1 | /* quotearg.c - quote arguments for output
|
---|
2 |
|
---|
3 | Copyright (C) 1998-2002, 2004-2012 Free Software Foundation, Inc.
|
---|
4 |
|
---|
5 | This program is free software: you can redistribute it and/or modify
|
---|
6 | it under the terms of the GNU General Public License as published by
|
---|
7 | the Free Software Foundation; either version 3 of the License, or
|
---|
8 | (at your option) any later version.
|
---|
9 |
|
---|
10 | This program is distributed in the hope that it will be useful,
|
---|
11 | but WITHOUT ANY WARRANTY; without even the implied warranty of
|
---|
12 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
---|
13 | GNU General Public License for more details.
|
---|
14 |
|
---|
15 | You should have received a copy of the GNU General Public License
|
---|
16 | along with this program. If not, see <http://www.gnu.org/licenses/>. */
|
---|
17 |
|
---|
18 | /* Written by Paul Eggert <[email protected]> */
|
---|
19 |
|
---|
20 | /* Without this pragma, gcc 4.7.0 20111124 mistakenly suggests that
|
---|
21 | the quoting_options_from_style function might be candidate for
|
---|
22 | attribute 'pure' */
|
---|
23 | #if (__GNUC__ == 4 && 6 <= __GNUC_MINOR__) || 4 < __GNUC__
|
---|
24 | # pragma GCC diagnostic ignored "-Wsuggest-attribute=pure"
|
---|
25 | #endif
|
---|
26 |
|
---|
27 | #include <config.h>
|
---|
28 |
|
---|
29 | #include "quotearg.h"
|
---|
30 | #include "quote.h"
|
---|
31 |
|
---|
32 | #include "xalloc.h"
|
---|
33 | #include "c-strcaseeq.h"
|
---|
34 | #include "localcharset.h"
|
---|
35 |
|
---|
36 | #include <ctype.h>
|
---|
37 | #include <errno.h>
|
---|
38 | #include <limits.h>
|
---|
39 | #include <stdbool.h>
|
---|
40 | #include <stdlib.h>
|
---|
41 | #include <string.h>
|
---|
42 | #include <wchar.h>
|
---|
43 | #include <wctype.h>
|
---|
44 |
|
---|
45 | #include "gettext.h"
|
---|
46 | #define _(msgid) gettext (msgid)
|
---|
47 | #define N_(msgid) msgid
|
---|
48 |
|
---|
49 | #ifndef SIZE_MAX
|
---|
50 | # define SIZE_MAX ((size_t) -1)
|
---|
51 | #endif
|
---|
52 |
|
---|
53 | #define INT_BITS (sizeof (int) * CHAR_BIT)
|
---|
54 |
|
---|
55 | struct quoting_options
|
---|
56 | {
|
---|
57 | /* Basic quoting style. */
|
---|
58 | enum quoting_style style;
|
---|
59 |
|
---|
60 | /* Additional flags. Bitwise combination of enum quoting_flags. */
|
---|
61 | int flags;
|
---|
62 |
|
---|
63 | /* Quote the characters indicated by this bit vector even if the
|
---|
64 | quoting style would not normally require them to be quoted. */
|
---|
65 | unsigned int quote_these_too[(UCHAR_MAX / INT_BITS) + 1];
|
---|
66 |
|
---|
67 | /* The left quote for custom_quoting_style. */
|
---|
68 | char const *left_quote;
|
---|
69 |
|
---|
70 | /* The right quote for custom_quoting_style. */
|
---|
71 | char const *right_quote;
|
---|
72 | };
|
---|
73 |
|
---|
74 | /* Names of quoting styles. */
|
---|
75 | char const *const quoting_style_args[] =
|
---|
76 | {
|
---|
77 | "literal",
|
---|
78 | "shell",
|
---|
79 | "shell-always",
|
---|
80 | "c",
|
---|
81 | "c-maybe",
|
---|
82 | "escape",
|
---|
83 | "locale",
|
---|
84 | "clocale",
|
---|
85 | 0
|
---|
86 | };
|
---|
87 |
|
---|
88 | /* Correspondences to quoting style names. */
|
---|
89 | enum quoting_style const quoting_style_vals[] =
|
---|
90 | {
|
---|
91 | literal_quoting_style,
|
---|
92 | shell_quoting_style,
|
---|
93 | shell_always_quoting_style,
|
---|
94 | c_quoting_style,
|
---|
95 | c_maybe_quoting_style,
|
---|
96 | escape_quoting_style,
|
---|
97 | locale_quoting_style,
|
---|
98 | clocale_quoting_style
|
---|
99 | };
|
---|
100 |
|
---|
101 | /* The default quoting options. */
|
---|
102 | static struct quoting_options default_quoting_options;
|
---|
103 |
|
---|
104 | /* Allocate a new set of quoting options, with contents initially identical
|
---|
105 | to O if O is not null, or to the default if O is null.
|
---|
106 | It is the caller's responsibility to free the result. */
|
---|
107 | struct quoting_options *
|
---|
108 | clone_quoting_options (struct quoting_options *o)
|
---|
109 | {
|
---|
110 | int e = errno;
|
---|
111 | struct quoting_options *p = xmemdup (o ? o : &default_quoting_options,
|
---|
112 | sizeof *o);
|
---|
113 | errno = e;
|
---|
114 | return p;
|
---|
115 | }
|
---|
116 |
|
---|
117 | /* Get the value of O's quoting style. If O is null, use the default. */
|
---|
118 | enum quoting_style
|
---|
119 | get_quoting_style (struct quoting_options *o)
|
---|
120 | {
|
---|
121 | return (o ? o : &default_quoting_options)->style;
|
---|
122 | }
|
---|
123 |
|
---|
124 | /* In O (or in the default if O is null),
|
---|
125 | set the value of the quoting style to S. */
|
---|
126 | void
|
---|
127 | set_quoting_style (struct quoting_options *o, enum quoting_style s)
|
---|
128 | {
|
---|
129 | (o ? o : &default_quoting_options)->style = s;
|
---|
130 | }
|
---|
131 |
|
---|
132 | /* In O (or in the default if O is null),
|
---|
133 | set the value of the quoting options for character C to I.
|
---|
134 | Return the old value. Currently, the only values defined for I are
|
---|
135 | 0 (the default) and 1 (which means to quote the character even if
|
---|
136 | it would not otherwise be quoted). */
|
---|
137 | int
|
---|
138 | set_char_quoting (struct quoting_options *o, char c, int i)
|
---|
139 | {
|
---|
140 | unsigned char uc = c;
|
---|
141 | unsigned int *p =
|
---|
142 | (o ? o : &default_quoting_options)->quote_these_too + uc / INT_BITS;
|
---|
143 | int shift = uc % INT_BITS;
|
---|
144 | int r = (*p >> shift) & 1;
|
---|
145 | *p ^= ((i & 1) ^ r) << shift;
|
---|
146 | return r;
|
---|
147 | }
|
---|
148 |
|
---|
149 | /* In O (or in the default if O is null),
|
---|
150 | set the value of the quoting options flag to I, which can be a
|
---|
151 | bitwise combination of enum quoting_flags, or 0 for default
|
---|
152 | behavior. Return the old value. */
|
---|
153 | int
|
---|
154 | set_quoting_flags (struct quoting_options *o, int i)
|
---|
155 | {
|
---|
156 | int r;
|
---|
157 | if (!o)
|
---|
158 | o = &default_quoting_options;
|
---|
159 | r = o->flags;
|
---|
160 | o->flags = i;
|
---|
161 | return r;
|
---|
162 | }
|
---|
163 |
|
---|
164 | void
|
---|
165 | set_custom_quoting (struct quoting_options *o,
|
---|
166 | char const *left_quote, char const *right_quote)
|
---|
167 | {
|
---|
168 | if (!o)
|
---|
169 | o = &default_quoting_options;
|
---|
170 | o->style = custom_quoting_style;
|
---|
171 | if (!left_quote || !right_quote)
|
---|
172 | abort ();
|
---|
173 | o->left_quote = left_quote;
|
---|
174 | o->right_quote = right_quote;
|
---|
175 | }
|
---|
176 |
|
---|
177 | /* Return quoting options for STYLE, with no extra quoting. */
|
---|
178 | static struct quoting_options /* NOT PURE!! */
|
---|
179 | quoting_options_from_style (enum quoting_style style)
|
---|
180 | {
|
---|
181 | struct quoting_options o = { 0, 0, { 0 }, NULL, NULL };
|
---|
182 | if (style == custom_quoting_style)
|
---|
183 | abort ();
|
---|
184 | o.style = style;
|
---|
185 | return o;
|
---|
186 | }
|
---|
187 |
|
---|
188 | /* MSGID approximates a quotation mark. Return its translation if it
|
---|
189 | has one; otherwise, return either it or "\"", depending on S.
|
---|
190 |
|
---|
191 | S is either clocale_quoting_style or locale_quoting_style. */
|
---|
192 | static char const *
|
---|
193 | gettext_quote (char const *msgid, enum quoting_style s)
|
---|
194 | {
|
---|
195 | char const *translation = _(msgid);
|
---|
196 | char const *locale_code;
|
---|
197 |
|
---|
198 | if (translation != msgid)
|
---|
199 | return translation;
|
---|
200 |
|
---|
201 | /* For UTF-8 and GB-18030, use single quotes U+2018 and U+2019.
|
---|
202 | Here is a list of other locales that include U+2018 and U+2019:
|
---|
203 |
|
---|
204 | ISO-8859-7 0xA1 KOI8-T 0x91
|
---|
205 | CP869 0x8B CP874 0x91
|
---|
206 | CP932 0x81 0x65 CP936 0xA1 0xAE
|
---|
207 | CP949 0xA1 0xAE CP950 0xA1 0xA5
|
---|
208 | CP1250 0x91 CP1251 0x91
|
---|
209 | CP1252 0x91 CP1253 0x91
|
---|
210 | CP1254 0x91 CP1255 0x91
|
---|
211 | CP1256 0x91 CP1257 0x91
|
---|
212 | EUC-JP 0xA1 0xC6 EUC-KR 0xA1 0xAE
|
---|
213 | EUC-TW 0xA1 0xE4 BIG5 0xA1 0xA5
|
---|
214 | BIG5-HKSCS 0xA1 0xA5 EUC-CN 0xA1 0xAE
|
---|
215 | GBK 0xA1 0xAE Georgian-PS 0x91
|
---|
216 | PT154 0x91
|
---|
217 |
|
---|
218 | None of these is still in wide use; using iconv is overkill. */
|
---|
219 | locale_code = locale_charset ();
|
---|
220 | if (STRCASEEQ (locale_code, "UTF-8", 'U','T','F','-','8',0,0,0,0))
|
---|
221 | return msgid[0] == '`' ? "\xe2\x80\x98": "\xe2\x80\x99";
|
---|
222 | if (STRCASEEQ (locale_code, "GB18030", 'G','B','1','8','0','3','0',0,0))
|
---|
223 | return msgid[0] == '`' ? "\xa1\ae": "\xa1\xaf";
|
---|
224 |
|
---|
225 | return (s == clocale_quoting_style ? "\"" : "'");
|
---|
226 | }
|
---|
227 |
|
---|
228 | /* Place into buffer BUFFER (of size BUFFERSIZE) a quoted version of
|
---|
229 | argument ARG (of size ARGSIZE), using QUOTING_STYLE, FLAGS, and
|
---|
230 | QUOTE_THESE_TOO to control quoting.
|
---|
231 | Terminate the output with a null character, and return the written
|
---|
232 | size of the output, not counting the terminating null.
|
---|
233 | If BUFFERSIZE is too small to store the output string, return the
|
---|
234 | value that would have been returned had BUFFERSIZE been large enough.
|
---|
235 | If ARGSIZE is SIZE_MAX, use the string length of the argument for ARGSIZE.
|
---|
236 |
|
---|
237 | This function acts like quotearg_buffer (BUFFER, BUFFERSIZE, ARG,
|
---|
238 | ARGSIZE, O), except it breaks O into its component pieces and is
|
---|
239 | not careful about errno. */
|
---|
240 |
|
---|
241 | static size_t
|
---|
242 | quotearg_buffer_restyled (char *buffer, size_t buffersize,
|
---|
243 | char const *arg, size_t argsize,
|
---|
244 | enum quoting_style quoting_style, int flags,
|
---|
245 | unsigned int const *quote_these_too,
|
---|
246 | char const *left_quote,
|
---|
247 | char const *right_quote)
|
---|
248 | {
|
---|
249 | size_t i;
|
---|
250 | size_t len = 0;
|
---|
251 | char const *quote_string = 0;
|
---|
252 | size_t quote_string_len = 0;
|
---|
253 | bool backslash_escapes = false;
|
---|
254 | bool unibyte_locale = MB_CUR_MAX == 1;
|
---|
255 | bool elide_outer_quotes = (flags & QA_ELIDE_OUTER_QUOTES) != 0;
|
---|
256 |
|
---|
257 | #define STORE(c) \
|
---|
258 | do \
|
---|
259 | { \
|
---|
260 | if (len < buffersize) \
|
---|
261 | buffer[len] = (c); \
|
---|
262 | len++; \
|
---|
263 | } \
|
---|
264 | while (0)
|
---|
265 |
|
---|
266 | switch (quoting_style)
|
---|
267 | {
|
---|
268 | case c_maybe_quoting_style:
|
---|
269 | quoting_style = c_quoting_style;
|
---|
270 | elide_outer_quotes = true;
|
---|
271 | /* Fall through. */
|
---|
272 | case c_quoting_style:
|
---|
273 | if (!elide_outer_quotes)
|
---|
274 | STORE ('"');
|
---|
275 | backslash_escapes = true;
|
---|
276 | quote_string = "\"";
|
---|
277 | quote_string_len = 1;
|
---|
278 | break;
|
---|
279 |
|
---|
280 | case escape_quoting_style:
|
---|
281 | backslash_escapes = true;
|
---|
282 | elide_outer_quotes = false;
|
---|
283 | break;
|
---|
284 |
|
---|
285 | case locale_quoting_style:
|
---|
286 | case clocale_quoting_style:
|
---|
287 | case custom_quoting_style:
|
---|
288 | {
|
---|
289 | if (quoting_style != custom_quoting_style)
|
---|
290 | {
|
---|
291 | /* TRANSLATORS:
|
---|
292 | Get translations for open and closing quotation marks.
|
---|
293 | The message catalog should translate "`" to a left
|
---|
294 | quotation mark suitable for the locale, and similarly for
|
---|
295 | "'". For example, a French Unicode local should translate
|
---|
296 | these to U+00AB (LEFT-POINTING DOUBLE ANGLE
|
---|
297 | QUOTATION MARK), and U+00BB (RIGHT-POINTING DOUBLE ANGLE
|
---|
298 | QUOTATION MARK), respectively.
|
---|
299 |
|
---|
300 | If the catalog has no translation, we will try to
|
---|
301 | use Unicode U+2018 (LEFT SINGLE QUOTATION MARK) and
|
---|
302 | Unicode U+2019 (RIGHT SINGLE QUOTATION MARK). If the
|
---|
303 | current locale is not Unicode, locale_quoting_style
|
---|
304 | will quote 'like this', and clocale_quoting_style will
|
---|
305 | quote "like this". You should always include translations
|
---|
306 | for "`" and "'" even if U+2018 and U+2019 are appropriate
|
---|
307 | for your locale.
|
---|
308 |
|
---|
309 | If you don't know what to put here, please see
|
---|
310 | <http://en.wikipedia.org/wiki/Quotation_marks_in_other_languages>
|
---|
311 | and use glyphs suitable for your language. */
|
---|
312 | left_quote = gettext_quote (N_("`"), quoting_style);
|
---|
313 | right_quote = gettext_quote (N_("'"), quoting_style);
|
---|
314 | }
|
---|
315 | if (!elide_outer_quotes)
|
---|
316 | for (quote_string = left_quote; *quote_string; quote_string++)
|
---|
317 | STORE (*quote_string);
|
---|
318 | backslash_escapes = true;
|
---|
319 | quote_string = right_quote;
|
---|
320 | quote_string_len = strlen (quote_string);
|
---|
321 | }
|
---|
322 | break;
|
---|
323 |
|
---|
324 | case shell_quoting_style:
|
---|
325 | quoting_style = shell_always_quoting_style;
|
---|
326 | elide_outer_quotes = true;
|
---|
327 | /* Fall through. */
|
---|
328 | case shell_always_quoting_style:
|
---|
329 | if (!elide_outer_quotes)
|
---|
330 | STORE ('\'');
|
---|
331 | quote_string = "'";
|
---|
332 | quote_string_len = 1;
|
---|
333 | break;
|
---|
334 |
|
---|
335 | case literal_quoting_style:
|
---|
336 | elide_outer_quotes = false;
|
---|
337 | break;
|
---|
338 |
|
---|
339 | default:
|
---|
340 | abort ();
|
---|
341 | }
|
---|
342 |
|
---|
343 | for (i = 0; ! (argsize == SIZE_MAX ? arg[i] == '\0' : i == argsize); i++)
|
---|
344 | {
|
---|
345 | unsigned char c;
|
---|
346 | unsigned char esc;
|
---|
347 | bool is_right_quote = false;
|
---|
348 |
|
---|
349 | if (backslash_escapes
|
---|
350 | && quote_string_len
|
---|
351 | && i + quote_string_len <= argsize
|
---|
352 | && memcmp (arg + i, quote_string, quote_string_len) == 0)
|
---|
353 | {
|
---|
354 | if (elide_outer_quotes)
|
---|
355 | goto force_outer_quoting_style;
|
---|
356 | is_right_quote = true;
|
---|
357 | }
|
---|
358 |
|
---|
359 | c = arg[i];
|
---|
360 | switch (c)
|
---|
361 | {
|
---|
362 | case '\0':
|
---|
363 | if (backslash_escapes)
|
---|
364 | {
|
---|
365 | if (elide_outer_quotes)
|
---|
366 | goto force_outer_quoting_style;
|
---|
367 | STORE ('\\');
|
---|
368 | /* If quote_string were to begin with digits, we'd need to
|
---|
369 | test for the end of the arg as well. However, it's
|
---|
370 | hard to imagine any locale that would use digits in
|
---|
371 | quotes, and set_custom_quoting is documented not to
|
---|
372 | accept them. */
|
---|
373 | if (i + 1 < argsize && '0' <= arg[i + 1] && arg[i + 1] <= '9')
|
---|
374 | {
|
---|
375 | STORE ('0');
|
---|
376 | STORE ('0');
|
---|
377 | }
|
---|
378 | c = '0';
|
---|
379 | /* We don't have to worry that this last '0' will be
|
---|
380 | backslash-escaped because, again, quote_string should
|
---|
381 | not start with it and because quote_these_too is
|
---|
382 | documented as not accepting it. */
|
---|
383 | }
|
---|
384 | else if (flags & QA_ELIDE_NULL_BYTES)
|
---|
385 | continue;
|
---|
386 | break;
|
---|
387 |
|
---|
388 | case '?':
|
---|
389 | switch (quoting_style)
|
---|
390 | {
|
---|
391 | case shell_always_quoting_style:
|
---|
392 | if (elide_outer_quotes)
|
---|
393 | goto force_outer_quoting_style;
|
---|
394 | break;
|
---|
395 |
|
---|
396 | case c_quoting_style:
|
---|
397 | if ((flags & QA_SPLIT_TRIGRAPHS)
|
---|
398 | && i + 2 < argsize && arg[i + 1] == '?')
|
---|
399 | switch (arg[i + 2])
|
---|
400 | {
|
---|
401 | case '!': case '\'':
|
---|
402 | case '(': case ')': case '-': case '/':
|
---|
403 | case '<': case '=': case '>':
|
---|
404 | /* Escape the second '?' in what would otherwise be
|
---|
405 | a trigraph. */
|
---|
406 | if (elide_outer_quotes)
|
---|
407 | goto force_outer_quoting_style;
|
---|
408 | c = arg[i + 2];
|
---|
409 | i += 2;
|
---|
410 | STORE ('?');
|
---|
411 | STORE ('"');
|
---|
412 | STORE ('"');
|
---|
413 | STORE ('?');
|
---|
414 | break;
|
---|
415 |
|
---|
416 | default:
|
---|
417 | break;
|
---|
418 | }
|
---|
419 | break;
|
---|
420 |
|
---|
421 | default:
|
---|
422 | break;
|
---|
423 | }
|
---|
424 | break;
|
---|
425 |
|
---|
426 | case '\a': esc = 'a'; goto c_escape;
|
---|
427 | case '\b': esc = 'b'; goto c_escape;
|
---|
428 | case '\f': esc = 'f'; goto c_escape;
|
---|
429 | case '\n': esc = 'n'; goto c_and_shell_escape;
|
---|
430 | case '\r': esc = 'r'; goto c_and_shell_escape;
|
---|
431 | case '\t': esc = 't'; goto c_and_shell_escape;
|
---|
432 | case '\v': esc = 'v'; goto c_escape;
|
---|
433 | case '\\': esc = c;
|
---|
434 | /* No need to escape the escape if we are trying to elide
|
---|
435 | outer quotes and nothing else is problematic. */
|
---|
436 | if (backslash_escapes && elide_outer_quotes && quote_string_len)
|
---|
437 | goto store_c;
|
---|
438 |
|
---|
439 | c_and_shell_escape:
|
---|
440 | if (quoting_style == shell_always_quoting_style
|
---|
441 | && elide_outer_quotes)
|
---|
442 | goto force_outer_quoting_style;
|
---|
443 | /* Fall through. */
|
---|
444 | c_escape:
|
---|
445 | if (backslash_escapes)
|
---|
446 | {
|
---|
447 | c = esc;
|
---|
448 | goto store_escape;
|
---|
449 | }
|
---|
450 | break;
|
---|
451 |
|
---|
452 | case '{': case '}': /* sometimes special if isolated */
|
---|
453 | if (! (argsize == SIZE_MAX ? arg[1] == '\0' : argsize == 1))
|
---|
454 | break;
|
---|
455 | /* Fall through. */
|
---|
456 | case '#': case '~':
|
---|
457 | if (i != 0)
|
---|
458 | break;
|
---|
459 | /* Fall through. */
|
---|
460 | case ' ':
|
---|
461 | case '!': /* special in bash */
|
---|
462 | case '"': case '$': case '&':
|
---|
463 | case '(': case ')': case '*': case ';':
|
---|
464 | case '<':
|
---|
465 | case '=': /* sometimes special in 0th or (with "set -k") later args */
|
---|
466 | case '>': case '[':
|
---|
467 | case '^': /* special in old /bin/sh, e.g. SunOS 4.1.4 */
|
---|
468 | case '`': case '|':
|
---|
469 | /* A shell special character. In theory, '$' and '`' could
|
---|
470 | be the first bytes of multibyte characters, which means
|
---|
471 | we should check them with mbrtowc, but in practice this
|
---|
472 | doesn't happen so it's not worth worrying about. */
|
---|
473 | if (quoting_style == shell_always_quoting_style
|
---|
474 | && elide_outer_quotes)
|
---|
475 | goto force_outer_quoting_style;
|
---|
476 | break;
|
---|
477 |
|
---|
478 | case '\'':
|
---|
479 | if (quoting_style == shell_always_quoting_style)
|
---|
480 | {
|
---|
481 | if (elide_outer_quotes)
|
---|
482 | goto force_outer_quoting_style;
|
---|
483 | STORE ('\'');
|
---|
484 | STORE ('\\');
|
---|
485 | STORE ('\'');
|
---|
486 | }
|
---|
487 | break;
|
---|
488 |
|
---|
489 | case '%': case '+': case ',': case '-': case '.': case '/':
|
---|
490 | case '0': case '1': case '2': case '3': case '4': case '5':
|
---|
491 | case '6': case '7': case '8': case '9': case ':':
|
---|
492 | case 'A': case 'B': case 'C': case 'D': case 'E': case 'F':
|
---|
493 | case 'G': case 'H': case 'I': case 'J': case 'K': case 'L':
|
---|
494 | case 'M': case 'N': case 'O': case 'P': case 'Q': case 'R':
|
---|
495 | case 'S': case 'T': case 'U': case 'V': case 'W': case 'X':
|
---|
496 | case 'Y': case 'Z': case ']': case '_': case 'a': case 'b':
|
---|
497 | case 'c': case 'd': case 'e': case 'f': case 'g': case 'h':
|
---|
498 | case 'i': case 'j': case 'k': case 'l': case 'm': case 'n':
|
---|
499 | case 'o': case 'p': case 'q': case 'r': case 's': case 't':
|
---|
500 | case 'u': case 'v': case 'w': case 'x': case 'y': case 'z':
|
---|
501 | /* These characters don't cause problems, no matter what the
|
---|
502 | quoting style is. They cannot start multibyte sequences.
|
---|
503 | A digit or a special letter would cause trouble if it
|
---|
504 | appeared at the beginning of quote_string because we'd then
|
---|
505 | escape by prepending a backslash. However, it's hard to
|
---|
506 | imagine any locale that would use digits or letters as
|
---|
507 | quotes, and set_custom_quoting is documented not to accept
|
---|
508 | them. Also, a digit or a special letter would cause
|
---|
509 | trouble if it appeared in quote_these_too, but that's also
|
---|
510 | documented as not accepting them. */
|
---|
511 | break;
|
---|
512 |
|
---|
513 | default:
|
---|
514 | /* If we have a multibyte sequence, copy it until we reach
|
---|
515 | its end, find an error, or come back to the initial shift
|
---|
516 | state. For C-like styles, if the sequence has
|
---|
517 | unprintable characters, escape the whole sequence, since
|
---|
518 | we can't easily escape single characters within it. */
|
---|
519 | {
|
---|
520 | /* Length of multibyte sequence found so far. */
|
---|
521 | size_t m;
|
---|
522 |
|
---|
523 | bool printable;
|
---|
524 |
|
---|
525 | if (unibyte_locale)
|
---|
526 | {
|
---|
527 | m = 1;
|
---|
528 | printable = isprint (c) != 0;
|
---|
529 | }
|
---|
530 | else
|
---|
531 | {
|
---|
532 | mbstate_t mbstate;
|
---|
533 | memset (&mbstate, 0, sizeof mbstate);
|
---|
534 |
|
---|
535 | m = 0;
|
---|
536 | printable = true;
|
---|
537 | if (argsize == SIZE_MAX)
|
---|
538 | argsize = strlen (arg);
|
---|
539 |
|
---|
540 | do
|
---|
541 | {
|
---|
542 | wchar_t w;
|
---|
543 | size_t bytes = mbrtowc (&w, &arg[i + m],
|
---|
544 | argsize - (i + m), &mbstate);
|
---|
545 | if (bytes == 0)
|
---|
546 | break;
|
---|
547 | else if (bytes == (size_t) -1)
|
---|
548 | {
|
---|
549 | printable = false;
|
---|
550 | break;
|
---|
551 | }
|
---|
552 | else if (bytes == (size_t) -2)
|
---|
553 | {
|
---|
554 | printable = false;
|
---|
555 | while (i + m < argsize && arg[i + m])
|
---|
556 | m++;
|
---|
557 | break;
|
---|
558 | }
|
---|
559 | else
|
---|
560 | {
|
---|
561 | /* Work around a bug with older shells that "see" a '\'
|
---|
562 | that is really the 2nd byte of a multibyte character.
|
---|
563 | In practice the problem is limited to ASCII
|
---|
564 | chars >= '@' that are shell special chars. */
|
---|
565 | if ('[' == 0x5b && elide_outer_quotes
|
---|
566 | && quoting_style == shell_always_quoting_style)
|
---|
567 | {
|
---|
568 | size_t j;
|
---|
569 | for (j = 1; j < bytes; j++)
|
---|
570 | switch (arg[i + m + j])
|
---|
571 | {
|
---|
572 | case '[': case '\\': case '^':
|
---|
573 | case '`': case '|':
|
---|
574 | goto force_outer_quoting_style;
|
---|
575 |
|
---|
576 | default:
|
---|
577 | break;
|
---|
578 | }
|
---|
579 | }
|
---|
580 |
|
---|
581 | if (! iswprint (w))
|
---|
582 | printable = false;
|
---|
583 | m += bytes;
|
---|
584 | }
|
---|
585 | }
|
---|
586 | while (! mbsinit (&mbstate));
|
---|
587 | }
|
---|
588 |
|
---|
589 | if (1 < m || (backslash_escapes && ! printable))
|
---|
590 | {
|
---|
591 | /* Output a multibyte sequence, or an escaped
|
---|
592 | unprintable unibyte character. */
|
---|
593 | size_t ilim = i + m;
|
---|
594 |
|
---|
595 | for (;;)
|
---|
596 | {
|
---|
597 | if (backslash_escapes && ! printable)
|
---|
598 | {
|
---|
599 | if (elide_outer_quotes)
|
---|
600 | goto force_outer_quoting_style;
|
---|
601 | STORE ('\\');
|
---|
602 | STORE ('0' + (c >> 6));
|
---|
603 | STORE ('0' + ((c >> 3) & 7));
|
---|
604 | c = '0' + (c & 7);
|
---|
605 | }
|
---|
606 | else if (is_right_quote)
|
---|
607 | {
|
---|
608 | STORE ('\\');
|
---|
609 | is_right_quote = false;
|
---|
610 | }
|
---|
611 | if (ilim <= i + 1)
|
---|
612 | break;
|
---|
613 | STORE (c);
|
---|
614 | c = arg[++i];
|
---|
615 | }
|
---|
616 |
|
---|
617 | goto store_c;
|
---|
618 | }
|
---|
619 | }
|
---|
620 | }
|
---|
621 |
|
---|
622 | if (! ((backslash_escapes || elide_outer_quotes)
|
---|
623 | && quote_these_too
|
---|
624 | && quote_these_too[c / INT_BITS] & (1 << (c % INT_BITS)))
|
---|
625 | && !is_right_quote)
|
---|
626 | goto store_c;
|
---|
627 |
|
---|
628 | store_escape:
|
---|
629 | if (elide_outer_quotes)
|
---|
630 | goto force_outer_quoting_style;
|
---|
631 | STORE ('\\');
|
---|
632 |
|
---|
633 | store_c:
|
---|
634 | STORE (c);
|
---|
635 | }
|
---|
636 |
|
---|
637 | if (len == 0 && quoting_style == shell_always_quoting_style
|
---|
638 | && elide_outer_quotes)
|
---|
639 | goto force_outer_quoting_style;
|
---|
640 |
|
---|
641 | if (quote_string && !elide_outer_quotes)
|
---|
642 | for (; *quote_string; quote_string++)
|
---|
643 | STORE (*quote_string);
|
---|
644 |
|
---|
645 | if (len < buffersize)
|
---|
646 | buffer[len] = '\0';
|
---|
647 | return len;
|
---|
648 |
|
---|
649 | force_outer_quoting_style:
|
---|
650 | /* Don't reuse quote_these_too, since the addition of outer quotes
|
---|
651 | sufficiently quotes the specified characters. */
|
---|
652 | return quotearg_buffer_restyled (buffer, buffersize, arg, argsize,
|
---|
653 | quoting_style,
|
---|
654 | flags & ~QA_ELIDE_OUTER_QUOTES, NULL,
|
---|
655 | left_quote, right_quote);
|
---|
656 | }
|
---|
657 |
|
---|
658 | /* Place into buffer BUFFER (of size BUFFERSIZE) a quoted version of
|
---|
659 | argument ARG (of size ARGSIZE), using O to control quoting.
|
---|
660 | If O is null, use the default.
|
---|
661 | Terminate the output with a null character, and return the written
|
---|
662 | size of the output, not counting the terminating null.
|
---|
663 | If BUFFERSIZE is too small to store the output string, return the
|
---|
664 | value that would have been returned had BUFFERSIZE been large enough.
|
---|
665 | If ARGSIZE is SIZE_MAX, use the string length of the argument for
|
---|
666 | ARGSIZE. */
|
---|
667 | size_t
|
---|
668 | quotearg_buffer (char *buffer, size_t buffersize,
|
---|
669 | char const *arg, size_t argsize,
|
---|
670 | struct quoting_options const *o)
|
---|
671 | {
|
---|
672 | struct quoting_options const *p = o ? o : &default_quoting_options;
|
---|
673 | int e = errno;
|
---|
674 | size_t r = quotearg_buffer_restyled (buffer, buffersize, arg, argsize,
|
---|
675 | p->style, p->flags, p->quote_these_too,
|
---|
676 | p->left_quote, p->right_quote);
|
---|
677 | errno = e;
|
---|
678 | return r;
|
---|
679 | }
|
---|
680 |
|
---|
681 | /* Equivalent to quotearg_alloc (ARG, ARGSIZE, NULL, O). */
|
---|
682 | char *
|
---|
683 | quotearg_alloc (char const *arg, size_t argsize,
|
---|
684 | struct quoting_options const *o)
|
---|
685 | {
|
---|
686 | return quotearg_alloc_mem (arg, argsize, NULL, o);
|
---|
687 | }
|
---|
688 |
|
---|
689 | /* Like quotearg_buffer (..., ARG, ARGSIZE, O), except return newly
|
---|
690 | allocated storage containing the quoted string, and store the
|
---|
691 | resulting size into *SIZE, if non-NULL. The result can contain
|
---|
692 | embedded null bytes only if ARGSIZE is not SIZE_MAX, SIZE is not
|
---|
693 | NULL, and set_quoting_flags has not set the null byte elision
|
---|
694 | flag. */
|
---|
695 | char *
|
---|
696 | quotearg_alloc_mem (char const *arg, size_t argsize, size_t *size,
|
---|
697 | struct quoting_options const *o)
|
---|
698 | {
|
---|
699 | struct quoting_options const *p = o ? o : &default_quoting_options;
|
---|
700 | int e = errno;
|
---|
701 | /* Elide embedded null bytes if we can't return a size. */
|
---|
702 | int flags = p->flags | (size ? 0 : QA_ELIDE_NULL_BYTES);
|
---|
703 | size_t bufsize = quotearg_buffer_restyled (0, 0, arg, argsize, p->style,
|
---|
704 | flags, p->quote_these_too,
|
---|
705 | p->left_quote,
|
---|
706 | p->right_quote) + 1;
|
---|
707 | char *buf = xcharalloc (bufsize);
|
---|
708 | quotearg_buffer_restyled (buf, bufsize, arg, argsize, p->style, flags,
|
---|
709 | p->quote_these_too,
|
---|
710 | p->left_quote, p->right_quote);
|
---|
711 | errno = e;
|
---|
712 | if (size)
|
---|
713 | *size = bufsize - 1;
|
---|
714 | return buf;
|
---|
715 | }
|
---|
716 |
|
---|
717 | /* A storage slot with size and pointer to a value. */
|
---|
718 | struct slotvec
|
---|
719 | {
|
---|
720 | size_t size;
|
---|
721 | char *val;
|
---|
722 | };
|
---|
723 |
|
---|
724 | /* Preallocate a slot 0 buffer, so that the caller can always quote
|
---|
725 | one small component of a "memory exhausted" message in slot 0. */
|
---|
726 | static char slot0[256];
|
---|
727 | static unsigned int nslots = 1;
|
---|
728 | static struct slotvec slotvec0 = {sizeof slot0, slot0};
|
---|
729 | static struct slotvec *slotvec = &slotvec0;
|
---|
730 |
|
---|
731 | void
|
---|
732 | quotearg_free (void)
|
---|
733 | {
|
---|
734 | struct slotvec *sv = slotvec;
|
---|
735 | unsigned int i;
|
---|
736 | for (i = 1; i < nslots; i++)
|
---|
737 | free (sv[i].val);
|
---|
738 | if (sv[0].val != slot0)
|
---|
739 | {
|
---|
740 | free (sv[0].val);
|
---|
741 | slotvec0.size = sizeof slot0;
|
---|
742 | slotvec0.val = slot0;
|
---|
743 | }
|
---|
744 | if (sv != &slotvec0)
|
---|
745 | {
|
---|
746 | free (sv);
|
---|
747 | slotvec = &slotvec0;
|
---|
748 | }
|
---|
749 | nslots = 1;
|
---|
750 | }
|
---|
751 |
|
---|
752 | /* Use storage slot N to return a quoted version of argument ARG.
|
---|
753 | ARG is of size ARGSIZE, but if that is SIZE_MAX, ARG is a
|
---|
754 | null-terminated string.
|
---|
755 | OPTIONS specifies the quoting options.
|
---|
756 | The returned value points to static storage that can be
|
---|
757 | reused by the next call to this function with the same value of N.
|
---|
758 | N must be nonnegative. N is deliberately declared with type "int"
|
---|
759 | to allow for future extensions (using negative values). */
|
---|
760 | static char *
|
---|
761 | quotearg_n_options (int n, char const *arg, size_t argsize,
|
---|
762 | struct quoting_options const *options)
|
---|
763 | {
|
---|
764 | int e = errno;
|
---|
765 |
|
---|
766 | unsigned int n0 = n;
|
---|
767 | struct slotvec *sv = slotvec;
|
---|
768 |
|
---|
769 | if (n < 0)
|
---|
770 | abort ();
|
---|
771 |
|
---|
772 | if (nslots <= n0)
|
---|
773 | {
|
---|
774 | /* FIXME: technically, the type of n1 should be 'unsigned int',
|
---|
775 | but that evokes an unsuppressible warning from gcc-4.0.1 and
|
---|
776 | older. If gcc ever provides an option to suppress that warning,
|
---|
777 | revert to the original type, so that the test in xalloc_oversized
|
---|
778 | is once again performed only at compile time. */
|
---|
779 | size_t n1 = n0 + 1;
|
---|
780 | bool preallocated = (sv == &slotvec0);
|
---|
781 |
|
---|
782 | if (xalloc_oversized (n1, sizeof *sv))
|
---|
783 | xalloc_die ();
|
---|
784 |
|
---|
785 | slotvec = sv = xrealloc (preallocated ? NULL : sv, n1 * sizeof *sv);
|
---|
786 | if (preallocated)
|
---|
787 | *sv = slotvec0;
|
---|
788 | memset (sv + nslots, 0, (n1 - nslots) * sizeof *sv);
|
---|
789 | nslots = n1;
|
---|
790 | }
|
---|
791 |
|
---|
792 | {
|
---|
793 | size_t size = sv[n].size;
|
---|
794 | char *val = sv[n].val;
|
---|
795 | /* Elide embedded null bytes since we don't return a size. */
|
---|
796 | int flags = options->flags | QA_ELIDE_NULL_BYTES;
|
---|
797 | size_t qsize = quotearg_buffer_restyled (val, size, arg, argsize,
|
---|
798 | options->style, flags,
|
---|
799 | options->quote_these_too,
|
---|
800 | options->left_quote,
|
---|
801 | options->right_quote);
|
---|
802 |
|
---|
803 | if (size <= qsize)
|
---|
804 | {
|
---|
805 | sv[n].size = size = qsize + 1;
|
---|
806 | if (val != slot0)
|
---|
807 | free (val);
|
---|
808 | sv[n].val = val = xcharalloc (size);
|
---|
809 | quotearg_buffer_restyled (val, size, arg, argsize, options->style,
|
---|
810 | flags, options->quote_these_too,
|
---|
811 | options->left_quote,
|
---|
812 | options->right_quote);
|
---|
813 | }
|
---|
814 |
|
---|
815 | errno = e;
|
---|
816 | return val;
|
---|
817 | }
|
---|
818 | }
|
---|
819 |
|
---|
820 | char *
|
---|
821 | quotearg_n (int n, char const *arg)
|
---|
822 | {
|
---|
823 | return quotearg_n_options (n, arg, SIZE_MAX, &default_quoting_options);
|
---|
824 | }
|
---|
825 |
|
---|
826 | char *
|
---|
827 | quotearg_n_mem (int n, char const *arg, size_t argsize)
|
---|
828 | {
|
---|
829 | return quotearg_n_options (n, arg, argsize, &default_quoting_options);
|
---|
830 | }
|
---|
831 |
|
---|
832 | char *
|
---|
833 | quotearg (char const *arg)
|
---|
834 | {
|
---|
835 | return quotearg_n (0, arg);
|
---|
836 | }
|
---|
837 |
|
---|
838 | char *
|
---|
839 | quotearg_mem (char const *arg, size_t argsize)
|
---|
840 | {
|
---|
841 | return quotearg_n_mem (0, arg, argsize);
|
---|
842 | }
|
---|
843 |
|
---|
844 | char *
|
---|
845 | quotearg_n_style (int n, enum quoting_style s, char const *arg)
|
---|
846 | {
|
---|
847 | struct quoting_options const o = quoting_options_from_style (s);
|
---|
848 | return quotearg_n_options (n, arg, SIZE_MAX, &o);
|
---|
849 | }
|
---|
850 |
|
---|
851 | char *
|
---|
852 | quotearg_n_style_mem (int n, enum quoting_style s,
|
---|
853 | char const *arg, size_t argsize)
|
---|
854 | {
|
---|
855 | struct quoting_options const o = quoting_options_from_style (s);
|
---|
856 | return quotearg_n_options (n, arg, argsize, &o);
|
---|
857 | }
|
---|
858 |
|
---|
859 | char *
|
---|
860 | quotearg_style (enum quoting_style s, char const *arg)
|
---|
861 | {
|
---|
862 | return quotearg_n_style (0, s, arg);
|
---|
863 | }
|
---|
864 |
|
---|
865 | char *
|
---|
866 | quotearg_style_mem (enum quoting_style s, char const *arg, size_t argsize)
|
---|
867 | {
|
---|
868 | return quotearg_n_style_mem (0, s, arg, argsize);
|
---|
869 | }
|
---|
870 |
|
---|
871 | char *
|
---|
872 | quotearg_char_mem (char const *arg, size_t argsize, char ch)
|
---|
873 | {
|
---|
874 | struct quoting_options options;
|
---|
875 | options = default_quoting_options;
|
---|
876 | set_char_quoting (&options, ch, 1);
|
---|
877 | return quotearg_n_options (0, arg, argsize, &options);
|
---|
878 | }
|
---|
879 |
|
---|
880 | char *
|
---|
881 | quotearg_char (char const *arg, char ch)
|
---|
882 | {
|
---|
883 | return quotearg_char_mem (arg, SIZE_MAX, ch);
|
---|
884 | }
|
---|
885 |
|
---|
886 | char *
|
---|
887 | quotearg_colon (char const *arg)
|
---|
888 | {
|
---|
889 | return quotearg_char (arg, ':');
|
---|
890 | }
|
---|
891 |
|
---|
892 | char *
|
---|
893 | quotearg_colon_mem (char const *arg, size_t argsize)
|
---|
894 | {
|
---|
895 | return quotearg_char_mem (arg, argsize, ':');
|
---|
896 | }
|
---|
897 |
|
---|
898 | char *
|
---|
899 | quotearg_n_custom (int n, char const *left_quote,
|
---|
900 | char const *right_quote, char const *arg)
|
---|
901 | {
|
---|
902 | return quotearg_n_custom_mem (n, left_quote, right_quote, arg,
|
---|
903 | SIZE_MAX);
|
---|
904 | }
|
---|
905 |
|
---|
906 | char *
|
---|
907 | quotearg_n_custom_mem (int n, char const *left_quote,
|
---|
908 | char const *right_quote,
|
---|
909 | char const *arg, size_t argsize)
|
---|
910 | {
|
---|
911 | struct quoting_options o = default_quoting_options;
|
---|
912 | set_custom_quoting (&o, left_quote, right_quote);
|
---|
913 | return quotearg_n_options (n, arg, argsize, &o);
|
---|
914 | }
|
---|
915 |
|
---|
916 | char *
|
---|
917 | quotearg_custom (char const *left_quote, char const *right_quote,
|
---|
918 | char const *arg)
|
---|
919 | {
|
---|
920 | return quotearg_n_custom (0, left_quote, right_quote, arg);
|
---|
921 | }
|
---|
922 |
|
---|
923 | char *
|
---|
924 | quotearg_custom_mem (char const *left_quote, char const *right_quote,
|
---|
925 | char const *arg, size_t argsize)
|
---|
926 | {
|
---|
927 | return quotearg_n_custom_mem (0, left_quote, right_quote, arg,
|
---|
928 | argsize);
|
---|
929 | }
|
---|
930 |
|
---|
931 |
|
---|
932 | /* The quoting option used by quote_n and quote. */
|
---|
933 | struct quoting_options quote_quoting_options =
|
---|
934 | {
|
---|
935 | locale_quoting_style,
|
---|
936 | 0,
|
---|
937 | { 0 },
|
---|
938 | NULL, NULL
|
---|
939 | };
|
---|
940 |
|
---|
941 | char const *
|
---|
942 | quote_n (int n, char const *name)
|
---|
943 | {
|
---|
944 | return quotearg_n_options (n, name, SIZE_MAX, "e_quoting_options);
|
---|
945 | }
|
---|
946 |
|
---|
947 | char const *
|
---|
948 | quote (char const *name)
|
---|
949 | {
|
---|
950 | return quote_n (0, name);
|
---|
951 | }
|
---|