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 /* Return pointer to a static char[] buffer in which zero-terminated
257 string-representation of TM (in form hh:mm:ss) is printed.
259 If TM is NULL, the current time will be used. */
262 time_str (time_t *tm)
264 static char output[15];
266 time_t secs = tm ? *tm : time (NULL);
270 /* In case of error, return the empty string. Maybe we should
271 just abort if this happens? */
275 ptm = localtime (&secs);
276 sprintf (output, "%02d:%02d:%02d", ptm->tm_hour, ptm->tm_min, ptm->tm_sec);
280 /* Like the above, but include the date: YYYY-MM-DD hh:mm:ss. */
283 datetime_str (time_t *tm)
285 static char output[20]; /* "YYYY-MM-DD hh:mm:ss" + \0 */
287 time_t secs = tm ? *tm : time (NULL);
291 /* In case of error, return the empty string. Maybe we should
292 just abort if this happens? */
296 ptm = localtime (&secs);
297 sprintf (output, "%04d-%02d-%02d %02d:%02d:%02d",
298 ptm->tm_year + 1900, ptm->tm_mon + 1, ptm->tm_mday,
299 ptm->tm_hour, ptm->tm_min, ptm->tm_sec);
303 /* The Windows versions of the following two functions are defined in
308 fork_to_background (void)
311 /* Whether we arrange our own version of opt.lfilename here. */
312 bool logfile_changed = false;
316 /* We must create the file immediately to avoid either a race
317 condition (which arises from using unique_name and failing to
318 use fopen_excl) or lying to the user about the log file name
319 (which arises from using unique_name, printing the name, and
320 using fopen_excl later on.) */
321 FILE *new_log_fp = unique_create (DEFAULT_LOGFILE, false, &opt.lfilename);
324 logfile_changed = true;
337 /* parent, no error */
338 printf (_("Continuing in background, pid %d.\n"), (int) pid);
340 printf (_("Output will be written to `%s'.\n"), opt.lfilename);
341 exit (0); /* #### should we use _exit()? */
344 /* child: give up the privileges and keep running. */
346 freopen ("/dev/null", "r", stdin);
347 freopen ("/dev/null", "w", stdout);
348 freopen ("/dev/null", "w", stderr);
350 #endif /* not WINDOWS */
352 /* "Touch" FILE, i.e. make its mtime ("modified time") equal the time
353 specified with TM. The atime ("access time") is set to the current
357 touch (const char *file, time_t tm)
359 #ifdef HAVE_STRUCT_UTIMBUF
360 struct utimbuf times;
368 times.actime = time (NULL);
369 if (utime (file, ×) == -1)
370 logprintf (LOG_NOTQUIET, "utime(%s): %s\n", file, strerror (errno));
373 /* Checks if FILE is a symbolic link, and removes it if it is. Does
374 nothing under MS-Windows. */
376 remove_link (const char *file)
381 if (lstat (file, &st) == 0 && S_ISLNK (st.st_mode))
383 DEBUGP (("Unlinking %s (symlink).\n", file));
386 logprintf (LOG_VERBOSE, _("Failed to unlink symlink `%s': %s\n"),
387 file, strerror (errno));
392 /* Does FILENAME exist? This is quite a lousy implementation, since
393 it supplies no error codes -- only a yes-or-no answer. Thus it
394 will return that a file does not exist if, e.g., the directory is
395 unreadable. I don't mind it too much currently, though. The
396 proper way should, of course, be to have a third, error state,
397 other than true/false, but that would introduce uncalled-for
398 additional complexity to the callers. */
400 file_exists_p (const char *filename)
403 return access (filename, F_OK) >= 0;
406 return stat (filename, &buf) >= 0;
410 /* Returns 0 if PATH is a directory, 1 otherwise (any kind of file).
411 Returns 0 on error. */
413 file_non_directory_p (const char *path)
416 /* Use lstat() rather than stat() so that symbolic links pointing to
417 directories can be identified correctly. */
418 if (lstat (path, &buf) != 0)
420 return S_ISDIR (buf.st_mode) ? false : true;
423 /* Return the size of file named by FILENAME, or -1 if it cannot be
424 opened or seeked into. */
426 file_size (const char *filename)
428 #if defined(HAVE_FSEEKO) && defined(HAVE_FTELLO)
430 /* We use fseek rather than stat to determine the file size because
431 that way we can also verify that the file is readable without
432 explicitly checking for permissions. Inspired by the POST patch
434 FILE *fp = fopen (filename, "rb");
437 fseeko (fp, 0, SEEK_END);
443 if (stat (filename, &st) < 0)
449 /* stat file names named PREFIX.1, PREFIX.2, etc., until one that
450 doesn't exist is found. Return a freshly allocated copy of the
454 unique_name_1 (const char *prefix)
457 int plen = strlen (prefix);
458 char *template = (char *)alloca (plen + 1 + 24);
459 char *template_tail = template + plen;
461 memcpy (template, prefix, plen);
462 *template_tail++ = '.';
465 number_to_string (template_tail, count++);
466 while (file_exists_p (template));
468 return xstrdup (template);
471 /* Return a unique file name, based on FILE.
473 More precisely, if FILE doesn't exist, it is returned unmodified.
474 If not, FILE.1 is tried, then FILE.2, etc. The first FILE.<number>
475 file name that doesn't exist is returned.
477 The resulting file is not created, only verified that it didn't
478 exist at the point in time when the function was called.
479 Therefore, where security matters, don't rely that the file created
480 by this function exists until you open it with O_EXCL or
483 If ALLOW_PASSTHROUGH is 0, it always returns a freshly allocated
484 string. Otherwise, it may return FILE if the file doesn't exist
485 (and therefore doesn't need changing). */
488 unique_name (const char *file, bool allow_passthrough)
490 /* If the FILE itself doesn't exist, return it without
492 if (!file_exists_p (file))
493 return allow_passthrough ? (char *)file : xstrdup (file);
495 /* Otherwise, find a numeric suffix that results in unused file name
497 return unique_name_1 (file);
500 /* Create a file based on NAME, except without overwriting an existing
501 file with that name. Providing O_EXCL is correctly implemented,
502 this function does not have the race condition associated with
503 opening the file returned by unique_name. */
506 unique_create (const char *name, bool binary, char **opened_name)
508 /* unique file name, based on NAME */
509 char *uname = unique_name (name, false);
511 while ((fp = fopen_excl (uname, binary)) == NULL && errno == EEXIST)
514 uname = unique_name (name, false);
516 if (opened_name && fp != NULL)
519 *opened_name = uname;
531 /* Open the file for writing, with the addition that the file is
532 opened "exclusively". This means that, if the file already exists,
533 this function will *fail* and errno will be set to EEXIST. If
534 BINARY is set, the file will be opened in binary mode, equivalent
537 If opening the file fails for any reason, including the file having
538 previously existed, this function returns NULL and sets errno
542 fopen_excl (const char *fname, bool binary)
546 int flags = O_WRONLY | O_CREAT | O_EXCL;
551 fd = open (fname, flags, 0666);
554 return fdopen (fd, binary ? "wb" : "w");
555 #else /* not O_EXCL */
556 /* Manually check whether the file exists. This is prone to race
557 conditions, but systems without O_EXCL haven't deserved
559 if (file_exists_p (fname))
564 return fopen (fname, binary ? "wb" : "w");
565 #endif /* not O_EXCL */
568 /* Create DIRECTORY. If some of the pathname components of DIRECTORY
569 are missing, create them first. In case any mkdir() call fails,
570 return its error status. Returns 0 on successful completion.
572 The behaviour of this function should be identical to the behaviour
573 of `mkdir -p' on systems where mkdir supports the `-p' option. */
575 make_directory (const char *directory)
577 int i, ret, quit = 0;
580 /* Make a copy of dir, to be able to write to it. Otherwise, the
581 function is unsafe if called with a read-only char *argument. */
582 STRDUP_ALLOCA (dir, directory);
584 /* If the first character of dir is '/', skip it (and thus enable
585 creation of absolute-pathname directories. */
586 for (i = (*dir == '/'); 1; ++i)
588 for (; dir[i] && dir[i] != '/'; i++)
593 /* Check whether the directory already exists. Allow creation of
594 of intermediate directories to fail, as the initial path components
595 are not necessarily directories! */
596 if (!file_exists_p (dir))
597 ret = mkdir (dir, 0777);
608 /* Merge BASE with FILE. BASE can be a directory or a file name, FILE
609 should be a file name.
611 file_merge("/foo/bar", "baz") => "/foo/baz"
612 file_merge("/foo/bar/", "baz") => "/foo/bar/baz"
613 file_merge("foo", "bar") => "bar"
615 In other words, it's a simpler and gentler version of uri_merge. */
618 file_merge (const char *base, const char *file)
621 const char *cut = (const char *)strrchr (base, '/');
624 return xstrdup (file);
626 result = xmalloc (cut - base + 1 + strlen (file) + 1);
627 memcpy (result, base, cut - base);
628 result[cut - base] = '/';
629 strcpy (result + (cut - base) + 1, file);
634 /* Like fnmatch, but performs a case-insensitive match. */
637 fnmatch_nocase (const char *pattern, const char *string, int flags)
640 /* The FNM_CASEFOLD flag started as a GNU extension, but it is now
641 also present on *BSD platforms, and possibly elsewhere. */
642 return fnmatch (pattern, string, flags | FNM_CASEFOLD);
644 /* Turn PATTERN and STRING to lower case and call fnmatch on them. */
645 char *patcopy = (char *) alloca (strlen (pattern) + 1);
646 char *strcopy = (char *) alloca (strlen (string) + 1);
648 for (p = patcopy; *pattern; pattern++, p++)
649 *p = TOLOWER (*pattern);
651 for (p = strcopy; *string; string++, p++)
652 *p = TOLOWER (*string);
654 return fnmatch (patcopy, strcopy, flags);
658 static bool in_acclist (const char *const *, const char *, bool);
660 /* Determine whether a file is acceptable to be followed, according to
661 lists of patterns to accept/reject. */
663 acceptable (const char *s)
667 while (l && s[l] != '/')
674 return (in_acclist ((const char *const *)opt.accepts, s, true)
675 && !in_acclist ((const char *const *)opt.rejects, s, true));
677 return in_acclist ((const char *const *)opt.accepts, s, true);
679 else if (opt.rejects)
680 return !in_acclist ((const char *const *)opt.rejects, s, true);
684 /* Check if D2 is a subdirectory of D1. E.g. if D1 is `/something', subdir_p()
685 will return true if and only if D2 begins with `/something/' or is exactly
688 subdir_p (const char *d1, const char *d2)
690 if (!opt.ignore_case)
691 for (; *d1 && *d2 && (*d1 == *d2); ++d1, ++d2)
694 for (; *d1 && *d2 && (TOLOWER (*d1) == TOLOWER (*d2)); ++d1, ++d2)
697 return *d1 == '\0' && (*d2 == '\0' || *d2 == '/');
700 /* Iterate through DIRLIST (which must be NULL-terminated), and return the
701 first element that matches DIR, through wildcards or front comparison (as
704 dir_matches_p (char **dirlist, const char *dir)
707 int (*matcher) (const char *, const char *, int)
708 = opt.ignore_case ? fnmatch_nocase : fnmatch;
710 for (x = dirlist; *x; x++)
712 /* Remove leading '/' */
713 char *p = *x + (**x == '/');
714 if (has_wildcards_p (p))
716 if (matcher (p, dir, FNM_PATHNAME) == 0)
721 if (subdir_p (p, dir))
726 return *x ? true : false;
729 /* Returns whether DIRECTORY is acceptable for download, wrt the
730 include/exclude lists.
732 The leading `/' is ignored in paths; relative and absolute paths
733 may be freely intermixed. */
736 accdir (const char *directory)
738 /* Remove starting '/'. */
739 if (*directory == '/')
743 if (!dir_matches_p (opt.includes, directory))
748 if (dir_matches_p (opt.excludes, directory))
754 /* Return true if STRING ends with TAIL. For instance:
756 match_tail ("abc", "bc", false) -> 1
757 match_tail ("abc", "ab", false) -> 0
758 match_tail ("abc", "abc", false) -> 1
760 If FOLD_CASE is true, the comparison will be case-insensitive. */
763 match_tail (const char *string, const char *tail, bool fold_case)
767 /* We want this to be fast, so we code two loops, one with
768 case-folding, one without. */
772 for (i = strlen (string), j = strlen (tail); i >= 0 && j >= 0; i--, j--)
773 if (string[i] != tail[j])
778 for (i = strlen (string), j = strlen (tail); i >= 0 && j >= 0; i--, j--)
779 if (TOLOWER (string[i]) != TOLOWER (tail[j]))
783 /* If the tail was exhausted, the match was succesful. */
790 /* Checks whether string S matches each element of ACCEPTS. A list
791 element are matched either with fnmatch() or match_tail(),
792 according to whether the element contains wildcards or not.
794 If the BACKWARD is false, don't do backward comparison -- just compare
797 in_acclist (const char *const *accepts, const char *s, bool backward)
799 for (; *accepts; accepts++)
801 if (has_wildcards_p (*accepts))
803 int res = opt.ignore_case
804 ? fnmatch_nocase (*accepts, s, 0) : fnmatch (*accepts, s, 0);
805 /* fnmatch returns 0 if the pattern *does* match the string. */
813 if (match_tail (s, *accepts, opt.ignore_case))
818 int cmp = opt.ignore_case
819 ? strcasecmp (s, *accepts) : strcmp (s, *accepts);
828 /* Return the location of STR's suffix (file extension). Examples:
829 suffix ("foo.bar") -> "bar"
830 suffix ("foo.bar.baz") -> "baz"
831 suffix ("/foo/bar") -> NULL
832 suffix ("/foo.bar/baz") -> NULL */
834 suffix (const char *str)
838 for (i = strlen (str); i && str[i] != '/' && str[i] != '.'; i--)
842 return (char *)str + i;
847 /* Return true if S contains globbing wildcards (`*', `?', `[' or
851 has_wildcards_p (const char *s)
854 if (*s == '*' || *s == '?' || *s == '[' || *s == ']')
859 /* Return true if FNAME ends with a typical HTML suffix. The
860 following (case-insensitive) suffixes are presumed to be HTML
865 ?html (`?' matches one character)
867 #### CAVEAT. This is not necessarily a good indication that FNAME
868 refers to a file that contains HTML! */
870 has_html_suffix_p (const char *fname)
874 if ((suf = suffix (fname)) == NULL)
876 if (!strcasecmp (suf, "html"))
878 if (!strcasecmp (suf, "htm"))
880 if (suf[0] && !strcasecmp (suf + 1, "html"))
885 /* Read a line from FP and return the pointer to freshly allocated
886 storage. The storage space is obtained through malloc() and should
887 be freed with free() when it is no longer needed.
889 The length of the line is not limited, except by available memory.
890 The newline character at the end of line is retained. The line is
891 terminated with a zero character.
893 After end-of-file is encountered without anything being read, NULL
894 is returned. NULL is also returned on error. To distinguish
895 between these two cases, use the stdio function ferror(). */
898 read_whole_line (FILE *fp)
902 char *line = xmalloc (bufsize);
904 while (fgets (line + length, bufsize - length, fp))
906 length += strlen (line + length);
908 /* Possible for example when reading from a binary file where
909 a line begins with \0. */
912 if (line[length - 1] == '\n')
915 /* fgets() guarantees to read the whole line, or to use up the
916 space we've given it. We can double the buffer
919 line = xrealloc (line, bufsize);
921 if (length == 0 || ferror (fp))
926 if (length + 1 < bufsize)
927 /* Relieve the memory from our exponential greediness. We say
928 `length + 1' because the terminating \0 is not included in
929 LENGTH. We don't need to zero-terminate the string ourselves,
930 though, because fgets() does that. */
931 line = xrealloc (line, length + 1);
935 /* Read FILE into memory. A pointer to `struct file_memory' are
936 returned; use struct element `content' to access file contents, and
937 the element `length' to know the file length. `content' is *not*
938 zero-terminated, and you should *not* read or write beyond the [0,
939 length) range of characters.
941 After you are done with the file contents, call read_file_free to
944 Depending on the operating system and the type of file that is
945 being read, read_file() either mmap's the file into memory, or
946 reads the file into the core using read().
948 If file is named "-", fileno(stdin) is used for reading instead.
949 If you want to read from a real file named "-", use "./-" instead. */
952 read_file (const char *file)
955 struct file_memory *fm;
957 bool inhibit_close = false;
959 /* Some magic in the finest tradition of Perl and its kin: if FILE
960 is "-", just use stdin. */
964 inhibit_close = true;
965 /* Note that we don't inhibit mmap() in this case. If stdin is
966 redirected from a regular file, mmap() will still work. */
969 fd = open (file, O_RDONLY);
972 fm = xnew (struct file_memory);
977 if (fstat (fd, &buf) < 0)
979 fm->length = buf.st_size;
980 /* NOTE: As far as I know, the callers of this function never
981 modify the file text. Relying on this would enable us to
982 specify PROT_READ and MAP_SHARED for a marginal gain in
983 efficiency, but at some cost to generality. */
984 fm->content = mmap (NULL, fm->length, PROT_READ | PROT_WRITE,
986 if (fm->content == (char *)MAP_FAILED)
996 /* The most common reason why mmap() fails is that FD does not point
997 to a plain file. However, it's also possible that mmap() doesn't
998 work for a particular type of file. Therefore, whenever mmap()
999 fails, we just fall back to the regular method. */
1000 #endif /* HAVE_MMAP */
1003 size = 512; /* number of bytes fm->contents can
1004 hold at any given time. */
1005 fm->content = xmalloc (size);
1009 if (fm->length > size / 2)
1011 /* #### I'm not sure whether the whole exponential-growth
1012 thing makes sense with kernel read. On Linux at least,
1013 read() refuses to read more than 4K from a file at a
1014 single chunk anyway. But other Unixes might optimize it
1015 better, and it doesn't *hurt* anything, so I'm leaving
1018 /* Normally, we grow SIZE exponentially to make the number
1019 of calls to read() and realloc() logarithmic in relation
1020 to file size. However, read() can read an amount of data
1021 smaller than requested, and it would be unreasonable to
1022 double SIZE every time *something* was read. Therefore,
1023 we double SIZE only when the length exceeds half of the
1024 entire allocated size. */
1026 fm->content = xrealloc (fm->content, size);
1028 nread = read (fd, fm->content + fm->length, size - fm->length);
1030 /* Successful read. */
1031 fm->length += nread;
1041 if (size > fm->length && fm->length != 0)
1042 /* Due to exponential growth of fm->content, the allocated region
1043 might be much larger than what is actually needed. */
1044 fm->content = xrealloc (fm->content, fm->length);
1051 xfree (fm->content);
1056 /* Release the resources held by FM. Specifically, this calls
1057 munmap() or xfree() on fm->content, depending whether mmap or
1058 malloc/read were used to read in the file. It also frees the
1059 memory needed to hold the FM structure itself. */
1062 read_file_free (struct file_memory *fm)
1067 munmap (fm->content, fm->length);
1072 xfree (fm->content);
1077 /* Free the pointers in a NULL-terminated vector of pointers, then
1078 free the pointer itself. */
1080 free_vec (char **vec)
1091 /* Append vector V2 to vector V1. The function frees V2 and
1092 reallocates V1 (thus you may not use the contents of neither
1093 pointer after the call). If V1 is NULL, V2 is returned. */
1095 merge_vecs (char **v1, char **v2)
1105 /* To avoid j == 0 */
1110 for (i = 0; v1[i]; i++)
1113 for (j = 0; v2[j]; j++)
1115 /* Reallocate v1. */
1116 v1 = xrealloc (v1, (i + j + 1) * sizeof (char **));
1117 memcpy (v1 + i, v2, (j + 1) * sizeof (char *));
1122 /* Append a freshly allocated copy of STR to VEC. If VEC is NULL, it
1123 is allocated as needed. Return the new value of the vector. */
1126 vec_append (char **vec, const char *str)
1128 int cnt; /* count of vector elements, including
1129 the one we're about to append */
1132 for (cnt = 0; vec[cnt]; cnt++)
1138 /* Reallocate the array to fit the new element and the NULL. */
1139 vec = xrealloc (vec, (cnt + 1) * sizeof (char *));
1140 /* Append a copy of STR to the vector. */
1141 vec[cnt - 1] = xstrdup (str);
1146 /* Sometimes it's useful to create "sets" of strings, i.e. special
1147 hash tables where you want to store strings as keys and merely
1148 query for their existence. Here is a set of utility routines that
1149 makes that transparent. */
1152 string_set_add (struct hash_table *ht, const char *s)
1154 /* First check whether the set element already exists. If it does,
1155 do nothing so that we don't have to free() the old element and
1156 then strdup() a new one. */
1157 if (hash_table_contains (ht, s))
1160 /* We use "1" as value. It provides us a useful and clear arbitrary
1161 value, and it consumes no memory -- the pointers to the same
1162 string "1" will be shared by all the key-value pairs in all `set'
1164 hash_table_put (ht, xstrdup (s), "1");
1167 /* Synonym for hash_table_contains... */
1170 string_set_contains (struct hash_table *ht, const char *s)
1172 return hash_table_contains (ht, s);
1175 /* Convert the specified string set to array. ARRAY should be large
1176 enough to hold hash_table_count(ht) char pointers. */
1178 void string_set_to_array (struct hash_table *ht, char **array)
1180 hash_table_iterator iter;
1181 for (hash_table_iterate (ht, &iter); hash_table_iter_next (&iter); )
1182 *array++ = iter.key;
1185 /* Free the string set. This frees both the storage allocated for
1186 keys and the actual hash table. (hash_table_destroy would only
1187 destroy the hash table.) */
1190 string_set_free (struct hash_table *ht)
1192 hash_table_iterator iter;
1193 for (hash_table_iterate (ht, &iter); hash_table_iter_next (&iter); )
1195 hash_table_destroy (ht);
1198 /* Utility function: simply call xfree() on all keys and values of HT. */
1201 free_keys_and_values (struct hash_table *ht)
1203 hash_table_iterator iter;
1204 for (hash_table_iterate (ht, &iter); hash_table_iter_next (&iter); )
1211 /* Get digit grouping data for thousand separors by calling
1212 localeconv(). The data includes separator string and grouping info
1213 and is cached after the first call to the function.
1215 In locales that don't set a thousand separator (such as the "C"
1216 locale), this forces it to be ",". We are now only showing
1217 thousand separators in one place, so this shouldn't be a problem in
1221 get_grouping_data (const char **sep, const char **grouping)
1223 static const char *cached_sep;
1224 static const char *cached_grouping;
1225 static bool initialized;
1228 /* Get the grouping info from the locale. */
1229 struct lconv *lconv = localeconv ();
1230 cached_sep = lconv->thousands_sep;
1231 cached_grouping = lconv->grouping;
1234 /* Many locales (such as "C" or "hr_HR") don't specify
1235 grouping, which we still want to use it for legibility.
1236 In those locales set the sep char to ',', unless that
1237 character is used for decimal point, in which case set it
1239 if (*lconv->decimal_point != ',')
1243 cached_grouping = "\x03";
1248 *grouping = cached_grouping;
1251 /* Return a printed representation of N with thousand separators.
1252 This should respect locale settings, with the exception of the "C"
1253 locale which mandates no separator, but we use one anyway.
1255 Unfortunately, we cannot use %'d (in fact it would be %'j) to get
1256 the separators because it's too non-portable, and it's hard to test
1257 for this feature at configure time. Besides, it wouldn't display
1258 separators in the "C" locale, still used by many Unix users. */
1261 with_thousand_seps (wgint n)
1263 static char outbuf[48];
1264 char *p = outbuf + sizeof outbuf;
1266 /* Info received from locale */
1267 const char *grouping, *sep;
1270 /* State information */
1271 int i = 0, groupsize;
1272 const char *atgroup;
1274 bool negative = n < 0;
1276 /* Initialize grouping data. */
1277 get_grouping_data (&sep, &grouping);
1278 seplen = strlen (sep);
1280 groupsize = *atgroup++;
1282 /* This would overflow on WGINT_MIN, but printing negative numbers
1283 is not an important goal of this fuinction. */
1287 /* Write the number into the buffer, backwards, inserting the
1288 separators as necessary. */
1292 *--p = n % 10 + '0';
1296 /* Prepend SEP to every groupsize'd digit and get new groupsize. */
1297 if (++i == groupsize)
1302 memcpy (p -= seplen, sep, seplen);
1305 groupsize = *atgroup++;
1314 /* N, a byte quantity, is converted to a human-readable abberviated
1315 form a la sizes printed by `ls -lh'. The result is written to a
1316 static buffer, a pointer to which is returned.
1318 Unlike `with_thousand_seps', this approximates to the nearest unit.
1319 Quoting GNU libit: "Most people visually process strings of 3-4
1320 digits effectively, but longer strings of digits are more prone to
1321 misinterpretation. Hence, converting to an abbreviated form
1322 usually improves readability."
1324 This intentionally uses kilobyte (KB), megabyte (MB), etc. in their
1325 original computer-related meaning of "powers of 1024". We don't
1326 use the "*bibyte" names invented in 1998, and seldom used in
1327 practice. Wikipedia's entry on "binary prefix" discusses this in
1331 human_readable (HR_NUMTYPE n)
1333 /* These suffixes are compatible with those of GNU `ls -lh'. */
1334 static char powers[] =
1336 'K', /* kilobyte, 2^10 bytes */
1337 'M', /* megabyte, 2^20 bytes */
1338 'G', /* gigabyte, 2^30 bytes */
1339 'T', /* terabyte, 2^40 bytes */
1340 'P', /* petabyte, 2^50 bytes */
1341 'E', /* exabyte, 2^60 bytes */
1346 /* If the quantity is smaller than 1K, just print it. */
1349 snprintf (buf, sizeof (buf), "%d", (int) n);
1353 /* Loop over powers, dividing N with 1024 in each iteration. This
1354 works unchanged for all sizes of wgint, while still avoiding
1355 non-portable `long double' arithmetic. */
1356 for (i = 0; i < countof (powers); i++)
1358 /* At each iteration N is greater than the *subsequent* power.
1359 That way N/1024.0 produces a decimal number in the units of
1361 if ((n / 1024) < 1024 || i == countof (powers) - 1)
1363 double val = n / 1024.0;
1364 /* Print values smaller than 10 with one decimal digits, and
1365 others without any decimals. */
1366 snprintf (buf, sizeof (buf), "%.*f%c",
1367 val < 10 ? 1 : 0, val, powers[i]);
1372 return NULL; /* unreached */
1375 /* Count the digits in the provided number. Used to allocate space
1376 when printing numbers. */
1379 numdigit (wgint number)
1383 ++cnt; /* accomodate '-' */
1384 while ((number /= 10) != 0)
1389 #define PR(mask) *p++ = n / (mask) + '0'
1391 /* DIGITS_<D> is used to print a D-digit number and should be called
1392 with mask==10^(D-1). It prints n/mask (the first digit), reducing
1393 n to n%mask (the remaining digits), and calling DIGITS_<D-1>.
1394 Recursively this continues until DIGITS_1 is invoked. */
1396 #define DIGITS_1(mask) PR (mask)
1397 #define DIGITS_2(mask) PR (mask), n %= (mask), DIGITS_1 ((mask) / 10)
1398 #define DIGITS_3(mask) PR (mask), n %= (mask), DIGITS_2 ((mask) / 10)
1399 #define DIGITS_4(mask) PR (mask), n %= (mask), DIGITS_3 ((mask) / 10)
1400 #define DIGITS_5(mask) PR (mask), n %= (mask), DIGITS_4 ((mask) / 10)
1401 #define DIGITS_6(mask) PR (mask), n %= (mask), DIGITS_5 ((mask) / 10)
1402 #define DIGITS_7(mask) PR (mask), n %= (mask), DIGITS_6 ((mask) / 10)
1403 #define DIGITS_8(mask) PR (mask), n %= (mask), DIGITS_7 ((mask) / 10)
1404 #define DIGITS_9(mask) PR (mask), n %= (mask), DIGITS_8 ((mask) / 10)
1405 #define DIGITS_10(mask) PR (mask), n %= (mask), DIGITS_9 ((mask) / 10)
1407 /* DIGITS_<11-20> are only used on machines with 64-bit wgints. */
1409 #define DIGITS_11(mask) PR (mask), n %= (mask), DIGITS_10 ((mask) / 10)
1410 #define DIGITS_12(mask) PR (mask), n %= (mask), DIGITS_11 ((mask) / 10)
1411 #define DIGITS_13(mask) PR (mask), n %= (mask), DIGITS_12 ((mask) / 10)
1412 #define DIGITS_14(mask) PR (mask), n %= (mask), DIGITS_13 ((mask) / 10)
1413 #define DIGITS_15(mask) PR (mask), n %= (mask), DIGITS_14 ((mask) / 10)
1414 #define DIGITS_16(mask) PR (mask), n %= (mask), DIGITS_15 ((mask) / 10)
1415 #define DIGITS_17(mask) PR (mask), n %= (mask), DIGITS_16 ((mask) / 10)
1416 #define DIGITS_18(mask) PR (mask), n %= (mask), DIGITS_17 ((mask) / 10)
1417 #define DIGITS_19(mask) PR (mask), n %= (mask), DIGITS_18 ((mask) / 10)
1419 /* Shorthand for casting to wgint. */
1422 /* Print NUMBER to BUFFER in base 10. This is equivalent to
1423 `sprintf(buffer, "%lld", (long long) number)', only typically much
1424 faster and portable to machines without long long.
1426 The speedup may make a difference in programs that frequently
1427 convert numbers to strings. Some implementations of sprintf,
1428 particularly the one in some versions of GNU libc, have been known
1429 to be quite slow when converting integers to strings.
1431 Return the pointer to the location where the terminating zero was
1432 printed. (Equivalent to calling buffer+strlen(buffer) after the
1435 BUFFER should be large enough to accept as many bytes as you expect
1436 the number to take up. On machines with 64-bit wgints the maximum
1437 needed size is 24 bytes. That includes the digits needed for the
1438 largest 64-bit number, the `-' sign in case it's negative, and the
1439 terminating '\0'. */
1442 number_to_string (char *buffer, wgint number)
1447 int last_digit_char = 0;
1449 #if (SIZEOF_WGINT != 4) && (SIZEOF_WGINT != 8)
1450 /* We are running in a very strange environment. Leave the correct
1451 printing to sprintf. */
1452 p += sprintf (buf, "%j", (intmax_t) (n));
1453 #else /* (SIZEOF_WGINT == 4) || (SIZEOF_WGINT == 8) */
1459 /* n = -n would overflow because -n would evaluate to a
1460 wgint value larger than WGINT_MAX. Need to make n
1461 smaller and handle the last digit separately. */
1462 int last_digit = n % 10;
1463 /* The sign of n%10 is implementation-defined. */
1465 last_digit_char = '0' - last_digit;
1467 last_digit_char = '0' + last_digit;
1468 /* After n is made smaller, -n will not overflow. */
1476 /* Use the DIGITS_ macro appropriate for N's number of digits. That
1477 way printing any N is fully open-coded without a loop or jump.
1478 (Also see description of DIGITS_*.) */
1480 if (n < 10) DIGITS_1 (1);
1481 else if (n < 100) DIGITS_2 (10);
1482 else if (n < 1000) DIGITS_3 (100);
1483 else if (n < 10000) DIGITS_4 (1000);
1484 else if (n < 100000) DIGITS_5 (10000);
1485 else if (n < 1000000) DIGITS_6 (100000);
1486 else if (n < 10000000) DIGITS_7 (1000000);
1487 else if (n < 100000000) DIGITS_8 (10000000);
1488 else if (n < 1000000000) DIGITS_9 (100000000);
1489 #if SIZEOF_WGINT == 4
1490 /* wgint is 32 bits wide: no number has more than 10 digits. */
1491 else DIGITS_10 (1000000000);
1493 /* wgint is 64 bits wide: handle numbers with 9-19 decimal digits.
1494 Constants are constructed by compile-time multiplication to avoid
1495 dealing with different notations for 64-bit constants
1496 (nL/nLL/nI64, depending on the compiler and architecture). */
1497 else if (n < 10*(W)1000000000) DIGITS_10 (1000000000);
1498 else if (n < 100*(W)1000000000) DIGITS_11 (10*(W)1000000000);
1499 else if (n < 1000*(W)1000000000) DIGITS_12 (100*(W)1000000000);
1500 else if (n < 10000*(W)1000000000) DIGITS_13 (1000*(W)1000000000);
1501 else if (n < 100000*(W)1000000000) DIGITS_14 (10000*(W)1000000000);
1502 else if (n < 1000000*(W)1000000000) DIGITS_15 (100000*(W)1000000000);
1503 else if (n < 10000000*(W)1000000000) DIGITS_16 (1000000*(W)1000000000);
1504 else if (n < 100000000*(W)1000000000) DIGITS_17 (10000000*(W)1000000000);
1505 else if (n < 1000000000*(W)1000000000) DIGITS_18 (100000000*(W)1000000000);
1506 else DIGITS_19 (1000000000*(W)1000000000);
1509 if (last_digit_char)
1510 *p++ = last_digit_char;
1513 #endif /* (SIZEOF_WGINT == 4) || (SIZEOF_WGINT == 8) */
1520 #undef SPRINTF_WGINT
1543 /* Print NUMBER to a statically allocated string and return a pointer
1544 to the printed representation.
1546 This function is intended to be used in conjunction with printf.
1547 It is hard to portably print wgint values:
1548 a) you cannot use printf("%ld", number) because wgint can be long
1549 long on 32-bit machines with LFS.
1550 b) you cannot use printf("%lld", number) because NUMBER could be
1551 long on 32-bit machines without LFS, or on 64-bit machines,
1552 which do not require LFS. Also, Windows doesn't support %lld.
1553 c) you cannot use printf("%j", (int_max_t) number) because not all
1554 versions of printf support "%j", the most notable being the one
1556 d) you cannot #define WGINT_FMT to the appropriate format and use
1557 printf(WGINT_FMT, number) because that would break translations
1558 for user-visible messages, such as printf("Downloaded: %d
1561 What you should use instead is printf("%s", number_to_static_string
1564 CAVEAT: since the function returns pointers to static data, you
1565 must be careful to copy its result before calling it again.
1566 However, to make it more useful with printf, the function maintains
1567 an internal ring of static buffers to return. That way things like
1568 printf("%s %s", number_to_static_string (num1),
1569 number_to_static_string (num2)) work as expected. Three buffers
1570 are currently used, which means that "%s %s %s" will work, but "%s
1571 %s %s %s" won't. If you need to print more than three wgints,
1572 bump the RING_SIZE (or rethink your message.) */
1575 number_to_static_string (wgint number)
1577 static char ring[RING_SIZE][24];
1579 char *buf = ring[ringpos];
1580 number_to_string (buf, number);
1581 ringpos = (ringpos + 1) % RING_SIZE;
1585 /* Determine the width of the terminal we're running on. If that's
1586 not possible, return 0. */
1589 determine_screen_width (void)
1591 /* If there's a way to get the terminal size using POSIX
1592 tcgetattr(), somebody please tell me. */
1597 if (opt.lfilename != NULL)
1600 fd = fileno (stderr);
1601 if (ioctl (fd, TIOCGWINSZ, &wsz) < 0)
1602 return 0; /* most likely ENOTTY */
1605 #elif defined(WINDOWS)
1606 CONSOLE_SCREEN_BUFFER_INFO csbi;
1607 if (!GetConsoleScreenBufferInfo (GetStdHandle (STD_ERROR_HANDLE), &csbi))
1609 return csbi.dwSize.X;
1610 #else /* neither TIOCGWINSZ nor WINDOWS */
1612 #endif /* neither TIOCGWINSZ nor WINDOWS */
1615 /* Whether the rnd system (either rand or [dl]rand48) has been
1617 static int rnd_seeded;
1619 /* Return a random number between 0 and MAX-1, inclusive.
1621 If the system does not support lrand48 and MAX is greater than the
1622 value of RAND_MAX+1 on the system, the returned value will be in
1623 the range [0, RAND_MAX]. This may be fixed in a future release.
1624 The random number generator is seeded automatically the first time
1627 This uses lrand48 where available, rand elsewhere. DO NOT use it
1628 for cryptography. It is only meant to be used in situations where
1629 quality of the random numbers returned doesn't really matter. */
1632 random_number (int max)
1637 srand48 ((long) time (NULL) ^ (long) getpid ());
1640 return lrand48 () % max;
1641 #else /* not HAVE_DRAND48 */
1647 srand ((unsigned) time (NULL) ^ (unsigned) getpid ());
1652 /* Like rand() % max, but uses the high-order bits for better
1653 randomness on architectures where rand() is implemented using a
1654 simple congruential generator. */
1656 bounded = (double) max * rnd / (RAND_MAX + 1.0);
1657 return (int) bounded;
1659 #endif /* not HAVE_DRAND48 */
1662 /* Return a random uniformly distributed floating point number in the
1663 [0, 1) range. Uses drand48 where available, and a really lame
1664 kludge elsewhere. */
1672 srand48 ((long) time (NULL) ^ (long) getpid ());
1676 #else /* not HAVE_DRAND48 */
1677 return ( random_number (10000) / 10000.0
1678 + random_number (10000) / (10000.0 * 10000.0)
1679 + random_number (10000) / (10000.0 * 10000.0 * 10000.0)
1680 + random_number (10000) / (10000.0 * 10000.0 * 10000.0 * 10000.0));
1681 #endif /* not HAVE_DRAND48 */
1684 /* Implementation of run_with_timeout, a generic timeout-forcing
1685 routine for systems with Unix-like signal handling. */
1687 #ifdef USE_SIGNAL_TIMEOUT
1688 # ifdef HAVE_SIGSETJMP
1689 # define SETJMP(env) sigsetjmp (env, 1)
1691 static sigjmp_buf run_with_timeout_env;
1694 abort_run_with_timeout (int sig)
1696 assert (sig == SIGALRM);
1697 siglongjmp (run_with_timeout_env, -1);
1699 # else /* not HAVE_SIGSETJMP */
1700 # define SETJMP(env) setjmp (env)
1702 static jmp_buf run_with_timeout_env;
1705 abort_run_with_timeout (int sig)
1707 assert (sig == SIGALRM);
1708 /* We don't have siglongjmp to preserve the set of blocked signals;
1709 if we longjumped out of the handler at this point, SIGALRM would
1710 remain blocked. We must unblock it manually. */
1711 int mask = siggetmask ();
1712 mask &= ~sigmask (SIGALRM);
1715 /* Now it's safe to longjump. */
1716 longjmp (run_with_timeout_env, -1);
1718 # endif /* not HAVE_SIGSETJMP */
1720 /* Arrange for SIGALRM to be delivered in TIMEOUT seconds. This uses
1721 setitimer where available, alarm otherwise.
1723 TIMEOUT should be non-zero. If the timeout value is so small that
1724 it would be rounded to zero, it is rounded to the least legal value
1725 instead (1us for setitimer, 1s for alarm). That ensures that
1726 SIGALRM will be delivered in all cases. */
1729 alarm_set (double timeout)
1732 /* Use the modern itimer interface. */
1733 struct itimerval itv;
1735 itv.it_value.tv_sec = (long) timeout;
1736 itv.it_value.tv_usec = 1000000 * (timeout - (long)timeout);
1737 if (itv.it_value.tv_sec == 0 && itv.it_value.tv_usec == 0)
1738 /* Ensure that we wait for at least the minimum interval.
1739 Specifying zero would mean "wait forever". */
1740 itv.it_value.tv_usec = 1;
1741 setitimer (ITIMER_REAL, &itv, NULL);
1742 #else /* not ITIMER_REAL */
1743 /* Use the old alarm() interface. */
1744 int secs = (int) timeout;
1746 /* Round TIMEOUTs smaller than 1 to 1, not to zero. This is
1747 because alarm(0) means "never deliver the alarm", i.e. "wait
1748 forever", which is not what someone who specifies a 0.5s
1749 timeout would expect. */
1752 #endif /* not ITIMER_REAL */
1755 /* Cancel the alarm set with alarm_set. */
1761 struct itimerval disable;
1763 setitimer (ITIMER_REAL, &disable, NULL);
1764 #else /* not ITIMER_REAL */
1766 #endif /* not ITIMER_REAL */
1769 /* Call FUN(ARG), but don't allow it to run for more than TIMEOUT
1770 seconds. Returns true if the function was interrupted with a
1771 timeout, false otherwise.
1773 This works by setting up SIGALRM to be delivered in TIMEOUT seconds
1774 using setitimer() or alarm(). The timeout is enforced by
1775 longjumping out of the SIGALRM handler. This has several
1776 advantages compared to the traditional approach of relying on
1777 signals causing system calls to exit with EINTR:
1779 * The callback function is *forcibly* interrupted after the
1780 timeout expires, (almost) regardless of what it was doing and
1781 whether it was in a syscall. For example, a calculation that
1782 takes a long time is interrupted as reliably as an IO
1785 * It works with both SYSV and BSD signals because it doesn't
1786 depend on the default setting of SA_RESTART.
1788 * It doesn't require special handler setup beyond a simple call
1789 to signal(). (It does use sigsetjmp/siglongjmp, but they're
1792 The only downside is that, if FUN allocates internal resources that
1793 are normally freed prior to exit from the functions, they will be
1794 lost in case of timeout. */
1797 run_with_timeout (double timeout, void (*fun) (void *), void *arg)
1807 signal (SIGALRM, abort_run_with_timeout);
1808 if (SETJMP (run_with_timeout_env) != 0)
1810 /* Longjumped out of FUN with a timeout. */
1811 signal (SIGALRM, SIG_DFL);
1814 alarm_set (timeout);
1817 /* Preserve errno in case alarm() or signal() modifies it. */
1818 saved_errno = errno;
1820 signal (SIGALRM, SIG_DFL);
1821 errno = saved_errno;
1826 #else /* not USE_SIGNAL_TIMEOUT */
1829 /* A stub version of run_with_timeout that just calls FUN(ARG). Don't
1830 define it under Windows, because Windows has its own version of
1831 run_with_timeout that uses threads. */
1834 run_with_timeout (double timeout, void (*fun) (void *), void *arg)
1839 #endif /* not WINDOWS */
1840 #endif /* not USE_SIGNAL_TIMEOUT */
1844 /* Sleep the specified amount of seconds. On machines without
1845 nanosleep(), this may sleep shorter if interrupted by signals. */
1848 xsleep (double seconds)
1850 #ifdef HAVE_NANOSLEEP
1851 /* nanosleep is the preferred interface because it offers high
1852 accuracy and, more importantly, because it allows us to reliably
1853 restart receiving a signal such as SIGWINCH. (There was an
1854 actual Debian bug report about --limit-rate malfunctioning while
1855 the terminal was being resized.) */
1856 struct timespec sleep, remaining;
1857 sleep.tv_sec = (long) seconds;
1858 sleep.tv_nsec = 1000000000 * (seconds - (long) seconds);
1859 while (nanosleep (&sleep, &remaining) < 0 && errno == EINTR)
1860 /* If nanosleep has been interrupted by a signal, adjust the
1861 sleeping period and return to sleep. */
1863 #elif defined(HAVE_USLEEP)
1864 /* If usleep is available, use it in preference to select. */
1867 /* On some systems, usleep cannot handle values larger than
1868 1,000,000. If the period is larger than that, use sleep
1869 first, then add usleep for subsecond accuracy. */
1871 seconds -= (long) seconds;
1873 usleep (seconds * 1000000);
1874 #else /* fall back select */
1875 /* Note that, although Windows supports select, it can't be used to
1876 implement sleeping because Winsock's select doesn't implement
1877 timeout when it is passed NULL pointers for all fd sets. (But it
1878 does under Cygwin, which implements Unix-compatible select.) */
1879 struct timeval sleep;
1880 sleep.tv_sec = (long) seconds;
1881 sleep.tv_usec = 1000000 * (seconds - (long) seconds);
1882 select (0, NULL, NULL, NULL, &sleep);
1883 /* If select returns -1 and errno is EINTR, it means we were
1884 interrupted by a signal. But without knowing how long we've
1885 actually slept, we can't return to sleep. Using gettimeofday to
1886 track sleeps is slow and unreliable due to clock skew. */
1890 #endif /* not WINDOWS */
1892 /* Encode the octets in DATA of length LENGTH to base64 format,
1893 storing the result to DEST. The output will be zero-terminated,
1894 and must point to a writable buffer of at least
1895 1+BASE64_LENGTH(length) bytes. The function returns the length of
1896 the resulting base64 data, not counting the terminating zero.
1898 This implementation does not emit newlines after 76 characters of
1902 base64_encode (const void *data, int length, char *dest)
1904 /* Conversion table. */
1905 static const char tbl[64] = {
1906 'A','B','C','D','E','F','G','H','I','J','K','L','M','N','O','P',
1907 'Q','R','S','T','U','V','W','X','Y','Z','a','b','c','d','e','f',
1908 'g','h','i','j','k','l','m','n','o','p','q','r','s','t','u','v',
1909 'w','x','y','z','0','1','2','3','4','5','6','7','8','9','+','/'
1911 /* Access bytes in DATA as unsigned char, otherwise the shifts below
1912 don't work for data with MSB set. */
1913 const unsigned char *s = data;
1914 /* Theoretical ANSI violation when length < 3. */
1915 const unsigned char *end = (const unsigned char *) data + length - 2;
1918 /* Transform the 3x8 bits to 4x6 bits, as required by base64. */
1919 for (; s < end; s += 3)
1921 *p++ = tbl[s[0] >> 2];
1922 *p++ = tbl[((s[0] & 3) << 4) + (s[1] >> 4)];
1923 *p++ = tbl[((s[1] & 0xf) << 2) + (s[2] >> 6)];
1924 *p++ = tbl[s[2] & 0x3f];
1927 /* Pad the result if necessary... */
1931 *p++ = tbl[s[0] >> 2];
1932 *p++ = tbl[(s[0] & 3) << 4];
1937 *p++ = tbl[s[0] >> 2];
1938 *p++ = tbl[((s[0] & 3) << 4) + (s[1] >> 4)];
1939 *p++ = tbl[((s[1] & 0xf) << 2)];
1943 /* ...and zero-terminate it. */
1949 /* Store in C the next non-whitespace character from the string, or \0
1950 when end of string is reached. */
1951 #define NEXT_CHAR(c, p) do { \
1952 c = (unsigned char) *p++; \
1953 } while (ISSPACE (c))
1955 #define IS_ASCII(c) (((c) & 0x80) == 0)
1957 /* Decode data from BASE64 (a null-terminated string) into memory
1958 pointed to by DEST. DEST is assumed to be large enough to
1959 accomodate the decoded data, which is guaranteed to be no more than
1962 Since DEST is assumed to contain binary data, it is not
1963 NUL-terminated. The function returns the length of the data
1964 written to TO. -1 is returned in case of error caused by malformed
1967 This function originates from Free Recode. */
1970 base64_decode (const char *base64, void *dest)
1972 /* Table of base64 values for first 128 characters. Note that this
1973 assumes ASCII (but so does Wget in other places). */
1974 static const signed char base64_char_to_value[128] =
1976 -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, /* 0- 9 */
1977 -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, /* 10- 19 */
1978 -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, /* 20- 29 */
1979 -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, /* 30- 39 */
1980 -1, -1, -1, 62, -1, -1, -1, 63, 52, 53, /* 40- 49 */
1981 54, 55, 56, 57, 58, 59, 60, 61, -1, -1, /* 50- 59 */
1982 -1, -1, -1, -1, -1, 0, 1, 2, 3, 4, /* 60- 69 */
1983 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, /* 70- 79 */
1984 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, /* 80- 89 */
1985 25, -1, -1, -1, -1, -1, -1, 26, 27, 28, /* 90- 99 */
1986 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, /* 100-109 */
1987 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, /* 110-119 */
1988 49, 50, 51, -1, -1, -1, -1, -1 /* 120-127 */
1990 #define BASE64_CHAR_TO_VALUE(c) ((int) base64_char_to_value[c])
1991 #define IS_BASE64(c) ((IS_ASCII (c) && BASE64_CHAR_TO_VALUE (c) >= 0) || c == '=')
1993 const char *p = base64;
1999 unsigned long value;
2001 /* Process first byte of a quadruplet. */
2005 if (c == '=' || !IS_BASE64 (c))
2006 return -1; /* illegal char while decoding base64 */
2007 value = BASE64_CHAR_TO_VALUE (c) << 18;
2009 /* Process second byte of a quadruplet. */
2012 return -1; /* premature EOF while decoding base64 */
2013 if (c == '=' || !IS_BASE64 (c))
2014 return -1; /* illegal char while decoding base64 */
2015 value |= BASE64_CHAR_TO_VALUE (c) << 12;
2018 /* Process third byte of a quadruplet. */
2021 return -1; /* premature EOF while decoding base64 */
2023 return -1; /* illegal char while decoding base64 */
2029 return -1; /* premature EOF while decoding base64 */
2031 return -1; /* padding `=' expected but not found */
2035 value |= BASE64_CHAR_TO_VALUE (c) << 6;
2036 *q++ = 0xff & value >> 8;
2038 /* Process fourth byte of a quadruplet. */
2041 return -1; /* premature EOF while decoding base64 */
2045 return -1; /* illegal char while decoding base64 */
2047 value |= BASE64_CHAR_TO_VALUE (c);
2048 *q++ = 0xff & value;
2051 #undef BASE64_CHAR_TO_VALUE
2053 return q - (char *) dest;
2059 /* Simple merge sort for use by stable_sort. Implementation courtesy
2060 Zeljko Vrba with additional debugging by Nenad Barbutov. */
2063 mergesort_internal (void *base, void *temp, size_t size, size_t from, size_t to,
2064 int (*cmpfun) (const void *, const void *))
2066 #define ELT(array, pos) ((char *)(array) + (pos) * size)
2070 size_t mid = (to + from) / 2;
2071 mergesort_internal (base, temp, size, from, mid, cmpfun);
2072 mergesort_internal (base, temp, size, mid + 1, to, cmpfun);
2075 for (k = from; (i <= mid) && (j <= to); k++)
2076 if (cmpfun (ELT (base, i), ELT (base, j)) <= 0)
2077 memcpy (ELT (temp, k), ELT (base, i++), size);
2079 memcpy (ELT (temp, k), ELT (base, j++), size);
2081 memcpy (ELT (temp, k++), ELT (base, i++), size);
2083 memcpy (ELT (temp, k++), ELT (base, j++), size);
2084 for (k = from; k <= to; k++)
2085 memcpy (ELT (base, k), ELT (temp, k), size);
2090 /* Stable sort with interface exactly like standard library's qsort.
2091 Uses mergesort internally, allocating temporary storage with
2095 stable_sort (void *base, size_t nmemb, size_t size,
2096 int (*cmpfun) (const void *, const void *))
2100 void *temp = alloca (nmemb * size * sizeof (void *));
2101 mergesort_internal (base, temp, size, 0, nmemb - 1, cmpfun);
2105 /* Print a decimal number. If it is equal to or larger than ten, the
2106 number is rounded. Otherwise it is printed with one significant
2107 digit without trailing zeros and with no more than three fractional
2108 digits total. For example, 0.1 is printed as "0.1", 0.035 is
2109 printed as "0.04", 0.0091 as "0.009", and 0.0003 as simply "0".
2111 This is useful for displaying durations because it provides
2112 order-of-magnitude information without unnecessary clutter --
2113 long-running downloads are shown without the fractional part, and
2114 short ones still retain one significant digit. */
2117 print_decimal (double number)
2119 static char buf[32];
2120 double n = number >= 0 ? number : -number;
2123 /* Cut off at 9.95 because the below %.1f would round 9.96 to
2124 "10.0" instead of "10". OTOH 9.94 will print as "9.9". */
2125 snprintf (buf, sizeof buf, "%.0f", number);
2127 snprintf (buf, sizeof buf, "%.1f", number);
2128 else if (n >= 0.001)
2129 snprintf (buf, sizeof buf, "%.1g", number);
2130 else if (n >= 0.0005)
2131 /* round [0.0005, 0.001) to 0.001 */
2132 snprintf (buf, sizeof buf, "%.3f", number);
2134 /* print numbers close to 0 as 0, not 0.000 */
2151 { "/somedir", "/somedir", true },
2152 { "/somedir", "/somedir/d2", true },
2153 { "/somedir/d1", "/somedir", false },
2156 for (i = 0; i < countof(test_array); ++i)
2158 bool res = subdir_p (test_array[i].d1, test_array[i].d2);
2160 mu_assert ("test_subdir_p: wrong result",
2161 res == test_array[i].result);
2168 test_dir_matches_p()
2176 { { "/somedir", "/someotherdir", NULL }, "somedir", true },
2177 { { "/somedir", "/someotherdir", NULL }, "anotherdir", false },
2178 { { "/somedir", "/*otherdir", NULL }, "anotherdir", true },
2179 { { "/somedir/d1", "/someotherdir", NULL }, "somedir/d1", true },
2180 { { "/somedir/d1", "/someotherdir", NULL }, "d1", false },
2183 for (i = 0; i < countof(test_array); ++i)
2185 bool res = dir_matches_p (test_array[i].dirlist, test_array[i].dir);
2187 mu_assert ("test_dir_matches_p: wrong result",
2188 res == test_array[i].result);
2194 #endif /* TESTING */