1 /* $Id: mandoc.h,v 1.62 2014/03/21 22:17:01 schwarze Exp $ */ 2 /* 3 * Copyright (c) 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv> 4 * Copyright (c) 2010-2014 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 #define ASCII_BREAK 29 /* breakable zero-width space */ 24 25 /* 26 * Status level. This refers to both internal status (i.e., whilst 27 * running, when warnings/errors are reported) and an indicator of a 28 * threshold of when to halt (when said internal state exceeds the 29 * threshold). 30 */ 31 enum mandoclevel { 32 MANDOCLEVEL_OK = 0, 33 MANDOCLEVEL_RESERVED, 34 MANDOCLEVEL_WARNING, /* warnings: syntax, whitespace, etc. */ 35 MANDOCLEVEL_ERROR, /* input has been thrown away */ 36 MANDOCLEVEL_FATAL, /* input is borked */ 37 MANDOCLEVEL_BADARG, /* bad argument in invocation */ 38 MANDOCLEVEL_SYSERR, /* system error */ 39 MANDOCLEVEL_MAX 40 }; 41 42 /* 43 * All possible things that can go wrong within a parse, be it libroff, 44 * libmdoc, or libman. 45 */ 46 enum mandocerr { 47 MANDOCERR_OK, 48 49 MANDOCERR_WARNING, /* ===== start of warnings ===== */ 50 51 /* related to the prologue */ 52 MANDOCERR_NOTITLE, /* no title in document */ 53 MANDOCERR_UPPERCASE, /* document title should be all caps */ 54 MANDOCERR_BADMSEC, /* unknown manual section */ 55 MANDOCERR_BADVOLARCH, /* unknown manual volume or arch */ 56 MANDOCERR_NODATE, /* date missing, using today's date */ 57 MANDOCERR_BADDATE, /* cannot parse date, using it verbatim */ 58 MANDOCERR_PROLOGOOO, /* prologue macros out of order */ 59 MANDOCERR_PROLOGREP, /* duplicate prologue macro */ 60 MANDOCERR_BADPROLOG, /* macro not allowed in prologue */ 61 MANDOCERR_BADBODY, /* macro not allowed in body */ 62 63 /* related to document structure */ 64 MANDOCERR_SO, /* .so is fragile, better use ln(1) */ 65 MANDOCERR_NAMESECFIRST, /* NAME section must come first */ 66 MANDOCERR_BADNAMESEC, /* bad NAME section contents */ 67 MANDOCERR_SECOOO, /* sections out of conventional order */ 68 MANDOCERR_SECREP, /* duplicate section name */ 69 MANDOCERR_SECMSEC, /* section header suited to sections ... */ 70 71 /* related to macros and nesting */ 72 MANDOCERR_MACROOBS, /* skipping obsolete macro */ 73 MANDOCERR_IGNPAR, /* skipping paragraph macro */ 74 MANDOCERR_MOVEPAR, /* moving paragraph macro out of list */ 75 MANDOCERR_IGNNS, /* skipping no-space macro */ 76 MANDOCERR_SCOPENEST, /* blocks badly nested */ 77 MANDOCERR_CHILD, /* child violates parent syntax */ 78 MANDOCERR_NESTEDDISP, /* nested displays are not portable */ 79 MANDOCERR_SCOPEREP, /* already in literal mode */ 80 MANDOCERR_LINESCOPE, /* line scope broken */ 81 82 /* related to missing macro arguments */ 83 MANDOCERR_MACROEMPTY, /* skipping empty macro */ 84 MANDOCERR_ARGCWARN, /* argument count wrong */ 85 MANDOCERR_DISPTYPE, /* missing display type */ 86 MANDOCERR_LISTFIRST, /* list type must come first */ 87 MANDOCERR_NOWIDTHARG, /* tag lists require a width argument */ 88 MANDOCERR_FONTTYPE, /* missing font type */ 89 MANDOCERR_WNOSCOPE, /* skipping end of block that is not open */ 90 91 /* related to bad macro arguments */ 92 MANDOCERR_IGNARGV, /* skipping argument */ 93 MANDOCERR_ARGVREP, /* duplicate argument */ 94 MANDOCERR_DISPREP, /* duplicate display type */ 95 MANDOCERR_LISTREP, /* duplicate list type */ 96 MANDOCERR_BADATT, /* unknown AT&T UNIX version */ 97 MANDOCERR_BADBOOL, /* bad Boolean value */ 98 MANDOCERR_BADFONT, /* unknown font */ 99 MANDOCERR_BADSTANDARD, /* unknown standard specifier */ 100 MANDOCERR_BADWIDTH, /* bad width argument */ 101 102 /* related to plain text */ 103 MANDOCERR_NOBLANKLN, /* blank line in non-literal context */ 104 MANDOCERR_BADTAB, /* tab in non-literal context */ 105 MANDOCERR_EOLNSPACE, /* end of line whitespace */ 106 MANDOCERR_BADCOMMENT, /* bad comment style */ 107 MANDOCERR_BADESCAPE, /* unknown escape sequence */ 108 MANDOCERR_BADQUOTE, /* unterminated quoted string */ 109 110 /* related to equations */ 111 MANDOCERR_EQNQUOTE, /* unexpected literal in equation */ 112 113 MANDOCERR_ERROR, /* ===== start of errors ===== */ 114 115 /* related to equations */ 116 MANDOCERR_EQNNSCOPE, /* unexpected equation scope closure*/ 117 MANDOCERR_EQNSCOPE, /* equation scope open on exit */ 118 MANDOCERR_EQNBADSCOPE, /* overlapping equation scopes */ 119 MANDOCERR_EQNEOF, /* unexpected end of equation */ 120 MANDOCERR_EQNSYNT, /* equation syntax error */ 121 122 /* related to tables */ 123 MANDOCERR_TBL, /* bad table syntax */ 124 MANDOCERR_TBLOPT, /* bad table option */ 125 MANDOCERR_TBLLAYOUT, /* bad table layout */ 126 MANDOCERR_TBLNOLAYOUT, /* no table layout cells specified */ 127 MANDOCERR_TBLNODATA, /* no table data cells specified */ 128 MANDOCERR_TBLIGNDATA, /* ignore data in cell */ 129 MANDOCERR_TBLBLOCK, /* data block still open */ 130 MANDOCERR_TBLEXTRADAT, /* ignoring extra data cells */ 131 132 MANDOCERR_ROFFLOOP, /* input stack limit exceeded, infinite loop? */ 133 MANDOCERR_BADCHAR, /* skipping bad character */ 134 MANDOCERR_NAMESC, /* escaped character not allowed in a name */ 135 MANDOCERR_NONAME, /* manual name not yet set */ 136 MANDOCERR_NOTEXT, /* skipping text before the first section header */ 137 MANDOCERR_MACRO, /* skipping unknown macro */ 138 MANDOCERR_REQUEST, /* NOT IMPLEMENTED: skipping request */ 139 MANDOCERR_ARGCOUNT, /* argument count wrong */ 140 MANDOCERR_STRAYTA, /* skipping column outside column list */ 141 MANDOCERR_NOSCOPE, /* skipping end of block that is not open */ 142 MANDOCERR_SCOPEBROKEN, /* missing end of block */ 143 MANDOCERR_SCOPEEXIT, /* scope open on exit */ 144 MANDOCERR_UNAME, /* uname(3) system call failed */ 145 /* FIXME: merge following with MANDOCERR_ARGCOUNT */ 146 MANDOCERR_NOARGS, /* macro requires line argument(s) */ 147 MANDOCERR_NOBODY, /* macro requires body argument(s) */ 148 MANDOCERR_NOARGV, /* macro requires argument(s) */ 149 MANDOCERR_NUMERIC, /* request requires a numeric argument */ 150 MANDOCERR_LISTTYPE, /* missing list type */ 151 MANDOCERR_ARGSLOST, /* line argument(s) will be lost */ 152 MANDOCERR_BODYLOST, /* body argument(s) will be lost */ 153 154 MANDOCERR_FATAL, /* ===== start of fatal errors ===== */ 155 156 MANDOCERR_TOOLARGE, /* input too large */ 157 MANDOCERR_NOTMANUAL, /* manual isn't really a manual */ 158 MANDOCERR_COLUMNS, /* column syntax is inconsistent */ 159 MANDOCERR_BADDISP, /* NOT IMPLEMENTED: .Bd -file */ 160 MANDOCERR_SYNTARGVCOUNT, /* argument count wrong, violates syntax */ 161 MANDOCERR_SYNTCHILD, /* child violates parent syntax */ 162 MANDOCERR_SYNTARGCOUNT, /* argument count wrong, violates syntax */ 163 MANDOCERR_SOPATH, /* NOT IMPLEMENTED: .so with absolute path or ".." */ 164 MANDOCERR_NODOCBODY, /* no document body */ 165 MANDOCERR_NODOCPROLOG, /* no document prologue */ 166 MANDOCERR_MEM, /* static buffer exhausted */ 167 168 /* ===== system errors ===== */ 169 170 MANDOCERR_SYSOPEN, /* cannot open file */ 171 MANDOCERR_SYSSTAT, /* cannot stat file */ 172 MANDOCERR_SYSREAD, /* cannot read file */ 173 174 MANDOCERR_MAX 175 }; 176 177 struct tbl_opts { 178 char tab; /* cell-separator */ 179 char decimal; /* decimal point */ 180 int linesize; 181 int opts; 182 #define TBL_OPT_CENTRE (1 << 0) 183 #define TBL_OPT_EXPAND (1 << 1) 184 #define TBL_OPT_BOX (1 << 2) 185 #define TBL_OPT_DBOX (1 << 3) 186 #define TBL_OPT_ALLBOX (1 << 4) 187 #define TBL_OPT_NOKEEP (1 << 5) 188 #define TBL_OPT_NOSPACE (1 << 6) 189 int cols; /* number of columns */ 190 }; 191 192 /* 193 * The head of a table specifies all of its columns. When formatting a 194 * tbl_span, iterate over these and plug in data from the tbl_span when 195 * appropriate, using tbl_cell as a guide to placement. 196 */ 197 struct tbl_head { 198 int ident; /* 0 <= unique id < cols */ 199 int vert; /* width of preceding vertical line */ 200 struct tbl_head *next; 201 struct tbl_head *prev; 202 }; 203 204 enum tbl_cellt { 205 TBL_CELL_CENTRE, /* c, C */ 206 TBL_CELL_RIGHT, /* r, R */ 207 TBL_CELL_LEFT, /* l, L */ 208 TBL_CELL_NUMBER, /* n, N */ 209 TBL_CELL_SPAN, /* s, S */ 210 TBL_CELL_LONG, /* a, A */ 211 TBL_CELL_DOWN, /* ^ */ 212 TBL_CELL_HORIZ, /* _, - */ 213 TBL_CELL_DHORIZ, /* = */ 214 TBL_CELL_MAX 215 }; 216 217 /* 218 * A cell in a layout row. 219 */ 220 struct tbl_cell { 221 struct tbl_cell *next; 222 int vert; /* width of preceding vertical line */ 223 enum tbl_cellt pos; 224 size_t spacing; 225 int flags; 226 #define TBL_CELL_TALIGN (1 << 0) /* t, T */ 227 #define TBL_CELL_BALIGN (1 << 1) /* d, D */ 228 #define TBL_CELL_BOLD (1 << 2) /* fB, B, b */ 229 #define TBL_CELL_ITALIC (1 << 3) /* fI, I, i */ 230 #define TBL_CELL_EQUAL (1 << 4) /* e, E */ 231 #define TBL_CELL_UP (1 << 5) /* u, U */ 232 #define TBL_CELL_WIGN (1 << 6) /* z, Z */ 233 struct tbl_head *head; 234 }; 235 236 /* 237 * A layout row. 238 */ 239 struct tbl_row { 240 struct tbl_row *next; 241 struct tbl_cell *first; 242 struct tbl_cell *last; 243 }; 244 245 enum tbl_datt { 246 TBL_DATA_NONE, /* has no data */ 247 TBL_DATA_DATA, /* consists of data/string */ 248 TBL_DATA_HORIZ, /* horizontal line */ 249 TBL_DATA_DHORIZ, /* double-horizontal line */ 250 TBL_DATA_NHORIZ, /* squeezed horizontal line */ 251 TBL_DATA_NDHORIZ /* squeezed double-horizontal line */ 252 }; 253 254 /* 255 * A cell within a row of data. The "string" field contains the actual 256 * string value that's in the cell. The rest is layout. 257 */ 258 struct tbl_dat { 259 struct tbl_cell *layout; /* layout cell */ 260 int spans; /* how many spans follow */ 261 struct tbl_dat *next; 262 char *string; /* data (NULL if not TBL_DATA_DATA) */ 263 enum tbl_datt pos; 264 }; 265 266 enum tbl_spant { 267 TBL_SPAN_DATA, /* span consists of data */ 268 TBL_SPAN_HORIZ, /* span is horizontal line */ 269 TBL_SPAN_DHORIZ /* span is double horizontal line */ 270 }; 271 272 /* 273 * A row of data in a table. 274 */ 275 struct tbl_span { 276 struct tbl_opts *opts; 277 struct tbl_head *head; 278 struct tbl_row *layout; /* layout row */ 279 struct tbl_dat *first; 280 struct tbl_dat *last; 281 int line; /* parse line */ 282 int flags; 283 #define TBL_SPAN_FIRST (1 << 0) 284 #define TBL_SPAN_LAST (1 << 1) 285 enum tbl_spant pos; 286 struct tbl_span *next; 287 }; 288 289 enum eqn_boxt { 290 EQN_ROOT, /* root of parse tree */ 291 EQN_TEXT, /* text (number, variable, whatever) */ 292 EQN_SUBEXPR, /* nested `eqn' subexpression */ 293 EQN_LIST, /* subexpressions list */ 294 EQN_MATRIX /* matrix subexpression */ 295 }; 296 297 enum eqn_markt { 298 EQNMARK_NONE = 0, 299 EQNMARK_DOT, 300 EQNMARK_DOTDOT, 301 EQNMARK_HAT, 302 EQNMARK_TILDE, 303 EQNMARK_VEC, 304 EQNMARK_DYAD, 305 EQNMARK_BAR, 306 EQNMARK_UNDER, 307 EQNMARK__MAX 308 }; 309 310 enum eqn_fontt { 311 EQNFONT_NONE = 0, 312 EQNFONT_ROMAN, 313 EQNFONT_BOLD, 314 EQNFONT_FAT, 315 EQNFONT_ITALIC, 316 EQNFONT__MAX 317 }; 318 319 enum eqn_post { 320 EQNPOS_NONE = 0, 321 EQNPOS_OVER, 322 EQNPOS_SUP, 323 EQNPOS_SUB, 324 EQNPOS_TO, 325 EQNPOS_FROM, 326 EQNPOS__MAX 327 }; 328 329 enum eqn_pilet { 330 EQNPILE_NONE = 0, 331 EQNPILE_PILE, 332 EQNPILE_CPILE, 333 EQNPILE_RPILE, 334 EQNPILE_LPILE, 335 EQNPILE_COL, 336 EQNPILE_CCOL, 337 EQNPILE_RCOL, 338 EQNPILE_LCOL, 339 EQNPILE__MAX 340 }; 341 342 /* 343 * A "box" is a parsed mathematical expression as defined by the eqn.7 344 * grammar. 345 */ 346 struct eqn_box { 347 int size; /* font size of expression */ 348 #define EQN_DEFSIZE INT_MIN 349 enum eqn_boxt type; /* type of node */ 350 struct eqn_box *first; /* first child node */ 351 struct eqn_box *last; /* last child node */ 352 struct eqn_box *next; /* node sibling */ 353 struct eqn_box *parent; /* node sibling */ 354 char *text; /* text (or NULL) */ 355 char *left; 356 char *right; 357 enum eqn_post pos; /* position of next box */ 358 enum eqn_markt mark; /* a mark about the box */ 359 enum eqn_fontt font; /* font of box */ 360 enum eqn_pilet pile; /* equation piling */ 361 }; 362 363 /* 364 * An equation consists of a tree of expressions starting at a given 365 * line and position. 366 */ 367 struct eqn { 368 char *name; /* identifier (or NULL) */ 369 struct eqn_box *root; /* root mathematical expression */ 370 int ln; /* invocation line */ 371 int pos; /* invocation position */ 372 }; 373 374 /* 375 * Parse options. 376 */ 377 #define MPARSE_MDOC 1 /* assume -mdoc */ 378 #define MPARSE_MAN 2 /* assume -man */ 379 #define MPARSE_SO 4 /* honour .so requests */ 380 #define MPARSE_QUICK 8 /* abort the parse early */ 381 382 enum mandoc_esc { 383 ESCAPE_ERROR = 0, /* bail! unparsable escape */ 384 ESCAPE_IGNORE, /* escape to be ignored */ 385 ESCAPE_SPECIAL, /* a regular special character */ 386 ESCAPE_FONT, /* a generic font mode */ 387 ESCAPE_FONTBOLD, /* bold font mode */ 388 ESCAPE_FONTITALIC, /* italic font mode */ 389 ESCAPE_FONTBI, /* bold italic font mode */ 390 ESCAPE_FONTROMAN, /* roman font mode */ 391 ESCAPE_FONTPREV, /* previous font mode */ 392 ESCAPE_NUMBERED, /* a numbered glyph */ 393 ESCAPE_UNICODE, /* a unicode codepoint */ 394 ESCAPE_NOSPACE, /* suppress space if the last on a line */ 395 ESCAPE_SKIPCHAR /* skip the next character */ 396 }; 397 398 typedef void (*mandocmsg)(enum mandocerr, enum mandoclevel, 399 const char *, int, int, const char *); 400 401 struct mparse; 402 struct mchars; 403 struct mdoc; 404 struct man; 405 406 __BEGIN_DECLS 407 408 enum mandoc_esc mandoc_escape(const char **, const char **, int *); 409 struct mchars *mchars_alloc(void); 410 void mchars_free(struct mchars *); 411 char mchars_num2char(const char *, size_t); 412 int mchars_num2uc(const char *, size_t); 413 int mchars_spec2cp(const struct mchars *, 414 const char *, size_t); 415 const char *mchars_spec2str(const struct mchars *, 416 const char *, size_t, size_t *); 417 struct mparse *mparse_alloc(int, enum mandoclevel, mandocmsg, char *); 418 void mparse_free(struct mparse *); 419 void mparse_keep(struct mparse *); 420 enum mandoclevel mparse_readfd(struct mparse *, int, const char *); 421 void mparse_reset(struct mparse *); 422 void mparse_result(struct mparse *, 423 struct mdoc **, struct man **, char **); 424 const char *mparse_getkeep(const struct mparse *); 425 const char *mparse_strerror(enum mandocerr); 426 const char *mparse_strlevel(enum mandoclevel); 427 428 __END_DECLS 429 430 #endif /*!MANDOC_H*/ 431