man_validate.c

   1 /*      $OpenBSD$ */
   2 /*
   3  * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
   4  * Copyright (c) 2010, 2012-2017 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 <stdlib.h>
  28 #include <string.h>
  29 #include <time.h>
  30
  31 #include "mandoc_aux.h"
  32 #include "mandoc.h"
  33 #include "roff.h"
  34 #include "man.h"
  35 #include "libmandoc.h"
  36 #include "roff_int.h"
  37 #include "libman.h"
  38
  39 #define CHKARGS   struct roff_man *man, struct roff_node *n
  40
  41 typedef void    (*v_check)(CHKARGS);
  42
  43 static  void      check_par(CHKARGS);
  44 static  void      check_part(CHKARGS);
  45 static  void      check_root(CHKARGS);
  46 static  void      check_text(CHKARGS);
  47
  48 static  void      post_AT(CHKARGS);
  49 static  void      post_IP(CHKARGS);
  50 static  void      post_OP(CHKARGS);
  51 static  void      post_TH(CHKARGS);
  52 static  void      post_UC(CHKARGS);
  53 static  void      post_UR(CHKARGS);
  54 static  void      post_in(CHKARGS);
  55 static  void      post_vs(CHKARGS);
  56
  57 static  const v_check __man_valids[MAN_MAX - MAN_TH] = {
  58         post_TH,    /* TH */
  59         NULL,       /* SH */
  60         NULL,       /* SS */
  61         NULL,       /* TP */
  62         check_par,  /* LP */
  63         check_par,  /* PP */
  64         check_par,  /* P */
  65         post_IP,    /* IP */
  66         NULL,       /* HP */
  67         NULL,       /* SM */
  68         NULL,       /* SB */
  69         NULL,       /* BI */
  70         NULL,       /* IB */
  71         NULL,       /* BR */
  72         NULL,       /* RB */
  73         NULL,       /* R */
  74         NULL,       /* B */
  75         NULL,       /* I */
  76         NULL,       /* IR */
  77         NULL,       /* RI */
  78         NULL,       /* nf */
  79         NULL,       /* fi */
  80         NULL,       /* RE */
  81         check_part, /* RS */
  82         NULL,       /* DT */
  83         post_UC,    /* UC */
  84         NULL,       /* PD */
  85         post_AT,    /* AT */
  86         post_in,    /* in */
  87         post_OP,    /* OP */
  88         NULL,       /* EX */
  89         NULL,       /* EE */
  90         post_UR,    /* UR */
  91         NULL,       /* UE */
  92         post_UR,    /* MT */
  93         NULL,       /* ME */
  94 };
  95 static  const v_check *man_valids = __man_valids - MAN_TH;
  96
  97
  98 void
  99 man_node_validate(struct roff_man *man)
 100 {
 101         struct roff_node *n;
 102         const v_check    *cp;
 103
 104         n = man->last;
 105         man->last = man->last->child;
 106         while (man->last != NULL) {
 107                 man_node_validate(man);
 108                 if (man->last == n)
 109                         man->last = man->last->child;
 110                 else
 111                         man->last = man->last->next;
 112         }
 113
 114         man->last = n;
 115         man->next = ROFF_NEXT_SIBLING;
 116         switch (n->type) {
 117         case ROFFT_TEXT:
 118                 check_text(man, n);
 119                 break;
 120         case ROFFT_ROOT:
 121                 check_root(man, n);
 122                 break;
 123         case ROFFT_EQN:
 124         case ROFFT_TBL:
 125                 break;
 126         default:
 127                 if (n->tok < ROFF_MAX) {
 128                         switch (n->tok) {
 129                         case ROFF_br:
 130                         case ROFF_sp:
 131                                 post_vs(man, n);
 132                                 break;
 133                         default:
 134                                 roff_validate(man);
 135                                 break;
 136                         }
 137                         break;
 138                 }
 139                 assert(n->tok >= MAN_TH && n->tok < MAN_MAX);
 140                 cp = man_valids + n->tok;
 141                 if (*cp)
 142                         (*cp)(man, n);
 143                 if (man->last == n)
 144                         man_state(man, n);
 145                 break;
 146         }
 147 }
 148
 149 static void
 150 check_root(CHKARGS)
 151 {
 152
 153         assert((man->flags & (MAN_BLINE | MAN_ELINE)) == 0);
 154
 155         if (NULL == man->first->child)
 156                 mandoc_msg(MANDOCERR_DOC_EMPTY, man->parse,
 157                     n->line, n->pos, NULL);
 158         else
 159                 man->meta.hasbody = 1;
 160
 161         if (NULL == man->meta.title) {
 162                 mandoc_msg(MANDOCERR_TH_NOTITLE, man->parse,
 163                     n->line, n->pos, NULL);
 164
 165                 /*
 166                  * If a title hasn't been set, do so now (by
 167                  * implication, date and section also aren't set).
 168                  */
 169
 170                 man->meta.title = mandoc_strdup("");
 171                 man->meta.msec = mandoc_strdup("");
 172                 man->meta.date = man->quick ? mandoc_strdup("") :
 173                     mandoc_normdate(man, NULL, n->line, n->pos);
 174         }
 175
 176         if (man->meta.os_e &&
 177             (man->meta.rcsids & (1 << man->meta.os_e)) == 0)
 178                 mandoc_msg(MANDOCERR_RCS_MISSING, man->parse, 0, 0,
 179                     man->meta.os_e == MANDOC_OS_OPENBSD ?
 180                     "(OpenBSD)" : "(NetBSD)");
 181 }
 182
 183 static void
 184 check_text(CHKARGS)
 185 {
 186         char            *cp, *p;
 187
 188         if (MAN_LITERAL & man->flags)
 189                 return;
 190
 191         cp = n->string;
 192         for (p = cp; NULL != (p = strchr(p, '\t')); p++)
 193                 mandoc_msg(MANDOCERR_FI_TAB, man->parse,
 194                     n->line, n->pos + (p - cp), NULL);
 195 }
 196
 197 static void
 198 post_OP(CHKARGS)
 199 {
 200
 201         if (n->child == NULL)
 202                 mandoc_msg(MANDOCERR_OP_EMPTY, man->parse,
 203                     n->line, n->pos, "OP");
 204         else if (n->child->next != NULL && n->child->next->next != NULL) {
 205                 n = n->child->next->next;
 206                 mandoc_vmsg(MANDOCERR_ARG_EXCESS, man->parse,
 207                     n->line, n->pos, "OP ... %s", n->string);
 208         }
 209 }
 210
 211 static void
 212 post_UR(CHKARGS)
 213 {
 214
 215         if (n->type == ROFFT_HEAD && n->child == NULL)
 216                 mandoc_vmsg(MANDOCERR_UR_NOHEAD, man->parse,
 217                     n->line, n->pos, roff_name[n->tok]);
 218         check_part(man, n);
 219 }
 220
 221 static void
 222 check_part(CHKARGS)
 223 {
 224
 225         if (n->type == ROFFT_BODY && n->child == NULL)
 226                 mandoc_msg(MANDOCERR_BLK_EMPTY, man->parse,
 227                     n->line, n->pos, roff_name[n->tok]);
 228 }
 229
 230 static void
 231 check_par(CHKARGS)
 232 {
 233
 234         switch (n->type) {
 235         case ROFFT_BLOCK:
 236                 if (n->body->child == NULL)
 237                         roff_node_delete(man, n);
 238                 break;
 239         case ROFFT_BODY:
 240                 if (n->child == NULL)
 241                         mandoc_vmsg(MANDOCERR_PAR_SKIP,
 242                             man->parse, n->line, n->pos,
 243                             "%s empty", roff_name[n->tok]);
 244                 break;
 245         case ROFFT_HEAD:
 246                 if (n->child != NULL)
 247                         mandoc_vmsg(MANDOCERR_ARG_SKIP,
 248                             man->parse, n->line, n->pos, "%s %s%s",
 249                             roff_name[n->tok], n->child->string,
 250                             n->child->next != NULL ? " ..." : "");
 251                 break;
 252         default:
 253                 break;
 254         }
 255 }
 256
 257 static void
 258 post_IP(CHKARGS)
 259 {
 260
 261         switch (n->type) {
 262         case ROFFT_BLOCK:
 263                 if (n->head->child == NULL && n->body->child == NULL)
 264                         roff_node_delete(man, n);
 265                 break;
 266         case ROFFT_BODY:
 267                 if (n->parent->head->child == NULL && n->child == NULL)
 268                         mandoc_vmsg(MANDOCERR_PAR_SKIP,
 269                             man->parse, n->line, n->pos,
 270                             "%s empty", roff_name[n->tok]);
 271                 break;
 272         default:
 273                 break;
 274         }
 275 }
 276
 277 static void
 278 post_TH(CHKARGS)
 279 {
 280         struct roff_node *nb;
 281         const char      *p;
 282
 283         free(man->meta.title);
 284         free(man->meta.vol);
 285         free(man->meta.os);
 286         free(man->meta.msec);
 287         free(man->meta.date);
 288
 289         man->meta.title = man->meta.vol = man->meta.date =
 290             man->meta.msec = man->meta.os = NULL;
 291
 292         nb = n;
 293
 294         /* ->TITLE<- MSEC DATE OS VOL */
 295
 296         n = n->child;
 297         if (n && n->string) {
 298                 for (p = n->string; '\0' != *p; p++) {
 299                         /* Only warn about this once... */
 300                         if (isalpha((unsigned char)*p) &&
 301                             ! isupper((unsigned char)*p)) {
 302                                 mandoc_vmsg(MANDOCERR_TITLE_CASE,
 303                                     man->parse, n->line,
 304                                     n->pos + (p - n->string),
 305                                     "TH %s", n->string);
 306                                 break;
 307                         }
 308                 }
 309                 man->meta.title = mandoc_strdup(n->string);
 310         } else {
 311                 man->meta.title = mandoc_strdup("");
 312                 mandoc_msg(MANDOCERR_TH_NOTITLE, man->parse,
 313                     nb->line, nb->pos, "TH");
 314         }
 315
 316         /* TITLE ->MSEC<- DATE OS VOL */
 317
 318         if (n)
 319                 n = n->next;
 320         if (n && n->string)
 321                 man->meta.msec = mandoc_strdup(n->string);
 322         else {
 323                 man->meta.msec = mandoc_strdup("");
 324                 mandoc_vmsg(MANDOCERR_MSEC_MISSING, man->parse,
 325                     nb->line, nb->pos, "TH %s", man->meta.title);
 326         }
 327
 328         /* TITLE MSEC ->DATE<- OS VOL */
 329
 330         if (n)
 331                 n = n->next;
 332         if (n && n->string && '\0' != n->string[0]) {
 333                 man->meta.date = man->quick ?
 334                     mandoc_strdup(n->string) :
 335                     mandoc_normdate(man, n->string, n->line, n->pos);
 336         } else {
 337                 man->meta.date = mandoc_strdup("");
 338                 mandoc_msg(MANDOCERR_DATE_MISSING, man->parse,
 339                     n ? n->line : nb->line,
 340                     n ? n->pos : nb->pos, "TH");
 341         }
 342
 343         /* TITLE MSEC DATE ->OS<- VOL */
 344
 345         if (n && (n = n->next))
 346                 man->meta.os = mandoc_strdup(n->string);
 347         else if (man->os_s != NULL)
 348                 man->meta.os = mandoc_strdup(man->os_s);
 349         if (man->meta.os_e == MANDOC_OS_OTHER && man->meta.os != NULL) {
 350                 if (strstr(man->meta.os, "OpenBSD") != NULL)
 351                         man->meta.os_e = MANDOC_OS_OPENBSD;
 352                 else if (strstr(man->meta.os, "NetBSD") != NULL)
 353                         man->meta.os_e = MANDOC_OS_NETBSD;
 354         }
 355
 356         /* TITLE MSEC DATE OS ->VOL<- */
 357         /* If missing, use the default VOL name for MSEC. */
 358
 359         if (n && (n = n->next))
 360                 man->meta.vol = mandoc_strdup(n->string);
 361         else if ('\0' != man->meta.msec[0] &&
 362             (NULL != (p = mandoc_a2msec(man->meta.msec))))
 363                 man->meta.vol = mandoc_strdup(p);
 364
 365         if (n != NULL && (n = n->next) != NULL)
 366                 mandoc_vmsg(MANDOCERR_ARG_EXCESS, man->parse,
 367                     n->line, n->pos, "TH ... %s", n->string);
 368
 369         /*
 370          * Remove the `TH' node after we've processed it for our
 371          * meta-data.
 372          */
 373         roff_node_delete(man, man->last);
 374 }
 375
 376 static void
 377 post_UC(CHKARGS)
 378 {
 379         static const char * const bsd_versions[] = {
 380             "3rd Berkeley Distribution",
 381             "4th Berkeley Distribution",
 382             "4.2 Berkeley Distribution",
 383             "4.3 Berkeley Distribution",
 384             "4.4 Berkeley Distribution",
 385         };
 386
 387         const char      *p, *s;
 388
 389         n = n->child;
 390
 391         if (n == NULL || n->type != ROFFT_TEXT)
 392                 p = bsd_versions[0];
 393         else {
 394                 s = n->string;
 395                 if (0 == strcmp(s, "3"))
 396                         p = bsd_versions[0];
 397                 else if (0 == strcmp(s, "4"))
 398                         p = bsd_versions[1];
 399                 else if (0 == strcmp(s, "5"))
 400                         p = bsd_versions[2];
 401                 else if (0 == strcmp(s, "6"))
 402                         p = bsd_versions[3];
 403                 else if (0 == strcmp(s, "7"))
 404                         p = bsd_versions[4];
 405                 else
 406                         p = bsd_versions[0];
 407         }
 408
 409         free(man->meta.os);
 410         man->meta.os = mandoc_strdup(p);
 411 }
 412
 413 static void
 414 post_AT(CHKARGS)
 415 {
 416         static const char * const unix_versions[] = {
 417             "7th Edition",
 418             "System III",
 419             "System V",
 420             "System V Release 2",
 421         };
 422
 423         struct roff_node *nn;
 424         const char      *p, *s;
 425
 426         n = n->child;
 427
 428         if (n == NULL || n->type != ROFFT_TEXT)
 429                 p = unix_versions[0];
 430         else {
 431                 s = n->string;
 432                 if (0 == strcmp(s, "3"))
 433                         p = unix_versions[0];
 434                 else if (0 == strcmp(s, "4"))
 435                         p = unix_versions[1];
 436                 else if (0 == strcmp(s, "5")) {
 437                         nn = n->next;
 438                         if (nn != NULL &&
 439                             nn->type == ROFFT_TEXT &&
 440                             nn->string[0] != '\0')
 441                                 p = unix_versions[3];
 442                         else
 443                                 p = unix_versions[2];
 444                 } else
 445                         p = unix_versions[0];
 446         }
 447
 448         free(man->meta.os);
 449         man->meta.os = mandoc_strdup(p);
 450 }
 451
 452 static void
 453 post_in(CHKARGS)
 454 {
 455         char    *s;
 456
 457         if (n->parent->tok != MAN_TP ||
 458             n->parent->type != ROFFT_HEAD ||
 459             n->child == NULL ||
 460             *n->child->string == '+' ||
 461             *n->child->string == '-')
 462                 return;
 463         mandoc_asprintf(&s, "+%s", n->child->string);
 464         free(n->child->string);
 465         n->child->string = s;
 466 }
 467
 468 static void
 469 post_vs(CHKARGS)
 470 {
 471
 472         if (NULL != n->prev)
 473                 return;
 474
 475         switch (n->parent->tok) {
 476         case MAN_SH:
 477         case MAN_SS:
 478         case MAN_PP:
 479         case MAN_LP:
 480         case MAN_P:
 481                 mandoc_vmsg(MANDOCERR_PAR_SKIP, man->parse, n->line, n->pos,
 482                     "%s after %s", roff_name[n->tok],
 483                     roff_name[n->parent->tok]);
 484                 /* FALLTHROUGH */
 485         case TOKEN_NONE:
 486                 /*
 487                  * Don't warn about this because it occurs in pod2man
 488                  * and would cause considerable (unfixable) warnage.
 489                  */
 490                 roff_node_delete(man, n);
 491                 break;
 492         default:
 493                 break;
 494         }
 495 }