]>
git.cameronkatri.com Git - mandoc.git/blob - mdoc.c
48756059b3ceaf5000aaa79c8bcc302234579abc
1 /* $Id: mdoc.c,v 1.14 2009/01/02 14:06:16 kristaps Exp $ */
3 * Copyright (c) 2008 Kristaps Dzonsons <kristaps@kth.se>
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
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.
29 const char *const __mdoc_macronames
[MDOC_MAX
] = {
30 "\\\"", "Dd", "Dt", "Os",
31 "Sh", "Ss", "Pp", "D1",
32 "Dl", "Bd", "Ed", "Bl",
33 "El", "It", "Ad", "An",
34 "Ar", "Cd", "Cm", "Dv",
35 "Er", "Ev", "Ex", "Fa",
36 "Fd", "Fl", "Fn", "Ft",
37 "Ic", "In", "Li", "Nd",
38 "Nm", "Op", "Ot", "Pa",
39 "Rv", "St", "Va", "Vt",
41 "Xr", "\%A", "\%B", "\%D",
43 "\%I", "\%J", "\%N", "\%O",
45 "\%P", "\%R", "\%T", "\%V",
46 "Ac", "Ao", "Aq", "At",
47 "Bc", "Bf", "Bo", "Bq",
48 "Bsx", "Bx", "Db", "Dc",
49 "Do", "Dq", "Ec", "Ef",
50 "Em", "Eo", "Fx", "Ms",
51 "No", "Ns", "Nx", "Ox",
52 "Pc", "Pf", "Po", "Pq",
53 "Qc", "Ql", "Qo", "Qq",
54 "Re", "Rs", "Sc", "So",
55 "Sq", "Sm", "Sx", "Sy",
56 "Tn", "Ux", "Xc", "Xo",
57 "Fo", "Fc", "Oo", "Oc",
58 "Bk", "Ek", "Bt", "Hf",
62 const char *const __mdoc_argnames
[MDOC_ARG_MAX
] = {
63 "split", "nosplit", "ragged",
64 "unfilled", "literal", "file",
65 "offset", "bullet", "dash",
66 "hyphen", "item", "enum",
67 "tag", "diag", "hang",
68 "ohang", "inset", "column",
69 "width", "compact", "std",
70 "p1003.1-88", "p1003.1-90", "p1003.1-96",
71 "p1003.1-2001", "p1003.1-2004", "p1003.1",
72 "p1003.1b", "p1003.1b-93", "p1003.1c-95",
73 "p1003.1g-2000", "p1003.2-92", "p1387.2-95",
74 "p1003.2", "p1387.2", "isoC-90",
75 "isoC-amd1", "isoC-tcor1", "isoC-tcor2",
76 "isoC-99", "ansiC", "ansiC-89",
77 "ansiC-99", "ieee754", "iso8802-3",
78 "xpg3", "xpg4", "xpg4.2",
79 "xpg4.3", "xbd5", "xcu5",
80 "xsh5", "xns5", "xns5.2d2.0",
81 "xcurses4.2", "susv2", "susv3",
82 "svid4", "filled", "words",
83 "emphasis", "symbolic",
86 const struct mdoc_macro __mdoc_macros
[MDOC_MAX
] = {
88 { NULL
, MDOC_PROLOGUE
}, /* Dd */
89 { NULL
, MDOC_PROLOGUE
}, /* Dt */
90 { NULL
, MDOC_PROLOGUE
}, /* Os */
91 { macro_scoped
, 0 }, /* Sh */
92 { macro_scoped
, 0 }, /* Ss */
93 { macro_text
, 0 }, /* Pp */
94 { macro_scoped_line
, 0 }, /* D1 */
95 { macro_scoped_line
, 0 }, /* Dl */
96 { macro_scoped
, MDOC_EXPLICIT
}, /* Bd */
97 { macro_close_explicit
, 0 }, /* Ed */
98 { macro_scoped
, MDOC_EXPLICIT
}, /* Bl */
99 { macro_close_explicit
, 0 }, /* El */
100 { NULL
, 0 }, /* It */
101 { macro_text
, MDOC_CALLABLE
}, /* Ad */
102 { macro_constant
, 0 }, /* An */
103 { macro_text
, MDOC_CALLABLE
}, /* Ar */
104 { macro_constant
, MDOC_QUOTABLE
}, /* Cd */
105 { macro_text
, MDOC_CALLABLE
}, /* Cm */
106 { macro_text
, MDOC_CALLABLE
}, /* Dv */
107 { macro_text
, MDOC_CALLABLE
}, /* Er */
108 { macro_text
, MDOC_CALLABLE
}, /* Ev */
109 { macro_constant
, 0 }, /* Ex */
110 { macro_text
, MDOC_CALLABLE
}, /* Fa */
111 { macro_constant
, 0 }, /* Fd */
112 { macro_text
, MDOC_CALLABLE
}, /* Fl */
113 { macro_text
, MDOC_CALLABLE
| MDOC_QUOTABLE
}, /* Fn */
114 { macro_text
, 0 }, /* Ft */
115 { macro_text
, MDOC_CALLABLE
}, /* Ic */
116 { macro_constant
, 0 }, /* In */
117 { macro_text
, MDOC_CALLABLE
}, /* Li */
118 { macro_constant
, 0 }, /* Nd */
119 { NULL
, 0 }, /* Nm */
120 { macro_scoped_line
, MDOC_CALLABLE
}, /* Op */
121 { macro_obsolete
, 0 }, /* Ot */
122 { macro_text
, MDOC_CALLABLE
}, /* Pa */
123 { macro_constant
, 0 }, /* Rv */
124 { macro_constant
, 0 }, /* St */
125 { macro_text
, MDOC_CALLABLE
}, /* Va */
126 { macro_text
, MDOC_CALLABLE
}, /* Vt */
127 { macro_text
, MDOC_CALLABLE
}, /* Xr */
128 { macro_constant
, MDOC_QUOTABLE
}, /* %A */
129 { macro_constant
, MDOC_QUOTABLE
}, /* %B */
130 { macro_constant
, MDOC_QUOTABLE
}, /* %D */
131 { macro_constant
, MDOC_QUOTABLE
}, /* %I */
132 { macro_constant
, MDOC_QUOTABLE
}, /* %J */
133 { macro_constant
, MDOC_QUOTABLE
}, /* %N */
134 { macro_constant
, MDOC_QUOTABLE
}, /* %O */
135 { macro_constant
, MDOC_QUOTABLE
}, /* %P */
136 { macro_constant
, MDOC_QUOTABLE
}, /* %R */
137 { macro_constant
, MDOC_QUOTABLE
}, /* %T */
138 { macro_constant
, MDOC_QUOTABLE
}, /* %V */
139 { NULL
, 0 }, /* Ac */
140 { NULL
, 0 }, /* Ao */
141 { macro_scoped_line
, MDOC_CALLABLE
}, /* Aq */
142 { macro_constant
, 0 }, /* At */
143 { NULL
, 0 }, /* Bc */
144 { macro_scoped
, 0 }, /* Bf */
145 { NULL
, 0 }, /* Bo */
146 { macro_scoped_line
, MDOC_CALLABLE
}, /* Bq */
147 { macro_constant_delimited
, 0 }, /* Bsx */
148 { macro_constant_delimited
, 0 }, /* Bx */
149 { macro_constant
, 0 }, /* Db */
150 { NULL
, 0 }, /* Dc */
151 { NULL
, 0 }, /* Do */
152 { macro_scoped_line
, MDOC_CALLABLE
}, /* Dq */
153 { NULL
, 0 }, /* Ec */
154 { macro_close_explicit
, 0 }, /* Ef */
155 { macro_text
, MDOC_CALLABLE
}, /* Em */
156 { NULL
, 0 }, /* Eo */
157 { macro_constant_delimited
, 0 }, /* Fx */
158 { macro_text
, 0 }, /* Ms */
159 { macro_constant_delimited
, MDOC_CALLABLE
}, /* No */
160 { macro_constant_delimited
, MDOC_CALLABLE
}, /* Ns */
161 { macro_constant_delimited
, 0 }, /* Nx */
162 { macro_constant_delimited
, 0 }, /* Ox */
163 { NULL
, 0 }, /* Pc */
164 { macro_constant
, 0 }, /* Pf */
165 { NULL
, 0 }, /* Po */
166 { macro_scoped_line
, MDOC_CALLABLE
}, /* Pq */
167 { NULL
, 0 }, /* Qc */
168 { macro_scoped_line
, MDOC_CALLABLE
}, /* Ql */
169 { NULL
, 0 }, /* Qo */
170 { macro_scoped_line
, MDOC_CALLABLE
}, /* Qq */
171 { macro_scoped
, MDOC_EXPLICIT
}, /* Re */
172 { macro_scoped
, MDOC_EXPLICIT
}, /* Rs */
173 { NULL
, 0 }, /* Sc */
174 { NULL
, 0 }, /* So */
175 { macro_scoped_line
, MDOC_CALLABLE
}, /* Sq */
176 { macro_constant
, 0 }, /* Sm */
177 { macro_text
, MDOC_CALLABLE
}, /* Sx */
178 { macro_text
, MDOC_CALLABLE
}, /* Sy */
179 { macro_text
, MDOC_CALLABLE
}, /* Tn */
180 { macro_constant_delimited
, 0 }, /* Ux */
181 { NULL
, 0 }, /* Xc */
182 { NULL
, 0 }, /* Xo */
183 { NULL
, 0 }, /* Fo */
184 { NULL
, 0 }, /* Fc */
185 { NULL
, 0 }, /* Oo */
186 { NULL
, 0 }, /* Oc */
187 { NULL
, 0 }, /* Bk */
188 { NULL
, 0 }, /* Ek */
189 { macro_constant
, 0 }, /* Bt */
190 { macro_constant
, 0 }, /* Hf */
191 { macro_obsolete
, 0 }, /* Fr */
192 { macro_constant
, 0 }, /* Ud */
195 const char * const *mdoc_macronames
= __mdoc_macronames
;
196 const char * const *mdoc_argnames
= __mdoc_argnames
;
197 const struct mdoc_macro
* const mdoc_macros
= __mdoc_macros
;
200 static struct mdoc_arg
*argdup(size_t, const struct mdoc_arg
*);
201 static void argfree(size_t, struct mdoc_arg
*);
202 static void argcpy(struct mdoc_arg
*,
203 const struct mdoc_arg
*);
204 static char **paramdup(size_t, const char **);
205 static void paramfree(size_t, char **);
207 static void mdoc_node_freelist(struct mdoc_node
*);
208 static void mdoc_node_append(struct mdoc
*, int,
210 static void mdoc_elem_free(struct mdoc_elem
*);
211 static void mdoc_text_free(struct mdoc_text
*);
214 const struct mdoc_node
*
215 mdoc_result(struct mdoc
*mdoc
)
223 mdoc_free(struct mdoc
*mdoc
)
227 mdoc_node_freelist(mdoc
->first
);
229 mdoc_tokhash_free(mdoc
->htab
);
236 mdoc_alloc(void *data
, const struct mdoc_cb
*cb
)
240 p
= xcalloc(1, sizeof(struct mdoc
));
243 (void)memcpy(&p
->cb
, cb
, sizeof(struct mdoc_cb
));
245 p
->htab
= mdoc_tokhash_alloc();
251 mdoc_parseln(struct mdoc
*mdoc
, char *buf
)
257 mdoc_word_alloc(mdoc
, 0, buf
);
258 mdoc
->next
= MDOC_NEXT_SIBLING
;
262 if (buf
[1] && '\\' == buf
[1])
263 if (buf
[2] && '\"' == buf
[2])
267 while (buf
[i
] && ! isspace(buf
[i
]) && i
< (int)sizeof(tmp
))
270 if (i
== (int)sizeof(tmp
))
271 return(mdoc_err(mdoc
, -1, 1, ERR_MACRO_NOTSUP
));
273 return(mdoc_err(mdoc
, -1, 1, ERR_MACRO_NOTSUP
));
277 (void)memcpy(tmp
, buf
+ 1, (size_t)i
);
280 if (MDOC_MAX
== (c
= mdoc_find(mdoc
, tmp
)))
281 return(mdoc_err(mdoc
, c
, 1, ERR_MACRO_NOTSUP
));
283 while (buf
[i
] && isspace(buf
[i
]))
286 return(mdoc_macro(mdoc
, c
, 1, &i
, buf
));
291 mdoc_msg(struct mdoc
*mdoc
, int pos
, const char *fmt
, ...)
296 if (NULL
== mdoc
->cb
.mdoc_msg
)
300 (void)vsnprintf(buf
, sizeof(buf
), fmt
, ap
);
303 (*mdoc
->cb
.mdoc_msg
)(mdoc
->data
, pos
, buf
);
308 mdoc_err(struct mdoc
*mdoc
, int tok
, int pos
, enum mdoc_err type
)
311 if (NULL
== mdoc
->cb
.mdoc_err
)
313 return((*mdoc
->cb
.mdoc_err
)(mdoc
->data
, tok
, pos
, type
));
318 mdoc_warn(struct mdoc
*mdoc
, int tok
, int pos
, enum mdoc_warn type
)
321 if (NULL
== mdoc
->cb
.mdoc_warn
)
323 return((*mdoc
->cb
.mdoc_warn
)(mdoc
->data
, tok
, pos
, type
));
328 mdoc_macro(struct mdoc
*mdoc
, int tok
, int ppos
, int *pos
, char *buf
)
331 if (MDOC_PROLOGUE
& mdoc_macros
[tok
].flags
&&
332 SEC_PROLOGUE
== mdoc
->sec_lastn
)
333 return(macro_prologue(mdoc
, tok
, ppos
, pos
, buf
));
335 if (SEC_PROLOGUE
== mdoc
->sec_lastn
)
336 return(mdoc_err(mdoc
, tok
, ppos
, ERR_SEC_PROLOGUE
));
338 if (NULL
== (mdoc_macros
[tok
].fp
)) {
339 (void)mdoc_err(mdoc
, tok
, ppos
, ERR_MACRO_NOTSUP
);
343 if (1 != ppos
&& ! (MDOC_CALLABLE
& mdoc_macros
[tok
].flags
)) {
344 (void)mdoc_err(mdoc
, tok
, ppos
, ERR_MACRO_NOTCALL
);
348 return((*mdoc_macros
[tok
].fp
)(mdoc
, tok
, ppos
, pos
, buf
));
353 mdoc_node_append(struct mdoc
*mdoc
, int pos
, struct mdoc_node
*p
)
355 const char *nn
, *on
, *nt
, *ot
, *act
;
359 nn
= p
->data
.text
.string
;
363 nn
= mdoc_macronames
[p
->data
.body
.tok
];
367 nn
= mdoc_macronames
[p
->data
.elem
.tok
];
371 nn
= mdoc_macronames
[p
->data
.head
.tok
];
375 nn
= mdoc_macronames
[p
->data
.block
.tok
];
383 if (NULL
== mdoc
->first
) {
384 assert(NULL
== mdoc
->last
);
387 mdoc_msg(mdoc
, pos
, "parse: root %s `%s'", nt
, nn
);
391 switch (mdoc
->last
->type
) {
397 on
= mdoc_macronames
[mdoc
->last
->data
.body
.tok
];
401 on
= mdoc_macronames
[mdoc
->last
->data
.elem
.tok
];
405 on
= mdoc_macronames
[mdoc
->last
->data
.head
.tok
];
409 on
= mdoc_macronames
[mdoc
->last
->data
.block
.tok
];
417 switch (mdoc
->next
) {
418 case (MDOC_NEXT_SIBLING
):
419 mdoc
->last
->next
= p
;
420 p
->prev
= mdoc
->last
;
421 p
->parent
= mdoc
->last
->parent
;
424 case (MDOC_NEXT_CHILD
):
425 mdoc
->last
->child
= p
;
426 p
->parent
= mdoc
->last
;
434 mdoc_msg(mdoc
, pos
, "parse: %s `%s' %s of %s `%s'",
435 nt
, nn
, act
, ot
, on
);
442 mdoc_head_alloc(struct mdoc
*mdoc
, int pos
, int tok
)
449 p
= xcalloc(1, sizeof(struct mdoc_node
));
452 p
->data
.head
.tok
= tok
;
454 mdoc_node_append(mdoc
, pos
, p
);
459 mdoc_body_alloc(struct mdoc
*mdoc
, int pos
, int tok
)
466 p
= xcalloc(1, sizeof(struct mdoc_node
));
469 p
->data
.body
.tok
= tok
;
471 mdoc_node_append(mdoc
, pos
, p
);
476 mdoc_block_alloc(struct mdoc
*mdoc
, int pos
, int tok
,
477 size_t argsz
, const struct mdoc_arg
*args
)
481 p
= xcalloc(1, sizeof(struct mdoc_node
));
483 p
->type
= MDOC_BLOCK
;
484 p
->data
.block
.tok
= tok
;
485 p
->data
.block
.argc
= argsz
;
486 p
->data
.block
.argv
= argdup(argsz
, args
);
488 mdoc_node_append(mdoc
, pos
, p
);
493 mdoc_elem_alloc(struct mdoc
*mdoc
, int pos
, int tok
,
494 size_t argsz
, const struct mdoc_arg
*args
,
495 size_t paramsz
, const char **params
)
499 p
= xcalloc(1, sizeof(struct mdoc_node
));
501 p
->data
.elem
.tok
= tok
;
502 p
->data
.elem
.sz
= paramsz
;
503 p
->data
.elem
.args
= paramdup(paramsz
, params
);
504 p
->data
.elem
.argc
= argsz
;
505 p
->data
.elem
.argv
= argdup(argsz
, args
);
507 mdoc_node_append(mdoc
, pos
, p
);
512 mdoc_word_alloc(struct mdoc
*mdoc
, int pos
, const char *word
)
516 p
= xcalloc(1, sizeof(struct mdoc_node
));
518 p
->data
.text
.string
= xstrdup(word
);
520 mdoc_node_append(mdoc
, pos
, p
);
525 argfree(size_t sz
, struct mdoc_arg
*p
)
534 for (i
= 0; i
< (int)sz
; i
++)
538 for (j
= 0; j
< (int)p
[i
].sz
; j
++)
546 mdoc_elem_free(struct mdoc_elem
*p
)
549 paramfree(p
->sz
, p
->args
);
550 argfree(p
->argc
, p
->argv
);
555 mdoc_block_free(struct mdoc_block
*p
)
558 argfree(p
->argc
, p
->argv
);
563 mdoc_text_free(struct mdoc_text
*p
)
572 mdoc_node_free(struct mdoc_node
*p
)
577 mdoc_text_free(&p
->data
.text
);
580 mdoc_elem_free(&p
->data
.elem
);
583 mdoc_block_free(&p
->data
.block
);
594 mdoc_node_freelist(struct mdoc_node
*p
)
598 mdoc_node_freelist(p
->child
);
600 mdoc_node_freelist(p
->next
);
607 mdoc_find(const struct mdoc
*mdoc
, const char *key
)
610 return(mdoc_tokhash_find(mdoc
->htab
, key
));
615 argcpy(struct mdoc_arg
*dst
, const struct mdoc_arg
*src
)
620 if (0 == (dst
->sz
= src
->sz
))
622 dst
->value
= xcalloc(dst
->sz
, sizeof(char *));
623 for (i
= 0; i
< (int)dst
->sz
; i
++)
624 dst
->value
[i
] = xstrdup(src
->value
[i
]);
628 static struct mdoc_arg
*
629 argdup(size_t argsz
, const struct mdoc_arg
*args
)
637 pp
= xcalloc((size_t)argsz
, sizeof(struct mdoc_arg
));
638 for (i
= 0; i
< (int)argsz
; i
++)
639 argcpy(&pp
[i
], &args
[i
]);
646 paramfree(size_t sz
, char **p
)
655 for (i
= 0; i
< (int)sz
; i
++)
662 paramdup(size_t sz
, const char **p
)
670 pp
= xcalloc(sz
, sizeof(char *));
671 for (i
= 0; i
< (int)sz
; i
++)
672 pp
[i
] = xstrdup(p
[i
]);