mdoc.h

   1 /*      $Id: mdoc.h,v 1.105 2010/10/04 07:01:02 kristaps Exp $ */
   2 /*
   3  * Copyright (c) 2008, 2009, 2010 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 MDOC_H
  18 #define MDOC_H
  19
  20 /*
  21  * What follows is a list of ALL possible macros.
  22  */
  23 enum    mdoct {
  24         MDOC_Ap = 0,
  25         MDOC_Dd,
  26         MDOC_Dt,
  27         MDOC_Os,
  28         MDOC_Sh,
  29         MDOC_Ss,
  30         MDOC_Pp,
  31         MDOC_D1,
  32         MDOC_Dl,
  33         MDOC_Bd,
  34         MDOC_Ed,
  35         MDOC_Bl,
  36         MDOC_El,
  37         MDOC_It,
  38         MDOC_Ad,
  39         MDOC_An,
  40         MDOC_Ar,
  41         MDOC_Cd,
  42         MDOC_Cm,
  43         MDOC_Dv,
  44         MDOC_Er,
  45         MDOC_Ev,
  46         MDOC_Ex,
  47         MDOC_Fa,
  48         MDOC_Fd,
  49         MDOC_Fl,
  50         MDOC_Fn,
  51         MDOC_Ft,
  52         MDOC_Ic,
  53         MDOC_In,
  54         MDOC_Li,
  55         MDOC_Nd,
  56         MDOC_Nm,
  57         MDOC_Op,
  58         MDOC_Ot,
  59         MDOC_Pa,
  60         MDOC_Rv,
  61         MDOC_St,
  62         MDOC_Va,
  63         MDOC_Vt,
  64         MDOC_Xr,
  65         MDOC__A,
  66         MDOC__B,
  67         MDOC__D,
  68         MDOC__I,
  69         MDOC__J,
  70         MDOC__N,
  71         MDOC__O,
  72         MDOC__P,
  73         MDOC__R,
  74         MDOC__T,
  75         MDOC__V,
  76         MDOC_Ac,
  77         MDOC_Ao,
  78         MDOC_Aq,
  79         MDOC_At,
  80         MDOC_Bc,
  81         MDOC_Bf,
  82         MDOC_Bo,
  83         MDOC_Bq,
  84         MDOC_Bsx,
  85         MDOC_Bx,
  86         MDOC_Db,
  87         MDOC_Dc,
  88         MDOC_Do,
  89         MDOC_Dq,
  90         MDOC_Ec,
  91         MDOC_Ef,
  92         MDOC_Em,
  93         MDOC_Eo,
  94         MDOC_Fx,
  95         MDOC_Ms,
  96         MDOC_No,
  97         MDOC_Ns,
  98         MDOC_Nx,
  99         MDOC_Ox,
 100         MDOC_Pc,
 101         MDOC_Pf,
 102         MDOC_Po,
 103         MDOC_Pq,
 104         MDOC_Qc,
 105         MDOC_Ql,
 106         MDOC_Qo,
 107         MDOC_Qq,
 108         MDOC_Re,
 109         MDOC_Rs,
 110         MDOC_Sc,
 111         MDOC_So,
 112         MDOC_Sq,
 113         MDOC_Sm,
 114         MDOC_Sx,
 115         MDOC_Sy,
 116         MDOC_Tn,
 117         MDOC_Ux,
 118         MDOC_Xc,
 119         MDOC_Xo,
 120         MDOC_Fo,
 121         MDOC_Fc,
 122         MDOC_Oo,
 123         MDOC_Oc,
 124         MDOC_Bk,
 125         MDOC_Ek,
 126         MDOC_Bt,
 127         MDOC_Hf,
 128         MDOC_Fr,
 129         MDOC_Ud,
 130         MDOC_Lb,
 131         MDOC_Lp,
 132         MDOC_Lk,
 133         MDOC_Mt,
 134         MDOC_Brq,
 135         MDOC_Bro,
 136         MDOC_Brc,
 137         MDOC__C,
 138         MDOC_Es,
 139         MDOC_En,
 140         MDOC_Dx,
 141         MDOC__Q,
 142         MDOC_br,
 143         MDOC_sp,
 144         MDOC__U,
 145         MDOC_Ta,
 146         MDOC_MAX
 147 };
 148
 149 /*
 150  * What follows is a list of ALL possible macro arguments.
 151  */
 152 enum    mdocargt {
 153         MDOC_Split,
 154         MDOC_Nosplit,
 155         MDOC_Ragged,
 156         MDOC_Unfilled,
 157         MDOC_Literal,
 158         MDOC_File,
 159         MDOC_Offset,
 160         MDOC_Bullet,
 161         MDOC_Dash,
 162         MDOC_Hyphen,
 163         MDOC_Item,
 164         MDOC_Enum,
 165         MDOC_Tag,
 166         MDOC_Diag,
 167         MDOC_Hang,
 168         MDOC_Ohang,
 169         MDOC_Inset,
 170         MDOC_Column,
 171         MDOC_Width,
 172         MDOC_Compact,
 173         MDOC_Std,
 174         MDOC_Filled,
 175         MDOC_Words,
 176         MDOC_Emphasis,
 177         MDOC_Symbolic,
 178         MDOC_Nested,
 179         MDOC_Centred,
 180         MDOC_ARG_MAX
 181 };
 182
 183 /*
 184  * Type of a syntax node.
 185  */
 186 enum    mdoc_type {
 187         MDOC_TEXT,
 188         MDOC_ELEM,
 189         MDOC_HEAD,
 190         MDOC_TAIL,
 191         MDOC_BODY,
 192         MDOC_BLOCK,
 193         MDOC_ROOT
 194 };
 195
 196 /*
 197  * Section (named/unnamed) of `Sh'.   Note that these appear in the
 198  * conventional order imposed by mdoc.7.
 199  */
 200 enum    mdoc_sec {
 201         SEC_NONE = 0, /* No section, yet. */
 202         SEC_NAME,
 203         SEC_LIBRARY,
 204         SEC_SYNOPSIS,
 205         SEC_DESCRIPTION,
 206         SEC_IMPLEMENTATION,
 207         SEC_RETURN_VALUES,
 208         SEC_ENVIRONMENT,
 209         SEC_FILES,
 210         SEC_EXIT_STATUS,
 211         SEC_EXAMPLES,
 212         SEC_DIAGNOSTICS,
 213         SEC_COMPATIBILITY,
 214         SEC_ERRORS,
 215         SEC_SEE_ALSO,
 216         SEC_STANDARDS,
 217         SEC_HISTORY,
 218         SEC_AUTHORS,
 219         SEC_CAVEATS,
 220         SEC_BUGS,
 221         SEC_SECURITY,
 222         SEC_CUSTOM, /* User-defined. */
 223         SEC__MAX
 224 };
 225
 226 /*
 227  * Information from prologue.
 228  */
 229 struct  mdoc_meta {
 230         char             *msec; /* `Dt' section (1, 3p, etc.) */
 231         char             *vol; /* `Dt' volume (implied) */
 232         char             *arch; /* `Dt' arch (i386, etc.) */
 233         time_t            date; /* `Dd' normalised date */
 234         char             *title; /* `Dt' title (FOO, etc.) */
 235         char             *os; /* `Os' system (OpenBSD, etc.) */
 236         char             *name; /* leading `Nm' name */
 237 };
 238
 239 /*
 240  * An argument to a macro (multiple values = `-column xxx yyy').
 241  */
 242 struct  mdoc_argv {
 243         enum mdocargt     arg; /* type of argument */
 244         int               line;
 245         int               pos;
 246         size_t            sz; /* elements in "value" */
 247         char            **value; /* argument strings */
 248 };
 249
 250 /*
 251  * Reference-counted macro arguments.  These are refcounted because
 252  * blocks have multiple instances of the same arguments spread across
 253  * the HEAD, BODY, TAIL, and BLOCK node types.
 254  */
 255 struct  mdoc_arg {
 256         size_t            argc;
 257         struct mdoc_argv *argv;
 258         unsigned int      refcnt;
 259 };
 260
 261 /*
 262  * Indicates that a BODY's formatting has ended, but the scope is still
 263  * open.  Used for syntax-broken blocks.
 264  */
 265 enum    mdoc_endbody {
 266         ENDBODY_NOT = 0,
 267         ENDBODY_SPACE, /* is broken: append a space */
 268         ENDBODY_NOSPACE /* is broken: don't append a space */
 269 };
 270
 271 /*
 272  * Normalised `Bl' list type.
 273  */
 274 enum    mdoc_list {
 275         LIST__NONE = 0,
 276         LIST_bullet,
 277         LIST_column,
 278         LIST_dash,
 279         LIST_diag,
 280         LIST_enum,
 281         LIST_hang,
 282         LIST_hyphen,
 283         LIST_inset,
 284         LIST_item,
 285         LIST_ohang,
 286         LIST_tag
 287 };
 288
 289 /*
 290  * Normalised `Bd' display type.
 291  */
 292 enum    mdoc_disp {
 293         DISP__NONE = 0,
 294         DISP_centred,
 295         DISP_ragged,
 296         DISP_unfilled,
 297         DISP_filled,
 298         DISP_literal
 299 };
 300
 301 /*
 302  * Normalised `An' splitting argument.
 303  */
 304 enum    mdoc_auth {
 305         AUTH__NONE = 0,
 306         AUTH_split,
 307         AUTH_nosplit
 308 };
 309
 310 /*
 311  * Normalised `Bf' font type.
 312  */
 313 enum    mdoc_font {
 314         FONT__NONE = 0,
 315         FONT_Em,
 316         FONT_Li,
 317         FONT_Sy
 318 };
 319
 320 /*
 321  * Normalised arguments for `Bd'.
 322  */
 323 struct  mdoc_bd {
 324         const char       *offs; /* -offset */
 325         enum mdoc_disp    type; /* -ragged, etc. */
 326         int               comp; /* -compact */
 327 };
 328
 329 /*
 330  * Normalised arguments for `Bl'.
 331  */
 332 struct  mdoc_bl {
 333         const char       *width; /* -width */
 334         const char       *offs; /* -offset */
 335         enum mdoc_list    type; /* -tag, -enum, etc. */
 336         int               comp; /* -compact */
 337         size_t            ncols; /* -column arg count */
 338         const char      **cols; /* -column val ptr */
 339 };
 340
 341 /*
 342  * Normalised arguments for `Bf'.
 343  */
 344 struct  mdoc_bf {
 345         enum mdoc_font    font; /* font */
 346 };
 347
 348 /*
 349  * Normalised arguments for `An'.
 350  */
 351 struct  mdoc_an {
 352         enum mdoc_auth    auth; /* -split, etc. */
 353 };
 354
 355 /*
 356  * Consists of normalised node arguments.  These should be used instead
 357  * of iterating through the mdoc_arg pointers of a node: defaults are
 358  * provided, etc.
 359  */
 360 union mdoc_data {
 361         struct mdoc_an    An;
 362         struct mdoc_bd   *Bd;
 363         struct mdoc_bf   *Bf;
 364         struct mdoc_bl   *Bl;
 365 };
 366
 367 /*
 368  * Single node in tree-linked AST.
 369  */
 370 struct  mdoc_node {
 371         struct mdoc_node *parent; /* parent AST node */
 372         struct mdoc_node *child; /* first child AST node */
 373         struct mdoc_node *next; /* sibling AST node */
 374         struct mdoc_node *prev; /* prior sibling AST node */
 375         int               nchild; /* number children */
 376         int               line; /* parse line */
 377         int               pos; /* parse column */
 378         enum mdoct        tok; /* tok or MDOC__MAX if none */
 379         int               flags;
 380 #define MDOC_VALID       (1 << 0) /* has been validated */
 381 #define MDOC_ACTED       (1 << 1) /* has been acted upon */
 382 #define MDOC_EOS         (1 << 2) /* at sentence boundary */
 383 #define MDOC_LINE        (1 << 3) /* first macro/text on line */
 384 #define MDOC_SYNPRETTY   (1 << 4) /* SYNOPSIS-style formatting */
 385 #define MDOC_ENDED       (1 << 5) /* rendering has been ended */
 386         enum mdoc_type    type; /* AST node type */
 387         enum mdoc_sec     sec; /* current named section */
 388         /* FIXME: these can be union'd to shave a few bytes. */
 389         struct mdoc_arg  *args; /* BLOCK/ELEM */
 390         struct mdoc_node *pending; /* BLOCK */
 391         struct mdoc_node *head; /* BLOCK */
 392         struct mdoc_node *body; /* BLOCK */
 393         struct mdoc_node *tail; /* BLOCK */
 394         char             *string; /* TEXT */
 395         enum mdoc_endbody end; /* BODY */
 396         union mdoc_data   data;
 397 };
 398
 399 /*
 400  * Names of macros.  Index is enum mdoct.  Indexing into this returns
 401  * the normalised name, e.g., mdoc_macronames[MDOC_Sh] -> "Sh".
 402  */
 403 extern  const char *const *mdoc_macronames;
 404
 405 /*
 406  * Names of macro args.  Index is enum mdocargt.  Indexing into this
 407  * returns the normalised name, e.g., mdoc_argnames[MDOC_File] ->
 408  * "file".
 409  */
 410 extern  const char *const *mdoc_argnames;
 411
 412 __BEGIN_DECLS
 413
 414 struct  mdoc;
 415
 416 void              mdoc_free(struct mdoc *);
 417 struct  mdoc     *mdoc_alloc(struct regset *, void *, mandocmsg);
 418 void              mdoc_reset(struct mdoc *);
 419 int               mdoc_parseln(struct mdoc *, int, char *, int);
 420 const struct mdoc_node *mdoc_node(const struct mdoc *);
 421 const struct mdoc_meta *mdoc_meta(const struct mdoc *);
 422 int               mdoc_endparse(struct mdoc *);
 423
 424 __END_DECLS
 425
 426 #endif /*!MDOC_H*/