1 /*      $Id: mandoc.h,v 1.99 2012/02/16 20:51:31 joerg Exp $ */
   2 /*
   3  * Copyright (c) 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
   4  *
   5  * Permission to use, copy, modify, and distribute this software for any
   6  * purpose with or without fee is hereby granted, provided that the above
   7  * copyright notice and this permission notice appear in all copies.
   8  *
   9  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
  10  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
  11  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
  12  * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
  13  * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
  14  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
  15  * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
  16  */
  17 #ifndef MANDOC_H
  18 #define MANDOC_H
  19 
  20 #define ASCII_NBRSP      31  /* non-breaking space */
  21 #define ASCII_HYPH       30  /* breakable hyphen */
  22 
  23 /*
  24  * Status level.  This refers to both internal status (i.e., whilst
  25  * running, when warnings/errors are reported) and an indicator of a
  26  * threshold of when to halt (when said internal state exceeds the
  27  * threshold).
  28  */
  29 enum    mandoclevel {
  30         MANDOCLEVEL_OK = 0,
  31         MANDOCLEVEL_RESERVED,
  32         MANDOCLEVEL_WARNING, /* warnings: syntax, whitespace, etc. */
  33         MANDOCLEVEL_ERROR, /* input has been thrown away */
  34         MANDOCLEVEL_FATAL, /* input is borked */
  35         MANDOCLEVEL_BADARG, /* bad argument in invocation */
  36         MANDOCLEVEL_SYSERR, /* system error */
  37         MANDOCLEVEL_MAX
  38 };
  39 
  40 /*
  41  * All possible things that can go wrong within a parse, be it libroff,
  42  * libmdoc, or libman.
  43  */
  44 enum    mandocerr {
  45         MANDOCERR_OK,
  46 
  47         MANDOCERR_WARNING, /* ===== start of warnings ===== */
  48 
  49         /* related to the prologue */
  50         MANDOCERR_NOTITLE, /* no title in document */
  51         MANDOCERR_UPPERCASE, /* document title should be all caps */
  52         MANDOCERR_BADMSEC, /* unknown manual section */
  53         MANDOCERR_NODATE, /* date missing, using today's date */
  54         MANDOCERR_BADDATE, /* cannot parse date, using it verbatim */
  55         MANDOCERR_PROLOGOOO, /* prologue macros out of order */
  56         MANDOCERR_PROLOGREP, /* duplicate prologue macro */
  57         MANDOCERR_BADPROLOG, /* macro not allowed in prologue */
  58         MANDOCERR_BADBODY, /* macro not allowed in body */
  59 
  60         /* related to document structure */
  61         MANDOCERR_SO, /* .so is fragile, better use ln(1) */
  62         MANDOCERR_NAMESECFIRST, /* NAME section must come first */
  63         MANDOCERR_BADNAMESEC, /* bad NAME section contents */
  64         MANDOCERR_NONAME, /* manual name not yet set */
  65         MANDOCERR_SECOOO, /* sections out of conventional order */
  66         MANDOCERR_SECREP, /* duplicate section name */
  67         MANDOCERR_SECMSEC, /* section not in conventional manual section */
  68 
  69         /* related to macros and nesting */
  70         MANDOCERR_MACROOBS, /* skipping obsolete macro */
  71         MANDOCERR_IGNPAR, /* skipping paragraph macro */
  72         MANDOCERR_IGNNS, /* skipping no-space macro */
  73         MANDOCERR_SCOPENEST, /* blocks badly nested */
  74         MANDOCERR_CHILD, /* child violates parent syntax */
  75         MANDOCERR_NESTEDDISP, /* nested displays are not portable */
  76         MANDOCERR_SCOPEREP, /* already in literal mode */
  77         MANDOCERR_LINESCOPE, /* line scope broken */
  78 
  79         /* related to missing macro arguments */
  80         MANDOCERR_MACROEMPTY, /* skipping empty macro */
  81         MANDOCERR_ARGCWARN, /* argument count wrong */
  82         MANDOCERR_DISPTYPE, /* missing display type */
  83         MANDOCERR_LISTFIRST, /* list type must come first */
  84         MANDOCERR_NOWIDTHARG, /* tag lists require a width argument */
  85         MANDOCERR_FONTTYPE, /* missing font type */
  86         MANDOCERR_WNOSCOPE, /* skipping end of block that is not open */
  87 
  88         /* related to bad macro arguments */
  89         MANDOCERR_IGNARGV, /* skipping argument */
  90         MANDOCERR_ARGVREP, /* duplicate argument */
  91         MANDOCERR_DISPREP, /* duplicate display type */
  92         MANDOCERR_LISTREP, /* duplicate list type */
  93         MANDOCERR_BADATT, /* unknown AT&T UNIX version */
  94         MANDOCERR_BADBOOL, /* bad Boolean value */
  95         MANDOCERR_BADFONT, /* unknown font */
  96         MANDOCERR_BADSTANDARD, /* unknown standard specifier */
  97         MANDOCERR_BADWIDTH, /* bad width argument */
  98 
  99         /* related to plain text */
 100         MANDOCERR_NOBLANKLN, /* blank line in non-literal context */
 101         MANDOCERR_BADTAB, /* tab in non-literal context */
 102         MANDOCERR_EOLNSPACE, /* end of line whitespace */
 103         MANDOCERR_BADCOMMENT, /* bad comment style */
 104         MANDOCERR_BADESCAPE, /* unknown escape sequence */
 105         MANDOCERR_BADQUOTE, /* unterminated quoted string */
 106 
 107         /* related to equations */
 108         MANDOCERR_EQNQUOTE, /* unexpected literal in equation */
 109 
 110         MANDOCERR_ERROR, /* ===== start of errors ===== */
 111 
 112         /* related to equations */
 113         MANDOCERR_EQNNSCOPE, /* unexpected equation scope closure*/
 114         MANDOCERR_EQNSCOPE, /* equation scope open on exit */
 115         MANDOCERR_EQNBADSCOPE, /* overlapping equation scopes */
 116         MANDOCERR_EQNEOF, /* unexpected end of equation */
 117         MANDOCERR_EQNSYNT, /* equation syntax error */
 118 
 119         /* related to tables */
 120         MANDOCERR_TBL, /* bad table syntax */
 121         MANDOCERR_TBLOPT, /* bad table option */
 122         MANDOCERR_TBLLAYOUT, /* bad table layout */
 123         MANDOCERR_TBLNOLAYOUT, /* no table layout cells specified */
 124         MANDOCERR_TBLNODATA, /* no table data cells specified */
 125         MANDOCERR_TBLIGNDATA, /* ignore data in cell */
 126         MANDOCERR_TBLBLOCK, /* data block still open */
 127         MANDOCERR_TBLEXTRADAT, /* ignoring extra data cells */
 128 
 129         MANDOCERR_ROFFLOOP, /* input stack limit exceeded, infinite loop? */
 130         MANDOCERR_BADCHAR, /* skipping bad character */
 131         MANDOCERR_NAMESC, /* escaped character not allowed in a name */
 132         MANDOCERR_NOTEXT, /* skipping text before the first section header */
 133         MANDOCERR_MACRO, /* skipping unknown macro */
 134         MANDOCERR_REQUEST, /* NOT IMPLEMENTED: skipping request */
 135         MANDOCERR_ARGCOUNT, /* argument count wrong */
 136         MANDOCERR_NOSCOPE, /* skipping end of block that is not open */
 137         MANDOCERR_SCOPEBROKEN, /* missing end of block */
 138         MANDOCERR_SCOPEEXIT, /* scope open on exit */
 139         MANDOCERR_UNAME, /* uname(3) system call failed */
 140         /* FIXME: merge following with MANDOCERR_ARGCOUNT */
 141         MANDOCERR_NOARGS, /* macro requires line argument(s) */
 142         MANDOCERR_NOBODY, /* macro requires body argument(s) */
 143         MANDOCERR_NOARGV, /* macro requires argument(s) */
 144         MANDOCERR_LISTTYPE, /* missing list type */
 145         MANDOCERR_ARGSLOST, /* line argument(s) will be lost */
 146         MANDOCERR_BODYLOST, /* body argument(s) will be lost */
 147 
 148         MANDOCERR_FATAL, /* ===== start of fatal errors ===== */
 149 
 150         MANDOCERR_NOTMANUAL, /* manual isn't really a manual */
 151         MANDOCERR_COLUMNS, /* column syntax is inconsistent */
 152         MANDOCERR_BADDISP, /* NOT IMPLEMENTED: .Bd -file */
 153         MANDOCERR_SYNTARGVCOUNT, /* argument count wrong, violates syntax */
 154         MANDOCERR_SYNTCHILD, /* child violates parent syntax */
 155         MANDOCERR_SYNTARGCOUNT, /* argument count wrong, violates syntax */
 156         MANDOCERR_SOPATH, /* NOT IMPLEMENTED: .so with absolute path or ".." */
 157         MANDOCERR_NODOCBODY, /* no document body */
 158         MANDOCERR_NODOCPROLOG, /* no document prologue */
 159         MANDOCERR_MEM, /* static buffer exhausted */
 160         MANDOCERR_MAX
 161 };
 162 
 163 struct  tbl {
 164         char              tab; /* cell-separator */
 165         char              decimal; /* decimal point */
 166         int               linesize;
 167         int               opts;
 168 #define TBL_OPT_CENTRE   (1 << 0)
 169 #define TBL_OPT_EXPAND   (1 << 1)
 170 #define TBL_OPT_BOX      (1 << 2)
 171 #define TBL_OPT_DBOX     (1 << 3)
 172 #define TBL_OPT_ALLBOX   (1 << 4)
 173 #define TBL_OPT_NOKEEP   (1 << 5)
 174 #define TBL_OPT_NOSPACE  (1 << 6)
 175         int               cols; /* number of columns */
 176 };
 177 
 178 enum    tbl_headt {
 179         TBL_HEAD_DATA, /* plug in data from tbl_dat */
 180         TBL_HEAD_VERT, /* vertical spacer */
 181         TBL_HEAD_DVERT  /* double-vertical spacer */
 182 };
 183 
 184 /*
 185  * The head of a table specifies all of its columns.  When formatting a
 186  * tbl_span, iterate over these and plug in data from the tbl_span when
 187  * appropriate, using tbl_cell as a guide to placement.
 188  */
 189 struct  tbl_head {
 190         enum tbl_headt    pos;
 191         int               ident; /* 0 <= unique id < cols */
 192         struct tbl_head  *next;
 193         struct tbl_head  *prev;
 194 };
 195 
 196 enum    tbl_cellt {
 197         TBL_CELL_CENTRE, /* c, C */
 198         TBL_CELL_RIGHT, /* r, R */
 199         TBL_CELL_LEFT, /* l, L */
 200         TBL_CELL_NUMBER, /* n, N */
 201         TBL_CELL_SPAN, /* s, S */
 202         TBL_CELL_LONG, /* a, A */
 203         TBL_CELL_DOWN, /* ^ */
 204         TBL_CELL_HORIZ, /* _, - */
 205         TBL_CELL_DHORIZ, /* = */
 206         TBL_CELL_VERT, /* | */
 207         TBL_CELL_DVERT, /* || */
 208         TBL_CELL_MAX
 209 };
 210 
 211 /*
 212  * A cell in a layout row.
 213  */
 214 struct  tbl_cell {
 215         struct tbl_cell  *next;
 216         enum tbl_cellt    pos;
 217         size_t            spacing;
 218         int               flags;
 219 #define TBL_CELL_TALIGN  (1 << 0) /* t, T */
 220 #define TBL_CELL_BALIGN  (1 << 1) /* d, D */
 221 #define TBL_CELL_BOLD    (1 << 2) /* fB, B, b */
 222 #define TBL_CELL_ITALIC  (1 << 3) /* fI, I, i */
 223 #define TBL_CELL_EQUAL   (1 << 4) /* e, E */
 224 #define TBL_CELL_UP      (1 << 5) /* u, U */
 225 #define TBL_CELL_WIGN    (1 << 6) /* z, Z */
 226         struct tbl_head  *head;
 227 };
 228 
 229 /*
 230  * A layout row.
 231  */
 232 struct  tbl_row {
 233         struct tbl_row   *next;
 234         struct tbl_cell  *first;
 235         struct tbl_cell  *last;
 236 };
 237 
 238 enum    tbl_datt {
 239         TBL_DATA_NONE, /* has no data */
 240         TBL_DATA_DATA, /* consists of data/string */
 241         TBL_DATA_HORIZ, /* horizontal line */
 242         TBL_DATA_DHORIZ, /* double-horizontal line */
 243         TBL_DATA_NHORIZ, /* squeezed horizontal line */
 244         TBL_DATA_NDHORIZ /* squeezed double-horizontal line */
 245 };
 246 
 247 /*
 248  * A cell within a row of data.  The "string" field contains the actual
 249  * string value that's in the cell.  The rest is layout.
 250  */
 251 struct  tbl_dat {
 252         struct tbl_cell  *layout; /* layout cell */
 253         int               spans; /* how many spans follow */
 254         struct tbl_dat   *next;
 255         char             *string; /* data (NULL if not TBL_DATA_DATA) */
 256         enum tbl_datt     pos;
 257 };
 258 
 259 enum    tbl_spant {
 260         TBL_SPAN_DATA, /* span consists of data */
 261         TBL_SPAN_HORIZ, /* span is horizontal line */
 262         TBL_SPAN_DHORIZ /* span is double horizontal line */
 263 };
 264 
 265 /*
 266  * A row of data in a table.
 267  */
 268 struct  tbl_span {
 269         struct tbl       *tbl;
 270         struct tbl_head  *head;
 271         struct tbl_row   *layout; /* layout row */
 272         struct tbl_dat   *first;
 273         struct tbl_dat   *last;
 274         int               line; /* parse line */
 275         int               flags;
 276 #define TBL_SPAN_FIRST   (1 << 0)
 277 #define TBL_SPAN_LAST    (1 << 1)
 278         enum tbl_spant    pos;
 279         struct tbl_span  *next;
 280 };
 281 
 282 enum    eqn_boxt {
 283         EQN_ROOT, /* root of parse tree */
 284         EQN_TEXT, /* text (number, variable, whatever) */
 285         EQN_SUBEXPR, /* nested `eqn' subexpression */
 286         EQN_LIST, /* subexpressions list */
 287         EQN_MATRIX /* matrix subexpression */
 288 };
 289 
 290 enum    eqn_markt {
 291         EQNMARK_NONE = 0,
 292         EQNMARK_DOT,
 293         EQNMARK_DOTDOT,
 294         EQNMARK_HAT,
 295         EQNMARK_TILDE,
 296         EQNMARK_VEC,
 297         EQNMARK_DYAD,
 298         EQNMARK_BAR,
 299         EQNMARK_UNDER,
 300         EQNMARK__MAX
 301 };
 302 
 303 enum    eqn_fontt {
 304         EQNFONT_NONE = 0,
 305         EQNFONT_ROMAN,
 306         EQNFONT_BOLD,
 307         EQNFONT_FAT,
 308         EQNFONT_ITALIC,
 309         EQNFONT__MAX
 310 };
 311 
 312 enum    eqn_post {
 313         EQNPOS_NONE = 0,
 314         EQNPOS_OVER,
 315         EQNPOS_SUP,
 316         EQNPOS_SUB,
 317         EQNPOS_TO,
 318         EQNPOS_FROM,
 319         EQNPOS__MAX
 320 };
 321 
 322 enum    eqn_pilet {
 323         EQNPILE_NONE = 0,
 324         EQNPILE_PILE,
 325         EQNPILE_CPILE,
 326         EQNPILE_RPILE,
 327         EQNPILE_LPILE,
 328         EQNPILE_COL,
 329         EQNPILE_CCOL,
 330         EQNPILE_RCOL,
 331         EQNPILE_LCOL,
 332         EQNPILE__MAX
 333 };
 334 
 335  /*
 336  * A "box" is a parsed mathematical expression as defined by the eqn.7
 337  * grammar.
 338  */
 339 struct  eqn_box {
 340         int               size; /* font size of expression */
 341 #define EQN_DEFSIZE       INT_MIN
 342         enum eqn_boxt     type; /* type of node */
 343         struct eqn_box   *first; /* first child node */
 344         struct eqn_box   *last; /* last child node */
 345         struct eqn_box   *next; /* node sibling */
 346         struct eqn_box   *parent; /* node sibling */
 347         char             *text; /* text (or NULL) */
 348         char             *left;
 349         char             *right;
 350         enum eqn_post     pos; /* position of next box */
 351         enum eqn_markt    mark; /* a mark about the box */
 352         enum eqn_fontt    font; /* font of box */
 353         enum eqn_pilet    pile; /* equation piling */
 354 };
 355 
 356 /*
 357  * An equation consists of a tree of expressions starting at a given
 358  * line and position. 
 359  */
 360 struct  eqn {
 361         char             *name; /* identifier (or NULL) */
 362         struct eqn_box   *root; /* root mathematical expression */
 363         int               ln; /* invocation line */
 364         int               pos; /* invocation position */
 365 };
 366 
 367 /*
 368  * The type of parse sequence.  This value is usually passed via the
 369  * mandoc(1) command line of -man and -mdoc.  It's almost exclusively
 370  * -mandoc but the others have been retained for compatibility.
 371  */
 372 enum    mparset {
 373         MPARSE_AUTO, /* magically determine the document type */
 374         MPARSE_MDOC, /* assume -mdoc */
 375         MPARSE_MAN /* assume -man */
 376 };
 377 
 378 enum    mandoc_esc {
 379         ESCAPE_ERROR = 0, /* bail! unparsable escape */
 380         ESCAPE_IGNORE, /* escape to be ignored */
 381         ESCAPE_SPECIAL, /* a regular special character */
 382         ESCAPE_FONT, /* a generic font mode */
 383         ESCAPE_FONTBOLD, /* bold font mode */
 384         ESCAPE_FONTITALIC, /* italic font mode */
 385         ESCAPE_FONTROMAN, /* roman font mode */
 386         ESCAPE_FONTPREV, /* previous font mode */
 387         ESCAPE_NUMBERED, /* a numbered glyph */
 388         ESCAPE_UNICODE, /* a unicode codepoint */
 389         ESCAPE_NOSPACE /* suppress space if the last on a line */
 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, 
 417                         enum mandoclevel, mandocmsg, void *);
 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*/