "IX", "IZ", "KD", "KE", "KF", "KQ", "KS", "LB",
161: "LC", "LD", "LE", "LG", "LI", "LP", "MC", "ME", "MF", "MH", "ML", "MR",
162: "MT", "ND", "NE", "NH", "NL", "NP", "NS", "OF", "OH", "OK", "OP", "P",
163: "P1", "PF", "PH", "PP", "PT", "PX", "PY", "QE", "QP", "QS", "R", "RA",
164: "RC", "RE", "RL", "RP", "RQ", "RS", "RT", "S", "S0", "S2", "S3", "SA",
165: "SG", "SH", "SK", "SM", "SP", "SY", "T&", "TA", "TB", "TC", "TD", "TE",
166: "TH", "TL", "TM", "TP", "TQ", "TR", "TS", "TX", "UL", "US", "UX", "VL",
167: "WC", "WH", "XA", "XD", "XE", "XF", "XK", "XP", "XS", "[", "[-", "[0",
168: "[1", "[2", "[3", "[4", "[5", "[<", "[>", "[]", "]", "]-", "]<", "]>",
169: "][", "ab", "ac", "ad", "af", "am", "ar", "as", "b", "ba", "bc", "bd",
170: "bi", "bl", "bp", "br", "bx", "c.", "c2", "cc", "ce", "cf", "ch", "cs",
171: "ct", "cu", "da", "de", "di", "dl", "dn", "ds", "dt", "dw", "dy", "ec",
172: "ef", "eh", "el", "em", "eo", "ep", "ev", "ex", "fc", "fi", "fl", "fo",
173: "fp", "ft", "fz", "hc", "he", "hl", "hp", "ht", "hw", "hx", "hy", "i",
174: "ie", "if", "ig", "in", "ip", "it", "ix", "lc", "lg", "li", "ll", "ln",
175: "lo", "lp", "ls", "lt", "m1", "m2", "m3", "m4", "mc", "mk", "mo", "n1",
176: "n2", "na", "ne", "nf", "nh", "nl", "nm", "nn", "np", "nr", "ns", "nx",
177: "of", "oh", "os", "pa", "pc", "pi", "pl", "pm", "pn", "po", "pp", "ps",
178: "q", "r", "rb", "rd", "re", "rm", "rn", "ro", "rr", "rs", "rt", "sb",
179: "sc", "sh", "sk", "so", "sp", "ss", "st", "sv", "sz", "ta", "tc", "th",
180: "ti", "tl", "tm", "tp", "tr", "u", "uf", "uh", "ul", "vs", "wh", "xp",
181: "yr", 0
182: };
183:
184: int lineno; /* current line number in input file */
185: char line[256]; /* the current line */
186: char *cfilename; /* name of current file */
187: int nfiles; /* number of files to process */
188: int fflag; /* -f: ignore \f */
189: int sflag; /* -s: ignore \s */
190: int ncmds; /* size of knowncmds */
191: int slot; /* slot in knowncmds found by binsrch */
192:
1.5 deraadt 193: int
1.10 deraadt 194: main(int argc, char *argv[])
1.1 deraadt 195: {
196: FILE *f;
197: int i;
198: char *cp;
199: char b1[4];
200:
201: /* Figure out how many known commands there are */
202: while (knowncmds[ncmds])
203: ncmds++;
204: while (argc > 1 && argv[1][0] == '-') {
205: switch(argv[1][1]) {
206:
207: /* -a: add pairs of macros */
208: case 'a':
209: i = strlen(argv[1]) - 2;
210: if (i % 6 != 0)
211: usage();
212: /* look for empty macro slots */
213: for (i=0; br[i].opbr; i++)
214: ;
215: for (cp=argv[1]+3; cp[-1]; cp += 6) {
216: br[i].opbr = malloc(3);
217: strncpy(br[i].opbr, cp, 2);
218: br[i].clbr = malloc(3);
219: strncpy(br[i].clbr, cp+3, 2);
220: addmac(br[i].opbr); /* knows pairs are also known cmds */
221: addmac(br[i].clbr);
222: i++;
223: }
224: break;
225:
226: /* -c: add known commands */
227: case 'c':
228: i = strlen(argv[1]) - 2;
229: if (i % 3 != 0)
230: usage();
231: for (cp=argv[1]+3; cp[-1]; cp += 3) {
232: if (cp[2] && cp[2] != '.')
233: usage();
234: strncpy(b1, cp, 2);
235: addmac(b1);
236: }
237: break;
238:
239: /* -f: ignore font changes */
240: case 'f':
241: fflag = 1;
242: break;
243:
244: /* -s: ignore size changes */
245: case 's':
246: sflag = 1;
247: break;
248: default:
249: usage();
250: }
251: argc--; argv++;
252: }
253:
254: nfiles = argc - 1;
255:
256: if (nfiles > 0) {
257: for (i=1; i<argc; i++) {
258: cfilename = argv[i];
259: f = fopen(cfilename, "r");
260: if (f == NULL)
261: perror(cfilename);
262: else
263: process(f);
264: }
265: } else {
266: cfilename = "stdin";
267: process(stdin);
268: }
269: exit(0);
270: }
271:
1.5 deraadt 272: void
1.10 deraadt 273: usage(void)
1.1 deraadt 274: {
1.3 aaron 275: (void)fprintf(stderr,
1.4 aaron 276: "usage: checknr [-fs] [-a.x1.y1.x2.y2. ... .xn.yn] "
1.3 aaron 277: "[-c.x1.x2.x3. ... .xn] [file]\n");
1.1 deraadt 278: exit(1);
279: }
280:
1.5 deraadt 281: void
1.10 deraadt 282: process(FILE *f)
1.1 deraadt 283: {
1.7 mpech 284: int i, n;
1.1 deraadt 285: char mac[5]; /* The current macro or nroff command */
286: int pl;
287:
288: stktop = -1;
289: for (lineno = 1; fgets(line, sizeof line, f); lineno++) {
290: if (line[0] == '.') {
291: /*
292: * find and isolate the macro/command name.
293: */
294: strncpy(mac, line+1, 4);
295: if (isspace(mac[0])) {
296: pe(lineno);
297: printf("Empty command\n");
298: } else if (isspace(mac[1])) {
299: mac[1] = 0;
300: } else if (isspace(mac[2])) {
301: mac[2] = 0;
302: } else if (mac[0] != '\\' || mac[1] != '\"') {
303: pe(lineno);
304: printf("Command too long\n");
305: }
306:
307: /*
308: * Is it a known command?
309: */
310: checkknown(mac);
311:
312: /*
313: * Should we add it?
314: */
315: if (eq(mac, "de"))
316: addcmd(line);
317:
318: chkcmd(line, mac);
319: }
320:
321: /*
322: * At this point we process the line looking
323: * for \s and \f.
324: */
325: for (i=0; line[i]; i++)
326: if (line[i]=='\\' && (i==0 || line[i-1]!='\\')) {
327: if (!sflag && line[++i]=='s') {
328: pl = line[++i];
329: if (isdigit(pl)) {
330: n = pl - '0';
331: pl = ' ';
332: } else
333: n = 0;
334: while (isdigit(line[++i]))
335: n = 10 * n + line[i] - '0';
336: i--;
337: if (n == 0) {
338: if (stk[stktop].opno == SZ) {
339: stktop--;
340: } else {
341: pe(lineno);
342: printf("unmatched \\s0\n");
343: }
344: } else {
345: stk[++stktop].opno = SZ;
346: stk[stktop].pl = pl;
347: stk[stktop].parm = n;
348: stk[stktop].lno = lineno;
349: }
350: } else if (!fflag && line[i]=='f') {
351: n = line[++i];
352: if (n == 'P') {
353: if (stk[stktop].opno == FT) {
354: stktop--;
355: } else {
356: pe(lineno);
357: printf("unmatched \\fP\n");
358: }
359: } else {
360: stk[++stktop].opno = FT;
361: stk[stktop].pl = 1;
362: stk[stktop].parm = n;
363: stk[stktop].lno = lineno;
364: }
365: }
366: }
367: }
368: /*
369: * We've hit the end and look at all this stuff that hasn't been
370: * matched yet! Complain, complain.
371: */
372: for (i=stktop; i>=0; i--) {
373: complain(i);
374: }
375: }
376:
1.5 deraadt 377: void
1.10 deraadt 378: complain(int i)
1.1 deraadt 379: {
380: pe(stk[i].lno);
381: printf("Unmatched ");
382: prop(i);
383: printf("\n");
384: }
385:
1.5 deraadt 386: void
1.10 deraadt 387: prop(int i)
1.1 deraadt 388: {
389: if (stk[i].pl == 0)
390: printf(".%s", br[stk[i].opno].opbr);
391: else switch(stk[i].opno) {
392: case SZ:
393: printf("\\s%c%d", stk[i].pl, stk[i].parm);
394: break;
395: case FT:
396: printf("\\f%c", stk[i].parm);
397: break;
398: default:
399: printf("Bug: stk[%d].opno = %d = .%s, .%s",
400: i, stk[i].opno, br[stk[i].opno].opbr, br[stk[i].opno].clbr);
401: }
402: }
403:
1.5 deraadt 404: void
1.10 deraadt 405: chkcmd(char *line, char *mac)
1.1 deraadt 406: {
1.7 mpech 407: int i;
1.1 deraadt 408:
409: /*
410: * Check to see if it matches top of stack.
411: */
412: if (stktop >= 0 && eq(mac, br[stk[stktop].opno].clbr))
413: stktop--; /* OK. Pop & forget */
414: else {
415: /* No. Maybe it's an opener */
416: for (i=0; br[i].opbr; i++) {
417: if (eq(mac, br[i].opbr)) {
418: /* Found. Push it. */
419: stktop++;
420: stk[stktop].opno = i;
421: stk[stktop].pl = 0;
422: stk[stktop].parm = 0;
423: stk[stktop].lno = lineno;
424: break;
425: }
426: /*
427: * Maybe it's an unmatched closer.
428: * NOTE: this depends on the fact
429: * that none of the closers can be
430: * openers too.
431: */
432: if (eq(mac, br[i].clbr)) {
433: nomatch(mac);
434: break;
435: }
436: }
437: }
438: }
439:
1.5 deraadt 440: void
1.10 deraadt 441: nomatch(char *mac)
1.1 deraadt 442: {
1.7 mpech 443: int i, j;
1.1 deraadt 444:
445: /*
446: * Look for a match further down on stack
447: * If we find one, it suggests that the stuff in
448: * between is supposed to match itself.
449: */
450: for (j=stktop; j>=0; j--)
451: if (eq(mac,br[stk[j].opno].clbr)) {
452: /* Found. Make a good diagnostic. */
453: if (j == stktop-2) {
454: /*
455: * Check for special case \fx..\fR and don't
456: * complain.
457: */
458: if (stk[j+1].opno==FT && stk[j+1].parm!='R'
459: && stk[j+2].opno==FT && stk[j+2].parm=='R') {
460: stktop = j -1;
461: return;
462: }
463: /*
464: * We have two unmatched frobs. Chances are
465: * they were intended to match, so we mention
466: * them together.
467: */
468: pe(stk[j+1].lno);
469: prop(j+1);
470: printf(" does not match %d: ", stk[j+2].lno);
471: prop(j+2);
472: printf("\n");
473: } else for (i=j+1; i <= stktop; i++) {
474: complain(i);
475: }
476: stktop = j-1;
477: return;
478: }
479: /* Didn't find one. Throw this away. */
480: pe(lineno);
481: printf("Unmatched .%s\n", mac);
482: }
483:
484: /* eq: are two strings equal? */
1.5 deraadt 485: int
1.10 deraadt 486: eq(char *s1, char *s2)
1.1 deraadt 487: {
488: return (strcmp(s1, s2) == 0);
489: }
490:
491: /* print the first part of an error message, given the line number */
1.5 deraadt 492: void
1.10 deraadt 493: pe(int lineno)
1.1 deraadt 494: {
495: if (nfiles > 1)
496: printf("%s: ", cfilename);
497: printf("%d: ", lineno);
498: }
499:
1.5 deraadt 500: void
1.10 deraadt 501: checkknown(char *mac)
1.1 deraadt 502: {
503:
504: if (eq(mac, "."))
505: return;
506: if (binsrch(mac) >= 0)
507: return;
508: if (mac[0] == '\\' && mac[1] == '"') /* comments */
509: return;
510:
511: pe(lineno);
512: printf("Unknown command: .%s\n", mac);
513: }
514:
515: /*
516: * We have a .de xx line in "line". Add xx to the list of known commands.
517: */
1.5 deraadt 518: void
1.10 deraadt 519: addcmd(char *line)
1.1 deraadt 520: {
521: char *mac;
522:
523: /* grab the macro being defined */
524: mac = line+4;
525: while (isspace(*mac))
526: mac++;
527: if (*mac == 0) {
528: pe(lineno);
529: printf("illegal define: %s\n", line);
530: return;
531: }
532: mac[2] = 0;
533: if (isspace(mac[1]) || mac[1] == '\\')
534: mac[1] = 0;
535: if (ncmds >= MAXCMDS) {
536: printf("Only %d known commands allowed\n", MAXCMDS);
537: exit(1);
538: }
539: addmac(mac);
540: }
541:
542: /*
543: * Add mac to the list. We should really have some kind of tree
544: * structure here but this is a quick-and-dirty job and I just don't
545: * have time to mess with it. (I wonder if this will come back to haunt
546: * me someday?) Anyway, I claim that .de is fairly rare in user
547: * nroff programs, and the register loop below is pretty fast.
548: */
1.5 deraadt 549: void
1.10 deraadt 550: addmac(char *mac)
1.1 deraadt 551: {
1.7 mpech 552: char **src, **dest, **loc;
1.1 deraadt 553:
554: if (binsrch(mac) >= 0){ /* it's OK to redefine something */
555: #ifdef DEBUG
556: printf("binsrch(%s) -> already in table\n", mac);
1.6 heko 557: #endif /* DEBUG */
1.1 deraadt 558: return;
559: }
560: /* binsrch sets slot as a side effect */
561: #ifdef DEBUG
562: printf("binsrch(%s) -> %d\n", mac, slot);
563: #endif
564: loc = &knowncmds[slot];
565: src = &knowncmds[ncmds-1];
566: dest = src+1;
567: while (dest > loc)
568: *dest-- = *src--;
1.8 deraadt 569: *loc = strdup(mac);
1.1 deraadt 570: ncmds++;
571: #ifdef DEBUG
572: printf("after: %s %s %s %s %s, %d cmds\n", knowncmds[slot-2], knowncmds[slot-1], knowncmds[slot], knowncmds[slot+1], knowncmds[slot+2], ncmds);
573: #endif
574: }
575:
576: /*
577: * Do a binary search in knowncmds for mac.
578: * If found, return the index. If not, return -1.
579: */
1.5 deraadt 580: int
1.10 deraadt 581: binsrch(char *mac)
1.1 deraadt 582: {
1.7 mpech 583: char *p; /* pointer to current cmd in list */
584: int d; /* difference if any */
585: int mid; /* mid point in binary search */
586: int top, bot; /* boundaries of bin search, inclusive */
1.1 deraadt 587:
588: top = ncmds-1;
589: bot = 0;
590: while (top >= bot) {
591: mid = (top+bot)/2;
592: p = knowncmds[mid];
593: d = p[0] - mac[0];
594: if (d == 0)
595: d = p[1] - mac[1];
596: if (d == 0)
597: return mid;
598: if (d < 0)
599: bot = mid + 1;
600: else
601: top = mid - 1;
602: }
603: slot = bot; /* place it would have gone */
604: return -1;
605: }