1 | /* |
---|
2 | * This file stores code common to the command line tools. |
---|
3 | * XXX: This should be converted to a proper library. |
---|
4 | * |
---|
5 | * Copyright (C) 2005-2008 Timothy D. Morgan |
---|
6 | * Copyright (C) 2002 Richard Sharpe, rsharpe@richardsharpe.com |
---|
7 | * |
---|
8 | * This program is free software; you can redistribute it and/or modify |
---|
9 | * it under the terms of the GNU General Public License as published by |
---|
10 | * the Free Software Foundation; version 3 of the License. |
---|
11 | * |
---|
12 | * This program is distributed in the hope that it will be useful, |
---|
13 | * but WITHOUT ANY WARRANTY; without even the implied warranty of |
---|
14 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
---|
15 | * GNU General Public License for more details. |
---|
16 | * |
---|
17 | * You should have received a copy of the GNU General Public License |
---|
18 | * along with this program; if not, write to the Free Software |
---|
19 | * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. |
---|
20 | * |
---|
21 | * $Id: common.c 121 2008-08-09 17:22:26Z tim $ |
---|
22 | */ |
---|
23 | |
---|
24 | #include <iconv.h> |
---|
25 | iconv_t conv_desc; |
---|
26 | |
---|
27 | const char* key_special_chars = ",\"\\/"; |
---|
28 | const char* subfield_special_chars = ",\"\\|"; |
---|
29 | const char* common_special_chars = ",\"\\"; |
---|
30 | |
---|
31 | #define REGLOOKUP_VERSION "0.9.0" |
---|
32 | |
---|
33 | |
---|
34 | void bailOut(int code, char* message) |
---|
35 | { |
---|
36 | fprintf(stderr, message); |
---|
37 | exit(code); |
---|
38 | } |
---|
39 | |
---|
40 | |
---|
41 | /* Returns a newly malloc()ed string which contains original buffer, |
---|
42 | * except for non-printable or special characters are quoted in hex |
---|
43 | * with the syntax '\xQQ' where QQ is the hex ascii value of the quoted |
---|
44 | * character. A null terminator is added, since only ascii, not binary, |
---|
45 | * is returned. |
---|
46 | */ |
---|
47 | static char* quote_buffer(const unsigned char* str, |
---|
48 | unsigned int len, const char* special) |
---|
49 | { |
---|
50 | unsigned int i, added_len; |
---|
51 | unsigned int num_written = 0; |
---|
52 | |
---|
53 | unsigned int buf_len = sizeof(char)*(len+1); |
---|
54 | char* ret_val = malloc(buf_len); |
---|
55 | char* tmp_buf; |
---|
56 | |
---|
57 | if(ret_val == NULL) |
---|
58 | return NULL; |
---|
59 | |
---|
60 | for(i=0; i<len; i++) |
---|
61 | { |
---|
62 | if(buf_len <= (num_written+5)) |
---|
63 | { |
---|
64 | /* Expand the buffer by the memory consumption rate seen so far |
---|
65 | * times the amount of input left to process. The expansion is bounded |
---|
66 | * below by a minimum safety increase, and above by the maximum possible |
---|
67 | * output string length. This should minimize both the number of |
---|
68 | * reallocs() and the amount of wasted memory. |
---|
69 | */ |
---|
70 | added_len = (len-i)*num_written/(i+1); |
---|
71 | if((buf_len+added_len) > (len*4+1)) |
---|
72 | buf_len = len*4+1; |
---|
73 | else |
---|
74 | { |
---|
75 | if (added_len < 5) |
---|
76 | buf_len += 5; |
---|
77 | else |
---|
78 | buf_len += added_len; |
---|
79 | } |
---|
80 | |
---|
81 | tmp_buf = realloc(ret_val, buf_len); |
---|
82 | if(tmp_buf == NULL) |
---|
83 | { |
---|
84 | free(ret_val); |
---|
85 | return NULL; |
---|
86 | } |
---|
87 | ret_val = tmp_buf; |
---|
88 | } |
---|
89 | |
---|
90 | if(str[i] < 32 || str[i] > 126 || strchr(special, str[i]) != NULL) |
---|
91 | { |
---|
92 | num_written += snprintf(ret_val + num_written, buf_len - num_written, |
---|
93 | "\\x%.2X", str[i]); |
---|
94 | } |
---|
95 | else |
---|
96 | ret_val[num_written++] = str[i]; |
---|
97 | } |
---|
98 | ret_val[num_written] = '\0'; |
---|
99 | |
---|
100 | return ret_val; |
---|
101 | } |
---|
102 | |
---|
103 | |
---|
104 | /* Returns a newly malloc()ed string which contains original string, |
---|
105 | * except for non-printable or special characters are quoted in hex |
---|
106 | * with the syntax '\xQQ' where QQ is the hex ascii value of the quoted |
---|
107 | * character. |
---|
108 | */ |
---|
109 | static char* quote_string(const char* str, const char* special) |
---|
110 | { |
---|
111 | unsigned int len; |
---|
112 | |
---|
113 | if(str == NULL) |
---|
114 | return NULL; |
---|
115 | |
---|
116 | len = strlen(str); |
---|
117 | return quote_buffer((const unsigned char*)str, len, special); |
---|
118 | } |
---|
119 | |
---|
120 | |
---|
121 | /* |
---|
122 | * Convert from UTF-16LE to ASCII. Accepts a Unicode buffer, uni, and |
---|
123 | * it's length, uni_max. Writes ASCII to the buffer ascii, whose size |
---|
124 | * is ascii_max. Writes at most (ascii_max-1) bytes to ascii, and null |
---|
125 | * terminates the string. Returns the length of the string stored in |
---|
126 | * ascii. On error, returns a negative errno code. |
---|
127 | */ |
---|
128 | static int uni_to_ascii(unsigned char* uni, char* ascii, |
---|
129 | unsigned int uni_max, unsigned int ascii_max) |
---|
130 | { |
---|
131 | char* inbuf = (char*)uni; |
---|
132 | char* outbuf = ascii; |
---|
133 | size_t in_len = (size_t)uni_max; |
---|
134 | size_t out_len = (size_t)(ascii_max-1); |
---|
135 | int ret; |
---|
136 | |
---|
137 | /* Set up conversion descriptor. */ |
---|
138 | conv_desc = iconv_open("US-ASCII", "UTF-16LE"); |
---|
139 | |
---|
140 | ret = iconv(conv_desc, &inbuf, &in_len, &outbuf, &out_len); |
---|
141 | if(ret == -1) |
---|
142 | { |
---|
143 | iconv_close(conv_desc); |
---|
144 | return -errno; |
---|
145 | } |
---|
146 | *outbuf = '\0'; |
---|
147 | |
---|
148 | iconv_close(conv_desc); |
---|
149 | return strlen(ascii); |
---|
150 | } |
---|
151 | |
---|
152 | |
---|
153 | /* |
---|
154 | * Convert a data value to a string for display. Returns NULL on error, |
---|
155 | * and the string to display if there is no error, or a non-fatal |
---|
156 | * error. On any error (fatal or non-fatal) occurs, (*error_msg) will |
---|
157 | * be set to a newly allocated string, containing an error message. If |
---|
158 | * a memory allocation failure occurs while generating the error |
---|
159 | * message, both the return value and (*error_msg) will be NULL. It |
---|
160 | * is the responsibility of the caller to free both a non-NULL return |
---|
161 | * value, and a non-NULL (*error_msg). |
---|
162 | */ |
---|
163 | static char* data_to_ascii(unsigned char* datap, uint32 len, uint32 type, |
---|
164 | char** error_msg) |
---|
165 | { |
---|
166 | char* asciip; |
---|
167 | char* ascii; |
---|
168 | unsigned char* cur_str; |
---|
169 | char* cur_ascii; |
---|
170 | char* cur_quoted; |
---|
171 | char* tmp_err; |
---|
172 | const char* str_type; |
---|
173 | uint32 i; |
---|
174 | uint32 cur_str_len; |
---|
175 | uint32 ascii_max, cur_str_max; |
---|
176 | uint32 str_rem, cur_str_rem, alen; |
---|
177 | int ret_err; |
---|
178 | unsigned short num_nulls; |
---|
179 | |
---|
180 | if(datap == NULL) |
---|
181 | { |
---|
182 | *error_msg = (char*)malloc(24); |
---|
183 | if(*error_msg == NULL) |
---|
184 | return NULL; |
---|
185 | strcpy(*error_msg, "Data pointer was NULL."); |
---|
186 | return NULL; |
---|
187 | } |
---|
188 | *error_msg = NULL; |
---|
189 | |
---|
190 | switch (type) |
---|
191 | { |
---|
192 | case REG_SZ: |
---|
193 | case REG_EXPAND_SZ: |
---|
194 | /* REG_LINK is a symbolic link, stored as a unicode string. */ |
---|
195 | case REG_LINK: |
---|
196 | ascii_max = sizeof(char)*(len+1); |
---|
197 | ascii = malloc(ascii_max); |
---|
198 | if(ascii == NULL) |
---|
199 | return NULL; |
---|
200 | |
---|
201 | /* Sometimes values have binary stored in them. If the unicode |
---|
202 | * conversion fails, just quote it raw. |
---|
203 | */ |
---|
204 | ret_err = uni_to_ascii(datap, ascii, len, ascii_max); |
---|
205 | if(ret_err < 0) |
---|
206 | { |
---|
207 | tmp_err = strerror(-ret_err); |
---|
208 | str_type = regfi_type_val2str(type); |
---|
209 | *error_msg = (char*)malloc(65+strlen(str_type)+strlen(tmp_err)+1); |
---|
210 | if(*error_msg == NULL) |
---|
211 | { |
---|
212 | free(ascii); |
---|
213 | return NULL; |
---|
214 | } |
---|
215 | sprintf(*error_msg, "Unicode conversion failed on %s field; " |
---|
216 | "printing as binary. Error: %s", str_type, tmp_err); |
---|
217 | |
---|
218 | cur_quoted = quote_buffer(datap, len, common_special_chars); |
---|
219 | } |
---|
220 | else |
---|
221 | cur_quoted = quote_string(ascii, common_special_chars); |
---|
222 | free(ascii); |
---|
223 | if(cur_quoted == NULL) |
---|
224 | { |
---|
225 | *error_msg = (char*)malloc(27+1); |
---|
226 | if(*error_msg != NULL) |
---|
227 | strcpy(*error_msg, "Buffer could not be quoted."); |
---|
228 | } |
---|
229 | return cur_quoted; |
---|
230 | break; |
---|
231 | |
---|
232 | case REG_DWORD: |
---|
233 | ascii_max = sizeof(char)*(8+2+1); |
---|
234 | ascii = malloc(ascii_max); |
---|
235 | if(ascii == NULL) |
---|
236 | return NULL; |
---|
237 | |
---|
238 | snprintf(ascii, ascii_max, "0x%.2X%.2X%.2X%.2X", |
---|
239 | datap[3], datap[2], datap[1], datap[0]); |
---|
240 | return ascii; |
---|
241 | break; |
---|
242 | |
---|
243 | case REG_DWORD_BE: |
---|
244 | ascii_max = sizeof(char)*(8+2+1); |
---|
245 | ascii = malloc(ascii_max); |
---|
246 | if(ascii == NULL) |
---|
247 | return NULL; |
---|
248 | |
---|
249 | snprintf(ascii, ascii_max, "0x%.2X%.2X%.2X%.2X", |
---|
250 | datap[0], datap[1], datap[2], datap[3]); |
---|
251 | return ascii; |
---|
252 | break; |
---|
253 | |
---|
254 | case REG_QWORD: |
---|
255 | ascii_max = sizeof(char)*(16+2+1); |
---|
256 | ascii = malloc(ascii_max); |
---|
257 | if(ascii == NULL) |
---|
258 | return NULL; |
---|
259 | |
---|
260 | snprintf(ascii, ascii_max, "0x%.2X%.2X%.2X%.2X%.2X%.2X%.2X%.2X", |
---|
261 | datap[7], datap[6], datap[5], datap[4], |
---|
262 | datap[3], datap[2], datap[1], datap[0]); |
---|
263 | return ascii; |
---|
264 | break; |
---|
265 | |
---|
266 | |
---|
267 | /* XXX: this MULTI_SZ parser is pretty inefficient. Should be |
---|
268 | * redone with fewer malloc calls and better string concatenation. |
---|
269 | * Also, gives lame output when "\0\0" is the string. |
---|
270 | */ |
---|
271 | case REG_MULTI_SZ: |
---|
272 | ascii_max = sizeof(char)*(len*4+1); |
---|
273 | cur_str_max = sizeof(char)*(len+1); |
---|
274 | cur_str = malloc(cur_str_max); |
---|
275 | cur_ascii = malloc(cur_str_max); |
---|
276 | ascii = malloc(ascii_max); |
---|
277 | if(ascii == NULL || cur_str == NULL || cur_ascii == NULL) |
---|
278 | return NULL; |
---|
279 | |
---|
280 | /* Reads until it reaches 4 consecutive NULLs, |
---|
281 | * which is two nulls in unicode, or until it reaches len, or until we |
---|
282 | * run out of buffer. The latter should never happen, but we shouldn't |
---|
283 | * trust our file to have the right lengths/delimiters. |
---|
284 | */ |
---|
285 | asciip = ascii; |
---|
286 | num_nulls = 0; |
---|
287 | str_rem = ascii_max; |
---|
288 | cur_str_rem = cur_str_max; |
---|
289 | cur_str_len = 0; |
---|
290 | |
---|
291 | for(i=0; (i < len) && str_rem > 0; i++) |
---|
292 | { |
---|
293 | *(cur_str+cur_str_len) = *(datap+i); |
---|
294 | if(*(cur_str+cur_str_len) == 0) |
---|
295 | num_nulls++; |
---|
296 | else |
---|
297 | num_nulls = 0; |
---|
298 | cur_str_len++; |
---|
299 | |
---|
300 | if(num_nulls == 2) |
---|
301 | { |
---|
302 | ret_err = uni_to_ascii(cur_str, cur_ascii, cur_str_len-1, cur_str_max); |
---|
303 | if(ret_err < 0) |
---|
304 | { |
---|
305 | /* XXX: should every sub-field error be enumerated? */ |
---|
306 | if(*error_msg == NULL) |
---|
307 | { |
---|
308 | tmp_err = strerror(-ret_err); |
---|
309 | *error_msg = (char*)malloc(90+strlen(tmp_err)+1); |
---|
310 | if(*error_msg == NULL) |
---|
311 | { |
---|
312 | free(cur_str); |
---|
313 | free(cur_ascii); |
---|
314 | free(ascii); |
---|
315 | return NULL; |
---|
316 | } |
---|
317 | sprintf(*error_msg, "Unicode conversion failed on at least one " |
---|
318 | "MULTI_SZ sub-field; printing as binary. Error: %s", |
---|
319 | tmp_err); |
---|
320 | } |
---|
321 | cur_quoted = quote_buffer(cur_str, cur_str_len-1, |
---|
322 | subfield_special_chars); |
---|
323 | } |
---|
324 | else |
---|
325 | cur_quoted = quote_string(cur_ascii, subfield_special_chars); |
---|
326 | |
---|
327 | alen = snprintf(asciip, str_rem, "%s", cur_quoted); |
---|
328 | asciip += alen; |
---|
329 | str_rem -= alen; |
---|
330 | free(cur_quoted); |
---|
331 | |
---|
332 | if(*(datap+i+1) == 0 && *(datap+i+2) == 0) |
---|
333 | break; |
---|
334 | else |
---|
335 | { |
---|
336 | if(str_rem > 0) |
---|
337 | { |
---|
338 | asciip[0] = '|'; |
---|
339 | asciip[1] = '\0'; |
---|
340 | asciip++; |
---|
341 | str_rem--; |
---|
342 | } |
---|
343 | memset(cur_str, 0, cur_str_max); |
---|
344 | cur_str_len = 0; |
---|
345 | num_nulls = 0; |
---|
346 | /* To eliminate leading nulls in subsequent strings. */ |
---|
347 | i++; |
---|
348 | } |
---|
349 | } |
---|
350 | } |
---|
351 | *asciip = 0; |
---|
352 | free(cur_str); |
---|
353 | free(cur_ascii); |
---|
354 | return ascii; |
---|
355 | break; |
---|
356 | |
---|
357 | /* XXX: Dont know what to do with these yet, just print as binary... */ |
---|
358 | default: |
---|
359 | /* XXX: It would be really nice if this message somehow included the |
---|
360 | * name of the current value we're having trouble with, since |
---|
361 | * stderr/stdout don't always sync nicely. |
---|
362 | */ |
---|
363 | fprintf(stderr, "WARNING: Unrecognized registry data type (0x%.8X); quoting as binary.\n", type); |
---|
364 | |
---|
365 | case REG_NONE: |
---|
366 | case REG_RESOURCE_LIST: |
---|
367 | case REG_FULL_RESOURCE_DESCRIPTOR: |
---|
368 | case REG_RESOURCE_REQUIREMENTS_LIST: |
---|
369 | |
---|
370 | case REG_BINARY: |
---|
371 | return quote_buffer(datap, len, common_special_chars); |
---|
372 | break; |
---|
373 | } |
---|
374 | |
---|
375 | return NULL; |
---|
376 | } |
---|