1 /* Various utility functions.
2 Copyright (C) 1996-2006 Free Software Foundation, Inc.
4 This file is part of GNU Wget.
6 GNU Wget is free software; you can redistribute it and/or modify
7 it under the terms of the GNU General Public License as published by
8 the Free Software Foundation; either version 2 of the License, or
9 (at your option) any later version.
11 GNU Wget is distributed in the hope that it will be useful,
12 but WITHOUT ANY WARRANTY; without even the implied warranty of
13 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
14 GNU General Public License for more details.
16 You should have received a copy of the GNU General Public License
17 along with Wget; if not, write to the Free Software Foundation, Inc.,
18 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA.
20 In addition, as a special exception, the Free Software Foundation
21 gives permission to link the code of its release of Wget with the
22 OpenSSL project's "OpenSSL" library (or with modified versions of it
23 that use the same license as the "OpenSSL" library), and distribute
24 the linked executables. You must obey the GNU General Public License
25 in all respects for all of the code used other than "OpenSSL". If you
26 modify this file, you may extend this exception to your version of the
27 file, but you are not obligated to do so. If you do not wish to do
28 so, delete this exception statement from your version. */
36 #ifdef HAVE_SYS_TIME_H
37 # include <sys/time.h>
43 # include <sys/mman.h>
48 #ifdef HAVE_SYS_UTIME_H
49 # include <sys/utime.h>
57 /* For TIOCGWINSZ and friends: */
58 #ifdef HAVE_SYS_IOCTL_H
59 # include <sys/ioctl.h>
65 /* Needed for Unix version of run_with_timeout. */
69 #ifndef HAVE_SIGSETJMP
70 /* If sigsetjmp is a macro, configure won't pick it up. */
72 # define HAVE_SIGSETJMP
76 #if defined HAVE_SIGSETJMP || defined HAVE_SIGBLOCK
77 # define USE_SIGNAL_TIMEOUT
88 /* Utility function: like xstrdup(), but also lowercases S. */
91 xstrdup_lower (const char *s)
93 char *copy = xstrdup (s);
100 /* Copy the string formed by two pointers (one on the beginning, other
101 on the char after the last char) to a new, malloc-ed location.
104 strdupdelim (const char *beg, const char *end)
106 char *res = xmalloc (end - beg + 1);
107 memcpy (res, beg, end - beg);
108 res[end - beg] = '\0';
112 /* Parse a string containing comma-separated elements, and return a
113 vector of char pointers with the elements. Spaces following the
114 commas are ignored. */
116 sepstring (const char *s)
130 res = xrealloc (res, (i + 2) * sizeof (char *));
131 res[i] = strdupdelim (p, s);
134 /* Skip the blanks following the ','. */
142 res = xrealloc (res, (i + 2) * sizeof (char *));
143 res[i] = strdupdelim (p, s);
148 /* Like sprintf, but prints into a string of sufficient size freshly
149 allocated with malloc, which is returned. If unable to print due
150 to invalid format, returns NULL. Inability to allocate needed
151 memory results in abort, as with xmalloc. This is in spirit
152 similar to the GNU/BSD extension asprintf, but somewhat easier to
155 Internally the function either calls vasprintf or loops around
156 vsnprintf until the correct size is found. Since Wget also ships a
157 fallback implementation of vsnprintf, this should be portable. */
160 aprintf (const char *fmt, ...)
162 #ifdef HAVE_VASPRINTF
167 va_start (args, fmt);
168 ret = vasprintf (&str, fmt, args);
170 if (ret < 0 && errno == ENOMEM)
171 abort (); /* for consistency with xmalloc/xrealloc */
175 #else /* not HAVE_VASPRINTF */
177 /* vasprintf is unavailable. snprintf into a small buffer and
178 resize it as necessary. */
180 char *str = xmalloc (size);
182 /* #### This code will infloop and eventually abort in xrealloc if
183 passed a FMT that causes snprintf to consistently return -1. */
190 va_start (args, fmt);
191 n = vsnprintf (str, size, fmt, args);
194 /* If the printing worked, return the string. */
195 if (n > -1 && n < size)
198 /* Else try again with a larger buffer. */
199 if (n > -1) /* C99 */
200 size = n + 1; /* precisely what is needed */
202 size <<= 1; /* twice the old size */
203 str = xrealloc (str, size);
205 #endif /* not HAVE_VASPRINTF */
208 /* Concatenate the NULL-terminated list of string arguments into
209 freshly allocated space. */
212 concat_strings (const char *str0, ...)
215 int saved_lengths[5]; /* inspired by Apache's apr_pstrcat */
218 const char *next_str;
219 int total_length = 0;
222 /* Calculate the length of and allocate the resulting string. */
225 va_start (args, str0);
226 for (next_str = str0; next_str != NULL; next_str = va_arg (args, char *))
228 int len = strlen (next_str);
229 if (argcount < countof (saved_lengths))
230 saved_lengths[argcount++] = len;
234 p = ret = xmalloc (total_length + 1);
236 /* Copy the strings into the allocated space. */
239 va_start (args, str0);
240 for (next_str = str0; next_str != NULL; next_str = va_arg (args, char *))
243 if (argcount < countof (saved_lengths))
244 len = saved_lengths[argcount++];
246 len = strlen (next_str);
247 memcpy (p, next_str, len);
256 /* Format the provided time according to the specified format. The
257 format is a string with format elements supported by strftime. */
260 fmttime (time_t t, const char *fmt)
262 static char output[32];
263 struct tm *tm = localtime(&t);
266 if (!strftime(output, sizeof(output), fmt, tm))
271 /* Return pointer to a static char[] buffer in which zero-terminated
272 string-representation of TM (in form hh:mm:ss) is printed.
274 If TM is NULL, the current time will be used. */
279 return fmttime(t, "%H:%M:%S");
282 /* Like the above, but include the date: YYYY-MM-DD hh:mm:ss. */
285 datetime_str (time_t t)
287 return fmttime(t, "%Y-%m-%d %H:%M:%S");
290 /* The Windows versions of the following two functions are defined in
295 fork_to_background (void)
298 /* Whether we arrange our own version of opt.lfilename here. */
299 bool logfile_changed = false;
303 /* We must create the file immediately to avoid either a race
304 condition (which arises from using unique_name and failing to
305 use fopen_excl) or lying to the user about the log file name
306 (which arises from using unique_name, printing the name, and
307 using fopen_excl later on.) */
308 FILE *new_log_fp = unique_create (DEFAULT_LOGFILE, false, &opt.lfilename);
311 logfile_changed = true;
324 /* parent, no error */
325 printf (_("Continuing in background, pid %d.\n"), (int) pid);
327 printf (_("Output will be written to `%s'.\n"), opt.lfilename);
328 exit (0); /* #### should we use _exit()? */
331 /* child: give up the privileges and keep running. */
333 freopen ("/dev/null", "r", stdin);
334 freopen ("/dev/null", "w", stdout);
335 freopen ("/dev/null", "w", stderr);
337 #endif /* not WINDOWS */
339 /* "Touch" FILE, i.e. make its mtime ("modified time") equal the time
340 specified with TM. The atime ("access time") is set to the current
344 touch (const char *file, time_t tm)
346 #ifdef HAVE_STRUCT_UTIMBUF
347 struct utimbuf times;
355 times.actime = time (NULL);
356 if (utime (file, ×) == -1)
357 logprintf (LOG_NOTQUIET, "utime(%s): %s\n", file, strerror (errno));
360 /* Checks if FILE is a symbolic link, and removes it if it is. Does
361 nothing under MS-Windows. */
363 remove_link (const char *file)
368 if (lstat (file, &st) == 0 && S_ISLNK (st.st_mode))
370 DEBUGP (("Unlinking %s (symlink).\n", file));
373 logprintf (LOG_VERBOSE, _("Failed to unlink symlink `%s': %s\n"),
374 file, strerror (errno));
379 /* Does FILENAME exist? This is quite a lousy implementation, since
380 it supplies no error codes -- only a yes-or-no answer. Thus it
381 will return that a file does not exist if, e.g., the directory is
382 unreadable. I don't mind it too much currently, though. The
383 proper way should, of course, be to have a third, error state,
384 other than true/false, but that would introduce uncalled-for
385 additional complexity to the callers. */
387 file_exists_p (const char *filename)
390 return access (filename, F_OK) >= 0;
393 return stat (filename, &buf) >= 0;
397 /* Returns 0 if PATH is a directory, 1 otherwise (any kind of file).
398 Returns 0 on error. */
400 file_non_directory_p (const char *path)
403 /* Use lstat() rather than stat() so that symbolic links pointing to
404 directories can be identified correctly. */
405 if (lstat (path, &buf) != 0)
407 return S_ISDIR (buf.st_mode) ? false : true;
410 /* Return the size of file named by FILENAME, or -1 if it cannot be
411 opened or seeked into. */
413 file_size (const char *filename)
415 #if defined(HAVE_FSEEKO) && defined(HAVE_FTELLO)
417 /* We use fseek rather than stat to determine the file size because
418 that way we can also verify that the file is readable without
419 explicitly checking for permissions. Inspired by the POST patch
421 FILE *fp = fopen (filename, "rb");
424 fseeko (fp, 0, SEEK_END);
430 if (stat (filename, &st) < 0)
436 /* stat file names named PREFIX.1, PREFIX.2, etc., until one that
437 doesn't exist is found. Return a freshly allocated copy of the
441 unique_name_1 (const char *prefix)
444 int plen = strlen (prefix);
445 char *template = (char *)alloca (plen + 1 + 24);
446 char *template_tail = template + plen;
448 memcpy (template, prefix, plen);
449 *template_tail++ = '.';
452 number_to_string (template_tail, count++);
453 while (file_exists_p (template));
455 return xstrdup (template);
458 /* Return a unique file name, based on FILE.
460 More precisely, if FILE doesn't exist, it is returned unmodified.
461 If not, FILE.1 is tried, then FILE.2, etc. The first FILE.<number>
462 file name that doesn't exist is returned.
464 The resulting file is not created, only verified that it didn't
465 exist at the point in time when the function was called.
466 Therefore, where security matters, don't rely that the file created
467 by this function exists until you open it with O_EXCL or
470 If ALLOW_PASSTHROUGH is 0, it always returns a freshly allocated
471 string. Otherwise, it may return FILE if the file doesn't exist
472 (and therefore doesn't need changing). */
475 unique_name (const char *file, bool allow_passthrough)
477 /* If the FILE itself doesn't exist, return it without
479 if (!file_exists_p (file))
480 return allow_passthrough ? (char *)file : xstrdup (file);
482 /* Otherwise, find a numeric suffix that results in unused file name
484 return unique_name_1 (file);
487 /* Create a file based on NAME, except without overwriting an existing
488 file with that name. Providing O_EXCL is correctly implemented,
489 this function does not have the race condition associated with
490 opening the file returned by unique_name. */
493 unique_create (const char *name, bool binary, char **opened_name)
495 /* unique file name, based on NAME */
496 char *uname = unique_name (name, false);
498 while ((fp = fopen_excl (uname, binary)) == NULL && errno == EEXIST)
501 uname = unique_name (name, false);
503 if (opened_name && fp != NULL)
506 *opened_name = uname;
518 /* Open the file for writing, with the addition that the file is
519 opened "exclusively". This means that, if the file already exists,
520 this function will *fail* and errno will be set to EEXIST. If
521 BINARY is set, the file will be opened in binary mode, equivalent
524 If opening the file fails for any reason, including the file having
525 previously existed, this function returns NULL and sets errno
529 fopen_excl (const char *fname, bool binary)
533 int flags = O_WRONLY | O_CREAT | O_EXCL;
538 fd = open (fname, flags, 0666);
541 return fdopen (fd, binary ? "wb" : "w");
542 #else /* not O_EXCL */
543 /* Manually check whether the file exists. This is prone to race
544 conditions, but systems without O_EXCL haven't deserved
546 if (file_exists_p (fname))
551 return fopen (fname, binary ? "wb" : "w");
552 #endif /* not O_EXCL */
555 /* Create DIRECTORY. If some of the pathname components of DIRECTORY
556 are missing, create them first. In case any mkdir() call fails,
557 return its error status. Returns 0 on successful completion.
559 The behaviour of this function should be identical to the behaviour
560 of `mkdir -p' on systems where mkdir supports the `-p' option. */
562 make_directory (const char *directory)
564 int i, ret, quit = 0;
567 /* Make a copy of dir, to be able to write to it. Otherwise, the
568 function is unsafe if called with a read-only char *argument. */
569 STRDUP_ALLOCA (dir, directory);
571 /* If the first character of dir is '/', skip it (and thus enable
572 creation of absolute-pathname directories. */
573 for (i = (*dir == '/'); 1; ++i)
575 for (; dir[i] && dir[i] != '/'; i++)
580 /* Check whether the directory already exists. Allow creation of
581 of intermediate directories to fail, as the initial path components
582 are not necessarily directories! */
583 if (!file_exists_p (dir))
584 ret = mkdir (dir, 0777);
595 /* Merge BASE with FILE. BASE can be a directory or a file name, FILE
596 should be a file name.
598 file_merge("/foo/bar", "baz") => "/foo/baz"
599 file_merge("/foo/bar/", "baz") => "/foo/bar/baz"
600 file_merge("foo", "bar") => "bar"
602 In other words, it's a simpler and gentler version of uri_merge. */
605 file_merge (const char *base, const char *file)
608 const char *cut = (const char *)strrchr (base, '/');
611 return xstrdup (file);
613 result = xmalloc (cut - base + 1 + strlen (file) + 1);
614 memcpy (result, base, cut - base);
615 result[cut - base] = '/';
616 strcpy (result + (cut - base) + 1, file);
621 /* Like fnmatch, but performs a case-insensitive match. */
624 fnmatch_nocase (const char *pattern, const char *string, int flags)
627 /* The FNM_CASEFOLD flag started as a GNU extension, but it is now
628 also present on *BSD platforms, and possibly elsewhere. */
629 return fnmatch (pattern, string, flags | FNM_CASEFOLD);
631 /* Turn PATTERN and STRING to lower case and call fnmatch on them. */
632 char *patcopy = (char *) alloca (strlen (pattern) + 1);
633 char *strcopy = (char *) alloca (strlen (string) + 1);
635 for (p = patcopy; *pattern; pattern++, p++)
636 *p = TOLOWER (*pattern);
638 for (p = strcopy; *string; string++, p++)
639 *p = TOLOWER (*string);
641 return fnmatch (patcopy, strcopy, flags);
645 static bool in_acclist (const char *const *, const char *, bool);
647 /* Determine whether a file is acceptable to be followed, according to
648 lists of patterns to accept/reject. */
650 acceptable (const char *s)
654 while (l && s[l] != '/')
661 return (in_acclist ((const char *const *)opt.accepts, s, true)
662 && !in_acclist ((const char *const *)opt.rejects, s, true));
664 return in_acclist ((const char *const *)opt.accepts, s, true);
666 else if (opt.rejects)
667 return !in_acclist ((const char *const *)opt.rejects, s, true);
671 /* Check if D2 is a subdirectory of D1. E.g. if D1 is `/something', subdir_p()
672 will return true if and only if D2 begins with `/something/' or is exactly
675 subdir_p (const char *d1, const char *d2)
677 if (!opt.ignore_case)
678 for (; *d1 && *d2 && (*d1 == *d2); ++d1, ++d2)
681 for (; *d1 && *d2 && (TOLOWER (*d1) == TOLOWER (*d2)); ++d1, ++d2)
684 return *d1 == '\0' && (*d2 == '\0' || *d2 == '/');
687 /* Iterate through DIRLIST (which must be NULL-terminated), and return the
688 first element that matches DIR, through wildcards or front comparison (as
691 dir_matches_p (char **dirlist, const char *dir)
694 int (*matcher) (const char *, const char *, int)
695 = opt.ignore_case ? fnmatch_nocase : fnmatch;
697 for (x = dirlist; *x; x++)
699 /* Remove leading '/' */
700 char *p = *x + (**x == '/');
701 if (has_wildcards_p (p))
703 if (matcher (p, dir, FNM_PATHNAME) == 0)
708 if (subdir_p (p, dir))
713 return *x ? true : false;
716 /* Returns whether DIRECTORY is acceptable for download, wrt the
717 include/exclude lists.
719 The leading `/' is ignored in paths; relative and absolute paths
720 may be freely intermixed. */
723 accdir (const char *directory)
725 /* Remove starting '/'. */
726 if (*directory == '/')
730 if (!dir_matches_p (opt.includes, directory))
735 if (dir_matches_p (opt.excludes, directory))
741 /* Return true if STRING ends with TAIL. For instance:
743 match_tail ("abc", "bc", false) -> 1
744 match_tail ("abc", "ab", false) -> 0
745 match_tail ("abc", "abc", false) -> 1
747 If FOLD_CASE is true, the comparison will be case-insensitive. */
750 match_tail (const char *string, const char *tail, bool fold_case)
754 /* We want this to be fast, so we code two loops, one with
755 case-folding, one without. */
759 for (i = strlen (string), j = strlen (tail); i >= 0 && j >= 0; i--, j--)
760 if (string[i] != tail[j])
765 for (i = strlen (string), j = strlen (tail); i >= 0 && j >= 0; i--, j--)
766 if (TOLOWER (string[i]) != TOLOWER (tail[j]))
770 /* If the tail was exhausted, the match was succesful. */
777 /* Checks whether string S matches each element of ACCEPTS. A list
778 element are matched either with fnmatch() or match_tail(),
779 according to whether the element contains wildcards or not.
781 If the BACKWARD is false, don't do backward comparison -- just compare
784 in_acclist (const char *const *accepts, const char *s, bool backward)
786 for (; *accepts; accepts++)
788 if (has_wildcards_p (*accepts))
790 int res = opt.ignore_case
791 ? fnmatch_nocase (*accepts, s, 0) : fnmatch (*accepts, s, 0);
792 /* fnmatch returns 0 if the pattern *does* match the string. */
800 if (match_tail (s, *accepts, opt.ignore_case))
805 int cmp = opt.ignore_case
806 ? strcasecmp (s, *accepts) : strcmp (s, *accepts);
815 /* Return the location of STR's suffix (file extension). Examples:
816 suffix ("foo.bar") -> "bar"
817 suffix ("foo.bar.baz") -> "baz"
818 suffix ("/foo/bar") -> NULL
819 suffix ("/foo.bar/baz") -> NULL */
821 suffix (const char *str)
825 for (i = strlen (str); i && str[i] != '/' && str[i] != '.'; i--)
829 return (char *)str + i;
834 /* Return true if S contains globbing wildcards (`*', `?', `[' or
838 has_wildcards_p (const char *s)
841 if (*s == '*' || *s == '?' || *s == '[' || *s == ']')
846 /* Return true if FNAME ends with a typical HTML suffix. The
847 following (case-insensitive) suffixes are presumed to be HTML
852 ?html (`?' matches one character)
854 #### CAVEAT. This is not necessarily a good indication that FNAME
855 refers to a file that contains HTML! */
857 has_html_suffix_p (const char *fname)
861 if ((suf = suffix (fname)) == NULL)
863 if (!strcasecmp (suf, "html"))
865 if (!strcasecmp (suf, "htm"))
867 if (suf[0] && !strcasecmp (suf + 1, "html"))
872 /* Read a line from FP and return the pointer to freshly allocated
873 storage. The storage space is obtained through malloc() and should
874 be freed with free() when it is no longer needed.
876 The length of the line is not limited, except by available memory.
877 The newline character at the end of line is retained. The line is
878 terminated with a zero character.
880 After end-of-file is encountered without anything being read, NULL
881 is returned. NULL is also returned on error. To distinguish
882 between these two cases, use the stdio function ferror(). */
885 read_whole_line (FILE *fp)
889 char *line = xmalloc (bufsize);
891 while (fgets (line + length, bufsize - length, fp))
893 length += strlen (line + length);
895 /* Possible for example when reading from a binary file where
896 a line begins with \0. */
899 if (line[length - 1] == '\n')
902 /* fgets() guarantees to read the whole line, or to use up the
903 space we've given it. We can double the buffer
906 line = xrealloc (line, bufsize);
908 if (length == 0 || ferror (fp))
913 if (length + 1 < bufsize)
914 /* Relieve the memory from our exponential greediness. We say
915 `length + 1' because the terminating \0 is not included in
916 LENGTH. We don't need to zero-terminate the string ourselves,
917 though, because fgets() does that. */
918 line = xrealloc (line, length + 1);
922 /* Read FILE into memory. A pointer to `struct file_memory' are
923 returned; use struct element `content' to access file contents, and
924 the element `length' to know the file length. `content' is *not*
925 zero-terminated, and you should *not* read or write beyond the [0,
926 length) range of characters.
928 After you are done with the file contents, call read_file_free to
931 Depending on the operating system and the type of file that is
932 being read, read_file() either mmap's the file into memory, or
933 reads the file into the core using read().
935 If file is named "-", fileno(stdin) is used for reading instead.
936 If you want to read from a real file named "-", use "./-" instead. */
939 read_file (const char *file)
942 struct file_memory *fm;
944 bool inhibit_close = false;
946 /* Some magic in the finest tradition of Perl and its kin: if FILE
947 is "-", just use stdin. */
951 inhibit_close = true;
952 /* Note that we don't inhibit mmap() in this case. If stdin is
953 redirected from a regular file, mmap() will still work. */
956 fd = open (file, O_RDONLY);
959 fm = xnew (struct file_memory);
964 if (fstat (fd, &buf) < 0)
966 fm->length = buf.st_size;
967 /* NOTE: As far as I know, the callers of this function never
968 modify the file text. Relying on this would enable us to
969 specify PROT_READ and MAP_SHARED for a marginal gain in
970 efficiency, but at some cost to generality. */
971 fm->content = mmap (NULL, fm->length, PROT_READ | PROT_WRITE,
973 if (fm->content == (char *)MAP_FAILED)
983 /* The most common reason why mmap() fails is that FD does not point
984 to a plain file. However, it's also possible that mmap() doesn't
985 work for a particular type of file. Therefore, whenever mmap()
986 fails, we just fall back to the regular method. */
987 #endif /* HAVE_MMAP */
990 size = 512; /* number of bytes fm->contents can
991 hold at any given time. */
992 fm->content = xmalloc (size);
996 if (fm->length > size / 2)
998 /* #### I'm not sure whether the whole exponential-growth
999 thing makes sense with kernel read. On Linux at least,
1000 read() refuses to read more than 4K from a file at a
1001 single chunk anyway. But other Unixes might optimize it
1002 better, and it doesn't *hurt* anything, so I'm leaving
1005 /* Normally, we grow SIZE exponentially to make the number
1006 of calls to read() and realloc() logarithmic in relation
1007 to file size. However, read() can read an amount of data
1008 smaller than requested, and it would be unreasonable to
1009 double SIZE every time *something* was read. Therefore,
1010 we double SIZE only when the length exceeds half of the
1011 entire allocated size. */
1013 fm->content = xrealloc (fm->content, size);
1015 nread = read (fd, fm->content + fm->length, size - fm->length);
1017 /* Successful read. */
1018 fm->length += nread;
1028 if (size > fm->length && fm->length != 0)
1029 /* Due to exponential growth of fm->content, the allocated region
1030 might be much larger than what is actually needed. */
1031 fm->content = xrealloc (fm->content, fm->length);
1038 xfree (fm->content);
1043 /* Release the resources held by FM. Specifically, this calls
1044 munmap() or xfree() on fm->content, depending whether mmap or
1045 malloc/read were used to read in the file. It also frees the
1046 memory needed to hold the FM structure itself. */
1049 read_file_free (struct file_memory *fm)
1054 munmap (fm->content, fm->length);
1059 xfree (fm->content);
1064 /* Free the pointers in a NULL-terminated vector of pointers, then
1065 free the pointer itself. */
1067 free_vec (char **vec)
1078 /* Append vector V2 to vector V1. The function frees V2 and
1079 reallocates V1 (thus you may not use the contents of neither
1080 pointer after the call). If V1 is NULL, V2 is returned. */
1082 merge_vecs (char **v1, char **v2)
1092 /* To avoid j == 0 */
1097 for (i = 0; v1[i]; i++)
1100 for (j = 0; v2[j]; j++)
1102 /* Reallocate v1. */
1103 v1 = xrealloc (v1, (i + j + 1) * sizeof (char **));
1104 memcpy (v1 + i, v2, (j + 1) * sizeof (char *));
1109 /* Append a freshly allocated copy of STR to VEC. If VEC is NULL, it
1110 is allocated as needed. Return the new value of the vector. */
1113 vec_append (char **vec, const char *str)
1115 int cnt; /* count of vector elements, including
1116 the one we're about to append */
1119 for (cnt = 0; vec[cnt]; cnt++)
1125 /* Reallocate the array to fit the new element and the NULL. */
1126 vec = xrealloc (vec, (cnt + 1) * sizeof (char *));
1127 /* Append a copy of STR to the vector. */
1128 vec[cnt - 1] = xstrdup (str);
1133 /* Sometimes it's useful to create "sets" of strings, i.e. special
1134 hash tables where you want to store strings as keys and merely
1135 query for their existence. Here is a set of utility routines that
1136 makes that transparent. */
1139 string_set_add (struct hash_table *ht, const char *s)
1141 /* First check whether the set element already exists. If it does,
1142 do nothing so that we don't have to free() the old element and
1143 then strdup() a new one. */
1144 if (hash_table_contains (ht, s))
1147 /* We use "1" as value. It provides us a useful and clear arbitrary
1148 value, and it consumes no memory -- the pointers to the same
1149 string "1" will be shared by all the key-value pairs in all `set'
1151 hash_table_put (ht, xstrdup (s), "1");
1154 /* Synonym for hash_table_contains... */
1157 string_set_contains (struct hash_table *ht, const char *s)
1159 return hash_table_contains (ht, s);
1162 /* Convert the specified string set to array. ARRAY should be large
1163 enough to hold hash_table_count(ht) char pointers. */
1165 void string_set_to_array (struct hash_table *ht, char **array)
1167 hash_table_iterator iter;
1168 for (hash_table_iterate (ht, &iter); hash_table_iter_next (&iter); )
1169 *array++ = iter.key;
1172 /* Free the string set. This frees both the storage allocated for
1173 keys and the actual hash table. (hash_table_destroy would only
1174 destroy the hash table.) */
1177 string_set_free (struct hash_table *ht)
1179 hash_table_iterator iter;
1180 for (hash_table_iterate (ht, &iter); hash_table_iter_next (&iter); )
1182 hash_table_destroy (ht);
1185 /* Utility function: simply call xfree() on all keys and values of HT. */
1188 free_keys_and_values (struct hash_table *ht)
1190 hash_table_iterator iter;
1191 for (hash_table_iterate (ht, &iter); hash_table_iter_next (&iter); )
1198 /* Get digit grouping data for thousand separors by calling
1199 localeconv(). The data includes separator string and grouping info
1200 and is cached after the first call to the function.
1202 In locales that don't set a thousand separator (such as the "C"
1203 locale), this forces it to be ",". We are now only showing
1204 thousand separators in one place, so this shouldn't be a problem in
1208 get_grouping_data (const char **sep, const char **grouping)
1210 static const char *cached_sep;
1211 static const char *cached_grouping;
1212 static bool initialized;
1215 /* Get the grouping info from the locale. */
1216 struct lconv *lconv = localeconv ();
1217 cached_sep = lconv->thousands_sep;
1218 cached_grouping = lconv->grouping;
1221 /* Many locales (such as "C" or "hr_HR") don't specify
1222 grouping, which we still want to use it for legibility.
1223 In those locales set the sep char to ',', unless that
1224 character is used for decimal point, in which case set it
1226 if (*lconv->decimal_point != ',')
1230 cached_grouping = "\x03";
1235 *grouping = cached_grouping;
1238 /* Return a printed representation of N with thousand separators.
1239 This should respect locale settings, with the exception of the "C"
1240 locale which mandates no separator, but we use one anyway.
1242 Unfortunately, we cannot use %'d (in fact it would be %'j) to get
1243 the separators because it's too non-portable, and it's hard to test
1244 for this feature at configure time. Besides, it wouldn't display
1245 separators in the "C" locale, still used by many Unix users. */
1248 with_thousand_seps (wgint n)
1250 static char outbuf[48];
1251 char *p = outbuf + sizeof outbuf;
1253 /* Info received from locale */
1254 const char *grouping, *sep;
1257 /* State information */
1258 int i = 0, groupsize;
1259 const char *atgroup;
1261 bool negative = n < 0;
1263 /* Initialize grouping data. */
1264 get_grouping_data (&sep, &grouping);
1265 seplen = strlen (sep);
1267 groupsize = *atgroup++;
1269 /* This would overflow on WGINT_MIN, but printing negative numbers
1270 is not an important goal of this fuinction. */
1274 /* Write the number into the buffer, backwards, inserting the
1275 separators as necessary. */
1279 *--p = n % 10 + '0';
1283 /* Prepend SEP to every groupsize'd digit and get new groupsize. */
1284 if (++i == groupsize)
1289 memcpy (p -= seplen, sep, seplen);
1292 groupsize = *atgroup++;
1301 /* N, a byte quantity, is converted to a human-readable abberviated
1302 form a la sizes printed by `ls -lh'. The result is written to a
1303 static buffer, a pointer to which is returned.
1305 Unlike `with_thousand_seps', this approximates to the nearest unit.
1306 Quoting GNU libit: "Most people visually process strings of 3-4
1307 digits effectively, but longer strings of digits are more prone to
1308 misinterpretation. Hence, converting to an abbreviated form
1309 usually improves readability."
1311 This intentionally uses kilobyte (KB), megabyte (MB), etc. in their
1312 original computer-related meaning of "powers of 1024". We don't
1313 use the "*bibyte" names invented in 1998, and seldom used in
1314 practice. Wikipedia's entry on "binary prefix" discusses this in
1318 human_readable (HR_NUMTYPE n)
1320 /* These suffixes are compatible with those of GNU `ls -lh'. */
1321 static char powers[] =
1323 'K', /* kilobyte, 2^10 bytes */
1324 'M', /* megabyte, 2^20 bytes */
1325 'G', /* gigabyte, 2^30 bytes */
1326 'T', /* terabyte, 2^40 bytes */
1327 'P', /* petabyte, 2^50 bytes */
1328 'E', /* exabyte, 2^60 bytes */
1333 /* If the quantity is smaller than 1K, just print it. */
1336 snprintf (buf, sizeof (buf), "%d", (int) n);
1340 /* Loop over powers, dividing N with 1024 in each iteration. This
1341 works unchanged for all sizes of wgint, while still avoiding
1342 non-portable `long double' arithmetic. */
1343 for (i = 0; i < countof (powers); i++)
1345 /* At each iteration N is greater than the *subsequent* power.
1346 That way N/1024.0 produces a decimal number in the units of
1348 if ((n / 1024) < 1024 || i == countof (powers) - 1)
1350 double val = n / 1024.0;
1351 /* Print values smaller than 10 with one decimal digits, and
1352 others without any decimals. */
1353 snprintf (buf, sizeof (buf), "%.*f%c",
1354 val < 10 ? 1 : 0, val, powers[i]);
1359 return NULL; /* unreached */
1362 /* Count the digits in the provided number. Used to allocate space
1363 when printing numbers. */
1366 numdigit (wgint number)
1370 ++cnt; /* accomodate '-' */
1371 while ((number /= 10) != 0)
1376 #define PR(mask) *p++ = n / (mask) + '0'
1378 /* DIGITS_<D> is used to print a D-digit number and should be called
1379 with mask==10^(D-1). It prints n/mask (the first digit), reducing
1380 n to n%mask (the remaining digits), and calling DIGITS_<D-1>.
1381 Recursively this continues until DIGITS_1 is invoked. */
1383 #define DIGITS_1(mask) PR (mask)
1384 #define DIGITS_2(mask) PR (mask), n %= (mask), DIGITS_1 ((mask) / 10)
1385 #define DIGITS_3(mask) PR (mask), n %= (mask), DIGITS_2 ((mask) / 10)
1386 #define DIGITS_4(mask) PR (mask), n %= (mask), DIGITS_3 ((mask) / 10)
1387 #define DIGITS_5(mask) PR (mask), n %= (mask), DIGITS_4 ((mask) / 10)
1388 #define DIGITS_6(mask) PR (mask), n %= (mask), DIGITS_5 ((mask) / 10)
1389 #define DIGITS_7(mask) PR (mask), n %= (mask), DIGITS_6 ((mask) / 10)
1390 #define DIGITS_8(mask) PR (mask), n %= (mask), DIGITS_7 ((mask) / 10)
1391 #define DIGITS_9(mask) PR (mask), n %= (mask), DIGITS_8 ((mask) / 10)
1392 #define DIGITS_10(mask) PR (mask), n %= (mask), DIGITS_9 ((mask) / 10)
1394 /* DIGITS_<11-20> are only used on machines with 64-bit wgints. */
1396 #define DIGITS_11(mask) PR (mask), n %= (mask), DIGITS_10 ((mask) / 10)
1397 #define DIGITS_12(mask) PR (mask), n %= (mask), DIGITS_11 ((mask) / 10)
1398 #define DIGITS_13(mask) PR (mask), n %= (mask), DIGITS_12 ((mask) / 10)
1399 #define DIGITS_14(mask) PR (mask), n %= (mask), DIGITS_13 ((mask) / 10)
1400 #define DIGITS_15(mask) PR (mask), n %= (mask), DIGITS_14 ((mask) / 10)
1401 #define DIGITS_16(mask) PR (mask), n %= (mask), DIGITS_15 ((mask) / 10)
1402 #define DIGITS_17(mask) PR (mask), n %= (mask), DIGITS_16 ((mask) / 10)
1403 #define DIGITS_18(mask) PR (mask), n %= (mask), DIGITS_17 ((mask) / 10)
1404 #define DIGITS_19(mask) PR (mask), n %= (mask), DIGITS_18 ((mask) / 10)
1406 /* Shorthand for casting to wgint. */
1409 /* Print NUMBER to BUFFER in base 10. This is equivalent to
1410 `sprintf(buffer, "%lld", (long long) number)', only typically much
1411 faster and portable to machines without long long.
1413 The speedup may make a difference in programs that frequently
1414 convert numbers to strings. Some implementations of sprintf,
1415 particularly the one in some versions of GNU libc, have been known
1416 to be quite slow when converting integers to strings.
1418 Return the pointer to the location where the terminating zero was
1419 printed. (Equivalent to calling buffer+strlen(buffer) after the
1422 BUFFER should be large enough to accept as many bytes as you expect
1423 the number to take up. On machines with 64-bit wgints the maximum
1424 needed size is 24 bytes. That includes the digits needed for the
1425 largest 64-bit number, the `-' sign in case it's negative, and the
1426 terminating '\0'. */
1429 number_to_string (char *buffer, wgint number)
1434 int last_digit_char = 0;
1436 #if (SIZEOF_WGINT != 4) && (SIZEOF_WGINT != 8)
1437 /* We are running in a very strange environment. Leave the correct
1438 printing to sprintf. */
1439 p += sprintf (buf, "%j", (intmax_t) (n));
1440 #else /* (SIZEOF_WGINT == 4) || (SIZEOF_WGINT == 8) */
1446 /* n = -n would overflow because -n would evaluate to a
1447 wgint value larger than WGINT_MAX. Need to make n
1448 smaller and handle the last digit separately. */
1449 int last_digit = n % 10;
1450 /* The sign of n%10 is implementation-defined. */
1452 last_digit_char = '0' - last_digit;
1454 last_digit_char = '0' + last_digit;
1455 /* After n is made smaller, -n will not overflow. */
1463 /* Use the DIGITS_ macro appropriate for N's number of digits. That
1464 way printing any N is fully open-coded without a loop or jump.
1465 (Also see description of DIGITS_*.) */
1467 if (n < 10) DIGITS_1 (1);
1468 else if (n < 100) DIGITS_2 (10);
1469 else if (n < 1000) DIGITS_3 (100);
1470 else if (n < 10000) DIGITS_4 (1000);
1471 else if (n < 100000) DIGITS_5 (10000);
1472 else if (n < 1000000) DIGITS_6 (100000);
1473 else if (n < 10000000) DIGITS_7 (1000000);
1474 else if (n < 100000000) DIGITS_8 (10000000);
1475 else if (n < 1000000000) DIGITS_9 (100000000);
1476 #if SIZEOF_WGINT == 4
1477 /* wgint is 32 bits wide: no number has more than 10 digits. */
1478 else DIGITS_10 (1000000000);
1480 /* wgint is 64 bits wide: handle numbers with 9-19 decimal digits.
1481 Constants are constructed by compile-time multiplication to avoid
1482 dealing with different notations for 64-bit constants
1483 (nL/nLL/nI64, depending on the compiler and architecture). */
1484 else if (n < 10*(W)1000000000) DIGITS_10 (1000000000);
1485 else if (n < 100*(W)1000000000) DIGITS_11 (10*(W)1000000000);
1486 else if (n < 1000*(W)1000000000) DIGITS_12 (100*(W)1000000000);
1487 else if (n < 10000*(W)1000000000) DIGITS_13 (1000*(W)1000000000);
1488 else if (n < 100000*(W)1000000000) DIGITS_14 (10000*(W)1000000000);
1489 else if (n < 1000000*(W)1000000000) DIGITS_15 (100000*(W)1000000000);
1490 else if (n < 10000000*(W)1000000000) DIGITS_16 (1000000*(W)1000000000);
1491 else if (n < 100000000*(W)1000000000) DIGITS_17 (10000000*(W)1000000000);
1492 else if (n < 1000000000*(W)1000000000) DIGITS_18 (100000000*(W)1000000000);
1493 else DIGITS_19 (1000000000*(W)1000000000);
1496 if (last_digit_char)
1497 *p++ = last_digit_char;
1500 #endif /* (SIZEOF_WGINT == 4) || (SIZEOF_WGINT == 8) */
1507 #undef SPRINTF_WGINT
1530 /* Print NUMBER to a statically allocated string and return a pointer
1531 to the printed representation.
1533 This function is intended to be used in conjunction with printf.
1534 It is hard to portably print wgint values:
1535 a) you cannot use printf("%ld", number) because wgint can be long
1536 long on 32-bit machines with LFS.
1537 b) you cannot use printf("%lld", number) because NUMBER could be
1538 long on 32-bit machines without LFS, or on 64-bit machines,
1539 which do not require LFS. Also, Windows doesn't support %lld.
1540 c) you cannot use printf("%j", (int_max_t) number) because not all
1541 versions of printf support "%j", the most notable being the one
1543 d) you cannot #define WGINT_FMT to the appropriate format and use
1544 printf(WGINT_FMT, number) because that would break translations
1545 for user-visible messages, such as printf("Downloaded: %d
1548 What you should use instead is printf("%s", number_to_static_string
1551 CAVEAT: since the function returns pointers to static data, you
1552 must be careful to copy its result before calling it again.
1553 However, to make it more useful with printf, the function maintains
1554 an internal ring of static buffers to return. That way things like
1555 printf("%s %s", number_to_static_string (num1),
1556 number_to_static_string (num2)) work as expected. Three buffers
1557 are currently used, which means that "%s %s %s" will work, but "%s
1558 %s %s %s" won't. If you need to print more than three wgints,
1559 bump the RING_SIZE (or rethink your message.) */
1562 number_to_static_string (wgint number)
1564 static char ring[RING_SIZE][24];
1566 char *buf = ring[ringpos];
1567 number_to_string (buf, number);
1568 ringpos = (ringpos + 1) % RING_SIZE;
1572 /* Determine the width of the terminal we're running on. If that's
1573 not possible, return 0. */
1576 determine_screen_width (void)
1578 /* If there's a way to get the terminal size using POSIX
1579 tcgetattr(), somebody please tell me. */
1584 if (opt.lfilename != NULL)
1587 fd = fileno (stderr);
1588 if (ioctl (fd, TIOCGWINSZ, &wsz) < 0)
1589 return 0; /* most likely ENOTTY */
1592 #elif defined(WINDOWS)
1593 CONSOLE_SCREEN_BUFFER_INFO csbi;
1594 if (!GetConsoleScreenBufferInfo (GetStdHandle (STD_ERROR_HANDLE), &csbi))
1596 return csbi.dwSize.X;
1597 #else /* neither TIOCGWINSZ nor WINDOWS */
1599 #endif /* neither TIOCGWINSZ nor WINDOWS */
1602 /* Whether the rnd system (either rand or [dl]rand48) has been
1604 static int rnd_seeded;
1606 /* Return a random number between 0 and MAX-1, inclusive.
1608 If the system does not support lrand48 and MAX is greater than the
1609 value of RAND_MAX+1 on the system, the returned value will be in
1610 the range [0, RAND_MAX]. This may be fixed in a future release.
1611 The random number generator is seeded automatically the first time
1614 This uses lrand48 where available, rand elsewhere. DO NOT use it
1615 for cryptography. It is only meant to be used in situations where
1616 quality of the random numbers returned doesn't really matter. */
1619 random_number (int max)
1624 srand48 ((long) time (NULL) ^ (long) getpid ());
1627 return lrand48 () % max;
1628 #else /* not HAVE_DRAND48 */
1634 srand ((unsigned) time (NULL) ^ (unsigned) getpid ());
1639 /* Like rand() % max, but uses the high-order bits for better
1640 randomness on architectures where rand() is implemented using a
1641 simple congruential generator. */
1643 bounded = (double) max * rnd / (RAND_MAX + 1.0);
1644 return (int) bounded;
1646 #endif /* not HAVE_DRAND48 */
1649 /* Return a random uniformly distributed floating point number in the
1650 [0, 1) range. Uses drand48 where available, and a really lame
1651 kludge elsewhere. */
1659 srand48 ((long) time (NULL) ^ (long) getpid ());
1663 #else /* not HAVE_DRAND48 */
1664 return ( random_number (10000) / 10000.0
1665 + random_number (10000) / (10000.0 * 10000.0)
1666 + random_number (10000) / (10000.0 * 10000.0 * 10000.0)
1667 + random_number (10000) / (10000.0 * 10000.0 * 10000.0 * 10000.0));
1668 #endif /* not HAVE_DRAND48 */
1671 /* Implementation of run_with_timeout, a generic timeout-forcing
1672 routine for systems with Unix-like signal handling. */
1674 #ifdef USE_SIGNAL_TIMEOUT
1675 # ifdef HAVE_SIGSETJMP
1676 # define SETJMP(env) sigsetjmp (env, 1)
1678 static sigjmp_buf run_with_timeout_env;
1681 abort_run_with_timeout (int sig)
1683 assert (sig == SIGALRM);
1684 siglongjmp (run_with_timeout_env, -1);
1686 # else /* not HAVE_SIGSETJMP */
1687 # define SETJMP(env) setjmp (env)
1689 static jmp_buf run_with_timeout_env;
1692 abort_run_with_timeout (int sig)
1694 assert (sig == SIGALRM);
1695 /* We don't have siglongjmp to preserve the set of blocked signals;
1696 if we longjumped out of the handler at this point, SIGALRM would
1697 remain blocked. We must unblock it manually. */
1698 int mask = siggetmask ();
1699 mask &= ~sigmask (SIGALRM);
1702 /* Now it's safe to longjump. */
1703 longjmp (run_with_timeout_env, -1);
1705 # endif /* not HAVE_SIGSETJMP */
1707 /* Arrange for SIGALRM to be delivered in TIMEOUT seconds. This uses
1708 setitimer where available, alarm otherwise.
1710 TIMEOUT should be non-zero. If the timeout value is so small that
1711 it would be rounded to zero, it is rounded to the least legal value
1712 instead (1us for setitimer, 1s for alarm). That ensures that
1713 SIGALRM will be delivered in all cases. */
1716 alarm_set (double timeout)
1719 /* Use the modern itimer interface. */
1720 struct itimerval itv;
1722 itv.it_value.tv_sec = (long) timeout;
1723 itv.it_value.tv_usec = 1000000 * (timeout - (long)timeout);
1724 if (itv.it_value.tv_sec == 0 && itv.it_value.tv_usec == 0)
1725 /* Ensure that we wait for at least the minimum interval.
1726 Specifying zero would mean "wait forever". */
1727 itv.it_value.tv_usec = 1;
1728 setitimer (ITIMER_REAL, &itv, NULL);
1729 #else /* not ITIMER_REAL */
1730 /* Use the old alarm() interface. */
1731 int secs = (int) timeout;
1733 /* Round TIMEOUTs smaller than 1 to 1, not to zero. This is
1734 because alarm(0) means "never deliver the alarm", i.e. "wait
1735 forever", which is not what someone who specifies a 0.5s
1736 timeout would expect. */
1739 #endif /* not ITIMER_REAL */
1742 /* Cancel the alarm set with alarm_set. */
1748 struct itimerval disable;
1750 setitimer (ITIMER_REAL, &disable, NULL);
1751 #else /* not ITIMER_REAL */
1753 #endif /* not ITIMER_REAL */
1756 /* Call FUN(ARG), but don't allow it to run for more than TIMEOUT
1757 seconds. Returns true if the function was interrupted with a
1758 timeout, false otherwise.
1760 This works by setting up SIGALRM to be delivered in TIMEOUT seconds
1761 using setitimer() or alarm(). The timeout is enforced by
1762 longjumping out of the SIGALRM handler. This has several
1763 advantages compared to the traditional approach of relying on
1764 signals causing system calls to exit with EINTR:
1766 * The callback function is *forcibly* interrupted after the
1767 timeout expires, (almost) regardless of what it was doing and
1768 whether it was in a syscall. For example, a calculation that
1769 takes a long time is interrupted as reliably as an IO
1772 * It works with both SYSV and BSD signals because it doesn't
1773 depend on the default setting of SA_RESTART.
1775 * It doesn't require special handler setup beyond a simple call
1776 to signal(). (It does use sigsetjmp/siglongjmp, but they're
1779 The only downside is that, if FUN allocates internal resources that
1780 are normally freed prior to exit from the functions, they will be
1781 lost in case of timeout. */
1784 run_with_timeout (double timeout, void (*fun) (void *), void *arg)
1794 signal (SIGALRM, abort_run_with_timeout);
1795 if (SETJMP (run_with_timeout_env) != 0)
1797 /* Longjumped out of FUN with a timeout. */
1798 signal (SIGALRM, SIG_DFL);
1801 alarm_set (timeout);
1804 /* Preserve errno in case alarm() or signal() modifies it. */
1805 saved_errno = errno;
1807 signal (SIGALRM, SIG_DFL);
1808 errno = saved_errno;
1813 #else /* not USE_SIGNAL_TIMEOUT */
1816 /* A stub version of run_with_timeout that just calls FUN(ARG). Don't
1817 define it under Windows, because Windows has its own version of
1818 run_with_timeout that uses threads. */
1821 run_with_timeout (double timeout, void (*fun) (void *), void *arg)
1826 #endif /* not WINDOWS */
1827 #endif /* not USE_SIGNAL_TIMEOUT */
1831 /* Sleep the specified amount of seconds. On machines without
1832 nanosleep(), this may sleep shorter if interrupted by signals. */
1835 xsleep (double seconds)
1837 #ifdef HAVE_NANOSLEEP
1838 /* nanosleep is the preferred interface because it offers high
1839 accuracy and, more importantly, because it allows us to reliably
1840 restart receiving a signal such as SIGWINCH. (There was an
1841 actual Debian bug report about --limit-rate malfunctioning while
1842 the terminal was being resized.) */
1843 struct timespec sleep, remaining;
1844 sleep.tv_sec = (long) seconds;
1845 sleep.tv_nsec = 1000000000 * (seconds - (long) seconds);
1846 while (nanosleep (&sleep, &remaining) < 0 && errno == EINTR)
1847 /* If nanosleep has been interrupted by a signal, adjust the
1848 sleeping period and return to sleep. */
1850 #elif defined(HAVE_USLEEP)
1851 /* If usleep is available, use it in preference to select. */
1854 /* On some systems, usleep cannot handle values larger than
1855 1,000,000. If the period is larger than that, use sleep
1856 first, then add usleep for subsecond accuracy. */
1858 seconds -= (long) seconds;
1860 usleep (seconds * 1000000);
1861 #else /* fall back select */
1862 /* Note that, although Windows supports select, it can't be used to
1863 implement sleeping because Winsock's select doesn't implement
1864 timeout when it is passed NULL pointers for all fd sets. (But it
1865 does under Cygwin, which implements Unix-compatible select.) */
1866 struct timeval sleep;
1867 sleep.tv_sec = (long) seconds;
1868 sleep.tv_usec = 1000000 * (seconds - (long) seconds);
1869 select (0, NULL, NULL, NULL, &sleep);
1870 /* If select returns -1 and errno is EINTR, it means we were
1871 interrupted by a signal. But without knowing how long we've
1872 actually slept, we can't return to sleep. Using gettimeofday to
1873 track sleeps is slow and unreliable due to clock skew. */
1877 #endif /* not WINDOWS */
1879 /* Encode the octets in DATA of length LENGTH to base64 format,
1880 storing the result to DEST. The output will be zero-terminated,
1881 and must point to a writable buffer of at least
1882 1+BASE64_LENGTH(length) bytes. The function returns the length of
1883 the resulting base64 data, not counting the terminating zero.
1885 This implementation does not emit newlines after 76 characters of
1889 base64_encode (const void *data, int length, char *dest)
1891 /* Conversion table. */
1892 static const char tbl[64] = {
1893 'A','B','C','D','E','F','G','H','I','J','K','L','M','N','O','P',
1894 'Q','R','S','T','U','V','W','X','Y','Z','a','b','c','d','e','f',
1895 'g','h','i','j','k','l','m','n','o','p','q','r','s','t','u','v',
1896 'w','x','y','z','0','1','2','3','4','5','6','7','8','9','+','/'
1898 /* Access bytes in DATA as unsigned char, otherwise the shifts below
1899 don't work for data with MSB set. */
1900 const unsigned char *s = data;
1901 /* Theoretical ANSI violation when length < 3. */
1902 const unsigned char *end = (const unsigned char *) data + length - 2;
1905 /* Transform the 3x8 bits to 4x6 bits, as required by base64. */
1906 for (; s < end; s += 3)
1908 *p++ = tbl[s[0] >> 2];
1909 *p++ = tbl[((s[0] & 3) << 4) + (s[1] >> 4)];
1910 *p++ = tbl[((s[1] & 0xf) << 2) + (s[2] >> 6)];
1911 *p++ = tbl[s[2] & 0x3f];
1914 /* Pad the result if necessary... */
1918 *p++ = tbl[s[0] >> 2];
1919 *p++ = tbl[(s[0] & 3) << 4];
1924 *p++ = tbl[s[0] >> 2];
1925 *p++ = tbl[((s[0] & 3) << 4) + (s[1] >> 4)];
1926 *p++ = tbl[((s[1] & 0xf) << 2)];
1930 /* ...and zero-terminate it. */
1936 /* Store in C the next non-whitespace character from the string, or \0
1937 when end of string is reached. */
1938 #define NEXT_CHAR(c, p) do { \
1939 c = (unsigned char) *p++; \
1940 } while (ISSPACE (c))
1942 #define IS_ASCII(c) (((c) & 0x80) == 0)
1944 /* Decode data from BASE64 (a null-terminated string) into memory
1945 pointed to by DEST. DEST is assumed to be large enough to
1946 accomodate the decoded data, which is guaranteed to be no more than
1949 Since DEST is assumed to contain binary data, it is not
1950 NUL-terminated. The function returns the length of the data
1951 written to TO. -1 is returned in case of error caused by malformed
1954 This function originates from Free Recode. */
1957 base64_decode (const char *base64, void *dest)
1959 /* Table of base64 values for first 128 characters. Note that this
1960 assumes ASCII (but so does Wget in other places). */
1961 static const signed char base64_char_to_value[128] =
1963 -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, /* 0- 9 */
1964 -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, /* 10- 19 */
1965 -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, /* 20- 29 */
1966 -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, /* 30- 39 */
1967 -1, -1, -1, 62, -1, -1, -1, 63, 52, 53, /* 40- 49 */
1968 54, 55, 56, 57, 58, 59, 60, 61, -1, -1, /* 50- 59 */
1969 -1, -1, -1, -1, -1, 0, 1, 2, 3, 4, /* 60- 69 */
1970 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, /* 70- 79 */
1971 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, /* 80- 89 */
1972 25, -1, -1, -1, -1, -1, -1, 26, 27, 28, /* 90- 99 */
1973 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, /* 100-109 */
1974 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, /* 110-119 */
1975 49, 50, 51, -1, -1, -1, -1, -1 /* 120-127 */
1977 #define BASE64_CHAR_TO_VALUE(c) ((int) base64_char_to_value[c])
1978 #define IS_BASE64(c) ((IS_ASCII (c) && BASE64_CHAR_TO_VALUE (c) >= 0) || c == '=')
1980 const char *p = base64;
1986 unsigned long value;
1988 /* Process first byte of a quadruplet. */
1992 if (c == '=' || !IS_BASE64 (c))
1993 return -1; /* illegal char while decoding base64 */
1994 value = BASE64_CHAR_TO_VALUE (c) << 18;
1996 /* Process second byte of a quadruplet. */
1999 return -1; /* premature EOF while decoding base64 */
2000 if (c == '=' || !IS_BASE64 (c))
2001 return -1; /* illegal char while decoding base64 */
2002 value |= BASE64_CHAR_TO_VALUE (c) << 12;
2005 /* Process third byte of a quadruplet. */
2008 return -1; /* premature EOF while decoding base64 */
2010 return -1; /* illegal char while decoding base64 */
2016 return -1; /* premature EOF while decoding base64 */
2018 return -1; /* padding `=' expected but not found */
2022 value |= BASE64_CHAR_TO_VALUE (c) << 6;
2023 *q++ = 0xff & value >> 8;
2025 /* Process fourth byte of a quadruplet. */
2028 return -1; /* premature EOF while decoding base64 */
2032 return -1; /* illegal char while decoding base64 */
2034 value |= BASE64_CHAR_TO_VALUE (c);
2035 *q++ = 0xff & value;
2038 #undef BASE64_CHAR_TO_VALUE
2040 return q - (char *) dest;
2046 /* Simple merge sort for use by stable_sort. Implementation courtesy
2047 Zeljko Vrba with additional debugging by Nenad Barbutov. */
2050 mergesort_internal (void *base, void *temp, size_t size, size_t from, size_t to,
2051 int (*cmpfun) (const void *, const void *))
2053 #define ELT(array, pos) ((char *)(array) + (pos) * size)
2057 size_t mid = (to + from) / 2;
2058 mergesort_internal (base, temp, size, from, mid, cmpfun);
2059 mergesort_internal (base, temp, size, mid + 1, to, cmpfun);
2062 for (k = from; (i <= mid) && (j <= to); k++)
2063 if (cmpfun (ELT (base, i), ELT (base, j)) <= 0)
2064 memcpy (ELT (temp, k), ELT (base, i++), size);
2066 memcpy (ELT (temp, k), ELT (base, j++), size);
2068 memcpy (ELT (temp, k++), ELT (base, i++), size);
2070 memcpy (ELT (temp, k++), ELT (base, j++), size);
2071 for (k = from; k <= to; k++)
2072 memcpy (ELT (base, k), ELT (temp, k), size);
2077 /* Stable sort with interface exactly like standard library's qsort.
2078 Uses mergesort internally, allocating temporary storage with
2082 stable_sort (void *base, size_t nmemb, size_t size,
2083 int (*cmpfun) (const void *, const void *))
2087 void *temp = alloca (nmemb * size * sizeof (void *));
2088 mergesort_internal (base, temp, size, 0, nmemb - 1, cmpfun);
2092 /* Print a decimal number. If it is equal to or larger than ten, the
2093 number is rounded. Otherwise it is printed with one significant
2094 digit without trailing zeros and with no more than three fractional
2095 digits total. For example, 0.1 is printed as "0.1", 0.035 is
2096 printed as "0.04", 0.0091 as "0.009", and 0.0003 as simply "0".
2098 This is useful for displaying durations because it provides
2099 order-of-magnitude information without unnecessary clutter --
2100 long-running downloads are shown without the fractional part, and
2101 short ones still retain one significant digit. */
2104 print_decimal (double number)
2106 static char buf[32];
2107 double n = number >= 0 ? number : -number;
2110 /* Cut off at 9.95 because the below %.1f would round 9.96 to
2111 "10.0" instead of "10". OTOH 9.94 will print as "9.9". */
2112 snprintf (buf, sizeof buf, "%.0f", number);
2114 snprintf (buf, sizeof buf, "%.1f", number);
2115 else if (n >= 0.001)
2116 snprintf (buf, sizeof buf, "%.1g", number);
2117 else if (n >= 0.0005)
2118 /* round [0.0005, 0.001) to 0.001 */
2119 snprintf (buf, sizeof buf, "%.3f", number);
2121 /* print numbers close to 0 as 0, not 0.000 */
2138 { "/somedir", "/somedir", true },
2139 { "/somedir", "/somedir/d2", true },
2140 { "/somedir/d1", "/somedir", false },
2143 for (i = 0; i < countof(test_array); ++i)
2145 bool res = subdir_p (test_array[i].d1, test_array[i].d2);
2147 mu_assert ("test_subdir_p: wrong result",
2148 res == test_array[i].result);
2155 test_dir_matches_p()
2163 { { "/somedir", "/someotherdir", NULL }, "somedir", true },
2164 { { "/somedir", "/someotherdir", NULL }, "anotherdir", false },
2165 { { "/somedir", "/*otherdir", NULL }, "anotherdir", true },
2166 { { "/somedir/d1", "/someotherdir", NULL }, "somedir/d1", true },
2167 { { "/somedir/d1", "/someotherdir", NULL }, "d1", false },
2170 for (i = 0; i < countof(test_array); ++i)
2172 bool res = dir_matches_p (test_array[i].dirlist, test_array[i].dir);
2174 mu_assert ("test_dir_matches_p: wrong result",
2175 res == test_array[i].result);
2181 #endif /* TESTING */