mandoc.h

   1 /*      $Id: mandoc.h,v 1.179 2015/01/14 22:02:50 schwarze Exp $ */
   2 /*
   3  * Copyright (c) 2010, 2011, 2014 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
  19 #define ASCII_NBRSP      31  /* non-breaking space */
  20 #define ASCII_HYPH       30  /* breakable hyphen */
  21 #define ASCII_BREAK      29  /* breakable zero-width space */
  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_DT_NOTITLE, /* missing manual title, using UNTITLED: line */
  51         MANDOCERR_TH_NOTITLE, /* missing manual title, using "": [macro] */
  52         MANDOCERR_TITLE_CASE, /* lower case character in document title */
  53         MANDOCERR_MSEC_MISSING, /* missing manual section, using "": macro */
  54         MANDOCERR_MSEC_BAD, /* unknown manual section: Dt ... section */
  55         MANDOCERR_DATE_MISSING, /* missing date, using today's date */
  56         MANDOCERR_DATE_BAD, /* cannot parse date, using it verbatim: date */
  57         MANDOCERR_OS_MISSING, /* missing Os macro, using "" */
  58         MANDOCERR_PROLOG_REP, /* duplicate prologue macro: macro */
  59         MANDOCERR_PROLOG_LATE, /* late prologue macro: macro */
  60         MANDOCERR_DT_LATE, /* skipping late title macro: Dt args */
  61         MANDOCERR_PROLOG_ORDER, /* prologue macros out of order: macros */
  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: Sh title */
  68         MANDOCERR_NAMESEC_BAD, /* bad NAME section contents: macro */
  69         MANDOCERR_SEC_ORDER, /* sections out of conventional order: Sh title */
  70         MANDOCERR_SEC_REP, /* duplicate section title: Sh title */
  71         MANDOCERR_SEC_MSEC, /* unexpected section: Sh title for ... only */
  72         MANDOCERR_XR_ORDER, /* unusual Xr order: ... after ... */
  73         MANDOCERR_XR_PUNCT, /* unusual Xr punctuation: ... after ... */
  74         MANDOCERR_AN_MISSING, /* AUTHORS section without An macro */
  75
  76         /* related to macros and nesting */
  77         MANDOCERR_MACRO_OBS, /* obsolete macro: macro */
  78         MANDOCERR_MACRO_CALL, /* macro neither callable nor escaped: macro */
  79         MANDOCERR_PAR_SKIP, /* skipping paragraph macro: macro ... */
  80         MANDOCERR_PAR_MOVE, /* moving paragraph macro out of list: macro */
  81         MANDOCERR_NS_SKIP, /* skipping no-space macro */
  82         MANDOCERR_BLK_NEST, /* blocks badly nested: macro ... */
  83         MANDOCERR_BD_NEST, /* nested displays are not portable: macro ... */
  84         MANDOCERR_BL_MOVE, /* moving content out of list: macro */
  85         MANDOCERR_VT_CHILD, /* .Vt block has child macro: macro */
  86         MANDOCERR_FI_SKIP, /* fill mode already enabled, skipping: fi */
  87         MANDOCERR_NF_SKIP, /* fill mode already disabled, skipping: nf */
  88         MANDOCERR_BLK_LINE, /* line scope broken: macro breaks macro */
  89
  90         /* related to missing arguments */
  91         MANDOCERR_REQ_EMPTY, /* skipping empty request: request */
  92         MANDOCERR_COND_EMPTY, /* conditional request controls empty scope */
  93         MANDOCERR_MACRO_EMPTY, /* skipping empty macro: macro */
  94         MANDOCERR_ARG_EMPTY, /* empty argument, using 0n: macro arg */
  95         MANDOCERR_ARGCWARN, /* argument count wrong */
  96         MANDOCERR_BD_NOTYPE, /* missing display type, using -ragged: Bd */
  97         MANDOCERR_BL_LATETYPE, /* list type is not the first argument: Bl arg */
  98         MANDOCERR_BL_NOWIDTH, /* missing -width in -tag list, using 8n */
  99         MANDOCERR_EX_NONAME, /* missing utility name, using "": Ex */
 100         MANDOCERR_IT_NOHEAD, /* empty head in list item: Bl -type It */
 101         MANDOCERR_IT_NOBODY, /* empty list item: Bl -type It */
 102         MANDOCERR_BF_NOFONT, /* missing font type, using \fR: Bf */
 103         MANDOCERR_BF_BADFONT, /* unknown font type, using \fR: Bf font */
 104         MANDOCERR_PF_SKIP, /* nothing follows prefix: Pf arg */
 105         MANDOCERR_ARG_STD, /* missing -std argument, adding it: macro */
 106         MANDOCERR_EQN_NOBOX, /* missing eqn box, using "": op */
 107
 108         /* related to bad arguments */
 109         MANDOCERR_ARG_QUOTE, /* unterminated quoted argument */
 110         MANDOCERR_ARG_REP, /* duplicate argument: macro arg */
 111         MANDOCERR_AN_REP, /* skipping duplicate argument: An -arg */
 112         MANDOCERR_BD_REP, /* skipping duplicate display type: Bd -type */
 113         MANDOCERR_BL_REP, /* skipping duplicate list type: Bl -type */
 114         MANDOCERR_BL_SKIPW, /* skipping -width argument: Bl -type */
 115         MANDOCERR_AT_BAD, /* unknown AT&T UNIX version: At version */
 116         MANDOCERR_FA_COMMA, /* comma in function argument: arg */
 117         MANDOCERR_FN_PAREN, /* parenthesis in function name: arg */
 118         MANDOCERR_RS_BAD, /* invalid content in Rs block: macro */
 119         MANDOCERR_SM_BAD, /* invalid Boolean argument: macro arg */
 120         MANDOCERR_FT_BAD, /* unknown font, skipping request: ft font */
 121
 122         /* related to plain text */
 123         MANDOCERR_FI_BLANK, /* blank line in fill mode, using .sp */
 124         MANDOCERR_FI_TAB, /* tab in filled text */
 125         MANDOCERR_SPACE_EOL, /* whitespace at end of input line */
 126         MANDOCERR_COMMENT_BAD, /* bad comment style */
 127         MANDOCERR_ESC_BAD, /* invalid escape sequence: esc */
 128         MANDOCERR_STR_UNDEF, /* undefined string, using "": name */
 129
 130         MANDOCERR_ERROR, /* ===== start of errors ===== */
 131
 132         /* related to equations */
 133         MANDOCERR_EQNNSCOPE, /* unexpected equation scope closure*/
 134         MANDOCERR_EQNSCOPE, /* equation scope open on exit */
 135         MANDOCERR_EQNBADSCOPE, /* overlapping equation scopes */
 136         MANDOCERR_EQNEOF, /* unexpected end of equation */
 137
 138         /* related to tables */
 139         MANDOCERR_TBL, /* bad table syntax */
 140         MANDOCERR_TBLOPT, /* bad table option */
 141         MANDOCERR_TBLLAYOUT, /* bad table layout */
 142         MANDOCERR_TBLNOLAYOUT, /* no table layout cells specified */
 143         MANDOCERR_TBLNODATA, /* no table data cells specified */
 144         MANDOCERR_TBLIGNDATA, /* ignore data in cell */
 145         MANDOCERR_TBLBLOCK, /* data block still open */
 146         MANDOCERR_TBLEXTRADAT, /* ignoring extra data cells */
 147         MANDOCERR_TBLMACRO, /* ignoring macro in table: macro */
 148
 149         /* related to document structure and macros */
 150         MANDOCERR_FILE, /* cannot open file */
 151         MANDOCERR_ROFFLOOP, /* input stack limit exceeded, infinite loop? */
 152         MANDOCERR_BADCHAR, /* skipping bad character: number */
 153         MANDOCERR_MACRO, /* skipping unknown macro: macro */
 154         MANDOCERR_IT_STRAY, /* skipping item outside list: It ... */
 155         MANDOCERR_TA_STRAY, /* skipping column outside column list: Ta */
 156         MANDOCERR_BLK_NOTOPEN, /* skipping end of block that is not open */
 157         MANDOCERR_BLK_BROKEN, /* inserting missing end of block: macro ... */
 158         MANDOCERR_BLK_NOEND, /* appending missing end of block: macro */
 159
 160         /* related to request and macro arguments */
 161         MANDOCERR_NAMESC, /* escaped character not allowed in a name: name */
 162         MANDOCERR_ARGCOUNT, /* argument count wrong */
 163         MANDOCERR_BD_FILE, /* NOT IMPLEMENTED: Bd -file */
 164         MANDOCERR_BL_NOTYPE, /* missing list type, using -item: Bl */
 165         MANDOCERR_NM_NONAME, /* missing manual name, using "": Nm */
 166         MANDOCERR_OS_UNAME, /* uname(3) system call failed, using UNKNOWN */
 167         MANDOCERR_ST_BAD, /* unknown standard specifier: St standard */
 168         MANDOCERR_IT_NONUM, /* skipping request without numeric argument */
 169         MANDOCERR_ARG_SKIP, /* skipping all arguments: macro args */
 170         MANDOCERR_ARG_EXCESS, /* skipping excess arguments: macro ... args */
 171         MANDOCERR_DIVZERO, /* divide by zero */
 172
 173         MANDOCERR_FATAL, /* ===== start of fatal errors ===== */
 174
 175         MANDOCERR_TOOLARGE, /* input too large */
 176         MANDOCERR_SO_PATH, /* NOT IMPLEMENTED: .so with absolute path or ".." */
 177         MANDOCERR_SO_FAIL, /* .so request failed */
 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 #define TBL_CELL_WMAX    (1 << 7) /* x, X */
 239         struct tbl_head  *head;
 240 };
 241
 242 /*
 243  * A layout row.
 244  */
 245 struct  tbl_row {
 246         struct tbl_row   *next;
 247         struct tbl_cell  *first;
 248         struct tbl_cell  *last;
 249         int               vert; /* trailing vertical line */
 250 };
 251
 252 enum    tbl_datt {
 253         TBL_DATA_NONE, /* has no data */
 254         TBL_DATA_DATA, /* consists of data/string */
 255         TBL_DATA_HORIZ, /* horizontal line */
 256         TBL_DATA_DHORIZ, /* double-horizontal line */
 257         TBL_DATA_NHORIZ, /* squeezed horizontal line */
 258         TBL_DATA_NDHORIZ /* squeezed double-horizontal line */
 259 };
 260
 261 /*
 262  * A cell within a row of data.  The "string" field contains the actual
 263  * string value that's in the cell.  The rest is layout.
 264  */
 265 struct  tbl_dat {
 266         struct tbl_cell  *layout; /* layout cell */
 267         int               spans; /* how many spans follow */
 268         struct tbl_dat   *next;
 269         char             *string; /* data (NULL if not TBL_DATA_DATA) */
 270         enum tbl_datt     pos;
 271 };
 272
 273 enum    tbl_spant {
 274         TBL_SPAN_DATA, /* span consists of data */
 275         TBL_SPAN_HORIZ, /* span is horizontal line */
 276         TBL_SPAN_DHORIZ /* span is double horizontal line */
 277 };
 278
 279 /*
 280  * A row of data in a table.
 281  */
 282 struct  tbl_span {
 283         struct tbl_opts  *opts;
 284         struct tbl_head  *head;
 285         struct tbl_row   *layout; /* layout row */
 286         struct tbl_dat   *first;
 287         struct tbl_dat   *last;
 288         int               line; /* parse line */
 289         int               flags;
 290 #define TBL_SPAN_FIRST   (1 << 0)
 291 #define TBL_SPAN_LAST    (1 << 1)
 292         enum tbl_spant    pos;
 293         struct tbl_span  *next;
 294 };
 295
 296 enum    eqn_boxt {
 297         EQN_ROOT, /* root of parse tree */
 298         EQN_TEXT, /* text (number, variable, whatever) */
 299         EQN_SUBEXPR, /* nested `eqn' subexpression */
 300         EQN_LIST, /* list (braces, etc.) */
 301         EQN_LISTONE, /* singleton list */
 302         EQN_PILE, /* vertical pile */
 303         EQN_MATRIX /* pile of piles */
 304 };
 305
 306 enum    eqn_fontt {
 307         EQNFONT_NONE = 0,
 308         EQNFONT_ROMAN,
 309         EQNFONT_BOLD,
 310         EQNFONT_FAT,
 311         EQNFONT_ITALIC,
 312         EQNFONT__MAX
 313 };
 314
 315 enum    eqn_post {
 316         EQNPOS_NONE = 0,
 317         EQNPOS_SUP,
 318         EQNPOS_SUBSUP,
 319         EQNPOS_SUB,
 320         EQNPOS_TO,
 321         EQNPOS_FROM,
 322         EQNPOS_FROMTO,
 323         EQNPOS_OVER,
 324         EQNPOS_SQRT,
 325         EQNPOS__MAX
 326 };
 327
 328 enum    eqn_pilet {
 329         EQNPILE_NONE = 0,
 330         EQNPILE_PILE,
 331         EQNPILE_CPILE,
 332         EQNPILE_RPILE,
 333         EQNPILE_LPILE,
 334         EQNPILE_COL,
 335         EQNPILE_CCOL,
 336         EQNPILE_RCOL,
 337         EQNPILE_LCOL,
 338         EQNPILE__MAX
 339 };
 340
 341  /*
 342  * A "box" is a parsed mathematical expression as defined by the eqn.7
 343  * grammar.
 344  */
 345 struct  eqn_box {
 346         int               size; /* font size of expression */
 347 #define EQN_DEFSIZE       INT_MIN
 348         enum eqn_boxt     type; /* type of node */
 349         struct eqn_box   *first; /* first child node */
 350         struct eqn_box   *last; /* last child node */
 351         struct eqn_box   *next; /* node sibling */
 352         struct eqn_box   *prev; /* node sibling */
 353         struct eqn_box   *parent; /* node sibling */
 354         char             *text; /* text (or NULL) */
 355         char             *left; /* fence left-hand */
 356         char             *right; /* fence right-hand */
 357         char             *top; /* expression over-symbol */
 358         char             *bottom; /* expression under-symbol */
 359         size_t            args; /* arguments in parent */
 360         size_t            expectargs; /* max arguments in parent */
 361         enum eqn_post     pos; /* position of next 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 #define MPARSE_UTF8     16 /* accept UTF-8 input */
 385 #define MPARSE_LATIN1   32 /* accept ISO-LATIN-1 input */
 386
 387 enum    mandoc_esc {
 388         ESCAPE_ERROR = 0, /* bail! unparsable escape */
 389         ESCAPE_IGNORE, /* escape to be ignored */
 390         ESCAPE_SPECIAL, /* a regular special character */
 391         ESCAPE_FONT, /* a generic font mode */
 392         ESCAPE_FONTBOLD, /* bold font mode */
 393         ESCAPE_FONTITALIC, /* italic font mode */
 394         ESCAPE_FONTBI, /* bold italic font mode */
 395         ESCAPE_FONTROMAN, /* roman font mode */
 396         ESCAPE_FONTPREV, /* previous font mode */
 397         ESCAPE_NUMBERED, /* a numbered glyph */
 398         ESCAPE_UNICODE, /* a unicode codepoint */
 399         ESCAPE_NOSPACE, /* suppress space if the last on a line */
 400         ESCAPE_SKIPCHAR /* skip the next character */
 401 };
 402
 403 typedef void    (*mandocmsg)(enum mandocerr, enum mandoclevel,
 404                         const char *, int, int, const char *);
 405
 406 __BEGIN_DECLS
 407
 408 struct  mparse;
 409 struct  mchars;
 410 struct  mdoc;
 411 struct  man;
 412
 413 enum mandoc_esc   mandoc_escape(const char **, const char **, int *);
 414 struct mchars    *mchars_alloc(void);
 415 void              mchars_free(struct mchars *);
 416 int               mchars_num2char(const char *, size_t);
 417 const char       *mchars_uc2str(int);
 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 struct mchars *, const char *);
 425 void              mparse_free(struct mparse *);
 426 void              mparse_keep(struct mparse *);
 427 enum mandoclevel  mparse_open(struct mparse *, int *, const char *);
 428 enum mandoclevel  mparse_readfd(struct mparse *, int, const char *);
 429 enum mandoclevel  mparse_readmem(struct mparse *, void *, size_t,
 430                         const char *);
 431 void              mparse_reset(struct mparse *);
 432 void              mparse_result(struct mparse *,
 433                         struct mdoc **, struct man **, char **);
 434 const char       *mparse_getkeep(const struct mparse *);
 435 const char       *mparse_strerror(enum mandocerr);
 436 const char       *mparse_strlevel(enum mandoclevel);
 437 enum mandoclevel  mparse_wait(struct mparse *);
 438
 439 __END_DECLS