mdoc.h

   1 /* $Id: mdoc.h,v 1.43 2009/03/09 13:04:01 kristaps Exp $ */
   2 /*
   3  * Copyright (c) 2008 Kristaps Dzonsons <kristaps@kth.se>
   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
   7  * above copyright notice and this permission notice appear in all
   8  * copies.
   9  *
  10  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL
  11  * WARRANTIES WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED
  12  * WARRANTIES OF MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE
  13  * AUTHOR BE LIABLE FOR ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL
  14  * DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR
  15  * PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER
  16  * TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR
  17  * PERFORMANCE OF THIS SOFTWARE.
  18  */
  19 #ifndef MDOC_H
  20 #define MDOC_H
  21
  22 /*
  23  * This library implements a validating scanner/parser for ``mdoc'' roff
  24  * macro documents, a.k.a. BSD manual page documents.  The mdoc.c file
  25  * drives the parser, while macro.c describes the macro ontologies.
  26  * validate.c pre- and post-validates parsed macros, and action.c
  27  * performs actions on parsed and validated macros.
  28  */
  29
  30 /* What follows is a list of ALL possible macros. */
  31
  32 /* TODO: Brq et al. */
  33
  34 #define MDOC___          0
  35 #define MDOC_Dd          1
  36 #define MDOC_Dt          2
  37 #define MDOC_Os          3
  38 #define MDOC_Sh          4
  39 #define MDOC_Ss          5
  40 #define MDOC_Pp          6
  41 #define MDOC_D1          7
  42 #define MDOC_Dl          8
  43 #define MDOC_Bd          9
  44 #define MDOC_Ed          10
  45 #define MDOC_Bl          11
  46 #define MDOC_El          12
  47 #define MDOC_It          13
  48 #define MDOC_Ad          14
  49 #define MDOC_An          15
  50 #define MDOC_Ar          16
  51 #define MDOC_Cd          17
  52 #define MDOC_Cm          18
  53 #define MDOC_Dv          19
  54 #define MDOC_Er          20
  55 #define MDOC_Ev          21
  56 #define MDOC_Ex          22
  57 #define MDOC_Fa          23
  58 #define MDOC_Fd          24
  59 #define MDOC_Fl          25
  60 #define MDOC_Fn          26
  61 #define MDOC_Ft          27
  62 #define MDOC_Ic          28
  63 #define MDOC_In          29
  64 #define MDOC_Li          30
  65 #define MDOC_Nd          31
  66 #define MDOC_Nm          32
  67 #define MDOC_Op          33
  68 #define MDOC_Ot          34
  69 #define MDOC_Pa          35
  70 #define MDOC_Rv          36
  71 #define MDOC_St          37
  72 #define MDOC_Va          38
  73 #define MDOC_Vt          39
  74 #define MDOC_Xr          40
  75 #define MDOC__A          41
  76 #define MDOC__B          42
  77 #define MDOC__D          43
  78 #define MDOC__I          44
  79 #define MDOC__J          45
  80 #define MDOC__N          46
  81 #define MDOC__O          47
  82 #define MDOC__P          48
  83 #define MDOC__R          49
  84 #define MDOC__T          50
  85 #define MDOC__V          51
  86 #define MDOC_Ac          52
  87 #define MDOC_Ao          53
  88 #define MDOC_Aq          54
  89 #define MDOC_At          55
  90 #define MDOC_Bc          56
  91 #define MDOC_Bf          57
  92 #define MDOC_Bo          58
  93 #define MDOC_Bq          59
  94 #define MDOC_Bsx         60
  95 #define MDOC_Bx          61
  96 #define MDOC_Db          62
  97 #define MDOC_Dc          63
  98 #define MDOC_Do          64
  99 #define MDOC_Dq          65
 100 #define MDOC_Ec          66
 101 #define MDOC_Ef          67
 102 #define MDOC_Em          68
 103 #define MDOC_Eo          69
 104 #define MDOC_Fx          70
 105 #define MDOC_Ms          71
 106 #define MDOC_No          72
 107 #define MDOC_Ns          73
 108 #define MDOC_Nx          74
 109 #define MDOC_Ox          75
 110 #define MDOC_Pc          76
 111 #define MDOC_Pf          77
 112 #define MDOC_Po          78
 113 #define MDOC_Pq          79
 114 #define MDOC_Qc          80
 115 #define MDOC_Ql          81
 116 #define MDOC_Qo          82
 117 #define MDOC_Qq          83
 118 #define MDOC_Re          84
 119 #define MDOC_Rs          85
 120 #define MDOC_Sc          86
 121 #define MDOC_So          87
 122 #define MDOC_Sq          88
 123 #define MDOC_Sm          89
 124 #define MDOC_Sx          90
 125 #define MDOC_Sy          91
 126 #define MDOC_Tn          92
 127 #define MDOC_Ux          93
 128 #define MDOC_Xc          94
 129 #define MDOC_Xo          95
 130 #define MDOC_Fo          96
 131 #define MDOC_Fc          97
 132 #define MDOC_Oo          98
 133 #define MDOC_Oc          99
 134 #define MDOC_Bk          100
 135 #define MDOC_Ek          101
 136 #define MDOC_Bt          102
 137 #define MDOC_Hf          103
 138 #define MDOC_Fr          104
 139 #define MDOC_Ud          105
 140 #define MDOC_Lb          106
 141 #define MDOC_Ap          107
 142 #define MDOC_Lp          108
 143 #define MDOC_MAX         109
 144
 145 /* What follows is a list of ALL possible macro arguments. */
 146
 147 #define MDOC_Split       0
 148 #define MDOC_Nosplit     1
 149 #define MDOC_Ragged      2
 150 #define MDOC_Unfilled    3
 151 #define MDOC_Literal     4
 152 #define MDOC_File        5
 153 #define MDOC_Offset      6
 154 #define MDOC_Bullet      7
 155 #define MDOC_Dash        8
 156 #define MDOC_Hyphen      9
 157 #define MDOC_Item        10
 158 #define MDOC_Enum        11
 159 #define MDOC_Tag         12
 160 #define MDOC_Diag        13
 161 #define MDOC_Hang        14
 162 #define MDOC_Ohang       15
 163 #define MDOC_Inset       16
 164 #define MDOC_Column      17
 165 #define MDOC_Width       18
 166 #define MDOC_Compact     19
 167 #define MDOC_Std         20
 168 #define MDOC_Filled      21
 169 #define MDOC_Words       22
 170 #define MDOC_Emphasis    23
 171 #define MDOC_Symbolic    24
 172 #define MDOC_ARG_MAX     25
 173
 174 /* Warnings are either syntax or groff-compatibility. */
 175 enum    mdoc_warn {
 176         WARN_SYNTAX,
 177         WARN_COMPAT
 178 };
 179
 180 /* Type of a syntax node. */
 181 enum    mdoc_type {
 182         MDOC_TEXT,
 183         MDOC_ELEM,
 184         MDOC_HEAD,
 185         MDOC_TAIL,
 186         MDOC_BODY,
 187         MDOC_BLOCK,
 188         MDOC_ROOT
 189 };
 190
 191 /* Section (named/unnamed) of `Sh'. */
 192 enum    mdoc_sec {
 193         SEC_PROLOGUE            = 0,
 194         SEC_BODY                = 1,
 195         SEC_NAME                = 2,
 196         SEC_LIBRARY             = 3,
 197         SEC_SYNOPSIS            = 4,
 198         SEC_DESCRIPTION         = 5,
 199         SEC_IMPLEMENTATION      = 6,
 200         SEC_RETURN_VALUES       = 7,
 201         SEC_ENVIRONMENT         = 8,
 202         SEC_FILES               = 9,
 203         SEC_EXAMPLES            = 10,
 204         SEC_DIAGNOSTICS         = 11,
 205         SEC_COMPATIBILITY       = 12,
 206         SEC_ERRORS              = 13,
 207         SEC_SEE_ALSO            = 14,
 208         SEC_STANDARDS           = 15,
 209         SEC_HISTORY             = 16,
 210         SEC_AUTHORS             = 17,
 211         SEC_CAVEATS             = 18,
 212         SEC_BUGS                = 19,
 213         SEC_CUSTOM
 214 };
 215
 216 /* Information from prologue. */
 217 struct  mdoc_meta {
 218         int               msec;
 219         char             *vol;
 220         char             *arch;
 221         time_t            date;
 222         char             *title;
 223         char             *os;
 224         char             *name;
 225 };
 226
 227 /* An argument to a macro (multiple values = `It -column'). */
 228 struct  mdoc_argv {
 229         int               arg;
 230         int               line;
 231         int               pos;
 232         size_t            sz;
 233         char            **value;
 234 };
 235
 236 struct  mdoc_arg {
 237         size_t            argc;
 238         struct mdoc_argv *argv;
 239         unsigned int      refcnt;
 240 };
 241
 242 /* Node in AST. */
 243 struct  mdoc_node {
 244         struct mdoc_node *parent;
 245         struct mdoc_node *child;
 246         struct mdoc_node *next;
 247         struct mdoc_node *prev;
 248         int               line;
 249         int               pos;
 250         int               tok;
 251         int               flags;
 252 #define MDOC_VALID       (1 << 0)
 253 #define MDOC_ACTED       (1 << 1)
 254         enum mdoc_type    type;
 255         enum mdoc_sec     sec;
 256
 257         /* FIXME: union/struct this with #defines. */
 258         struct mdoc_arg  *args;         /* BLOCK/ELEM */
 259         struct mdoc_node *head;         /* BLOCK */
 260         struct mdoc_node *body;         /* BLOCK */
 261         struct mdoc_node *tail;         /* BLOCK */
 262         char             *string;       /* TEXT */
 263 };
 264
 265 #define MDOC_IGN_SCOPE   (1 << 0) /* Ignore scope violations. */
 266 #define MDOC_IGN_ESCAPE  (1 << 1) /* Ignore bad escape sequences. */
 267 #define MDOC_IGN_MACRO   (1 << 2) /* Ignore unknown macros. */
 268
 269 /* Call-backs for parse messages. */
 270 struct  mdoc_cb {
 271         void    (*mdoc_msg)(void *, int, int, const char *);
 272         int     (*mdoc_err)(void *, int, int, const char *);
 273         int     (*mdoc_warn)(void *, int, int,
 274                         enum mdoc_warn, const char *);
 275 };
 276
 277 /* Global table of macro names (`Bd', `Ed', etc.). */
 278 extern  const char *const *mdoc_macronames;
 279
 280 /* Global table of argument names (`column', `tag', etc.). */
 281 extern  const char *const *mdoc_argnames;
 282
 283 __BEGIN_DECLS
 284
 285 struct  mdoc;
 286
 287 /* Free memory allocated with mdoc_alloc. */
 288 void              mdoc_free(struct mdoc *);
 289
 290 /* Allocate a new parser instance. */
 291 struct  mdoc     *mdoc_alloc(void *, int, const struct mdoc_cb *);
 292
 293 /* Set parse options. */
 294 void              mdoc_setflags(struct mdoc *, int);
 295
 296 /* Parse a single line in a stream (boolean retval). */
 297 int               mdoc_parseln(struct mdoc *, int, char *buf);
 298
 299 /* Get result first node (after mdoc_endparse!). */
 300 const struct mdoc_node *mdoc_node(const struct mdoc *);
 301
 302 /* Get result meta-information (after mdoc_endparse!). */
 303 const struct mdoc_meta *mdoc_meta(const struct mdoc *);
 304
 305 /* Signal end of parse sequence (boolean retval). */
 306 int               mdoc_endparse(struct mdoc *);
 307
 308 /* The following are utility functions. */
 309
 310 const char       *mdoc_a2att(const char *);
 311 const char       *mdoc_a2lib(const char *);
 312 const char       *mdoc_a2st(const char *);
 313
 314 int               mdoc_isdelim(const char *);
 315
 316 __END_DECLS
 317
 318 #endif /*!MDOC_H*/