X-Git-Url: http://sjero.net/git/?p=wget;a=blobdiff_plain;f=src%2Furl.c;h=1edfbae38c2fc9a9f0bfc5271e1ace7e1da7404b;hp=4a1c9f101278cadba3e08cf5d5f777e0dee6508d;hb=38a7829dcb4eb5dba28dbf0f05c6a80fea9217f8;hpb=67e6027ea130d06aeff365adfbc83f34d019b968 diff --git a/src/url.c b/src/url.c index 4a1c9f10..1edfbae3 100644 --- a/src/url.c +++ b/src/url.c @@ -169,7 +169,7 @@ static const unsigned char urlchr_table[256] = The transformation is done in place. If you need the original string intact, make a copy before calling this function. */ -static void +void url_unescape (char *s) { char *t = s; /* t - tortoise */ @@ -701,7 +701,10 @@ url_parse (const char *url, int *error, struct iri *iri, bool percent_encode) if (!iri->utf8_encode) new_url = NULL; else - iri->orig_url = xstrdup (url); + { + iri->orig_url = xstrdup (url); + percent_encode = true; + } } /* XXX XXX Could that change introduce (security) bugs ??? XXX XXX*/ @@ -1245,12 +1248,16 @@ mkalldirs (const char *path) The idea is to have a convenient and efficient way to construct a string by having various functions append data to it. Instead of passing the obligatory BASEVAR, SIZEVAR and TAILPOS to all the - functions in questions, we pass the pointer to this struct. */ + functions in questions, we pass the pointer to this struct. + + Functions that write to the members in this struct must make sure + that base remains null terminated by calling append_null(). + */ struct growable { char *base; - int size; - int tail; + int size; /* memory allocated */ + int tail; /* string length */ }; /* Ensure that the string can accept APPEND_COUNT more characters past @@ -1268,29 +1275,46 @@ struct growable { /* Move the tail position by APPEND_COUNT characters. */ #define TAIL_INCR(r, append_count) ((r)->tail += append_count) -/* Append the string STR to DEST. NOTICE: the string in DEST is not - terminated. */ +/* Append NULL to DEST. */ static void -append_string (const char *str, struct growable *dest) +append_null (struct growable *dest) { - int l = strlen (str); - GROW (dest, l); - memcpy (TAIL (dest), str, l); - TAIL_INCR (dest, l); + GROW (dest, 1); + *TAIL (dest) = 0; } -/* Append CH to DEST. For example, append_char (0, DEST) - zero-terminates DEST. */ - +/* Append CH to DEST. */ static void append_char (char ch, struct growable *dest) { - GROW (dest, 1); - *TAIL (dest) = ch; - TAIL_INCR (dest, 1); + if (ch) + { + GROW (dest, 1); + *TAIL (dest) = ch; + TAIL_INCR (dest, 1); + } + + append_null (dest); +} + +/* Append the string STR to DEST. */ +static void +append_string (const char *str, struct growable *dest) +{ + int l = strlen (str); + + if (l) + { + GROW (dest, l); + memcpy (TAIL (dest), str, l); + TAIL_INCR (dest, l); + } + + append_null (dest); } + enum { filechr_not_unix = 1, /* unusable on Unix, / and \0 */ filechr_not_windows = 2, /* unusable on Windows, one of \|/<>?:*" */ @@ -1455,6 +1479,7 @@ append_uri_pathel (const char *b, const char *e, bool escaped, } TAIL_INCR (dest, outlen); + append_null (dest); } /* Append to DEST the directory structure that corresponds the @@ -1585,7 +1610,26 @@ url_file_name (const struct url *u, char *replaced_filename) append_char ('\0', &temp_fnres); /* Check that the length of the file name is acceptable. */ +#ifdef WINDOWS + if (MAX_PATH > (fnres.tail + CHOMP_BUFFER + 2)) + { + max_length = MAX_PATH - (fnres.tail + CHOMP_BUFFER + 2); + /* FIXME: In Windows a filename is usually limited to 255 characters. + To really be accurate you could call GetVolumeInformation() to get + lpMaximumComponentLength + */ + if (max_length > 255) + { + max_length = 255; + } + } + else + { + max_length = 0; + } +#else max_length = get_max_length (fnres.base, fnres.tail, _PC_NAME_MAX) - CHOMP_BUFFER; +#endif if (max_length > 0 && strlen (temp_fnres.base) > max_length) { logprintf (LOG_NOTQUIET, "The name is too long, %lu chars total.\n", @@ -1606,9 +1650,6 @@ url_file_name (const struct url *u, char *replaced_filename) append_char ('/', &fnres); append_string (temp_fnres.base, &fnres); - /* Zero-terminate the file name. */ - append_char ('\0', &fnres); - fname = fnres.base; /* Make a final check that the path length is acceptable? */ @@ -1621,11 +1662,12 @@ url_file_name (const struct url *u, char *replaced_filename) 2) Retrieval with regetting. 3) Timestamping is used. 4) Hierarchy is built. + 5) Backups are specified. The exception is the case when file does exist and is a directory (see `mkalldirs' for explanation). */ - if ((opt.noclobber || opt.always_rest || opt.timestamping || opt.dirstruct) + if (ALLOW_CLOBBER && !(file_exists_p (fname) && !file_non_directory_p (fname))) { unique = fname; @@ -1994,7 +2036,7 @@ url_string (const struct url *url, enum url_auth_mode auth_mode) if (url->passwd) { if (auth_mode == URL_AUTH_HIDE_PASSWD) - quoted_passwd = HIDDEN_PASSWORD; + quoted_passwd = (char *) HIDDEN_PASSWORD; else quoted_passwd = url_escape_allow_passthrough (url->passwd); } @@ -2163,7 +2205,7 @@ ps (char *path) #endif static const char * -run_test (char *test, char *expected_result, enum url_scheme scheme, +run_test (const char *test, const char *expected_result, enum url_scheme scheme, bool expected_change) { char *test_copy = xstrdup (test); @@ -2192,8 +2234,8 @@ run_test (char *test, char *expected_result, enum url_scheme scheme, const char * test_path_simplify (void) { - static struct { - char *test, *result; + static const struct { + const char *test, *result; enum url_scheme scheme; bool should_modify; } tests[] = { @@ -2225,15 +2267,16 @@ test_path_simplify (void) { "a/b/../../c", "c", SCHEME_HTTP, true }, { "./a/../b", "b", SCHEME_HTTP, true } }; - int i; + unsigned i; for (i = 0; i < countof (tests); i++) { const char *message; - char *test = tests[i].test; - char *expected_result = tests[i].result; + const char *test = tests[i].test; + const char *expected_result = tests[i].result; enum url_scheme scheme = tests[i].scheme; bool expected_change = tests[i].should_modify; + message = run_test (test, expected_result, scheme, expected_change); if (message) return message; } @@ -2241,19 +2284,19 @@ test_path_simplify (void) } const char * -test_append_uri_pathel() +test_append_uri_pathel(void) { - int i; - struct { - char *original_url; - char *input; + unsigned i; + static const struct { + const char *original_url; + const char *input; bool escaped; - char *expected_result; + const char *expected_result; } test_array[] = { { "http://www.yoyodyne.com/path/", "somepage.html", false, "http://www.yoyodyne.com/path/somepage.html" }, }; - for (i = 0; i < sizeof(test_array)/sizeof(test_array[0]); ++i) + for (i = 0; i < countof(test_array); ++i) { struct growable dest; const char *p = test_array[i].input; @@ -2262,7 +2305,6 @@ test_append_uri_pathel() append_string (test_array[i].original_url, &dest); append_uri_pathel (p, p + strlen(p), test_array[i].escaped, &dest); - append_char ('\0', &dest); mu_assert ("test_append_uri_pathel: wrong result", strcmp (dest.base, test_array[i].expected_result) == 0); @@ -2271,13 +2313,13 @@ test_append_uri_pathel() return NULL; } -const char* -test_are_urls_equal() +const char * +test_are_urls_equal(void) { - int i; - struct { - char *url1; - char *url2; + unsigned i; + static const struct { + const char *url1; + const char *url2; bool expected_result; } test_array[] = { { "http://www.adomain.com/apath/", "http://www.adomain.com/apath/", true }, @@ -2288,7 +2330,7 @@ test_are_urls_equal() { "http://www.adomain.com/path%2f", "http://www.adomain.com/path/", false }, }; - for (i = 0; i < sizeof(test_array)/sizeof(test_array[0]); ++i) + for (i = 0; i < countof(test_array); ++i) { mu_assert ("test_are_urls_equal: wrong result", are_urls_equal (test_array[i].url1, test_array[i].url2) == test_array[i].expected_result);