mandoc.h

   1 /*      $Id: mandoc.h,v 1.112 2013/12/30 18:30:32 schwarze Exp $ */
   2 /*
   3  * Copyright (c) 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
   4  * Copyright (c) 2012, 2013 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_NOTMANUAL, /* manual isn't really a manual */
 156         MANDOCERR_COLUMNS, /* column syntax is inconsistent */
 157         MANDOCERR_BADDISP, /* NOT IMPLEMENTED: .Bd -file */
 158         MANDOCERR_SYNTARGVCOUNT, /* argument count wrong, violates syntax */
 159         MANDOCERR_SYNTCHILD, /* child violates parent syntax */
 160         MANDOCERR_SYNTARGCOUNT, /* argument count wrong, violates syntax */
 161         MANDOCERR_SOPATH, /* NOT IMPLEMENTED: .so with absolute path or ".." */
 162         MANDOCERR_NODOCBODY, /* no document body */
 163         MANDOCERR_NODOCPROLOG, /* no document prologue */
 164         MANDOCERR_MEM, /* static buffer exhausted */
 165         MANDOCERR_MAX
 166 };
 167
 168 struct  tbl_opts {
 169         char              tab; /* cell-separator */
 170         char              decimal; /* decimal point */
 171         int               linesize;
 172         int               opts;
 173 #define TBL_OPT_CENTRE   (1 << 0)
 174 #define TBL_OPT_EXPAND   (1 << 1)
 175 #define TBL_OPT_BOX      (1 << 2)
 176 #define TBL_OPT_DBOX     (1 << 3)
 177 #define TBL_OPT_ALLBOX   (1 << 4)
 178 #define TBL_OPT_NOKEEP   (1 << 5)
 179 #define TBL_OPT_NOSPACE  (1 << 6)
 180         int               cols; /* number of columns */
 181 };
 182
 183 /*
 184  * The head of a table specifies all of its columns.  When formatting a
 185  * tbl_span, iterate over these and plug in data from the tbl_span when
 186  * appropriate, using tbl_cell as a guide to placement.
 187  */
 188 struct  tbl_head {
 189         int               ident; /* 0 <= unique id < cols */
 190         int               vert; /* width of preceding vertical line */
 191         struct tbl_head  *next;
 192         struct tbl_head  *prev;
 193 };
 194
 195 enum    tbl_cellt {
 196         TBL_CELL_CENTRE, /* c, C */
 197         TBL_CELL_RIGHT, /* r, R */
 198         TBL_CELL_LEFT, /* l, L */
 199         TBL_CELL_NUMBER, /* n, N */
 200         TBL_CELL_SPAN, /* s, S */
 201         TBL_CELL_LONG, /* a, A */
 202         TBL_CELL_DOWN, /* ^ */
 203         TBL_CELL_HORIZ, /* _, - */
 204         TBL_CELL_DHORIZ, /* = */
 205         TBL_CELL_MAX
 206 };
 207
 208 /*
 209  * A cell in a layout row.
 210  */
 211 struct  tbl_cell {
 212         struct tbl_cell  *next;
 213         int               vert; /* width of preceding vertical line */
 214         enum tbl_cellt    pos;
 215         size_t            spacing;
 216         int               flags;
 217 #define TBL_CELL_TALIGN  (1 << 0) /* t, T */
 218 #define TBL_CELL_BALIGN  (1 << 1) /* d, D */
 219 #define TBL_CELL_BOLD    (1 << 2) /* fB, B, b */
 220 #define TBL_CELL_ITALIC  (1 << 3) /* fI, I, i */
 221 #define TBL_CELL_EQUAL   (1 << 4) /* e, E */
 222 #define TBL_CELL_UP      (1 << 5) /* u, U */
 223 #define TBL_CELL_WIGN    (1 << 6) /* z, Z */
 224         struct tbl_head  *head;
 225 };
 226
 227 /*
 228  * A layout row.
 229  */
 230 struct  tbl_row {
 231         struct tbl_row   *next;
 232         struct tbl_cell  *first;
 233         struct tbl_cell  *last;
 234 };
 235
 236 enum    tbl_datt {
 237         TBL_DATA_NONE, /* has no data */
 238         TBL_DATA_DATA, /* consists of data/string */
 239         TBL_DATA_HORIZ, /* horizontal line */
 240         TBL_DATA_DHORIZ, /* double-horizontal line */
 241         TBL_DATA_NHORIZ, /* squeezed horizontal line */
 242         TBL_DATA_NDHORIZ /* squeezed double-horizontal line */
 243 };
 244
 245 /*
 246  * A cell within a row of data.  The "string" field contains the actual
 247  * string value that's in the cell.  The rest is layout.
 248  */
 249 struct  tbl_dat {
 250         struct tbl_cell  *layout; /* layout cell */
 251         int               spans; /* how many spans follow */
 252         struct tbl_dat   *next;
 253         char             *string; /* data (NULL if not TBL_DATA_DATA) */
 254         enum tbl_datt     pos;
 255 };
 256
 257 enum    tbl_spant {
 258         TBL_SPAN_DATA, /* span consists of data */
 259         TBL_SPAN_HORIZ, /* span is horizontal line */
 260         TBL_SPAN_DHORIZ /* span is double horizontal line */
 261 };
 262
 263 /*
 264  * A row of data in a table.
 265  */
 266 struct  tbl_span {
 267         struct tbl_opts  *opts;
 268         struct tbl_head  *head;
 269         struct tbl_row   *layout; /* layout row */
 270         struct tbl_dat   *first;
 271         struct tbl_dat   *last;
 272         int               line; /* parse line */
 273         int               flags;
 274 #define TBL_SPAN_FIRST   (1 << 0)
 275 #define TBL_SPAN_LAST    (1 << 1)
 276         enum tbl_spant    pos;
 277         struct tbl_span  *next;
 278 };
 279
 280 enum    eqn_boxt {
 281         EQN_ROOT, /* root of parse tree */
 282         EQN_TEXT, /* text (number, variable, whatever) */
 283         EQN_SUBEXPR, /* nested `eqn' subexpression */
 284         EQN_LIST, /* subexpressions list */
 285         EQN_MATRIX /* matrix subexpression */
 286 };
 287
 288 enum    eqn_markt {
 289         EQNMARK_NONE = 0,
 290         EQNMARK_DOT,
 291         EQNMARK_DOTDOT,
 292         EQNMARK_HAT,
 293         EQNMARK_TILDE,
 294         EQNMARK_VEC,
 295         EQNMARK_DYAD,
 296         EQNMARK_BAR,
 297         EQNMARK_UNDER,
 298         EQNMARK__MAX
 299 };
 300
 301 enum    eqn_fontt {
 302         EQNFONT_NONE = 0,
 303         EQNFONT_ROMAN,
 304         EQNFONT_BOLD,
 305         EQNFONT_FAT,
 306         EQNFONT_ITALIC,
 307         EQNFONT__MAX
 308 };
 309
 310 enum    eqn_post {
 311         EQNPOS_NONE = 0,
 312         EQNPOS_OVER,
 313         EQNPOS_SUP,
 314         EQNPOS_SUB,
 315         EQNPOS_TO,
 316         EQNPOS_FROM,
 317         EQNPOS__MAX
 318 };
 319
 320 enum    eqn_pilet {
 321         EQNPILE_NONE = 0,
 322         EQNPILE_PILE,
 323         EQNPILE_CPILE,
 324         EQNPILE_RPILE,
 325         EQNPILE_LPILE,
 326         EQNPILE_COL,
 327         EQNPILE_CCOL,
 328         EQNPILE_RCOL,
 329         EQNPILE_LCOL,
 330         EQNPILE__MAX
 331 };
 332
 333  /*
 334  * A "box" is a parsed mathematical expression as defined by the eqn.7
 335  * grammar.
 336  */
 337 struct  eqn_box {
 338         int               size; /* font size of expression */
 339 #define EQN_DEFSIZE       INT_MIN
 340         enum eqn_boxt     type; /* type of node */
 341         struct eqn_box   *first; /* first child node */
 342         struct eqn_box   *last; /* last child node */
 343         struct eqn_box   *next; /* node sibling */
 344         struct eqn_box   *parent; /* node sibling */
 345         char             *text; /* text (or NULL) */
 346         char             *left;
 347         char             *right;
 348         enum eqn_post     pos; /* position of next box */
 349         enum eqn_markt    mark; /* a mark about the box */
 350         enum eqn_fontt    font; /* font of box */
 351         enum eqn_pilet    pile; /* equation piling */
 352 };
 353
 354 /*
 355  * An equation consists of a tree of expressions starting at a given
 356  * line and position.
 357  */
 358 struct  eqn {
 359         char             *name; /* identifier (or NULL) */
 360         struct eqn_box   *root; /* root mathematical expression */
 361         int               ln; /* invocation line */
 362         int               pos; /* invocation position */
 363 };
 364
 365 /*
 366  * The type of parse sequence.  This value is usually passed via the
 367  * mandoc(1) command line of -man and -mdoc.  It's almost exclusively
 368  * -mandoc but the others have been retained for compatibility.
 369  */
 370 enum    mparset {
 371         MPARSE_AUTO, /* magically determine the document type */
 372         MPARSE_MDOC, /* assume -mdoc */
 373         MPARSE_MAN /* assume -man */
 374 };
 375
 376 enum    mandoc_esc {
 377         ESCAPE_ERROR = 0, /* bail! unparsable escape */
 378         ESCAPE_IGNORE, /* escape to be ignored */
 379         ESCAPE_SPECIAL, /* a regular special character */
 380         ESCAPE_FONT, /* a generic font mode */
 381         ESCAPE_FONTBOLD, /* bold font mode */
 382         ESCAPE_FONTITALIC, /* italic font mode */
 383         ESCAPE_FONTBI, /* bold italic font mode */
 384         ESCAPE_FONTROMAN, /* roman font mode */
 385         ESCAPE_FONTPREV, /* previous font mode */
 386         ESCAPE_NUMBERED, /* a numbered glyph */
 387         ESCAPE_UNICODE, /* a unicode codepoint */
 388         ESCAPE_NOSPACE, /* suppress space if the last on a line */
 389         ESCAPE_SKIPCHAR /* skip the next character */
 390 };
 391
 392 typedef void    (*mandocmsg)(enum mandocerr, enum mandoclevel,
 393                         const char *, int, int, const char *);
 394
 395 struct  mparse;
 396 struct  mchars;
 397 struct  mdoc;
 398 struct  man;
 399
 400 __BEGIN_DECLS
 401
 402 void             *mandoc_calloc(size_t, size_t);
 403 enum mandoc_esc   mandoc_escape(const char **, const char **, int *);
 404 void             *mandoc_malloc(size_t);
 405 void             *mandoc_realloc(void *, size_t);
 406 char             *mandoc_strdup(const char *);
 407 char             *mandoc_strndup(const char *, size_t);
 408 struct mchars    *mchars_alloc(void);
 409 void              mchars_free(struct mchars *);
 410 char              mchars_num2char(const char *, size_t);
 411 int               mchars_num2uc(const char *, size_t);
 412 int               mchars_spec2cp(const struct mchars *,
 413                         const char *, size_t);
 414 const char       *mchars_spec2str(const struct mchars *,
 415                         const char *, size_t, size_t *);
 416 struct mparse    *mparse_alloc(enum mparset, enum mandoclevel,
 417                         mandocmsg, void *, char *);
 418 void              mparse_free(struct mparse *);
 419 void              mparse_keep(struct mparse *);
 420 enum mandoclevel  mparse_readfd(struct mparse *, int, const char *);
 421 enum mandoclevel  mparse_readmem(struct mparse *, const void *, size_t,
 422                         const char *);
 423 void              mparse_reset(struct mparse *);
 424 void              mparse_result(struct mparse *,
 425                         struct mdoc **, struct man **);
 426 const char       *mparse_getkeep(const struct mparse *);
 427 const char       *mparse_strerror(enum mandocerr);
 428 const char       *mparse_strlevel(enum mandoclevel);
 429
 430 __END_DECLS
 431
 432 #endif /*!MANDOC_H*/