mandoc.h

   1 /*      $Id: mandoc.h,v 1.130 2014/07/02 20:19:11 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, /* missing .TH macro, using "unknown 1" */
  53         MANDOCERR_TITLE_CASE, /* lower case character in document title */
  54         MANDOCERR_MSEC_BAD, /* unknown manual section: section */
  55         MANDOCERR_ARCH_BAD, /* unknown manual volume or arch: volume */
  56         MANDOCERR_DATE_MISSING, /* missing date, using today's date */
  57         MANDOCERR_DATE_BAD, /* cannot parse date, using it verbatim: date */
  58         MANDOCERR_PROLOG_ORDER, /* prologue macros out of order: macro */
  59         MANDOCERR_PROLOG_REP, /* duplicate prologue macro: macro */
  60         MANDOCERR_PROLOG_BAD, /* incomplete prologue, terminated by: macro */
  61         MANDOCERR_PROLOG_ONLY, /* skipping prologue macro in body: macro */
  62
  63         /* related to document structure */
  64         MANDOCERR_SO, /* .so is fragile, better use ln(1): .so path */
  65         MANDOCERR_DOC_EMPTY, /* no document body */
  66         MANDOCERR_SEC_BEFORE, /* content before first section header: macro */
  67         MANDOCERR_NAMESEC_FIRST, /* first section is not "NAME": title */
  68         MANDOCERR_NAMESEC_BAD, /* bad NAME section contents: macro */
  69         MANDOCERR_SEC_ORDER, /* sections out of conventional order: title */
  70         MANDOCERR_SEC_REP, /* duplicate section title: title */
  71         MANDOCERR_SEC_MSEC, /* unexpected section: title for ... only */
  72
  73         /* related to macros and nesting */
  74         MANDOCERR_MACRO_OBS, /* obsolete macro: macro */
  75         MANDOCERR_PAR_SKIP, /* skipping paragraph macro: macro ... */
  76         MANDOCERR_PAR_MOVE, /* moving paragraph macro out of list: macro */
  77         MANDOCERR_NS_SKIP, /* skipping no-space macro */
  78         MANDOCERR_BLOCK_NEST, /* blocks badly nested: macro ... */
  79         MANDOCERR_BD_NEST, /* nested displays are not portable: macro ... */
  80         MANDOCERR_BL_MOVE, /* moving content out of list: macro */
  81         MANDOCERR_VT_CHILD, /* .Vt block has child macro: macro */
  82         MANDOCERR_FI_SKIP, /* fill mode already enabled, skipping .fi */
  83         MANDOCERR_NF_SKIP, /* fill mode already disabled, skipping .nf */
  84         MANDOCERR_LINESCOPE, /* line scope broken: macro breaks macro */
  85
  86         /* related to missing macro arguments */
  87         MANDOCERR_MACRO_EMPTY, /* skipping empty macro: macro */
  88         MANDOCERR_ARGCWARN, /* argument count wrong */
  89         MANDOCERR_DISPTYPE, /* missing display type */
  90         MANDOCERR_LISTFIRST, /* list type must come first */
  91         MANDOCERR_NOWIDTHARG, /* tag lists require a width argument */
  92         MANDOCERR_FONTTYPE, /* missing font type */
  93
  94         /* related to bad macro arguments */
  95         MANDOCERR_IGNARGV, /* skipping argument */
  96         MANDOCERR_ARGVREP, /* duplicate argument */
  97         MANDOCERR_DISPREP, /* duplicate display type */
  98         MANDOCERR_LISTREP, /* duplicate list type */
  99         MANDOCERR_BADATT, /* unknown AT&T UNIX version */
 100         MANDOCERR_BADBOOL, /* bad Boolean value */
 101         MANDOCERR_BADFONT, /* unknown font */
 102         MANDOCERR_BADSTANDARD, /* unknown standard specifier */
 103         MANDOCERR_BADWIDTH, /* bad width argument */
 104
 105         /* related to plain text */
 106         MANDOCERR_NOBLANKLN, /* blank line in non-literal context */
 107         MANDOCERR_BADTAB, /* tab in non-literal context */
 108         MANDOCERR_EOLNSPACE, /* end of line whitespace */
 109         MANDOCERR_BADCOMMENT, /* bad comment style */
 110         MANDOCERR_BADESCAPE, /* bad escape sequence */
 111         MANDOCERR_BADQUOTE, /* unterminated quoted string */
 112
 113         MANDOCERR_ERROR, /* ===== start of errors ===== */
 114
 115         /* related to equations */
 116         MANDOCERR_EQNNSCOPE, /* unexpected equation scope closure*/
 117         MANDOCERR_EQNSCOPE, /* equation scope open on exit */
 118         MANDOCERR_EQNBADSCOPE, /* overlapping equation scopes */
 119         MANDOCERR_EQNEOF, /* unexpected end of equation */
 120         MANDOCERR_EQNSYNT, /* equation syntax error */
 121
 122         /* related to tables */
 123         MANDOCERR_TBL, /* bad table syntax */
 124         MANDOCERR_TBLOPT, /* bad table option */
 125         MANDOCERR_TBLLAYOUT, /* bad table layout */
 126         MANDOCERR_TBLNOLAYOUT, /* no table layout cells specified */
 127         MANDOCERR_TBLNODATA, /* no table data cells specified */
 128         MANDOCERR_TBLIGNDATA, /* ignore data in cell */
 129         MANDOCERR_TBLBLOCK, /* data block still open */
 130         MANDOCERR_TBLEXTRADAT, /* ignoring extra data cells */
 131
 132         MANDOCERR_ROFFLOOP, /* input stack limit exceeded, infinite loop? */
 133         MANDOCERR_BADCHAR, /* skipping bad character */
 134         MANDOCERR_NAMESC, /* escaped character not allowed in a name */
 135         MANDOCERR_NONAME, /* manual name not yet set */
 136         MANDOCERR_NOTEXT, /* skipping text before first section header */
 137         MANDOCERR_MACRO, /* skipping unknown macro */
 138         MANDOCERR_REQUEST, /* NOT IMPLEMENTED: skipping request */
 139         MANDOCERR_ARGCOUNT, /* argument count wrong */
 140         MANDOCERR_RS_SKIP, /* skipping invalid content in .Rs block: macro */
 141         MANDOCERR_STRAYTA, /* skipping column outside column list */
 142         MANDOCERR_NOSCOPE, /* skipping end of block that is not open */
 143         MANDOCERR_SCOPEBROKEN, /* missing end of block */
 144         MANDOCERR_SCOPEEXIT, /* scope open on exit */
 145         MANDOCERR_UNAME, /* uname(3) system call failed */
 146         /* FIXME: merge following with MANDOCERR_ARGCOUNT */
 147         MANDOCERR_NOARGS, /* macro requires line argument(s) */
 148         MANDOCERR_NOBODY, /* macro requires body argument(s) */
 149         MANDOCERR_NOARGV, /* macro requires argument(s) */
 150         MANDOCERR_NUMERIC, /* request requires a numeric argument */
 151         MANDOCERR_LISTTYPE, /* missing list type */
 152         MANDOCERR_ARGSLOST, /* line argument(s) will be lost */
 153
 154         MANDOCERR_FATAL, /* ===== start of fatal errors ===== */
 155
 156         MANDOCERR_TOOLARGE, /* input too large */
 157         MANDOCERR_NOTMANUAL, /* not a manual */
 158         MANDOCERR_COLUMNS, /* column syntax is inconsistent */
 159         MANDOCERR_BADDISP, /* NOT IMPLEMENTED: .Bd -file */
 160         MANDOCERR_SYNTARGVCOUNT, /* argument count wrong, violates syntax */
 161         MANDOCERR_SYNTCHILD, /* child violates parent syntax */
 162         MANDOCERR_SYNTARGCOUNT, /* argument count wrong, violates syntax */
 163         MANDOCERR_SO_PATH, /* NOT IMPLEMENTED: .so with absolute path or ".." */
 164         MANDOCERR_SO_FAIL, /* .so request failed */
 165         MANDOCERR_NODOCPROLOG, /* no document prologue */
 166         MANDOCERR_MEM, /* static buffer exhausted */
 167
 168         /* ===== system errors ===== */
 169
 170         MANDOCERR_SYSOPEN, /* cannot open file */
 171         MANDOCERR_SYSSTAT, /* cannot stat file */
 172         MANDOCERR_SYSREAD, /* cannot read file */
 173
 174         MANDOCERR_MAX
 175 };
 176
 177 struct  tbl_opts {
 178         char              tab; /* cell-separator */
 179         char              decimal; /* decimal point */
 180         int               linesize;
 181         int               opts;
 182 #define TBL_OPT_CENTRE   (1 << 0)
 183 #define TBL_OPT_EXPAND   (1 << 1)
 184 #define TBL_OPT_BOX      (1 << 2)
 185 #define TBL_OPT_DBOX     (1 << 3)
 186 #define TBL_OPT_ALLBOX   (1 << 4)
 187 #define TBL_OPT_NOKEEP   (1 << 5)
 188 #define TBL_OPT_NOSPACE  (1 << 6)
 189         int               cols; /* number of columns */
 190 };
 191
 192 /*
 193  * The head of a table specifies all of its columns.  When formatting a
 194  * tbl_span, iterate over these and plug in data from the tbl_span when
 195  * appropriate, using tbl_cell as a guide to placement.
 196  */
 197 struct  tbl_head {
 198         int               ident; /* 0 <= unique id < cols */
 199         int               vert; /* width of preceding vertical line */
 200         struct tbl_head  *next;
 201         struct tbl_head  *prev;
 202 };
 203
 204 enum    tbl_cellt {
 205         TBL_CELL_CENTRE, /* c, C */
 206         TBL_CELL_RIGHT, /* r, R */
 207         TBL_CELL_LEFT, /* l, L */
 208         TBL_CELL_NUMBER, /* n, N */
 209         TBL_CELL_SPAN, /* s, S */
 210         TBL_CELL_LONG, /* a, A */
 211         TBL_CELL_DOWN, /* ^ */
 212         TBL_CELL_HORIZ, /* _, - */
 213         TBL_CELL_DHORIZ, /* = */
 214         TBL_CELL_MAX
 215 };
 216
 217 /*
 218  * A cell in a layout row.
 219  */
 220 struct  tbl_cell {
 221         struct tbl_cell  *next;
 222         int               vert; /* width of preceding vertical line */
 223         enum tbl_cellt    pos;
 224         size_t            spacing;
 225         int               flags;
 226 #define TBL_CELL_TALIGN  (1 << 0) /* t, T */
 227 #define TBL_CELL_BALIGN  (1 << 1) /* d, D */
 228 #define TBL_CELL_BOLD    (1 << 2) /* fB, B, b */
 229 #define TBL_CELL_ITALIC  (1 << 3) /* fI, I, i */
 230 #define TBL_CELL_EQUAL   (1 << 4) /* e, E */
 231 #define TBL_CELL_UP      (1 << 5) /* u, U */
 232 #define TBL_CELL_WIGN    (1 << 6) /* z, Z */
 233         struct tbl_head  *head;
 234 };
 235
 236 /*
 237  * A layout row.
 238  */
 239 struct  tbl_row {
 240         struct tbl_row   *next;
 241         struct tbl_cell  *first;
 242         struct tbl_cell  *last;
 243         int               vert; /* trailing vertical line */
 244 };
 245
 246 enum    tbl_datt {
 247         TBL_DATA_NONE, /* has no data */
 248         TBL_DATA_DATA, /* consists of data/string */
 249         TBL_DATA_HORIZ, /* horizontal line */
 250         TBL_DATA_DHORIZ, /* double-horizontal line */
 251         TBL_DATA_NHORIZ, /* squeezed horizontal line */
 252         TBL_DATA_NDHORIZ /* squeezed double-horizontal line */
 253 };
 254
 255 /*
 256  * A cell within a row of data.  The "string" field contains the actual
 257  * string value that's in the cell.  The rest is layout.
 258  */
 259 struct  tbl_dat {
 260         struct tbl_cell  *layout; /* layout cell */
 261         int               spans; /* how many spans follow */
 262         struct tbl_dat   *next;
 263         char             *string; /* data (NULL if not TBL_DATA_DATA) */
 264         enum tbl_datt     pos;
 265 };
 266
 267 enum    tbl_spant {
 268         TBL_SPAN_DATA, /* span consists of data */
 269         TBL_SPAN_HORIZ, /* span is horizontal line */
 270         TBL_SPAN_DHORIZ /* span is double horizontal line */
 271 };
 272
 273 /*
 274  * A row of data in a table.
 275  */
 276 struct  tbl_span {
 277         struct tbl_opts  *opts;
 278         struct tbl_head  *head;
 279         struct tbl_row   *layout; /* layout row */
 280         struct tbl_dat   *first;
 281         struct tbl_dat   *last;
 282         int               line; /* parse line */
 283         int               flags;
 284 #define TBL_SPAN_FIRST   (1 << 0)
 285 #define TBL_SPAN_LAST    (1 << 1)
 286         enum tbl_spant    pos;
 287         struct tbl_span  *next;
 288 };
 289
 290 enum    eqn_boxt {
 291         EQN_ROOT, /* root of parse tree */
 292         EQN_TEXT, /* text (number, variable, whatever) */
 293         EQN_SUBEXPR, /* nested `eqn' subexpression */
 294         EQN_LIST, /* subexpressions list */
 295         EQN_MATRIX /* matrix subexpression */
 296 };
 297
 298 enum    eqn_markt {
 299         EQNMARK_NONE = 0,
 300         EQNMARK_DOT,
 301         EQNMARK_DOTDOT,
 302         EQNMARK_HAT,
 303         EQNMARK_TILDE,
 304         EQNMARK_VEC,
 305         EQNMARK_DYAD,
 306         EQNMARK_BAR,
 307         EQNMARK_UNDER,
 308         EQNMARK__MAX
 309 };
 310
 311 enum    eqn_fontt {
 312         EQNFONT_NONE = 0,
 313         EQNFONT_ROMAN,
 314         EQNFONT_BOLD,
 315         EQNFONT_FAT,
 316         EQNFONT_ITALIC,
 317         EQNFONT__MAX
 318 };
 319
 320 enum    eqn_post {
 321         EQNPOS_NONE = 0,
 322         EQNPOS_OVER,
 323         EQNPOS_SUP,
 324         EQNPOS_SUB,
 325         EQNPOS_TO,
 326         EQNPOS_FROM,
 327         EQNPOS__MAX
 328 };
 329
 330 enum    eqn_pilet {
 331         EQNPILE_NONE = 0,
 332         EQNPILE_PILE,
 333         EQNPILE_CPILE,
 334         EQNPILE_RPILE,
 335         EQNPILE_LPILE,
 336         EQNPILE_COL,
 337         EQNPILE_CCOL,
 338         EQNPILE_RCOL,
 339         EQNPILE_LCOL,
 340         EQNPILE__MAX
 341 };
 342
 343  /*
 344  * A "box" is a parsed mathematical expression as defined by the eqn.7
 345  * grammar.
 346  */
 347 struct  eqn_box {
 348         int               size; /* font size of expression */
 349 #define EQN_DEFSIZE       INT_MIN
 350         enum eqn_boxt     type; /* type of node */
 351         struct eqn_box   *first; /* first child node */
 352         struct eqn_box   *last; /* last child node */
 353         struct eqn_box   *next; /* node sibling */
 354         struct eqn_box   *parent; /* node sibling */
 355         char             *text; /* text (or NULL) */
 356         char             *left;
 357         char             *right;
 358         enum eqn_post     pos; /* position of next box */
 359         enum eqn_markt    mark; /* a mark about the box */
 360         enum eqn_fontt    font; /* font of box */
 361         enum eqn_pilet    pile; /* equation piling */
 362 };
 363
 364 /*
 365  * An equation consists of a tree of expressions starting at a given
 366  * line and position.
 367  */
 368 struct  eqn {
 369         char             *name; /* identifier (or NULL) */
 370         struct eqn_box   *root; /* root mathematical expression */
 371         int               ln; /* invocation line */
 372         int               pos; /* invocation position */
 373 };
 374
 375 /*
 376  * Parse options.
 377  */
 378 #define MPARSE_MDOC     1  /* assume -mdoc */
 379 #define MPARSE_MAN      2  /* assume -man */
 380 #define MPARSE_SO       4  /* honour .so requests */
 381 #define MPARSE_QUICK    8  /* abort the parse early */
 382
 383 enum    mandoc_esc {
 384         ESCAPE_ERROR = 0, /* bail! unparsable escape */
 385         ESCAPE_IGNORE, /* escape to be ignored */
 386         ESCAPE_SPECIAL, /* a regular special character */
 387         ESCAPE_FONT, /* a generic font mode */
 388         ESCAPE_FONTBOLD, /* bold font mode */
 389         ESCAPE_FONTITALIC, /* italic font mode */
 390         ESCAPE_FONTBI, /* bold italic font mode */
 391         ESCAPE_FONTROMAN, /* roman font mode */
 392         ESCAPE_FONTPREV, /* previous font mode */
 393         ESCAPE_NUMBERED, /* a numbered glyph */
 394         ESCAPE_UNICODE, /* a unicode codepoint */
 395         ESCAPE_NOSPACE, /* suppress space if the last on a line */
 396         ESCAPE_SKIPCHAR /* skip the next character */
 397 };
 398
 399 typedef void    (*mandocmsg)(enum mandocerr, enum mandoclevel,
 400                         const char *, int, int, const char *);
 401
 402 struct  mparse;
 403 struct  mchars;
 404 struct  mdoc;
 405 struct  man;
 406
 407 __BEGIN_DECLS
 408
 409 enum mandoc_esc   mandoc_escape(const char **, const char **, int *);
 410 struct mchars    *mchars_alloc(void);
 411 void              mchars_free(struct mchars *);
 412 char              mchars_num2char(const char *, size_t);
 413 int               mchars_num2uc(const char *, size_t);
 414 int               mchars_spec2cp(const struct mchars *,
 415                         const char *, size_t);
 416 const char       *mchars_spec2str(const struct mchars *,
 417                         const char *, size_t, size_t *);
 418 struct mparse    *mparse_alloc(int, enum mandoclevel, mandocmsg, char *);
 419 void              mparse_free(struct mparse *);
 420 void              mparse_keep(struct mparse *);
 421 enum mandoclevel  mparse_readfd(struct mparse *, int, const char *);
 422 enum mandoclevel  mparse_readmem(struct mparse *, const void *, size_t,
 423                         const char *);
 424 void              mparse_reset(struct mparse *);
 425 void              mparse_result(struct mparse *,
 426                         struct mdoc **, struct man **, char **);
 427 const char       *mparse_getkeep(const struct mparse *);
 428 const char       *mparse_strerror(enum mandocerr);
 429 const char       *mparse_strlevel(enum mandoclevel);
 430
 431 __END_DECLS
 432
 433 #endif /*!MANDOC_H*/