mandoc.h

   1 /*      $Id: mandoc.h,v 1.144 2014/07/30 12:58:21 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_BLK_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_BLK_LINE, /* line scope broken: macro breaks macro */
  85
  86         /* related to missing arguments */
  87         MANDOCERR_REQ_EMPTY, /* skipping empty request: request */
  88         MANDOCERR_COND_EMPTY, /* conditional request controls empty scope */
  89         MANDOCERR_MACRO_EMPTY, /* skipping empty macro: macro */
  90         MANDOCERR_ARG_EMPTY, /* empty argument, using 0n: macro arg */
  91         MANDOCERR_ARGCWARN, /* argument count wrong */
  92         MANDOCERR_BD_NOTYPE, /* missing display type, using -ragged */
  93         MANDOCERR_BL_LATETYPE, /* list type is not the first argument: arg */
  94         MANDOCERR_BL_NOWIDTH, /* missing -width in -tag list, using 8n */
  95         MANDOCERR_EX_NONAME, /* missing name for .Ex, using "" */
  96         MANDOCERR_IT_NOHEAD, /* empty head in list item: type */
  97         MANDOCERR_IT_NOBODY, /* empty list item: type */
  98         MANDOCERR_BF_NOFONT, /* missing font type, using \fR */
  99         MANDOCERR_BF_BADFONT, /* unknown font type, using \fR: macro font */
 100         MANDOCERR_ARG_STD, /* missing -std argument, adding it: macro */
 101
 102         /* related to bad arguments */
 103         MANDOCERR_IGNARGV, /* skipping argument */
 104         MANDOCERR_ARG_QUOTE, /* unterminated quoted argument */
 105         MANDOCERR_ARG_REP, /* duplicate argument: macro arg */
 106         MANDOCERR_BD_REP, /* skipping duplicate display type: type */
 107         MANDOCERR_BL_REP, /* skipping duplicate list type: type */
 108         MANDOCERR_AT_BAD, /* unknown AT&T UNIX version: version */
 109         MANDOCERR_RS_BAD, /* invalid content in Rs block: macro */
 110         MANDOCERR_SM_BAD, /* invalid Boolean argument: macro arg */
 111         MANDOCERR_FT_BAD, /* unknown font, skipping request: request font */
 112
 113         /* related to plain text */
 114         MANDOCERR_FI_BLANK, /* blank line in fill mode, using .sp */
 115         MANDOCERR_FI_TAB, /* tab in filled text */
 116         MANDOCERR_SPACE_EOL, /* whitespace at end of input line */
 117         MANDOCERR_COMMENT_BAD, /* bad comment style */
 118         MANDOCERR_ESC_BAD, /* invalid escape sequence: esc */
 119         MANDOCERR_STR_UNDEF, /* undefined string, using "": name */
 120
 121         MANDOCERR_ERROR, /* ===== start of errors ===== */
 122
 123         /* related to equations */
 124         MANDOCERR_EQNNSCOPE, /* unexpected equation scope closure*/
 125         MANDOCERR_EQNSCOPE, /* equation scope open on exit */
 126         MANDOCERR_EQNBADSCOPE, /* overlapping equation scopes */
 127         MANDOCERR_EQNEOF, /* unexpected end of equation */
 128         MANDOCERR_EQNSYNT, /* equation syntax error */
 129
 130         /* related to tables */
 131         MANDOCERR_TBL, /* bad table syntax */
 132         MANDOCERR_TBLOPT, /* bad table option */
 133         MANDOCERR_TBLLAYOUT, /* bad table layout */
 134         MANDOCERR_TBLNOLAYOUT, /* no table layout cells specified */
 135         MANDOCERR_TBLNODATA, /* no table data cells specified */
 136         MANDOCERR_TBLIGNDATA, /* ignore data in cell */
 137         MANDOCERR_TBLBLOCK, /* data block still open */
 138         MANDOCERR_TBLEXTRADAT, /* ignoring extra data cells */
 139
 140         /* related to document structure and macros */
 141         MANDOCERR_ROFFLOOP, /* input stack limit exceeded, infinite loop? */
 142         MANDOCERR_BADCHAR, /* skipping bad character */
 143         MANDOCERR_MACRO, /* skipping unknown macro */
 144         MANDOCERR_TA_STRAY, /* skipping column outside column list */
 145         MANDOCERR_BLK_NOTOPEN, /* skipping end of block that is not open */
 146         MANDOCERR_BLK_BROKEN, /* inserting missing end of block: macro ... */
 147         MANDOCERR_BLK_NOEND, /* appending missing end of block: macro */
 148
 149         /* related to request and macro arguments */
 150         MANDOCERR_NAMESC, /* escaped character not allowed in a name */
 151         MANDOCERR_ARGCOUNT, /* argument count wrong */
 152         MANDOCERR_BL_NOTYPE, /* missing list type, using -item */
 153         MANDOCERR_NM_NONAME, /* missing manual name, using "" */
 154         MANDOCERR_OS_UNAME, /* uname(3) system call failed, using UNKNOWN */
 155         MANDOCERR_ST_BAD, /* unknown standard specifier: standard */
 156         MANDOCERR_IT_NONUM, /* skipping request without numeric argument */
 157         MANDOCERR_ARG_SKIP, /* skipping all arguments: macro args */
 158         MANDOCERR_ARG_EXCESS, /* skipping excess arguments: macro ... args */
 159
 160         MANDOCERR_FATAL, /* ===== start of fatal errors ===== */
 161
 162         MANDOCERR_TOOLARGE, /* input too large */
 163         MANDOCERR_NOTMANUAL, /* not a manual */
 164         MANDOCERR_COLUMNS, /* column syntax is inconsistent */
 165         MANDOCERR_BADDISP, /* NOT IMPLEMENTED: .Bd -file */
 166         MANDOCERR_SYNTCHILD, /* child violates parent syntax */
 167         MANDOCERR_SYNTARGCOUNT, /* argument count wrong, violates syntax */
 168         MANDOCERR_SO_PATH, /* NOT IMPLEMENTED: .so with absolute path or ".." */
 169         MANDOCERR_SO_FAIL, /* .so request failed */
 170         MANDOCERR_NODOCPROLOG, /* no document prologue */
 171         MANDOCERR_MEM, /* static buffer exhausted */
 172
 173         /* ===== system errors ===== */
 174
 175         MANDOCERR_SYSOPEN, /* cannot open file */
 176         MANDOCERR_SYSSTAT, /* cannot stat file */
 177         MANDOCERR_SYSREAD, /* cannot read file */
 178
 179         MANDOCERR_MAX
 180 };
 181
 182 struct  tbl_opts {
 183         char              tab; /* cell-separator */
 184         char              decimal; /* decimal point */
 185         int               linesize;
 186         int               opts;
 187 #define TBL_OPT_CENTRE   (1 << 0)
 188 #define TBL_OPT_EXPAND   (1 << 1)
 189 #define TBL_OPT_BOX      (1 << 2)
 190 #define TBL_OPT_DBOX     (1 << 3)
 191 #define TBL_OPT_ALLBOX   (1 << 4)
 192 #define TBL_OPT_NOKEEP   (1 << 5)
 193 #define TBL_OPT_NOSPACE  (1 << 6)
 194         int               cols; /* number of columns */
 195 };
 196
 197 /*
 198  * The head of a table specifies all of its columns.  When formatting a
 199  * tbl_span, iterate over these and plug in data from the tbl_span when
 200  * appropriate, using tbl_cell as a guide to placement.
 201  */
 202 struct  tbl_head {
 203         int               ident; /* 0 <= unique id < cols */
 204         int               vert; /* width of preceding vertical line */
 205         struct tbl_head  *next;
 206         struct tbl_head  *prev;
 207 };
 208
 209 enum    tbl_cellt {
 210         TBL_CELL_CENTRE, /* c, C */
 211         TBL_CELL_RIGHT, /* r, R */
 212         TBL_CELL_LEFT, /* l, L */
 213         TBL_CELL_NUMBER, /* n, N */
 214         TBL_CELL_SPAN, /* s, S */
 215         TBL_CELL_LONG, /* a, A */
 216         TBL_CELL_DOWN, /* ^ */
 217         TBL_CELL_HORIZ, /* _, - */
 218         TBL_CELL_DHORIZ, /* = */
 219         TBL_CELL_MAX
 220 };
 221
 222 /*
 223  * A cell in a layout row.
 224  */
 225 struct  tbl_cell {
 226         struct tbl_cell  *next;
 227         int               vert; /* width of preceding vertical line */
 228         enum tbl_cellt    pos;
 229         size_t            spacing;
 230         int               flags;
 231 #define TBL_CELL_TALIGN  (1 << 0) /* t, T */
 232 #define TBL_CELL_BALIGN  (1 << 1) /* d, D */
 233 #define TBL_CELL_BOLD    (1 << 2) /* fB, B, b */
 234 #define TBL_CELL_ITALIC  (1 << 3) /* fI, I, i */
 235 #define TBL_CELL_EQUAL   (1 << 4) /* e, E */
 236 #define TBL_CELL_UP      (1 << 5) /* u, U */
 237 #define TBL_CELL_WIGN    (1 << 6) /* z, Z */
 238         struct tbl_head  *head;
 239 };
 240
 241 /*
 242  * A layout row.
 243  */
 244 struct  tbl_row {
 245         struct tbl_row   *next;
 246         struct tbl_cell  *first;
 247         struct tbl_cell  *last;
 248         int               vert; /* trailing vertical line */
 249 };
 250
 251 enum    tbl_datt {
 252         TBL_DATA_NONE, /* has no data */
 253         TBL_DATA_DATA, /* consists of data/string */
 254         TBL_DATA_HORIZ, /* horizontal line */
 255         TBL_DATA_DHORIZ, /* double-horizontal line */
 256         TBL_DATA_NHORIZ, /* squeezed horizontal line */
 257         TBL_DATA_NDHORIZ /* squeezed double-horizontal line */
 258 };
 259
 260 /*
 261  * A cell within a row of data.  The "string" field contains the actual
 262  * string value that's in the cell.  The rest is layout.
 263  */
 264 struct  tbl_dat {
 265         struct tbl_cell  *layout; /* layout cell */
 266         int               spans; /* how many spans follow */
 267         struct tbl_dat   *next;
 268         char             *string; /* data (NULL if not TBL_DATA_DATA) */
 269         enum tbl_datt     pos;
 270 };
 271
 272 enum    tbl_spant {
 273         TBL_SPAN_DATA, /* span consists of data */
 274         TBL_SPAN_HORIZ, /* span is horizontal line */
 275         TBL_SPAN_DHORIZ /* span is double horizontal line */
 276 };
 277
 278 /*
 279  * A row of data in a table.
 280  */
 281 struct  tbl_span {
 282         struct tbl_opts  *opts;
 283         struct tbl_head  *head;
 284         struct tbl_row   *layout; /* layout row */
 285         struct tbl_dat   *first;
 286         struct tbl_dat   *last;
 287         int               line; /* parse line */
 288         int               flags;
 289 #define TBL_SPAN_FIRST   (1 << 0)
 290 #define TBL_SPAN_LAST    (1 << 1)
 291         enum tbl_spant    pos;
 292         struct tbl_span  *next;
 293 };
 294
 295 enum    eqn_boxt {
 296         EQN_ROOT, /* root of parse tree */
 297         EQN_TEXT, /* text (number, variable, whatever) */
 298         EQN_SUBEXPR, /* nested `eqn' subexpression */
 299         EQN_LIST, /* subexpressions list */
 300         EQN_MATRIX /* matrix subexpression */
 301 };
 302
 303 enum    eqn_markt {
 304         EQNMARK_NONE = 0,
 305         EQNMARK_DOT,
 306         EQNMARK_DOTDOT,
 307         EQNMARK_HAT,
 308         EQNMARK_TILDE,
 309         EQNMARK_VEC,
 310         EQNMARK_DYAD,
 311         EQNMARK_BAR,
 312         EQNMARK_UNDER,
 313         EQNMARK__MAX
 314 };
 315
 316 enum    eqn_fontt {
 317         EQNFONT_NONE = 0,
 318         EQNFONT_ROMAN,
 319         EQNFONT_BOLD,
 320         EQNFONT_FAT,
 321         EQNFONT_ITALIC,
 322         EQNFONT__MAX
 323 };
 324
 325 enum    eqn_post {
 326         EQNPOS_NONE = 0,
 327         EQNPOS_OVER,
 328         EQNPOS_SUP,
 329         EQNPOS_SUB,
 330         EQNPOS_TO,
 331         EQNPOS_FROM,
 332         EQNPOS__MAX
 333 };
 334
 335 enum    eqn_pilet {
 336         EQNPILE_NONE = 0,
 337         EQNPILE_PILE,
 338         EQNPILE_CPILE,
 339         EQNPILE_RPILE,
 340         EQNPILE_LPILE,
 341         EQNPILE_COL,
 342         EQNPILE_CCOL,
 343         EQNPILE_RCOL,
 344         EQNPILE_LCOL,
 345         EQNPILE__MAX
 346 };
 347
 348  /*
 349  * A "box" is a parsed mathematical expression as defined by the eqn.7
 350  * grammar.
 351  */
 352 struct  eqn_box {
 353         int               size; /* font size of expression */
 354 #define EQN_DEFSIZE       INT_MIN
 355         enum eqn_boxt     type; /* type of node */
 356         struct eqn_box   *first; /* first child node */
 357         struct eqn_box   *last; /* last child node */
 358         struct eqn_box   *next; /* node sibling */
 359         struct eqn_box   *parent; /* node sibling */
 360         char             *text; /* text (or NULL) */
 361         char             *left;
 362         char             *right;
 363         enum eqn_post     pos; /* position of next box */
 364         enum eqn_markt    mark; /* a mark about the box */
 365         enum eqn_fontt    font; /* font of box */
 366         enum eqn_pilet    pile; /* equation piling */
 367 };
 368
 369 /*
 370  * An equation consists of a tree of expressions starting at a given
 371  * line and position.
 372  */
 373 struct  eqn {
 374         char             *name; /* identifier (or NULL) */
 375         struct eqn_box   *root; /* root mathematical expression */
 376         int               ln; /* invocation line */
 377         int               pos; /* invocation position */
 378 };
 379
 380 /*
 381  * Parse options.
 382  */
 383 #define MPARSE_MDOC     1  /* assume -mdoc */
 384 #define MPARSE_MAN      2  /* assume -man */
 385 #define MPARSE_SO       4  /* honour .so requests */
 386 #define MPARSE_QUICK    8  /* abort the parse early */
 387
 388 enum    mandoc_esc {
 389         ESCAPE_ERROR = 0, /* bail! unparsable escape */
 390         ESCAPE_IGNORE, /* escape to be ignored */
 391         ESCAPE_SPECIAL, /* a regular special character */
 392         ESCAPE_FONT, /* a generic font mode */
 393         ESCAPE_FONTBOLD, /* bold font mode */
 394         ESCAPE_FONTITALIC, /* italic font mode */
 395         ESCAPE_FONTBI, /* bold italic font mode */
 396         ESCAPE_FONTROMAN, /* roman font mode */
 397         ESCAPE_FONTPREV, /* previous font mode */
 398         ESCAPE_NUMBERED, /* a numbered glyph */
 399         ESCAPE_UNICODE, /* a unicode codepoint */
 400         ESCAPE_NOSPACE, /* suppress space if the last on a line */
 401         ESCAPE_SKIPCHAR /* skip the next character */
 402 };
 403
 404 typedef void    (*mandocmsg)(enum mandocerr, enum mandoclevel,
 405                         const char *, int, int, const char *);
 406
 407 struct  mparse;
 408 struct  mchars;
 409 struct  mdoc;
 410 struct  man;
 411
 412 __BEGIN_DECLS
 413
 414 enum mandoc_esc   mandoc_escape(const char **, const char **, int *);
 415 struct mchars    *mchars_alloc(void);
 416 void              mchars_free(struct mchars *);
 417 char              mchars_num2char(const char *, size_t);
 418 int               mchars_num2uc(const char *, size_t);
 419 int               mchars_spec2cp(const struct mchars *,
 420                         const char *, size_t);
 421 const char       *mchars_spec2str(const struct mchars *,
 422                         const char *, size_t, size_t *);
 423 struct mparse    *mparse_alloc(int, enum mandoclevel, mandocmsg,
 424                         const char *);
 425 void              mparse_free(struct mparse *);
 426 void              mparse_keep(struct mparse *);
 427 enum mandoclevel  mparse_readfd(struct mparse *, int, const char *);
 428 enum mandoclevel  mparse_readmem(struct mparse *, const void *, size_t,
 429                         const char *);
 430 void              mparse_reset(struct mparse *);
 431 void              mparse_result(struct mparse *,
 432                         struct mdoc **, struct man **, char **);
 433 const char       *mparse_getkeep(const struct mparse *);
 434 const char       *mparse_strerror(enum mandocerr);
 435 const char       *mparse_strlevel(enum mandoclevel);
 436
 437 __END_DECLS
 438
 439 #endif /*!MANDOC_H*/