1 |
|
|
/* $OpenBSD: mdoc_argv.c,v 1.64 2016/07/15 18:02:32 schwarze Exp $ */ |
2 |
|
|
/* |
3 |
|
|
* Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv> |
4 |
|
|
* Copyright (c) 2012, 2014, 2015 Ingo Schwarze <schwarze@openbsd.org> |
5 |
|
|
* |
6 |
|
|
* Permission to use, copy, modify, and distribute this software for any |
7 |
|
|
* purpose with or without fee is hereby granted, provided that the above |
8 |
|
|
* copyright notice and this permission notice appear in all copies. |
9 |
|
|
* |
10 |
|
|
* THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHORS DISCLAIM ALL WARRANTIES |
11 |
|
|
* WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF |
12 |
|
|
* MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHORS BE LIABLE FOR |
13 |
|
|
* ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES |
14 |
|
|
* WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN |
15 |
|
|
* ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF |
16 |
|
|
* OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE. |
17 |
|
|
*/ |
18 |
|
|
#include <sys/types.h> |
19 |
|
|
|
20 |
|
|
#include <assert.h> |
21 |
|
|
#include <stdlib.h> |
22 |
|
|
#include <stdio.h> |
23 |
|
|
#include <string.h> |
24 |
|
|
|
25 |
|
|
#include "mandoc_aux.h" |
26 |
|
|
#include "mandoc.h" |
27 |
|
|
#include "roff.h" |
28 |
|
|
#include "mdoc.h" |
29 |
|
|
#include "libmandoc.h" |
30 |
|
|
#include "roff_int.h" |
31 |
|
|
#include "libmdoc.h" |
32 |
|
|
|
33 |
|
|
#define MULTI_STEP 5 /* pre-allocate argument values */ |
34 |
|
|
#define DELIMSZ 6 /* max possible size of a delimiter */ |
35 |
|
|
|
36 |
|
|
enum argsflag { |
37 |
|
|
ARGSFL_NONE = 0, |
38 |
|
|
ARGSFL_DELIM, /* handle delimiters of [[::delim::][ ]+]+ */ |
39 |
|
|
ARGSFL_TABSEP /* handle tab/`Ta' separated phrases */ |
40 |
|
|
}; |
41 |
|
|
|
42 |
|
|
enum argvflag { |
43 |
|
|
ARGV_NONE, /* no args to flag (e.g., -split) */ |
44 |
|
|
ARGV_SINGLE, /* one arg to flag (e.g., -file xxx) */ |
45 |
|
|
ARGV_MULTI /* multiple args (e.g., -column xxx yyy) */ |
46 |
|
|
}; |
47 |
|
|
|
48 |
|
|
struct mdocarg { |
49 |
|
|
enum argsflag flags; |
50 |
|
|
const enum mdocargt *argvs; |
51 |
|
|
}; |
52 |
|
|
|
53 |
|
|
static void argn_free(struct mdoc_arg *, int); |
54 |
|
|
static enum margserr args(struct roff_man *, int, int *, |
55 |
|
|
char *, enum argsflag, char **); |
56 |
|
|
static int args_checkpunct(const char *, int); |
57 |
|
|
static void argv_multi(struct roff_man *, int, |
58 |
|
|
struct mdoc_argv *, int *, char *); |
59 |
|
|
static void argv_single(struct roff_man *, int, |
60 |
|
|
struct mdoc_argv *, int *, char *); |
61 |
|
|
|
62 |
|
|
static const enum argvflag argvflags[MDOC_ARG_MAX] = { |
63 |
|
|
ARGV_NONE, /* MDOC_Split */ |
64 |
|
|
ARGV_NONE, /* MDOC_Nosplit */ |
65 |
|
|
ARGV_NONE, /* MDOC_Ragged */ |
66 |
|
|
ARGV_NONE, /* MDOC_Unfilled */ |
67 |
|
|
ARGV_NONE, /* MDOC_Literal */ |
68 |
|
|
ARGV_SINGLE, /* MDOC_File */ |
69 |
|
|
ARGV_SINGLE, /* MDOC_Offset */ |
70 |
|
|
ARGV_NONE, /* MDOC_Bullet */ |
71 |
|
|
ARGV_NONE, /* MDOC_Dash */ |
72 |
|
|
ARGV_NONE, /* MDOC_Hyphen */ |
73 |
|
|
ARGV_NONE, /* MDOC_Item */ |
74 |
|
|
ARGV_NONE, /* MDOC_Enum */ |
75 |
|
|
ARGV_NONE, /* MDOC_Tag */ |
76 |
|
|
ARGV_NONE, /* MDOC_Diag */ |
77 |
|
|
ARGV_NONE, /* MDOC_Hang */ |
78 |
|
|
ARGV_NONE, /* MDOC_Ohang */ |
79 |
|
|
ARGV_NONE, /* MDOC_Inset */ |
80 |
|
|
ARGV_MULTI, /* MDOC_Column */ |
81 |
|
|
ARGV_SINGLE, /* MDOC_Width */ |
82 |
|
|
ARGV_NONE, /* MDOC_Compact */ |
83 |
|
|
ARGV_NONE, /* MDOC_Std */ |
84 |
|
|
ARGV_NONE, /* MDOC_Filled */ |
85 |
|
|
ARGV_NONE, /* MDOC_Words */ |
86 |
|
|
ARGV_NONE, /* MDOC_Emphasis */ |
87 |
|
|
ARGV_NONE, /* MDOC_Symbolic */ |
88 |
|
|
ARGV_NONE /* MDOC_Symbolic */ |
89 |
|
|
}; |
90 |
|
|
|
91 |
|
|
static const enum mdocargt args_Ex[] = { |
92 |
|
|
MDOC_Std, |
93 |
|
|
MDOC_ARG_MAX |
94 |
|
|
}; |
95 |
|
|
|
96 |
|
|
static const enum mdocargt args_An[] = { |
97 |
|
|
MDOC_Split, |
98 |
|
|
MDOC_Nosplit, |
99 |
|
|
MDOC_ARG_MAX |
100 |
|
|
}; |
101 |
|
|
|
102 |
|
|
static const enum mdocargt args_Bd[] = { |
103 |
|
|
MDOC_Ragged, |
104 |
|
|
MDOC_Unfilled, |
105 |
|
|
MDOC_Filled, |
106 |
|
|
MDOC_Literal, |
107 |
|
|
MDOC_File, |
108 |
|
|
MDOC_Offset, |
109 |
|
|
MDOC_Compact, |
110 |
|
|
MDOC_Centred, |
111 |
|
|
MDOC_ARG_MAX |
112 |
|
|
}; |
113 |
|
|
|
114 |
|
|
static const enum mdocargt args_Bf[] = { |
115 |
|
|
MDOC_Emphasis, |
116 |
|
|
MDOC_Literal, |
117 |
|
|
MDOC_Symbolic, |
118 |
|
|
MDOC_ARG_MAX |
119 |
|
|
}; |
120 |
|
|
|
121 |
|
|
static const enum mdocargt args_Bk[] = { |
122 |
|
|
MDOC_Words, |
123 |
|
|
MDOC_ARG_MAX |
124 |
|
|
}; |
125 |
|
|
|
126 |
|
|
static const enum mdocargt args_Bl[] = { |
127 |
|
|
MDOC_Bullet, |
128 |
|
|
MDOC_Dash, |
129 |
|
|
MDOC_Hyphen, |
130 |
|
|
MDOC_Item, |
131 |
|
|
MDOC_Enum, |
132 |
|
|
MDOC_Tag, |
133 |
|
|
MDOC_Diag, |
134 |
|
|
MDOC_Hang, |
135 |
|
|
MDOC_Ohang, |
136 |
|
|
MDOC_Inset, |
137 |
|
|
MDOC_Column, |
138 |
|
|
MDOC_Width, |
139 |
|
|
MDOC_Offset, |
140 |
|
|
MDOC_Compact, |
141 |
|
|
MDOC_Nested, |
142 |
|
|
MDOC_ARG_MAX |
143 |
|
|
}; |
144 |
|
|
|
145 |
|
|
static const struct mdocarg mdocargs[MDOC_MAX] = { |
146 |
|
|
{ ARGSFL_DELIM, NULL }, /* Ap */ |
147 |
|
|
{ ARGSFL_NONE, NULL }, /* Dd */ |
148 |
|
|
{ ARGSFL_NONE, NULL }, /* Dt */ |
149 |
|
|
{ ARGSFL_NONE, NULL }, /* Os */ |
150 |
|
|
{ ARGSFL_NONE, NULL }, /* Sh */ |
151 |
|
|
{ ARGSFL_NONE, NULL }, /* Ss */ |
152 |
|
|
{ ARGSFL_NONE, NULL }, /* Pp */ |
153 |
|
|
{ ARGSFL_DELIM, NULL }, /* D1 */ |
154 |
|
|
{ ARGSFL_DELIM, NULL }, /* Dl */ |
155 |
|
|
{ ARGSFL_NONE, args_Bd }, /* Bd */ |
156 |
|
|
{ ARGSFL_NONE, NULL }, /* Ed */ |
157 |
|
|
{ ARGSFL_NONE, args_Bl }, /* Bl */ |
158 |
|
|
{ ARGSFL_NONE, NULL }, /* El */ |
159 |
|
|
{ ARGSFL_NONE, NULL }, /* It */ |
160 |
|
|
{ ARGSFL_DELIM, NULL }, /* Ad */ |
161 |
|
|
{ ARGSFL_DELIM, args_An }, /* An */ |
162 |
|
|
{ ARGSFL_DELIM, NULL }, /* Ar */ |
163 |
|
|
{ ARGSFL_DELIM, NULL }, /* Cd */ |
164 |
|
|
{ ARGSFL_DELIM, NULL }, /* Cm */ |
165 |
|
|
{ ARGSFL_DELIM, NULL }, /* Dv */ |
166 |
|
|
{ ARGSFL_DELIM, NULL }, /* Er */ |
167 |
|
|
{ ARGSFL_DELIM, NULL }, /* Ev */ |
168 |
|
|
{ ARGSFL_NONE, args_Ex }, /* Ex */ |
169 |
|
|
{ ARGSFL_DELIM, NULL }, /* Fa */ |
170 |
|
|
{ ARGSFL_NONE, NULL }, /* Fd */ |
171 |
|
|
{ ARGSFL_DELIM, NULL }, /* Fl */ |
172 |
|
|
{ ARGSFL_DELIM, NULL }, /* Fn */ |
173 |
|
|
{ ARGSFL_DELIM, NULL }, /* Ft */ |
174 |
|
|
{ ARGSFL_DELIM, NULL }, /* Ic */ |
175 |
|
|
{ ARGSFL_DELIM, NULL }, /* In */ |
176 |
|
|
{ ARGSFL_DELIM, NULL }, /* Li */ |
177 |
|
|
{ ARGSFL_NONE, NULL }, /* Nd */ |
178 |
|
|
{ ARGSFL_DELIM, NULL }, /* Nm */ |
179 |
|
|
{ ARGSFL_DELIM, NULL }, /* Op */ |
180 |
|
|
{ ARGSFL_DELIM, NULL }, /* Ot */ |
181 |
|
|
{ ARGSFL_DELIM, NULL }, /* Pa */ |
182 |
|
|
{ ARGSFL_NONE, args_Ex }, /* Rv */ |
183 |
|
|
{ ARGSFL_DELIM, NULL }, /* St */ |
184 |
|
|
{ ARGSFL_DELIM, NULL }, /* Va */ |
185 |
|
|
{ ARGSFL_DELIM, NULL }, /* Vt */ |
186 |
|
|
{ ARGSFL_DELIM, NULL }, /* Xr */ |
187 |
|
|
{ ARGSFL_NONE, NULL }, /* %A */ |
188 |
|
|
{ ARGSFL_NONE, NULL }, /* %B */ |
189 |
|
|
{ ARGSFL_NONE, NULL }, /* %D */ |
190 |
|
|
{ ARGSFL_NONE, NULL }, /* %I */ |
191 |
|
|
{ ARGSFL_NONE, NULL }, /* %J */ |
192 |
|
|
{ ARGSFL_NONE, NULL }, /* %N */ |
193 |
|
|
{ ARGSFL_NONE, NULL }, /* %O */ |
194 |
|
|
{ ARGSFL_NONE, NULL }, /* %P */ |
195 |
|
|
{ ARGSFL_NONE, NULL }, /* %R */ |
196 |
|
|
{ ARGSFL_NONE, NULL }, /* %T */ |
197 |
|
|
{ ARGSFL_NONE, NULL }, /* %V */ |
198 |
|
|
{ ARGSFL_DELIM, NULL }, /* Ac */ |
199 |
|
|
{ ARGSFL_NONE, NULL }, /* Ao */ |
200 |
|
|
{ ARGSFL_DELIM, NULL }, /* Aq */ |
201 |
|
|
{ ARGSFL_DELIM, NULL }, /* At */ |
202 |
|
|
{ ARGSFL_DELIM, NULL }, /* Bc */ |
203 |
|
|
{ ARGSFL_NONE, args_Bf }, /* Bf */ |
204 |
|
|
{ ARGSFL_NONE, NULL }, /* Bo */ |
205 |
|
|
{ ARGSFL_DELIM, NULL }, /* Bq */ |
206 |
|
|
{ ARGSFL_DELIM, NULL }, /* Bsx */ |
207 |
|
|
{ ARGSFL_DELIM, NULL }, /* Bx */ |
208 |
|
|
{ ARGSFL_NONE, NULL }, /* Db */ |
209 |
|
|
{ ARGSFL_DELIM, NULL }, /* Dc */ |
210 |
|
|
{ ARGSFL_NONE, NULL }, /* Do */ |
211 |
|
|
{ ARGSFL_DELIM, NULL }, /* Dq */ |
212 |
|
|
{ ARGSFL_DELIM, NULL }, /* Ec */ |
213 |
|
|
{ ARGSFL_NONE, NULL }, /* Ef */ |
214 |
|
|
{ ARGSFL_DELIM, NULL }, /* Em */ |
215 |
|
|
{ ARGSFL_NONE, NULL }, /* Eo */ |
216 |
|
|
{ ARGSFL_DELIM, NULL }, /* Fx */ |
217 |
|
|
{ ARGSFL_DELIM, NULL }, /* Ms */ |
218 |
|
|
{ ARGSFL_DELIM, NULL }, /* No */ |
219 |
|
|
{ ARGSFL_DELIM, NULL }, /* Ns */ |
220 |
|
|
{ ARGSFL_DELIM, NULL }, /* Nx */ |
221 |
|
|
{ ARGSFL_DELIM, NULL }, /* Ox */ |
222 |
|
|
{ ARGSFL_DELIM, NULL }, /* Pc */ |
223 |
|
|
{ ARGSFL_DELIM, NULL }, /* Pf */ |
224 |
|
|
{ ARGSFL_NONE, NULL }, /* Po */ |
225 |
|
|
{ ARGSFL_DELIM, NULL }, /* Pq */ |
226 |
|
|
{ ARGSFL_DELIM, NULL }, /* Qc */ |
227 |
|
|
{ ARGSFL_DELIM, NULL }, /* Ql */ |
228 |
|
|
{ ARGSFL_NONE, NULL }, /* Qo */ |
229 |
|
|
{ ARGSFL_DELIM, NULL }, /* Qq */ |
230 |
|
|
{ ARGSFL_NONE, NULL }, /* Re */ |
231 |
|
|
{ ARGSFL_NONE, NULL }, /* Rs */ |
232 |
|
|
{ ARGSFL_DELIM, NULL }, /* Sc */ |
233 |
|
|
{ ARGSFL_NONE, NULL }, /* So */ |
234 |
|
|
{ ARGSFL_DELIM, NULL }, /* Sq */ |
235 |
|
|
{ ARGSFL_NONE, NULL }, /* Sm */ |
236 |
|
|
{ ARGSFL_DELIM, NULL }, /* Sx */ |
237 |
|
|
{ ARGSFL_DELIM, NULL }, /* Sy */ |
238 |
|
|
{ ARGSFL_DELIM, NULL }, /* Tn */ |
239 |
|
|
{ ARGSFL_DELIM, NULL }, /* Ux */ |
240 |
|
|
{ ARGSFL_DELIM, NULL }, /* Xc */ |
241 |
|
|
{ ARGSFL_NONE, NULL }, /* Xo */ |
242 |
|
|
{ ARGSFL_NONE, NULL }, /* Fo */ |
243 |
|
|
{ ARGSFL_DELIM, NULL }, /* Fc */ |
244 |
|
|
{ ARGSFL_NONE, NULL }, /* Oo */ |
245 |
|
|
{ ARGSFL_DELIM, NULL }, /* Oc */ |
246 |
|
|
{ ARGSFL_NONE, args_Bk }, /* Bk */ |
247 |
|
|
{ ARGSFL_NONE, NULL }, /* Ek */ |
248 |
|
|
{ ARGSFL_NONE, NULL }, /* Bt */ |
249 |
|
|
{ ARGSFL_NONE, NULL }, /* Hf */ |
250 |
|
|
{ ARGSFL_DELIM, NULL }, /* Fr */ |
251 |
|
|
{ ARGSFL_NONE, NULL }, /* Ud */ |
252 |
|
|
{ ARGSFL_DELIM, NULL }, /* Lb */ |
253 |
|
|
{ ARGSFL_NONE, NULL }, /* Lp */ |
254 |
|
|
{ ARGSFL_DELIM, NULL }, /* Lk */ |
255 |
|
|
{ ARGSFL_DELIM, NULL }, /* Mt */ |
256 |
|
|
{ ARGSFL_DELIM, NULL }, /* Brq */ |
257 |
|
|
{ ARGSFL_NONE, NULL }, /* Bro */ |
258 |
|
|
{ ARGSFL_DELIM, NULL }, /* Brc */ |
259 |
|
|
{ ARGSFL_NONE, NULL }, /* %C */ |
260 |
|
|
{ ARGSFL_NONE, NULL }, /* Es */ |
261 |
|
|
{ ARGSFL_DELIM, NULL }, /* En */ |
262 |
|
|
{ ARGSFL_DELIM, NULL }, /* Dx */ |
263 |
|
|
{ ARGSFL_NONE, NULL }, /* %Q */ |
264 |
|
|
{ ARGSFL_NONE, NULL }, /* br */ |
265 |
|
|
{ ARGSFL_NONE, NULL }, /* sp */ |
266 |
|
|
{ ARGSFL_NONE, NULL }, /* %U */ |
267 |
|
|
{ ARGSFL_NONE, NULL }, /* Ta */ |
268 |
|
|
{ ARGSFL_NONE, NULL }, /* ll */ |
269 |
|
|
}; |
270 |
|
|
|
271 |
|
|
|
272 |
|
|
/* |
273 |
|
|
* Parse flags and their arguments from the input line. |
274 |
|
|
* These come in the form -flag [argument ...]. |
275 |
|
|
* Some flags take no argument, some one, some multiple. |
276 |
|
|
*/ |
277 |
|
|
void |
278 |
|
|
mdoc_argv(struct roff_man *mdoc, int line, int tok, |
279 |
|
|
struct mdoc_arg **reta, int *pos, char *buf) |
280 |
|
|
{ |
281 |
|
|
struct mdoc_argv tmpv; |
282 |
|
|
struct mdoc_argv **retv; |
283 |
|
|
const enum mdocargt *argtable; |
284 |
|
|
char *argname; |
285 |
|
|
int ipos, retc; |
286 |
|
|
char savechar; |
287 |
|
|
|
288 |
|
|
*reta = NULL; |
289 |
|
|
|
290 |
|
|
/* Which flags does this macro support? */ |
291 |
|
|
|
292 |
|
|
argtable = mdocargs[tok].argvs; |
293 |
|
|
if (argtable == NULL) |
294 |
|
|
return; |
295 |
|
|
|
296 |
|
|
/* Loop over the flags on the input line. */ |
297 |
|
|
|
298 |
|
|
ipos = *pos; |
299 |
|
|
while (buf[ipos] == '-') { |
300 |
|
|
|
301 |
|
|
/* Seek to the first unescaped space. */ |
302 |
|
|
|
303 |
|
|
for (argname = buf + ++ipos; buf[ipos] != '\0'; ipos++) |
304 |
|
|
if (buf[ipos] == ' ' && buf[ipos - 1] != '\\') |
305 |
|
|
break; |
306 |
|
|
|
307 |
|
|
/* |
308 |
|
|
* We want to nil-terminate the word to look it up. |
309 |
|
|
* But we may not have a flag, in which case we need |
310 |
|
|
* to restore the line as-is. So keep around the |
311 |
|
|
* stray byte, which we'll reset upon exiting. |
312 |
|
|
*/ |
313 |
|
|
|
314 |
|
|
if ((savechar = buf[ipos]) != '\0') |
315 |
|
|
buf[ipos++] = '\0'; |
316 |
|
|
|
317 |
|
|
/* |
318 |
|
|
* Now look up the word as a flag. Use temporary |
319 |
|
|
* storage that we'll copy into the node's flags. |
320 |
|
|
*/ |
321 |
|
|
|
322 |
|
|
while ((tmpv.arg = *argtable++) != MDOC_ARG_MAX) |
323 |
|
|
if ( ! strcmp(argname, mdoc_argnames[tmpv.arg])) |
324 |
|
|
break; |
325 |
|
|
|
326 |
|
|
/* If it isn't a flag, restore the saved byte. */ |
327 |
|
|
|
328 |
|
|
if (tmpv.arg == MDOC_ARG_MAX) { |
329 |
|
|
if (savechar != '\0') |
330 |
|
|
buf[ipos - 1] = savechar; |
331 |
|
|
break; |
332 |
|
|
} |
333 |
|
|
|
334 |
|
|
/* Read to the next word (the first argument). */ |
335 |
|
|
|
336 |
|
|
while (buf[ipos] == ' ') |
337 |
|
|
ipos++; |
338 |
|
|
|
339 |
|
|
/* Parse the arguments of the flag. */ |
340 |
|
|
|
341 |
|
|
tmpv.line = line; |
342 |
|
|
tmpv.pos = *pos; |
343 |
|
|
tmpv.sz = 0; |
344 |
|
|
tmpv.value = NULL; |
345 |
|
|
|
346 |
|
|
switch (argvflags[tmpv.arg]) { |
347 |
|
|
case ARGV_SINGLE: |
348 |
|
|
argv_single(mdoc, line, &tmpv, &ipos, buf); |
349 |
|
|
break; |
350 |
|
|
case ARGV_MULTI: |
351 |
|
|
argv_multi(mdoc, line, &tmpv, &ipos, buf); |
352 |
|
|
break; |
353 |
|
|
case ARGV_NONE: |
354 |
|
|
break; |
355 |
|
|
} |
356 |
|
|
|
357 |
|
|
/* Append to the return values. */ |
358 |
|
|
|
359 |
|
|
if (*reta == NULL) |
360 |
|
|
*reta = mandoc_calloc(1, sizeof(**reta)); |
361 |
|
|
|
362 |
|
|
retc = ++(*reta)->argc; |
363 |
|
|
retv = &(*reta)->argv; |
364 |
|
|
*retv = mandoc_reallocarray(*retv, retc, sizeof(**retv)); |
365 |
|
|
memcpy(*retv + retc - 1, &tmpv, sizeof(**retv)); |
366 |
|
|
|
367 |
|
|
/* Prepare for parsing the next flag. */ |
368 |
|
|
|
369 |
|
|
*pos = ipos; |
370 |
|
|
argtable = mdocargs[tok].argvs; |
371 |
|
|
} |
372 |
|
|
} |
373 |
|
|
|
374 |
|
|
void |
375 |
|
|
mdoc_argv_free(struct mdoc_arg *p) |
376 |
|
|
{ |
377 |
|
|
int i; |
378 |
|
|
|
379 |
|
|
if (NULL == p) |
380 |
|
|
return; |
381 |
|
|
|
382 |
|
|
if (p->refcnt) { |
383 |
|
|
--(p->refcnt); |
384 |
|
|
if (p->refcnt) |
385 |
|
|
return; |
386 |
|
|
} |
387 |
|
|
assert(p->argc); |
388 |
|
|
|
389 |
|
|
for (i = (int)p->argc - 1; i >= 0; i--) |
390 |
|
|
argn_free(p, i); |
391 |
|
|
|
392 |
|
|
free(p->argv); |
393 |
|
|
free(p); |
394 |
|
|
} |
395 |
|
|
|
396 |
|
|
static void |
397 |
|
|
argn_free(struct mdoc_arg *p, int iarg) |
398 |
|
|
{ |
399 |
|
|
struct mdoc_argv *arg; |
400 |
|
|
int j; |
401 |
|
|
|
402 |
|
|
arg = &p->argv[iarg]; |
403 |
|
|
|
404 |
|
|
if (arg->sz && arg->value) { |
405 |
|
|
for (j = (int)arg->sz - 1; j >= 0; j--) |
406 |
|
|
free(arg->value[j]); |
407 |
|
|
free(arg->value); |
408 |
|
|
} |
409 |
|
|
|
410 |
|
|
for (--p->argc; iarg < (int)p->argc; iarg++) |
411 |
|
|
p->argv[iarg] = p->argv[iarg+1]; |
412 |
|
|
} |
413 |
|
|
|
414 |
|
|
enum margserr |
415 |
|
|
mdoc_args(struct roff_man *mdoc, int line, int *pos, |
416 |
|
|
char *buf, int tok, char **v) |
417 |
|
|
{ |
418 |
|
|
struct roff_node *n; |
419 |
|
|
char *v_local; |
420 |
|
|
enum argsflag fl; |
421 |
|
|
|
422 |
|
|
if (v == NULL) |
423 |
|
|
v = &v_local; |
424 |
|
|
fl = tok == TOKEN_NONE ? ARGSFL_NONE : mdocargs[tok].flags; |
425 |
|
|
if (tok != MDOC_It) |
426 |
|
|
return args(mdoc, line, pos, buf, fl, v); |
427 |
|
|
|
428 |
|
|
/* |
429 |
|
|
* We know that we're in an `It', so it's reasonable to expect |
430 |
|
|
* us to be sitting in a `Bl'. Someday this may not be the case |
431 |
|
|
* (if we allow random `It's sitting out there), so provide a |
432 |
|
|
* safe fall-back into the default behaviour. |
433 |
|
|
*/ |
434 |
|
|
|
435 |
|
|
for (n = mdoc->last; n; n = n->parent) |
436 |
|
|
if (MDOC_Bl == n->tok) |
437 |
|
|
if (LIST_column == n->norm->Bl.type) { |
438 |
|
|
fl = ARGSFL_TABSEP; |
439 |
|
|
break; |
440 |
|
|
} |
441 |
|
|
|
442 |
|
|
return args(mdoc, line, pos, buf, fl, v); |
443 |
|
|
} |
444 |
|
|
|
445 |
|
|
static enum margserr |
446 |
|
|
args(struct roff_man *mdoc, int line, int *pos, |
447 |
|
|
char *buf, enum argsflag fl, char **v) |
448 |
|
|
{ |
449 |
|
|
char *p; |
450 |
|
|
int pairs; |
451 |
|
|
|
452 |
|
|
if (buf[*pos] == '\0') { |
453 |
|
|
if (mdoc->flags & MDOC_PHRASELIT && |
454 |
|
|
! (mdoc->flags & MDOC_PHRASE)) { |
455 |
|
|
mandoc_msg(MANDOCERR_ARG_QUOTE, |
456 |
|
|
mdoc->parse, line, *pos, NULL); |
457 |
|
|
mdoc->flags &= ~MDOC_PHRASELIT; |
458 |
|
|
} |
459 |
|
|
return ARGS_EOLN; |
460 |
|
|
} |
461 |
|
|
|
462 |
|
|
*v = buf + *pos; |
463 |
|
|
|
464 |
|
|
if (fl == ARGSFL_DELIM && args_checkpunct(buf, *pos)) |
465 |
|
|
return ARGS_PUNCT; |
466 |
|
|
|
467 |
|
|
/* |
468 |
|
|
* Tabs in `It' lines in `Bl -column' can't be escaped. |
469 |
|
|
* Phrases are reparsed for `Ta' and other macros later. |
470 |
|
|
*/ |
471 |
|
|
|
472 |
|
|
if (fl == ARGSFL_TABSEP) { |
473 |
|
|
if ((p = strchr(*v, '\t')) != NULL) { |
474 |
|
|
|
475 |
|
|
/* |
476 |
|
|
* Words right before and right after |
477 |
|
|
* tab characters are not parsed, |
478 |
|
|
* unless there is a blank in between. |
479 |
|
|
*/ |
480 |
|
|
|
481 |
|
|
if (p[-1] != ' ') |
482 |
|
|
mdoc->flags |= MDOC_PHRASEQL; |
483 |
|
|
if (p[1] != ' ') |
484 |
|
|
mdoc->flags |= MDOC_PHRASEQN; |
485 |
|
|
|
486 |
|
|
/* |
487 |
|
|
* One or more blanks after a tab cause |
488 |
|
|
* one leading blank in the next column. |
489 |
|
|
* So skip all but one of them. |
490 |
|
|
*/ |
491 |
|
|
|
492 |
|
|
*pos += (int)(p - *v) + 1; |
493 |
|
|
while (buf[*pos] == ' ' && buf[*pos + 1] == ' ') |
494 |
|
|
(*pos)++; |
495 |
|
|
|
496 |
|
|
/* |
497 |
|
|
* A tab at the end of an input line |
498 |
|
|
* switches to the next column. |
499 |
|
|
*/ |
500 |
|
|
|
501 |
|
|
if (buf[*pos] == '\0' || buf[*pos + 1] == '\0') |
502 |
|
|
mdoc->flags |= MDOC_PHRASEQN; |
503 |
|
|
} else { |
504 |
|
|
p = strchr(*v, '\0'); |
505 |
|
|
if (p[-1] == ' ') |
506 |
|
|
mandoc_msg(MANDOCERR_SPACE_EOL, |
507 |
|
|
mdoc->parse, line, *pos, NULL); |
508 |
|
|
*pos += (int)(p - *v); |
509 |
|
|
} |
510 |
|
|
|
511 |
|
|
/* Skip any trailing blank characters. */ |
512 |
|
|
while (p > *v && p[-1] == ' ' && |
513 |
|
|
(p - 1 == *v || p[-2] != '\\')) |
514 |
|
|
p--; |
515 |
|
|
*p = '\0'; |
516 |
|
|
|
517 |
|
|
return ARGS_PHRASE; |
518 |
|
|
} |
519 |
|
|
|
520 |
|
|
/* |
521 |
|
|
* Process a quoted literal. A quote begins with a double-quote |
522 |
|
|
* and ends with a double-quote NOT preceded by a double-quote. |
523 |
|
|
* NUL-terminate the literal in place. |
524 |
|
|
* Collapse pairs of quotes inside quoted literals. |
525 |
|
|
* Whitespace is NOT involved in literal termination. |
526 |
|
|
*/ |
527 |
|
|
|
528 |
|
|
if (mdoc->flags & MDOC_PHRASELIT || buf[*pos] == '\"') { |
529 |
|
|
if ( ! (mdoc->flags & MDOC_PHRASELIT)) |
530 |
|
|
*v = &buf[++(*pos)]; |
531 |
|
|
|
532 |
|
|
if (mdoc->flags & MDOC_PHRASE) |
533 |
|
|
mdoc->flags |= MDOC_PHRASELIT; |
534 |
|
|
|
535 |
|
|
pairs = 0; |
536 |
|
|
for ( ; buf[*pos]; (*pos)++) { |
537 |
|
|
/* Move following text left after quoted quotes. */ |
538 |
|
|
if (pairs) |
539 |
|
|
buf[*pos - pairs] = buf[*pos]; |
540 |
|
|
if ('\"' != buf[*pos]) |
541 |
|
|
continue; |
542 |
|
|
/* Unquoted quotes end quoted args. */ |
543 |
|
|
if ('\"' != buf[*pos + 1]) |
544 |
|
|
break; |
545 |
|
|
/* Quoted quotes collapse. */ |
546 |
|
|
pairs++; |
547 |
|
|
(*pos)++; |
548 |
|
|
} |
549 |
|
|
if (pairs) |
550 |
|
|
buf[*pos - pairs] = '\0'; |
551 |
|
|
|
552 |
|
|
if (buf[*pos] == '\0') { |
553 |
|
|
if ( ! (mdoc->flags & MDOC_PHRASE)) |
554 |
|
|
mandoc_msg(MANDOCERR_ARG_QUOTE, |
555 |
|
|
mdoc->parse, line, *pos, NULL); |
556 |
|
|
return ARGS_QWORD; |
557 |
|
|
} |
558 |
|
|
|
559 |
|
|
mdoc->flags &= ~MDOC_PHRASELIT; |
560 |
|
|
buf[(*pos)++] = '\0'; |
561 |
|
|
|
562 |
|
|
if ('\0' == buf[*pos]) |
563 |
|
|
return ARGS_QWORD; |
564 |
|
|
|
565 |
|
|
while (' ' == buf[*pos]) |
566 |
|
|
(*pos)++; |
567 |
|
|
|
568 |
|
|
if ('\0' == buf[*pos]) |
569 |
|
|
mandoc_msg(MANDOCERR_SPACE_EOL, mdoc->parse, |
570 |
|
|
line, *pos, NULL); |
571 |
|
|
|
572 |
|
|
return ARGS_QWORD; |
573 |
|
|
} |
574 |
|
|
|
575 |
|
|
p = &buf[*pos]; |
576 |
|
|
*v = mandoc_getarg(mdoc->parse, &p, line, pos); |
577 |
|
|
|
578 |
|
|
/* |
579 |
|
|
* After parsing the last word in this phrase, |
580 |
|
|
* tell lookup() whether or not to interpret it. |
581 |
|
|
*/ |
582 |
|
|
|
583 |
|
|
if (*p == '\0' && mdoc->flags & MDOC_PHRASEQL) { |
584 |
|
|
mdoc->flags &= ~MDOC_PHRASEQL; |
585 |
|
|
mdoc->flags |= MDOC_PHRASEQF; |
586 |
|
|
} |
587 |
|
|
return ARGS_WORD; |
588 |
|
|
} |
589 |
|
|
|
590 |
|
|
/* |
591 |
|
|
* Check if the string consists only of space-separated closing |
592 |
|
|
* delimiters. This is a bit of a dance: the first must be a close |
593 |
|
|
* delimiter, but it may be followed by middle delimiters. Arbitrary |
594 |
|
|
* whitespace may separate these tokens. |
595 |
|
|
*/ |
596 |
|
|
static int |
597 |
|
|
args_checkpunct(const char *buf, int i) |
598 |
|
|
{ |
599 |
|
|
int j; |
600 |
|
|
char dbuf[DELIMSZ]; |
601 |
|
|
enum mdelim d; |
602 |
|
|
|
603 |
|
|
/* First token must be a close-delimiter. */ |
604 |
|
|
|
605 |
|
|
for (j = 0; buf[i] && ' ' != buf[i] && j < DELIMSZ; j++, i++) |
606 |
|
|
dbuf[j] = buf[i]; |
607 |
|
|
|
608 |
|
|
if (DELIMSZ == j) |
609 |
|
|
return 0; |
610 |
|
|
|
611 |
|
|
dbuf[j] = '\0'; |
612 |
|
|
if (DELIM_CLOSE != mdoc_isdelim(dbuf)) |
613 |
|
|
return 0; |
614 |
|
|
|
615 |
|
|
while (' ' == buf[i]) |
616 |
|
|
i++; |
617 |
|
|
|
618 |
|
|
/* Remaining must NOT be open/none. */ |
619 |
|
|
|
620 |
|
|
while (buf[i]) { |
621 |
|
|
j = 0; |
622 |
|
|
while (buf[i] && ' ' != buf[i] && j < DELIMSZ) |
623 |
|
|
dbuf[j++] = buf[i++]; |
624 |
|
|
|
625 |
|
|
if (DELIMSZ == j) |
626 |
|
|
return 0; |
627 |
|
|
|
628 |
|
|
dbuf[j] = '\0'; |
629 |
|
|
d = mdoc_isdelim(dbuf); |
630 |
|
|
if (DELIM_NONE == d || DELIM_OPEN == d) |
631 |
|
|
return 0; |
632 |
|
|
|
633 |
|
|
while (' ' == buf[i]) |
634 |
|
|
i++; |
635 |
|
|
} |
636 |
|
|
|
637 |
|
|
return '\0' == buf[i]; |
638 |
|
|
} |
639 |
|
|
|
640 |
|
|
static void |
641 |
|
|
argv_multi(struct roff_man *mdoc, int line, |
642 |
|
|
struct mdoc_argv *v, int *pos, char *buf) |
643 |
|
|
{ |
644 |
|
|
enum margserr ac; |
645 |
|
|
char *p; |
646 |
|
|
|
647 |
|
|
for (v->sz = 0; ; v->sz++) { |
648 |
|
|
if (buf[*pos] == '-') |
649 |
|
|
break; |
650 |
|
|
ac = args(mdoc, line, pos, buf, ARGSFL_NONE, &p); |
651 |
|
|
if (ac == ARGS_EOLN) |
652 |
|
|
break; |
653 |
|
|
|
654 |
|
|
if (v->sz % MULTI_STEP == 0) |
655 |
|
|
v->value = mandoc_reallocarray(v->value, |
656 |
|
|
v->sz + MULTI_STEP, sizeof(char *)); |
657 |
|
|
|
658 |
|
|
v->value[(int)v->sz] = mandoc_strdup(p); |
659 |
|
|
} |
660 |
|
|
} |
661 |
|
|
|
662 |
|
|
static void |
663 |
|
|
argv_single(struct roff_man *mdoc, int line, |
664 |
|
|
struct mdoc_argv *v, int *pos, char *buf) |
665 |
|
|
{ |
666 |
|
|
enum margserr ac; |
667 |
|
|
char *p; |
668 |
|
|
|
669 |
|
|
ac = args(mdoc, line, pos, buf, ARGSFL_NONE, &p); |
670 |
|
|
if (ac == ARGS_EOLN) |
671 |
|
|
return; |
672 |
|
|
|
673 |
|
|
v->sz = 1; |
674 |
|
|
v->value = mandoc_malloc(sizeof(char *)); |
675 |
|
|
v->value[0] = mandoc_strdup(p); |
676 |
|
|
} |