bin/mandoc/mandoc.h

   1 /*      $Id: mandoc.h,v 1.214 2017/01/28 23:30:08 schwarze Exp $ */
   2 /*
   3  * Copyright (c) 2010, 2011, 2014 Kristaps Dzonsons <kristaps@bsd.lv>
   4  * Copyright (c) 2010-2017 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 AUTHORS DISCLAIM ALL WARRANTIES
  11  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
  12  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHORS 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_UNSUPP, /* input needs unimplemented features */
  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_NONM, /* NAME section without Nm before Nd */
  69         MANDOCERR_NAMESEC_NOND, /* NAME section without description */
  70         MANDOCERR_NAMESEC_ND, /* description not at the end of NAME */
  71         MANDOCERR_NAMESEC_BAD, /* bad NAME section content: macro */
  72         MANDOCERR_NAMESEC_PUNCT, /* missing comma before name: Nm name */
  73         MANDOCERR_ND_EMPTY, /* missing description line, using "" */
  74         MANDOCERR_SEC_ORDER, /* sections out of conventional order: Sh title */
  75         MANDOCERR_SEC_REP, /* duplicate section title: Sh title */
  76         MANDOCERR_SEC_MSEC, /* unexpected section: Sh title for ... only */
  77         MANDOCERR_XR_ORDER, /* unusual Xr order: ... after ... */
  78         MANDOCERR_XR_PUNCT, /* unusual Xr punctuation: ... after ... */
  79         MANDOCERR_AN_MISSING, /* AUTHORS section without An macro */
  80
  81         /* related to macros and nesting */
  82         MANDOCERR_MACRO_OBS, /* obsolete macro: macro */
  83         MANDOCERR_MACRO_CALL, /* macro neither callable nor escaped: macro */
  84         MANDOCERR_PAR_SKIP, /* skipping paragraph macro: macro ... */
  85         MANDOCERR_PAR_MOVE, /* moving paragraph macro out of list: macro */
  86         MANDOCERR_NS_SKIP, /* skipping no-space macro */
  87         MANDOCERR_BLK_NEST, /* blocks badly nested: macro ... */
  88         MANDOCERR_BD_NEST, /* nested displays are not portable: macro ... */
  89         MANDOCERR_BL_MOVE, /* moving content out of list: macro */
  90         MANDOCERR_FI_SKIP, /* fill mode already enabled, skipping: fi */
  91         MANDOCERR_NF_SKIP, /* fill mode already disabled, skipping: nf */
  92         MANDOCERR_BLK_LINE, /* line scope broken: macro breaks macro */
  93
  94         /* related to missing arguments */
  95         MANDOCERR_REQ_EMPTY, /* skipping empty request: request */
  96         MANDOCERR_COND_EMPTY, /* conditional request controls empty scope */
  97         MANDOCERR_MACRO_EMPTY, /* skipping empty macro: macro */
  98         MANDOCERR_BLK_EMPTY, /* empty block: macro */
  99         MANDOCERR_ARG_EMPTY, /* empty argument, using 0n: macro arg */
 100         MANDOCERR_BD_NOTYPE, /* missing display type, using -ragged: Bd */
 101         MANDOCERR_BL_LATETYPE, /* list type is not the first argument: Bl arg */
 102         MANDOCERR_BL_NOWIDTH, /* missing -width in -tag list, using 6n */
 103         MANDOCERR_EX_NONAME, /* missing utility name, using "": Ex */
 104         MANDOCERR_FO_NOHEAD, /* missing function name, using "": Fo */
 105         MANDOCERR_IT_NOHEAD, /* empty head in list item: Bl -type It */
 106         MANDOCERR_IT_NOBODY, /* empty list item: Bl -type It */
 107         MANDOCERR_BF_NOFONT, /* missing font type, using \fR: Bf */
 108         MANDOCERR_BF_BADFONT, /* unknown font type, using \fR: Bf font */
 109         MANDOCERR_PF_SKIP, /* nothing follows prefix: Pf arg */
 110         MANDOCERR_RS_EMPTY, /* empty reference block: Rs */
 111         MANDOCERR_XR_NOSEC, /* missing section argument: Xr arg */
 112         MANDOCERR_ARG_STD, /* missing -std argument, adding it: macro */
 113         MANDOCERR_OP_EMPTY, /* missing option string, using "": OP */
 114         MANDOCERR_UR_NOHEAD, /* missing resource identifier, using "": UR */
 115         MANDOCERR_EQN_NOBOX, /* missing eqn box, using "": op */
 116
 117         /* related to bad arguments */
 118         MANDOCERR_ARG_QUOTE, /* unterminated quoted argument */
 119         MANDOCERR_ARG_REP, /* duplicate argument: macro arg */
 120         MANDOCERR_AN_REP, /* skipping duplicate argument: An -arg */
 121         MANDOCERR_BD_REP, /* skipping duplicate display type: Bd -type */
 122         MANDOCERR_BL_REP, /* skipping duplicate list type: Bl -type */
 123         MANDOCERR_BL_SKIPW, /* skipping -width argument: Bl -type */
 124         MANDOCERR_BL_COL, /* wrong number of cells */
 125         MANDOCERR_AT_BAD, /* unknown AT&T UNIX version: At version */
 126         MANDOCERR_FA_COMMA, /* comma in function argument: arg */
 127         MANDOCERR_FN_PAREN, /* parenthesis in function name: arg */
 128         MANDOCERR_RS_BAD, /* invalid content in Rs block: macro */
 129         MANDOCERR_SM_BAD, /* invalid Boolean argument: macro arg */
 130         MANDOCERR_FT_BAD, /* unknown font, skipping request: ft font */
 131         MANDOCERR_TR_ODD, /* odd number of characters in request: tr char */
 132
 133         /* related to plain text */
 134         MANDOCERR_FI_BLANK, /* blank line in fill mode, using .sp */
 135         MANDOCERR_FI_TAB, /* tab in filled text */
 136         MANDOCERR_SPACE_EOL, /* whitespace at end of input line */
 137         MANDOCERR_EOS, /* new sentence, new line */
 138         MANDOCERR_COMMENT_BAD, /* bad comment style */
 139         MANDOCERR_ESC_BAD, /* invalid escape sequence: esc */
 140         MANDOCERR_STR_UNDEF, /* undefined string, using "": name */
 141
 142         /* related to tables */
 143         MANDOCERR_TBLLAYOUT_SPAN, /* tbl line starts with span */
 144         MANDOCERR_TBLLAYOUT_DOWN, /* tbl column starts with span */
 145         MANDOCERR_TBLLAYOUT_VERT, /* skipping vertical bar in tbl layout */
 146
 147         MANDOCERR_ERROR, /* ===== start of errors ===== */
 148
 149         /* related to tables */
 150         MANDOCERR_TBLOPT_ALPHA, /* non-alphabetic character in tbl options */
 151         MANDOCERR_TBLOPT_BAD, /* skipping unknown tbl option: option */
 152         MANDOCERR_TBLOPT_NOARG, /* missing tbl option argument: option */
 153         MANDOCERR_TBLOPT_ARGSZ, /* wrong tbl option argument size: option */
 154         MANDOCERR_TBLLAYOUT_NONE, /* empty tbl layout */
 155         MANDOCERR_TBLLAYOUT_CHAR, /* invalid character in tbl layout: char */
 156         MANDOCERR_TBLLAYOUT_PAR, /* unmatched parenthesis in tbl layout */
 157         MANDOCERR_TBLDATA_NONE, /* tbl without any data cells */
 158         MANDOCERR_TBLDATA_SPAN, /* ignoring data in spanned tbl cell: data */
 159         MANDOCERR_TBLDATA_EXTRA, /* ignoring extra tbl data cells: data */
 160         MANDOCERR_TBLDATA_BLK, /* data block open at end of tbl: macro */
 161
 162         /* related to document structure and macros */
 163         MANDOCERR_FILE, /* cannot open file */
 164         MANDOCERR_ROFFLOOP, /* input stack limit exceeded, infinite loop? */
 165         MANDOCERR_CHAR_BAD, /* skipping bad character: number */
 166         MANDOCERR_MACRO, /* skipping unknown macro: macro */
 167         MANDOCERR_REQ_INSEC, /* skipping insecure request: request */
 168         MANDOCERR_IT_STRAY, /* skipping item outside list: It ... */
 169         MANDOCERR_TA_STRAY, /* skipping column outside column list: Ta */
 170         MANDOCERR_BLK_NOTOPEN, /* skipping end of block that is not open */
 171         MANDOCERR_RE_NOTOPEN, /* fewer RS blocks open, skipping: RE arg */
 172         MANDOCERR_BLK_BROKEN, /* inserting missing end of block: macro ... */
 173         MANDOCERR_BLK_NOEND, /* appending missing end of block: macro */
 174
 175         /* related to request and macro arguments */
 176         MANDOCERR_NAMESC, /* escaped character not allowed in a name: name */
 177         MANDOCERR_BD_FILE, /* NOT IMPLEMENTED: Bd -file */
 178         MANDOCERR_BD_NOARG, /* skipping display without arguments: Bd */
 179         MANDOCERR_BL_NOTYPE, /* missing list type, using -item: Bl */
 180         MANDOCERR_NM_NONAME, /* missing manual name, using "": Nm */
 181         MANDOCERR_OS_UNAME, /* uname(3) system call failed, using UNKNOWN */
 182         MANDOCERR_ST_BAD, /* unknown standard specifier: St standard */
 183         MANDOCERR_IT_NONUM, /* skipping request without numeric argument */
 184         MANDOCERR_SO_PATH, /* NOT IMPLEMENTED: .so with absolute path or ".." */
 185         MANDOCERR_SO_FAIL, /* .so request failed */
 186         MANDOCERR_ARG_SKIP, /* skipping all arguments: macro args */
 187         MANDOCERR_ARG_EXCESS, /* skipping excess arguments: macro ... args */
 188         MANDOCERR_DIVZERO, /* divide by zero */
 189
 190         MANDOCERR_UNSUPP, /* ===== start of unsupported features ===== */
 191
 192         MANDOCERR_TOOLARGE, /* input too large */
 193         MANDOCERR_CHAR_UNSUPP, /* unsupported control character: number */
 194         MANDOCERR_REQ_UNSUPP, /* unsupported roff request: request */
 195         MANDOCERR_TBLOPT_EQN, /* eqn delim option in tbl: arg */
 196         MANDOCERR_TBLLAYOUT_MOD, /* unsupported tbl layout modifier: m */
 197         MANDOCERR_TBLMACRO, /* ignoring macro in table: macro */
 198
 199         MANDOCERR_MAX
 200 };
 201
 202 struct  tbl_opts {
 203         char              tab; /* cell-separator */
 204         char              decimal; /* decimal point */
 205         int               opts;
 206 #define TBL_OPT_CENTRE   (1 << 0)
 207 #define TBL_OPT_EXPAND   (1 << 1)
 208 #define TBL_OPT_BOX      (1 << 2)
 209 #define TBL_OPT_DBOX     (1 << 3)
 210 #define TBL_OPT_ALLBOX   (1 << 4)
 211 #define TBL_OPT_NOKEEP   (1 << 5)
 212 #define TBL_OPT_NOSPACE  (1 << 6)
 213 #define TBL_OPT_NOWARN   (1 << 7)
 214         int               cols; /* number of columns */
 215         int               lvert; /* width of left vertical line */
 216         int               rvert; /* width of right vertical line */
 217 };
 218
 219 enum    tbl_cellt {
 220         TBL_CELL_CENTRE, /* c, C */
 221         TBL_CELL_RIGHT, /* r, R */
 222         TBL_CELL_LEFT, /* l, L */
 223         TBL_CELL_NUMBER, /* n, N */
 224         TBL_CELL_SPAN, /* s, S */
 225         TBL_CELL_LONG, /* a, A */
 226         TBL_CELL_DOWN, /* ^ */
 227         TBL_CELL_HORIZ, /* _, - */
 228         TBL_CELL_DHORIZ, /* = */
 229         TBL_CELL_MAX
 230 };
 231
 232 /*
 233  * A cell in a layout row.
 234  */
 235 struct  tbl_cell {
 236         struct tbl_cell  *next;
 237         int               vert; /* width of subsequent vertical line */
 238         enum tbl_cellt    pos;
 239         size_t            spacing;
 240         int               col; /* column number, starting from 0 */
 241         int               flags;
 242 #define TBL_CELL_TALIGN  (1 << 0) /* t, T */
 243 #define TBL_CELL_BALIGN  (1 << 1) /* d, D */
 244 #define TBL_CELL_BOLD    (1 << 2) /* fB, B, b */
 245 #define TBL_CELL_ITALIC  (1 << 3) /* fI, I, i */
 246 #define TBL_CELL_EQUAL   (1 << 4) /* e, E */
 247 #define TBL_CELL_UP      (1 << 5) /* u, U */
 248 #define TBL_CELL_WIGN    (1 << 6) /* z, Z */
 249 #define TBL_CELL_WMAX    (1 << 7) /* x, X */
 250 };
 251
 252 /*
 253  * A layout row.
 254  */
 255 struct  tbl_row {
 256         struct tbl_row   *next;
 257         struct tbl_cell  *first;
 258         struct tbl_cell  *last;
 259         int               vert; /* width of left vertical line */
 260 };
 261
 262 enum    tbl_datt {
 263         TBL_DATA_NONE, /* has no data */
 264         TBL_DATA_DATA, /* consists of data/string */
 265         TBL_DATA_HORIZ, /* horizontal line */
 266         TBL_DATA_DHORIZ, /* double-horizontal line */
 267         TBL_DATA_NHORIZ, /* squeezed horizontal line */
 268         TBL_DATA_NDHORIZ /* squeezed double-horizontal line */
 269 };
 270
 271 /*
 272  * A cell within a row of data.  The "string" field contains the actual
 273  * string value that's in the cell.  The rest is layout.
 274  */
 275 struct  tbl_dat {
 276         struct tbl_cell  *layout; /* layout cell */
 277         int               spans; /* how many spans follow */
 278         struct tbl_dat   *next;
 279         char             *string; /* data (NULL if not TBL_DATA_DATA) */
 280         enum tbl_datt     pos;
 281 };
 282
 283 enum    tbl_spant {
 284         TBL_SPAN_DATA, /* span consists of data */
 285         TBL_SPAN_HORIZ, /* span is horizontal line */
 286         TBL_SPAN_DHORIZ /* span is double horizontal line */
 287 };
 288
 289 /*
 290  * A row of data in a table.
 291  */
 292 struct  tbl_span {
 293         struct tbl_opts  *opts;
 294         struct tbl_row   *layout; /* layout row */
 295         struct tbl_dat   *first;
 296         struct tbl_dat   *last;
 297         struct tbl_span  *prev;
 298         struct tbl_span  *next;
 299         int               line; /* parse line */
 300         enum tbl_spant    pos;
 301 };
 302
 303 enum    eqn_boxt {
 304         EQN_ROOT, /* root of parse tree */
 305         EQN_TEXT, /* text (number, variable, whatever) */
 306         EQN_SUBEXPR, /* nested `eqn' subexpression */
 307         EQN_LIST, /* list (braces, etc.) */
 308         EQN_LISTONE, /* singleton list */
 309         EQN_PILE, /* vertical pile */
 310         EQN_MATRIX /* pile of piles */
 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_SUP,
 325         EQNPOS_SUBSUP,
 326         EQNPOS_SUB,
 327         EQNPOS_TO,
 328         EQNPOS_FROM,
 329         EQNPOS_FROMTO,
 330         EQNPOS_OVER,
 331         EQNPOS_SQRT,
 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   *prev; /* node sibling */
 360         struct eqn_box   *parent; /* node sibling */
 361         char             *text; /* text (or NULL) */
 362         char             *left; /* fence left-hand */
 363         char             *right; /* fence right-hand */
 364         char             *top; /* expression over-symbol */
 365         char             *bottom; /* expression under-symbol */
 366         size_t            args; /* arguments in parent */
 367         size_t            expectargs; /* max arguments in parent */
 368         enum eqn_post     pos; /* position of next box */
 369         enum eqn_fontt    font; /* font of box */
 370         enum eqn_pilet    pile; /* equation piling */
 371 };
 372
 373 /*
 374  * An equation consists of a tree of expressions starting at a given
 375  * line and position.
 376  */
 377 struct  eqn {
 378         char             *name; /* identifier (or NULL) */
 379         struct eqn_box   *root; /* root mathematical expression */
 380         int               ln; /* invocation line */
 381         int               pos; /* invocation position */
 382 };
 383
 384 /*
 385  * Parse options.
 386  */
 387 #define MPARSE_MDOC     1  /* assume -mdoc */
 388 #define MPARSE_MAN      2  /* assume -man */
 389 #define MPARSE_SO       4  /* honour .so requests */
 390 #define MPARSE_QUICK    8  /* abort the parse early */
 391 #define MPARSE_UTF8     16 /* accept UTF-8 input */
 392 #define MPARSE_LATIN1   32 /* accept ISO-LATIN-1 input */
 393
 394 enum    mandoc_esc {
 395         ESCAPE_ERROR = 0, /* bail! unparsable escape */
 396         ESCAPE_IGNORE, /* escape to be ignored */
 397         ESCAPE_SPECIAL, /* a regular special character */
 398         ESCAPE_FONT, /* a generic font mode */
 399         ESCAPE_FONTBOLD, /* bold font mode */
 400         ESCAPE_FONTITALIC, /* italic font mode */
 401         ESCAPE_FONTBI, /* bold italic font mode */
 402         ESCAPE_FONTROMAN, /* roman font mode */
 403         ESCAPE_FONTPREV, /* previous font mode */
 404         ESCAPE_NUMBERED, /* a numbered glyph */
 405         ESCAPE_UNICODE, /* a unicode codepoint */
 406         ESCAPE_NOSPACE, /* suppress space if the last on a line */
 407         ESCAPE_SKIPCHAR, /* skip the next character */
 408         ESCAPE_OVERSTRIKE /* overstrike all chars in the argument */
 409 };
 410
 411 typedef void    (*mandocmsg)(enum mandocerr, enum mandoclevel,
 412                         const char *, int, int, const char *);
 413
 414
 415 struct  mparse;
 416 struct  roff_man;
 417
 418 enum mandoc_esc   mandoc_escape(const char **, const char **, int *);
 419 void              mchars_alloc(void);
 420 void              mchars_free(void);
 421 int               mchars_num2char(const char *, size_t);
 422 const char       *mchars_uc2str(int);
 423 int               mchars_num2uc(const char *, size_t);
 424 int               mchars_spec2cp(const char *, size_t);
 425 const char       *mchars_spec2str(const char *, size_t, size_t *);
 426 struct mparse    *mparse_alloc(int, enum mandoclevel, mandocmsg, const char *);
 427 void              mparse_free(struct mparse *);
 428 void              mparse_keep(struct mparse *);
 429 int               mparse_open(struct mparse *, const char *);
 430 enum mandoclevel  mparse_readfd(struct mparse *, int, const char *);
 431 enum mandoclevel  mparse_readmem(struct mparse *, void *, size_t,
 432                         const char *);
 433 void              mparse_reset(struct mparse *);
 434 void              mparse_result(struct mparse *,
 435                         struct roff_man **, char **);
 436 const char       *mparse_getkeep(const struct mparse *);
 437 const char       *mparse_strerror(enum mandocerr);
 438 const char       *mparse_strlevel(enum mandoclevel);
 439 void              mparse_updaterc(struct mparse *, enum mandoclevel *);