mandoc.h

   1 /*      $Id: mandoc.h,v 1.92 2011/07/23 12:01:54 kristaps 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_EQNNEST, /* too many nested equation defines */
 114         MANDOCERR_EQNNSCOPE, /* unexpected equation scope closure*/
 115         MANDOCERR_EQNSCOPE, /* equation scope open on exit */
 116         MANDOCERR_EQNBADSCOPE, /* overlapping equation scopes */
 117         MANDOCERR_EQNEOF, /* unexpected end of equation */
 118         MANDOCERR_EQNSYNT, /* equation syntax error */
 119
 120         /* related to tables */
 121         MANDOCERR_TBL, /* bad table syntax */
 122         MANDOCERR_TBLOPT, /* bad table option */
 123         MANDOCERR_TBLLAYOUT, /* bad table layout */
 124         MANDOCERR_TBLNOLAYOUT, /* no table layout cells specified */
 125         MANDOCERR_TBLNODATA, /* no table data cells specified */
 126         MANDOCERR_TBLIGNDATA, /* ignore data in cell */
 127         MANDOCERR_TBLBLOCK, /* data block still open */
 128         MANDOCERR_TBLEXTRADAT, /* ignoring extra data cells */
 129
 130         MANDOCERR_ROFFLOOP, /* input stack limit exceeded, infinite loop? */
 131         MANDOCERR_BADCHAR, /* skipping bad character */
 132         MANDOCERR_NAMESC, /* escaped character not allowed in a name */
 133         MANDOCERR_NOTEXT, /* skipping text before the first section header */
 134         MANDOCERR_MACRO, /* skipping unknown macro */
 135         MANDOCERR_REQUEST, /* NOT IMPLEMENTED: skipping request */
 136         MANDOCERR_ARGCOUNT, /* argument count wrong */
 137         MANDOCERR_NOSCOPE, /* skipping end of block that is not open */
 138         MANDOCERR_SCOPEBROKEN, /* missing end of block */
 139         MANDOCERR_SCOPEEXIT, /* scope open on exit */
 140         MANDOCERR_UNAME, /* uname(3) system call failed */
 141         /* FIXME: merge following with MANDOCERR_ARGCOUNT */
 142         MANDOCERR_NOARGS, /* macro requires line argument(s) */
 143         MANDOCERR_NOBODY, /* macro requires body argument(s) */
 144         MANDOCERR_NOARGV, /* macro requires argument(s) */
 145         MANDOCERR_LISTTYPE, /* missing list type */
 146         MANDOCERR_ARGSLOST, /* line argument(s) will be lost */
 147         MANDOCERR_BODYLOST, /* body argument(s) will be lost */
 148
 149         MANDOCERR_FATAL, /* ===== start of fatal errors ===== */
 150
 151         MANDOCERR_NOTMANUAL, /* manual isn't really a manual */
 152         MANDOCERR_COLUMNS, /* column syntax is inconsistent */
 153         MANDOCERR_BADDISP, /* NOT IMPLEMENTED: .Bd -file */
 154         MANDOCERR_SYNTLINESCOPE, /* line scope broken, syntax violated */
 155         MANDOCERR_SYNTARGVCOUNT, /* argument count wrong, violates syntax */
 156         MANDOCERR_SYNTCHILD, /* child violates parent syntax */
 157         MANDOCERR_SYNTARGCOUNT, /* argument count wrong, violates syntax */
 158         MANDOCERR_SOPATH, /* NOT IMPLEMENTED: .so with absolute path or ".." */
 159         MANDOCERR_NODOCBODY, /* no document body */
 160         MANDOCERR_NODOCPROLOG, /* no document prologue */
 161         MANDOCERR_MEM, /* static buffer exhausted */
 162         MANDOCERR_MAX
 163 };
 164
 165 struct  tbl {
 166         char              tab; /* cell-separator */
 167         char              decimal; /* decimal point */
 168         int               linesize;
 169         int               opts;
 170 #define TBL_OPT_CENTRE   (1 << 0)
 171 #define TBL_OPT_EXPAND   (1 << 1)
 172 #define TBL_OPT_BOX      (1 << 2)
 173 #define TBL_OPT_DBOX     (1 << 3)
 174 #define TBL_OPT_ALLBOX   (1 << 4)
 175 #define TBL_OPT_NOKEEP   (1 << 5)
 176 #define TBL_OPT_NOSPACE  (1 << 6)
 177         int               cols; /* number of columns */
 178 };
 179
 180 enum    tbl_headt {
 181         TBL_HEAD_DATA, /* plug in data from tbl_dat */
 182         TBL_HEAD_VERT, /* vertical spacer */
 183         TBL_HEAD_DVERT  /* double-vertical spacer */
 184 };
 185
 186 /*
 187  * The head of a table specifies all of its columns.  When formatting a
 188  * tbl_span, iterate over these and plug in data from the tbl_span when
 189  * appropriate, using tbl_cell as a guide to placement.
 190  */
 191 struct  tbl_head {
 192         enum tbl_headt    pos;
 193         int               ident; /* 0 <= unique id < cols */
 194         struct tbl_head  *next;
 195         struct tbl_head  *prev;
 196 };
 197
 198 enum    tbl_cellt {
 199         TBL_CELL_CENTRE, /* c, C */
 200         TBL_CELL_RIGHT, /* r, R */
 201         TBL_CELL_LEFT, /* l, L */
 202         TBL_CELL_NUMBER, /* n, N */
 203         TBL_CELL_SPAN, /* s, S */
 204         TBL_CELL_LONG, /* a, A */
 205         TBL_CELL_DOWN, /* ^ */
 206         TBL_CELL_HORIZ, /* _, - */
 207         TBL_CELL_DHORIZ, /* = */
 208         TBL_CELL_VERT, /* | */
 209         TBL_CELL_DVERT, /* || */
 210         TBL_CELL_MAX
 211 };
 212
 213 /*
 214  * A cell in a layout row.
 215  */
 216 struct  tbl_cell {
 217         struct tbl_cell  *next;
 218         enum tbl_cellt    pos;
 219         size_t            spacing;
 220         int               flags;
 221 #define TBL_CELL_TALIGN  (1 << 0) /* t, T */
 222 #define TBL_CELL_BALIGN  (1 << 1) /* d, D */
 223 #define TBL_CELL_BOLD    (1 << 2) /* fB, B, b */
 224 #define TBL_CELL_ITALIC  (1 << 3) /* fI, I, i */
 225 #define TBL_CELL_EQUAL   (1 << 4) /* e, E */
 226 #define TBL_CELL_UP      (1 << 5) /* u, U */
 227 #define TBL_CELL_WIGN    (1 << 6) /* z, Z */
 228         struct tbl_head  *head;
 229 };
 230
 231 /*
 232  * A layout row.
 233  */
 234 struct  tbl_row {
 235         struct tbl_row   *next;
 236         struct tbl_cell  *first;
 237         struct tbl_cell  *last;
 238 };
 239
 240 enum    tbl_datt {
 241         TBL_DATA_NONE, /* has no data */
 242         TBL_DATA_DATA, /* consists of data/string */
 243         TBL_DATA_HORIZ, /* horizontal line */
 244         TBL_DATA_DHORIZ, /* double-horizontal line */
 245         TBL_DATA_NHORIZ, /* squeezed horizontal line */
 246         TBL_DATA_NDHORIZ /* squeezed double-horizontal line */
 247 };
 248
 249 /*
 250  * A cell within a row of data.  The "string" field contains the actual
 251  * string value that's in the cell.  The rest is layout.
 252  */
 253 struct  tbl_dat {
 254         struct tbl_cell  *layout; /* layout cell */
 255         int               spans; /* how many spans follow */
 256         struct tbl_dat   *next;
 257         char             *string; /* data (NULL if not TBL_DATA_DATA) */
 258         enum tbl_datt     pos;
 259 };
 260
 261 enum    tbl_spant {
 262         TBL_SPAN_DATA, /* span consists of data */
 263         TBL_SPAN_HORIZ, /* span is horizontal line */
 264         TBL_SPAN_DHORIZ /* span is double horizontal line */
 265 };
 266
 267 /*
 268  * A row of data in a table.
 269  */
 270 struct  tbl_span {
 271         struct tbl       *tbl;
 272         struct tbl_head  *head;
 273         struct tbl_row   *layout; /* layout row */
 274         struct tbl_dat   *first;
 275         struct tbl_dat   *last;
 276         int               line; /* parse line */
 277         int               flags;
 278 #define TBL_SPAN_FIRST   (1 << 0)
 279 #define TBL_SPAN_LAST    (1 << 1)
 280         enum tbl_spant    pos;
 281         struct tbl_span  *next;
 282 };
 283
 284 enum    eqn_boxt {
 285         EQN_ROOT, /* root of parse tree */
 286         EQN_TEXT, /* text (number, variable, whatever) */
 287         EQN_SUBEXPR, /* nested `eqn' subexpression */
 288         EQN_LIST, /* subexpressions list */
 289         EQN_MATRIX /* matrix subexpression */
 290 };
 291
 292 enum    eqn_markt {
 293         EQNMARK_NONE = 0,
 294         EQNMARK_DOT,
 295         EQNMARK_DOTDOT,
 296         EQNMARK_HAT,
 297         EQNMARK_TILDE,
 298         EQNMARK_VEC,
 299         EQNMARK_DYAD,
 300         EQNMARK_BAR,
 301         EQNMARK_UNDER,
 302         EQNMARK__MAX
 303 };
 304
 305 enum    eqn_fontt {
 306         EQNFONT_NONE = 0,
 307         EQNFONT_ROMAN,
 308         EQNFONT_BOLD,
 309         EQNFONT_ITALIC,
 310         EQNFONT__MAX
 311 };
 312
 313 enum    eqn_post {
 314         EQNPOS_NONE = 0,
 315         EQNPOS_OVER,
 316         EQNPOS_SUP,
 317         EQNPOS_SUB,
 318         EQNPOS_TO,
 319         EQNPOS_FROM,
 320         EQNPOS__MAX
 321 };
 322
 323 enum    eqn_pilet {
 324         EQNPILE_NONE = 0,
 325         EQNPILE_CPILE,
 326         EQNPILE_RPILE,
 327         EQNPILE_LPILE,
 328         EQNPILE_CCOL,
 329         EQNPILE_RCOL,
 330         EQNPILE_LCOL,
 331         EQNPILE__MAX
 332 };
 333
 334  /*
 335  * A "box" is a parsed mathematical expression as defined by the eqn.7
 336  * grammar.
 337  */
 338 struct  eqn_box {
 339         int               size; /* font size of expression */
 340 #define EQN_DEFSIZE       INT_MIN
 341         enum eqn_boxt     type; /* type of node */
 342         struct eqn_box   *first; /* first child node */
 343         struct eqn_box   *last; /* last child node */
 344         struct eqn_box   *next; /* node sibling */
 345         struct eqn_box   *parent; /* node sibling */
 346         char             *text; /* text (or NULL) */
 347         char             *left;
 348         char             *right;
 349         enum eqn_post     pos; /* position of next box */
 350         enum eqn_markt    mark; /* a mark about the box */
 351         enum eqn_fontt    font; /* font of box */
 352         enum eqn_pilet    pile; /* equation piling */
 353 };
 354
 355 /*
 356  * An equation consists of a tree of expressions starting at a given
 357  * line and position.
 358  */
 359 struct  eqn {
 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_FONTROMAN, /* roman font mode */
 384         ESCAPE_FONTPREV, /* previous font mode */
 385         ESCAPE_NUMBERED, /* a numbered glyph */
 386         ESCAPE_UNICODE, /* a unicode codepoint */
 387         ESCAPE_NOSPACE /* suppress space if the last on a line */
 388 };
 389
 390 typedef void    (*mandocmsg)(enum mandocerr, enum mandoclevel,
 391                         const char *, int, int, const char *);
 392
 393 struct  mparse;
 394 struct  mchars;
 395 struct  mdoc;
 396 struct  man;
 397
 398 __BEGIN_DECLS
 399
 400 void              mparse_free(struct mparse *);
 401 void              mparse_reset(struct mparse *);
 402 struct mparse    *mparse_alloc(enum mparset,
 403                         enum mandoclevel, mandocmsg, void *);
 404 enum mandoclevel  mparse_readfd(struct mparse *, int, const char *);
 405 void              mparse_result(struct mparse *, struct mdoc **, struct man **);
 406 const char       *mparse_strerror(enum mandocerr);
 407 const char       *mparse_strlevel(enum mandoclevel);
 408
 409 void             *mandoc_calloc(size_t, size_t);
 410 void             *mandoc_malloc(size_t);
 411 void             *mandoc_realloc(void *, size_t);
 412 char             *mandoc_strdup(const char *);
 413 char             *mandoc_strndup(const char *, size_t);
 414
 415 enum mandoc_esc   mandoc_escape(const char **, const char **, int *);
 416
 417 struct mchars    *mchars_alloc(void);
 418 char              mchars_num2char(const char *, size_t);
 419 int               mchars_num2uc(const char *, size_t);
 420 const char       *mchars_spec2str(struct mchars *, const char *, size_t, size_t *);
 421 int               mchars_spec2cp(struct mchars *, const char *, size_t);
 422 void              mchars_free(struct mchars *);
 423
 424
 425 __END_DECLS
 426
 427 #endif /*!MANDOC_H*/