contrib/mdocml/mandoc.h

   1 /*      $Id: mandoc.h,v 1.77 2011/05/24 21:31:23 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         MANDOCERR_LINESCOPE, /* line scope broken */
  78
  79         /* related to missing macro arguments */
  80         MANDOCERR_MACROEMPTY, /* skipping empty macro */
  81         MANDOCERR_ARGCWARN, /* argument count wrong */
  82         MANDOCERR_DISPTYPE, /* missing display type */
  83         MANDOCERR_LISTFIRST, /* list type must come first */
  84         MANDOCERR_NOWIDTHARG, /* tag lists require a width argument */
  85         MANDOCERR_FONTTYPE, /* missing font type */
  86         MANDOCERR_WNOSCOPE, /* skipping end of block that is not open */
  87
  88         /* related to bad macro arguments */
  89         MANDOCERR_IGNARGV, /* skipping argument */
  90         MANDOCERR_ARGVREP, /* duplicate argument */
  91         MANDOCERR_DISPREP, /* duplicate display type */
  92         MANDOCERR_LISTREP, /* duplicate list type */
  93         MANDOCERR_BADATT, /* unknown AT&T UNIX version */
  94         MANDOCERR_BADBOOL, /* bad Boolean value */
  95         MANDOCERR_BADFONT, /* unknown font */
  96         MANDOCERR_BADSTANDARD, /* unknown standard specifier */
  97         MANDOCERR_BADWIDTH, /* bad width argument */
  98
  99         /* related to plain text */
 100         MANDOCERR_NOBLANKLN, /* blank line in non-literal context */
 101         MANDOCERR_BADTAB, /* tab in non-literal context */
 102         MANDOCERR_EOLNSPACE, /* end of line whitespace */
 103         MANDOCERR_BADCOMMENT, /* bad comment style */
 104         MANDOCERR_BADESCAPE, /* unknown escape sequence */
 105         MANDOCERR_BADQUOTE, /* unterminated quoted string */
 106
 107         MANDOCERR_ERROR, /* ===== start of errors ===== */
 108
 109         /* related to tables */
 110         MANDOCERR_TBL, /* bad table syntax */
 111         MANDOCERR_TBLOPT, /* bad table option */
 112         MANDOCERR_TBLLAYOUT, /* bad table layout */
 113         MANDOCERR_TBLNOLAYOUT, /* no table layout cells specified */
 114         MANDOCERR_TBLNODATA, /* no table data cells specified */
 115         MANDOCERR_TBLIGNDATA, /* ignore data in cell */
 116         MANDOCERR_TBLBLOCK, /* data block still open */
 117         MANDOCERR_TBLEXTRADAT, /* ignoring extra data cells */
 118
 119         MANDOCERR_ROFFLOOP, /* input stack limit exceeded, infinite loop? */
 120         MANDOCERR_BADCHAR, /* skipping bad character */
 121         MANDOCERR_NAMESC, /* escaped character not allowed in a name */
 122         MANDOCERR_NOTEXT, /* skipping text before the first section header */
 123         MANDOCERR_MACRO, /* skipping unknown macro */
 124         MANDOCERR_REQUEST, /* NOT IMPLEMENTED: skipping request */
 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_NOTMANUAL, /* manual isn't really a manual */
 141         MANDOCERR_COLUMNS, /* column syntax is inconsistent */
 142         MANDOCERR_BADDISP, /* NOT IMPLEMENTED: .Bd -file */
 143         MANDOCERR_SYNTLINESCOPE, /* line scope broken, syntax violated */
 144         MANDOCERR_SYNTARGVCOUNT, /* argument count wrong, violates syntax */
 145         MANDOCERR_SYNTCHILD, /* child violates parent syntax */
 146         MANDOCERR_SYNTARGCOUNT, /* argument count wrong, violates syntax */
 147         MANDOCERR_SOPATH, /* NOT IMPLEMENTED: .so with absolute path or ".." */
 148         MANDOCERR_NODOCBODY, /* no document body */
 149         MANDOCERR_NODOCPROLOG, /* no document prologue */
 150         MANDOCERR_MEM, /* static buffer exhausted */
 151         MANDOCERR_MAX
 152 };
 153
 154 struct  tbl {
 155         char              tab; /* cell-separator */
 156         char              decimal; /* decimal point */
 157         int               linesize;
 158         int               opts;
 159 #define TBL_OPT_CENTRE   (1 << 0)
 160 #define TBL_OPT_EXPAND   (1 << 1)
 161 #define TBL_OPT_BOX      (1 << 2)
 162 #define TBL_OPT_DBOX     (1 << 3)
 163 #define TBL_OPT_ALLBOX   (1 << 4)
 164 #define TBL_OPT_NOKEEP   (1 << 5)
 165 #define TBL_OPT_NOSPACE  (1 << 6)
 166         int               cols; /* number of columns */
 167 };
 168
 169 enum    tbl_headt {
 170         TBL_HEAD_DATA, /* plug in data from tbl_dat */
 171         TBL_HEAD_VERT, /* vertical spacer */
 172         TBL_HEAD_DVERT  /* double-vertical spacer */
 173 };
 174
 175 /*
 176  * The head of a table specifies all of its columns.  When formatting a
 177  * tbl_span, iterate over these and plug in data from the tbl_span when
 178  * appropriate, using tbl_cell as a guide to placement.
 179  */
 180 struct  tbl_head {
 181         enum tbl_headt    pos;
 182         int               ident; /* 0 <= unique id < cols */
 183         struct tbl_head  *next;
 184         struct tbl_head  *prev;
 185 };
 186
 187 enum    tbl_cellt {
 188         TBL_CELL_CENTRE, /* c, C */
 189         TBL_CELL_RIGHT, /* r, R */
 190         TBL_CELL_LEFT, /* l, L */
 191         TBL_CELL_NUMBER, /* n, N */
 192         TBL_CELL_SPAN, /* s, S */
 193         TBL_CELL_LONG, /* a, A */
 194         TBL_CELL_DOWN, /* ^ */
 195         TBL_CELL_HORIZ, /* _, - */
 196         TBL_CELL_DHORIZ, /* = */
 197         TBL_CELL_VERT, /* | */
 198         TBL_CELL_DVERT, /* || */
 199         TBL_CELL_MAX
 200 };
 201
 202 /*
 203  * A cell in a layout row.
 204  */
 205 struct  tbl_cell {
 206         struct tbl_cell  *next;
 207         enum tbl_cellt    pos;
 208         size_t            spacing;
 209         int               flags;
 210 #define TBL_CELL_TALIGN  (1 << 0) /* t, T */
 211 #define TBL_CELL_BALIGN  (1 << 1) /* d, D */
 212 #define TBL_CELL_BOLD    (1 << 2) /* fB, B, b */
 213 #define TBL_CELL_ITALIC  (1 << 3) /* fI, I, i */
 214 #define TBL_CELL_EQUAL   (1 << 4) /* e, E */
 215 #define TBL_CELL_UP      (1 << 5) /* u, U */
 216 #define TBL_CELL_WIGN    (1 << 6) /* z, Z */
 217         struct tbl_head  *head;
 218 };
 219
 220 /*
 221  * A layout row.
 222  */
 223 struct  tbl_row {
 224         struct tbl_row   *next;
 225         struct tbl_cell  *first;
 226         struct tbl_cell  *last;
 227 };
 228
 229 enum    tbl_datt {
 230         TBL_DATA_NONE, /* has no data */
 231         TBL_DATA_DATA, /* consists of data/string */
 232         TBL_DATA_HORIZ, /* horizontal line */
 233         TBL_DATA_DHORIZ, /* double-horizontal line */
 234         TBL_DATA_NHORIZ, /* squeezed horizontal line */
 235         TBL_DATA_NDHORIZ /* squeezed double-horizontal line */
 236 };
 237
 238 /*
 239  * A cell within a row of data.  The "string" field contains the actual
 240  * string value that's in the cell.  The rest is layout.
 241  */
 242 struct  tbl_dat {
 243         struct tbl_cell  *layout; /* layout cell */
 244         int               spans; /* how many spans follow */
 245         struct tbl_dat   *next;
 246         char             *string; /* data (NULL if not TBL_DATA_DATA) */
 247         enum tbl_datt     pos;
 248 };
 249
 250 enum    tbl_spant {
 251         TBL_SPAN_DATA, /* span consists of data */
 252         TBL_SPAN_HORIZ, /* span is horizontal line */
 253         TBL_SPAN_DHORIZ /* span is double horizontal line */
 254 };
 255
 256 /*
 257  * A row of data in a table.
 258  */
 259 struct  tbl_span {
 260         struct tbl       *tbl;
 261         struct tbl_head  *head;
 262         struct tbl_row   *layout; /* layout row */
 263         struct tbl_dat   *first;
 264         struct tbl_dat   *last;
 265         int               line; /* parse line */
 266         int               flags;
 267 #define TBL_SPAN_FIRST   (1 << 0)
 268 #define TBL_SPAN_LAST    (1 << 1)
 269         enum tbl_spant    pos;
 270         struct tbl_span  *next;
 271 };
 272
 273 struct  eqn {
 274         size_t            sz;
 275         char             *data;
 276         int               line; /* invocation line */
 277         int               pos; /* invocation position */
 278 };
 279
 280 /*
 281  * The type of parse sequence.  This value is usually passed via the
 282  * mandoc(1) command line of -man and -mdoc.  It's almost exclusively
 283  * -mandoc but the others have been retained for compatibility.
 284  */
 285 enum    mparset {
 286         MPARSE_AUTO, /* magically determine the document type */
 287         MPARSE_MDOC, /* assume -mdoc */
 288         MPARSE_MAN /* assume -man */
 289 };
 290
 291 enum    mandoc_esc {
 292         ESCAPE_ERROR = 0, /* bail! unparsable escape */
 293         ESCAPE_IGNORE, /* escape to be ignored */
 294         ESCAPE_SPECIAL, /* a regular special character */
 295         ESCAPE_FONT, /* a generic font mode */
 296         ESCAPE_FONTBOLD, /* bold font mode */
 297         ESCAPE_FONTITALIC, /* italic font mode */
 298         ESCAPE_FONTROMAN, /* roman font mode */
 299         ESCAPE_FONTPREV, /* previous font mode */
 300         ESCAPE_NUMBERED, /* a numbered glyph */
 301         ESCAPE_UNICODE, /* a unicode codepoint */
 302         ESCAPE_NOSPACE /* suppress space if the last on a line */
 303 };
 304
 305 typedef void    (*mandocmsg)(enum mandocerr, enum mandoclevel,
 306                         const char *, int, int, const char *);
 307
 308 struct  mparse;
 309 struct  mchars;
 310 struct  mdoc;
 311 struct  man;
 312
 313 __BEGIN_DECLS
 314
 315 void              mparse_free(struct mparse *);
 316 void              mparse_reset(struct mparse *);
 317 struct mparse    *mparse_alloc(enum mparset,
 318                         enum mandoclevel, mandocmsg, void *);
 319 enum mandoclevel  mparse_readfd(struct mparse *, int, const char *);
 320 void              mparse_result(struct mparse *, struct mdoc **, struct man **);
 321 const char       *mparse_strerror(enum mandocerr);
 322 const char       *mparse_strlevel(enum mandoclevel);
 323
 324 void             *mandoc_calloc(size_t, size_t);
 325 void             *mandoc_malloc(size_t);
 326 void             *mandoc_realloc(void *, size_t);
 327
 328 enum mandoc_esc   mandoc_escape(const char **, const char **, int *);
 329
 330 struct mchars    *mchars_alloc(void);
 331 char              mchars_num2char(const char *, size_t);
 332 int               mchars_num2uc(const char *, size_t);
 333 const char       *mchars_spec2str(struct mchars *, const char *, size_t, size_t *);
 334 int               mchars_spec2cp(struct mchars *, const char *, size_t);
 335 void              mchars_free(struct mchars *);
 336
 337
 338 __END_DECLS
 339
 340 #endif /*!MANDOC_H*/