1 /* $Id: tbl_layout.c,v 1.11 2012/05/26 20:03:34 schwarze Exp $ */ 2 /* 3 * Copyright (c) 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv> 4 * Copyright (c) 2012 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 AUTHOR DISCLAIMS ALL WARRANTIES 11 * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF 12 * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR 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 <assert.h> 19 #include <ctype.h> 20 #include <stdlib.h> 21 #include <string.h> 22 #include <time.h> 23 24 #include "mandoc.h" 25 #include "libmandoc.h" 26 #include "libroff.h" 27 28 struct tbl_phrase { 29 char name; 30 enum tbl_cellt key; 31 }; 32 33 /* 34 * FIXME: we can make this parse a lot nicer by, when an error is 35 * encountered in a layout key, bailing to the next key (i.e. to the 36 * next whitespace then continuing). 37 */ 38 39 #define KEYS_MAX 11 40 41 static const struct tbl_phrase keys[KEYS_MAX] = { 42 { 'c', TBL_CELL_CENTRE }, 43 { 'r', TBL_CELL_RIGHT }, 44 { 'l', TBL_CELL_LEFT }, 45 { 'n', TBL_CELL_NUMBER }, 46 { 's', TBL_CELL_SPAN }, 47 { 'a', TBL_CELL_LONG }, 48 { '^', TBL_CELL_DOWN }, 49 { '-', TBL_CELL_HORIZ }, 50 { '_', TBL_CELL_HORIZ }, 51 { '=', TBL_CELL_DHORIZ } 52 }; 53 54 static int mods(struct tbl_node *, struct tbl_cell *, 55 int, const char *, int *); 56 static int cell(struct tbl_node *, struct tbl_row *, 57 int, const char *, int *); 58 static void row(struct tbl_node *, int, const char *, int *); 59 static struct tbl_cell *cell_alloc(struct tbl_node *, struct tbl_row *, 60 enum tbl_cellt, int vert); 61 62 static int 63 mods(struct tbl_node *tbl, struct tbl_cell *cp, 64 int ln, const char *p, int *pos) 65 { 66 char buf[5]; 67 int i; 68 69 /* Not all types accept modifiers. */ 70 71 switch (cp->pos) { 72 case (TBL_CELL_DOWN): 73 /* FALLTHROUGH */ 74 case (TBL_CELL_HORIZ): 75 /* FALLTHROUGH */ 76 case (TBL_CELL_DHORIZ): 77 return(1); 78 default: 79 break; 80 } 81 82 mod: 83 /* 84 * XXX: since, at least for now, modifiers are non-conflicting 85 * (are separable by value, regardless of position), we let 86 * modifiers come in any order. The existing tbl doesn't let 87 * this happen. 88 */ 89 switch (p[*pos]) { 90 case ('\0'): 91 /* FALLTHROUGH */ 92 case (' '): 93 /* FALLTHROUGH */ 94 case ('\t'): 95 /* FALLTHROUGH */ 96 case (','): 97 /* FALLTHROUGH */ 98 case ('.'): 99 return(1); 100 default: 101 break; 102 } 103 104 /* Throw away parenthesised expression. */ 105 106 if ('(' == p[*pos]) { 107 (*pos)++; 108 while (p[*pos] && ')' != p[*pos]) 109 (*pos)++; 110 if (')' == p[*pos]) { 111 (*pos)++; 112 goto mod; 113 } 114 mandoc_msg(MANDOCERR_TBLLAYOUT, 115 tbl->parse, ln, *pos, NULL); 116 return(0); 117 } 118 119 /* Parse numerical spacing from modifier string. */ 120 121 if (isdigit((unsigned char)p[*pos])) { 122 for (i = 0; i < 4; i++) { 123 if ( ! isdigit((unsigned char)p[*pos + i])) 124 break; 125 buf[i] = p[*pos + i]; 126 } 127 buf[i] = '\0'; 128 129 /* No greater than 4 digits. */ 130 131 if (4 == i) { 132 mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse, 133 ln, *pos, NULL); 134 return(0); 135 } 136 137 *pos += i; 138 cp->spacing = (size_t)atoi(buf); 139 140 goto mod; 141 /* NOTREACHED */ 142 } 143 144 /* TODO: GNU has many more extensions. */ 145 146 switch (tolower((unsigned char)p[(*pos)++])) { 147 case ('z'): 148 cp->flags |= TBL_CELL_WIGN; 149 goto mod; 150 case ('u'): 151 cp->flags |= TBL_CELL_UP; 152 goto mod; 153 case ('e'): 154 cp->flags |= TBL_CELL_EQUAL; 155 goto mod; 156 case ('t'): 157 cp->flags |= TBL_CELL_TALIGN; 158 goto mod; 159 case ('d'): 160 cp->flags |= TBL_CELL_BALIGN; 161 goto mod; 162 case ('w'): /* XXX for now, ignore minimal column width */ 163 goto mod; 164 case ('f'): 165 break; 166 case ('r'): 167 /* FALLTHROUGH */ 168 case ('b'): 169 /* FALLTHROUGH */ 170 case ('i'): 171 (*pos)--; 172 break; 173 default: 174 mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse, 175 ln, *pos - 1, NULL); 176 return(0); 177 } 178 179 switch (tolower((unsigned char)p[(*pos)++])) { 180 case ('3'): 181 /* FALLTHROUGH */ 182 case ('b'): 183 cp->flags |= TBL_CELL_BOLD; 184 goto mod; 185 case ('2'): 186 /* FALLTHROUGH */ 187 case ('i'): 188 cp->flags |= TBL_CELL_ITALIC; 189 goto mod; 190 case ('1'): 191 /* FALLTHROUGH */ 192 case ('r'): 193 goto mod; 194 default: 195 break; 196 } 197 198 mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse, 199 ln, *pos - 1, NULL); 200 return(0); 201 } 202 203 static int 204 cell(struct tbl_node *tbl, struct tbl_row *rp, 205 int ln, const char *p, int *pos) 206 { 207 int vert, i; 208 enum tbl_cellt c; 209 210 /* Handle vertical lines. */ 211 212 for (vert = 0; '|' == p[*pos]; ++*pos) 213 vert++; 214 while (' ' == p[*pos]) 215 (*pos)++; 216 217 /* Parse the column position (`c', `l', `r', ...). */ 218 219 for (i = 0; i < KEYS_MAX; i++) 220 if (tolower((unsigned char)p[*pos]) == keys[i].name) 221 break; 222 223 if (KEYS_MAX == i) { 224 mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse, 225 ln, *pos, NULL); 226 return(0); 227 } 228 229 c = keys[i].key; 230 231 /* 232 * If a span cell is found first, raise a warning and abort the 233 * parse. If a span cell is found and the last layout element 234 * isn't a "normal" layout, bail. 235 * 236 * FIXME: recover from this somehow? 237 */ 238 239 if (TBL_CELL_SPAN == c) { 240 if (NULL == rp->first) { 241 mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse, 242 ln, *pos, NULL); 243 return(0); 244 } else if (rp->last) 245 switch (rp->last->pos) { 246 case (TBL_CELL_HORIZ): 247 case (TBL_CELL_DHORIZ): 248 mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse, 249 ln, *pos, NULL); 250 return(0); 251 default: 252 break; 253 } 254 } 255 256 /* 257 * If a vertical spanner is found, we may not be in the first 258 * row. 259 */ 260 261 if (TBL_CELL_DOWN == c && rp == tbl->first_row) { 262 mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse, ln, *pos, NULL); 263 return(0); 264 } 265 266 (*pos)++; 267 268 /* Disallow adjacent spacers. */ 269 270 if (vert > 2) { 271 mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse, ln, *pos - 1, NULL); 272 return(0); 273 } 274 275 /* Allocate cell then parse its modifiers. */ 276 277 return(mods(tbl, cell_alloc(tbl, rp, c, vert), ln, p, pos)); 278 } 279 280 281 static void 282 row(struct tbl_node *tbl, int ln, const char *p, int *pos) 283 { 284 struct tbl_row *rp; 285 286 row: /* 287 * EBNF describing this section: 288 * 289 * row ::= row_list [:space:]* [.]?[\n] 290 * row_list ::= [:space:]* row_elem row_tail 291 * row_tail ::= [:space:]*[,] row_list | 292 * epsilon 293 * row_elem ::= [\t\ ]*[:alpha:]+ 294 */ 295 296 rp = mandoc_calloc(1, sizeof(struct tbl_row)); 297 if (tbl->last_row) 298 tbl->last_row->next = rp; 299 else 300 tbl->first_row = rp; 301 tbl->last_row = rp; 302 303 cell: 304 while (isspace((unsigned char)p[*pos])) 305 (*pos)++; 306 307 /* Safely exit layout context. */ 308 309 if ('.' == p[*pos]) { 310 tbl->part = TBL_PART_DATA; 311 if (NULL == tbl->first_row) 312 mandoc_msg(MANDOCERR_TBLNOLAYOUT, tbl->parse, 313 ln, *pos, NULL); 314 (*pos)++; 315 return; 316 } 317 318 /* End (and possibly restart) a row. */ 319 320 if (',' == p[*pos]) { 321 (*pos)++; 322 goto row; 323 } else if ('\0' == p[*pos]) 324 return; 325 326 if ( ! cell(tbl, rp, ln, p, pos)) 327 return; 328 329 goto cell; 330 /* NOTREACHED */ 331 } 332 333 int 334 tbl_layout(struct tbl_node *tbl, int ln, const char *p) 335 { 336 int pos; 337 338 pos = 0; 339 row(tbl, ln, p, &pos); 340 341 /* Always succeed. */ 342 return(1); 343 } 344 345 static struct tbl_cell * 346 cell_alloc(struct tbl_node *tbl, struct tbl_row *rp, enum tbl_cellt pos, 347 int vert) 348 { 349 struct tbl_cell *p, *pp; 350 struct tbl_head *h, *hp; 351 352 p = mandoc_calloc(1, sizeof(struct tbl_cell)); 353 354 if (NULL != (pp = rp->last)) { 355 pp->next = p; 356 h = pp->head->next; 357 } else { 358 rp->first = p; 359 h = tbl->first_head; 360 } 361 rp->last = p; 362 363 p->pos = pos; 364 p->vert = vert; 365 366 /* Re-use header. */ 367 368 if (h) { 369 p->head = h; 370 return(p); 371 } 372 373 hp = mandoc_calloc(1, sizeof(struct tbl_head)); 374 hp->ident = tbl->opts.cols++; 375 hp->vert = vert; 376 377 if (tbl->last_head) { 378 hp->prev = tbl->last_head; 379 tbl->last_head->next = hp; 380 } else 381 tbl->first_head = hp; 382 tbl->last_head = hp; 383 384 p->head = hp; 385 return(p); 386 } 387