1 /* Miscellaneous declarations.
2 Copyright (C) 1996-2011, 2015, 2018-2021 Free Software Foundation,
3 Inc.
4
5 This file is part of GNU Wget.
6
7 GNU Wget is free software; you can redistribute it and/or modify
8 it under the terms of the GNU General Public License as published by
9 the Free Software Foundation; either version 3 of the License, or
10 (at your option) any later version.
11
12 GNU Wget 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 Wget. If not, see <http://www.gnu.org/licenses/>.
19
20 Additional permission under GNU GPL version 3 section 7
21
22 If you modify this program, or any covered work, by linking or
23 combining it with the OpenSSL project's OpenSSL library (or a
24 modified version of that library), containing parts covered by the
25 terms of the OpenSSL or SSLeay licenses, the Free Software Foundation
26 grants you additional permission to convey the resulting work.
27 Corresponding Source for a non-source form of such a combination
28 shall include the source code for the parts of OpenSSL used as well
29 as that of the covered work. */
30
31 /* This file contains declarations that are universally useful and
32 those that don't fit elsewhere. It also includes sysdep.h which
33 includes some often-needed system includes, like the obnoxious
34 <time.h> inclusion. */
35
36 #ifndef WGET_H
37 #define WGET_H
38
39 #include "config.h"
40
41 #if ((defined _WIN32 || defined __WIN32__) && !defined __CYGWIN__)
42 # define WINDOWS
43 #endif
44
45 /* Include these, so random files need not include them. */
46 #include "sysdep.h"
47
48 /* Disable assertions when debug support is not compiled in. */
49 #ifndef ENABLE_DEBUG
50 #ifndef NDEBUG
51 # define NDEBUG
52 #endif
53 #endif
54
55 /* Is OpenSSL or GNUTLS available? */
56 #if defined HAVE_LIBSSL || defined HAVE_LIBSSL32 || defined HAVE_LIBGNUTLS
57 # define HAVE_SSL
58 # define HAVE_HSTS /* There's no sense in enabling HSTS without SSL */
59 #endif
60
61 /* `gettext (FOO)' is long to write, so we use `_(FOO)'. If NLS is
62 unavailable, _(STRING) simply returns STRING. */
63 #include "gettext.h"
64 #define _(STRING) gettext(STRING)
65
66 /* A pseudo function call that serves as a marker for the automated
67 extraction of messages, but does not call gettext(). The run-time
68 translation is done at a different place in the code. The purpose
69 of the N_("...") call is to make the message snarfer aware that the
70 "..." string needs to be translated. STRING should be a string
71 literal. Concatenated strings and other string expressions won't
72 work. The macro's expansion is not parenthesized, so that it is
73 suitable as initializer for static 'char[]' or 'const char[]'
74 variables. -- explanation partly taken from GNU make. */
75 #define N_(string) string
76
77 #if HAVE_WCWIDTH && HAVE_MBTOWC
78 # define USE_NLS_PROGRESS_BAR 1
79 #else
80 /* Just to be a little paranoid about it. */
81 # undef USE_NLS_PROGRESS_BAR
82 #endif
83
84 /* I18N NOTE: You will notice that none of the DEBUGP messages are
85 marked as translatable. This is intentional, for a few reasons:
86
87 1) The debug messages are not meant for the users to look at, but
88 for the developers; as such, they should be considered more like
89 source comments than real program output.
90
91 2) The messages are numerous, and yet they are random and frivolous
92 ("double yuck!" and such). There would be a lot of work with no
93 gain.
94
95 3) Finally, the debug messages are meant to be a clue for me to
96 debug problems with Wget. If I get them in a language I don't
97 understand, debugging will become a new challenge of its own! */
98
99 /* locale independent replacement for ctype.h */
100 #include "c-ctype.h"
101
102 /* Conditionalize the use of GCC's __attribute__((format)) and
103 __builtin_expect features using macros. */
104
105 #if defined(__GNUC__) && __GNUC__ >= 3
106 # define GCC_FORMAT_ATTR(a, b) __attribute__ ((format (printf, a, b)))
107 # define LIKELY(exp) __builtin_expect (!!(exp), 1)
108 # define UNLIKELY(exp) __builtin_expect ((exp), 0)
109 #else
110 # define GCC_FORMAT_ATTR(a, b)
111 # define LIKELY(exp) (exp)
112 # define UNLIKELY(exp) (exp)
113 #endif
114
115 /* Execute the following statement if debugging is both enabled at
116 compile-time and requested at run-time; a no-op otherwise. */
117
118 #ifdef ENABLE_DEBUG
119 # define IF_DEBUG if (UNLIKELY (opt.debug))
120 #else
121 # define IF_DEBUG if (0)
122 #endif
123
124 /* Print ARGS if debugging is enabled and requested, otherwise do
125 nothing. This must be called with an extra level of parentheses
126 because it's not possible to pass a variable number of arguments to
127 a macro (in portable C89). ARGS are like arguments to printf. */
128
129 #define DEBUGP(args) do { IF_DEBUG { debug_logprintf args; } } while (0)
130
131 /* Pick an integer type large enough for file sizes, content lengths,
132 and such. Because today's files can be very large, it should be a
133 signed integer at least 64 bits wide. This can't be typedeffed to
134 off_t because: a) off_t is always 32-bit on Windows, and b) we
135 don't necessarily want to tie having a 64-bit type for internal
136 calculations to having LFS support. */
137
138 /* Gnulib's stdint.h module essentially guarantees the existence of int64_t.
139 * Thus we can simply assume it always exists and use it.
140 */
141 #include <stdint.h>
142
143 typedef int64_t wgint;
144 #define WGINT_MIN INT64_MIN
145 #define WGINT_MAX INT64_MAX
146
147 #define str_to_wgint strtoll
148
149 #include "options.h"
150
151 /* Everything uses this, so include them here directly. */
152 #ifdef __cplusplus
153 # undef _Noreturn
154 #endif
155 #include "xalloc.h"
156
157 /* Likewise for logging functions. */
158 #include "log.h"
159
160 /* Likewise for quoting functions. */
161 #include "quote.h"
162 #include "quotearg.h"
163
164 /* Likewise for struct iri definition */
165 #include "iri.h"
166
167 /* Useful macros used across the code: */
168
169 /* The number of elements in an array. For example:
170 static char a[] = "foo"; -- countof(a) == 4 (note terminating \0)
171 int a[5] = {1, 2}; -- countof(a) == 5
172 char *a[] = { -- countof(a) == 3
173 "foo", "bar", "baz"
174 }; */
175 #define countof(array) (sizeof (array) / sizeof ((array)[0]))
176
177 /* Zero out a value. */
178 #define xzero(x) memset (&(x), '\0', sizeof (x))
179
180 /* Convert an ASCII hex digit to the corresponding number between 0
181 and 15. c should be a hexadecimal digit that satisfies c_isxdigit;
182 otherwise, the result is undefined. */
_unhex(unsigned char c)183 static inline unsigned char _unhex(unsigned char c)
184 {
185 return c <= '9' ? c - '0' : (c <= 'F' ? c - 'A' + 10 : c - 'a' + 10);
186 }
187 #define X2DIGITS_TO_NUM(h1, h2) ((_unhex (h1) << 4) + _unhex (h2))
188
189 /* The reverse of the above: convert a number in the [0, 16) range to
190 the ASCII representation of the corresponding hexadecimal digit.
191 `+ 0' is there so you can't accidentally use it as an lvalue. */
192 #define XNUM_TO_DIGIT(x) ("0123456789ABCDEF"[x] + 0)
193 #define XNUM_TO_digit(x) ("0123456789abcdef"[x] + 0)
194
195 /* Return non-zero if string bounded between BEG and END is equal to
196 STRING_LITERAL. The comparison is case-sensitive. */
197 #define BOUNDED_EQUAL(beg, end, string_literal) \
198 ((end) - (beg) == sizeof (string_literal) - 1 \
199 && !memcmp (beg, string_literal, sizeof (string_literal) - 1))
200
201 /* The same as above, except the comparison is case-insensitive. */
202 #define BOUNDED_EQUAL_NO_CASE(beg, end, string_literal) \
203 ((end) - (beg) == sizeof (string_literal) - 1 \
204 && !c_strncasecmp (beg, string_literal, sizeof (string_literal) - 1))
205
206 /* Generally useful if you want to avoid arbitrary size limits but
207 don't need a full dynamic array. Assumes that BASEVAR points to a
208 malloced array of TYPE objects (or possibly a NULL pointer, if
209 SIZEVAR is 0), with the total size stored in SIZEVAR. This macro
210 will realloc BASEVAR as necessary so that it can hold at least
211 NEEDED_SIZE objects. The reallocing is done by doubling, which
212 ensures constant amortized time per element. */
213
214 #define DO_REALLOC(basevar, sizevar, needed_size, type) do { \
215 long DR_needed_size = (needed_size); \
216 long DR_newsize = 0; \
217 while ((sizevar) < (DR_needed_size)) { \
218 DR_newsize = sizevar << 1; \
219 if (DR_newsize < 16) \
220 DR_newsize = 16; \
221 (sizevar) = DR_newsize; \
222 } \
223 if (DR_newsize) \
224 basevar = xrealloc (basevar, DR_newsize * sizeof (type)); \
225 } while (0)
226
227 /* Used to print pointers (usually for debugging). Print pointers
228 using printf ("0x%0*lx", PTR_FORMAT (p)). (%p is too unpredictable;
229 some implementations prepend 0x, while some don't, and most don't
230 0-pad the address.) */
231 #define PTR_FORMAT(p) (int) (2 * sizeof (void *)), (unsigned long) (p)
232
233 /* Find the maximum buffer length needed to print an integer of type `x'
234 in base 10. 24082 / 10000 = 8*log_{10}(2). */
235 #define MAX_INT_TO_STRING_LEN(x) ((sizeof(x) * 24082 / 10000) + 2)
236
237 /* Find the minimum or maximum of two provided values */
238 # define MIN(i, j) ((i) <= (j) ? (i) : (j))
239 # define MAX(i, j) ((i) >= (j) ? (i) : (j))
240
241
242 extern const char *exec_name;
243 extern const char *program_name;
244 extern const char *program_argstring;
245
246 /* Document type ("dt") flags */
247 enum
248 {
249 TEXTHTML = 0x0001, /* document is of type text/html
250 or application/xhtml+xml */
251 RETROKF = 0x0002, /* retrieval was OK */
252 HEAD_ONLY = 0x0004, /* only send the HEAD request */
253 SEND_NOCACHE = 0x0008, /* send Cache-Control: no-cache and Pragma: no-cache directive */
254 ACCEPTRANGES = 0x0010, /* Accept-ranges header was found */
255 ADDED_HTML_EXTENSION = 0x0020, /* added ".html" extension due to -E */
256 TEXTCSS = 0x0040, /* document is of type text/css */
257 IF_MODIFIED_SINCE = 0x0080, /* use if-modified-since header */
258 METALINK_METADATA = 0x0100 /* use HTTP response for Metalink metadata */
259 };
260
261 /* Universal error type -- used almost everywhere. Error reporting of
262 this detail is not generally used or needed and should be
263 simplified. */
264 typedef enum
265 {
266 NOCONERROR, HOSTERR, CONSOCKERR, CONERROR, CONSSLERR,
267 CONIMPOSSIBLE, NEWLOCATION,
268 FTPOK, FTPLOGINC, FTPLOGREFUSED, FTPPORTERR, FTPSYSERR,
269 FTPNSFOD, FTPUNKNOWNTYPE, FTPRERR,
270 FTPSRVERR, FTPRETRINT, FTPRESTFAIL, URLERROR, FOPENERR,
271 FOPEN_EXCL_ERR, FWRITEERR, HEOF, GATEWAYTIMEOUT,
272 HERR, RETROK, RECLEVELEXC, WRONGCODE,
273 FTPINVPASV, FTPNOPASV, FTPNOPBSZ, FTPNOPROT, FTPNOAUTH,
274 CONTNOTSUPPORTED, RETRUNNEEDED, RETRFINISHED,
275 READERR, TRYLIMEXC, FILEBADFILE, RANGEERR,
276 RETRBADPATTERN, PROXERR,
277 AUTHFAILED, QUOTEXC, WRITEFAILED, SSLINITFAILED, VERIFCERTERR,
278 UNLINKERR, NEWLOCATION_KEEP_POST, CLOSEFAILED, ATTRMISSING, UNKNOWNATTR,
279 WARC_ERR, WARC_TMP_FOPENERR, WARC_TMP_FWRITEERR,
280 TIMECONV_ERR,
281 METALINK_PARSE_ERROR, METALINK_RETR_ERROR,
282 METALINK_CHKSUM_ERROR, METALINK_SIG_ERROR, METALINK_MISSING_RESOURCE,
283 RETR_WITH_METALINK,
284 METALINK_SIZE_ERROR
285 } uerr_t;
286
287 /* 2005-02-19 SMS.
288 Select an appropriate "orig" suffix and a separator character for
289 adding a unique suffix to a file name.
290
291 A VMS ODS2 file system can't tolerate multiple dots. An ODS5 file
292 system can, but even there not all dots are equal, and heroic effort
293 would be needed to get ".html^.orig" rather than (the less desirable)
294 "^.html.orig". It's more satisfactory always to use "_orig" on VMS
295 (rather than including "vms.h", testing "ods5_dest", and acting
296 accordingly).
297
298 Note that code in various places assumes that this string is five
299 characters long.
300 */
301 # ifdef __VMS
302 # define ORIG_SFX "_orig"
303 # else /* def __VMS */
304 # define ORIG_SFX ".orig"
305 # endif /* def __VMS [else] */
306
307 /* ".NNN" unique-ifying suffix separator character for unique_name() in
308 url.c (and anywhere else). Note that on VMS, the file system's
309 version numbers solve the problem that unique_name() is designed to
310 handle, obviating this whole exercise. Other systems may specify a
311 character different from "." here, if desired.
312 */
313 # ifndef __VMS
314 # define UNIQ_SEP '.'
315 # endif /* ndef __VMS */
316
317 #if defined FUZZING && defined TESTING
318 /* Rename fopen so we can have our own version in fuzz/main.c to
319 not create random files. */
320 # define fopen(fp, mode) fopen_wget(fp, mode)
321 # define exit(status) exit_wget(status)
322
323 /* In run_wgetrc() we call fopen_wgetrc() instead of fopen, so we can catch
324 the call in our fuzzers. */
325 FILE *fopen_wget(const char *pathname, const char *mode);
326 FILE *fopen_wgetrc(const char *pathname, const char *mode);
327 void exit_wget(int status);
328 #else
329 /* When not fuzzing, we want to call fopen() instead of fopen_wgetrc() */
330 # define fopen_wgetrc(fp, mode) fopen(fp, mode)
331 #endif /* FUZZING && TESTING */
332
333 #endif /* WGET_H */
334