man_validate.c

   1 /*      $Id: man_validate.c,v 1.142 2018/12/16 00:21:05 schwarze Exp $ */
   2 /*
   3  * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
   4  * Copyright (c) 2010, 2012-2018 Ingo Schwarze <schwarze@openbsd.org>
   5  *
   6  * Permission to use, copy, modify, and distribute this software for any
   7  * purpose with or without fee is hereby granted, provided that the above
   8  * copyright notice and this permission notice appear in all copies.
   9  *
  10  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHORS DISCLAIM ALL WARRANTIES
  11  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
  12  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHORS BE LIABLE FOR
  13  * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
  14  * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
  15  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
  16  * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
  17  */
  18 #include "config.h"
  19
  20 #include <sys/types.h>
  21
  22 #include <assert.h>
  23 #include <ctype.h>
  24 #include <errno.h>
  25 #include <limits.h>
  26 #include <stdarg.h>
  27 #include <stdio.h>
  28 #include <stdlib.h>
  29 #include <string.h>
  30 #include <time.h>
  31
  32 #include "mandoc_aux.h"
  33 #include "mandoc.h"
  34 #include "roff.h"
  35 #include "man.h"
  36 #include "libmandoc.h"
  37 #include "roff_int.h"
  38 #include "libman.h"
  39
  40 #define CHKARGS   struct roff_man *man, struct roff_node *n
  41
  42 typedef void    (*v_check)(CHKARGS);
  43
  44 static  void      check_abort(CHKARGS);
  45 static  void      check_par(CHKARGS);
  46 static  void      check_part(CHKARGS);
  47 static  void      check_root(CHKARGS);
  48 static  void      check_text(CHKARGS);
  49
  50 static  void      post_AT(CHKARGS);
  51 static  void      post_IP(CHKARGS);
  52 static  void      post_OP(CHKARGS);
  53 static  void      post_SH(CHKARGS);
  54 static  void      post_TH(CHKARGS);
  55 static  void      post_UC(CHKARGS);
  56 static  void      post_UR(CHKARGS);
  57 static  void      post_in(CHKARGS);
  58
  59 static  const v_check man_valids[MAN_MAX - MAN_TH] = {
  60         post_TH,    /* TH */
  61         post_SH,    /* SH */
  62         post_SH,    /* SS */
  63         NULL,       /* TP */
  64         NULL,       /* TQ */
  65         check_abort,/* LP */
  66         check_par,  /* PP */
  67         check_abort,/* P */
  68         post_IP,    /* IP */
  69         NULL,       /* HP */
  70         NULL,       /* SM */
  71         NULL,       /* SB */
  72         NULL,       /* BI */
  73         NULL,       /* IB */
  74         NULL,       /* BR */
  75         NULL,       /* RB */
  76         NULL,       /* R */
  77         NULL,       /* B */
  78         NULL,       /* I */
  79         NULL,       /* IR */
  80         NULL,       /* RI */
  81         NULL,       /* nf */
  82         NULL,       /* fi */
  83         NULL,       /* RE */
  84         check_part, /* RS */
  85         NULL,       /* DT */
  86         post_UC,    /* UC */
  87         NULL,       /* PD */
  88         post_AT,    /* AT */
  89         post_in,    /* in */
  90         NULL,       /* SY */
  91         NULL,       /* YS */
  92         post_OP,    /* OP */
  93         NULL,       /* EX */
  94         NULL,       /* EE */
  95         post_UR,    /* UR */
  96         NULL,       /* UE */
  97         post_UR,    /* MT */
  98         NULL,       /* ME */
  99 };
 100
 101
 102 /* Validate the subtree rooted at man->last. */
 103 void
 104 man_node_validate(struct roff_man *man)
 105 {
 106         struct roff_node *n;
 107         const v_check    *cp;
 108
 109         /*
 110          * Translate obsolete macros such that later code
 111          * does not need to look for them.
 112          */
 113
 114         n = man->last;
 115         switch (n->tok) {
 116         case MAN_LP:
 117         case MAN_P:
 118                 n->tok = MAN_PP;
 119                 break;
 120         default:
 121                 break;
 122         }
 123
 124         /*
 125          * Iterate over all children, recursing into each one
 126          * in turn, depth-first.
 127          */
 128
 129         man->last = man->last->child;
 130         while (man->last != NULL) {
 131                 man_node_validate(man);
 132                 if (man->last == n)
 133                         man->last = man->last->child;
 134                 else
 135                         man->last = man->last->next;
 136         }
 137
 138         /* Finally validate the macro itself. */
 139
 140         man->last = n;
 141         man->next = ROFF_NEXT_SIBLING;
 142         switch (n->type) {
 143         case ROFFT_TEXT:
 144                 check_text(man, n);
 145                 break;
 146         case ROFFT_ROOT:
 147                 check_root(man, n);
 148                 break;
 149         case ROFFT_COMMENT:
 150         case ROFFT_EQN:
 151         case ROFFT_TBL:
 152                 break;
 153         default:
 154                 if (n->tok < ROFF_MAX) {
 155                         roff_validate(man);
 156                         break;
 157                 }
 158                 assert(n->tok >= MAN_TH && n->tok < MAN_MAX);
 159                 cp = man_valids + (n->tok - MAN_TH);
 160                 if (*cp)
 161                         (*cp)(man, n);
 162                 if (man->last == n)
 163                         man_state(man, n);
 164                 break;
 165         }
 166 }
 167
 168 static void
 169 check_root(CHKARGS)
 170 {
 171         assert((man->flags & (MAN_BLINE | MAN_ELINE)) == 0);
 172
 173         if (n->last == NULL || n->last->type == ROFFT_COMMENT)
 174                 mandoc_msg(MANDOCERR_DOC_EMPTY, n->line, n->pos, NULL);
 175         else
 176                 man->meta.hasbody = 1;
 177
 178         if (NULL == man->meta.title) {
 179                 mandoc_msg(MANDOCERR_TH_NOTITLE, n->line, n->pos, NULL);
 180
 181                 /*
 182                  * If a title hasn't been set, do so now (by
 183                  * implication, date and section also aren't set).
 184                  */
 185
 186                 man->meta.title = mandoc_strdup("");
 187                 man->meta.msec = mandoc_strdup("");
 188                 man->meta.date = man->quick ? mandoc_strdup("") :
 189                     mandoc_normdate(man, NULL, n->line, n->pos);
 190         }
 191
 192         if (man->meta.os_e &&
 193             (man->meta.rcsids & (1 << man->meta.os_e)) == 0)
 194                 mandoc_msg(MANDOCERR_RCS_MISSING, 0, 0,
 195                     man->meta.os_e == MANDOC_OS_OPENBSD ?
 196                     "(OpenBSD)" : "(NetBSD)");
 197 }
 198
 199 static void
 200 check_abort(CHKARGS)
 201 {
 202         abort();
 203 }
 204
 205 static void
 206 check_text(CHKARGS)
 207 {
 208         char            *cp, *p;
 209
 210         if (MAN_LITERAL & man->flags)
 211                 return;
 212
 213         cp = n->string;
 214         for (p = cp; NULL != (p = strchr(p, '\t')); p++)
 215                 mandoc_msg(MANDOCERR_FI_TAB,
 216                     n->line, n->pos + (int)(p - cp), NULL);
 217 }
 218
 219 static void
 220 post_OP(CHKARGS)
 221 {
 222
 223         if (n->child == NULL)
 224                 mandoc_msg(MANDOCERR_OP_EMPTY, n->line, n->pos, "OP");
 225         else if (n->child->next != NULL && n->child->next->next != NULL) {
 226                 n = n->child->next->next;
 227                 mandoc_msg(MANDOCERR_ARG_EXCESS,
 228                     n->line, n->pos, "OP ... %s", n->string);
 229         }
 230 }
 231
 232 static void
 233 post_SH(CHKARGS)
 234 {
 235         struct roff_node        *nc;
 236
 237         if (n->type != ROFFT_BODY || (nc = n->child) == NULL)
 238                 return;
 239
 240         if (nc->tok == MAN_PP && nc->body->child != NULL) {
 241                 while (nc->body->last != NULL) {
 242                         man->next = ROFF_NEXT_CHILD;
 243                         roff_node_relink(man, nc->body->last);
 244                         man->last = n;
 245                 }
 246         }
 247
 248         if (nc->tok == MAN_PP || nc->tok == ROFF_sp || nc->tok == ROFF_br) {
 249                 mandoc_msg(MANDOCERR_PAR_SKIP, nc->line, nc->pos,
 250                     "%s after %s", roff_name[nc->tok], roff_name[n->tok]);
 251                 roff_node_delete(man, nc);
 252         }
 253
 254         /*
 255          * Trailing PP is empty, so it is deleted by check_par().
 256          * Trailing sp is significant.
 257          */
 258
 259         if ((nc = n->last) != NULL && nc->tok == ROFF_br) {
 260                 mandoc_msg(MANDOCERR_PAR_SKIP,
 261                     nc->line, nc->pos, "%s at the end of %s",
 262                     roff_name[nc->tok], roff_name[n->tok]);
 263                 roff_node_delete(man, nc);
 264         }
 265 }
 266
 267 static void
 268 post_UR(CHKARGS)
 269 {
 270         if (n->type == ROFFT_HEAD && n->child == NULL)
 271                 mandoc_msg(MANDOCERR_UR_NOHEAD, n->line, n->pos,
 272                     "%s", roff_name[n->tok]);
 273         check_part(man, n);
 274 }
 275
 276 static void
 277 check_part(CHKARGS)
 278 {
 279
 280         if (n->type == ROFFT_BODY && n->child == NULL)
 281                 mandoc_msg(MANDOCERR_BLK_EMPTY, n->line, n->pos,
 282                     "%s", roff_name[n->tok]);
 283 }
 284
 285 static void
 286 check_par(CHKARGS)
 287 {
 288
 289         switch (n->type) {
 290         case ROFFT_BLOCK:
 291                 if (n->body->child == NULL)
 292                         roff_node_delete(man, n);
 293                 break;
 294         case ROFFT_BODY:
 295                 if (n->child != NULL &&
 296                     (n->child->tok == ROFF_sp || n->child->tok == ROFF_br)) {
 297                         mandoc_msg(MANDOCERR_PAR_SKIP,
 298                             n->child->line, n->child->pos,
 299                             "%s after %s", roff_name[n->child->tok],
 300                             roff_name[n->tok]);
 301                         roff_node_delete(man, n->child);
 302                 }
 303                 if (n->child == NULL)
 304                         mandoc_msg(MANDOCERR_PAR_SKIP, n->line, n->pos,
 305                             "%s empty", roff_name[n->tok]);
 306                 break;
 307         case ROFFT_HEAD:
 308                 if (n->child != NULL)
 309                         mandoc_msg(MANDOCERR_ARG_SKIP,
 310                             n->line, n->pos, "%s %s%s",
 311                             roff_name[n->tok], n->child->string,
 312                             n->child->next != NULL ? " ..." : "");
 313                 break;
 314         default:
 315                 break;
 316         }
 317 }
 318
 319 static void
 320 post_IP(CHKARGS)
 321 {
 322
 323         switch (n->type) {
 324         case ROFFT_BLOCK:
 325                 if (n->head->child == NULL && n->body->child == NULL)
 326                         roff_node_delete(man, n);
 327                 break;
 328         case ROFFT_BODY:
 329                 if (n->parent->head->child == NULL && n->child == NULL)
 330                         mandoc_msg(MANDOCERR_PAR_SKIP, n->line, n->pos,
 331                             "%s empty", roff_name[n->tok]);
 332                 break;
 333         default:
 334                 break;
 335         }
 336 }
 337
 338 static void
 339 post_TH(CHKARGS)
 340 {
 341         struct roff_node *nb;
 342         const char      *p;
 343
 344         free(man->meta.title);
 345         free(man->meta.vol);
 346         free(man->meta.os);
 347         free(man->meta.msec);
 348         free(man->meta.date);
 349
 350         man->meta.title = man->meta.vol = man->meta.date =
 351             man->meta.msec = man->meta.os = NULL;
 352
 353         nb = n;
 354
 355         /* ->TITLE<- MSEC DATE OS VOL */
 356
 357         n = n->child;
 358         if (n && n->string) {
 359                 for (p = n->string; '\0' != *p; p++) {
 360                         /* Only warn about this once... */
 361                         if (isalpha((unsigned char)*p) &&
 362                             ! isupper((unsigned char)*p)) {
 363                                 mandoc_msg(MANDOCERR_TITLE_CASE, n->line,
 364                                     n->pos + (int)(p - n->string),
 365                                     "TH %s", n->string);
 366                                 break;
 367                         }
 368                 }
 369                 man->meta.title = mandoc_strdup(n->string);
 370         } else {
 371                 man->meta.title = mandoc_strdup("");
 372                 mandoc_msg(MANDOCERR_TH_NOTITLE, nb->line, nb->pos, "TH");
 373         }
 374
 375         /* TITLE ->MSEC<- DATE OS VOL */
 376
 377         if (n)
 378                 n = n->next;
 379         if (n && n->string)
 380                 man->meta.msec = mandoc_strdup(n->string);
 381         else {
 382                 man->meta.msec = mandoc_strdup("");
 383                 mandoc_msg(MANDOCERR_MSEC_MISSING,
 384                     nb->line, nb->pos, "TH %s", man->meta.title);
 385         }
 386
 387         /* TITLE MSEC ->DATE<- OS VOL */
 388
 389         if (n)
 390                 n = n->next;
 391         if (n && n->string && '\0' != n->string[0]) {
 392                 man->meta.date = man->quick ?
 393                     mandoc_strdup(n->string) :
 394                     mandoc_normdate(man, n->string, n->line, n->pos);
 395         } else {
 396                 man->meta.date = mandoc_strdup("");
 397                 mandoc_msg(MANDOCERR_DATE_MISSING,
 398                     n ? n->line : nb->line,
 399                     n ? n->pos : nb->pos, "TH");
 400         }
 401
 402         /* TITLE MSEC DATE ->OS<- VOL */
 403
 404         if (n && (n = n->next))
 405                 man->meta.os = mandoc_strdup(n->string);
 406         else if (man->os_s != NULL)
 407                 man->meta.os = mandoc_strdup(man->os_s);
 408         if (man->meta.os_e == MANDOC_OS_OTHER && man->meta.os != NULL) {
 409                 if (strstr(man->meta.os, "OpenBSD") != NULL)
 410                         man->meta.os_e = MANDOC_OS_OPENBSD;
 411                 else if (strstr(man->meta.os, "NetBSD") != NULL)
 412                         man->meta.os_e = MANDOC_OS_NETBSD;
 413         }
 414
 415         /* TITLE MSEC DATE OS ->VOL<- */
 416         /* If missing, use the default VOL name for MSEC. */
 417
 418         if (n && (n = n->next))
 419                 man->meta.vol = mandoc_strdup(n->string);
 420         else if ('\0' != man->meta.msec[0] &&
 421             (NULL != (p = mandoc_a2msec(man->meta.msec))))
 422                 man->meta.vol = mandoc_strdup(p);
 423
 424         if (n != NULL && (n = n->next) != NULL)
 425                 mandoc_msg(MANDOCERR_ARG_EXCESS,
 426                     n->line, n->pos, "TH ... %s", n->string);
 427
 428         /*
 429          * Remove the `TH' node after we've processed it for our
 430          * meta-data.
 431          */
 432         roff_node_delete(man, man->last);
 433 }
 434
 435 static void
 436 post_UC(CHKARGS)
 437 {
 438         static const char * const bsd_versions[] = {
 439             "3rd Berkeley Distribution",
 440             "4th Berkeley Distribution",
 441             "4.2 Berkeley Distribution",
 442             "4.3 Berkeley Distribution",
 443             "4.4 Berkeley Distribution",
 444         };
 445
 446         const char      *p, *s;
 447
 448         n = n->child;
 449
 450         if (n == NULL || n->type != ROFFT_TEXT)
 451                 p = bsd_versions[0];
 452         else {
 453                 s = n->string;
 454                 if (0 == strcmp(s, "3"))
 455                         p = bsd_versions[0];
 456                 else if (0 == strcmp(s, "4"))
 457                         p = bsd_versions[1];
 458                 else if (0 == strcmp(s, "5"))
 459                         p = bsd_versions[2];
 460                 else if (0 == strcmp(s, "6"))
 461                         p = bsd_versions[3];
 462                 else if (0 == strcmp(s, "7"))
 463                         p = bsd_versions[4];
 464                 else
 465                         p = bsd_versions[0];
 466         }
 467
 468         free(man->meta.os);
 469         man->meta.os = mandoc_strdup(p);
 470 }
 471
 472 static void
 473 post_AT(CHKARGS)
 474 {
 475         static const char * const unix_versions[] = {
 476             "7th Edition",
 477             "System III",
 478             "System V",
 479             "System V Release 2",
 480         };
 481
 482         struct roff_node *nn;
 483         const char      *p, *s;
 484
 485         n = n->child;
 486
 487         if (n == NULL || n->type != ROFFT_TEXT)
 488                 p = unix_versions[0];
 489         else {
 490                 s = n->string;
 491                 if (0 == strcmp(s, "3"))
 492                         p = unix_versions[0];
 493                 else if (0 == strcmp(s, "4"))
 494                         p = unix_versions[1];
 495                 else if (0 == strcmp(s, "5")) {
 496                         nn = n->next;
 497                         if (nn != NULL &&
 498                             nn->type == ROFFT_TEXT &&
 499                             nn->string[0] != '\0')
 500                                 p = unix_versions[3];
 501                         else
 502                                 p = unix_versions[2];
 503                 } else
 504                         p = unix_versions[0];
 505         }
 506
 507         free(man->meta.os);
 508         man->meta.os = mandoc_strdup(p);
 509 }
 510
 511 static void
 512 post_in(CHKARGS)
 513 {
 514         char    *s;
 515
 516         if (n->parent->tok != MAN_TP ||
 517             n->parent->type != ROFFT_HEAD ||
 518             n->child == NULL ||
 519             *n->child->string == '+' ||
 520             *n->child->string == '-')
 521                 return;
 522         mandoc_asprintf(&s, "+%s", n->child->string);
 523         free(n->child->string);
 524         n->child->string = s;
 525 }