mandoc.h

   1 /*      $Id: mandoc.h,v 1.123 2014/06/25 00:20:19 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_TH_MISSING, /* no TH macro in document */
  53         MANDOCERR_TITLE_CASE, /* document title should be all caps */
  54         MANDOCERR_MSEC_BAD, /* unknown manual section */
  55         MANDOCERR_ARCH_BAD, /* unknown manual volume or arch */
  56         MANDOCERR_DATE_MISSING, /* date missing, using today's date */
  57         MANDOCERR_DATE_BAD, /* cannot parse date, using it verbatim */
  58         MANDOCERR_PROLOG_ORDER, /* prologue macros out of order */
  59         MANDOCERR_PROLOG_REP, /* duplicate prologue macro */
  60         MANDOCERR_PROLOG_BAD, /* macro not allowed in prologue */
  61         MANDOCERR_PROLOG_ONLY, /* macro not allowed in body */
  62
  63         /* related to document structure */
  64         MANDOCERR_SO, /* .so is fragile, better use ln(1) */
  65         MANDOCERR_DOC_EMPTY, /* no document body */
  66         MANDOCERR_SEC_BEFORE, /* content before the first section header */
  67         MANDOCERR_NAMESECFIRST, /* NAME section must come first */
  68         MANDOCERR_BADNAMESEC, /* bad NAME section contents */
  69         MANDOCERR_SECOOO, /* sections out of conventional order */
  70         MANDOCERR_SECREP, /* duplicate section name */
  71         MANDOCERR_SECMSEC, /* section header suited to sections ... */
  72
  73         /* related to macros and nesting */
  74         MANDOCERR_MACROOBS, /* skipping obsolete macro */
  75         MANDOCERR_IGNPAR, /* skipping paragraph macro */
  76         MANDOCERR_MOVEPAR, /* moving paragraph macro out of list */
  77         MANDOCERR_IGNNS, /* skipping no-space macro */
  78         MANDOCERR_SCOPENEST, /* blocks badly nested */
  79         MANDOCERR_CHILD, /* child violates parent syntax */
  80         MANDOCERR_NESTEDDISP, /* nested displays are not portable */
  81         MANDOCERR_SCOPEREP, /* already in literal mode */
  82         MANDOCERR_LINESCOPE, /* line scope broken */
  83
  84         /* related to missing macro arguments */
  85         MANDOCERR_MACROEMPTY, /* skipping empty macro */
  86         MANDOCERR_ARGCWARN, /* argument count wrong */
  87         MANDOCERR_DISPTYPE, /* missing display type */
  88         MANDOCERR_LISTFIRST, /* list type must come first */
  89         MANDOCERR_NOWIDTHARG, /* tag lists require a width argument */
  90         MANDOCERR_FONTTYPE, /* missing font type */
  91         MANDOCERR_WNOSCOPE, /* skipping end of block that is not open */
  92
  93         /* related to bad macro arguments */
  94         MANDOCERR_IGNARGV, /* skipping argument */
  95         MANDOCERR_ARGVREP, /* duplicate argument */
  96         MANDOCERR_DISPREP, /* duplicate display type */
  97         MANDOCERR_LISTREP, /* duplicate list type */
  98         MANDOCERR_BADATT, /* unknown AT&T UNIX version */
  99         MANDOCERR_BADBOOL, /* bad Boolean value */
 100         MANDOCERR_BADFONT, /* unknown font */
 101         MANDOCERR_BADSTANDARD, /* unknown standard specifier */
 102         MANDOCERR_BADWIDTH, /* bad width argument */
 103
 104         /* related to plain text */
 105         MANDOCERR_NOBLANKLN, /* blank line in non-literal context */
 106         MANDOCERR_BADTAB, /* tab in non-literal context */
 107         MANDOCERR_EOLNSPACE, /* end of line whitespace */
 108         MANDOCERR_BADCOMMENT, /* bad comment style */
 109         MANDOCERR_BADESCAPE, /* unknown escape sequence */
 110         MANDOCERR_BADQUOTE, /* unterminated quoted string */
 111
 112         /* related to equations */
 113         MANDOCERR_EQNQUOTE, /* unexpected literal in equation */
 114
 115         MANDOCERR_ERROR, /* ===== start of errors ===== */
 116
 117         /* related to equations */
 118         MANDOCERR_EQNNSCOPE, /* unexpected equation scope closure*/
 119         MANDOCERR_EQNSCOPE, /* equation scope open on exit */
 120         MANDOCERR_EQNBADSCOPE, /* overlapping equation scopes */
 121         MANDOCERR_EQNEOF, /* unexpected end of equation */
 122         MANDOCERR_EQNSYNT, /* equation syntax error */
 123
 124         /* related to tables */
 125         MANDOCERR_TBL, /* bad table syntax */
 126         MANDOCERR_TBLOPT, /* bad table option */
 127         MANDOCERR_TBLLAYOUT, /* bad table layout */
 128         MANDOCERR_TBLNOLAYOUT, /* no table layout cells specified */
 129         MANDOCERR_TBLNODATA, /* no table data cells specified */
 130         MANDOCERR_TBLIGNDATA, /* ignore data in cell */
 131         MANDOCERR_TBLBLOCK, /* data block still open */
 132         MANDOCERR_TBLEXTRADAT, /* ignoring extra data cells */
 133
 134         MANDOCERR_ROFFLOOP, /* input stack limit exceeded, infinite loop? */
 135         MANDOCERR_BADCHAR, /* skipping bad character */
 136         MANDOCERR_NAMESC, /* escaped character not allowed in a name */
 137         MANDOCERR_NONAME, /* manual name not yet set */
 138         MANDOCERR_NOTEXT, /* skipping text before the first section header */
 139         MANDOCERR_MACRO, /* skipping unknown macro */
 140         MANDOCERR_REQUEST, /* NOT IMPLEMENTED: skipping request */
 141         MANDOCERR_ARGCOUNT, /* argument count wrong */
 142         MANDOCERR_STRAYTA, /* skipping column outside column list */
 143         MANDOCERR_NOSCOPE, /* skipping end of block that is not open */
 144         MANDOCERR_SCOPEBROKEN, /* missing end of block */
 145         MANDOCERR_SCOPEEXIT, /* scope open on exit */
 146         MANDOCERR_UNAME, /* uname(3) system call failed */
 147         /* FIXME: merge following with MANDOCERR_ARGCOUNT */
 148         MANDOCERR_NOARGS, /* macro requires line argument(s) */
 149         MANDOCERR_NOBODY, /* macro requires body argument(s) */
 150         MANDOCERR_NOARGV, /* macro requires argument(s) */
 151         MANDOCERR_NUMERIC, /* request requires a numeric argument */
 152         MANDOCERR_LISTTYPE, /* missing list type */
 153         MANDOCERR_ARGSLOST, /* line argument(s) will be lost */
 154         MANDOCERR_BODYLOST, /* body argument(s) will be lost */
 155
 156         MANDOCERR_FATAL, /* ===== start of fatal errors ===== */
 157
 158         MANDOCERR_TOOLARGE, /* input too large */
 159         MANDOCERR_NOTMANUAL, /* manual isn't really a manual */
 160         MANDOCERR_COLUMNS, /* column syntax is inconsistent */
 161         MANDOCERR_BADDISP, /* NOT IMPLEMENTED: .Bd -file */
 162         MANDOCERR_SYNTARGVCOUNT, /* argument count wrong, violates syntax */
 163         MANDOCERR_SYNTCHILD, /* child violates parent syntax */
 164         MANDOCERR_SYNTARGCOUNT, /* argument count wrong, violates syntax */
 165         MANDOCERR_SO_PATH, /* NOT IMPLEMENTED: .so with absolute path or ".." */
 166         MANDOCERR_SO_FAIL, /* .so request failed */
 167         MANDOCERR_NODOCPROLOG, /* no document prologue */
 168         MANDOCERR_MEM, /* static buffer exhausted */
 169
 170         /* ===== system errors ===== */
 171
 172         MANDOCERR_SYSOPEN, /* cannot open file */
 173         MANDOCERR_SYSSTAT, /* cannot stat file */
 174         MANDOCERR_SYSREAD, /* cannot read file */
 175
 176         MANDOCERR_MAX
 177 };
 178
 179 struct  tbl_opts {
 180         char              tab; /* cell-separator */
 181         char              decimal; /* decimal point */
 182         int               linesize;
 183         int               opts;
 184 #define TBL_OPT_CENTRE   (1 << 0)
 185 #define TBL_OPT_EXPAND   (1 << 1)
 186 #define TBL_OPT_BOX      (1 << 2)
 187 #define TBL_OPT_DBOX     (1 << 3)
 188 #define TBL_OPT_ALLBOX   (1 << 4)
 189 #define TBL_OPT_NOKEEP   (1 << 5)
 190 #define TBL_OPT_NOSPACE  (1 << 6)
 191         int               cols; /* number of columns */
 192 };
 193
 194 /*
 195  * The head of a table specifies all of its columns.  When formatting a
 196  * tbl_span, iterate over these and plug in data from the tbl_span when
 197  * appropriate, using tbl_cell as a guide to placement.
 198  */
 199 struct  tbl_head {
 200         int               ident; /* 0 <= unique id < cols */
 201         int               vert; /* width of preceding vertical line */
 202         struct tbl_head  *next;
 203         struct tbl_head  *prev;
 204 };
 205
 206 enum    tbl_cellt {
 207         TBL_CELL_CENTRE, /* c, C */
 208         TBL_CELL_RIGHT, /* r, R */
 209         TBL_CELL_LEFT, /* l, L */
 210         TBL_CELL_NUMBER, /* n, N */
 211         TBL_CELL_SPAN, /* s, S */
 212         TBL_CELL_LONG, /* a, A */
 213         TBL_CELL_DOWN, /* ^ */
 214         TBL_CELL_HORIZ, /* _, - */
 215         TBL_CELL_DHORIZ, /* = */
 216         TBL_CELL_MAX
 217 };
 218
 219 /*
 220  * A cell in a layout row.
 221  */
 222 struct  tbl_cell {
 223         struct tbl_cell  *next;
 224         int               vert; /* width of preceding vertical line */
 225         enum tbl_cellt    pos;
 226         size_t            spacing;
 227         int               flags;
 228 #define TBL_CELL_TALIGN  (1 << 0) /* t, T */
 229 #define TBL_CELL_BALIGN  (1 << 1) /* d, D */
 230 #define TBL_CELL_BOLD    (1 << 2) /* fB, B, b */
 231 #define TBL_CELL_ITALIC  (1 << 3) /* fI, I, i */
 232 #define TBL_CELL_EQUAL   (1 << 4) /* e, E */
 233 #define TBL_CELL_UP      (1 << 5) /* u, U */
 234 #define TBL_CELL_WIGN    (1 << 6) /* z, Z */
 235         struct tbl_head  *head;
 236 };
 237
 238 /*
 239  * A layout row.
 240  */
 241 struct  tbl_row {
 242         struct tbl_row   *next;
 243         struct tbl_cell  *first;
 244         struct tbl_cell  *last;
 245         int               vert; /* trailing vertical line */
 246 };
 247
 248 enum    tbl_datt {
 249         TBL_DATA_NONE, /* has no data */
 250         TBL_DATA_DATA, /* consists of data/string */
 251         TBL_DATA_HORIZ, /* horizontal line */
 252         TBL_DATA_DHORIZ, /* double-horizontal line */
 253         TBL_DATA_NHORIZ, /* squeezed horizontal line */
 254         TBL_DATA_NDHORIZ /* squeezed double-horizontal line */
 255 };
 256
 257 /*
 258  * A cell within a row of data.  The "string" field contains the actual
 259  * string value that's in the cell.  The rest is layout.
 260  */
 261 struct  tbl_dat {
 262         struct tbl_cell  *layout; /* layout cell */
 263         int               spans; /* how many spans follow */
 264         struct tbl_dat   *next;
 265         char             *string; /* data (NULL if not TBL_DATA_DATA) */
 266         enum tbl_datt     pos;
 267 };
 268
 269 enum    tbl_spant {
 270         TBL_SPAN_DATA, /* span consists of data */
 271         TBL_SPAN_HORIZ, /* span is horizontal line */
 272         TBL_SPAN_DHORIZ /* span is double horizontal line */
 273 };
 274
 275 /*
 276  * A row of data in a table.
 277  */
 278 struct  tbl_span {
 279         struct tbl_opts  *opts;
 280         struct tbl_head  *head;
 281         struct tbl_row   *layout; /* layout row */
 282         struct tbl_dat   *first;
 283         struct tbl_dat   *last;
 284         int               line; /* parse line */
 285         int               flags;
 286 #define TBL_SPAN_FIRST   (1 << 0)
 287 #define TBL_SPAN_LAST    (1 << 1)
 288         enum tbl_spant    pos;
 289         struct tbl_span  *next;
 290 };
 291
 292 enum    eqn_boxt {
 293         EQN_ROOT, /* root of parse tree */
 294         EQN_TEXT, /* text (number, variable, whatever) */
 295         EQN_SUBEXPR, /* nested `eqn' subexpression */
 296         EQN_LIST, /* subexpressions list */
 297         EQN_MATRIX /* matrix subexpression */
 298 };
 299
 300 enum    eqn_markt {
 301         EQNMARK_NONE = 0,
 302         EQNMARK_DOT,
 303         EQNMARK_DOTDOT,
 304         EQNMARK_HAT,
 305         EQNMARK_TILDE,
 306         EQNMARK_VEC,
 307         EQNMARK_DYAD,
 308         EQNMARK_BAR,
 309         EQNMARK_UNDER,
 310         EQNMARK__MAX
 311 };
 312
 313 enum    eqn_fontt {
 314         EQNFONT_NONE = 0,
 315         EQNFONT_ROMAN,
 316         EQNFONT_BOLD,
 317         EQNFONT_FAT,
 318         EQNFONT_ITALIC,
 319         EQNFONT__MAX
 320 };
 321
 322 enum    eqn_post {
 323         EQNPOS_NONE = 0,
 324         EQNPOS_OVER,
 325         EQNPOS_SUP,
 326         EQNPOS_SUB,
 327         EQNPOS_TO,
 328         EQNPOS_FROM,
 329         EQNPOS__MAX
 330 };
 331
 332 enum    eqn_pilet {
 333         EQNPILE_NONE = 0,
 334         EQNPILE_PILE,
 335         EQNPILE_CPILE,
 336         EQNPILE_RPILE,
 337         EQNPILE_LPILE,
 338         EQNPILE_COL,
 339         EQNPILE_CCOL,
 340         EQNPILE_RCOL,
 341         EQNPILE_LCOL,
 342         EQNPILE__MAX
 343 };
 344
 345  /*
 346  * A "box" is a parsed mathematical expression as defined by the eqn.7
 347  * grammar.
 348  */
 349 struct  eqn_box {
 350         int               size; /* font size of expression */
 351 #define EQN_DEFSIZE       INT_MIN
 352         enum eqn_boxt     type; /* type of node */
 353         struct eqn_box   *first; /* first child node */
 354         struct eqn_box   *last; /* last child node */
 355         struct eqn_box   *next; /* node sibling */
 356         struct eqn_box   *parent; /* node sibling */
 357         char             *text; /* text (or NULL) */
 358         char             *left;
 359         char             *right;
 360         enum eqn_post     pos; /* position of next box */
 361         enum eqn_markt    mark; /* a mark about the box */
 362         enum eqn_fontt    font; /* font of box */
 363         enum eqn_pilet    pile; /* equation piling */
 364 };
 365
 366 /*
 367  * An equation consists of a tree of expressions starting at a given
 368  * line and position.
 369  */
 370 struct  eqn {
 371         char             *name; /* identifier (or NULL) */
 372         struct eqn_box   *root; /* root mathematical expression */
 373         int               ln; /* invocation line */
 374         int               pos; /* invocation position */
 375 };
 376
 377 /*
 378  * Parse options.
 379  */
 380 #define MPARSE_MDOC     1  /* assume -mdoc */
 381 #define MPARSE_MAN      2  /* assume -man */
 382 #define MPARSE_SO       4  /* honour .so requests */
 383 #define MPARSE_QUICK    8  /* abort the parse early */
 384
 385 enum    mandoc_esc {
 386         ESCAPE_ERROR = 0, /* bail! unparsable escape */
 387         ESCAPE_IGNORE, /* escape to be ignored */
 388         ESCAPE_SPECIAL, /* a regular special character */
 389         ESCAPE_FONT, /* a generic font mode */
 390         ESCAPE_FONTBOLD, /* bold font mode */
 391         ESCAPE_FONTITALIC, /* italic font mode */
 392         ESCAPE_FONTBI, /* bold italic font mode */
 393         ESCAPE_FONTROMAN, /* roman font mode */
 394         ESCAPE_FONTPREV, /* previous font mode */
 395         ESCAPE_NUMBERED, /* a numbered glyph */
 396         ESCAPE_UNICODE, /* a unicode codepoint */
 397         ESCAPE_NOSPACE, /* suppress space if the last on a line */
 398         ESCAPE_SKIPCHAR /* skip the next character */
 399 };
 400
 401 typedef void    (*mandocmsg)(enum mandocerr, enum mandoclevel,
 402                         const char *, int, int, const char *);
 403
 404 struct  mparse;
 405 struct  mchars;
 406 struct  mdoc;
 407 struct  man;
 408
 409 __BEGIN_DECLS
 410
 411 enum mandoc_esc   mandoc_escape(const char **, const char **, int *);
 412 struct mchars    *mchars_alloc(void);
 413 void              mchars_free(struct mchars *);
 414 char              mchars_num2char(const char *, size_t);
 415 int               mchars_num2uc(const char *, size_t);
 416 int               mchars_spec2cp(const struct mchars *,
 417                         const char *, size_t);
 418 const char       *mchars_spec2str(const struct mchars *,
 419                         const char *, size_t, size_t *);
 420 struct mparse    *mparse_alloc(int, enum mandoclevel, mandocmsg, char *);
 421 void              mparse_free(struct mparse *);
 422 void              mparse_keep(struct mparse *);
 423 enum mandoclevel  mparse_readfd(struct mparse *, int, const char *);
 424 enum mandoclevel  mparse_readmem(struct mparse *, const void *, size_t,
 425                         const char *);
 426 void              mparse_reset(struct mparse *);
 427 void              mparse_result(struct mparse *,
 428                         struct mdoc **, struct man **, char **);
 429 const char       *mparse_getkeep(const struct mparse *);
 430 const char       *mparse_strerror(enum mandocerr);
 431 const char       *mparse_strlevel(enum mandoclevel);
 432
 433 __END_DECLS
 434
 435 #endif /*!MANDOC_H*/