Merge branch 'sk/gitweb-highlight-encoding' into HEAD
[git/git.git] / quote.c
1 #include "cache.h"
2 #include "quote.h"
3 #include "argv-array.h"
4
5 int quote_path_fully = 1;
6
7 static inline int need_bs_quote(char c)
8 {
9 return (c == '\'' || c == '!');
10 }
11
12 /* Help to copy the thing properly quoted for the shell safety.
13 * any single quote is replaced with '\'', any exclamation point
14 * is replaced with '\!', and the whole thing is enclosed in a
15 * single quote pair.
16 *
17 * E.g.
18 * original sq_quote result
19 * name ==> name ==> 'name'
20 * a b ==> a b ==> 'a b'
21 * a'b ==> a'\''b ==> 'a'\''b'
22 * a!b ==> a'\!'b ==> 'a'\!'b'
23 */
24 void sq_quote_buf(struct strbuf *dst, const char *src)
25 {
26 char *to_free = NULL;
27
28 if (dst->buf == src)
29 to_free = strbuf_detach(dst, NULL);
30
31 strbuf_addch(dst, '\'');
32 while (*src) {
33 size_t len = strcspn(src, "'!");
34 strbuf_add(dst, src, len);
35 src += len;
36 while (need_bs_quote(*src)) {
37 strbuf_addstr(dst, "'\\");
38 strbuf_addch(dst, *src++);
39 strbuf_addch(dst, '\'');
40 }
41 }
42 strbuf_addch(dst, '\'');
43 free(to_free);
44 }
45
46 void sq_quote_argv(struct strbuf *dst, const char** argv, size_t maxlen)
47 {
48 int i;
49
50 /* Copy into destination buffer. */
51 strbuf_grow(dst, 255);
52 for (i = 0; argv[i]; ++i) {
53 strbuf_addch(dst, ' ');
54 sq_quote_buf(dst, argv[i]);
55 if (maxlen && dst->len > maxlen)
56 die("Too many or long arguments");
57 }
58 }
59
60 static char *sq_dequote_step(char *arg, char **next)
61 {
62 char *dst = arg;
63 char *src = arg;
64 char c;
65
66 if (*src != '\'')
67 return NULL;
68 for (;;) {
69 c = *++src;
70 if (!c)
71 return NULL;
72 if (c != '\'') {
73 *dst++ = c;
74 continue;
75 }
76 /* We stepped out of sq */
77 switch (*++src) {
78 case '\0':
79 *dst = 0;
80 if (next)
81 *next = NULL;
82 return arg;
83 case '\\':
84 c = *++src;
85 if (need_bs_quote(c) && *++src == '\'') {
86 *dst++ = c;
87 continue;
88 }
89 /* Fallthrough */
90 default:
91 if (!next || !isspace(*src))
92 return NULL;
93 do {
94 c = *++src;
95 } while (isspace(c));
96 *dst = 0;
97 *next = src;
98 return arg;
99 }
100 }
101 }
102
103 char *sq_dequote(char *arg)
104 {
105 return sq_dequote_step(arg, NULL);
106 }
107
108 static int sq_dequote_to_argv_internal(char *arg,
109 const char ***argv, int *nr, int *alloc,
110 struct argv_array *array)
111 {
112 char *next = arg;
113
114 if (!*arg)
115 return 0;
116 do {
117 char *dequoted = sq_dequote_step(next, &next);
118 if (!dequoted)
119 return -1;
120 if (argv) {
121 ALLOC_GROW(*argv, *nr + 1, *alloc);
122 (*argv)[(*nr)++] = dequoted;
123 }
124 if (array)
125 argv_array_push(array, dequoted);
126 } while (next);
127
128 return 0;
129 }
130
131 int sq_dequote_to_argv(char *arg, const char ***argv, int *nr, int *alloc)
132 {
133 return sq_dequote_to_argv_internal(arg, argv, nr, alloc, NULL);
134 }
135
136 int sq_dequote_to_argv_array(char *arg, struct argv_array *array)
137 {
138 return sq_dequote_to_argv_internal(arg, NULL, NULL, NULL, array);
139 }
140
141 /* 1 means: quote as octal
142 * 0 means: quote as octal if (quote_path_fully)
143 * -1 means: never quote
144 * c: quote as "\\c"
145 */
146 #define X8(x) x, x, x, x, x, x, x, x
147 #define X16(x) X8(x), X8(x)
148 static signed char const sq_lookup[256] = {
149 /* 0 1 2 3 4 5 6 7 */
150 /* 0x00 */ 1, 1, 1, 1, 1, 1, 1, 'a',
151 /* 0x08 */ 'b', 't', 'n', 'v', 'f', 'r', 1, 1,
152 /* 0x10 */ X16(1),
153 /* 0x20 */ -1, -1, '"', -1, -1, -1, -1, -1,
154 /* 0x28 */ X16(-1), X16(-1), X16(-1),
155 /* 0x58 */ -1, -1, -1, -1,'\\', -1, -1, -1,
156 /* 0x60 */ X16(-1), X8(-1),
157 /* 0x78 */ -1, -1, -1, -1, -1, -1, -1, 1,
158 /* 0x80 */ /* set to 0 */
159 };
160
161 static inline int sq_must_quote(char c)
162 {
163 return sq_lookup[(unsigned char)c] + quote_path_fully > 0;
164 }
165
166 /* returns the longest prefix not needing a quote up to maxlen if positive.
167 This stops at the first \0 because it's marked as a character needing an
168 escape */
169 static size_t next_quote_pos(const char *s, ssize_t maxlen)
170 {
171 size_t len;
172 if (maxlen < 0) {
173 for (len = 0; !sq_must_quote(s[len]); len++);
174 } else {
175 for (len = 0; len < maxlen && !sq_must_quote(s[len]); len++);
176 }
177 return len;
178 }
179
180 /*
181 * C-style name quoting.
182 *
183 * (1) if sb and fp are both NULL, inspect the input name and counts the
184 * number of bytes that are needed to hold c_style quoted version of name,
185 * counting the double quotes around it but not terminating NUL, and
186 * returns it.
187 * However, if name does not need c_style quoting, it returns 0.
188 *
189 * (2) if sb or fp are not NULL, it emits the c_style quoted version
190 * of name, enclosed with double quotes if asked and needed only.
191 * Return value is the same as in (1).
192 */
193 static size_t quote_c_style_counted(const char *name, ssize_t maxlen,
194 struct strbuf *sb, FILE *fp, int no_dq)
195 {
196 #undef EMIT
197 #define EMIT(c) \
198 do { \
199 if (sb) strbuf_addch(sb, (c)); \
200 if (fp) fputc((c), fp); \
201 count++; \
202 } while (0)
203 #define EMITBUF(s, l) \
204 do { \
205 if (sb) strbuf_add(sb, (s), (l)); \
206 if (fp) fwrite((s), (l), 1, fp); \
207 count += (l); \
208 } while (0)
209
210 size_t len, count = 0;
211 const char *p = name;
212
213 for (;;) {
214 int ch;
215
216 len = next_quote_pos(p, maxlen);
217 if (len == maxlen || (maxlen < 0 && !p[len]))
218 break;
219
220 if (!no_dq && p == name)
221 EMIT('"');
222
223 EMITBUF(p, len);
224 EMIT('\\');
225 p += len;
226 ch = (unsigned char)*p++;
227 if (maxlen >= 0)
228 maxlen -= len + 1;
229 if (sq_lookup[ch] >= ' ') {
230 EMIT(sq_lookup[ch]);
231 } else {
232 EMIT(((ch >> 6) & 03) + '0');
233 EMIT(((ch >> 3) & 07) + '0');
234 EMIT(((ch >> 0) & 07) + '0');
235 }
236 }
237
238 EMITBUF(p, len);
239 if (p == name) /* no ending quote needed */
240 return 0;
241
242 if (!no_dq)
243 EMIT('"');
244 return count;
245 }
246
247 size_t quote_c_style(const char *name, struct strbuf *sb, FILE *fp, int nodq)
248 {
249 return quote_c_style_counted(name, -1, sb, fp, nodq);
250 }
251
252 void quote_two_c_style(struct strbuf *sb, const char *prefix, const char *path, int nodq)
253 {
254 if (quote_c_style(prefix, NULL, NULL, 0) ||
255 quote_c_style(path, NULL, NULL, 0)) {
256 if (!nodq)
257 strbuf_addch(sb, '"');
258 quote_c_style(prefix, sb, NULL, 1);
259 quote_c_style(path, sb, NULL, 1);
260 if (!nodq)
261 strbuf_addch(sb, '"');
262 } else {
263 strbuf_addstr(sb, prefix);
264 strbuf_addstr(sb, path);
265 }
266 }
267
268 void write_name_quoted(const char *name, FILE *fp, int terminator)
269 {
270 if (terminator) {
271 quote_c_style(name, NULL, fp, 0);
272 } else {
273 fputs(name, fp);
274 }
275 fputc(terminator, fp);
276 }
277
278 void write_name_quoted_relative(const char *name, const char *prefix,
279 FILE *fp, int terminator)
280 {
281 struct strbuf sb = STRBUF_INIT;
282
283 name = relative_path(name, prefix, &sb);
284 write_name_quoted(name, fp, terminator);
285
286 strbuf_release(&sb);
287 }
288
289 /* quote path as relative to the given prefix */
290 char *quote_path_relative(const char *in, const char *prefix,
291 struct strbuf *out)
292 {
293 struct strbuf sb = STRBUF_INIT;
294 const char *rel = relative_path(in, prefix, &sb);
295 strbuf_reset(out);
296 quote_c_style_counted(rel, strlen(rel), out, NULL, 0);
297 strbuf_release(&sb);
298
299 return out->buf;
300 }
301
302 /*
303 * C-style name unquoting.
304 *
305 * Quoted should point at the opening double quote.
306 * + Returns 0 if it was able to unquote the string properly, and appends the
307 * result in the strbuf `sb'.
308 * + Returns -1 in case of error, and doesn't touch the strbuf. Though note
309 * that this function will allocate memory in the strbuf, so calling
310 * strbuf_release is mandatory whichever result unquote_c_style returns.
311 *
312 * Updates endp pointer to point at one past the ending double quote if given.
313 */
314 int unquote_c_style(struct strbuf *sb, const char *quoted, const char **endp)
315 {
316 size_t oldlen = sb->len, len;
317 int ch, ac;
318
319 if (*quoted++ != '"')
320 return -1;
321
322 for (;;) {
323 len = strcspn(quoted, "\"\\");
324 strbuf_add(sb, quoted, len);
325 quoted += len;
326
327 switch (*quoted++) {
328 case '"':
329 if (endp)
330 *endp = quoted;
331 return 0;
332 case '\\':
333 break;
334 default:
335 goto error;
336 }
337
338 switch ((ch = *quoted++)) {
339 case 'a': ch = '\a'; break;
340 case 'b': ch = '\b'; break;
341 case 'f': ch = '\f'; break;
342 case 'n': ch = '\n'; break;
343 case 'r': ch = '\r'; break;
344 case 't': ch = '\t'; break;
345 case 'v': ch = '\v'; break;
346
347 case '\\': case '"':
348 break; /* verbatim */
349
350 /* octal values with first digit over 4 overflow */
351 case '0': case '1': case '2': case '3':
352 ac = ((ch - '0') << 6);
353 if ((ch = *quoted++) < '0' || '7' < ch)
354 goto error;
355 ac |= ((ch - '0') << 3);
356 if ((ch = *quoted++) < '0' || '7' < ch)
357 goto error;
358 ac |= (ch - '0');
359 ch = ac;
360 break;
361 default:
362 goto error;
363 }
364 strbuf_addch(sb, ch);
365 }
366
367 error:
368 strbuf_setlen(sb, oldlen);
369 return -1;
370 }
371
372 /* quoting as a string literal for other languages */
373
374 void perl_quote_buf(struct strbuf *sb, const char *src)
375 {
376 const char sq = '\'';
377 const char bq = '\\';
378 char c;
379
380 strbuf_addch(sb, sq);
381 while ((c = *src++)) {
382 if (c == sq || c == bq)
383 strbuf_addch(sb, bq);
384 strbuf_addch(sb, c);
385 }
386 strbuf_addch(sb, sq);
387 }
388
389 void python_quote_buf(struct strbuf *sb, const char *src)
390 {
391 const char sq = '\'';
392 const char bq = '\\';
393 const char nl = '\n';
394 char c;
395
396 strbuf_addch(sb, sq);
397 while ((c = *src++)) {
398 if (c == nl) {
399 strbuf_addch(sb, bq);
400 strbuf_addch(sb, 'n');
401 continue;
402 }
403 if (c == sq || c == bq)
404 strbuf_addch(sb, bq);
405 strbuf_addch(sb, c);
406 }
407 strbuf_addch(sb, sq);
408 }
409
410 void tcl_quote_buf(struct strbuf *sb, const char *src)
411 {
412 char c;
413
414 strbuf_addch(sb, '"');
415 while ((c = *src++)) {
416 switch (c) {
417 case '[': case ']':
418 case '{': case '}':
419 case '$': case '\\': case '"':
420 strbuf_addch(sb, '\\');
421 default:
422 strbuf_addch(sb, c);
423 break;
424 case '\f':
425 strbuf_addstr(sb, "\\f");
426 break;
427 case '\r':
428 strbuf_addstr(sb, "\\r");
429 break;
430 case '\n':
431 strbuf_addstr(sb, "\\n");
432 break;
433 case '\t':
434 strbuf_addstr(sb, "\\t");
435 break;
436 case '\v':
437 strbuf_addstr(sb, "\\v");
438 break;
439 }
440 }
441 strbuf_addch(sb, '"');
442 }