1 /* Id: mandoc.h,v 1.113 2014/01/02 16:29:55 schwarze Exp */ 2 /* 3 * Copyright (c) 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv> 4 * Copyright (c) 2012, 2013, 2014 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 #ifndef MANDOC_H 19 #define MANDOC_H 20 21 #define ASCII_NBRSP 31 /* non-breaking space */ 22 #define ASCII_HYPH 30 /* breakable hyphen */ 23 24 /* 25 * Status level. This refers to both internal status (i.e., whilst 26 * running, when warnings/errors are reported) and an indicator of a 27 * threshold of when to halt (when said internal state exceeds the 28 * threshold). 29 */ 30 enum mandoclevel { 31 MANDOCLEVEL_OK = 0, 32 MANDOCLEVEL_RESERVED, 33 MANDOCLEVEL_WARNING, /* warnings: syntax, whitespace, etc. */ 34 MANDOCLEVEL_ERROR, /* input has been thrown away */ 35 MANDOCLEVEL_FATAL, /* input is borked */ 36 MANDOCLEVEL_BADARG, /* bad argument in invocation */ 37 MANDOCLEVEL_SYSERR, /* system error */ 38 MANDOCLEVEL_MAX 39 }; 40 41 /* 42 * All possible things that can go wrong within a parse, be it libroff, 43 * libmdoc, or libman. 44 */ 45 enum mandocerr { 46 MANDOCERR_OK, 47 48 MANDOCERR_WARNING, /* ===== start of warnings ===== */ 49 50 /* related to the prologue */ 51 MANDOCERR_NOTITLE, /* no title in document */ 52 MANDOCERR_UPPERCASE, /* document title should be all caps */ 53 MANDOCERR_BADMSEC, /* unknown manual section */ 54 MANDOCERR_BADVOLARCH, /* unknown manual volume or arch */ 55 MANDOCERR_NODATE, /* date missing, using today's date */ 56 MANDOCERR_BADDATE, /* cannot parse date, using it verbatim */ 57 MANDOCERR_PROLOGOOO, /* prologue macros out of order */ 58 MANDOCERR_PROLOGREP, /* duplicate prologue macro */ 59 MANDOCERR_BADPROLOG, /* macro not allowed in prologue */ 60 MANDOCERR_BADBODY, /* macro not allowed in body */ 61 62 /* related to document structure */ 63 MANDOCERR_SO, /* .so is fragile, better use ln(1) */ 64 MANDOCERR_NAMESECFIRST, /* NAME section must come first */ 65 MANDOCERR_BADNAMESEC, /* bad NAME section contents */ 66 MANDOCERR_SECOOO, /* sections out of conventional order */ 67 MANDOCERR_SECREP, /* duplicate section name */ 68 MANDOCERR_SECMSEC, /* section header suited to sections ... */ 69 70 /* related to macros and nesting */ 71 MANDOCERR_MACROOBS, /* skipping obsolete macro */ 72 MANDOCERR_IGNPAR, /* skipping paragraph macro */ 73 MANDOCERR_MOVEPAR, /* moving paragraph macro out of list */ 74 MANDOCERR_IGNNS, /* skipping no-space macro */ 75 MANDOCERR_SCOPENEST, /* blocks badly nested */ 76 MANDOCERR_CHILD, /* child violates parent syntax */ 77 MANDOCERR_NESTEDDISP, /* nested displays are not portable */ 78 MANDOCERR_SCOPEREP, /* already in literal mode */ 79 MANDOCERR_LINESCOPE, /* line scope broken */ 80 81 /* related to missing macro arguments */ 82 MANDOCERR_MACROEMPTY, /* skipping empty macro */ 83 MANDOCERR_ARGCWARN, /* argument count wrong */ 84 MANDOCERR_DISPTYPE, /* missing display type */ 85 MANDOCERR_LISTFIRST, /* list type must come first */ 86 MANDOCERR_NOWIDTHARG, /* tag lists require a width argument */ 87 MANDOCERR_FONTTYPE, /* missing font type */ 88 MANDOCERR_WNOSCOPE, /* skipping end of block that is not open */ 89 90 /* related to bad macro arguments */ 91 MANDOCERR_IGNARGV, /* skipping argument */ 92 MANDOCERR_ARGVREP, /* duplicate argument */ 93 MANDOCERR_DISPREP, /* duplicate display type */ 94 MANDOCERR_LISTREP, /* duplicate list type */ 95 MANDOCERR_BADATT, /* unknown AT&T UNIX version */ 96 MANDOCERR_BADBOOL, /* bad Boolean value */ 97 MANDOCERR_BADFONT, /* unknown font */ 98 MANDOCERR_BADSTANDARD, /* unknown standard specifier */ 99 MANDOCERR_BADWIDTH, /* bad width argument */ 100 101 /* related to plain text */ 102 MANDOCERR_NOBLANKLN, /* blank line in non-literal context */ 103 MANDOCERR_BADTAB, /* tab in non-literal context */ 104 MANDOCERR_EOLNSPACE, /* end of line whitespace */ 105 MANDOCERR_BADCOMMENT, /* bad comment style */ 106 MANDOCERR_BADESCAPE, /* unknown escape sequence */ 107 MANDOCERR_BADQUOTE, /* unterminated quoted string */ 108 109 /* related to equations */ 110 MANDOCERR_EQNQUOTE, /* unexpected literal in equation */ 111 112 MANDOCERR_ERROR, /* ===== start of errors ===== */ 113 114 /* related to equations */ 115 MANDOCERR_EQNNSCOPE, /* unexpected equation scope closure*/ 116 MANDOCERR_EQNSCOPE, /* equation scope open on exit */ 117 MANDOCERR_EQNBADSCOPE, /* overlapping equation scopes */ 118 MANDOCERR_EQNEOF, /* unexpected end of equation */ 119 MANDOCERR_EQNSYNT, /* equation syntax error */ 120 121 /* related to tables */ 122 MANDOCERR_TBL, /* bad table syntax */ 123 MANDOCERR_TBLOPT, /* bad table option */ 124 MANDOCERR_TBLLAYOUT, /* bad table layout */ 125 MANDOCERR_TBLNOLAYOUT, /* no table layout cells specified */ 126 MANDOCERR_TBLNODATA, /* no table data cells specified */ 127 MANDOCERR_TBLIGNDATA, /* ignore data in cell */ 128 MANDOCERR_TBLBLOCK, /* data block still open */ 129 MANDOCERR_TBLEXTRADAT, /* ignoring extra data cells */ 130 131 MANDOCERR_ROFFLOOP, /* input stack limit exceeded, infinite loop? */ 132 MANDOCERR_BADCHAR, /* skipping bad character */ 133 MANDOCERR_NAMESC, /* escaped character not allowed in a name */ 134 MANDOCERR_NONAME, /* manual name not yet set */ 135 MANDOCERR_NOTEXT, /* skipping text before the first section header */ 136 MANDOCERR_MACRO, /* skipping unknown macro */ 137 MANDOCERR_REQUEST, /* NOT IMPLEMENTED: skipping request */ 138 MANDOCERR_ARGCOUNT, /* argument count wrong */ 139 MANDOCERR_STRAYTA, /* skipping column outside column list */ 140 MANDOCERR_NOSCOPE, /* skipping end of block that is not open */ 141 MANDOCERR_SCOPEBROKEN, /* missing end of block */ 142 MANDOCERR_SCOPEEXIT, /* scope open on exit */ 143 MANDOCERR_UNAME, /* uname(3) system call failed */ 144 /* FIXME: merge following with MANDOCERR_ARGCOUNT */ 145 MANDOCERR_NOARGS, /* macro requires line argument(s) */ 146 MANDOCERR_NOBODY, /* macro requires body argument(s) */ 147 MANDOCERR_NOARGV, /* macro requires argument(s) */ 148 MANDOCERR_NUMERIC, /* request requires a numeric argument */ 149 MANDOCERR_LISTTYPE, /* missing list type */ 150 MANDOCERR_ARGSLOST, /* line argument(s) will be lost */ 151 MANDOCERR_BODYLOST, /* body argument(s) will be lost */ 152 153 MANDOCERR_FATAL, /* ===== start of fatal errors ===== */ 154 155 MANDOCERR_TOOLARGE, /* input too large */ 156 MANDOCERR_NOTMANUAL, /* manual isn't really a manual */ 157 MANDOCERR_COLUMNS, /* column syntax is inconsistent */ 158 MANDOCERR_BADDISP, /* NOT IMPLEMENTED: .Bd -file */ 159 MANDOCERR_SYNTARGVCOUNT, /* argument count wrong, violates syntax */ 160 MANDOCERR_SYNTCHILD, /* child violates parent syntax */ 161 MANDOCERR_SYNTARGCOUNT, /* argument count wrong, violates syntax */ 162 MANDOCERR_SOPATH, /* NOT IMPLEMENTED: .so with absolute path or ".." */ 163 MANDOCERR_NODOCBODY, /* no document body */ 164 MANDOCERR_NODOCPROLOG, /* no document prologue */ 165 MANDOCERR_MEM, /* static buffer exhausted */ 166 167 /* ===== system errors ===== */ 168 169 MANDOCERR_SYSOPEN, /* cannot open file */ 170 MANDOCERR_SYSSTAT, /* cannot stat file */ 171 MANDOCERR_SYSREAD, /* cannot read file */ 172 173 MANDOCERR_MAX 174 }; 175 176 struct tbl_opts { 177 char tab; /* cell-separator */ 178 char decimal; /* decimal point */ 179 int linesize; 180 int opts; 181 #define TBL_OPT_CENTRE (1 << 0) 182 #define TBL_OPT_EXPAND (1 << 1) 183 #define TBL_OPT_BOX (1 << 2) 184 #define TBL_OPT_DBOX (1 << 3) 185 #define TBL_OPT_ALLBOX (1 << 4) 186 #define TBL_OPT_NOKEEP (1 << 5) 187 #define TBL_OPT_NOSPACE (1 << 6) 188 int cols; /* number of columns */ 189 }; 190 191 /* 192 * The head of a table specifies all of its columns. When formatting a 193 * tbl_span, iterate over these and plug in data from the tbl_span when 194 * appropriate, using tbl_cell as a guide to placement. 195 */ 196 struct tbl_head { 197 int ident; /* 0 <= unique id < cols */ 198 int vert; /* width of preceding vertical line */ 199 struct tbl_head *next; 200 struct tbl_head *prev; 201 }; 202 203 enum tbl_cellt { 204 TBL_CELL_CENTRE, /* c, C */ 205 TBL_CELL_RIGHT, /* r, R */ 206 TBL_CELL_LEFT, /* l, L */ 207 TBL_CELL_NUMBER, /* n, N */ 208 TBL_CELL_SPAN, /* s, S */ 209 TBL_CELL_LONG, /* a, A */ 210 TBL_CELL_DOWN, /* ^ */ 211 TBL_CELL_HORIZ, /* _, - */ 212 TBL_CELL_DHORIZ, /* = */ 213 TBL_CELL_MAX 214 }; 215 216 /* 217 * A cell in a layout row. 218 */ 219 struct tbl_cell { 220 struct tbl_cell *next; 221 int vert; /* width of preceding vertical line */ 222 enum tbl_cellt pos; 223 size_t spacing; 224 int flags; 225 #define TBL_CELL_TALIGN (1 << 0) /* t, T */ 226 #define TBL_CELL_BALIGN (1 << 1) /* d, D */ 227 #define TBL_CELL_BOLD (1 << 2) /* fB, B, b */ 228 #define TBL_CELL_ITALIC (1 << 3) /* fI, I, i */ 229 #define TBL_CELL_EQUAL (1 << 4) /* e, E */ 230 #define TBL_CELL_UP (1 << 5) /* u, U */ 231 #define TBL_CELL_WIGN (1 << 6) /* z, Z */ 232 struct tbl_head *head; 233 }; 234 235 /* 236 * A layout row. 237 */ 238 struct tbl_row { 239 struct tbl_row *next; 240 struct tbl_cell *first; 241 struct tbl_cell *last; 242 }; 243 244 enum tbl_datt { 245 TBL_DATA_NONE, /* has no data */ 246 TBL_DATA_DATA, /* consists of data/string */ 247 TBL_DATA_HORIZ, /* horizontal line */ 248 TBL_DATA_DHORIZ, /* double-horizontal line */ 249 TBL_DATA_NHORIZ, /* squeezed horizontal line */ 250 TBL_DATA_NDHORIZ /* squeezed double-horizontal line */ 251 }; 252 253 /* 254 * A cell within a row of data. The "string" field contains the actual 255 * string value that's in the cell. The rest is layout. 256 */ 257 struct tbl_dat { 258 struct tbl_cell *layout; /* layout cell */ 259 int spans; /* how many spans follow */ 260 struct tbl_dat *next; 261 char *string; /* data (NULL if not TBL_DATA_DATA) */ 262 enum tbl_datt pos; 263 }; 264 265 enum tbl_spant { 266 TBL_SPAN_DATA, /* span consists of data */ 267 TBL_SPAN_HORIZ, /* span is horizontal line */ 268 TBL_SPAN_DHORIZ /* span is double horizontal line */ 269 }; 270 271 /* 272 * A row of data in a table. 273 */ 274 struct tbl_span { 275 struct tbl_opts *opts; 276 struct tbl_head *head; 277 struct tbl_row *layout; /* layout row */ 278 struct tbl_dat *first; 279 struct tbl_dat *last; 280 int line; /* parse line */ 281 int flags; 282 #define TBL_SPAN_FIRST (1 << 0) 283 #define TBL_SPAN_LAST (1 << 1) 284 enum tbl_spant pos; 285 struct tbl_span *next; 286 }; 287 288 enum eqn_boxt { 289 EQN_ROOT, /* root of parse tree */ 290 EQN_TEXT, /* text (number, variable, whatever) */ 291 EQN_SUBEXPR, /* nested `eqn' subexpression */ 292 EQN_LIST, /* subexpressions list */ 293 EQN_MATRIX /* matrix subexpression */ 294 }; 295 296 enum eqn_markt { 297 EQNMARK_NONE = 0, 298 EQNMARK_DOT, 299 EQNMARK_DOTDOT, 300 EQNMARK_HAT, 301 EQNMARK_TILDE, 302 EQNMARK_VEC, 303 EQNMARK_DYAD, 304 EQNMARK_BAR, 305 EQNMARK_UNDER, 306 EQNMARK__MAX 307 }; 308 309 enum eqn_fontt { 310 EQNFONT_NONE = 0, 311 EQNFONT_ROMAN, 312 EQNFONT_BOLD, 313 EQNFONT_FAT, 314 EQNFONT_ITALIC, 315 EQNFONT__MAX 316 }; 317 318 enum eqn_post { 319 EQNPOS_NONE = 0, 320 EQNPOS_OVER, 321 EQNPOS_SUP, 322 EQNPOS_SUB, 323 EQNPOS_TO, 324 EQNPOS_FROM, 325 EQNPOS__MAX 326 }; 327 328 enum eqn_pilet { 329 EQNPILE_NONE = 0, 330 EQNPILE_PILE, 331 EQNPILE_CPILE, 332 EQNPILE_RPILE, 333 EQNPILE_LPILE, 334 EQNPILE_COL, 335 EQNPILE_CCOL, 336 EQNPILE_RCOL, 337 EQNPILE_LCOL, 338 EQNPILE__MAX 339 }; 340 341 /* 342 * A "box" is a parsed mathematical expression as defined by the eqn.7 343 * grammar. 344 */ 345 struct eqn_box { 346 int size; /* font size of expression */ 347 #define EQN_DEFSIZE INT_MIN 348 enum eqn_boxt type; /* type of node */ 349 struct eqn_box *first; /* first child node */ 350 struct eqn_box *last; /* last child node */ 351 struct eqn_box *next; /* node sibling */ 352 struct eqn_box *parent; /* node sibling */ 353 char *text; /* text (or NULL) */ 354 char *left; 355 char *right; 356 enum eqn_post pos; /* position of next box */ 357 enum eqn_markt mark; /* a mark about the box */ 358 enum eqn_fontt font; /* font of box */ 359 enum eqn_pilet pile; /* equation piling */ 360 }; 361 362 /* 363 * An equation consists of a tree of expressions starting at a given 364 * line and position. 365 */ 366 struct eqn { 367 char *name; /* identifier (or NULL) */ 368 struct eqn_box *root; /* root mathematical expression */ 369 int ln; /* invocation line */ 370 int pos; /* invocation position */ 371 }; 372 373 /* 374 * The type of parse sequence. This value is usually passed via the 375 * mandoc(1) command line of -man and -mdoc. It's almost exclusively 376 * -mandoc but the others have been retained for compatibility. 377 */ 378 enum mparset { 379 MPARSE_AUTO, /* magically determine the document type */ 380 MPARSE_MDOC, /* assume -mdoc */ 381 MPARSE_MAN /* assume -man */ 382 }; 383 384 enum mandoc_esc { 385 ESCAPE_ERROR = 0, /* bail! unparsable escape */ 386 ESCAPE_IGNORE, /* escape to be ignored */ 387 ESCAPE_SPECIAL, /* a regular special character */ 388 ESCAPE_FONT, /* a generic font mode */ 389 ESCAPE_FONTBOLD, /* bold font mode */ 390 ESCAPE_FONTITALIC, /* italic font mode */ 391 ESCAPE_FONTBI, /* bold italic font mode */ 392 ESCAPE_FONTROMAN, /* roman font mode */ 393 ESCAPE_FONTPREV, /* previous font mode */ 394 ESCAPE_NUMBERED, /* a numbered glyph */ 395 ESCAPE_UNICODE, /* a unicode codepoint */ 396 ESCAPE_NOSPACE, /* suppress space if the last on a line */ 397 ESCAPE_SKIPCHAR /* skip the next character */ 398 }; 399 400 typedef void (*mandocmsg)(enum mandocerr, enum mandoclevel, 401 const char *, int, int, const char *); 402 403 struct mparse; 404 struct mchars; 405 struct mdoc; 406 struct man; 407 408 __BEGIN_DECLS 409 410 void *mandoc_calloc(size_t, size_t); 411 enum mandoc_esc mandoc_escape(const char **, const char **, int *); 412 void *mandoc_malloc(size_t); 413 void *mandoc_realloc(void *, size_t); 414 char *mandoc_strdup(const char *); 415 char *mandoc_strndup(const char *, size_t); 416 struct mchars *mchars_alloc(void); 417 void mchars_free(struct mchars *); 418 char mchars_num2char(const char *, size_t); 419 int mchars_num2uc(const char *, size_t); 420 int mchars_spec2cp(const struct mchars *, 421 const char *, size_t); 422 const char *mchars_spec2str(const struct mchars *, 423 const char *, size_t, size_t *); 424 struct mparse *mparse_alloc(enum mparset, enum mandoclevel, 425 mandocmsg, void *, char *); 426 void mparse_free(struct mparse *); 427 void mparse_keep(struct mparse *); 428 enum mandoclevel mparse_readfd(struct mparse *, int, const char *); 429 enum mandoclevel mparse_readmem(struct mparse *, const void *, size_t, 430 const char *); 431 void mparse_reset(struct mparse *); 432 void mparse_result(struct mparse *, 433 struct mdoc **, struct man **); 434 const char *mparse_getkeep(const struct mparse *); 435 const char *mparse_strerror(enum mandocerr); 436 const char *mparse_strlevel(enum mandoclevel); 437 438 __END_DECLS 439 440 #endif /*!MANDOC_H*/ 441