]>
git.cameronkatri.com Git - mandoc.git/blob - argv.c
1 /* $Id: argv.c,v 1.52 2009/03/16 22:19:19 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.
19 #include <sys/types.h>
31 * Routines to parse arguments of macros. Arguments follow the syntax
32 * of `-arg [val [valN...]]'. Arguments come in all types: quoted
33 * arguments, multiple arguments per value, no-value arguments, etc.
35 * There's no limit to the number or arguments that may be allocated.
38 #define ARGS_QUOTED (1 << 0)
39 #define ARGS_DELIM (1 << 1)
40 #define ARGS_TABSEP (1 << 2)
41 #define ARGS_ARGVLIKE (1 << 3)
43 #define ARGV_NONE (1 << 0)
44 #define ARGV_SINGLE (1 << 1)
45 #define ARGV_MULTI (1 << 2)
46 #define ARGV_OPT_SINGLE (1 << 3)
62 static int argv_a2arg(int, const char *);
63 static int args(struct mdoc
*, int, int *,
64 char *, int, char **);
65 static int argv(struct mdoc
*, int,
66 struct mdoc_argv
*, int *, char *);
67 static int argv_single(struct mdoc
*, int,
68 struct mdoc_argv
*, int *, char *);
69 static int argv_opt_single(struct mdoc
*, int,
70 struct mdoc_argv
*, int *, char *);
71 static int argv_multi(struct mdoc
*, int,
72 struct mdoc_argv
*, int *, char *);
73 static int pwarn(struct mdoc
*, int, int, enum mwarn
);
74 static int perr(struct mdoc
*, int, int, enum merr
);
76 /* Per-argument flags. */
78 static int mdoc_argvflags
[MDOC_ARG_MAX
] = {
79 ARGV_NONE
, /* MDOC_Split */
80 ARGV_NONE
, /* MDOC_Nosplit */
81 ARGV_NONE
, /* MDOC_Ragged */
82 ARGV_NONE
, /* MDOC_Unfilled */
83 ARGV_NONE
, /* MDOC_Literal */
84 ARGV_NONE
, /* MDOC_File */
85 ARGV_SINGLE
, /* MDOC_Offset */
86 ARGV_NONE
, /* MDOC_Bullet */
87 ARGV_NONE
, /* MDOC_Dash */
88 ARGV_NONE
, /* MDOC_Hyphen */
89 ARGV_NONE
, /* MDOC_Item */
90 ARGV_NONE
, /* MDOC_Enum */
91 ARGV_NONE
, /* MDOC_Tag */
92 ARGV_NONE
, /* MDOC_Diag */
93 ARGV_NONE
, /* MDOC_Hang */
94 ARGV_NONE
, /* MDOC_Ohang */
95 ARGV_NONE
, /* MDOC_Inset */
96 ARGV_MULTI
, /* MDOC_Column */
97 ARGV_SINGLE
, /* MDOC_Width */
98 ARGV_NONE
, /* MDOC_Compact */
99 ARGV_OPT_SINGLE
, /* MDOC_Std */
100 ARGV_NONE
, /* MDOC_Filled */
101 ARGV_NONE
, /* MDOC_Words */
102 ARGV_NONE
, /* MDOC_Emphasis */
103 ARGV_NONE
, /* MDOC_Symbolic */
104 ARGV_NONE
/* MDOC_Symbolic */
107 static int mdoc_argflags
[MDOC_MAX
] = {
125 ARGS_QUOTED
, /* Cd */
131 ARGS_DELIM
| ARGS_QUOTED
, /* Fa */
134 ARGS_DELIM
| ARGS_QUOTED
, /* Fn */
135 ARGS_DELIM
| ARGS_QUOTED
, /* Ft */
145 ARGS_DELIM
| ARGS_ARGVLIKE
, /* St */
149 ARGS_QUOTED
, /* %A */
150 ARGS_QUOTED
, /* %B */
151 ARGS_QUOTED
, /* %D */
152 ARGS_QUOTED
, /* %I */
153 ARGS_QUOTED
, /* %J */
154 ARGS_QUOTED
, /* %N */
155 ARGS_QUOTED
, /* %O */
156 ARGS_QUOTED
, /* %P */
157 ARGS_QUOTED
, /* %R */
158 ARGS_QUOTED
, /* %T */
159 ARGS_QUOTED
, /* %V */
168 ARGS_DELIM
, /* Bsx */
217 ARGS_DELIM
| ARGS_QUOTED
, /* Lk */
218 ARGS_DELIM
| ARGS_QUOTED
, /* Mt */
219 ARGS_DELIM
, /* Brq */
221 ARGS_DELIM
, /* Brc */
222 ARGS_QUOTED
, /* %C */
229 * Parse an argument from line text. This comes in the form of -key
230 * [value0...], which may either have a single mandatory value, at least
231 * one mandatory value, an optional single value, or no value.
234 mdoc_argv(struct mdoc
*mdoc
, int line
, int tok
,
235 struct mdoc_arg
**v
, int *pos
, char *buf
)
239 struct mdoc_argv tmp
;
240 struct mdoc_arg
*arg
;
245 assert(' ' != buf
[*pos
]);
247 if ('-' != buf
[*pos
] || ARGS_ARGVLIKE
& mdoc_argflags
[tok
])
250 /* Parse through to the first unescaped space. */
259 if (' ' == buf
[*pos
])
260 if ('\\' != buf
[*pos
- 1])
265 /* XXX - save zeroed byte, if not an argument. */
273 (void)memset(&tmp
, 0, sizeof(struct mdoc_argv
));
277 /* See if our token accepts the argument. */
279 if (MDOC_ARG_MAX
== (tmp
.arg
= argv_a2arg(tok
, p
))) {
280 /* XXX - restore saved zeroed byte. */
283 if ( ! pwarn(mdoc
, line
, i
, WARGVPARM
))
288 while (buf
[*pos
] && ' ' == buf
[*pos
])
291 if ( ! argv(mdoc
, line
, &tmp
, pos
, buf
))
294 if (NULL
== (arg
= *v
)) {
295 *v
= xcalloc(1, sizeof(struct mdoc_arg
));
300 arg
->argv
= xrealloc(arg
->argv
, arg
->argc
*
301 sizeof(struct mdoc_argv
));
303 (void)memcpy(&arg
->argv
[(int)arg
->argc
- 1],
304 &tmp
, sizeof(struct mdoc_argv
));
311 mdoc_argv_free(struct mdoc_arg
*p
)
327 for (i
= 0; i
< (int)p
->argc
; i
++) {
328 if (0 == p
->argv
[i
].sz
)
331 for (j
= 0; j
< (int)p
->argv
[i
].sz
; j
++)
332 free(p
->argv
[i
].value
[j
]);
334 free(p
->argv
[i
].value
);
344 perr(struct mdoc
*mdoc
, int line
, int pos
, enum merr code
)
352 p
= "unterminated quoted parameter";
355 p
= "argument requires a value";
360 return(mdoc_perr(mdoc
, line
, pos
, p
));
365 pwarn(struct mdoc
*mdoc
, int line
, int pos
, enum mwarn code
)
375 p
= "unexpected quoted parameter";
378 p
= "argument-like parameter";
381 p
= "last list column is empty";
385 p
= "trailing whitespace";
391 return(mdoc_pwarn(mdoc
, line
, pos
, c
, p
));
396 mdoc_args(struct mdoc
*mdoc
, int line
,
397 int *pos
, char *buf
, int tok
, char **v
)
402 fl
= (0 == tok
) ? 0 : mdoc_argflags
[tok
];
405 * Override per-macro argument flags with context-specific ones.
406 * As of now, this is only valid for `It' depending on its list
412 for (n
= mdoc
->last
; n
; n
= n
->parent
)
413 if (MDOC_BLOCK
== n
->type
&& MDOC_Bl
== n
->tok
)
417 c
= (int)(n
->args
? n
->args
->argc
: 0);
421 * Using `Bl -column' adds ARGS_TABSEP to the arguments
422 * and invalidates ARGS_DELIM. Using `Bl -diag' allows
423 * for quoted arguments.
427 for (i
= 0; i
< c
; i
++) {
428 switch (n
->args
->argv
[i
].arg
) {
447 return(args(mdoc
, line
, pos
, buf
, fl
, v
));
452 args(struct mdoc
*mdoc
, int line
,
453 int *pos
, char *buf
, int fl
, char **v
)
463 if ('\"' == buf
[*pos
] && ! (fl
& ARGS_QUOTED
))
464 if ( ! pwarn(mdoc
, line
, *pos
, WQUOTPARM
))
467 if ( ! (fl
& ARGS_ARGVLIKE
) && '-' == buf
[*pos
])
468 if ( ! pwarn(mdoc
, line
, *pos
, WARGVPARM
))
472 * If the first character is a delimiter and we're to look for
473 * delimited strings, then pass down the buffer seeing if it
474 * follows the pattern of [[::delim::][ ]+]+.
477 if ((fl
& ARGS_DELIM
) && mdoc_iscdelim(buf
[*pos
])) {
478 for (i
= *pos
; buf
[i
]; ) {
479 if ( ! mdoc_iscdelim(buf
[i
]))
482 /* There must be at least one space... */
483 if (0 == buf
[i
] || ' ' != buf
[i
])
486 while (buf
[i
] && ' ' == buf
[i
])
495 /* First parse non-quoted strings. */
497 if ('\"' != buf
[*pos
] || ! (ARGS_QUOTED
& fl
)) {
501 * Thar be dragons here! If we're tab-separated, search
502 * ahead for either a tab or the `Ta' macro.
503 * If a `Ta' is detected, it must be space-buffered before and
504 * after. If either of these hold true, then prune out the
505 * extra spaces and call it an argument.
508 if (ARGS_TABSEP
& fl
) {
509 /* Scan ahead to unescaped tab. */
511 p
= strchr(*v
, '\t');
513 /* Scan ahead to unescaped `Ta'. */
515 for (pp
= *v
; ; pp
++) {
516 if (NULL
== (pp
= strstr(pp
, "Ta")))
518 if (pp
> *v
&& ' ' != *(pp
- 1))
520 if (' ' == *(pp
+ 2) || 0 == *(pp
+ 2))
524 /* Choose delimiter tab/Ta. */
527 p
= (p
< pp
? p
: pp
);
531 /* Strip delimiter's preceding whitespace. */
535 while (pp
> *v
&& ' ' == *pp
)
537 if (pp
== *v
&& ' ' == *pp
)
543 /* ...in- and proceding whitespace. */
545 if (p
&& ('\t' != *p
)) {
556 *pos
+= (int)(p
- *v
);
560 if ( ! pwarn(mdoc
, line
, *pos
, WCOLEMPTY
))
562 if (p
&& 0 == *p
&& p
> *v
&& ' ' == *(p
- 1))
563 if ( ! pwarn(mdoc
, line
, *pos
, WTAILWS
))
569 /* Configure the eoln case, too. */
574 if (p
> *v
&& ' ' == *(p
- 1))
575 if ( ! pwarn(mdoc
, line
, *pos
, WTAILWS
))
577 *pos
+= (int)(p
- *v
);
582 /* Do non-tabsep look-ahead here. */
584 if ( ! (ARGS_TABSEP
& fl
))
586 if (' ' == buf
[*pos
])
587 if ('\\' != buf
[*pos
- 1])
600 if ( ! (ARGS_TABSEP
& fl
))
601 while (buf
[*pos
] && ' ' == buf
[*pos
])
607 if ( ! pwarn(mdoc
, line
, *pos
, WTAILWS
))
614 * If we're a quoted string (and quoted strings are allowed),
615 * then parse ahead to the next quote. If none's found, it's an
616 * error. After, parse to the next word.
621 while (buf
[*pos
] && '\"' != buf
[*pos
])
624 if (0 == buf
[*pos
]) {
625 (void)perr(mdoc
, line
, *pos
, EQUOTTERM
);
633 while (buf
[*pos
] && ' ' == buf
[*pos
])
639 if ( ! pwarn(mdoc
, line
, *pos
, WTAILWS
))
647 argv_a2arg(int tok
, const char *argv
)
651 * Parse an argument identifier from its text. XXX - this
652 * should really be table-driven to clarify the code.
654 * If you add an argument to the list, make sure that you
655 * register it here with its one or more macros!
660 if (xstrcmp(argv
, "split"))
662 else if (xstrcmp(argv
, "nosplit"))
663 return(MDOC_Nosplit
);
667 if (xstrcmp(argv
, "ragged"))
669 else if (xstrcmp(argv
, "unfilled"))
670 return(MDOC_Unfilled
);
671 else if (xstrcmp(argv
, "filled"))
673 else if (xstrcmp(argv
, "literal"))
674 return(MDOC_Literal
);
675 else if (xstrcmp(argv
, "file"))
677 else if (xstrcmp(argv
, "offset"))
679 else if (xstrcmp(argv
, "compact"))
680 return(MDOC_Compact
);
684 if (xstrcmp(argv
, "emphasis"))
685 return(MDOC_Emphasis
);
686 else if (xstrcmp(argv
, "literal"))
687 return(MDOC_Literal
);
688 else if (xstrcmp(argv
, "symbolic"))
689 return(MDOC_Symbolic
);
693 if (xstrcmp(argv
, "words"))
698 if (xstrcmp(argv
, "bullet"))
700 else if (xstrcmp(argv
, "dash"))
702 else if (xstrcmp(argv
, "hyphen"))
704 else if (xstrcmp(argv
, "item"))
706 else if (xstrcmp(argv
, "enum"))
708 else if (xstrcmp(argv
, "tag"))
710 else if (xstrcmp(argv
, "diag"))
712 else if (xstrcmp(argv
, "hang"))
714 else if (xstrcmp(argv
, "ohang"))
716 else if (xstrcmp(argv
, "inset"))
718 else if (xstrcmp(argv
, "column"))
720 else if (xstrcmp(argv
, "width"))
722 else if (xstrcmp(argv
, "offset"))
724 else if (xstrcmp(argv
, "compact"))
725 return(MDOC_Compact
);
726 else if (xstrcmp(argv
, "nested"))
733 if (xstrcmp(argv
, "std"))
740 return(MDOC_ARG_MAX
);
745 argv_multi(struct mdoc
*mdoc
, int line
,
746 struct mdoc_argv
*v
, int *pos
, char *buf
)
753 for (v
->sz
= 0; ; v
->sz
++) {
754 if ('-' == buf
[*pos
])
756 c
= args(mdoc
, line
, pos
, buf
, ARGS_QUOTED
, &p
);
759 else if (ARGS_EOLN
== c
)
762 if (0 == v
->sz
% MULTI_STEP
)
763 v
->value
= xrealloc(v
->value
,
764 (v
->sz
+ MULTI_STEP
) * sizeof(char *));
766 v
->value
[(int)v
->sz
] = xstrdup(p
);
772 return(perr(mdoc
, line
, ppos
, EARGVAL
));
777 argv_opt_single(struct mdoc
*mdoc
, int line
,
778 struct mdoc_argv
*v
, int *pos
, char *buf
)
783 if ('-' == buf
[*pos
])
786 c
= args(mdoc
, line
, pos
, buf
, ARGS_QUOTED
, &p
);
793 v
->value
= xcalloc(1, sizeof(char *));
794 v
->value
[0] = xstrdup(p
);
800 * Parse a single, mandatory value from the stream.
803 argv_single(struct mdoc
*mdoc
, int line
,
804 struct mdoc_argv
*v
, int *pos
, char *buf
)
811 c
= args(mdoc
, line
, pos
, buf
, ARGS_QUOTED
, &p
);
815 return(perr(mdoc
, line
, ppos
, EARGVAL
));
818 v
->value
= xcalloc(1, sizeof(char *));
819 v
->value
[0] = xstrdup(p
);
825 * Determine rules for parsing arguments. Arguments can either accept
826 * no parameters, an optional single parameter, one parameter, or
827 * multiple parameters.
830 argv(struct mdoc
*mdoc
, int line
,
831 struct mdoc_argv
*v
, int *pos
, char *buf
)
837 switch (mdoc_argvflags
[v
->arg
]) {
839 return(argv_single(mdoc
, line
, v
, pos
, buf
));
841 return(argv_multi(mdoc
, line
, v
, pos
, buf
));
842 case (ARGV_OPT_SINGLE
):
843 return(argv_opt_single(mdoc
, line
, v
, pos
, buf
));