mandoc.h

   1 /*      $Id: mandoc.h,v 1.62 2011/03/20 11:41:24 kristaps Exp $ */
   2 /*
   3  * Copyright (c) 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
   4  *
   5  * Permission to use, copy, modify, and distribute this software for any
   6  * purpose with or without fee is hereby granted, provided that the above
   7  * copyright notice and this permission notice appear in all copies.
   8  *
   9  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
  10  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
  11  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
  12  * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
  13  * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
  14  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
  15  * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
  16  */
  17 #ifndef MANDOC_H
  18 #define MANDOC_H
  19
  20 #define ASCII_NBRSP      31  /* non-breaking space */
  21 #define ASCII_HYPH       30  /* breakable hyphen */
  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_NOTITLE, /* no title in document */
  51         MANDOCERR_UPPERCASE, /* document title should be all caps */
  52         MANDOCERR_BADMSEC, /* unknown manual section */
  53         MANDOCERR_NODATE, /* date missing, using today's date */
  54         MANDOCERR_BADDATE, /* cannot parse date, using it verbatim */
  55         MANDOCERR_PROLOGOOO, /* prologue macros out of order */
  56         MANDOCERR_PROLOGREP, /* duplicate prologue macro */
  57         MANDOCERR_BADPROLOG, /* macro not allowed in prologue */
  58         MANDOCERR_BADBODY, /* macro not allowed in body */
  59
  60         /* related to document structure */
  61         MANDOCERR_SO, /* .so is fragile, better use ln(1) */
  62         MANDOCERR_NAMESECFIRST, /* NAME section must come first */
  63         MANDOCERR_BADNAMESEC, /* bad NAME section contents */
  64         MANDOCERR_NONAME, /* manual name not yet set */
  65         MANDOCERR_SECOOO, /* sections out of conventional order */
  66         MANDOCERR_SECREP, /* duplicate section name */
  67         MANDOCERR_SECMSEC, /* section not in conventional manual section */
  68
  69         /* related to macros and nesting */
  70         MANDOCERR_MACROOBS, /* skipping obsolete macro */
  71         MANDOCERR_IGNPAR, /* skipping paragraph macro */
  72         MANDOCERR_IGNNS, /* skipping no-space macro */
  73         MANDOCERR_SCOPENEST, /* blocks badly nested */
  74         MANDOCERR_CHILD, /* child violates parent syntax */
  75         MANDOCERR_NESTEDDISP, /* nested displays are not portable */
  76         MANDOCERR_SCOPEREP, /* already in literal mode */
  77
  78         /* related to missing macro arguments */
  79         MANDOCERR_MACROEMPTY, /* skipping empty macro */
  80         MANDOCERR_ARGCWARN, /* argument count wrong */
  81         MANDOCERR_DISPTYPE, /* missing display type */
  82         MANDOCERR_LISTFIRST, /* list type must come first */
  83         MANDOCERR_NOWIDTHARG, /* tag lists require a width argument */
  84         MANDOCERR_FONTTYPE, /* missing font type */
  85         MANDOCERR_WNOSCOPE, /* skipping end of block that is not open */
  86
  87         /* related to bad macro arguments */
  88         MANDOCERR_IGNARGV, /* skipping argument */
  89         MANDOCERR_ARGVREP, /* duplicate argument */
  90         MANDOCERR_DISPREP, /* duplicate display type */
  91         MANDOCERR_LISTREP, /* duplicate list type */
  92         MANDOCERR_BADATT, /* unknown AT&T UNIX version */
  93         MANDOCERR_BADBOOL, /* bad Boolean value */
  94         MANDOCERR_BADFONT, /* unknown font */
  95         MANDOCERR_BADSTANDARD, /* unknown standard specifier */
  96         MANDOCERR_BADWIDTH, /* bad width argument */
  97
  98         /* related to plain text */
  99         MANDOCERR_NOBLANKLN, /* blank line in non-literal context */
 100         MANDOCERR_BADTAB, /* tab in non-literal context */
 101         MANDOCERR_EOLNSPACE, /* end of line whitespace */
 102         MANDOCERR_BADCOMMENT, /* bad comment style */
 103         MANDOCERR_BADESCAPE, /* unknown escape sequence */
 104         MANDOCERR_BADQUOTE, /* unterminated quoted string */
 105
 106         MANDOCERR_ERROR, /* ===== start of errors ===== */
 107
 108         /* related to tables */
 109         MANDOCERR_TBL, /* bad table syntax */
 110         MANDOCERR_TBLOPT, /* bad table option */
 111         MANDOCERR_TBLLAYOUT, /* bad table layout */
 112         MANDOCERR_TBLNOLAYOUT, /* no table layout cells specified */
 113         MANDOCERR_TBLNODATA, /* no table data cells specified */
 114         MANDOCERR_TBLIGNDATA, /* ignore data in cell */
 115         MANDOCERR_TBLBLOCK, /* data block still open */
 116         MANDOCERR_TBLEXTRADAT, /* ignoring extra data cells */
 117
 118         MANDOCERR_ROFFLOOP, /* input stack limit exceeded, infinite loop? */
 119         MANDOCERR_BADCHAR, /* skipping bad character */
 120         MANDOCERR_NAMESC, /* escaped character not allowed in a name */
 121         MANDOCERR_NOTEXT, /* skipping text before the first section header */
 122         MANDOCERR_MACRO, /* skipping unknown macro */
 123         MANDOCERR_REQUEST, /* NOT IMPLEMENTED: skipping request */
 124         MANDOCERR_LINESCOPE, /* line scope broken */
 125         MANDOCERR_ARGCOUNT, /* argument count wrong */
 126         MANDOCERR_NOSCOPE, /* skipping end of block that is not open */
 127         MANDOCERR_SCOPEBROKEN, /* missing end of block */
 128         MANDOCERR_SCOPEEXIT, /* scope open on exit */
 129         MANDOCERR_UNAME, /* uname(3) system call failed */
 130         /* FIXME: merge following with MANDOCERR_ARGCOUNT */
 131         MANDOCERR_NOARGS, /* macro requires line argument(s) */
 132         MANDOCERR_NOBODY, /* macro requires body argument(s) */
 133         MANDOCERR_NOARGV, /* macro requires argument(s) */
 134         MANDOCERR_LISTTYPE, /* missing list type */
 135         MANDOCERR_ARGSLOST, /* line argument(s) will be lost */
 136         MANDOCERR_BODYLOST, /* body argument(s) will be lost */
 137
 138         MANDOCERR_FATAL, /* ===== start of fatal errors ===== */
 139
 140         MANDOCERR_COLUMNS, /* column syntax is inconsistent */
 141         MANDOCERR_BADDISP, /* NOT IMPLEMENTED: .Bd -file */
 142         MANDOCERR_SYNTLINESCOPE, /* line scope broken, syntax violated */
 143         MANDOCERR_SYNTARGVCOUNT, /* argument count wrong, violates syntax */
 144         MANDOCERR_SYNTCHILD, /* child violates parent syntax */
 145         MANDOCERR_SYNTARGCOUNT, /* argument count wrong, violates syntax */
 146         MANDOCERR_SOPATH, /* NOT IMPLEMENTED: .so with absolute path or ".." */
 147         MANDOCERR_NODOCBODY, /* no document body */
 148         MANDOCERR_NODOCPROLOG, /* no document prologue */
 149         MANDOCERR_MEM, /* static buffer exhausted */
 150         MANDOCERR_MAX
 151 };
 152
 153 struct  tbl {
 154         char              tab; /* cell-separator */
 155         char              decimal; /* decimal point */
 156         int               linesize;
 157         int               opts;
 158 #define TBL_OPT_CENTRE   (1 << 0)
 159 #define TBL_OPT_EXPAND   (1 << 1)
 160 #define TBL_OPT_BOX      (1 << 2)
 161 #define TBL_OPT_DBOX     (1 << 3)
 162 #define TBL_OPT_ALLBOX   (1 << 4)
 163 #define TBL_OPT_NOKEEP   (1 << 5)
 164 #define TBL_OPT_NOSPACE  (1 << 6)
 165         int               cols; /* number of columns */
 166 };
 167
 168 enum    tbl_headt {
 169         TBL_HEAD_DATA, /* plug in data from tbl_dat */
 170         TBL_HEAD_VERT, /* vertical spacer */
 171         TBL_HEAD_DVERT  /* double-vertical spacer */
 172 };
 173
 174 /*
 175  * The head of a table specifies all of its columns.  When formatting a
 176  * tbl_span, iterate over these and plug in data from the tbl_span when
 177  * appropriate, using tbl_cell as a guide to placement.
 178  */
 179 struct  tbl_head {
 180         enum tbl_headt    pos;
 181         int               ident; /* 0 <= unique id < cols */
 182         struct tbl_head  *next;
 183         struct tbl_head  *prev;
 184 };
 185
 186 enum    tbl_cellt {
 187         TBL_CELL_CENTRE, /* c, C */
 188         TBL_CELL_RIGHT, /* r, R */
 189         TBL_CELL_LEFT, /* l, L */
 190         TBL_CELL_NUMBER, /* n, N */
 191         TBL_CELL_SPAN, /* s, S */
 192         TBL_CELL_LONG, /* a, A */
 193         TBL_CELL_DOWN, /* ^ */
 194         TBL_CELL_HORIZ, /* _, - */
 195         TBL_CELL_DHORIZ, /* = */
 196         TBL_CELL_VERT, /* | */
 197         TBL_CELL_DVERT, /* || */
 198         TBL_CELL_MAX
 199 };
 200
 201 /*
 202  * A cell in a layout row.
 203  */
 204 struct  tbl_cell {
 205         struct tbl_cell  *next;
 206         enum tbl_cellt    pos;
 207         size_t            spacing;
 208         int               flags;
 209 #define TBL_CELL_TALIGN  (1 << 0) /* t, T */
 210 #define TBL_CELL_BALIGN  (1 << 1) /* d, D */
 211 #define TBL_CELL_BOLD    (1 << 2) /* fB, B, b */
 212 #define TBL_CELL_ITALIC  (1 << 3) /* fI, I, i */
 213 #define TBL_CELL_EQUAL   (1 << 4) /* e, E */
 214 #define TBL_CELL_UP      (1 << 5) /* u, U */
 215 #define TBL_CELL_WIGN    (1 << 6) /* z, Z */
 216         struct tbl_head  *head;
 217 };
 218
 219 /*
 220  * A layout row.
 221  */
 222 struct  tbl_row {
 223         struct tbl_row   *next;
 224         struct tbl_cell  *first;
 225         struct tbl_cell  *last;
 226 };
 227
 228 enum    tbl_datt {
 229         TBL_DATA_NONE, /* has no data */
 230         TBL_DATA_DATA, /* consists of data/string */
 231         TBL_DATA_HORIZ, /* horizontal line */
 232         TBL_DATA_DHORIZ, /* double-horizontal line */
 233         TBL_DATA_NHORIZ, /* squeezed horizontal line */
 234         TBL_DATA_NDHORIZ /* squeezed double-horizontal line */
 235 };
 236
 237 /*
 238  * A cell within a row of data.  The "string" field contains the actual
 239  * string value that's in the cell.  The rest is layout.
 240  */
 241 struct  tbl_dat {
 242         struct tbl_cell  *layout; /* layout cell */
 243         int               spans; /* how many spans follow */
 244         struct tbl_dat   *next;
 245         char             *string; /* data (NULL if not TBL_DATA_DATA) */
 246         enum tbl_datt     pos;
 247 };
 248
 249 enum    tbl_spant {
 250         TBL_SPAN_DATA, /* span consists of data */
 251         TBL_SPAN_HORIZ, /* span is horizontal line */
 252         TBL_SPAN_DHORIZ /* span is double horizontal line */
 253 };
 254
 255 /*
 256  * A row of data in a table.
 257  */
 258 struct  tbl_span {
 259         struct tbl       *tbl;
 260         struct tbl_head  *head;
 261         struct tbl_row   *layout; /* layout row */
 262         struct tbl_dat   *first;
 263         struct tbl_dat   *last;
 264         int               line; /* parse line */
 265         int               flags;
 266 #define TBL_SPAN_FIRST   (1 << 0)
 267 #define TBL_SPAN_LAST    (1 << 1)
 268         enum tbl_spant    pos;
 269         struct tbl_span  *next;
 270 };
 271
 272 struct  eqn {
 273         size_t            sz;
 274         char             *data;
 275         int               line; /* invocation line */
 276         int               pos; /* invocation position */
 277 };
 278
 279 /*
 280  * Available registers (set in libroff, accessed elsewhere).
 281  */
 282 enum    regs {
 283         REG_nS = 0,
 284         REG__MAX
 285 };
 286
 287 /*
 288  * A register (struct reg) can consist of many types: this consists of
 289  * normalised types from the original string form.
 290  */
 291 union   regval {
 292         unsigned  u; /* unsigned integer */
 293 };
 294
 295 /*
 296  * A single register entity.  If "set" is zero, the value of the
 297  * register should be the default one, which is per-register.  It's
 298  * assumed that callers know which type in "v" corresponds to which
 299  * register value.
 300  */
 301 struct  reg {
 302         int               set; /* whether set or not */
 303         union regval      v; /* parsed data */
 304 };
 305
 306 /*
 307  * The primary interface to setting register values is in libroff,
 308  * although libmdoc and libman from time to time will manipulate
 309  * registers (such as `.Sh SYNOPSIS' enabling REG_nS).
 310  */
 311 struct  regset {
 312         struct reg        regs[REG__MAX];
 313 };
 314
 315 /*
 316  * A punctuation delimiter, used only in mdoc(7) documents, is opening,
 317  * closing, or "middle mark" punctuation.  These govern spacing.
 318  * Opening punctuation (e.g., the opening parenthesis) suppresses the
 319  * following space; closing punctuation (e.g., the closing parenthesis)
 320  * suppresses the leading space; middle punctuation (e.g., the vertical
 321  * bar) can do either.  The middle punctuation delimiter bends the rules
 322  * depending on usage.
 323  */
 324 enum    mdelim {
 325         DELIM_NONE = 0,
 326         DELIM_OPEN,
 327         DELIM_MIDDLE,
 328         DELIM_CLOSE
 329 };
 330
 331 /*
 332  * The type of parse sequence.  This value is usually passed via the
 333  * mandoc(1) command line of -man and -mdoc.  It's almost exclusively
 334  * -mandoc but the others have been retained for compatibility.
 335  */
 336 enum    mparset {
 337         MPARSE_AUTO, /* magically determine the document type */
 338         MPARSE_MDOC, /* assume -mdoc */
 339         MPARSE_MAN /* assume -man */
 340 };
 341
 342 typedef void    (*mandocmsg)(enum mandocerr, void *,
 343                         int, int, const char *);
 344 typedef int     (*mevt_open)(void *, const char *);
 345 typedef void    (*mevt_close)(void *, const char *);
 346
 347 struct  mparse;
 348 struct  mdoc;
 349 struct  man;
 350
 351 __BEGIN_DECLS
 352
 353 void              mparse_free(struct mparse *);
 354 void              mparse_reset(struct mparse *);
 355 struct mparse    *mparse_alloc(enum mparset, mevt_open,
 356                         mevt_close, mandocmsg, void *);
 357 void              mparse_setstatus(struct mparse *, enum mandoclevel);
 358 enum mandoclevel  mparse_readfd(struct mparse *, int, const char *);
 359 void              mparse_result(struct mparse *, struct mdoc **, struct man **);
 360
 361 void             *mandoc_calloc(size_t, size_t);
 362 void             *mandoc_malloc(size_t);
 363 void             *mandoc_realloc(void *, size_t);
 364 #define DELIMSZ   6 /* hint: max possible size of a delimiter */
 365 enum mdelim       mandoc_isdelim(const char *);
 366
 367 __END_DECLS
 368
 369 #endif /*!MANDOC_H*/