man_validate.c

   1 /*      $Id: man_validate.c,v 1.144 2018/12/31 07:08:12 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,       /* RE */
  82         check_part, /* RS */
  83         NULL,       /* DT */
  84         post_UC,    /* UC */
  85         NULL,       /* PD */
  86         post_AT,    /* AT */
  87         post_in,    /* in */
  88         NULL,       /* SY */
  89         NULL,       /* YS */
  90         post_OP,    /* OP */
  91         NULL,       /* EX */
  92         NULL,       /* EE */
  93         post_UR,    /* UR */
  94         NULL,       /* UE */
  95         post_UR,    /* MT */
  96         NULL,       /* ME */
  97 };
  98
  99
 100 /* Validate the subtree rooted at man->last. */
 101 void
 102 man_validate(struct roff_man *man)
 103 {
 104         struct roff_node *n;
 105         const v_check    *cp;
 106
 107         /*
 108          * Translate obsolete macros such that later code
 109          * does not need to look for them.
 110          */
 111
 112         n = man->last;
 113         switch (n->tok) {
 114         case MAN_LP:
 115         case MAN_P:
 116                 n->tok = MAN_PP;
 117                 break;
 118         default:
 119                 break;
 120         }
 121
 122         /*
 123          * Iterate over all children, recursing into each one
 124          * in turn, depth-first.
 125          */
 126
 127         man->last = man->last->child;
 128         while (man->last != NULL) {
 129                 man_validate(man);
 130                 if (man->last == n)
 131                         man->last = man->last->child;
 132                 else
 133                         man->last = man->last->next;
 134         }
 135
 136         /* Finally validate the macro itself. */
 137
 138         man->last = n;
 139         man->next = ROFF_NEXT_SIBLING;
 140         switch (n->type) {
 141         case ROFFT_TEXT:
 142                 check_text(man, n);
 143                 break;
 144         case ROFFT_ROOT:
 145                 check_root(man, n);
 146                 break;
 147         case ROFFT_COMMENT:
 148         case ROFFT_EQN:
 149         case ROFFT_TBL:
 150                 break;
 151         default:
 152                 if (n->tok < ROFF_MAX) {
 153                         roff_validate(man);
 154                         man_state(man, n);
 155                         break;
 156                 }
 157                 assert(n->tok >= MAN_TH && n->tok < MAN_MAX);
 158                 cp = man_valids + (n->tok - MAN_TH);
 159                 if (*cp)
 160                         (*cp)(man, n);
 161                 if (man->last == n)
 162                         man_state(man, n);
 163                 break;
 164         }
 165 }
 166
 167 static void
 168 check_root(CHKARGS)
 169 {
 170         assert((man->flags & (MAN_BLINE | MAN_ELINE)) == 0);
 171
 172         if (n->last == NULL || n->last->type == ROFFT_COMMENT)
 173                 mandoc_msg(MANDOCERR_DOC_EMPTY, n->line, n->pos, NULL);
 174         else
 175                 man->meta.hasbody = 1;
 176
 177         if (NULL == man->meta.title) {
 178                 mandoc_msg(MANDOCERR_TH_NOTITLE, n->line, n->pos, NULL);
 179
 180                 /*
 181                  * If a title hasn't been set, do so now (by
 182                  * implication, date and section also aren't set).
 183                  */
 184
 185                 man->meta.title = mandoc_strdup("");
 186                 man->meta.msec = mandoc_strdup("");
 187                 man->meta.date = man->quick ? mandoc_strdup("") :
 188                     mandoc_normdate(man, NULL, n->line, n->pos);
 189         }
 190
 191         if (man->meta.os_e &&
 192             (man->meta.rcsids & (1 << man->meta.os_e)) == 0)
 193                 mandoc_msg(MANDOCERR_RCS_MISSING, 0, 0,
 194                     man->meta.os_e == MANDOC_OS_OPENBSD ?
 195                     "(OpenBSD)" : "(NetBSD)");
 196 }
 197
 198 static void
 199 check_abort(CHKARGS)
 200 {
 201         abort();
 202 }
 203
 204 static void
 205 check_text(CHKARGS)
 206 {
 207         char            *cp, *p;
 208
 209         if (man->flags & ROFF_NOFILL)
 210                 return;
 211
 212         cp = n->string;
 213         for (p = cp; NULL != (p = strchr(p, '\t')); p++)
 214                 mandoc_msg(MANDOCERR_FI_TAB,
 215                     n->line, n->pos + (int)(p - cp), NULL);
 216 }
 217
 218 static void
 219 post_OP(CHKARGS)
 220 {
 221
 222         if (n->child == NULL)
 223                 mandoc_msg(MANDOCERR_OP_EMPTY, n->line, n->pos, "OP");
 224         else if (n->child->next != NULL && n->child->next->next != NULL) {
 225                 n = n->child->next->next;
 226                 mandoc_msg(MANDOCERR_ARG_EXCESS,
 227                     n->line, n->pos, "OP ... %s", n->string);
 228         }
 229 }
 230
 231 static void
 232 post_SH(CHKARGS)
 233 {
 234         struct roff_node        *nc;
 235
 236         if (n->type != ROFFT_BODY || (nc = n->child) == NULL)
 237                 return;
 238
 239         if (nc->tok == MAN_PP && nc->body->child != NULL) {
 240                 while (nc->body->last != NULL) {
 241                         man->next = ROFF_NEXT_CHILD;
 242                         roff_node_relink(man, nc->body->last);
 243                         man->last = n;
 244                 }
 245         }
 246
 247         if (nc->tok == MAN_PP || nc->tok == ROFF_sp || nc->tok == ROFF_br) {
 248                 mandoc_msg(MANDOCERR_PAR_SKIP, nc->line, nc->pos,
 249                     "%s after %s", roff_name[nc->tok], roff_name[n->tok]);
 250                 roff_node_delete(man, nc);
 251         }
 252
 253         /*
 254          * Trailing PP is empty, so it is deleted by check_par().
 255          * Trailing sp is significant.
 256          */
 257
 258         if ((nc = n->last) != NULL && nc->tok == ROFF_br) {
 259                 mandoc_msg(MANDOCERR_PAR_SKIP,
 260                     nc->line, nc->pos, "%s at the end of %s",
 261                     roff_name[nc->tok], roff_name[n->tok]);
 262                 roff_node_delete(man, nc);
 263         }
 264 }
 265
 266 static void
 267 post_UR(CHKARGS)
 268 {
 269         if (n->type == ROFFT_HEAD && n->child == NULL)
 270                 mandoc_msg(MANDOCERR_UR_NOHEAD, n->line, n->pos,
 271                     "%s", roff_name[n->tok]);
 272         check_part(man, n);
 273 }
 274
 275 static void
 276 check_part(CHKARGS)
 277 {
 278
 279         if (n->type == ROFFT_BODY && n->child == NULL)
 280                 mandoc_msg(MANDOCERR_BLK_EMPTY, n->line, n->pos,
 281                     "%s", roff_name[n->tok]);
 282 }
 283
 284 static void
 285 check_par(CHKARGS)
 286 {
 287
 288         switch (n->type) {
 289         case ROFFT_BLOCK:
 290                 if (n->body->child == NULL)
 291                         roff_node_delete(man, n);
 292                 break;
 293         case ROFFT_BODY:
 294                 if (n->child != NULL &&
 295                     (n->child->tok == ROFF_sp || n->child->tok == ROFF_br)) {
 296                         mandoc_msg(MANDOCERR_PAR_SKIP,
 297                             n->child->line, n->child->pos,
 298                             "%s after %s", roff_name[n->child->tok],
 299                             roff_name[n->tok]);
 300                         roff_node_delete(man, n->child);
 301                 }
 302                 if (n->child == NULL)
 303                         mandoc_msg(MANDOCERR_PAR_SKIP, n->line, n->pos,
 304                             "%s empty", roff_name[n->tok]);
 305                 break;
 306         case ROFFT_HEAD:
 307                 if (n->child != NULL)
 308                         mandoc_msg(MANDOCERR_ARG_SKIP,
 309                             n->line, n->pos, "%s %s%s",
 310                             roff_name[n->tok], n->child->string,
 311                             n->child->next != NULL ? " ..." : "");
 312                 break;
 313         default:
 314                 break;
 315         }
 316 }
 317
 318 static void
 319 post_IP(CHKARGS)
 320 {
 321
 322         switch (n->type) {
 323         case ROFFT_BLOCK:
 324                 if (n->head->child == NULL && n->body->child == NULL)
 325                         roff_node_delete(man, n);
 326                 break;
 327         case ROFFT_BODY:
 328                 if (n->parent->head->child == NULL && n->child == NULL)
 329                         mandoc_msg(MANDOCERR_PAR_SKIP, n->line, n->pos,
 330                             "%s empty", roff_name[n->tok]);
 331                 break;
 332         default:
 333                 break;
 334         }
 335 }
 336
 337 static void
 338 post_TH(CHKARGS)
 339 {
 340         struct roff_node *nb;
 341         const char      *p;
 342
 343         free(man->meta.title);
 344         free(man->meta.vol);
 345         free(man->meta.os);
 346         free(man->meta.msec);
 347         free(man->meta.date);
 348
 349         man->meta.title = man->meta.vol = man->meta.date =
 350             man->meta.msec = man->meta.os = NULL;
 351
 352         nb = n;
 353
 354         /* ->TITLE<- MSEC DATE OS VOL */
 355
 356         n = n->child;
 357         if (n && n->string) {
 358                 for (p = n->string; '\0' != *p; p++) {
 359                         /* Only warn about this once... */
 360                         if (isalpha((unsigned char)*p) &&
 361                             ! isupper((unsigned char)*p)) {
 362                                 mandoc_msg(MANDOCERR_TITLE_CASE, n->line,
 363                                     n->pos + (int)(p - n->string),
 364                                     "TH %s", n->string);
 365                                 break;
 366                         }
 367                 }
 368                 man->meta.title = mandoc_strdup(n->string);
 369         } else {
 370                 man->meta.title = mandoc_strdup("");
 371                 mandoc_msg(MANDOCERR_TH_NOTITLE, nb->line, nb->pos, "TH");
 372         }
 373
 374         /* TITLE ->MSEC<- DATE OS VOL */
 375
 376         if (n)
 377                 n = n->next;
 378         if (n && n->string)
 379                 man->meta.msec = mandoc_strdup(n->string);
 380         else {
 381                 man->meta.msec = mandoc_strdup("");
 382                 mandoc_msg(MANDOCERR_MSEC_MISSING,
 383                     nb->line, nb->pos, "TH %s", man->meta.title);
 384         }
 385
 386         /* TITLE MSEC ->DATE<- OS VOL */
 387
 388         if (n)
 389                 n = n->next;
 390         if (n && n->string && '\0' != n->string[0]) {
 391                 man->meta.date = man->quick ?
 392                     mandoc_strdup(n->string) :
 393                     mandoc_normdate(man, n->string, n->line, n->pos);
 394         } else {
 395                 man->meta.date = mandoc_strdup("");
 396                 mandoc_msg(MANDOCERR_DATE_MISSING,
 397                     n ? n->line : nb->line,
 398                     n ? n->pos : nb->pos, "TH");
 399         }
 400
 401         /* TITLE MSEC DATE ->OS<- VOL */
 402
 403         if (n && (n = n->next))
 404                 man->meta.os = mandoc_strdup(n->string);
 405         else if (man->os_s != NULL)
 406                 man->meta.os = mandoc_strdup(man->os_s);
 407         if (man->meta.os_e == MANDOC_OS_OTHER && man->meta.os != NULL) {
 408                 if (strstr(man->meta.os, "OpenBSD") != NULL)
 409                         man->meta.os_e = MANDOC_OS_OPENBSD;
 410                 else if (strstr(man->meta.os, "NetBSD") != NULL)
 411                         man->meta.os_e = MANDOC_OS_NETBSD;
 412         }
 413
 414         /* TITLE MSEC DATE OS ->VOL<- */
 415         /* If missing, use the default VOL name for MSEC. */
 416
 417         if (n && (n = n->next))
 418                 man->meta.vol = mandoc_strdup(n->string);
 419         else if ('\0' != man->meta.msec[0] &&
 420             (NULL != (p = mandoc_a2msec(man->meta.msec))))
 421                 man->meta.vol = mandoc_strdup(p);
 422
 423         if (n != NULL && (n = n->next) != NULL)
 424                 mandoc_msg(MANDOCERR_ARG_EXCESS,
 425                     n->line, n->pos, "TH ... %s", n->string);
 426
 427         /*
 428          * Remove the `TH' node after we've processed it for our
 429          * meta-data.
 430          */
 431         roff_node_delete(man, man->last);
 432 }
 433
 434 static void
 435 post_UC(CHKARGS)
 436 {
 437         static const char * const bsd_versions[] = {
 438             "3rd Berkeley Distribution",
 439             "4th Berkeley Distribution",
 440             "4.2 Berkeley Distribution",
 441             "4.3 Berkeley Distribution",
 442             "4.4 Berkeley Distribution",
 443         };
 444
 445         const char      *p, *s;
 446
 447         n = n->child;
 448
 449         if (n == NULL || n->type != ROFFT_TEXT)
 450                 p = bsd_versions[0];
 451         else {
 452                 s = n->string;
 453                 if (0 == strcmp(s, "3"))
 454                         p = bsd_versions[0];
 455                 else if (0 == strcmp(s, "4"))
 456                         p = bsd_versions[1];
 457                 else if (0 == strcmp(s, "5"))
 458                         p = bsd_versions[2];
 459                 else if (0 == strcmp(s, "6"))
 460                         p = bsd_versions[3];
 461                 else if (0 == strcmp(s, "7"))
 462                         p = bsd_versions[4];
 463                 else
 464                         p = bsd_versions[0];
 465         }
 466
 467         free(man->meta.os);
 468         man->meta.os = mandoc_strdup(p);
 469 }
 470
 471 static void
 472 post_AT(CHKARGS)
 473 {
 474         static const char * const unix_versions[] = {
 475             "7th Edition",
 476             "System III",
 477             "System V",
 478             "System V Release 2",
 479         };
 480
 481         struct roff_node *nn;
 482         const char      *p, *s;
 483
 484         n = n->child;
 485
 486         if (n == NULL || n->type != ROFFT_TEXT)
 487                 p = unix_versions[0];
 488         else {
 489                 s = n->string;
 490                 if (0 == strcmp(s, "3"))
 491                         p = unix_versions[0];
 492                 else if (0 == strcmp(s, "4"))
 493                         p = unix_versions[1];
 494                 else if (0 == strcmp(s, "5")) {
 495                         nn = n->next;
 496                         if (nn != NULL &&
 497                             nn->type == ROFFT_TEXT &&
 498                             nn->string[0] != '\0')
 499                                 p = unix_versions[3];
 500                         else
 501                                 p = unix_versions[2];
 502                 } else
 503                         p = unix_versions[0];
 504         }
 505
 506         free(man->meta.os);
 507         man->meta.os = mandoc_strdup(p);
 508 }
 509
 510 static void
 511 post_in(CHKARGS)
 512 {
 513         char    *s;
 514
 515         if (n->parent->tok != MAN_TP ||
 516             n->parent->type != ROFFT_HEAD ||
 517             n->child == NULL ||
 518             *n->child->string == '+' ||
 519             *n->child->string == '-')
 520                 return;
 521         mandoc_asprintf(&s, "+%s", n->child->string);
 522         free(n->child->string);
 523         n->child->string = s;
 524 }