2 * Copyright 2010 Jacek Caban for CodeWeavers
3 * Copyright 2010 Thomas Mullaly
5 * This library is free software; you can redistribute it and/or
6 * modify it under the terms of the GNU Lesser General Public
7 * License as published by the Free Software Foundation; either
8 * version 2.1 of the License, or (at your option) any later version.
10 * This library is distributed in the hope that it will be useful,
11 * but WITHOUT ANY WARRANTY; without even the implied warranty of
12 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
13 * Lesser General Public License for more details.
15 * You should have received a copy of the GNU Lesser General Public
16 * License along with this library; if not, write to the Free Software
17 * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA
20 #include "urlmon_main.h"
21 #include "wine/debug.h"
23 #define NO_SHLWAPI_REG
28 #define UINT_MAX 0xffffffff
29 #define USHORT_MAX 0xffff
31 #define URI_DISPLAY_NO_ABSOLUTE_URI 0x1
32 #define URI_DISPLAY_NO_DEFAULT_PORT_AUTH 0x2
34 #define ALLOW_NULL_TERM_SCHEME 0x01
35 #define ALLOW_NULL_TERM_USER_NAME 0x02
36 #define ALLOW_NULL_TERM_PASSWORD 0x04
37 #define ALLOW_BRACKETLESS_IP_LITERAL 0x08
38 #define SKIP_IP_FUTURE_CHECK 0x10
39 #define IGNORE_PORT_DELIMITER 0x20
41 #define RAW_URI_FORCE_PORT_DISP 0x1
42 #define RAW_URI_CONVERT_TO_DOS_PATH 0x2
44 #define COMBINE_URI_FORCE_FLAG_USE 0x1
46 WINE_DEFAULT_DEBUG_CHANNEL(urlmon);
48 static const IID IID_IUriObj = {0x4b364760,0x9f51,0x11df,{0x98,0x1c,0x08,0x00,0x20,0x0c,0x9a,0x66}};
52 IUriBuilderFactory IUriBuilderFactory_iface;
58 /* Information about the canonicalized URI's buffer. */
62 BOOL display_modifiers;
67 URL_SCHEME scheme_type;
75 Uri_HOST_TYPE host_type;
98 IUriBuilder IUriBuilder_iface;
102 DWORD modified_props;
135 /* IPv6 addresses can hold up to 8 h16 components. */
139 /* An IPv6 can have 1 elision ("::"). */
140 const WCHAR *elision;
142 /* An IPv6 can contain 1 IPv4 address as the last 32bits of the address. */
155 BOOL has_implicit_scheme;
156 BOOL has_implicit_ip;
162 URL_SCHEME scheme_type;
164 const WCHAR *username;
167 const WCHAR *password;
172 Uri_HOST_TYPE host_type;
175 ipv6_address ipv6_address;
188 const WCHAR *fragment;
192 static const CHAR hexDigits[] = "0123456789ABCDEF";
194 /* List of scheme types/scheme names that are recognized by the IUri interface as of IE 7. */
195 static const struct {
197 WCHAR scheme_name[16];
198 } recognized_schemes[] = {
199 {URL_SCHEME_FTP, {'f','t','p',0}},
200 {URL_SCHEME_HTTP, {'h','t','t','p',0}},
201 {URL_SCHEME_GOPHER, {'g','o','p','h','e','r',0}},
202 {URL_SCHEME_MAILTO, {'m','a','i','l','t','o',0}},
203 {URL_SCHEME_NEWS, {'n','e','w','s',0}},
204 {URL_SCHEME_NNTP, {'n','n','t','p',0}},
205 {URL_SCHEME_TELNET, {'t','e','l','n','e','t',0}},
206 {URL_SCHEME_WAIS, {'w','a','i','s',0}},
207 {URL_SCHEME_FILE, {'f','i','l','e',0}},
208 {URL_SCHEME_MK, {'m','k',0}},
209 {URL_SCHEME_HTTPS, {'h','t','t','p','s',0}},
210 {URL_SCHEME_SHELL, {'s','h','e','l','l',0}},
211 {URL_SCHEME_SNEWS, {'s','n','e','w','s',0}},
212 {URL_SCHEME_LOCAL, {'l','o','c','a','l',0}},
213 {URL_SCHEME_JAVASCRIPT, {'j','a','v','a','s','c','r','i','p','t',0}},
214 {URL_SCHEME_VBSCRIPT, {'v','b','s','c','r','i','p','t',0}},
215 {URL_SCHEME_ABOUT, {'a','b','o','u','t',0}},
216 {URL_SCHEME_RES, {'r','e','s',0}},
217 {URL_SCHEME_MSSHELLROOTED, {'m','s','-','s','h','e','l','l','-','r','o','o','t','e','d',0}},
218 {URL_SCHEME_MSSHELLIDLIST, {'m','s','-','s','h','e','l','l','-','i','d','l','i','s','t',0}},
219 {URL_SCHEME_MSHELP, {'h','c','p',0}},
220 {URL_SCHEME_WILDCARD, {'*',0}}
223 /* List of default ports Windows recognizes. */
224 static const struct {
227 } default_ports[] = {
228 {URL_SCHEME_FTP, 21},
229 {URL_SCHEME_HTTP, 80},
230 {URL_SCHEME_GOPHER, 70},
231 {URL_SCHEME_NNTP, 119},
232 {URL_SCHEME_TELNET, 23},
233 {URL_SCHEME_WAIS, 210},
234 {URL_SCHEME_HTTPS, 443},
237 /* List of 3-character top level domain names Windows seems to recognize.
238 * There might be more, but, these are the only ones I've found so far.
240 static const struct {
242 } recognized_tlds[] = {
252 static Uri *get_uri_obj(IUri *uri)
257 hres = IUri_QueryInterface(uri, &IID_IUriObj, (void**)&ret);
258 return SUCCEEDED(hres) ? ret : NULL;
261 static inline BOOL is_alpha(WCHAR val) {
262 return ((val >= 'a' && val <= 'z') || (val >= 'A' && val <= 'Z'));
265 static inline BOOL is_num(WCHAR val) {
266 return (val >= '0' && val <= '9');
269 static inline BOOL is_drive_path(const WCHAR *str) {
270 return (is_alpha(str[0]) && (str[1] == ':' || str[1] == '|'));
273 static inline BOOL is_unc_path(const WCHAR *str) {
274 return (str[0] == '\\' && str[0] == '\\');
277 static inline BOOL is_forbidden_dos_path_char(WCHAR val) {
278 return (val == '>' || val == '<' || val == '\"');
281 /* A URI is implicitly a file path if it begins with
282 * a drive letter (e.g. X:) or starts with "\\" (UNC path).
284 static inline BOOL is_implicit_file_path(const WCHAR *str) {
285 return (is_unc_path(str) || (is_alpha(str[0]) && str[1] == ':'));
288 /* Checks if the URI is a hierarchical URI. A hierarchical
289 * URI is one that has "//" after the scheme.
291 static BOOL check_hierarchical(const WCHAR **ptr) {
292 const WCHAR *start = *ptr;
307 /* unreserved = ALPHA / DIGIT / "-" / "." / "_" / "~" */
308 static inline BOOL is_unreserved(WCHAR val) {
309 return (is_alpha(val) || is_num(val) || val == '-' || val == '.' ||
310 val == '_' || val == '~');
313 /* sub-delims = "!" / "$" / "&" / "'" / "(" / ")"
314 * / "*" / "+" / "," / ";" / "="
316 static inline BOOL is_subdelim(WCHAR val) {
317 return (val == '!' || val == '$' || val == '&' ||
318 val == '\'' || val == '(' || val == ')' ||
319 val == '*' || val == '+' || val == ',' ||
320 val == ';' || val == '=');
323 /* gen-delims = ":" / "/" / "?" / "#" / "[" / "]" / "@" */
324 static inline BOOL is_gendelim(WCHAR val) {
325 return (val == ':' || val == '/' || val == '?' ||
326 val == '#' || val == '[' || val == ']' ||
330 /* Characters that delimit the end of the authority
331 * section of a URI. Sometimes a '\\' is considered
332 * an authority delimiter.
334 static inline BOOL is_auth_delim(WCHAR val, BOOL acceptSlash) {
335 return (val == '#' || val == '/' || val == '?' ||
336 val == '\0' || (acceptSlash && val == '\\'));
339 /* reserved = gen-delims / sub-delims */
340 static inline BOOL is_reserved(WCHAR val) {
341 return (is_subdelim(val) || is_gendelim(val));
344 static inline BOOL is_hexdigit(WCHAR val) {
345 return ((val >= 'a' && val <= 'f') ||
346 (val >= 'A' && val <= 'F') ||
347 (val >= '0' && val <= '9'));
350 static inline BOOL is_path_delim(WCHAR val) {
351 return (!val || val == '#' || val == '?');
354 static inline BOOL is_slash(WCHAR c)
356 return c == '/' || c == '\\';
359 static BOOL is_default_port(URL_SCHEME scheme, DWORD port) {
362 for(i = 0; i < sizeof(default_ports)/sizeof(default_ports[0]); ++i) {
363 if(default_ports[i].scheme == scheme && default_ports[i].port)
370 /* List of schemes types Windows seems to expect to be hierarchical. */
371 static inline BOOL is_hierarchical_scheme(URL_SCHEME type) {
372 return(type == URL_SCHEME_HTTP || type == URL_SCHEME_FTP ||
373 type == URL_SCHEME_GOPHER || type == URL_SCHEME_NNTP ||
374 type == URL_SCHEME_TELNET || type == URL_SCHEME_WAIS ||
375 type == URL_SCHEME_FILE || type == URL_SCHEME_HTTPS ||
376 type == URL_SCHEME_RES);
379 /* Checks if 'flags' contains an invalid combination of Uri_CREATE flags. */
380 static inline BOOL has_invalid_flag_combination(DWORD flags) {
381 return((flags & Uri_CREATE_DECODE_EXTRA_INFO && flags & Uri_CREATE_NO_DECODE_EXTRA_INFO) ||
382 (flags & Uri_CREATE_CANONICALIZE && flags & Uri_CREATE_NO_CANONICALIZE) ||
383 (flags & Uri_CREATE_CRACK_UNKNOWN_SCHEMES && flags & Uri_CREATE_NO_CRACK_UNKNOWN_SCHEMES) ||
384 (flags & Uri_CREATE_PRE_PROCESS_HTML_URI && flags & Uri_CREATE_NO_PRE_PROCESS_HTML_URI) ||
385 (flags & Uri_CREATE_IE_SETTINGS && flags & Uri_CREATE_NO_IE_SETTINGS));
388 /* Applies each default Uri_CREATE flags to 'flags' if it
389 * doesn't cause a flag conflict.
391 static void apply_default_flags(DWORD *flags) {
392 if(!(*flags & Uri_CREATE_NO_CANONICALIZE))
393 *flags |= Uri_CREATE_CANONICALIZE;
394 if(!(*flags & Uri_CREATE_NO_DECODE_EXTRA_INFO))
395 *flags |= Uri_CREATE_DECODE_EXTRA_INFO;
396 if(!(*flags & Uri_CREATE_NO_CRACK_UNKNOWN_SCHEMES))
397 *flags |= Uri_CREATE_CRACK_UNKNOWN_SCHEMES;
398 if(!(*flags & Uri_CREATE_NO_PRE_PROCESS_HTML_URI))
399 *flags |= Uri_CREATE_PRE_PROCESS_HTML_URI;
400 if(!(*flags & Uri_CREATE_IE_SETTINGS))
401 *flags |= Uri_CREATE_NO_IE_SETTINGS;
404 /* Determines if the URI is hierarchical using the information already parsed into
405 * data and using the current location of parsing in the URI string.
407 * Windows considers a URI hierarchical if one of the following is true:
408 * A.) It's a wildcard scheme.
409 * B.) It's an implicit file scheme.
410 * C.) It's a known hierarchical scheme and it has two '\\' after the scheme name.
411 * (the '\\' will be converted into "//" during canonicalization).
412 * D.) It's not a relative URI and "//" appears after the scheme name.
414 static inline BOOL is_hierarchical_uri(const WCHAR **ptr, const parse_data *data) {
415 const WCHAR *start = *ptr;
417 if(data->scheme_type == URL_SCHEME_WILDCARD)
419 else if(data->scheme_type == URL_SCHEME_FILE && data->has_implicit_scheme)
421 else if(is_hierarchical_scheme(data->scheme_type) && (*ptr)[0] == '\\' && (*ptr)[1] == '\\') {
424 } else if(!data->is_relative && check_hierarchical(ptr))
431 /* Checks if the two Uri's are logically equivalent. It's a simple
432 * comparison, since they are both of type Uri, and it can access
433 * the properties of each Uri directly without the need to go
434 * through the "IUri_Get*" interface calls.
436 static BOOL are_equal_simple(const Uri *a, const Uri *b) {
437 if(a->scheme_type == b->scheme_type) {
438 const BOOL known_scheme = a->scheme_type != URL_SCHEME_UNKNOWN;
439 const BOOL are_hierarchical =
440 (a->authority_start > -1 && b->authority_start > -1);
442 if(a->scheme_type == URL_SCHEME_FILE) {
443 if(a->canon_len == b->canon_len)
444 return !StrCmpIW(a->canon_uri, b->canon_uri);
447 /* Only compare the scheme names (if any) if their unknown scheme types. */
449 if((a->scheme_start > -1 && b->scheme_start > -1) &&
450 (a->scheme_len == b->scheme_len)) {
451 /* Make sure the schemes are the same. */
452 if(StrCmpNW(a->canon_uri+a->scheme_start, b->canon_uri+b->scheme_start, a->scheme_len))
454 } else if(a->scheme_len != b->scheme_len)
455 /* One of the Uri's has a scheme name, while the other doesn't. */
459 /* If they have a userinfo component, perform case sensitive compare. */
460 if((a->userinfo_start > -1 && b->userinfo_start > -1) &&
461 (a->userinfo_len == b->userinfo_len)) {
462 if(StrCmpNW(a->canon_uri+a->userinfo_start, b->canon_uri+b->userinfo_start, a->userinfo_len))
464 } else if(a->userinfo_len != b->userinfo_len)
465 /* One of the Uri's had a userinfo, while the other one doesn't. */
468 /* Check if they have a host name. */
469 if((a->host_start > -1 && b->host_start > -1) &&
470 (a->host_len == b->host_len)) {
471 /* Perform a case insensitive compare if they are a known scheme type. */
473 if(StrCmpNIW(a->canon_uri+a->host_start, b->canon_uri+b->host_start, a->host_len))
475 } else if(StrCmpNW(a->canon_uri+a->host_start, b->canon_uri+b->host_start, a->host_len))
477 } else if(a->host_len != b->host_len)
478 /* One of the Uri's had a host, while the other one didn't. */
481 if(a->has_port && b->has_port) {
482 if(a->port != b->port)
484 } else if(a->has_port || b->has_port)
485 /* One had a port, while the other one didn't. */
488 /* Windows is weird with how it handles paths. For example
489 * One URI could be "http://google.com" (after canonicalization)
490 * and one could be "http://google.com/" and the IsEqual function
491 * would still evaluate to TRUE, but, only if they are both hierarchical
494 if((a->path_start > -1 && b->path_start > -1) &&
495 (a->path_len == b->path_len)) {
496 if(StrCmpNW(a->canon_uri+a->path_start, b->canon_uri+b->path_start, a->path_len))
498 } else if(are_hierarchical && a->path_len == -1 && b->path_len == 0) {
499 if(*(a->canon_uri+a->path_start) != '/')
501 } else if(are_hierarchical && b->path_len == 1 && a->path_len == 0) {
502 if(*(b->canon_uri+b->path_start) != '/')
504 } else if(a->path_len != b->path_len)
507 /* Compare the query strings of the two URIs. */
508 if((a->query_start > -1 && b->query_start > -1) &&
509 (a->query_len == b->query_len)) {
510 if(StrCmpNW(a->canon_uri+a->query_start, b->canon_uri+b->query_start, a->query_len))
512 } else if(a->query_len != b->query_len)
515 if((a->fragment_start > -1 && b->fragment_start > -1) &&
516 (a->fragment_len == b->fragment_len)) {
517 if(StrCmpNW(a->canon_uri+a->fragment_start, b->canon_uri+b->fragment_start, a->fragment_len))
519 } else if(a->fragment_len != b->fragment_len)
522 /* If we get here, the two URIs are equivalent. */
529 /* Computes the size of the given IPv6 address.
530 * Each h16 component is 16 bits. If there is an IPv4 address, it's
531 * 32 bits. If there's an elision it can be 16 to 128 bits, depending
532 * on the number of other components.
534 * Modeled after google-url's CheckIPv6ComponentsSize function
536 static void compute_ipv6_comps_size(ipv6_address *address) {
537 address->components_size = address->h16_count * 2;
540 /* IPv4 address is 4 bytes. */
541 address->components_size += 4;
543 if(address->elision) {
544 /* An elision can be anywhere from 2 bytes up to 16 bytes.
545 * Its size depends on the size of the h16 and IPv4 components.
547 address->elision_size = 16 - address->components_size;
548 if(address->elision_size < 2)
549 address->elision_size = 2;
551 address->elision_size = 0;
554 /* Taken from dlls/jscript/lex.c */
555 static int hex_to_int(WCHAR val) {
556 if(val >= '0' && val <= '9')
558 else if(val >= 'a' && val <= 'f')
559 return val - 'a' + 10;
560 else if(val >= 'A' && val <= 'F')
561 return val - 'A' + 10;
566 /* Helper function for converting a percent encoded string
567 * representation of a WCHAR value into its actual WCHAR value. If
568 * the two characters following the '%' aren't valid hex values then
569 * this function returns the NULL character.
572 * "%2E" will result in '.' being returned by this function.
574 static WCHAR decode_pct_val(const WCHAR *ptr) {
577 if(*ptr == '%' && is_hexdigit(*(ptr + 1)) && is_hexdigit(*(ptr + 2))) {
578 INT a = hex_to_int(*(ptr + 1));
579 INT b = hex_to_int(*(ptr + 2));
588 /* Helper function for percent encoding a given character
589 * and storing the encoded value into a given buffer (dest).
591 * It's up to the calling function to ensure that there is
592 * at least enough space in 'dest' for the percent encoded
593 * value to be stored (so dest + 3 spaces available).
595 static inline void pct_encode_val(WCHAR val, WCHAR *dest) {
597 dest[1] = hexDigits[(val >> 4) & 0xf];
598 dest[2] = hexDigits[val & 0xf];
601 /* Attempts to parse the domain name from the host.
603 * This function also includes the Top-level Domain (TLD) name
604 * of the host when it tries to find the domain name. If it finds
605 * a valid domain name it will assign 'domain_start' the offset
606 * into 'host' where the domain name starts.
608 * It's implied that if there is a domain name its range is:
609 * [host+domain_start, host+host_len).
611 void find_domain_name(const WCHAR *host, DWORD host_len,
613 const WCHAR *last_tld, *sec_last_tld, *end;
615 end = host+host_len-1;
619 /* There has to be at least enough room for a '.' followed by a
620 * 3-character TLD for a domain to even exist in the host name.
625 last_tld = memrchrW(host, '.', host_len);
627 /* http://hostname -> has no domain name. */
630 sec_last_tld = memrchrW(host, '.', last_tld-host);
632 /* If the '.' is at the beginning of the host there
633 * has to be at least 3 characters in the TLD for it
635 * Ex: .com -> .com as the domain name.
636 * .co -> has no domain name.
638 if(last_tld-host == 0) {
639 if(end-(last_tld-1) < 3)
641 } else if(last_tld-host == 3) {
644 /* If there are three characters in front of last_tld and
645 * they are on the list of recognized TLDs, then this
646 * host doesn't have a domain (since the host only contains
648 * Ex: edu.uk -> has no domain name.
649 * foo.uk -> foo.uk as the domain name.
651 for(i = 0; i < sizeof(recognized_tlds)/sizeof(recognized_tlds[0]); ++i) {
652 if(!StrCmpNIW(host, recognized_tlds[i].tld_name, 3))
655 } else if(last_tld-host < 3)
656 /* Anything less than 3 characters is considered part
658 * Ex: ak.uk -> Has no domain name.
662 /* Otherwise the domain name is the whole host name. */
664 } else if(end+1-last_tld > 3) {
665 /* If the last_tld has more than 3 characters, then it's automatically
666 * considered the TLD of the domain name.
667 * Ex: www.winehq.org.uk.test -> uk.test as the domain name.
669 *domain_start = (sec_last_tld+1)-host;
670 } else if(last_tld - (sec_last_tld+1) < 4) {
672 /* If the sec_last_tld is 3 characters long it HAS to be on the list of
673 * recognized to still be considered part of the TLD name, otherwise
674 * its considered the domain name.
675 * Ex: www.google.com.uk -> google.com.uk as the domain name.
676 * www.google.foo.uk -> foo.uk as the domain name.
678 if(last_tld - (sec_last_tld+1) == 3) {
679 for(i = 0; i < sizeof(recognized_tlds)/sizeof(recognized_tlds[0]); ++i) {
680 if(!StrCmpNIW(sec_last_tld+1, recognized_tlds[i].tld_name, 3)) {
681 const WCHAR *domain = memrchrW(host, '.', sec_last_tld-host);
686 *domain_start = (domain+1) - host;
687 TRACE("Found domain name %s\n", debugstr_wn(host+*domain_start,
688 (host+host_len)-(host+*domain_start)));
693 *domain_start = (sec_last_tld+1)-host;
695 /* Since the sec_last_tld is less than 3 characters it's considered
697 * Ex: www.google.fo.uk -> google.fo.uk as the domain name.
699 const WCHAR *domain = memrchrW(host, '.', sec_last_tld-host);
704 *domain_start = (domain+1) - host;
707 /* The second to last TLD has more than 3 characters making it
709 * Ex: www.google.test.us -> test.us as the domain name.
711 *domain_start = (sec_last_tld+1)-host;
714 TRACE("Found domain name %s\n", debugstr_wn(host+*domain_start,
715 (host+host_len)-(host+*domain_start)));
718 /* Removes the dot segments from a hierarchical URIs path component. This
719 * function performs the removal in place.
721 * This function returns the new length of the path string.
723 static DWORD remove_dot_segments(WCHAR *path, DWORD path_len) {
725 const WCHAR *in = out;
726 const WCHAR *end = out + path_len;
730 /* Move the first path segment in the input buffer to the end of
731 * the output buffer, and any subsequent characters up to, including
732 * the next "/" character (if any) or the end of the input buffer.
734 while(in < end && !is_slash(*in))
744 /* Handle ending "/." */
751 if(is_slash(in[1])) {
756 /* If we don't have "/../" or ending "/.." */
757 if(in[1] != '.' || (in + 2 != end && !is_slash(in[2])))
760 /* Find the slash preceding out pointer and move out pointer to it */
761 if(out > path+1 && is_slash(*--out))
763 while(out > path && !is_slash(*(--out)));
773 TRACE("(%p %d): Path after dot segments removed %s len=%d\n", path, path_len,
774 debugstr_wn(path, len), len);
778 /* Attempts to find the file extension in a given path. */
779 static INT find_file_extension(const WCHAR *path, DWORD path_len) {
782 for(end = path+path_len-1; end >= path && *end != '/' && *end != '\\'; --end) {
790 /* Computes the location where the elision should occur in the IPv6
791 * address using the numerical values of each component stored in
792 * 'values'. If the address shouldn't contain an elision then 'index'
793 * is assigned -1 as its value. Otherwise 'index' will contain the
794 * starting index (into values) where the elision should be, and 'count'
795 * will contain the number of cells the elision covers.
798 * Windows will expand an elision if the elision only represents one h16
799 * component of the address.
801 * Ex: [1::2:3:4:5:6:7] -> [1:0:2:3:4:5:6:7]
803 * If the IPv6 address contains an IPv4 address, the IPv4 address is also
804 * considered for being included as part of an elision if all its components
807 * Ex: [1:2:3:4:5:6:0.0.0.0] -> [1:2:3:4:5:6::]
809 static void compute_elision_location(const ipv6_address *address, const USHORT values[8],
810 INT *index, DWORD *count) {
811 DWORD i, max_len, cur_len;
812 INT max_index, cur_index;
814 max_len = cur_len = 0;
815 max_index = cur_index = -1;
816 for(i = 0; i < 8; ++i) {
817 BOOL check_ipv4 = (address->ipv4 && i == 6);
818 BOOL is_end = (check_ipv4 || i == 7);
821 /* Check if the IPv4 address contains only zeros. */
822 if(values[i] == 0 && values[i+1] == 0) {
829 } else if(values[i] == 0) {
836 if(is_end || values[i] != 0) {
837 /* We only consider it for an elision if it's
838 * more than 1 component long.
840 if(cur_len > 1 && cur_len > max_len) {
841 /* Found the new elision location. */
843 max_index = cur_index;
846 /* Reset the current range for the next range of zeros. */
856 /* Removes all the leading and trailing white spaces or
857 * control characters from the URI and removes all control
858 * characters inside of the URI string.
860 static BSTR pre_process_uri(LPCWSTR uri) {
861 const WCHAR *start, *end, *ptr;
867 /* Skip leading controls and whitespace. */
868 while(*start && (iscntrlW(*start) || isspaceW(*start))) ++start;
870 /* URI consisted only of control/whitespace. */
872 return SysAllocStringLen(NULL, 0);
874 end = start + strlenW(start);
875 while(--end > start && (iscntrlW(*end) || isspaceW(*end)));
878 for(ptr = start; ptr < end; ptr++) {
883 ret = SysAllocStringLen(NULL, len);
887 for(ptr = start, ptr2=ret; ptr < end; ptr++) {
895 /* Converts the specified IPv4 address into an uint value.
897 * This function assumes that the IPv4 address has already been validated.
899 static UINT ipv4toui(const WCHAR *ip, DWORD len) {
901 DWORD comp_value = 0;
904 for(ptr = ip; ptr < ip+len; ++ptr) {
910 comp_value = comp_value*10 + (*ptr-'0');
919 /* Converts an IPv4 address in numerical form into its fully qualified
920 * string form. This function returns the number of characters written
921 * to 'dest'. If 'dest' is NULL this function will return the number of
922 * characters that would have been written.
924 * It's up to the caller to ensure there's enough space in 'dest' for the
927 static DWORD ui2ipv4(WCHAR *dest, UINT address) {
928 static const WCHAR formatW[] =
929 {'%','u','.','%','u','.','%','u','.','%','u',0};
933 digits[0] = (address >> 24) & 0xff;
934 digits[1] = (address >> 16) & 0xff;
935 digits[2] = (address >> 8) & 0xff;
936 digits[3] = address & 0xff;
940 ret = sprintfW(tmp, formatW, digits[0], digits[1], digits[2], digits[3]);
942 ret = sprintfW(dest, formatW, digits[0], digits[1], digits[2], digits[3]);
947 static DWORD ui2str(WCHAR *dest, UINT value) {
948 static const WCHAR formatW[] = {'%','u',0};
953 ret = sprintfW(tmp, formatW, value);
955 ret = sprintfW(dest, formatW, value);
960 /* Converts a h16 component (from an IPv6 address) into its
963 * This function assumes that the h16 component has already been validated.
965 static USHORT h16tous(h16 component) {
969 for(i = 0; i < component.len; ++i) {
971 ret += hex_to_int(component.str[i]);
977 /* Converts an IPv6 address into its 128 bits (16 bytes) numerical value.
979 * This function assumes that the ipv6_address has already been validated.
981 static BOOL ipv6_to_number(const ipv6_address *address, USHORT number[8]) {
982 DWORD i, cur_component = 0;
983 BOOL already_passed_elision = FALSE;
985 for(i = 0; i < address->h16_count; ++i) {
986 if(address->elision) {
987 if(address->components[i].str > address->elision && !already_passed_elision) {
988 /* Means we just passed the elision and need to add its values to
989 * 'number' before we do anything else.
992 for(j = 0; j < address->elision_size; j+=2)
993 number[cur_component++] = 0;
995 already_passed_elision = TRUE;
999 number[cur_component++] = h16tous(address->components[i]);
1002 /* Case when the elision appears after the h16 components. */
1003 if(!already_passed_elision && address->elision) {
1004 for(i = 0; i < address->elision_size; i+=2)
1005 number[cur_component++] = 0;
1009 UINT value = ipv4toui(address->ipv4, address->ipv4_len);
1011 if(cur_component != 6) {
1012 ERR("(%p %p): Failed sanity check with %d\n", address, number, cur_component);
1016 number[cur_component++] = (value >> 16) & 0xffff;
1017 number[cur_component] = value & 0xffff;
1023 /* Checks if the characters pointed to by 'ptr' are
1024 * a percent encoded data octet.
1026 * pct-encoded = "%" HEXDIG HEXDIG
1028 static BOOL check_pct_encoded(const WCHAR **ptr) {
1029 const WCHAR *start = *ptr;
1035 if(!is_hexdigit(**ptr)) {
1041 if(!is_hexdigit(**ptr)) {
1050 /* dec-octet = DIGIT ; 0-9
1051 * / %x31-39 DIGIT ; 10-99
1052 * / "1" 2DIGIT ; 100-199
1053 * / "2" %x30-34 DIGIT ; 200-249
1054 * / "25" %x30-35 ; 250-255
1056 static BOOL check_dec_octet(const WCHAR **ptr) {
1057 const WCHAR *c1, *c2, *c3;
1060 /* A dec-octet must be at least 1 digit long. */
1061 if(*c1 < '0' || *c1 > '9')
1067 /* Since the 1-digit requirement was met, it doesn't
1068 * matter if this is a DIGIT value, it's considered a
1071 if(*c2 < '0' || *c2 > '9')
1077 /* Same explanation as above. */
1078 if(*c3 < '0' || *c3 > '9')
1081 /* Anything > 255 isn't a valid IP dec-octet. */
1082 if(*c1 >= '2' && *c2 >= '5' && *c3 >= '5') {
1091 /* Checks if there is an implicit IPv4 address in the host component of the URI.
1092 * The max value of an implicit IPv4 address is UINT_MAX.
1095 * "234567" would be considered an implicit IPv4 address.
1097 static BOOL check_implicit_ipv4(const WCHAR **ptr, UINT *val) {
1098 const WCHAR *start = *ptr;
1102 while(is_num(**ptr)) {
1103 ret = ret*10 + (**ptr - '0');
1105 if(ret > UINT_MAX) {
1119 /* Checks if the string contains an IPv4 address.
1121 * This function has a strict mode or a non-strict mode of operation
1122 * When 'strict' is set to FALSE this function will return TRUE if
1123 * the string contains at least 'dec-octet "." dec-octet' since partial
1124 * IPv4 addresses will be normalized out into full IPv4 addresses. When
1125 * 'strict' is set this function expects there to be a full IPv4 address.
1127 * IPv4address = dec-octet "." dec-octet "." dec-octet "." dec-octet
1129 static BOOL check_ipv4address(const WCHAR **ptr, BOOL strict) {
1130 const WCHAR *start = *ptr;
1132 if(!check_dec_octet(ptr)) {
1143 if(!check_dec_octet(ptr)) {
1157 if(!check_dec_octet(ptr)) {
1171 if(!check_dec_octet(ptr)) {
1176 /* Found a four digit ip address. */
1179 /* Tries to parse the scheme name of the URI.
1181 * scheme = ALPHA *(ALPHA | NUM | '+' | '-' | '.') as defined by RFC 3896.
1182 * NOTE: Windows accepts a number as the first character of a scheme.
1184 static BOOL parse_scheme_name(const WCHAR **ptr, parse_data *data, DWORD extras) {
1185 const WCHAR *start = *ptr;
1187 data->scheme = NULL;
1188 data->scheme_len = 0;
1191 if(**ptr == '*' && *ptr == start) {
1192 /* Might have found a wildcard scheme. If it is the next
1193 * char has to be a ':' for it to be a valid URI
1197 } else if(!is_num(**ptr) && !is_alpha(**ptr) && **ptr != '+' &&
1198 **ptr != '-' && **ptr != '.')
1207 /* Schemes must end with a ':' */
1208 if(**ptr != ':' && !((extras & ALLOW_NULL_TERM_SCHEME) && !**ptr)) {
1213 data->scheme = start;
1214 data->scheme_len = *ptr - start;
1220 /* Tries to deduce the corresponding URL_SCHEME for the given URI. Stores
1221 * the deduced URL_SCHEME in data->scheme_type.
1223 static BOOL parse_scheme_type(parse_data *data) {
1224 /* If there's scheme data then see if it's a recognized scheme. */
1225 if(data->scheme && data->scheme_len) {
1228 for(i = 0; i < sizeof(recognized_schemes)/sizeof(recognized_schemes[0]); ++i) {
1229 if(lstrlenW(recognized_schemes[i].scheme_name) == data->scheme_len) {
1230 /* Has to be a case insensitive compare. */
1231 if(!StrCmpNIW(recognized_schemes[i].scheme_name, data->scheme, data->scheme_len)) {
1232 data->scheme_type = recognized_schemes[i].scheme;
1238 /* If we get here it means it's not a recognized scheme. */
1239 data->scheme_type = URL_SCHEME_UNKNOWN;
1241 } else if(data->is_relative) {
1242 /* Relative URI's have no scheme. */
1243 data->scheme_type = URL_SCHEME_UNKNOWN;
1246 /* Should never reach here! what happened... */
1247 FIXME("(%p): Unable to determine scheme type for URI %s\n", data, debugstr_w(data->uri));
1252 /* Tries to parse (or deduce) the scheme_name of a URI. If it can't
1253 * parse a scheme from the URI it will try to deduce the scheme_name and scheme_type
1254 * using the flags specified in 'flags' (if any). Flags that affect how this function
1255 * operates are the Uri_CREATE_ALLOW_* flags.
1257 * All parsed/deduced information will be stored in 'data' when the function returns.
1259 * Returns TRUE if it was able to successfully parse the information.
1261 static BOOL parse_scheme(const WCHAR **ptr, parse_data *data, DWORD flags, DWORD extras) {
1262 static const WCHAR fileW[] = {'f','i','l','e',0};
1263 static const WCHAR wildcardW[] = {'*',0};
1265 /* First check to see if the uri could implicitly be a file path. */
1266 if(is_implicit_file_path(*ptr)) {
1267 if(flags & Uri_CREATE_ALLOW_IMPLICIT_FILE_SCHEME) {
1268 data->scheme = fileW;
1269 data->scheme_len = lstrlenW(fileW);
1270 data->has_implicit_scheme = TRUE;
1272 TRACE("(%p %p %x): URI is an implicit file path.\n", ptr, data, flags);
1274 /* Windows does not consider anything that can implicitly be a file
1275 * path to be a valid URI if the ALLOW_IMPLICIT_FILE_SCHEME flag is not set...
1277 TRACE("(%p %p %x): URI is implicitly a file path, but, the ALLOW_IMPLICIT_FILE_SCHEME flag wasn't set.\n",
1281 } else if(!parse_scheme_name(ptr, data, extras)) {
1282 /* No scheme was found, this means it could be:
1283 * a) an implicit Wildcard scheme
1285 * c) an invalid URI.
1287 if(flags & Uri_CREATE_ALLOW_IMPLICIT_WILDCARD_SCHEME) {
1288 data->scheme = wildcardW;
1289 data->scheme_len = lstrlenW(wildcardW);
1290 data->has_implicit_scheme = TRUE;
1292 TRACE("(%p %p %x): URI is an implicit wildcard scheme.\n", ptr, data, flags);
1293 } else if (flags & Uri_CREATE_ALLOW_RELATIVE) {
1294 data->is_relative = TRUE;
1295 TRACE("(%p %p %x): URI is relative.\n", ptr, data, flags);
1297 TRACE("(%p %p %x): Malformed URI found. Unable to deduce scheme name.\n", ptr, data, flags);
1302 if(!data->is_relative)
1303 TRACE("(%p %p %x): Found scheme=%s scheme_len=%d\n", ptr, data, flags,
1304 debugstr_wn(data->scheme, data->scheme_len), data->scheme_len);
1306 if(!parse_scheme_type(data))
1309 TRACE("(%p %p %x): Assigned %d as the URL_SCHEME.\n", ptr, data, flags, data->scheme_type);
1313 static BOOL parse_username(const WCHAR **ptr, parse_data *data, DWORD flags, DWORD extras) {
1314 data->username = *ptr;
1316 while(**ptr != ':' && **ptr != '@') {
1318 if(!check_pct_encoded(ptr)) {
1319 if(data->scheme_type != URL_SCHEME_UNKNOWN) {
1320 *ptr = data->username;
1321 data->username = NULL;
1326 } else if(extras & ALLOW_NULL_TERM_USER_NAME && !**ptr)
1328 else if(is_auth_delim(**ptr, data->scheme_type != URL_SCHEME_UNKNOWN)) {
1329 *ptr = data->username;
1330 data->username = NULL;
1337 data->username_len = *ptr - data->username;
1341 static BOOL parse_password(const WCHAR **ptr, parse_data *data, DWORD flags, DWORD extras) {
1342 data->password = *ptr;
1344 while(**ptr != '@') {
1346 if(!check_pct_encoded(ptr)) {
1347 if(data->scheme_type != URL_SCHEME_UNKNOWN) {
1348 *ptr = data->password;
1349 data->password = NULL;
1354 } else if(extras & ALLOW_NULL_TERM_PASSWORD && !**ptr)
1356 else if(is_auth_delim(**ptr, data->scheme_type != URL_SCHEME_UNKNOWN)) {
1357 *ptr = data->password;
1358 data->password = NULL;
1365 data->password_len = *ptr - data->password;
1369 /* Parses the userinfo part of the URI (if it exists). The userinfo field of
1370 * a URI can consist of "username:password@", or just "username@".
1373 * userinfo = *( unreserved / pct-encoded / sub-delims / ":" )
1376 * 1) If there is more than one ':' in the userinfo part of the URI Windows
1377 * uses the first occurrence of ':' to delimit the username and password
1381 * ftp://user:pass:word@winehq.org
1383 * would yield "user" as the username and "pass:word" as the password.
1385 * 2) Windows allows any character to appear in the "userinfo" part of
1386 * a URI, as long as it's not an authority delimiter character set.
1388 static void parse_userinfo(const WCHAR **ptr, parse_data *data, DWORD flags) {
1389 const WCHAR *start = *ptr;
1391 if(!parse_username(ptr, data, flags, 0)) {
1392 TRACE("(%p %p %x): URI contained no userinfo.\n", ptr, data, flags);
1398 if(!parse_password(ptr, data, flags, 0)) {
1400 data->username = NULL;
1401 data->username_len = 0;
1402 TRACE("(%p %p %x): URI contained no userinfo.\n", ptr, data, flags);
1409 data->username = NULL;
1410 data->username_len = 0;
1411 data->password = NULL;
1412 data->password_len = 0;
1414 TRACE("(%p %p %x): URI contained no userinfo.\n", ptr, data, flags);
1419 TRACE("(%p %p %x): Found username %s len=%d.\n", ptr, data, flags,
1420 debugstr_wn(data->username, data->username_len), data->username_len);
1423 TRACE("(%p %p %x): Found password %s len=%d.\n", ptr, data, flags,
1424 debugstr_wn(data->password, data->password_len), data->password_len);
1429 /* Attempts to parse a port from the URI.
1432 * Windows seems to have a cap on what the maximum value
1433 * for a port can be. The max value is USHORT_MAX.
1437 static BOOL parse_port(const WCHAR **ptr, parse_data *data, DWORD flags) {
1441 while(!is_auth_delim(**ptr, data->scheme_type != URL_SCHEME_UNKNOWN)) {
1442 if(!is_num(**ptr)) {
1448 port = port*10 + (**ptr-'0');
1450 if(port > USHORT_MAX) {
1459 data->has_port = TRUE;
1460 data->port_value = port;
1461 data->port_len = *ptr - data->port;
1463 TRACE("(%p %p %x): Found port %s len=%d value=%u\n", ptr, data, flags,
1464 debugstr_wn(data->port, data->port_len), data->port_len, data->port_value);
1468 /* Attempts to parse a IPv4 address from the URI.
1471 * Windows normalizes IPv4 addresses, This means there are three
1472 * possibilities for the URI to contain an IPv4 address.
1473 * 1) A well formed address (ex. 192.2.2.2).
1474 * 2) A partially formed address. For example "192.0" would
1475 * normalize to "192.0.0.0" during canonicalization.
1476 * 3) An implicit IPv4 address. For example "256" would
1477 * normalize to "0.0.1.0" during canonicalization. Also
1478 * note that the maximum value for an implicit IP address
1479 * is UINT_MAX, if the value in the URI exceeds this then
1480 * it is not considered an IPv4 address.
1482 static BOOL parse_ipv4address(const WCHAR **ptr, parse_data *data, DWORD flags) {
1483 const BOOL is_unknown = data->scheme_type == URL_SCHEME_UNKNOWN;
1486 if(!check_ipv4address(ptr, FALSE)) {
1487 if(!check_implicit_ipv4(ptr, &data->implicit_ipv4)) {
1488 TRACE("(%p %p %x): URI didn't contain anything looking like an IPv4 address.\n",
1494 data->has_implicit_ip = TRUE;
1497 data->host_len = *ptr - data->host;
1498 data->host_type = Uri_HOST_IPV4;
1500 /* Check if what we found is the only part of the host name (if it isn't
1501 * we don't have an IPv4 address).
1505 if(!parse_port(ptr, data, flags)) {
1510 } else if(!is_auth_delim(**ptr, !is_unknown)) {
1511 /* Found more data which belongs to the host, so this isn't an IPv4. */
1514 data->has_implicit_ip = FALSE;
1518 TRACE("(%p %p %x): IPv4 address found. host=%s host_len=%d host_type=%d\n",
1519 ptr, data, flags, debugstr_wn(data->host, data->host_len),
1520 data->host_len, data->host_type);
1524 /* Attempts to parse the reg-name from the URI.
1526 * Because of the way Windows handles ':' this function also
1527 * handles parsing the port.
1529 * reg-name = *( unreserved / pct-encoded / sub-delims )
1532 * Windows allows everything, but, the characters in "auth_delims" and ':'
1533 * to appear in a reg-name, unless it's an unknown scheme type then ':' is
1534 * allowed to appear (even if a valid port isn't after it).
1536 * Windows doesn't like host names which start with '[' and end with ']'
1537 * and don't contain a valid IP literal address in between them.
1539 * On Windows if a '[' is encountered in the host name the ':' no longer
1540 * counts as a delimiter until you reach the next ']' or an "authority delimiter".
1542 * A reg-name CAN be empty.
1544 static BOOL parse_reg_name(const WCHAR **ptr, parse_data *data, DWORD flags, DWORD extras) {
1545 const BOOL has_start_bracket = **ptr == '[';
1546 const BOOL known_scheme = data->scheme_type != URL_SCHEME_UNKNOWN;
1547 const BOOL is_res = data->scheme_type == URL_SCHEME_RES;
1548 BOOL inside_brackets = has_start_bracket;
1550 /* res URIs don't have ports. */
1551 BOOL ignore_col = (extras & IGNORE_PORT_DELIMITER) || is_res;
1553 /* We have to be careful with file schemes. */
1554 if(data->scheme_type == URL_SCHEME_FILE) {
1555 /* This is because an implicit file scheme could be "C:\\test" and it
1556 * would trick this function into thinking the host is "C", when after
1557 * canonicalization the host would end up being an empty string. A drive
1558 * path can also have a '|' instead of a ':' after the drive letter.
1560 if(is_drive_path(*ptr)) {
1561 /* Regular old drive paths have no host type (or host name). */
1562 data->host_type = Uri_HOST_UNKNOWN;
1566 } else if(is_unc_path(*ptr))
1567 /* Skip past the "\\" of a UNC path. */
1573 /* For res URIs, everything before the first '/' is
1574 * considered the host.
1576 while((!is_res && !is_auth_delim(**ptr, known_scheme)) ||
1577 (is_res && **ptr && **ptr != '/')) {
1578 if(**ptr == ':' && !ignore_col) {
1579 /* We can ignore ':' if were inside brackets.*/
1580 if(!inside_brackets) {
1581 const WCHAR *tmp = (*ptr)++;
1583 /* Attempt to parse the port. */
1584 if(!parse_port(ptr, data, flags)) {
1585 /* Windows expects there to be a valid port for known scheme types. */
1586 if(data->scheme_type != URL_SCHEME_UNKNOWN) {
1589 TRACE("(%p %p %x %x): Expected valid port\n", ptr, data, flags, extras);
1592 /* Windows gives up on trying to parse a port when it
1593 * encounters an invalid port.
1597 data->host_len = tmp - data->host;
1601 } else if(**ptr == '%' && (known_scheme && !is_res)) {
1602 /* Has to be a legit % encoded value. */
1603 if(!check_pct_encoded(ptr)) {
1609 } else if(is_res && is_forbidden_dos_path_char(**ptr)) {
1613 } else if(**ptr == ']')
1614 inside_brackets = FALSE;
1615 else if(**ptr == '[')
1616 inside_brackets = TRUE;
1621 if(has_start_bracket) {
1622 /* Make sure the last character of the host wasn't a ']'. */
1623 if(*(*ptr-1) == ']') {
1624 TRACE("(%p %p %x %x): Expected an IP literal inside of the host\n",
1625 ptr, data, flags, extras);
1632 /* Don't overwrite our length if we found a port earlier. */
1634 data->host_len = *ptr - data->host;
1636 /* If the host is empty, then it's an unknown host type. */
1637 if(data->host_len == 0 || is_res)
1638 data->host_type = Uri_HOST_UNKNOWN;
1640 data->host_type = Uri_HOST_DNS;
1642 TRACE("(%p %p %x %x): Parsed reg-name. host=%s len=%d\n", ptr, data, flags, extras,
1643 debugstr_wn(data->host, data->host_len), data->host_len);
1647 /* Attempts to parse an IPv6 address out of the URI.
1649 * IPv6address = 6( h16 ":" ) ls32
1650 * / "::" 5( h16 ":" ) ls32
1651 * / [ h16 ] "::" 4( h16 ":" ) ls32
1652 * / [ *1( h16 ":" ) h16 ] "::" 3( h16 ":" ) ls32
1653 * / [ *2( h16 ":" ) h16 ] "::" 2( h16 ":" ) ls32
1654 * / [ *3( h16 ":" ) h16 ] "::" h16 ":" ls32
1655 * / [ *4( h16 ":" ) h16 ] "::" ls32
1656 * / [ *5( h16 ":" ) h16 ] "::" h16
1657 * / [ *6( h16 ":" ) h16 ] "::"
1659 * ls32 = ( h16 ":" h16 ) / IPv4address
1660 * ; least-significant 32 bits of address.
1663 * ; 16 bits of address represented in hexadecimal.
1665 * Modeled after google-url's 'DoParseIPv6' function.
1667 static BOOL parse_ipv6address(const WCHAR **ptr, parse_data *data, DWORD flags) {
1668 const WCHAR *start, *cur_start;
1671 start = cur_start = *ptr;
1672 memset(&ip, 0, sizeof(ipv6_address));
1675 /* Check if we're on the last character of the host. */
1676 BOOL is_end = (is_auth_delim(**ptr, data->scheme_type != URL_SCHEME_UNKNOWN)
1679 BOOL is_split = (**ptr == ':');
1680 BOOL is_elision = (is_split && !is_end && *(*ptr+1) == ':');
1682 /* Check if we're at the end of a component, or
1683 * if we're at the end of the IPv6 address.
1685 if(is_split || is_end) {
1688 cur_len = *ptr - cur_start;
1690 /* h16 can't have a length > 4. */
1694 TRACE("(%p %p %x): h16 component to long.\n",
1700 /* An h16 component can't have the length of 0 unless
1701 * the elision is at the beginning of the address, or
1702 * at the end of the address.
1704 if(!((*ptr == start && is_elision) ||
1705 (is_end && (*ptr-2) == ip.elision))) {
1707 TRACE("(%p %p %x): IPv6 component cannot have a length of 0.\n",
1714 /* An IPv6 address can have no more than 8 h16 components. */
1715 if(ip.h16_count >= 8) {
1717 TRACE("(%p %p %x): Not a IPv6 address, to many h16 components.\n",
1722 ip.components[ip.h16_count].str = cur_start;
1723 ip.components[ip.h16_count].len = cur_len;
1725 TRACE("(%p %p %x): Found h16 component %s, len=%d, h16_count=%d\n",
1726 ptr, data, flags, debugstr_wn(cur_start, cur_len), cur_len,
1736 /* A IPv6 address can only have 1 elision ('::'). */
1740 TRACE("(%p %p %x): IPv6 address cannot have 2 elisions.\n",
1752 if(!check_ipv4address(ptr, TRUE)) {
1753 if(!is_hexdigit(**ptr)) {
1754 /* Not a valid character for an IPv6 address. */
1759 /* Found an IPv4 address. */
1760 ip.ipv4 = cur_start;
1761 ip.ipv4_len = *ptr - cur_start;
1763 TRACE("(%p %p %x): Found an attached IPv4 address %s len=%d.\n",
1764 ptr, data, flags, debugstr_wn(ip.ipv4, ip.ipv4_len),
1767 /* IPv4 addresses can only appear at the end of a IPv6. */
1773 compute_ipv6_comps_size(&ip);
1775 /* Make sure the IPv6 address adds up to 16 bytes. */
1776 if(ip.components_size + ip.elision_size != 16) {
1778 TRACE("(%p %p %x): Invalid IPv6 address, did not add up to 16 bytes.\n",
1783 if(ip.elision_size == 2) {
1784 /* For some reason on Windows if an elision that represents
1785 * only one h16 component is encountered at the very begin or
1786 * end of an IPv6 address, Windows does not consider it a
1787 * valid IPv6 address.
1789 * Ex: [::2:3:4:5:6:7] is not valid, even though the sum
1790 * of all the components == 128bits.
1792 if(ip.elision < ip.components[0].str ||
1793 ip.elision > ip.components[ip.h16_count-1].str) {
1795 TRACE("(%p %p %x): Invalid IPv6 address. Detected elision of 2 bytes at the beginning or end of the address.\n",
1801 data->host_type = Uri_HOST_IPV6;
1802 data->has_ipv6 = TRUE;
1803 data->ipv6_address = ip;
1805 TRACE("(%p %p %x): Found valid IPv6 literal %s len=%d\n",
1806 ptr, data, flags, debugstr_wn(start, *ptr-start),
1811 /* IPvFuture = "v" 1*HEXDIG "." 1*( unreserved / sub-delims / ":" ) */
1812 static BOOL parse_ipvfuture(const WCHAR **ptr, parse_data *data, DWORD flags) {
1813 const WCHAR *start = *ptr;
1815 /* IPvFuture has to start with a 'v' or 'V'. */
1816 if(**ptr != 'v' && **ptr != 'V')
1819 /* Following the v there must be at least 1 hex digit. */
1821 if(!is_hexdigit(**ptr)) {
1827 while(is_hexdigit(**ptr))
1830 /* End of the hexdigit sequence must be a '.' */
1837 if(!is_unreserved(**ptr) && !is_subdelim(**ptr) && **ptr != ':') {
1843 while(is_unreserved(**ptr) || is_subdelim(**ptr) || **ptr == ':')
1846 data->host_type = Uri_HOST_UNKNOWN;
1848 TRACE("(%p %p %x): Parsed IPvFuture address %s len=%d\n", ptr, data, flags,
1849 debugstr_wn(start, *ptr-start), (int)(*ptr-start));
1854 /* IP-literal = "[" ( IPv6address / IPvFuture ) "]" */
1855 static BOOL parse_ip_literal(const WCHAR **ptr, parse_data *data, DWORD flags, DWORD extras) {
1858 if(**ptr != '[' && !(extras & ALLOW_BRACKETLESS_IP_LITERAL)) {
1861 } else if(**ptr == '[')
1864 if(!parse_ipv6address(ptr, data, flags)) {
1865 if(extras & SKIP_IP_FUTURE_CHECK || !parse_ipvfuture(ptr, data, flags)) {
1872 if(**ptr != ']' && !(extras & ALLOW_BRACKETLESS_IP_LITERAL)) {
1876 } else if(!**ptr && extras & ALLOW_BRACKETLESS_IP_LITERAL) {
1877 /* The IP literal didn't contain brackets and was followed by
1878 * a NULL terminator, so no reason to even check the port.
1880 data->host_len = *ptr - data->host;
1887 /* If a valid port is not found, then let it trickle down to
1890 if(!parse_port(ptr, data, flags)) {
1896 data->host_len = *ptr - data->host;
1901 /* Parses the host information from the URI.
1903 * host = IP-literal / IPv4address / reg-name
1905 static BOOL parse_host(const WCHAR **ptr, parse_data *data, DWORD flags, DWORD extras) {
1906 if(!parse_ip_literal(ptr, data, flags, extras)) {
1907 if(!parse_ipv4address(ptr, data, flags)) {
1908 if(!parse_reg_name(ptr, data, flags, extras)) {
1909 TRACE("(%p %p %x %x): Malformed URI, Unknown host type.\n",
1910 ptr, data, flags, extras);
1919 /* Parses the authority information from the URI.
1921 * authority = [ userinfo "@" ] host [ ":" port ]
1923 static BOOL parse_authority(const WCHAR **ptr, parse_data *data, DWORD flags) {
1924 parse_userinfo(ptr, data, flags);
1926 /* Parsing the port will happen during one of the host parsing
1927 * routines (if the URI has a port).
1929 if(!parse_host(ptr, data, flags, 0))
1935 /* Attempts to parse the path information of a hierarchical URI. */
1936 static BOOL parse_path_hierarchical(const WCHAR **ptr, parse_data *data, DWORD flags) {
1937 const WCHAR *start = *ptr;
1938 static const WCHAR slash[] = {'/',0};
1939 const BOOL is_file = data->scheme_type == URL_SCHEME_FILE;
1941 if(is_path_delim(**ptr)) {
1942 if(data->scheme_type == URL_SCHEME_WILDCARD && !data->must_have_path) {
1945 } else if(!(flags & Uri_CREATE_NO_CANONICALIZE)) {
1946 /* If the path component is empty, then a '/' is added. */
1951 while(!is_path_delim(**ptr)) {
1952 if(**ptr == '%' && data->scheme_type != URL_SCHEME_UNKNOWN && !is_file) {
1953 if(!check_pct_encoded(ptr)) {
1958 } else if(is_forbidden_dos_path_char(**ptr) && is_file &&
1959 (flags & Uri_CREATE_FILE_USE_DOS_PATH)) {
1960 /* File schemes with USE_DOS_PATH set aren't allowed to have
1961 * a '<' or '>' or '\"' appear in them.
1965 } else if(**ptr == '\\') {
1966 /* Not allowed to have a backslash if NO_CANONICALIZE is set
1967 * and the scheme is known type (but not a file scheme).
1969 if(flags & Uri_CREATE_NO_CANONICALIZE) {
1970 if(data->scheme_type != URL_SCHEME_FILE &&
1971 data->scheme_type != URL_SCHEME_UNKNOWN) {
1981 /* The only time a URI doesn't have a path is when
1982 * the NO_CANONICALIZE flag is set and the raw URI
1983 * didn't contain one.
1990 data->path_len = *ptr - start;
1995 TRACE("(%p %p %x): Parsed path %s len=%d\n", ptr, data, flags,
1996 debugstr_wn(data->path, data->path_len), data->path_len);
1998 TRACE("(%p %p %x): The URI contained no path\n", ptr, data, flags);
2003 /* Parses the path of an opaque URI (much less strict then the parser
2004 * for a hierarchical URI).
2007 * Windows allows invalid % encoded data to appear in opaque URI paths
2008 * for unknown scheme types.
2010 * File schemes with USE_DOS_PATH set aren't allowed to have '<', '>', or '\"'
2013 static BOOL parse_path_opaque(const WCHAR **ptr, parse_data *data, DWORD flags) {
2014 const BOOL known_scheme = data->scheme_type != URL_SCHEME_UNKNOWN;
2015 const BOOL is_file = data->scheme_type == URL_SCHEME_FILE;
2019 while(!is_path_delim(**ptr)) {
2020 if(**ptr == '%' && known_scheme) {
2021 if(!check_pct_encoded(ptr)) {
2027 } else if(is_forbidden_dos_path_char(**ptr) && is_file &&
2028 (flags & Uri_CREATE_FILE_USE_DOS_PATH)) {
2037 data->path_len = *ptr - data->path;
2038 TRACE("(%p %p %x): Parsed opaque URI path %s len=%d\n", ptr, data, flags,
2039 debugstr_wn(data->path, data->path_len), data->path_len);
2043 /* Determines how the URI should be parsed after the scheme information.
2045 * If the scheme is followed by "//", then it is treated as a hierarchical URI
2046 * which then the authority and path information will be parsed out. Otherwise, the
2047 * URI will be treated as an opaque URI which the authority information is not parsed
2050 * RFC 3896 definition of hier-part:
2052 * hier-part = "//" authority path-abempty
2057 * MSDN opaque URI definition:
2058 * scheme ":" path [ "#" fragment ]
2061 * If the URI is of an unknown scheme type and has a "//" following the scheme then it
2062 * is treated as a hierarchical URI, but, if the CREATE_NO_CRACK_UNKNOWN_SCHEMES flag is
2063 * set then it is considered an opaque URI regardless of what follows the scheme information
2064 * (per MSDN documentation).
2066 static BOOL parse_hierpart(const WCHAR **ptr, parse_data *data, DWORD flags) {
2067 const WCHAR *start = *ptr;
2069 data->must_have_path = FALSE;
2071 /* For javascript: URIs, simply set everything as a path */
2072 if(data->scheme_type == URL_SCHEME_JAVASCRIPT) {
2074 data->path_len = strlenW(*ptr);
2075 data->is_opaque = TRUE;
2076 *ptr += data->path_len;
2080 /* Checks if the authority information needs to be parsed. */
2081 if(is_hierarchical_uri(ptr, data)) {
2082 /* Only treat it as a hierarchical URI if the scheme_type is known or
2083 * the Uri_CREATE_NO_CRACK_UNKNOWN_SCHEMES flag is not set.
2085 if(data->scheme_type != URL_SCHEME_UNKNOWN ||
2086 !(flags & Uri_CREATE_NO_CRACK_UNKNOWN_SCHEMES)) {
2087 TRACE("(%p %p %x): Treating URI as an hierarchical URI.\n", ptr, data, flags);
2088 data->is_opaque = FALSE;
2090 if(data->scheme_type == URL_SCHEME_WILDCARD && !data->has_implicit_scheme) {
2091 if(**ptr == '/' && *(*ptr+1) == '/') {
2092 data->must_have_path = TRUE;
2097 /* TODO: Handle hierarchical URI's, parse authority then parse the path. */
2098 if(!parse_authority(ptr, data, flags))
2101 return parse_path_hierarchical(ptr, data, flags);
2103 /* Reset ptr to its starting position so opaque path parsing
2104 * begins at the correct location.
2109 /* If it reaches here, then the URI will be treated as an opaque
2113 TRACE("(%p %p %x): Treating URI as an opaque URI.\n", ptr, data, flags);
2115 data->is_opaque = TRUE;
2116 if(!parse_path_opaque(ptr, data, flags))
2122 /* Attempts to parse the query string from the URI.
2125 * If NO_DECODE_EXTRA_INFO flag is set, then invalid percent encoded
2126 * data is allowed to appear in the query string. For unknown scheme types
2127 * invalid percent encoded data is allowed to appear regardless.
2129 static BOOL parse_query(const WCHAR **ptr, parse_data *data, DWORD flags) {
2130 const BOOL known_scheme = data->scheme_type != URL_SCHEME_UNKNOWN;
2133 TRACE("(%p %p %x): URI didn't contain a query string.\n", ptr, data, flags);
2140 while(**ptr && **ptr != '#') {
2141 if(**ptr == '%' && known_scheme &&
2142 !(flags & Uri_CREATE_NO_DECODE_EXTRA_INFO)) {
2143 if(!check_pct_encoded(ptr)) {
2154 data->query_len = *ptr - data->query;
2156 TRACE("(%p %p %x): Parsed query string %s len=%d\n", ptr, data, flags,
2157 debugstr_wn(data->query, data->query_len), data->query_len);
2161 /* Attempts to parse the fragment from the URI.
2164 * If NO_DECODE_EXTRA_INFO flag is set, then invalid percent encoded
2165 * data is allowed to appear in the query string. For unknown scheme types
2166 * invalid percent encoded data is allowed to appear regardless.
2168 static BOOL parse_fragment(const WCHAR **ptr, parse_data *data, DWORD flags) {
2169 const BOOL known_scheme = data->scheme_type != URL_SCHEME_UNKNOWN;
2172 TRACE("(%p %p %x): URI didn't contain a fragment.\n", ptr, data, flags);
2176 data->fragment = *ptr;
2180 if(**ptr == '%' && known_scheme &&
2181 !(flags & Uri_CREATE_NO_DECODE_EXTRA_INFO)) {
2182 if(!check_pct_encoded(ptr)) {
2183 *ptr = data->fragment;
2184 data->fragment = NULL;
2193 data->fragment_len = *ptr - data->fragment;
2195 TRACE("(%p %p %x): Parsed fragment %s len=%d\n", ptr, data, flags,
2196 debugstr_wn(data->fragment, data->fragment_len), data->fragment_len);
2200 /* Parses and validates the components of the specified by data->uri
2201 * and stores the information it parses into 'data'.
2203 * Returns TRUE if it successfully parsed the URI. False otherwise.
2205 static BOOL parse_uri(parse_data *data, DWORD flags) {
2212 TRACE("(%p %x): BEGINNING TO PARSE URI %s.\n", data, flags, debugstr_w(data->uri));
2214 if(!parse_scheme(pptr, data, flags, 0))
2217 if(!parse_hierpart(pptr, data, flags))
2220 if(!parse_query(pptr, data, flags))
2223 if(!parse_fragment(pptr, data, flags))
2226 TRACE("(%p %x): FINISHED PARSING URI.\n", data, flags);
2230 static BOOL canonicalize_username(const parse_data *data, Uri *uri, DWORD flags, BOOL computeOnly) {
2233 if(!data->username) {
2234 uri->userinfo_start = -1;
2238 uri->userinfo_start = uri->canon_len;
2239 for(ptr = data->username; ptr < data->username+data->username_len; ++ptr) {
2241 /* Only decode % encoded values for known scheme types. */
2242 if(data->scheme_type != URL_SCHEME_UNKNOWN) {
2243 /* See if the value really needs decoding. */
2244 WCHAR val = decode_pct_val(ptr);
2245 if(is_unreserved(val)) {
2247 uri->canon_uri[uri->canon_len] = val;
2251 /* Move pass the hex characters. */
2256 } else if(!is_reserved(*ptr) && !is_unreserved(*ptr) && *ptr != '\\') {
2257 /* Only percent encode forbidden characters if the NO_ENCODE_FORBIDDEN_CHARACTERS flag
2260 if(!(flags & Uri_CREATE_NO_ENCODE_FORBIDDEN_CHARACTERS)) {
2262 pct_encode_val(*ptr, uri->canon_uri + uri->canon_len);
2264 uri->canon_len += 3;
2270 /* Nothing special, so just copy the character over. */
2271 uri->canon_uri[uri->canon_len] = *ptr;
2278 static BOOL canonicalize_password(const parse_data *data, Uri *uri, DWORD flags, BOOL computeOnly) {
2281 if(!data->password) {
2282 uri->userinfo_split = -1;
2286 if(uri->userinfo_start == -1)
2287 /* Has a password, but, doesn't have a username. */
2288 uri->userinfo_start = uri->canon_len;
2290 uri->userinfo_split = uri->canon_len - uri->userinfo_start;
2292 /* Add the ':' to the userinfo component. */
2294 uri->canon_uri[uri->canon_len] = ':';
2297 for(ptr = data->password; ptr < data->password+data->password_len; ++ptr) {
2299 /* Only decode % encoded values for known scheme types. */
2300 if(data->scheme_type != URL_SCHEME_UNKNOWN) {
2301 /* See if the value really needs decoding. */
2302 WCHAR val = decode_pct_val(ptr);
2303 if(is_unreserved(val)) {
2305 uri->canon_uri[uri->canon_len] = val;
2309 /* Move pass the hex characters. */
2314 } else if(!is_reserved(*ptr) && !is_unreserved(*ptr) && *ptr != '\\') {
2315 /* Only percent encode forbidden characters if the NO_ENCODE_FORBIDDEN_CHARACTERS flag
2318 if(!(flags & Uri_CREATE_NO_ENCODE_FORBIDDEN_CHARACTERS)) {
2320 pct_encode_val(*ptr, uri->canon_uri + uri->canon_len);
2322 uri->canon_len += 3;
2328 /* Nothing special, so just copy the character over. */
2329 uri->canon_uri[uri->canon_len] = *ptr;
2336 /* Canonicalizes the userinfo of the URI represented by the parse_data.
2338 * Canonicalization of the userinfo is a simple process. If there are any percent
2339 * encoded characters that fall in the "unreserved" character set, they are decoded
2340 * to their actual value. If a character is not in the "unreserved" or "reserved" sets
2341 * then it is percent encoded. Other than that the characters are copied over without
2344 static BOOL canonicalize_userinfo(const parse_data *data, Uri *uri, DWORD flags, BOOL computeOnly) {
2345 uri->userinfo_start = uri->userinfo_split = -1;
2346 uri->userinfo_len = 0;
2348 if(!data->username && !data->password)
2349 /* URI doesn't have userinfo, so nothing to do here. */
2352 if(!canonicalize_username(data, uri, flags, computeOnly))
2355 if(!canonicalize_password(data, uri, flags, computeOnly))
2358 uri->userinfo_len = uri->canon_len - uri->userinfo_start;
2360 TRACE("(%p %p %x %d): Canonicalized userinfo, userinfo_start=%d, userinfo=%s, userinfo_split=%d userinfo_len=%d.\n",
2361 data, uri, flags, computeOnly, uri->userinfo_start, debugstr_wn(uri->canon_uri + uri->userinfo_start, uri->userinfo_len),
2362 uri->userinfo_split, uri->userinfo_len);
2364 /* Now insert the '@' after the userinfo. */
2366 uri->canon_uri[uri->canon_len] = '@';
2372 /* Attempts to canonicalize a reg_name.
2374 * Things that happen:
2375 * 1) If Uri_CREATE_NO_CANONICALIZE flag is not set, then the reg_name is
2376 * lower cased. Unless it's an unknown scheme type, which case it's
2377 * no lower cased regardless.
2379 * 2) Unreserved % encoded characters are decoded for known
2382 * 3) Forbidden characters are % encoded as long as
2383 * Uri_CREATE_NO_ENCODE_FORBIDDEN_CHARACTERS flag is not set and
2384 * it isn't an unknown scheme type.
2386 * 4) If it's a file scheme and the host is "localhost" it's removed.
2388 * 5) If it's a file scheme and Uri_CREATE_FILE_USE_DOS_PATH is set,
2389 * then the UNC path characters are added before the host name.
2391 static BOOL canonicalize_reg_name(const parse_data *data, Uri *uri,
2392 DWORD flags, BOOL computeOnly) {
2393 static const WCHAR localhostW[] =
2394 {'l','o','c','a','l','h','o','s','t',0};
2396 const BOOL known_scheme = data->scheme_type != URL_SCHEME_UNKNOWN;
2398 if(data->scheme_type == URL_SCHEME_FILE &&
2399 data->host_len == lstrlenW(localhostW)) {
2400 if(!StrCmpNIW(data->host, localhostW, data->host_len)) {
2401 uri->host_start = -1;
2403 uri->host_type = Uri_HOST_UNKNOWN;
2408 if(data->scheme_type == URL_SCHEME_FILE && flags & Uri_CREATE_FILE_USE_DOS_PATH) {
2410 uri->canon_uri[uri->canon_len] = '\\';
2411 uri->canon_uri[uri->canon_len+1] = '\\';
2413 uri->canon_len += 2;
2414 uri->authority_start = uri->canon_len;
2417 uri->host_start = uri->canon_len;
2419 for(ptr = data->host; ptr < data->host+data->host_len; ++ptr) {
2420 if(*ptr == '%' && known_scheme) {
2421 WCHAR val = decode_pct_val(ptr);
2422 if(is_unreserved(val)) {
2423 /* If NO_CANONICALIZE is not set, then windows lower cases the
2426 if(!(flags & Uri_CREATE_NO_CANONICALIZE) && isupperW(val)) {
2428 uri->canon_uri[uri->canon_len] = tolowerW(val);
2431 uri->canon_uri[uri->canon_len] = val;
2435 /* Skip past the % encoded character. */
2439 /* Just copy the % over. */
2441 uri->canon_uri[uri->canon_len] = *ptr;
2444 } else if(*ptr == '\\') {
2445 /* Only unknown scheme types could have made it here with a '\\' in the host name. */
2447 uri->canon_uri[uri->canon_len] = *ptr;
2449 } else if(!(flags & Uri_CREATE_NO_ENCODE_FORBIDDEN_CHARACTERS) &&
2450 !is_unreserved(*ptr) && !is_reserved(*ptr) && known_scheme) {
2452 pct_encode_val(*ptr, uri->canon_uri+uri->canon_len);
2454 /* The percent encoded value gets lower cased also. */
2455 if(!(flags & Uri_CREATE_NO_CANONICALIZE)) {
2456 uri->canon_uri[uri->canon_len+1] = tolowerW(uri->canon_uri[uri->canon_len+1]);
2457 uri->canon_uri[uri->canon_len+2] = tolowerW(uri->canon_uri[uri->canon_len+2]);
2461 uri->canon_len += 3;
2464 if(!(flags & Uri_CREATE_NO_CANONICALIZE) && known_scheme)
2465 uri->canon_uri[uri->canon_len] = tolowerW(*ptr);
2467 uri->canon_uri[uri->canon_len] = *ptr;
2474 uri->host_len = uri->canon_len - uri->host_start;
2477 TRACE("(%p %p %x %d): Canonicalize reg_name=%s len=%d\n", data, uri, flags,
2478 computeOnly, debugstr_wn(uri->canon_uri+uri->host_start, uri->host_len),
2482 find_domain_name(uri->canon_uri+uri->host_start, uri->host_len,
2483 &(uri->domain_offset));
2488 /* Attempts to canonicalize an implicit IPv4 address. */
2489 static BOOL canonicalize_implicit_ipv4address(const parse_data *data, Uri *uri, DWORD flags, BOOL computeOnly) {
2490 uri->host_start = uri->canon_len;
2492 TRACE("%u\n", data->implicit_ipv4);
2493 /* For unknown scheme types Windows doesn't convert
2494 * the value into an IP address, but it still considers
2495 * it an IPv4 address.
2497 if(data->scheme_type == URL_SCHEME_UNKNOWN) {
2499 memcpy(uri->canon_uri+uri->canon_len, data->host, data->host_len*sizeof(WCHAR));
2500 uri->canon_len += data->host_len;
2503 uri->canon_len += ui2ipv4(uri->canon_uri+uri->canon_len, data->implicit_ipv4);
2505 uri->canon_len += ui2ipv4(NULL, data->implicit_ipv4);
2508 uri->host_len = uri->canon_len - uri->host_start;
2509 uri->host_type = Uri_HOST_IPV4;
2512 TRACE("%p %p %x %d): Canonicalized implicit IP address=%s len=%d\n",
2513 data, uri, flags, computeOnly,
2514 debugstr_wn(uri->canon_uri+uri->host_start, uri->host_len),
2520 /* Attempts to canonicalize an IPv4 address.
2522 * If the parse_data represents a URI that has an implicit IPv4 address
2523 * (ex. http://256/, this function will convert 256 into 0.0.1.0). If
2524 * the implicit IP address exceeds the value of UINT_MAX (maximum value
2525 * for an IPv4 address) it's canonicalized as if it were a reg-name.
2527 * If the parse_data contains a partial or full IPv4 address it normalizes it.
2528 * A partial IPv4 address is something like "192.0" and would be normalized to
2529 * "192.0.0.0". With a full (or partial) IPv4 address like "192.002.01.003" would
2530 * be normalized to "192.2.1.3".
2533 * Windows ONLY normalizes IPv4 address for known scheme types (one that isn't
2534 * URL_SCHEME_UNKNOWN). For unknown scheme types, it simply copies the data from
2535 * the original URI into the canonicalized URI, but, it still recognizes URI's
2536 * host type as HOST_IPV4.
2538 static BOOL canonicalize_ipv4address(const parse_data *data, Uri *uri, DWORD flags, BOOL computeOnly) {
2539 if(data->has_implicit_ip)
2540 return canonicalize_implicit_ipv4address(data, uri, flags, computeOnly);
2542 uri->host_start = uri->canon_len;
2544 /* Windows only normalizes for known scheme types. */
2545 if(data->scheme_type != URL_SCHEME_UNKNOWN) {
2546 /* parse_data contains a partial or full IPv4 address, so normalize it. */
2547 DWORD i, octetDigitCount = 0, octetCount = 0;
2548 BOOL octetHasDigit = FALSE;
2550 for(i = 0; i < data->host_len; ++i) {
2551 if(data->host[i] == '0' && !octetHasDigit) {
2552 /* Can ignore leading zeros if:
2553 * 1) It isn't the last digit of the octet.
2554 * 2) i+1 != data->host_len
2557 if(octetDigitCount == 2 ||
2558 i+1 == data->host_len ||
2559 data->host[i+1] == '.') {
2561 uri->canon_uri[uri->canon_len] = data->host[i];
2563 TRACE("Adding zero\n");
2565 } else if(data->host[i] == '.') {
2567 uri->canon_uri[uri->canon_len] = data->host[i];
2570 octetDigitCount = 0;
2571 octetHasDigit = FALSE;
2575 uri->canon_uri[uri->canon_len] = data->host[i];
2579 octetHasDigit = TRUE;
2583 /* Make sure the canonicalized IP address has 4 dec-octets.
2584 * If doesn't add "0" ones until there is 4;
2586 for( ; octetCount < 3; ++octetCount) {
2588 uri->canon_uri[uri->canon_len] = '.';
2589 uri->canon_uri[uri->canon_len+1] = '0';
2592 uri->canon_len += 2;
2595 /* Windows doesn't normalize addresses in unknown schemes. */
2597 memcpy(uri->canon_uri+uri->canon_len, data->host, data->host_len*sizeof(WCHAR));
2598 uri->canon_len += data->host_len;
2601 uri->host_len = uri->canon_len - uri->host_start;
2603 TRACE("(%p %p %x %d): Canonicalized IPv4 address, ip=%s len=%d\n",
2604 data, uri, flags, computeOnly,
2605 debugstr_wn(uri->canon_uri+uri->host_start, uri->host_len),
2612 /* Attempts to canonicalize the IPv6 address of the URI.
2614 * Multiple things happen during the canonicalization of an IPv6 address:
2615 * 1) Any leading zero's in a h16 component are removed.
2616 * Ex: [0001:0022::] -> [1:22::]
2618 * 2) The longest sequence of zero h16 components are compressed
2619 * into a "::" (elision). If there's a tie, the first is chosen.
2621 * Ex: [0:0:0:0:1:6:7:8] -> [::1:6:7:8]
2622 * [0:0:0:0:1:2::] -> [::1:2:0:0]
2623 * [0:0:1:2:0:0:7:8] -> [::1:2:0:0:7:8]
2625 * 3) If an IPv4 address is attached to the IPv6 address, it's
2627 * Ex: [::001.002.022.000] -> [::1.2.22.0]
2629 * 4) If an elision is present, but, only represents one h16 component
2632 * Ex: [1::2:3:4:5:6:7] -> [1:0:2:3:4:5:6:7]
2634 * 5) If the IPv6 address contains an IPv4 address and there exists
2635 * at least 1 non-zero h16 component the IPv4 address is converted
2636 * into two h16 components, otherwise it's normalized and kept as is.
2638 * Ex: [::192.200.003.4] -> [::192.200.3.4]
2639 * [ffff::192.200.003.4] -> [ffff::c0c8:3041]
2642 * For unknown scheme types Windows simply copies the address over without any
2645 * IPv4 address can be included in an elision if all its components are 0's.
2647 static BOOL canonicalize_ipv6address(const parse_data *data, Uri *uri,
2648 DWORD flags, BOOL computeOnly) {
2649 uri->host_start = uri->canon_len;
2651 if(data->scheme_type == URL_SCHEME_UNKNOWN) {
2653 memcpy(uri->canon_uri+uri->canon_len, data->host, data->host_len*sizeof(WCHAR));
2654 uri->canon_len += data->host_len;
2658 DWORD i, elision_len;
2660 if(!ipv6_to_number(&(data->ipv6_address), values)) {
2661 TRACE("(%p %p %x %d): Failed to compute numerical value for IPv6 address.\n",
2662 data, uri, flags, computeOnly);
2667 uri->canon_uri[uri->canon_len] = '[';
2670 /* Find where the elision should occur (if any). */
2671 compute_elision_location(&(data->ipv6_address), values, &elision_start, &elision_len);
2673 TRACE("%p %p %x %d): Elision starts at %d, len=%u\n", data, uri, flags,
2674 computeOnly, elision_start, elision_len);
2676 for(i = 0; i < 8; ++i) {
2677 BOOL in_elision = (elision_start > -1 && i >= elision_start &&
2678 i < elision_start+elision_len);
2679 BOOL do_ipv4 = (i == 6 && data->ipv6_address.ipv4 && !in_elision &&
2680 data->ipv6_address.h16_count == 0);
2682 if(i == elision_start) {
2684 uri->canon_uri[uri->canon_len] = ':';
2685 uri->canon_uri[uri->canon_len+1] = ':';
2687 uri->canon_len += 2;
2690 /* We can ignore the current component if we're in the elision. */
2694 /* We only add a ':' if we're not at i == 0, or when we're at
2695 * the very end of elision range since the ':' colon was handled
2696 * earlier. Otherwise we would end up with ":::" after elision.
2698 if(i != 0 && !(elision_start > -1 && i == elision_start+elision_len)) {
2700 uri->canon_uri[uri->canon_len] = ':';
2708 /* Combine the two parts of the IPv4 address values. */
2714 len = ui2ipv4(uri->canon_uri+uri->canon_len, val);
2716 len = ui2ipv4(NULL, val);
2718 uri->canon_len += len;
2721 /* Write a regular h16 component to the URI. */
2723 /* Short circuit for the trivial case. */
2724 if(values[i] == 0) {
2726 uri->canon_uri[uri->canon_len] = '0';
2729 static const WCHAR formatW[] = {'%','x',0};
2732 uri->canon_len += sprintfW(uri->canon_uri+uri->canon_len,
2733 formatW, values[i]);
2736 uri->canon_len += sprintfW(tmp, formatW, values[i]);
2742 /* Add the closing ']'. */
2744 uri->canon_uri[uri->canon_len] = ']';
2748 uri->host_len = uri->canon_len - uri->host_start;
2751 TRACE("(%p %p %x %d): Canonicalized IPv6 address %s, len=%d\n", data, uri, flags,
2752 computeOnly, debugstr_wn(uri->canon_uri+uri->host_start, uri->host_len),
2758 /* Attempts to canonicalize the host of the URI (if any). */
2759 static BOOL canonicalize_host(const parse_data *data, Uri *uri, DWORD flags, BOOL computeOnly) {
2760 uri->host_start = -1;
2762 uri->domain_offset = -1;
2765 switch(data->host_type) {
2767 uri->host_type = Uri_HOST_DNS;
2768 if(!canonicalize_reg_name(data, uri, flags, computeOnly))
2773 uri->host_type = Uri_HOST_IPV4;
2774 if(!canonicalize_ipv4address(data, uri, flags, computeOnly))
2779 if(!canonicalize_ipv6address(data, uri, flags, computeOnly))
2782 uri->host_type = Uri_HOST_IPV6;
2784 case Uri_HOST_UNKNOWN:
2785 if(data->host_len > 0 || data->scheme_type != URL_SCHEME_FILE) {
2786 uri->host_start = uri->canon_len;
2788 /* Nothing happens to unknown host types. */
2790 memcpy(uri->canon_uri+uri->canon_len, data->host, data->host_len*sizeof(WCHAR));
2791 uri->canon_len += data->host_len;
2792 uri->host_len = data->host_len;
2795 uri->host_type = Uri_HOST_UNKNOWN;
2798 FIXME("(%p %p %x %d): Canonicalization for host type %d not supported.\n", data,
2799 uri, flags, computeOnly, data->host_type);
2807 static BOOL canonicalize_port(const parse_data *data, Uri *uri, DWORD flags, BOOL computeOnly) {
2808 BOOL has_default_port = FALSE;
2809 USHORT default_port = 0;
2812 uri->port_offset = -1;
2814 /* Check if the scheme has a default port. */
2815 for(i = 0; i < sizeof(default_ports)/sizeof(default_ports[0]); ++i) {
2816 if(default_ports[i].scheme == data->scheme_type) {
2817 has_default_port = TRUE;
2818 default_port = default_ports[i].port;
2823 uri->has_port = data->has_port || has_default_port;
2826 * 1) Has a port which is the default port.
2827 * 2) Has a port (not the default).
2828 * 3) Doesn't have a port, but, scheme has a default port.
2831 if(has_default_port && data->has_port && data->port_value == default_port) {
2832 /* If it's the default port and this flag isn't set, don't do anything. */
2833 if(flags & Uri_CREATE_NO_CANONICALIZE) {
2834 uri->port_offset = uri->canon_len-uri->authority_start;
2836 uri->canon_uri[uri->canon_len] = ':';
2840 /* Copy the original port over. */
2842 memcpy(uri->canon_uri+uri->canon_len, data->port, data->port_len*sizeof(WCHAR));
2843 uri->canon_len += data->port_len;
2846 uri->canon_len += ui2str(uri->canon_uri+uri->canon_len, data->port_value);
2848 uri->canon_len += ui2str(NULL, data->port_value);
2852 uri->port = default_port;
2853 } else if(data->has_port) {
2854 uri->port_offset = uri->canon_len-uri->authority_start;
2856 uri->canon_uri[uri->canon_len] = ':';
2859 if(flags & Uri_CREATE_NO_CANONICALIZE && data->port) {
2860 /* Copy the original over without changes. */
2862 memcpy(uri->canon_uri+uri->canon_len, data->port, data->port_len*sizeof(WCHAR));
2863 uri->canon_len += data->port_len;
2866 uri->canon_len += ui2str(uri->canon_uri+uri->canon_len, data->port_value);
2868 uri->canon_len += ui2str(NULL, data->port_value);
2871 uri->port = data->port_value;
2872 } else if(has_default_port)
2873 uri->port = default_port;
2878 /* Canonicalizes the authority of the URI represented by the parse_data. */
2879 static BOOL canonicalize_authority(const parse_data *data, Uri *uri, DWORD flags, BOOL computeOnly) {
2880 uri->authority_start = uri->canon_len;
2881 uri->authority_len = 0;
2883 if(!canonicalize_userinfo(data, uri, flags, computeOnly))
2886 if(!canonicalize_host(data, uri, flags, computeOnly))
2889 if(!canonicalize_port(data, uri, flags, computeOnly))
2892 if(uri->host_start != -1 || (data->is_relative && (data->password || data->username)))
2893 uri->authority_len = uri->canon_len - uri->authority_start;
2895 uri->authority_start = -1;
2900 /* Attempts to canonicalize the path of a hierarchical URI.
2902 * Things that happen:
2903 * 1). Forbidden characters are percent encoded, unless the NO_ENCODE_FORBIDDEN
2904 * flag is set or it's a file URI. Forbidden characters are always encoded
2905 * for file schemes regardless and forbidden characters are never encoded
2906 * for unknown scheme types.
2908 * 2). For known scheme types '\\' are changed to '/'.
2910 * 3). Percent encoded, unreserved characters are decoded to their actual values.
2911 * Unless the scheme type is unknown. For file schemes any percent encoded
2912 * character in the unreserved or reserved set is decoded.
2914 * 4). For File schemes if the path is starts with a drive letter and doesn't
2915 * start with a '/' then one is appended.
2916 * Ex: file://c:/test.mp3 -> file:///c:/test.mp3
2918 * 5). Dot segments are removed from the path for all scheme types
2919 * unless NO_CANONICALIZE flag is set. Dot segments aren't removed
2920 * for wildcard scheme types.
2923 * file://c:/test%20test -> file:///c:/test%2520test
2924 * file://c:/test%3Etest -> file:///c:/test%253Etest
2925 * if Uri_CREATE_FILE_USE_DOS_PATH is not set:
2926 * file:///c:/test%20test -> file:///c:/test%20test
2927 * file:///c:/test%test -> file:///c:/test%25test
2929 static BOOL canonicalize_path_hierarchical(const parse_data *data, Uri *uri,
2930 DWORD flags, BOOL computeOnly) {
2932 const BOOL known_scheme = data->scheme_type != URL_SCHEME_UNKNOWN;
2933 const BOOL is_file = data->scheme_type == URL_SCHEME_FILE;
2934 const BOOL is_res = data->scheme_type == URL_SCHEME_RES;
2936 BOOL escape_pct = FALSE;
2939 uri->path_start = -1;
2944 uri->path_start = uri->canon_len;
2947 if(is_file && uri->host_start == -1) {
2948 /* Check if a '/' needs to be appended for the file scheme. */
2949 if(data->path_len > 1 && is_drive_path(ptr) && !(flags & Uri_CREATE_FILE_USE_DOS_PATH)) {
2951 uri->canon_uri[uri->canon_len] = '/';
2954 } else if(*ptr == '/') {
2955 if(!(flags & Uri_CREATE_FILE_USE_DOS_PATH)) {
2956 /* Copy the extra '/' over. */
2958 uri->canon_uri[uri->canon_len] = '/';
2964 if(is_drive_path(ptr)) {
2966 uri->canon_uri[uri->canon_len] = *ptr;
2967 /* If there's a '|' after the drive letter, convert it to a ':'. */
2968 uri->canon_uri[uri->canon_len+1] = ':';
2971 uri->canon_len += 2;
2975 if(!is_file && *(data->path) && *(data->path) != '/') {
2976 /* Prepend a '/' to the path if it doesn't have one. */
2978 uri->canon_uri[uri->canon_len] = '/';
2982 for(; ptr < data->path+data->path_len; ++ptr) {
2983 BOOL do_default_action = TRUE;
2985 if(*ptr == '%' && !is_res) {
2986 const WCHAR *tmp = ptr;
2989 /* Check if the % represents a valid encoded char, or if it needs encoding. */
2990 BOOL force_encode = !check_pct_encoded(&tmp) && is_file && !(flags&Uri_CREATE_FILE_USE_DOS_PATH);
2991 val = decode_pct_val(ptr);
2993 if(force_encode || escape_pct) {
2994 /* Escape the percent sign in the file URI. */
2996 pct_encode_val(*ptr, uri->canon_uri+uri->canon_len);
2997 uri->canon_len += 3;
2998 do_default_action = FALSE;
2999 } else if((is_unreserved(val) && known_scheme) ||
3000 (is_file && (is_unreserved(val) || is_reserved(val) ||
3001 (val && flags&Uri_CREATE_FILE_USE_DOS_PATH && !is_forbidden_dos_path_char(val))))) {
3003 uri->canon_uri[uri->canon_len] = val;
3009 } else if(*ptr == '/' && is_file && (flags & Uri_CREATE_FILE_USE_DOS_PATH)) {
3010 /* Convert the '/' back to a '\\'. */
3012 uri->canon_uri[uri->canon_len] = '\\';
3014 do_default_action = FALSE;
3015 } else if(*ptr == '\\' && known_scheme) {
3016 if(!(is_file && (flags & Uri_CREATE_FILE_USE_DOS_PATH))) {
3017 /* Convert '\\' into a '/'. */
3019 uri->canon_uri[uri->canon_len] = '/';
3021 do_default_action = FALSE;
3023 } else if(known_scheme && !is_res && !is_unreserved(*ptr) && !is_reserved(*ptr) &&
3024 (!(flags & Uri_CREATE_NO_ENCODE_FORBIDDEN_CHARACTERS) || is_file)) {
3025 if(!(is_file && (flags & Uri_CREATE_FILE_USE_DOS_PATH))) {
3026 /* Escape the forbidden character. */
3028 pct_encode_val(*ptr, uri->canon_uri+uri->canon_len);
3029 uri->canon_len += 3;
3030 do_default_action = FALSE;
3034 if(do_default_action) {
3036 uri->canon_uri[uri->canon_len] = *ptr;
3041 uri->path_len = uri->canon_len - uri->path_start;
3043 /* Removing the dot segments only happens when it's not in
3044 * computeOnly mode and it's not a wildcard scheme. File schemes
3045 * with USE_DOS_PATH set don't get dot segments removed.
3047 if(!(is_file && (flags & Uri_CREATE_FILE_USE_DOS_PATH)) &&
3048 data->scheme_type != URL_SCHEME_WILDCARD) {
3049 if(!(flags & Uri_CREATE_NO_CANONICALIZE) && !computeOnly) {
3050 /* Remove the dot segments (if any) and reset everything to the new
3053 DWORD new_len = remove_dot_segments(uri->canon_uri+uri->path_start, uri->path_len);
3054 uri->canon_len -= uri->path_len-new_len;
3055 uri->path_len = new_len;
3060 TRACE("Canonicalized path %s len=%d\n",
3061 debugstr_wn(uri->canon_uri+uri->path_start, uri->path_len),
3067 /* Attempts to canonicalize the path for an opaque URI.
3069 * For known scheme types:
3070 * 1) forbidden characters are percent encoded if
3071 * NO_ENCODE_FORBIDDEN_CHARACTERS isn't set.
3073 * 2) Percent encoded, unreserved characters are decoded
3074 * to their actual values, for known scheme types.
3076 * 3) '\\' are changed to '/' for known scheme types
3077 * except for mailto schemes.
3079 * 4) For file schemes, if USE_DOS_PATH is set all '/'
3080 * are converted to backslashes.
3082 * 5) For file schemes, if USE_DOS_PATH isn't set all '\'
3083 * are converted to forward slashes.
3085 static BOOL canonicalize_path_opaque(const parse_data *data, Uri *uri, DWORD flags, BOOL computeOnly) {
3087 const BOOL known_scheme = data->scheme_type != URL_SCHEME_UNKNOWN;
3088 const BOOL is_file = data->scheme_type == URL_SCHEME_FILE;
3091 uri->path_start = -1;
3096 uri->path_start = uri->canon_len;
3098 /* For javascript: URIs, simply copy path part without any canonicalization */
3099 if(data->scheme_type == URL_SCHEME_JAVASCRIPT) {
3101 memcpy(uri->canon_uri+uri->canon_len, data->path, data->path_len*sizeof(WCHAR));
3102 uri->path_len = data->path_len;
3103 uri->canon_len += data->path_len;
3107 /* Windows doesn't allow a "//" to appear after the scheme
3108 * of a URI, if it's an opaque URI.
3110 if(data->scheme && *(data->path) == '/' && *(data->path+1) == '/') {
3111 /* So it inserts a "/." before the "//" if it exists. */
3113 uri->canon_uri[uri->canon_len] = '/';
3114 uri->canon_uri[uri->canon_len+1] = '.';
3117 uri->canon_len += 2;
3120 for(ptr = data->path; ptr < data->path+data->path_len; ++ptr) {
3121 BOOL do_default_action = TRUE;
3123 if(*ptr == '%' && known_scheme) {
3124 WCHAR val = decode_pct_val(ptr);
3126 if(is_unreserved(val)) {
3128 uri->canon_uri[uri->canon_len] = val;
3134 } else if(*ptr == '/' && is_file && (flags & Uri_CREATE_FILE_USE_DOS_PATH)) {
3136 uri->canon_uri[uri->canon_len] = '\\';
3138 do_default_action = FALSE;
3139 } else if(*ptr == '\\') {
3140 if(is_file && !(flags & Uri_CREATE_FILE_USE_DOS_PATH)) {
3141 /* Convert to a '/'. */
3143 uri->canon_uri[uri->canon_len] = '/';
3145 do_default_action = FALSE;
3147 } else if(known_scheme && !is_unreserved(*ptr) && !is_reserved(*ptr) &&
3148 !(flags & Uri_CREATE_NO_ENCODE_FORBIDDEN_CHARACTERS)) {
3149 if(!(is_file && (flags & Uri_CREATE_FILE_USE_DOS_PATH))) {
3151 pct_encode_val(*ptr, uri->canon_uri+uri->canon_len);
3152 uri->canon_len += 3;
3153 do_default_action = FALSE;
3157 if(do_default_action) {
3159 uri->canon_uri[uri->canon_len] = *ptr;
3164 if(data->scheme_type == URL_SCHEME_MK && !computeOnly && !(flags & Uri_CREATE_NO_CANONICALIZE)) {
3165 DWORD new_len = remove_dot_segments(uri->canon_uri + uri->path_start,
3166 uri->canon_len - uri->path_start);
3167 uri->canon_len = uri->path_start + new_len;
3170 uri->path_len = uri->canon_len - uri->path_start;
3173 TRACE("(%p %p %x %d): Canonicalized opaque URI path %s len=%d\n", data, uri, flags, computeOnly,
3174 debugstr_wn(uri->canon_uri+uri->path_start, uri->path_len), uri->path_len);
3178 /* Determines how the URI represented by the parse_data should be canonicalized.
3180 * Essentially, if the parse_data represents an hierarchical URI then it calls
3181 * canonicalize_authority and the canonicalization functions for the path. If the
3182 * URI is opaque it canonicalizes the path of the URI.
3184 static BOOL canonicalize_hierpart(const parse_data *data, Uri *uri, DWORD flags, BOOL computeOnly) {
3185 if(!data->is_opaque || (data->is_relative && (data->password || data->username))) {
3186 /* "//" is only added for non-wildcard scheme types.
3188 * A "//" is only added to a relative URI if it has a
3189 * host or port component (this only happens if a IUriBuilder
3190 * is generating an IUri).
3192 if((data->is_relative && (data->host || data->has_port)) ||
3193 (!data->is_relative && data->scheme_type != URL_SCHEME_WILDCARD)) {
3194 if(data->scheme_type == URL_SCHEME_WILDCARD)
3198 INT pos = uri->canon_len;
3200 uri->canon_uri[pos] = '/';
3201 uri->canon_uri[pos+1] = '/';
3203 uri->canon_len += 2;
3206 if(!canonicalize_authority(data, uri, flags, computeOnly))
3209 if(data->is_relative && (data->password || data->username)) {
3210 if(!canonicalize_path_opaque(data, uri, flags, computeOnly))
3213 if(!canonicalize_path_hierarchical(data, uri, flags, computeOnly))
3217 /* Opaque URI's don't have an authority. */
3218 uri->userinfo_start = uri->userinfo_split = -1;
3219 uri->userinfo_len = 0;
3220 uri->host_start = -1;
3222 uri->host_type = Uri_HOST_UNKNOWN;
3223 uri->has_port = FALSE;
3224 uri->authority_start = -1;
3225 uri->authority_len = 0;
3226 uri->domain_offset = -1;
3227 uri->port_offset = -1;
3229 if(is_hierarchical_scheme(data->scheme_type)) {
3232 /* Absolute URIs aren't displayed for known scheme types
3233 * which should be hierarchical URIs.
3235 uri->display_modifiers |= URI_DISPLAY_NO_ABSOLUTE_URI;
3237 /* Windows also sets the port for these (if they have one). */
3238 for(i = 0; i < sizeof(default_ports)/sizeof(default_ports[0]); ++i) {
3239 if(data->scheme_type == default_ports[i].scheme) {
3240 uri->has_port = TRUE;
3241 uri->port = default_ports[i].port;
3247 if(!canonicalize_path_opaque(data, uri, flags, computeOnly))
3251 if(uri->path_start > -1 && !computeOnly)
3252 /* Finding file extensions happens for both types of URIs. */
3253 uri->extension_offset = find_file_extension(uri->canon_uri+uri->path_start, uri->path_len);
3255 uri->extension_offset = -1;
3260 /* Attempts to canonicalize the query string of the URI.
3262 * Things that happen:
3263 * 1) For known scheme types forbidden characters
3264 * are percent encoded, unless the NO_DECODE_EXTRA_INFO flag is set
3265 * or NO_ENCODE_FORBIDDEN_CHARACTERS is set.
3267 * 2) For known scheme types, percent encoded, unreserved characters
3268 * are decoded as long as the NO_DECODE_EXTRA_INFO flag isn't set.
3270 static BOOL canonicalize_query(const parse_data *data, Uri *uri, DWORD flags, BOOL computeOnly) {
3271 const WCHAR *ptr, *end;
3272 const BOOL known_scheme = data->scheme_type != URL_SCHEME_UNKNOWN;
3275 uri->query_start = -1;
3280 uri->query_start = uri->canon_len;
3282 end = data->query+data->query_len;
3283 for(ptr = data->query; ptr < end; ++ptr) {
3285 if(known_scheme && !(flags & Uri_CREATE_NO_DECODE_EXTRA_INFO)) {
3286 WCHAR val = decode_pct_val(ptr);
3287 if(is_unreserved(val)) {
3289 uri->canon_uri[uri->canon_len] = val;
3296 } else if(known_scheme && !is_unreserved(*ptr) && !is_reserved(*ptr)) {
3297 if(!(flags & Uri_CREATE_NO_ENCODE_FORBIDDEN_CHARACTERS) &&
3298 !(flags & Uri_CREATE_NO_DECODE_EXTRA_INFO)) {
3300 pct_encode_val(*ptr, uri->canon_uri+uri->canon_len);
3301 uri->canon_len += 3;
3307 uri->canon_uri[uri->canon_len] = *ptr;
3311 uri->query_len = uri->canon_len - uri->query_start;
3314 TRACE("(%p %p %x %d): Canonicalized query string %s len=%d\n", data, uri, flags,
3315 computeOnly, debugstr_wn(uri->canon_uri+uri->query_start, uri->query_len),
3320 static BOOL canonicalize_fragment(const parse_data *data, Uri *uri, DWORD flags, BOOL computeOnly) {
3321 const WCHAR *ptr, *end;
3322 const BOOL known_scheme = data->scheme_type != URL_SCHEME_UNKNOWN;
3324 if(!data->fragment) {
3325 uri->fragment_start = -1;
3326 uri->fragment_len = 0;
3330 uri->fragment_start = uri->canon_len;
3332 end = data->fragment + data->fragment_len;
3333 for(ptr = data->fragment; ptr < end; ++ptr) {
3335 if(known_scheme && !(flags & Uri_CREATE_NO_DECODE_EXTRA_INFO)) {
3336 WCHAR val = decode_pct_val(ptr);
3337 if(is_unreserved(val)) {
3339 uri->canon_uri[uri->canon_len] = val;
3346 } else if(known_scheme && !is_unreserved(*ptr) && !is_reserved(*ptr)) {
3347 if(!(flags & Uri_CREATE_NO_ENCODE_FORBIDDEN_CHARACTERS) &&
3348 !(flags & Uri_CREATE_NO_DECODE_EXTRA_INFO)) {
3350 pct_encode_val(*ptr, uri->canon_uri+uri->canon_len);
3351 uri->canon_len += 3;
3357 uri->canon_uri[uri->canon_len] = *ptr;
3361 uri->fragment_len = uri->canon_len - uri->fragment_start;
3364 TRACE("(%p %p %x %d): Canonicalized fragment %s len=%d\n", data, uri, flags,
3365 computeOnly, debugstr_wn(uri->canon_uri+uri->fragment_start, uri->fragment_len),
3370 /* Canonicalizes the scheme information specified in the parse_data using the specified flags. */
3371 static BOOL canonicalize_scheme(const parse_data *data, Uri *uri, DWORD flags, BOOL computeOnly) {
3372 uri->scheme_start = -1;
3373 uri->scheme_len = 0;
3376 /* The only type of URI that doesn't have to have a scheme is a relative
3379 if(!data->is_relative) {
3380 FIXME("(%p %p %x): Unable to determine the scheme type of %s.\n", data,
3381 uri, flags, debugstr_w(data->uri));
3387 INT pos = uri->canon_len;
3389 for(i = 0; i < data->scheme_len; ++i) {
3390 /* Scheme name must be lower case after canonicalization. */
3391 uri->canon_uri[i + pos] = tolowerW(data->scheme[i]);
3394 uri->canon_uri[i + pos] = ':';
3395 uri->scheme_start = pos;
3397 TRACE("(%p %p %x): Canonicalized scheme=%s, len=%d.\n", data, uri, flags,
3398 debugstr_wn(uri->canon_uri, uri->scheme_len), data->scheme_len);
3401 /* This happens in both computation modes. */
3402 uri->canon_len += data->scheme_len + 1;
3403 uri->scheme_len = data->scheme_len;
3408 /* Computes what the length of the URI specified by the parse_data will be
3409 * after canonicalization occurs using the specified flags.
3411 * This function will return a non-zero value indicating the length of the canonicalized
3412 * URI, or -1 on error.
3414 static int compute_canonicalized_length(const parse_data *data, DWORD flags) {
3417 memset(&uri, 0, sizeof(Uri));
3419 TRACE("(%p %x): Beginning to compute canonicalized length for URI %s\n", data, flags,
3420 debugstr_w(data->uri));
3422 if(!canonicalize_scheme(data, &uri, flags, TRUE)) {
3423 ERR("(%p %x): Failed to compute URI scheme length.\n", data, flags);
3427 if(!canonicalize_hierpart(data, &uri, flags, TRUE)) {
3428 ERR("(%p %x): Failed to compute URI hierpart length.\n", data, flags);
3432 if(!canonicalize_query(data, &uri, flags, TRUE)) {
3433 ERR("(%p %x): Failed to compute query string length.\n", data, flags);
3437 if(!canonicalize_fragment(data, &uri, flags, TRUE)) {
3438 ERR("(%p %x): Failed to compute fragment length.\n", data, flags);
3442 TRACE("(%p %x): Finished computing canonicalized URI length. length=%d\n", data, flags, uri.canon_len);
3444 return uri.canon_len;
3447 /* Canonicalizes the URI data specified in the parse_data, using the given flags. If the
3448 * canonicalization succeeds it will store all the canonicalization information
3449 * in the pointer to the Uri.
3451 * To canonicalize a URI this function first computes what the length of the URI
3452 * specified by the parse_data will be. Once this is done it will then perform the actual
3453 * canonicalization of the URI.
3455 static HRESULT canonicalize_uri(const parse_data *data, Uri *uri, DWORD flags) {
3458 uri->canon_uri = NULL;
3459 uri->canon_size = uri->canon_len = 0;
3461 TRACE("(%p %p %x): beginning to canonicalize URI %s.\n", data, uri, flags, debugstr_w(data->uri));
3463 /* First try to compute the length of the URI. */
3464 len = compute_canonicalized_length(data, flags);
3466 ERR("(%p %p %x): Could not compute the canonicalized length of %s.\n", data, uri, flags,
3467 debugstr_w(data->uri));
3468 return E_INVALIDARG;
3471 uri->canon_uri = heap_alloc((len+1)*sizeof(WCHAR));
3473 return E_OUTOFMEMORY;
3475 uri->canon_size = len;
3476 if(!canonicalize_scheme(data, uri, flags, FALSE)) {
3477 ERR("(%p %p %x): Unable to canonicalize the scheme of the URI.\n", data, uri, flags);
3478 return E_INVALIDARG;
3480 uri->scheme_type = data->scheme_type;
3482 if(!canonicalize_hierpart(data, uri, flags, FALSE)) {
3483 ERR("(%p %p %x): Unable to canonicalize the heirpart of the URI\n", data, uri, flags);
3484 return E_INVALIDARG;
3487 if(!canonicalize_query(data, uri, flags, FALSE)) {
3488 ERR("(%p %p %x): Unable to canonicalize query string of the URI.\n",
3490 return E_INVALIDARG;
3493 if(!canonicalize_fragment(data, uri, flags, FALSE)) {
3494 ERR("(%p %p %x): Unable to canonicalize fragment of the URI.\n",
3496 return E_INVALIDARG;
3499 /* There's a possibility we didn't use all the space we allocated
3502 if(uri->canon_len < uri->canon_size) {
3503 /* This happens if the URI is hierarchical and dot
3504 * segments were removed from its path.
3506 WCHAR *tmp = heap_realloc(uri->canon_uri, (uri->canon_len+1)*sizeof(WCHAR));
3508 return E_OUTOFMEMORY;
3510 uri->canon_uri = tmp;
3511 uri->canon_size = uri->canon_len;
3514 uri->canon_uri[uri->canon_len] = '\0';
3515 TRACE("(%p %p %x): finished canonicalizing the URI. uri=%s\n", data, uri, flags, debugstr_w(uri->canon_uri));
3520 static HRESULT get_builder_component(LPWSTR *component, DWORD *component_len,
3521 LPCWSTR source, DWORD source_len,
3522 LPCWSTR *output, DWORD *output_len)
3535 if(!(*component) && source) {
3536 /* Allocate 'component', and copy the contents from 'source'
3537 * into the new allocation.
3539 *component = heap_alloc((source_len+1)*sizeof(WCHAR));
3541 return E_OUTOFMEMORY;
3543 memcpy(*component, source, source_len*sizeof(WCHAR));
3544 (*component)[source_len] = '\0';
3545 *component_len = source_len;
3548 *output = *component;
3549 *output_len = *component_len;
3550 return *output ? S_OK : S_FALSE;
3553 /* Allocates 'component' and copies the string from 'new_value' into 'component'.
3554 * If 'prefix' is set and 'new_value' isn't NULL, then it checks if 'new_value'
3555 * starts with 'prefix'. If it doesn't then 'prefix' is prepended to 'component'.
3557 * If everything is successful, then will set 'success_flag' in 'flags'.
3559 static HRESULT set_builder_component(LPWSTR *component, DWORD *component_len, LPCWSTR new_value,
3560 WCHAR prefix, DWORD *flags, DWORD success_flag)
3562 heap_free(*component);
3568 BOOL add_prefix = FALSE;
3569 DWORD len = lstrlenW(new_value);
3572 if(prefix && *new_value != prefix) {
3574 *component = heap_alloc((len+2)*sizeof(WCHAR));
3576 *component = heap_alloc((len+1)*sizeof(WCHAR));
3579 return E_OUTOFMEMORY;
3582 (*component)[pos++] = prefix;
3584 memcpy(*component+pos, new_value, (len+1)*sizeof(WCHAR));
3585 *component_len = len+pos;
3588 *flags |= success_flag;
3592 static void reset_builder(UriBuilder *builder) {
3594 IUri_Release(&builder->uri->IUri_iface);
3595 builder->uri = NULL;
3597 heap_free(builder->fragment);
3598 builder->fragment = NULL;
3599 builder->fragment_len = 0;
3601 heap_free(builder->host);
3602 builder->host = NULL;
3603 builder->host_len = 0;
3605 heap_free(builder->password);
3606 builder->password = NULL;
3607 builder->password_len = 0;
3609 heap_free(builder->path);
3610 builder->path = NULL;
3611 builder->path_len = 0;
3613 heap_free(builder->query);
3614 builder->query = NULL;
3615 builder->query_len = 0;
3617 heap_free(builder->scheme);
3618 builder->scheme = NULL;
3619 builder->scheme_len = 0;
3621 heap_free(builder->username);
3622 builder->username = NULL;
3623 builder->username_len = 0;
3625 builder->has_port = FALSE;
3627 builder->modified_props = 0;
3630 static HRESULT validate_scheme_name(const UriBuilder *builder, parse_data *data, DWORD flags) {
3631 const WCHAR *component;
3636 if(builder->scheme) {
3637 ptr = builder->scheme;
3638 expected_len = builder->scheme_len;
3639 } else if(builder->uri && builder->uri->scheme_start > -1) {
3640 ptr = builder->uri->canon_uri+builder->uri->scheme_start;
3641 expected_len = builder->uri->scheme_len;
3643 static const WCHAR nullW[] = {0};
3650 if(parse_scheme(pptr, data, flags, ALLOW_NULL_TERM_SCHEME) &&
3651 data->scheme_len == expected_len) {
3653 TRACE("(%p %p %x): Found valid scheme component %s len=%d.\n", builder, data, flags,
3654 debugstr_wn(data->scheme, data->scheme_len), data->scheme_len);
3656 TRACE("(%p %p %x): Invalid scheme component found %s.\n", builder, data, flags,
3657 debugstr_wn(component, expected_len));
3658 return INET_E_INVALID_URL;
3664 static HRESULT validate_username(const UriBuilder *builder, parse_data *data, DWORD flags) {
3669 if(builder->username) {
3670 ptr = builder->username;
3671 expected_len = builder->username_len;
3672 } else if(!(builder->modified_props & Uri_HAS_USER_NAME) && builder->uri &&
3673 builder->uri->userinfo_start > -1 && builder->uri->userinfo_split != 0) {
3674 /* Just use the username from the base Uri. */
3675 data->username = builder->uri->canon_uri+builder->uri->userinfo_start;
3676 data->username_len = (builder->uri->userinfo_split > -1) ?
3677 builder->uri->userinfo_split : builder->uri->userinfo_len;
3685 const WCHAR *component = ptr;
3687 if(parse_username(pptr, data, flags, ALLOW_NULL_TERM_USER_NAME) &&
3688 data->username_len == expected_len)
3689 TRACE("(%p %p %x): Found valid username component %s len=%d.\n", builder, data, flags,
3690 debugstr_wn(data->username, data->username_len), data->username_len);
3692 TRACE("(%p %p %x): Invalid username component found %s.\n", builder, data, flags,
3693 debugstr_wn(component, expected_len));
3694 return INET_E_INVALID_URL;
3701 static HRESULT validate_password(const UriBuilder *builder, parse_data *data, DWORD flags) {
3706 if(builder->password) {
3707 ptr = builder->password;
3708 expected_len = builder->password_len;
3709 } else if(!(builder->modified_props & Uri_HAS_PASSWORD) && builder->uri &&
3710 builder->uri->userinfo_split > -1) {
3711 data->password = builder->uri->canon_uri+builder->uri->userinfo_start+builder->uri->userinfo_split+1;
3712 data->password_len = builder->uri->userinfo_len-builder->uri->userinfo_split-1;
3720 const WCHAR *component = ptr;
3722 if(parse_password(pptr, data, flags, ALLOW_NULL_TERM_PASSWORD) &&
3723 data->password_len == expected_len)
3724 TRACE("(%p %p %x): Found valid password component %s len=%d.\n", builder, data, flags,
3725 debugstr_wn(data->password, data->password_len), data->password_len);
3727 TRACE("(%p %p %x): Invalid password component found %s.\n", builder, data, flags,
3728 debugstr_wn(component, expected_len));
3729 return INET_E_INVALID_URL;
3736 static HRESULT validate_userinfo(const UriBuilder *builder, parse_data *data, DWORD flags) {
3739 hr = validate_username(builder, data, flags);
3743 hr = validate_password(builder, data, flags);
3750 static HRESULT validate_host(const UriBuilder *builder, parse_data *data, DWORD flags) {
3756 ptr = builder->host;
3757 expected_len = builder->host_len;
3758 } else if(!(builder->modified_props & Uri_HAS_HOST) && builder->uri && builder->uri->host_start > -1) {
3759 ptr = builder->uri->canon_uri + builder->uri->host_start;
3760 expected_len = builder->uri->host_len;
3765 const WCHAR *component = ptr;
3766 DWORD extras = ALLOW_BRACKETLESS_IP_LITERAL|IGNORE_PORT_DELIMITER|SKIP_IP_FUTURE_CHECK;
3769 if(parse_host(pptr, data, flags, extras) && data->host_len == expected_len)
3770 TRACE("(%p %p %x): Found valid host name %s len=%d type=%d.\n", builder, data, flags,
3771 debugstr_wn(data->host, data->host_len), data->host_len, data->host_type);
3773 TRACE("(%p %p %x): Invalid host name found %s.\n", builder, data, flags,
3774 debugstr_wn(component, expected_len));
3775 return INET_E_INVALID_URL;
3782 static void setup_port(const UriBuilder *builder, parse_data *data, DWORD flags) {
3783 if(builder->modified_props & Uri_HAS_PORT) {
3784 if(builder->has_port) {
3785 data->has_port = TRUE;
3786 data->port_value = builder->port;
3788 } else if(builder->uri && builder->uri->has_port) {
3789 data->has_port = TRUE;
3790 data->port_value = builder->uri->port;
3794 TRACE("(%p %p %x): Using %u as port for IUri.\n", builder, data, flags, data->port_value);
3797 static HRESULT validate_path(const UriBuilder *builder, parse_data *data, DWORD flags) {
3798 const WCHAR *ptr = NULL;
3799 const WCHAR *component;
3802 BOOL check_len = TRUE;
3806 ptr = builder->path;
3807 expected_len = builder->path_len;
3808 } else if(!(builder->modified_props & Uri_HAS_PATH) &&
3809 builder->uri && builder->uri->path_start > -1) {
3810 ptr = builder->uri->canon_uri+builder->uri->path_start;
3811 expected_len = builder->uri->path_len;
3813 static const WCHAR nullW[] = {0};
3822 /* How the path is validated depends on what type of
3825 valid = data->is_opaque ?
3826 parse_path_opaque(pptr, data, flags) : parse_path_hierarchical(pptr, data, flags);
3828 if(!valid || (check_len && expected_len != data->path_len)) {
3829 TRACE("(%p %p %x): Invalid path component %s.\n", builder, data, flags,
3830 debugstr_wn(component, expected_len) );
3831 return INET_E_INVALID_URL;
3834 TRACE("(%p %p %x): Valid path component %s len=%d.\n", builder, data, flags,
3835 debugstr_wn(data->path, data->path_len), data->path_len);
3840 static HRESULT validate_query(const UriBuilder *builder, parse_data *data, DWORD flags) {
3841 const WCHAR *ptr = NULL;
3845 if(builder->query) {
3846 ptr = builder->query;
3847 expected_len = builder->query_len;
3848 } else if(!(builder->modified_props & Uri_HAS_QUERY) && builder->uri &&
3849 builder->uri->query_start > -1) {
3850 ptr = builder->uri->canon_uri+builder->uri->query_start;
3851 expected_len = builder->uri->query_len;
3855 const WCHAR *component = ptr;
3858 if(parse_query(pptr, data, flags) && expected_len == data->query_len)
3859 TRACE("(%p %p %x): Valid query component %s len=%d.\n", builder, data, flags,
3860 debugstr_wn(data->query, data->query_len), data->query_len);
3862 TRACE("(%p %p %x): Invalid query component %s.\n", builder, data, flags,
3863 debugstr_wn(component, expected_len));
3864 return INET_E_INVALID_URL;
3871 static HRESULT validate_fragment(const UriBuilder *builder, parse_data *data, DWORD flags) {
3872 const WCHAR *ptr = NULL;
3876 if(builder->fragment) {
3877 ptr = builder->fragment;
3878 expected_len = builder->fragment_len;
3879 } else if(!(builder->modified_props & Uri_HAS_FRAGMENT) && builder->uri &&
3880 builder->uri->fragment_start > -1) {
3881 ptr = builder->uri->canon_uri+builder->uri->fragment_start;
3882 expected_len = builder->uri->fragment_len;
3886 const WCHAR *component = ptr;
3889 if(parse_fragment(pptr, data, flags) && expected_len == data->fragment_len)
3890 TRACE("(%p %p %x): Valid fragment component %s len=%d.\n", builder, data, flags,
3891 debugstr_wn(data->fragment, data->fragment_len), data->fragment_len);
3893 TRACE("(%p %p %x): Invalid fragment component %s.\n", builder, data, flags,
3894 debugstr_wn(component, expected_len));
3895 return INET_E_INVALID_URL;
3902 static HRESULT validate_components(const UriBuilder *builder, parse_data *data, DWORD flags) {
3905 memset(data, 0, sizeof(parse_data));
3907 TRACE("(%p %p %x): Beginning to validate builder components.\n", builder, data, flags);
3909 hr = validate_scheme_name(builder, data, flags);
3913 /* Extra validation for file schemes. */
3914 if(data->scheme_type == URL_SCHEME_FILE) {
3915 if((builder->password || (builder->uri && builder->uri->userinfo_split > -1)) ||
3916 (builder->username || (builder->uri && builder->uri->userinfo_start > -1))) {
3917 TRACE("(%p %p %x): File schemes can't contain a username or password.\n",
3918 builder, data, flags);
3919 return INET_E_INVALID_URL;
3923 hr = validate_userinfo(builder, data, flags);
3927 hr = validate_host(builder, data, flags);
3931 setup_port(builder, data, flags);
3933 /* The URI is opaque if it doesn't have an authority component. */
3934 if(!data->is_relative)
3935 data->is_opaque = !data->username && !data->password && !data->host && !data->has_port;
3937 data->is_opaque = !data->host && !data->has_port;
3939 hr = validate_path(builder, data, flags);
3943 hr = validate_query(builder, data, flags);
3947 hr = validate_fragment(builder, data, flags);
3951 TRACE("(%p %p %x): Finished validating builder components.\n", builder, data, flags);
3956 static void convert_to_dos_path(const WCHAR *path, DWORD path_len,
3957 WCHAR *output, DWORD *output_len)
3959 const WCHAR *ptr = path;
3961 if(path_len > 3 && *ptr == '/' && is_drive_path(path+1))
3962 /* Skip over the leading / before the drive path. */
3965 for(; ptr < path+path_len; ++ptr) {
3978 /* Generates a raw uri string using the parse_data. */
3979 static DWORD generate_raw_uri(const parse_data *data, BSTR uri, DWORD flags) {
3984 memcpy(uri, data->scheme, data->scheme_len*sizeof(WCHAR));
3985 uri[data->scheme_len] = ':';
3987 length += data->scheme_len+1;
3990 if(!data->is_opaque) {
3991 /* For the "//" which appears before the authority component. */
3994 uri[length+1] = '/';
3998 /* Check if we need to add the "\\" before the host name
3999 * of a UNC server name in a DOS path.
4001 if(flags & RAW_URI_CONVERT_TO_DOS_PATH &&
4002 data->scheme_type == URL_SCHEME_FILE && data->host) {
4005 uri[length+1] = '\\';
4011 if(data->username) {
4013 memcpy(uri+length, data->username, data->username_len*sizeof(WCHAR));
4014 length += data->username_len;
4017 if(data->password) {
4020 memcpy(uri+length+1, data->password, data->password_len*sizeof(WCHAR));
4022 length += data->password_len+1;
4025 if(data->password || data->username) {
4032 /* IPv6 addresses get the brackets added around them if they don't already
4035 const BOOL add_brackets = data->host_type == Uri_HOST_IPV6 && *(data->host) != '[';
4043 memcpy(uri+length, data->host, data->host_len*sizeof(WCHAR));
4044 length += data->host_len;
4053 if(data->has_port) {
4054 /* The port isn't included in the raw uri if it's the default
4055 * port for the scheme type.
4058 BOOL is_default = FALSE;
4060 for(i = 0; i < sizeof(default_ports)/sizeof(default_ports[0]); ++i) {
4061 if(data->scheme_type == default_ports[i].scheme &&
4062 data->port_value == default_ports[i].port)
4066 if(!is_default || flags & RAW_URI_FORCE_PORT_DISP) {
4072 length += ui2str(uri+length, data->port_value);
4074 length += ui2str(NULL, data->port_value);
4078 /* Check if a '/' should be added before the path for hierarchical URIs. */
4079 if(!data->is_opaque && data->path && *(data->path) != '/') {
4086 if(!data->is_opaque && data->scheme_type == URL_SCHEME_FILE &&
4087 flags & RAW_URI_CONVERT_TO_DOS_PATH) {
4091 convert_to_dos_path(data->path, data->path_len, uri+length, &len);
4093 convert_to_dos_path(data->path, data->path_len, NULL, &len);
4098 memcpy(uri+length, data->path, data->path_len*sizeof(WCHAR));
4099 length += data->path_len;
4105 memcpy(uri+length, data->query, data->query_len*sizeof(WCHAR));
4106 length += data->query_len;
4109 if(data->fragment) {
4111 memcpy(uri+length, data->fragment, data->fragment_len*sizeof(WCHAR));
4112 length += data->fragment_len;
4116 TRACE("(%p %p): Generated raw uri=%s len=%d\n", data, uri, debugstr_wn(uri, length), length);
4118 TRACE("(%p %p): Computed raw uri len=%d\n", data, uri, length);
4123 static HRESULT generate_uri(const UriBuilder *builder, const parse_data *data, Uri *uri, DWORD flags) {
4125 DWORD length = generate_raw_uri(data, NULL, 0);
4126 uri->raw_uri = SysAllocStringLen(NULL, length);
4128 return E_OUTOFMEMORY;
4130 generate_raw_uri(data, uri->raw_uri, 0);
4132 hr = canonicalize_uri(data, uri, flags);
4134 if(hr == E_INVALIDARG)
4135 return INET_E_INVALID_URL;
4139 uri->create_flags = flags;
4143 static inline Uri* impl_from_IUri(IUri *iface)
4145 return CONTAINING_RECORD(iface, Uri, IUri_iface);
4148 static inline void destory_uri_obj(Uri *This)
4150 SysFreeString(This->raw_uri);
4151 heap_free(This->canon_uri);
4155 static HRESULT WINAPI Uri_QueryInterface(IUri *iface, REFIID riid, void **ppv)
4157 Uri *This = impl_from_IUri(iface);
4159 if(IsEqualGUID(&IID_IUnknown, riid)) {
4160 TRACE("(%p)->(IID_IUnknown %p)\n", This, ppv);
4161 *ppv = &This->IUri_iface;
4162 }else if(IsEqualGUID(&IID_IUri, riid)) {
4163 TRACE("(%p)->(IID_IUri %p)\n", This, ppv);
4164 *ppv = &This->IUri_iface;
4165 }else if(IsEqualGUID(&IID_IUriBuilderFactory, riid)) {
4166 TRACE("(%p)->(IID_IUriBuilderFactory %p)\n", This, riid);
4167 *ppv = &This->IUriBuilderFactory_iface;
4168 }else if(IsEqualGUID(&IID_IUriObj, riid)) {
4169 TRACE("(%p)->(IID_IUriObj %p)\n", This, ppv);
4173 TRACE("(%p)->(%s %p)\n", This, debugstr_guid(riid), ppv);
4175 return E_NOINTERFACE;
4178 IUnknown_AddRef((IUnknown*)*ppv);
4182 static ULONG WINAPI Uri_AddRef(IUri *iface)
4184 Uri *This = impl_from_IUri(iface);
4185 LONG ref = InterlockedIncrement(&This->ref);
4187 TRACE("(%p) ref=%d\n", This, ref);
4192 static ULONG WINAPI Uri_Release(IUri *iface)
4194 Uri *This = impl_from_IUri(iface);
4195 LONG ref = InterlockedDecrement(&This->ref);
4197 TRACE("(%p) ref=%d\n", This, ref);
4200 destory_uri_obj(This);
4205 static HRESULT WINAPI Uri_GetPropertyBSTR(IUri *iface, Uri_PROPERTY uriProp, BSTR *pbstrProperty, DWORD dwFlags)
4207 Uri *This = impl_from_IUri(iface);
4209 TRACE("(%p %s)->(%d %p %x)\n", This, debugstr_w(This->canon_uri), uriProp, pbstrProperty, dwFlags);
4214 if(uriProp > Uri_PROPERTY_STRING_LAST) {
4215 /* Windows allocates an empty BSTR for invalid Uri_PROPERTY's. */
4216 *pbstrProperty = SysAllocStringLen(NULL, 0);
4217 if(!(*pbstrProperty))
4218 return E_OUTOFMEMORY;
4220 /* It only returns S_FALSE for the ZONE property... */
4221 if(uriProp == Uri_PROPERTY_ZONE)
4227 /* Don't have support for flags yet. */
4229 FIXME("(%p)->(%d %p %x)\n", This, uriProp, pbstrProperty, dwFlags);
4234 case Uri_PROPERTY_ABSOLUTE_URI:
4235 if(This->display_modifiers & URI_DISPLAY_NO_ABSOLUTE_URI) {
4236 *pbstrProperty = SysAllocStringLen(NULL, 0);
4239 if(This->scheme_type != URL_SCHEME_UNKNOWN && This->userinfo_start > -1) {
4240 if(This->userinfo_len == 0) {
4241 /* Don't include the '@' after the userinfo component. */
4242 *pbstrProperty = SysAllocStringLen(NULL, This->canon_len-1);
4244 if(*pbstrProperty) {
4245 /* Copy everything before it. */
4246 memcpy(*pbstrProperty, This->canon_uri, This->userinfo_start*sizeof(WCHAR));
4248 /* And everything after it. */
4249 memcpy(*pbstrProperty+This->userinfo_start, This->canon_uri+This->userinfo_start+1,
4250 (This->canon_len-This->userinfo_start-1)*sizeof(WCHAR));
4252 } else if(This->userinfo_split == 0 && This->userinfo_len == 1) {
4253 /* Don't include the ":@" */
4254 *pbstrProperty = SysAllocStringLen(NULL, This->canon_len-2);
4256 if(*pbstrProperty) {
4257 memcpy(*pbstrProperty, This->canon_uri, This->userinfo_start*sizeof(WCHAR));
4258 memcpy(*pbstrProperty+This->userinfo_start, This->canon_uri+This->userinfo_start+2,
4259 (This->canon_len-This->userinfo_start-2)*sizeof(WCHAR));
4262 *pbstrProperty = SysAllocString(This->canon_uri);
4266 *pbstrProperty = SysAllocString(This->canon_uri);
4271 if(!(*pbstrProperty))
4272 hres = E_OUTOFMEMORY;
4275 case Uri_PROPERTY_AUTHORITY:
4276 if(This->authority_start > -1) {
4277 if(This->port_offset > -1 && is_default_port(This->scheme_type, This->port) &&
4278 This->display_modifiers & URI_DISPLAY_NO_DEFAULT_PORT_AUTH)
4279 /* Don't include the port in the authority component. */
4280 *pbstrProperty = SysAllocStringLen(This->canon_uri+This->authority_start, This->port_offset);
4282 *pbstrProperty = SysAllocStringLen(This->canon_uri+This->authority_start, This->authority_len);
4285 *pbstrProperty = SysAllocStringLen(NULL, 0);
4289 if(!(*pbstrProperty))
4290 hres = E_OUTOFMEMORY;
4293 case Uri_PROPERTY_DISPLAY_URI:
4294 /* The Display URI contains everything except for the userinfo for known
4297 if(This->scheme_type != URL_SCHEME_UNKNOWN && This->userinfo_start > -1) {
4298 *pbstrProperty = SysAllocStringLen(NULL, This->canon_len-This->userinfo_len);
4300 if(*pbstrProperty) {
4301 /* Copy everything before the userinfo over. */
4302 memcpy(*pbstrProperty, This->canon_uri, This->userinfo_start*sizeof(WCHAR));
4303 /* Copy everything after the userinfo over. */
4304 memcpy(*pbstrProperty+This->userinfo_start,
4305 This->canon_uri+This->userinfo_start+This->userinfo_len+1,
4306 (This->canon_len-(This->userinfo_start+This->userinfo_len+1))*sizeof(WCHAR));
4309 *pbstrProperty = SysAllocString(This->canon_uri);
4311 if(!(*pbstrProperty))
4312 hres = E_OUTOFMEMORY;
4317 case Uri_PROPERTY_DOMAIN:
4318 if(This->domain_offset > -1) {
4319 *pbstrProperty = SysAllocStringLen(This->canon_uri+This->host_start+This->domain_offset,
4320 This->host_len-This->domain_offset);
4323 *pbstrProperty = SysAllocStringLen(NULL, 0);
4327 if(!(*pbstrProperty))
4328 hres = E_OUTOFMEMORY;
4331 case Uri_PROPERTY_EXTENSION:
4332 if(This->extension_offset > -1) {
4333 *pbstrProperty = SysAllocStringLen(This->canon_uri+This->path_start+This->extension_offset,
4334 This->path_len-This->extension_offset);
4337 *pbstrProperty = SysAllocStringLen(NULL, 0);
4341 if(!(*pbstrProperty))
4342 hres = E_OUTOFMEMORY;
4345 case Uri_PROPERTY_FRAGMENT:
4346 if(This->fragment_start > -1) {
4347 *pbstrProperty = SysAllocStringLen(This->canon_uri+This->fragment_start, This->fragment_len);
4350 *pbstrProperty = SysAllocStringLen(NULL, 0);
4354 if(!(*pbstrProperty))
4355 hres = E_OUTOFMEMORY;
4358 case Uri_PROPERTY_HOST:
4359 if(This->host_start > -1) {
4360 /* The '[' and ']' aren't included for IPv6 addresses. */
4361 if(This->host_type == Uri_HOST_IPV6)
4362 *pbstrProperty = SysAllocStringLen(This->canon_uri+This->host_start+1, This->host_len-2);
4364 *pbstrProperty = SysAllocStringLen(This->canon_uri+This->host_start, This->host_len);
4368 *pbstrProperty = SysAllocStringLen(NULL, 0);
4372 if(!(*pbstrProperty))
4373 hres = E_OUTOFMEMORY;
4376 case Uri_PROPERTY_PASSWORD:
4377 if(This->userinfo_split > -1) {
4378 *pbstrProperty = SysAllocStringLen(
4379 This->canon_uri+This->userinfo_start+This->userinfo_split+1,
4380 This->userinfo_len-This->userinfo_split-1);
4383 *pbstrProperty = SysAllocStringLen(NULL, 0);
4387 if(!(*pbstrProperty))
4388 return E_OUTOFMEMORY;
4391 case Uri_PROPERTY_PATH:
4392 if(This->path_start > -1) {
4393 *pbstrProperty = SysAllocStringLen(This->canon_uri+This->path_start, This->path_len);
4396 *pbstrProperty = SysAllocStringLen(NULL, 0);
4400 if(!(*pbstrProperty))
4401 hres = E_OUTOFMEMORY;
4404 case Uri_PROPERTY_PATH_AND_QUERY:
4405 if(This->path_start > -1) {
4406 *pbstrProperty = SysAllocStringLen(This->canon_uri+This->path_start, This->path_len+This->query_len);
4408 } else if(This->query_start > -1) {
4409 *pbstrProperty = SysAllocStringLen(This->canon_uri+This->query_start, This->query_len);
4412 *pbstrProperty = SysAllocStringLen(NULL, 0);
4416 if(!(*pbstrProperty))
4417 hres = E_OUTOFMEMORY;
4420 case Uri_PROPERTY_QUERY:
4421 if(This->query_start > -1) {
4422 *pbstrProperty = SysAllocStringLen(This->canon_uri+This->query_start, This->query_len);
4425 *pbstrProperty = SysAllocStringLen(NULL, 0);
4429 if(!(*pbstrProperty))
4430 hres = E_OUTOFMEMORY;
4433 case Uri_PROPERTY_RAW_URI:
4434 *pbstrProperty = SysAllocString(This->raw_uri);
4435 if(!(*pbstrProperty))
4436 hres = E_OUTOFMEMORY;
4440 case Uri_PROPERTY_SCHEME_NAME:
4441 if(This->scheme_start > -1) {
4442 *pbstrProperty = SysAllocStringLen(This->canon_uri + This->scheme_start, This->scheme_len);
4445 *pbstrProperty = SysAllocStringLen(NULL, 0);
4449 if(!(*pbstrProperty))
4450 hres = E_OUTOFMEMORY;
4453 case Uri_PROPERTY_USER_INFO:
4454 if(This->userinfo_start > -1) {
4455 *pbstrProperty = SysAllocStringLen(This->canon_uri+This->userinfo_start, This->userinfo_len);
4458 *pbstrProperty = SysAllocStringLen(NULL, 0);
4462 if(!(*pbstrProperty))
4463 hres = E_OUTOFMEMORY;
4466 case Uri_PROPERTY_USER_NAME:
4467 if(This->userinfo_start > -1 && This->userinfo_split != 0) {
4468 /* If userinfo_split is set, that means a password exists
4469 * so the username is only from userinfo_start to userinfo_split.
4471 if(This->userinfo_split > -1) {
4472 *pbstrProperty = SysAllocStringLen(This->canon_uri + This->userinfo_start, This->userinfo_split);
4475 *pbstrProperty = SysAllocStringLen(This->canon_uri + This->userinfo_start, This->userinfo_len);
4479 *pbstrProperty = SysAllocStringLen(NULL, 0);
4483 if(!(*pbstrProperty))
4484 return E_OUTOFMEMORY;
4488 FIXME("(%p)->(%d %p %x)\n", This, uriProp, pbstrProperty, dwFlags);
4495 static HRESULT WINAPI Uri_GetPropertyLength(IUri *iface, Uri_PROPERTY uriProp, DWORD *pcchProperty, DWORD dwFlags)
4497 Uri *This = impl_from_IUri(iface);
4499 TRACE("(%p %s)->(%d %p %x)\n", This, debugstr_w(This->canon_uri), uriProp, pcchProperty, dwFlags);
4502 return E_INVALIDARG;
4504 /* Can only return a length for a property if it's a string. */
4505 if(uriProp > Uri_PROPERTY_STRING_LAST)
4506 return E_INVALIDARG;
4508 /* Don't have support for flags yet. */
4510 FIXME("(%p)->(%d %p %x)\n", This, uriProp, pcchProperty, dwFlags);
4515 case Uri_PROPERTY_ABSOLUTE_URI:
4516 if(This->display_modifiers & URI_DISPLAY_NO_ABSOLUTE_URI) {
4520 if(This->scheme_type != URL_SCHEME_UNKNOWN) {
4521 if(This->userinfo_start > -1 && This->userinfo_len == 0)
4522 /* Don't include the '@' in the length. */
4523 *pcchProperty = This->canon_len-1;
4524 else if(This->userinfo_start > -1 && This->userinfo_len == 1 &&
4525 This->userinfo_split == 0)
4526 /* Don't include the ":@" in the length. */
4527 *pcchProperty = This->canon_len-2;
4529 *pcchProperty = This->canon_len;
4531 *pcchProperty = This->canon_len;
4537 case Uri_PROPERTY_AUTHORITY:
4538 if(This->port_offset > -1 &&
4539 This->display_modifiers & URI_DISPLAY_NO_DEFAULT_PORT_AUTH &&
4540 is_default_port(This->scheme_type, This->port))
4541 /* Only count up until the port in the authority. */
4542 *pcchProperty = This->port_offset;
4544 *pcchProperty = This->authority_len;
4545 hres = (This->authority_start > -1) ? S_OK : S_FALSE;
4547 case Uri_PROPERTY_DISPLAY_URI:
4548 if(This->scheme_type != URL_SCHEME_UNKNOWN && This->userinfo_start > -1)
4549 *pcchProperty = This->canon_len-This->userinfo_len-1;
4551 *pcchProperty = This->canon_len;
4555 case Uri_PROPERTY_DOMAIN:
4556 if(This->domain_offset > -1)
4557 *pcchProperty = This->host_len - This->domain_offset;
4561 hres = (This->domain_offset > -1) ? S_OK : S_FALSE;
4563 case Uri_PROPERTY_EXTENSION:
4564 if(This->extension_offset > -1) {
4565 *pcchProperty = This->path_len - This->extension_offset;
4573 case Uri_PROPERTY_FRAGMENT:
4574 *pcchProperty = This->fragment_len;
4575 hres = (This->fragment_start > -1) ? S_OK : S_FALSE;
4577 case Uri_PROPERTY_HOST:
4578 *pcchProperty = This->host_len;
4580 /* '[' and ']' aren't included in the length. */
4581 if(This->host_type == Uri_HOST_IPV6)
4584 hres = (This->host_start > -1) ? S_OK : S_FALSE;
4586 case Uri_PROPERTY_PASSWORD:
4587 *pcchProperty = (This->userinfo_split > -1) ? This->userinfo_len-This->userinfo_split-1 : 0;
4588 hres = (This->userinfo_split > -1) ? S_OK : S_FALSE;
4590 case Uri_PROPERTY_PATH:
4591 *pcchProperty = This->path_len;
4592 hres = (This->path_start > -1) ? S_OK : S_FALSE;
4594 case Uri_PROPERTY_PATH_AND_QUERY:
4595 *pcchProperty = This->path_len+This->query_len;
4596 hres = (This->path_start > -1 || This->query_start > -1) ? S_OK : S_FALSE;
4598 case Uri_PROPERTY_QUERY:
4599 *pcchProperty = This->query_len;
4600 hres = (This->query_start > -1) ? S_OK : S_FALSE;
4602 case Uri_PROPERTY_RAW_URI:
4603 *pcchProperty = SysStringLen(This->raw_uri);
4606 case Uri_PROPERTY_SCHEME_NAME:
4607 *pcchProperty = This->scheme_len;
4608 hres = (This->scheme_start > -1) ? S_OK : S_FALSE;
4610 case Uri_PROPERTY_USER_INFO:
4611 *pcchProperty = This->userinfo_len;
4612 hres = (This->userinfo_start > -1) ? S_OK : S_FALSE;
4614 case Uri_PROPERTY_USER_NAME:
4615 *pcchProperty = (This->userinfo_split > -1) ? This->userinfo_split : This->userinfo_len;
4616 if(This->userinfo_split == 0)
4619 hres = (This->userinfo_start > -1) ? S_OK : S_FALSE;
4622 FIXME("(%p)->(%d %p %x)\n", This, uriProp, pcchProperty, dwFlags);
4629 static HRESULT WINAPI Uri_GetPropertyDWORD(IUri *iface, Uri_PROPERTY uriProp, DWORD *pcchProperty, DWORD dwFlags)
4631 Uri *This = impl_from_IUri(iface);
4634 TRACE("(%p %s)->(%d %p %x)\n", This, debugstr_w(This->canon_uri), uriProp, pcchProperty, dwFlags);
4637 return E_INVALIDARG;
4639 /* Microsoft's implementation for the ZONE property of a URI seems to be lacking...
4640 * From what I can tell, instead of checking which URLZONE the URI belongs to it
4641 * simply assigns URLZONE_INVALID and returns E_NOTIMPL. This also applies to the GetZone
4644 if(uriProp == Uri_PROPERTY_ZONE) {
4645 *pcchProperty = URLZONE_INVALID;
4649 if(uriProp < Uri_PROPERTY_DWORD_START) {
4651 return E_INVALIDARG;
4655 case Uri_PROPERTY_HOST_TYPE:
4656 *pcchProperty = This->host_type;
4659 case Uri_PROPERTY_PORT:
4660 if(!This->has_port) {
4664 *pcchProperty = This->port;
4669 case Uri_PROPERTY_SCHEME:
4670 *pcchProperty = This->scheme_type;
4674 FIXME("(%p)->(%d %p %x)\n", This, uriProp, pcchProperty, dwFlags);
4681 static HRESULT WINAPI Uri_HasProperty(IUri *iface, Uri_PROPERTY uriProp, BOOL *pfHasProperty)
4683 Uri *This = impl_from_IUri(iface);
4685 TRACE("(%p %s)->(%d %p)\n", This, debugstr_w(This->canon_uri), uriProp, pfHasProperty);
4688 return E_INVALIDARG;
4691 case Uri_PROPERTY_ABSOLUTE_URI:
4692 *pfHasProperty = !(This->display_modifiers & URI_DISPLAY_NO_ABSOLUTE_URI);
4694 case Uri_PROPERTY_AUTHORITY:
4695 *pfHasProperty = This->authority_start > -1;
4697 case Uri_PROPERTY_DISPLAY_URI:
4698 *pfHasProperty = TRUE;
4700 case Uri_PROPERTY_DOMAIN:
4701 *pfHasProperty = This->domain_offset > -1;
4703 case Uri_PROPERTY_EXTENSION:
4704 *pfHasProperty = This->extension_offset > -1;
4706 case Uri_PROPERTY_FRAGMENT:
4707 *pfHasProperty = This->fragment_start > -1;
4709 case Uri_PROPERTY_HOST:
4710 *pfHasProperty = This->host_start > -1;
4712 case Uri_PROPERTY_PASSWORD:
4713 *pfHasProperty = This->userinfo_split > -1;
4715 case Uri_PROPERTY_PATH:
4716 *pfHasProperty = This->path_start > -1;
4718 case Uri_PROPERTY_PATH_AND_QUERY:
4719 *pfHasProperty = (This->path_start > -1 || This->query_start > -1);
4721 case Uri_PROPERTY_QUERY:
4722 *pfHasProperty = This->query_start > -1;
4724 case Uri_PROPERTY_RAW_URI:
4725 *pfHasProperty = TRUE;
4727 case Uri_PROPERTY_SCHEME_NAME:
4728 *pfHasProperty = This->scheme_start > -1;
4730 case Uri_PROPERTY_USER_INFO:
4731 *pfHasProperty = This->userinfo_start > -1;
4733 case Uri_PROPERTY_USER_NAME:
4734 if(This->userinfo_split == 0)
4735 *pfHasProperty = FALSE;
4737 *pfHasProperty = This->userinfo_start > -1;
4739 case Uri_PROPERTY_HOST_TYPE:
4740 *pfHasProperty = TRUE;
4742 case Uri_PROPERTY_PORT:
4743 *pfHasProperty = This->has_port;
4745 case Uri_PROPERTY_SCHEME:
4746 *pfHasProperty = TRUE;
4748 case Uri_PROPERTY_ZONE:
4749 *pfHasProperty = FALSE;
4752 FIXME("(%p)->(%d %p): Unsupported property type.\n", This, uriProp, pfHasProperty);
4759 static HRESULT WINAPI Uri_GetAbsoluteUri(IUri *iface, BSTR *pstrAbsoluteUri)
4761 TRACE("(%p)->(%p)\n", iface, pstrAbsoluteUri);
4762 return IUri_GetPropertyBSTR(iface, Uri_PROPERTY_ABSOLUTE_URI, pstrAbsoluteUri, 0);
4765 static HRESULT WINAPI Uri_GetAuthority(IUri *iface, BSTR *pstrAuthority)
4767 TRACE("(%p)->(%p)\n", iface, pstrAuthority);
4768 return IUri_GetPropertyBSTR(iface, Uri_PROPERTY_AUTHORITY, pstrAuthority, 0);
4771 static HRESULT WINAPI Uri_GetDisplayUri(IUri *iface, BSTR *pstrDisplayUri)
4773 TRACE("(%p)->(%p)\n", iface, pstrDisplayUri);
4774 return IUri_GetPropertyBSTR(iface, Uri_PROPERTY_DISPLAY_URI, pstrDisplayUri, 0);
4777 static HRESULT WINAPI Uri_GetDomain(IUri *iface, BSTR *pstrDomain)
4779 TRACE("(%p)->(%p)\n", iface, pstrDomain);
4780 return IUri_GetPropertyBSTR(iface, Uri_PROPERTY_DOMAIN, pstrDomain, 0);
4783 static HRESULT WINAPI Uri_GetExtension(IUri *iface, BSTR *pstrExtension)
4785 TRACE("(%p)->(%p)\n", iface, pstrExtension);
4786 return IUri_GetPropertyBSTR(iface, Uri_PROPERTY_EXTENSION, pstrExtension, 0);
4789 static HRESULT WINAPI Uri_GetFragment(IUri *iface, BSTR *pstrFragment)
4791 TRACE("(%p)->(%p)\n", iface, pstrFragment);
4792 return IUri_GetPropertyBSTR(iface, Uri_PROPERTY_FRAGMENT, pstrFragment, 0);
4795 static HRESULT WINAPI Uri_GetHost(IUri *iface, BSTR *pstrHost)
4797 TRACE("(%p)->(%p)\n", iface, pstrHost);
4798 return IUri_GetPropertyBSTR(iface, Uri_PROPERTY_HOST, pstrHost, 0);
4801 static HRESULT WINAPI Uri_GetPassword(IUri *iface, BSTR *pstrPassword)
4803 TRACE("(%p)->(%p)\n", iface, pstrPassword);
4804 return IUri_GetPropertyBSTR(iface, Uri_PROPERTY_PASSWORD, pstrPassword, 0);
4807 static HRESULT WINAPI Uri_GetPath(IUri *iface, BSTR *pstrPath)
4809 TRACE("(%p)->(%p)\n", iface, pstrPath);
4810 return IUri_GetPropertyBSTR(iface, Uri_PROPERTY_PATH, pstrPath, 0);
4813 static HRESULT WINAPI Uri_GetPathAndQuery(IUri *iface, BSTR *pstrPathAndQuery)
4815 TRACE("(%p)->(%p)\n", iface, pstrPathAndQuery);
4816 return IUri_GetPropertyBSTR(iface, Uri_PROPERTY_PATH_AND_QUERY, pstrPathAndQuery, 0);
4819 static HRESULT WINAPI Uri_GetQuery(IUri *iface, BSTR *pstrQuery)
4821 TRACE("(%p)->(%p)\n", iface, pstrQuery);
4822 return IUri_GetPropertyBSTR(iface, Uri_PROPERTY_QUERY, pstrQuery, 0);
4825 static HRESULT WINAPI Uri_GetRawUri(IUri *iface, BSTR *pstrRawUri)
4827 TRACE("(%p)->(%p)\n", iface, pstrRawUri);
4828 return IUri_GetPropertyBSTR(iface, Uri_PROPERTY_RAW_URI, pstrRawUri, 0);
4831 static HRESULT WINAPI Uri_GetSchemeName(IUri *iface, BSTR *pstrSchemeName)
4833 TRACE("(%p)->(%p)\n", iface, pstrSchemeName);
4834 return IUri_GetPropertyBSTR(iface, Uri_PROPERTY_SCHEME_NAME, pstrSchemeName, 0);
4837 static HRESULT WINAPI Uri_GetUserInfo(IUri *iface, BSTR *pstrUserInfo)
4839 TRACE("(%p)->(%p)\n", iface, pstrUserInfo);
4840 return IUri_GetPropertyBSTR(iface, Uri_PROPERTY_USER_INFO, pstrUserInfo, 0);
4843 static HRESULT WINAPI Uri_GetUserName(IUri *iface, BSTR *pstrUserName)
4845 TRACE("(%p)->(%p)\n", iface, pstrUserName);
4846 return IUri_GetPropertyBSTR(iface, Uri_PROPERTY_USER_NAME, pstrUserName, 0);
4849 static HRESULT WINAPI Uri_GetHostType(IUri *iface, DWORD *pdwHostType)
4851 TRACE("(%p)->(%p)\n", iface, pdwHostType);
4852 return IUri_GetPropertyDWORD(iface, Uri_PROPERTY_HOST_TYPE, pdwHostType, 0);
4855 static HRESULT WINAPI Uri_GetPort(IUri *iface, DWORD *pdwPort)
4857 TRACE("(%p)->(%p)\n", iface, pdwPort);
4858 return IUri_GetPropertyDWORD(iface, Uri_PROPERTY_PORT, pdwPort, 0);
4861 static HRESULT WINAPI Uri_GetScheme(IUri *iface, DWORD *pdwScheme)
4863 TRACE("(%p)->(%p)\n", iface, pdwScheme);
4864 return IUri_GetPropertyDWORD(iface, Uri_PROPERTY_SCHEME, pdwScheme, 0);
4867 static HRESULT WINAPI Uri_GetZone(IUri *iface, DWORD *pdwZone)
4869 TRACE("(%p)->(%p)\n", iface, pdwZone);
4870 return IUri_GetPropertyDWORD(iface, Uri_PROPERTY_ZONE,pdwZone, 0);
4873 static HRESULT WINAPI Uri_GetProperties(IUri *iface, DWORD *pdwProperties)
4875 Uri *This = impl_from_IUri(iface);
4876 TRACE("(%p %s)->(%p)\n", This, debugstr_w(This->canon_uri), pdwProperties);
4879 return E_INVALIDARG;
4881 /* All URIs have these. */
4882 *pdwProperties = Uri_HAS_DISPLAY_URI|Uri_HAS_RAW_URI|Uri_HAS_SCHEME|Uri_HAS_HOST_TYPE;
4884 if(!(This->display_modifiers & URI_DISPLAY_NO_ABSOLUTE_URI))
4885 *pdwProperties |= Uri_HAS_ABSOLUTE_URI;
4887 if(This->scheme_start > -1)
4888 *pdwProperties |= Uri_HAS_SCHEME_NAME;
4890 if(This->authority_start > -1) {
4891 *pdwProperties |= Uri_HAS_AUTHORITY;
4892 if(This->userinfo_start > -1) {
4893 *pdwProperties |= Uri_HAS_USER_INFO;
4894 if(This->userinfo_split != 0)
4895 *pdwProperties |= Uri_HAS_USER_NAME;
4897 if(This->userinfo_split > -1)
4898 *pdwProperties |= Uri_HAS_PASSWORD;
4899 if(This->host_start > -1)
4900 *pdwProperties |= Uri_HAS_HOST;
4901 if(This->domain_offset > -1)
4902 *pdwProperties |= Uri_HAS_DOMAIN;
4906 *pdwProperties |= Uri_HAS_PORT;
4907 if(This->path_start > -1)
4908 *pdwProperties |= Uri_HAS_PATH|Uri_HAS_PATH_AND_QUERY;
4909 if(This->query_start > -1)
4910 *pdwProperties |= Uri_HAS_QUERY|Uri_HAS_PATH_AND_QUERY;
4912 if(This->extension_offset > -1)
4913 *pdwProperties |= Uri_HAS_EXTENSION;
4915 if(This->fragment_start > -1)
4916 *pdwProperties |= Uri_HAS_FRAGMENT;
4921 static HRESULT WINAPI Uri_IsEqual(IUri *iface, IUri *pUri, BOOL *pfEqual)
4923 Uri *This = impl_from_IUri(iface);
4926 TRACE("(%p %s)->(%p %p)\n", This, debugstr_w(This->canon_uri), pUri, pfEqual);
4934 /* For some reason Windows returns S_OK here... */
4938 /* Try to convert it to a Uri (allows for a more simple comparison). */
4939 if((other = get_uri_obj(pUri))) {
4940 TRACE("comparing to %s\n", debugstr_w(other->canon_uri));
4941 *pfEqual = are_equal_simple(This, other);
4943 /* Do it the hard way. */
4944 FIXME("(%p)->(%p %p) No support for unknown IUri's yet.\n", iface, pUri, pfEqual);
4951 static const IUriVtbl UriVtbl = {
4955 Uri_GetPropertyBSTR,
4956 Uri_GetPropertyLength,
4957 Uri_GetPropertyDWORD,
4968 Uri_GetPathAndQuery,
4982 static inline Uri* impl_from_IUriBuilderFactory(IUriBuilderFactory *iface)
4984 return CONTAINING_RECORD(iface, Uri, IUriBuilderFactory_iface);
4987 static HRESULT WINAPI UriBuilderFactory_QueryInterface(IUriBuilderFactory *iface, REFIID riid, void **ppv)
4989 Uri *This = impl_from_IUriBuilderFactory(iface);
4991 if(IsEqualGUID(&IID_IUnknown, riid)) {
4992 TRACE("(%p)->(IID_IUnknown %p)\n", This, ppv);
4993 *ppv = &This->IUriBuilderFactory_iface;
4994 }else if(IsEqualGUID(&IID_IUriBuilderFactory, riid)) {
4995 TRACE("(%p)->(IID_IUriBuilderFactory %p)\n", This, ppv);
4996 *ppv = &This->IUriBuilderFactory_iface;
4997 }else if(IsEqualGUID(&IID_IUri, riid)) {
4998 TRACE("(%p)->(IID_IUri %p)\n", This, ppv);
4999 *ppv = &This->IUri_iface;
5001 TRACE("(%p)->(%s %p)\n", This, debugstr_guid(riid), ppv);
5003 return E_NOINTERFACE;
5006 IUnknown_AddRef((IUnknown*)*ppv);
5010 static ULONG WINAPI UriBuilderFactory_AddRef(IUriBuilderFactory *iface)
5012 Uri *This = impl_from_IUriBuilderFactory(iface);
5013 LONG ref = InterlockedIncrement(&This->ref);
5015 TRACE("(%p) ref=%d\n", This, ref);
5020 static ULONG WINAPI UriBuilderFactory_Release(IUriBuilderFactory *iface)
5022 Uri *This = impl_from_IUriBuilderFactory(iface);
5023 LONG ref = InterlockedDecrement(&This->ref);
5025 TRACE("(%p) ref=%d\n", This, ref);
5028 destory_uri_obj(This);
5033 static HRESULT WINAPI UriBuilderFactory_CreateIUriBuilder(IUriBuilderFactory *iface,
5035 DWORD_PTR dwReserved,
5036 IUriBuilder **ppIUriBuilder)
5038 Uri *This = impl_from_IUriBuilderFactory(iface);
5039 TRACE("(%p)->(%08x %08x %p)\n", This, dwFlags, (DWORD)dwReserved, ppIUriBuilder);
5044 if(dwFlags || dwReserved) {
5045 *ppIUriBuilder = NULL;
5046 return E_INVALIDARG;
5049 return CreateIUriBuilder(NULL, 0, 0, ppIUriBuilder);
5052 static HRESULT WINAPI UriBuilderFactory_CreateInitializedIUriBuilder(IUriBuilderFactory *iface,
5054 DWORD_PTR dwReserved,
5055 IUriBuilder **ppIUriBuilder)
5057 Uri *This = impl_from_IUriBuilderFactory(iface);
5058 TRACE("(%p)->(%08x %08x %p)\n", This, dwFlags, (DWORD)dwReserved, ppIUriBuilder);
5063 if(dwFlags || dwReserved) {
5064 *ppIUriBuilder = NULL;
5065 return E_INVALIDARG;
5068 return CreateIUriBuilder(&This->IUri_iface, 0, 0, ppIUriBuilder);
5071 static const IUriBuilderFactoryVtbl UriBuilderFactoryVtbl = {
5072 UriBuilderFactory_QueryInterface,
5073 UriBuilderFactory_AddRef,
5074 UriBuilderFactory_Release,
5075 UriBuilderFactory_CreateIUriBuilder,
5076 UriBuilderFactory_CreateInitializedIUriBuilder
5079 static Uri* create_uri_obj(void) {
5080 Uri *ret = heap_alloc_zero(sizeof(Uri));
5082 ret->IUri_iface.lpVtbl = &UriVtbl;
5083 ret->IUriBuilderFactory_iface.lpVtbl = &UriBuilderFactoryVtbl;
5090 /***********************************************************************
5091 * CreateUri (urlmon.@)
5093 * Creates a new IUri object using the URI represented by pwzURI. This function
5094 * parses and validates the components of pwzURI and then canonicalizes the
5095 * parsed components.
5098 * pwzURI [I] The URI to parse, validate, and canonicalize.
5099 * dwFlags [I] Flags which can affect how the parsing/canonicalization is performed.
5100 * dwReserved [I] Reserved (not used).
5101 * ppURI [O] The resulting IUri after parsing/canonicalization occurs.
5104 * Success: Returns S_OK. ppURI contains the pointer to the newly allocated IUri.
5105 * Failure: E_INVALIDARG if there are invalid flag combinations in dwFlags, or an
5106 * invalid parameter, or pwzURI doesn't represent a valid URI.
5107 * E_OUTOFMEMORY if any memory allocation fails.
5111 * Uri_CREATE_CANONICALIZE, Uri_CREATE_DECODE_EXTRA_INFO, Uri_CREATE_CRACK_UNKNOWN_SCHEMES,
5112 * Uri_CREATE_PRE_PROCESS_HTML_URI, Uri_CREATE_NO_IE_SETTINGS.
5114 HRESULT WINAPI CreateUri(LPCWSTR pwzURI, DWORD dwFlags, DWORD_PTR dwReserved, IUri **ppURI)
5116 const DWORD supported_flags = Uri_CREATE_ALLOW_RELATIVE|Uri_CREATE_ALLOW_IMPLICIT_WILDCARD_SCHEME|
5117 Uri_CREATE_ALLOW_IMPLICIT_FILE_SCHEME|Uri_CREATE_NO_CANONICALIZE|Uri_CREATE_CANONICALIZE|
5118 Uri_CREATE_DECODE_EXTRA_INFO|Uri_CREATE_NO_DECODE_EXTRA_INFO|Uri_CREATE_CRACK_UNKNOWN_SCHEMES|
5119 Uri_CREATE_NO_CRACK_UNKNOWN_SCHEMES|Uri_CREATE_PRE_PROCESS_HTML_URI|Uri_CREATE_NO_PRE_PROCESS_HTML_URI|
5120 Uri_CREATE_NO_IE_SETTINGS|Uri_CREATE_NO_ENCODE_FORBIDDEN_CHARACTERS|Uri_CREATE_FILE_USE_DOS_PATH;
5125 TRACE("(%s %x %x %p)\n", debugstr_w(pwzURI), dwFlags, (DWORD)dwReserved, ppURI);
5128 return E_INVALIDARG;
5132 return E_INVALIDARG;
5135 /* Check for invalid flags. */
5136 if(has_invalid_flag_combination(dwFlags)) {
5138 return E_INVALIDARG;
5141 /* Currently unsupported. */
5142 if(dwFlags & ~supported_flags)
5143 FIXME("Ignoring unsupported flag(s) %x\n", dwFlags & ~supported_flags);
5145 ret = create_uri_obj();
5148 return E_OUTOFMEMORY;
5151 /* Explicitly set the default flags if it doesn't cause a flag conflict. */
5152 apply_default_flags(&dwFlags);
5154 /* Pre process the URI, unless told otherwise. */
5155 if(!(dwFlags & Uri_CREATE_NO_PRE_PROCESS_HTML_URI))
5156 ret->raw_uri = pre_process_uri(pwzURI);
5158 ret->raw_uri = SysAllocString(pwzURI);
5162 return E_OUTOFMEMORY;
5165 memset(&data, 0, sizeof(parse_data));
5166 data.uri = ret->raw_uri;
5168 /* Validate and parse the URI into it's components. */
5169 if(!parse_uri(&data, dwFlags)) {
5170 /* Encountered an unsupported or invalid URI */
5171 IUri_Release(&ret->IUri_iface);
5173 return E_INVALIDARG;
5176 /* Canonicalize the URI. */
5177 hr = canonicalize_uri(&data, ret, dwFlags);
5179 IUri_Release(&ret->IUri_iface);
5184 ret->create_flags = dwFlags;
5186 *ppURI = &ret->IUri_iface;
5190 /***********************************************************************
5191 * CreateUriWithFragment (urlmon.@)
5193 * Creates a new IUri object. This is almost the same as CreateUri, expect that
5194 * it allows you to explicitly specify a fragment (pwzFragment) for pwzURI.
5197 * pwzURI [I] The URI to parse and perform canonicalization on.
5198 * pwzFragment [I] The explicit fragment string which should be added to pwzURI.
5199 * dwFlags [I] The flags which will be passed to CreateUri.
5200 * dwReserved [I] Reserved (not used).
5201 * ppURI [O] The resulting IUri after parsing/canonicalization.
5204 * Success: S_OK. ppURI contains the pointer to the newly allocated IUri.
5205 * Failure: E_INVALIDARG if pwzURI already contains a fragment and pwzFragment
5206 * isn't NULL. Will also return E_INVALIDARG for the same reasons as
5207 * CreateUri will. E_OUTOFMEMORY if any allocation fails.
5209 HRESULT WINAPI CreateUriWithFragment(LPCWSTR pwzURI, LPCWSTR pwzFragment, DWORD dwFlags,
5210 DWORD_PTR dwReserved, IUri **ppURI)
5213 TRACE("(%s %s %x %x %p)\n", debugstr_w(pwzURI), debugstr_w(pwzFragment), dwFlags, (DWORD)dwReserved, ppURI);
5216 return E_INVALIDARG;
5220 return E_INVALIDARG;
5223 /* Check if a fragment should be appended to the URI string. */
5226 DWORD uri_len, frag_len;
5229 /* Check if the original URI already has a fragment component. */
5230 if(StrChrW(pwzURI, '#')) {
5232 return E_INVALIDARG;
5235 uri_len = lstrlenW(pwzURI);
5236 frag_len = lstrlenW(pwzFragment);
5238 /* If the fragment doesn't start with a '#', one will be added. */
5239 add_pound = *pwzFragment != '#';
5242 uriW = heap_alloc((uri_len+frag_len+2)*sizeof(WCHAR));
5244 uriW = heap_alloc((uri_len+frag_len+1)*sizeof(WCHAR));
5247 return E_OUTOFMEMORY;
5249 memcpy(uriW, pwzURI, uri_len*sizeof(WCHAR));
5251 uriW[uri_len++] = '#';
5252 memcpy(uriW+uri_len, pwzFragment, (frag_len+1)*sizeof(WCHAR));
5254 hres = CreateUri(uriW, dwFlags, 0, ppURI);
5258 /* A fragment string wasn't specified, so just forward the call. */
5259 hres = CreateUri(pwzURI, dwFlags, 0, ppURI);
5264 static HRESULT build_uri(const UriBuilder *builder, IUri **uri, DWORD create_flags,
5265 DWORD use_orig_flags, DWORD encoding_mask)
5274 if(encoding_mask && (!builder->uri || builder->modified_props)) {
5279 /* Decide what flags should be used when creating the Uri. */
5280 if((use_orig_flags & UriBuilder_USE_ORIGINAL_FLAGS) && builder->uri)
5281 create_flags = builder->uri->create_flags;
5283 if(has_invalid_flag_combination(create_flags)) {
5285 return E_INVALIDARG;
5288 /* Set the default flags if they don't cause a conflict. */
5289 apply_default_flags(&create_flags);
5292 /* Return the base IUri if no changes have been made and the create_flags match. */
5293 if(builder->uri && !builder->modified_props && builder->uri->create_flags == create_flags) {
5294 *uri = &builder->uri->IUri_iface;
5299 hr = validate_components(builder, &data, create_flags);
5305 ret = create_uri_obj();
5308 return E_OUTOFMEMORY;
5311 hr = generate_uri(builder, &data, ret, create_flags);
5313 IUri_Release(&ret->IUri_iface);
5318 *uri = &ret->IUri_iface;
5322 static inline UriBuilder* impl_from_IUriBuilder(IUriBuilder *iface)
5324 return CONTAINING_RECORD(iface, UriBuilder, IUriBuilder_iface);
5327 static HRESULT WINAPI UriBuilder_QueryInterface(IUriBuilder *iface, REFIID riid, void **ppv)
5329 UriBuilder *This = impl_from_IUriBuilder(iface);
5331 if(IsEqualGUID(&IID_IUnknown, riid)) {
5332 TRACE("(%p)->(IID_IUnknown %p)\n", This, ppv);
5333 *ppv = &This->IUriBuilder_iface;
5334 }else if(IsEqualGUID(&IID_IUriBuilder, riid)) {
5335 TRACE("(%p)->(IID_IUriBuilder %p)\n", This, ppv);
5336 *ppv = &This->IUriBuilder_iface;
5338 TRACE("(%p)->(%s %p)\n", This, debugstr_guid(riid), ppv);
5340 return E_NOINTERFACE;
5343 IUnknown_AddRef((IUnknown*)*ppv);
5347 static ULONG WINAPI UriBuilder_AddRef(IUriBuilder *iface)
5349 UriBuilder *This = impl_from_IUriBuilder(iface);
5350 LONG ref = InterlockedIncrement(&This->ref);
5352 TRACE("(%p) ref=%d\n", This, ref);
5357 static ULONG WINAPI UriBuilder_Release(IUriBuilder *iface)
5359 UriBuilder *This = impl_from_IUriBuilder(iface);
5360 LONG ref = InterlockedDecrement(&This->ref);
5362 TRACE("(%p) ref=%d\n", This, ref);
5365 if(This->uri) IUri_Release(&This->uri->IUri_iface);
5366 heap_free(This->fragment);
5367 heap_free(This->host);
5368 heap_free(This->password);
5369 heap_free(This->path);
5370 heap_free(This->query);
5371 heap_free(This->scheme);
5372 heap_free(This->username);
5379 static HRESULT WINAPI UriBuilder_CreateUriSimple(IUriBuilder *iface,
5380 DWORD dwAllowEncodingPropertyMask,
5381 DWORD_PTR dwReserved,
5384 UriBuilder *This = impl_from_IUriBuilder(iface);
5386 TRACE("(%p)->(%d %d %p)\n", This, dwAllowEncodingPropertyMask, (DWORD)dwReserved, ppIUri);
5388 hr = build_uri(This, ppIUri, 0, UriBuilder_USE_ORIGINAL_FLAGS, dwAllowEncodingPropertyMask);
5390 FIXME("(%p)->(%d %d %p)\n", This, dwAllowEncodingPropertyMask, (DWORD)dwReserved, ppIUri);
5394 static HRESULT WINAPI UriBuilder_CreateUri(IUriBuilder *iface,
5395 DWORD dwCreateFlags,
5396 DWORD dwAllowEncodingPropertyMask,
5397 DWORD_PTR dwReserved,
5400 UriBuilder *This = impl_from_IUriBuilder(iface);
5402 TRACE("(%p)->(0x%08x %d %d %p)\n", This, dwCreateFlags, dwAllowEncodingPropertyMask, (DWORD)dwReserved, ppIUri);
5404 if(dwCreateFlags == -1)
5405 hr = build_uri(This, ppIUri, 0, UriBuilder_USE_ORIGINAL_FLAGS, dwAllowEncodingPropertyMask);
5407 hr = build_uri(This, ppIUri, dwCreateFlags, 0, dwAllowEncodingPropertyMask);
5410 FIXME("(%p)->(0x%08x %d %d %p)\n", This, dwCreateFlags, dwAllowEncodingPropertyMask, (DWORD)dwReserved, ppIUri);
5414 static HRESULT WINAPI UriBuilder_CreateUriWithFlags(IUriBuilder *iface,
5415 DWORD dwCreateFlags,
5416 DWORD dwUriBuilderFlags,
5417 DWORD dwAllowEncodingPropertyMask,
5418 DWORD_PTR dwReserved,
5421 UriBuilder *This = impl_from_IUriBuilder(iface);
5423 TRACE("(%p)->(0x%08x 0x%08x %d %d %p)\n", This, dwCreateFlags, dwUriBuilderFlags,
5424 dwAllowEncodingPropertyMask, (DWORD)dwReserved, ppIUri);
5426 hr = build_uri(This, ppIUri, dwCreateFlags, dwUriBuilderFlags, dwAllowEncodingPropertyMask);
5428 FIXME("(%p)->(0x%08x 0x%08x %d %d %p)\n", This, dwCreateFlags, dwUriBuilderFlags,
5429 dwAllowEncodingPropertyMask, (DWORD)dwReserved, ppIUri);
5433 static HRESULT WINAPI UriBuilder_GetIUri(IUriBuilder *iface, IUri **ppIUri)
5435 UriBuilder *This = impl_from_IUriBuilder(iface);
5436 TRACE("(%p)->(%p)\n", This, ppIUri);
5442 IUri *uri = &This->uri->IUri_iface;
5451 static HRESULT WINAPI UriBuilder_SetIUri(IUriBuilder *iface, IUri *pIUri)
5453 UriBuilder *This = impl_from_IUriBuilder(iface);
5454 TRACE("(%p)->(%p)\n", This, pIUri);
5459 if((uri = get_uri_obj(pIUri))) {
5460 /* Only reset the builder if it's Uri isn't the same as
5461 * the Uri passed to the function.
5463 if(This->uri != uri) {
5464 reset_builder(This);
5468 This->port = uri->port;
5473 FIXME("(%p)->(%p) Unknown IUri types not supported yet.\n", This, pIUri);
5476 } else if(This->uri)
5477 /* Only reset the builder if it's Uri isn't NULL. */
5478 reset_builder(This);
5483 static HRESULT WINAPI UriBuilder_GetFragment(IUriBuilder *iface, DWORD *pcchFragment, LPCWSTR *ppwzFragment)
5485 UriBuilder *This = impl_from_IUriBuilder(iface);
5486 TRACE("(%p)->(%p %p)\n", This, pcchFragment, ppwzFragment);
5488 if(!This->uri || This->uri->fragment_start == -1 || This->modified_props & Uri_HAS_FRAGMENT)
5489 return get_builder_component(&This->fragment, &This->fragment_len, NULL, 0, ppwzFragment, pcchFragment);
5491 return get_builder_component(&This->fragment, &This->fragment_len, This->uri->canon_uri+This->uri->fragment_start,
5492 This->uri->fragment_len, ppwzFragment, pcchFragment);
5495 static HRESULT WINAPI UriBuilder_GetHost(IUriBuilder *iface, DWORD *pcchHost, LPCWSTR *ppwzHost)
5497 UriBuilder *This = impl_from_IUriBuilder(iface);
5498 TRACE("(%p)->(%p %p)\n", This, pcchHost, ppwzHost);
5500 if(!This->uri || This->uri->host_start == -1 || This->modified_props & Uri_HAS_HOST)
5501 return get_builder_component(&This->host, &This->host_len, NULL, 0, ppwzHost, pcchHost);
5503 if(This->uri->host_type == Uri_HOST_IPV6)
5504 /* Don't include the '[' and ']' around the address. */
5505 return get_builder_component(&This->host, &This->host_len, This->uri->canon_uri+This->uri->host_start+1,
5506 This->uri->host_len-2, ppwzHost, pcchHost);
5508 return get_builder_component(&This->host, &This->host_len, This->uri->canon_uri+This->uri->host_start,
5509 This->uri->host_len, ppwzHost, pcchHost);
5513 static HRESULT WINAPI UriBuilder_GetPassword(IUriBuilder *iface, DWORD *pcchPassword, LPCWSTR *ppwzPassword)
5515 UriBuilder *This = impl_from_IUriBuilder(iface);
5516 TRACE("(%p)->(%p %p)\n", This, pcchPassword, ppwzPassword);
5518 if(!This->uri || This->uri->userinfo_split == -1 || This->modified_props & Uri_HAS_PASSWORD)
5519 return get_builder_component(&This->password, &This->password_len, NULL, 0, ppwzPassword, pcchPassword);
5521 const WCHAR *start = This->uri->canon_uri+This->uri->userinfo_start+This->uri->userinfo_split+1;
5522 DWORD len = This->uri->userinfo_len-This->uri->userinfo_split-1;
5523 return get_builder_component(&This->password, &This->password_len, start, len, ppwzPassword, pcchPassword);
5527 static HRESULT WINAPI UriBuilder_GetPath(IUriBuilder *iface, DWORD *pcchPath, LPCWSTR *ppwzPath)
5529 UriBuilder *This = impl_from_IUriBuilder(iface);
5530 TRACE("(%p)->(%p %p)\n", This, pcchPath, ppwzPath);
5532 if(!This->uri || This->uri->path_start == -1 || This->modified_props & Uri_HAS_PATH)
5533 return get_builder_component(&This->path, &This->path_len, NULL, 0, ppwzPath, pcchPath);
5535 return get_builder_component(&This->path, &This->path_len, This->uri->canon_uri+This->uri->path_start,
5536 This->uri->path_len, ppwzPath, pcchPath);
5539 static HRESULT WINAPI UriBuilder_GetPort(IUriBuilder *iface, BOOL *pfHasPort, DWORD *pdwPort)
5541 UriBuilder *This = impl_from_IUriBuilder(iface);
5542 TRACE("(%p)->(%p %p)\n", This, pfHasPort, pdwPort);
5555 *pfHasPort = This->has_port;
5556 *pdwPort = This->port;
5560 static HRESULT WINAPI UriBuilder_GetQuery(IUriBuilder *iface, DWORD *pcchQuery, LPCWSTR *ppwzQuery)
5562 UriBuilder *This = impl_from_IUriBuilder(iface);
5563 TRACE("(%p)->(%p %p)\n", This, pcchQuery, ppwzQuery);
5565 if(!This->uri || This->uri->query_start == -1 || This->modified_props & Uri_HAS_QUERY)
5566 return get_builder_component(&This->query, &This->query_len, NULL, 0, ppwzQuery, pcchQuery);
5568 return get_builder_component(&This->query, &This->query_len, This->uri->canon_uri+This->uri->query_start,
5569 This->uri->query_len, ppwzQuery, pcchQuery);
5572 static HRESULT WINAPI UriBuilder_GetSchemeName(IUriBuilder *iface, DWORD *pcchSchemeName, LPCWSTR *ppwzSchemeName)
5574 UriBuilder *This = impl_from_IUriBuilder(iface);
5575 TRACE("(%p)->(%p %p)\n", This, pcchSchemeName, ppwzSchemeName);
5577 if(!This->uri || This->uri->scheme_start == -1 || This->modified_props & Uri_HAS_SCHEME_NAME)
5578 return get_builder_component(&This->scheme, &This->scheme_len, NULL, 0, ppwzSchemeName, pcchSchemeName);
5580 return get_builder_component(&This->scheme, &This->scheme_len, This->uri->canon_uri+This->uri->scheme_start,
5581 This->uri->scheme_len, ppwzSchemeName, pcchSchemeName);
5584 static HRESULT WINAPI UriBuilder_GetUserName(IUriBuilder *iface, DWORD *pcchUserName, LPCWSTR *ppwzUserName)
5586 UriBuilder *This = impl_from_IUriBuilder(iface);
5587 TRACE("(%p)->(%p %p)\n", This, pcchUserName, ppwzUserName);
5589 if(!This->uri || This->uri->userinfo_start == -1 || This->uri->userinfo_split == 0 ||
5590 This->modified_props & Uri_HAS_USER_NAME)
5591 return get_builder_component(&This->username, &This->username_len, NULL, 0, ppwzUserName, pcchUserName);
5593 const WCHAR *start = This->uri->canon_uri+This->uri->userinfo_start;
5595 /* Check if there's a password in the userinfo section. */
5596 if(This->uri->userinfo_split > -1)
5597 /* Don't include the password. */
5598 return get_builder_component(&This->username, &This->username_len, start,
5599 This->uri->userinfo_split, ppwzUserName, pcchUserName);
5601 return get_builder_component(&This->username, &This->username_len, start,
5602 This->uri->userinfo_len, ppwzUserName, pcchUserName);
5606 static HRESULT WINAPI UriBuilder_SetFragment(IUriBuilder *iface, LPCWSTR pwzNewValue)
5608 UriBuilder *This = impl_from_IUriBuilder(iface);
5609 TRACE("(%p)->(%s)\n", This, debugstr_w(pwzNewValue));
5610 return set_builder_component(&This->fragment, &This->fragment_len, pwzNewValue, '#',
5611 &This->modified_props, Uri_HAS_FRAGMENT);
5614 static HRESULT WINAPI UriBuilder_SetHost(IUriBuilder *iface, LPCWSTR pwzNewValue)
5616 UriBuilder *This = impl_from_IUriBuilder(iface);
5617 TRACE("(%p)->(%s)\n", This, debugstr_w(pwzNewValue));
5619 /* Host name can't be set to NULL. */
5621 return E_INVALIDARG;
5623 return set_builder_component(&This->host, &This->host_len, pwzNewValue, 0,
5624 &This->modified_props, Uri_HAS_HOST);
5627 static HRESULT WINAPI UriBuilder_SetPassword(IUriBuilder *iface, LPCWSTR pwzNewValue)
5629 UriBuilder *This = impl_from_IUriBuilder(iface);
5630 TRACE("(%p)->(%s)\n", This, debugstr_w(pwzNewValue));
5631 return set_builder_component(&This->password, &This->password_len, pwzNewValue, 0,
5632 &This->modified_props, Uri_HAS_PASSWORD);
5635 static HRESULT WINAPI UriBuilder_SetPath(IUriBuilder *iface, LPCWSTR pwzNewValue)
5637 UriBuilder *This = impl_from_IUriBuilder(iface);
5638 TRACE("(%p)->(%s)\n", This, debugstr_w(pwzNewValue));
5639 return set_builder_component(&This->path, &This->path_len, pwzNewValue, 0,
5640 &This->modified_props, Uri_HAS_PATH);
5643 static HRESULT WINAPI UriBuilder_SetPort(IUriBuilder *iface, BOOL fHasPort, DWORD dwNewValue)
5645 UriBuilder *This = impl_from_IUriBuilder(iface);
5646 TRACE("(%p)->(%d %d)\n", This, fHasPort, dwNewValue);
5648 This->has_port = fHasPort;
5649 This->port = dwNewValue;
5650 This->modified_props |= Uri_HAS_PORT;
5654 static HRESULT WINAPI UriBuilder_SetQuery(IUriBuilder *iface, LPCWSTR pwzNewValue)
5656 UriBuilder *This = impl_from_IUriBuilder(iface);
5657 TRACE("(%p)->(%s)\n", This, debugstr_w(pwzNewValue));
5658 return set_builder_component(&This->query, &This->query_len, pwzNewValue, '?',
5659 &This->modified_props, Uri_HAS_QUERY);
5662 static HRESULT WINAPI UriBuilder_SetSchemeName(IUriBuilder *iface, LPCWSTR pwzNewValue)
5664 UriBuilder *This = impl_from_IUriBuilder(iface);
5665 TRACE("(%p)->(%s)\n", This, debugstr_w(pwzNewValue));
5667 /* Only set the scheme name if it's not NULL or empty. */
5668 if(!pwzNewValue || !*pwzNewValue)
5669 return E_INVALIDARG;
5671 return set_builder_component(&This->scheme, &This->scheme_len, pwzNewValue, 0,
5672 &This->modified_props, Uri_HAS_SCHEME_NAME);
5675 static HRESULT WINAPI UriBuilder_SetUserName(IUriBuilder *iface, LPCWSTR pwzNewValue)
5677 UriBuilder *This = impl_from_IUriBuilder(iface);
5678 TRACE("(%p)->(%s)\n", This, debugstr_w(pwzNewValue));
5679 return set_builder_component(&This->username, &This->username_len, pwzNewValue, 0,
5680 &This->modified_props, Uri_HAS_USER_NAME);
5683 static HRESULT WINAPI UriBuilder_RemoveProperties(IUriBuilder *iface, DWORD dwPropertyMask)
5685 const DWORD accepted_flags = Uri_HAS_AUTHORITY|Uri_HAS_DOMAIN|Uri_HAS_EXTENSION|Uri_HAS_FRAGMENT|Uri_HAS_HOST|
5686 Uri_HAS_PASSWORD|Uri_HAS_PATH|Uri_HAS_PATH_AND_QUERY|Uri_HAS_QUERY|
5687 Uri_HAS_USER_INFO|Uri_HAS_USER_NAME;
5689 UriBuilder *This = impl_from_IUriBuilder(iface);
5690 TRACE("(%p)->(0x%08x)\n", This, dwPropertyMask);
5692 if(dwPropertyMask & ~accepted_flags)
5693 return E_INVALIDARG;
5695 if(dwPropertyMask & Uri_HAS_FRAGMENT)
5696 UriBuilder_SetFragment(iface, NULL);
5698 /* Even though you can't set the host name to NULL or an
5699 * empty string, you can still remove it... for some reason.
5701 if(dwPropertyMask & Uri_HAS_HOST)
5702 set_builder_component(&This->host, &This->host_len, NULL, 0,
5703 &This->modified_props, Uri_HAS_HOST);
5705 if(dwPropertyMask & Uri_HAS_PASSWORD)
5706 UriBuilder_SetPassword(iface, NULL);
5708 if(dwPropertyMask & Uri_HAS_PATH)
5709 UriBuilder_SetPath(iface, NULL);
5711 if(dwPropertyMask & Uri_HAS_PORT)
5712 UriBuilder_SetPort(iface, FALSE, 0);
5714 if(dwPropertyMask & Uri_HAS_QUERY)
5715 UriBuilder_SetQuery(iface, NULL);
5717 if(dwPropertyMask & Uri_HAS_USER_NAME)
5718 UriBuilder_SetUserName(iface, NULL);
5723 static HRESULT WINAPI UriBuilder_HasBeenModified(IUriBuilder *iface, BOOL *pfModified)
5725 UriBuilder *This = impl_from_IUriBuilder(iface);
5726 TRACE("(%p)->(%p)\n", This, pfModified);
5731 *pfModified = This->modified_props > 0;
5735 static const IUriBuilderVtbl UriBuilderVtbl = {
5736 UriBuilder_QueryInterface,
5739 UriBuilder_CreateUriSimple,
5740 UriBuilder_CreateUri,
5741 UriBuilder_CreateUriWithFlags,
5744 UriBuilder_GetFragment,
5746 UriBuilder_GetPassword,
5749 UriBuilder_GetQuery,
5750 UriBuilder_GetSchemeName,
5751 UriBuilder_GetUserName,
5752 UriBuilder_SetFragment,
5754 UriBuilder_SetPassword,
5757 UriBuilder_SetQuery,
5758 UriBuilder_SetSchemeName,
5759 UriBuilder_SetUserName,
5760 UriBuilder_RemoveProperties,
5761 UriBuilder_HasBeenModified,
5764 /***********************************************************************
5765 * CreateIUriBuilder (urlmon.@)
5767 HRESULT WINAPI CreateIUriBuilder(IUri *pIUri, DWORD dwFlags, DWORD_PTR dwReserved, IUriBuilder **ppIUriBuilder)
5771 TRACE("(%p %x %x %p)\n", pIUri, dwFlags, (DWORD)dwReserved, ppIUriBuilder);
5776 ret = heap_alloc_zero(sizeof(UriBuilder));
5778 return E_OUTOFMEMORY;
5780 ret->IUriBuilder_iface.lpVtbl = &UriBuilderVtbl;
5786 if((uri = get_uri_obj(pIUri))) {
5791 /* Windows doesn't set 'has_port' to TRUE in this case. */
5792 ret->port = uri->port;
5796 *ppIUriBuilder = NULL;
5797 FIXME("(%p %x %x %p): Unknown IUri types not supported yet.\n", pIUri, dwFlags,
5798 (DWORD)dwReserved, ppIUriBuilder);
5803 *ppIUriBuilder = &ret->IUriBuilder_iface;
5807 /* Merges the base path with the relative path and stores the resulting path
5808 * and path len in 'result' and 'result_len'.
5810 static HRESULT merge_paths(parse_data *data, const WCHAR *base, DWORD base_len, const WCHAR *relative,
5811 DWORD relative_len, WCHAR **result, DWORD *result_len, DWORD flags)
5813 const WCHAR *end = NULL;
5814 DWORD base_copy_len = 0;
5818 /* Find the characters that will be copied over from
5821 end = memrchrW(base, '/', base_len);
5822 if(!end && data->scheme_type == URL_SCHEME_FILE)
5823 /* Try looking for a '\\'. */
5824 end = memrchrW(base, '\\', base_len);
5828 base_copy_len = (end+1)-base;
5829 *result = heap_alloc((base_copy_len+relative_len+1)*sizeof(WCHAR));
5831 *result = heap_alloc((relative_len+1)*sizeof(WCHAR));
5835 return E_OUTOFMEMORY;
5840 memcpy(ptr, base, base_copy_len*sizeof(WCHAR));
5841 ptr += base_copy_len;
5844 memcpy(ptr, relative, relative_len*sizeof(WCHAR));
5845 ptr += relative_len;
5848 *result_len = (ptr-*result);
5852 static HRESULT combine_uri(Uri *base, Uri *relative, DWORD flags, IUri **result, DWORD extras) {
5856 DWORD create_flags = 0, len = 0;
5858 memset(&data, 0, sizeof(parse_data));
5860 /* Base case is when the relative Uri has a scheme name,
5861 * if it does, then 'result' will contain the same data
5862 * as the relative Uri.
5864 if(relative->scheme_start > -1) {
5865 data.uri = SysAllocString(relative->raw_uri);
5868 return E_OUTOFMEMORY;
5871 parse_uri(&data, 0);
5873 ret = create_uri_obj();
5876 return E_OUTOFMEMORY;
5879 if(extras & COMBINE_URI_FORCE_FLAG_USE) {
5880 if(flags & URL_DONT_SIMPLIFY)
5881 create_flags |= Uri_CREATE_NO_CANONICALIZE;
5882 if(flags & URL_DONT_UNESCAPE_EXTRA_INFO)
5883 create_flags |= Uri_CREATE_NO_DECODE_EXTRA_INFO;
5886 ret->raw_uri = data.uri;
5887 hr = canonicalize_uri(&data, ret, create_flags);
5889 IUri_Release(&ret->IUri_iface);
5894 apply_default_flags(&create_flags);
5895 ret->create_flags = create_flags;
5897 *result = &ret->IUri_iface;
5900 DWORD raw_flags = 0;
5902 if(base->scheme_start > -1) {
5903 data.scheme = base->canon_uri+base->scheme_start;
5904 data.scheme_len = base->scheme_len;
5905 data.scheme_type = base->scheme_type;
5907 data.is_relative = TRUE;
5908 data.scheme_type = URL_SCHEME_UNKNOWN;
5909 create_flags |= Uri_CREATE_ALLOW_RELATIVE;
5912 if(base->authority_start > -1) {
5913 if(base->userinfo_start > -1 && base->userinfo_split != 0) {
5914 data.username = base->canon_uri+base->userinfo_start;
5915 data.username_len = (base->userinfo_split > -1) ? base->userinfo_split : base->userinfo_len;
5918 if(base->userinfo_split > -1) {
5919 data.password = base->canon_uri+base->userinfo_start+base->userinfo_split+1;
5920 data.password_len = base->userinfo_len-base->userinfo_split-1;
5923 if(base->host_start > -1) {
5924 data.host = base->canon_uri+base->host_start;
5925 data.host_len = base->host_len;
5926 data.host_type = base->host_type;
5929 if(base->has_port) {
5930 data.has_port = TRUE;
5931 data.port_value = base->port;
5933 } else if(base->scheme_type != URL_SCHEME_FILE)
5934 data.is_opaque = TRUE;
5936 if(relative->path_start == -1 || !relative->path_len) {
5937 if(base->path_start > -1) {
5938 data.path = base->canon_uri+base->path_start;
5939 data.path_len = base->path_len;
5940 } else if((base->path_start == -1 || !base->path_len) && !data.is_opaque) {
5941 /* Just set the path as a '/' if the base didn't have
5942 * one and if it's an hierarchical URI.
5944 static const WCHAR slashW[] = {'/',0};
5949 if(relative->query_start > -1) {
5950 data.query = relative->canon_uri+relative->query_start;
5951 data.query_len = relative->query_len;
5952 } else if(base->query_start > -1) {
5953 data.query = base->canon_uri+base->query_start;
5954 data.query_len = base->query_len;
5957 const WCHAR *ptr, **pptr;
5958 DWORD path_offset = 0, path_len = 0;
5960 /* There's two possibilities on what will happen to the path component
5961 * of the result IUri. First, if the relative path begins with a '/'
5962 * then the resulting path will just be the relative path. Second, if
5963 * relative path doesn't begin with a '/' then the base path and relative
5964 * path are merged together.
5966 if(relative->path_len && *(relative->canon_uri+relative->path_start) == '/') {
5968 BOOL copy_drive_path = FALSE;
5970 /* If the relative IUri's path starts with a '/', then we
5971 * don't use the base IUri's path. Unless the base IUri
5972 * is a file URI, in which case it uses the drive path of
5973 * the base IUri (if it has any) in the new path.
5975 if(base->scheme_type == URL_SCHEME_FILE) {
5976 if(base->path_len > 3 && *(base->canon_uri+base->path_start) == '/' &&
5977 is_drive_path(base->canon_uri+base->path_start+1)) {
5979 copy_drive_path = TRUE;
5983 path_len += relative->path_len;
5985 path = heap_alloc((path_len+1)*sizeof(WCHAR));
5988 return E_OUTOFMEMORY;
5993 /* Copy the base paths, drive path over. */
5994 if(copy_drive_path) {
5995 memcpy(tmp, base->canon_uri+base->path_start, 3*sizeof(WCHAR));
5999 memcpy(tmp, relative->canon_uri+relative->path_start, relative->path_len*sizeof(WCHAR));
6000 path[path_len] = '\0';
6002 /* Merge the base path with the relative path. */
6003 hr = merge_paths(&data, base->canon_uri+base->path_start, base->path_len,
6004 relative->canon_uri+relative->path_start, relative->path_len,
6005 &path, &path_len, flags);
6011 /* If the resulting IUri is a file URI, the drive path isn't
6012 * reduced out when the dot segments are removed.
6014 if(path_len >= 3 && data.scheme_type == URL_SCHEME_FILE && !data.host) {
6015 if(*path == '/' && is_drive_path(path+1))
6017 else if(is_drive_path(path))
6022 /* Check if the dot segments need to be removed from the path. */
6023 if(!(flags & URL_DONT_SIMPLIFY) && !data.is_opaque) {
6024 DWORD offset = (path_offset > 0) ? path_offset+1 : 0;
6025 DWORD new_len = remove_dot_segments(path+offset,path_len-offset);
6027 if(new_len != path_len) {
6028 WCHAR *tmp = heap_realloc(path, (offset+new_len+1)*sizeof(WCHAR));
6032 return E_OUTOFMEMORY;
6035 tmp[new_len+offset] = '\0';
6037 path_len = new_len+offset;
6041 if(relative->query_start > -1) {
6042 data.query = relative->canon_uri+relative->query_start;
6043 data.query_len = relative->query_len;
6046 /* Make sure the path component is valid. */
6049 if((data.is_opaque && !parse_path_opaque(pptr, &data, 0)) ||
6050 (!data.is_opaque && !parse_path_hierarchical(pptr, &data, 0))) {
6053 return E_INVALIDARG;
6057 if(relative->fragment_start > -1) {
6058 data.fragment = relative->canon_uri+relative->fragment_start;
6059 data.fragment_len = relative->fragment_len;
6062 if(flags & URL_DONT_SIMPLIFY)
6063 raw_flags |= RAW_URI_FORCE_PORT_DISP;
6064 if(flags & URL_FILE_USE_PATHURL)
6065 raw_flags |= RAW_URI_CONVERT_TO_DOS_PATH;
6067 len = generate_raw_uri(&data, data.uri, raw_flags);
6068 data.uri = SysAllocStringLen(NULL, len);
6072 return E_OUTOFMEMORY;
6075 generate_raw_uri(&data, data.uri, raw_flags);
6077 ret = create_uri_obj();
6079 SysFreeString(data.uri);
6082 return E_OUTOFMEMORY;
6085 if(flags & URL_DONT_SIMPLIFY)
6086 create_flags |= Uri_CREATE_NO_CANONICALIZE;
6087 if(flags & URL_FILE_USE_PATHURL)
6088 create_flags |= Uri_CREATE_FILE_USE_DOS_PATH;
6090 ret->raw_uri = data.uri;
6091 hr = canonicalize_uri(&data, ret, create_flags);
6093 IUri_Release(&ret->IUri_iface);
6098 if(flags & URL_DONT_SIMPLIFY)
6099 ret->display_modifiers |= URI_DISPLAY_NO_DEFAULT_PORT_AUTH;
6101 apply_default_flags(&create_flags);
6102 ret->create_flags = create_flags;
6103 *result = &ret->IUri_iface;
6111 /***********************************************************************
6112 * CoInternetCombineIUri (urlmon.@)
6114 HRESULT WINAPI CoInternetCombineIUri(IUri *pBaseUri, IUri *pRelativeUri, DWORD dwCombineFlags,
6115 IUri **ppCombinedUri, DWORD_PTR dwReserved)
6118 IInternetProtocolInfo *info;
6119 Uri *relative, *base;
6120 TRACE("(%p %p %x %p %x)\n", pBaseUri, pRelativeUri, dwCombineFlags, ppCombinedUri, (DWORD)dwReserved);
6123 return E_INVALIDARG;
6125 if(!pBaseUri || !pRelativeUri) {
6126 *ppCombinedUri = NULL;
6127 return E_INVALIDARG;
6130 relative = get_uri_obj(pRelativeUri);
6131 base = get_uri_obj(pBaseUri);
6132 if(!relative || !base) {
6133 *ppCombinedUri = NULL;
6134 FIXME("(%p %p %x %p %x) Unknown IUri types not supported yet.\n",
6135 pBaseUri, pRelativeUri, dwCombineFlags, ppCombinedUri, (DWORD)dwReserved);
6139 info = get_protocol_info(base->canon_uri);
6141 WCHAR result[INTERNET_MAX_URL_LENGTH+1];
6142 DWORD result_len = 0;
6144 hr = IInternetProtocolInfo_CombineUrl(info, base->canon_uri, relative->canon_uri, dwCombineFlags,
6145 result, INTERNET_MAX_URL_LENGTH+1, &result_len, 0);
6146 IInternetProtocolInfo_Release(info);
6148 hr = CreateUri(result, Uri_CREATE_ALLOW_RELATIVE, 0, ppCombinedUri);
6154 return combine_uri(base, relative, dwCombineFlags, ppCombinedUri, 0);
6157 /***********************************************************************
6158 * CoInternetCombineUrlEx (urlmon.@)
6160 HRESULT WINAPI CoInternetCombineUrlEx(IUri *pBaseUri, LPCWSTR pwzRelativeUrl, DWORD dwCombineFlags,
6161 IUri **ppCombinedUri, DWORD_PTR dwReserved)
6166 IInternetProtocolInfo *info;
6168 TRACE("(%p %s %x %p %x) stub\n", pBaseUri, debugstr_w(pwzRelativeUrl), dwCombineFlags,
6169 ppCombinedUri, (DWORD)dwReserved);
6174 if(!pwzRelativeUrl) {
6175 *ppCombinedUri = NULL;
6176 return E_UNEXPECTED;
6180 *ppCombinedUri = NULL;
6181 return E_INVALIDARG;
6184 base = get_uri_obj(pBaseUri);
6186 *ppCombinedUri = NULL;
6187 FIXME("(%p %s %x %p %x) Unknown IUri's not supported yet.\n", pBaseUri, debugstr_w(pwzRelativeUrl),
6188 dwCombineFlags, ppCombinedUri, (DWORD)dwReserved);
6192 info = get_protocol_info(base->canon_uri);
6194 WCHAR result[INTERNET_MAX_URL_LENGTH+1];
6195 DWORD result_len = 0;
6197 hr = IInternetProtocolInfo_CombineUrl(info, base->canon_uri, pwzRelativeUrl, dwCombineFlags,
6198 result, INTERNET_MAX_URL_LENGTH+1, &result_len, 0);
6199 IInternetProtocolInfo_Release(info);
6201 hr = CreateUri(result, Uri_CREATE_ALLOW_RELATIVE, 0, ppCombinedUri);
6207 hr = CreateUri(pwzRelativeUrl, Uri_CREATE_ALLOW_RELATIVE, 0, &relative);
6209 *ppCombinedUri = NULL;
6213 hr = combine_uri(base, get_uri_obj(relative), dwCombineFlags, ppCombinedUri, COMBINE_URI_FORCE_FLAG_USE);
6215 IUri_Release(relative);
6219 static HRESULT parse_canonicalize(const Uri *uri, DWORD flags, LPWSTR output,
6220 DWORD output_len, DWORD *result_len)
6222 const WCHAR *ptr = NULL;
6225 WCHAR buffer[INTERNET_MAX_URL_LENGTH+1];
6229 /* URL_UNESCAPE only has effect if none of the URL_ESCAPE flags are set. */
6230 const BOOL allow_unescape = !(flags & URL_ESCAPE_UNSAFE) &&
6231 !(flags & URL_ESCAPE_SPACES_ONLY) &&
6232 !(flags & URL_ESCAPE_PERCENT);
6235 /* Check if the dot segments need to be removed from the
6238 if(uri->scheme_start > -1 && uri->path_start > -1) {
6239 ptr = uri->canon_uri+uri->scheme_start+uri->scheme_len+1;
6242 reduce_path = !(flags & URL_NO_META) &&
6243 !(flags & URL_DONT_SIMPLIFY) &&
6244 ptr && check_hierarchical(pptr);
6246 for(ptr = uri->canon_uri; ptr < uri->canon_uri+uri->canon_len; ++ptr) {
6247 BOOL do_default_action = TRUE;
6249 /* Keep track of the path if we need to remove dot segments from
6252 if(reduce_path && !path && ptr == uri->canon_uri+uri->path_start)
6255 /* Check if it's time to reduce the path. */
6256 if(reduce_path && ptr == uri->canon_uri+uri->path_start+uri->path_len) {
6257 DWORD current_path_len = (buffer+len) - path;
6258 DWORD new_path_len = remove_dot_segments(path, current_path_len);
6260 /* Update the current length. */
6261 len -= (current_path_len-new_path_len);
6262 reduce_path = FALSE;
6266 const WCHAR decoded = decode_pct_val(ptr);
6268 if(allow_unescape && (flags & URL_UNESCAPE)) {
6269 buffer[len++] = decoded;
6271 do_default_action = FALSE;
6275 /* See if %'s needed to encoded. */
6276 if(do_default_action && (flags & URL_ESCAPE_PERCENT)) {
6277 pct_encode_val(*ptr, buffer+len);
6279 do_default_action = FALSE;
6281 } else if(*ptr == ' ') {
6282 if((flags & URL_ESCAPE_SPACES_ONLY) &&
6283 !(flags & URL_ESCAPE_UNSAFE)) {
6284 pct_encode_val(*ptr, buffer+len);
6286 do_default_action = FALSE;
6288 } else if(!is_reserved(*ptr) && !is_unreserved(*ptr)) {
6289 if(flags & URL_ESCAPE_UNSAFE) {
6290 pct_encode_val(*ptr, buffer+len);
6292 do_default_action = FALSE;
6296 if(do_default_action)
6297 buffer[len++] = *ptr;
6300 /* Sometimes the path is the very last component of the IUri, so
6301 * see if the dot segments need to be reduced now.
6303 if(reduce_path && path) {
6304 DWORD current_path_len = (buffer+len) - path;
6305 DWORD new_path_len = remove_dot_segments(path, current_path_len);
6307 /* Update the current length. */
6308 len -= (current_path_len-new_path_len);
6313 /* The null terminator isn't included in the length. */
6314 *result_len = len-1;
6315 if(len > output_len)
6316 return STRSAFE_E_INSUFFICIENT_BUFFER;
6318 memcpy(output, buffer, len*sizeof(WCHAR));
6323 static HRESULT parse_friendly(IUri *uri, LPWSTR output, DWORD output_len,
6330 hr = IUri_GetPropertyLength(uri, Uri_PROPERTY_DISPLAY_URI, &display_len, 0);
6336 *result_len = display_len;
6337 if(display_len+1 > output_len)
6338 return STRSAFE_E_INSUFFICIENT_BUFFER;
6340 hr = IUri_GetDisplayUri(uri, &display);
6346 memcpy(output, display, (display_len+1)*sizeof(WCHAR));
6347 SysFreeString(display);
6351 static HRESULT parse_rootdocument(const Uri *uri, LPWSTR output, DWORD output_len,
6354 static const WCHAR colon_slashesW[] = {':','/','/'};
6359 /* Windows only returns the root document if the URI has an authority
6360 * and it's not an unknown scheme type or a file scheme type.
6362 if(uri->authority_start == -1 ||
6363 uri->scheme_type == URL_SCHEME_UNKNOWN ||
6364 uri->scheme_type == URL_SCHEME_FILE) {
6367 return STRSAFE_E_INSUFFICIENT_BUFFER;
6373 len = uri->scheme_len+uri->authority_len;
6374 /* For the "://" and '/' which will be added. */
6377 if(len+1 > output_len) {
6379 return STRSAFE_E_INSUFFICIENT_BUFFER;
6383 memcpy(ptr, uri->canon_uri+uri->scheme_start, uri->scheme_len*sizeof(WCHAR));
6385 /* Add the "://". */
6386 ptr += uri->scheme_len;
6387 memcpy(ptr, colon_slashesW, sizeof(colon_slashesW));
6389 /* Add the authority. */
6390 ptr += sizeof(colon_slashesW)/sizeof(WCHAR);
6391 memcpy(ptr, uri->canon_uri+uri->authority_start, uri->authority_len*sizeof(WCHAR));
6393 /* Add the '/' after the authority. */
6394 ptr += uri->authority_len;
6402 static HRESULT parse_document(const Uri *uri, LPWSTR output, DWORD output_len,
6407 /* It has to be a known scheme type, but, it can't be a file
6408 * scheme. It also has to hierarchical.
6410 if(uri->scheme_type == URL_SCHEME_UNKNOWN ||
6411 uri->scheme_type == URL_SCHEME_FILE ||
6412 uri->authority_start == -1) {
6415 return STRSAFE_E_INSUFFICIENT_BUFFER;
6421 if(uri->fragment_start > -1)
6422 len = uri->fragment_start;
6424 len = uri->canon_len;
6427 if(len+1 > output_len)
6428 return STRSAFE_E_INSUFFICIENT_BUFFER;
6430 memcpy(output, uri->canon_uri, len*sizeof(WCHAR));
6435 static HRESULT parse_path_from_url(const Uri *uri, LPWSTR output, DWORD output_len,
6438 const WCHAR *path_ptr;
6439 WCHAR buffer[INTERNET_MAX_URL_LENGTH+1];
6442 if(uri->scheme_type != URL_SCHEME_FILE) {
6446 return E_INVALIDARG;
6450 if(uri->host_start > -1) {
6451 static const WCHAR slash_slashW[] = {'\\','\\'};
6453 memcpy(ptr, slash_slashW, sizeof(slash_slashW));
6454 ptr += sizeof(slash_slashW)/sizeof(WCHAR);
6455 memcpy(ptr, uri->canon_uri+uri->host_start, uri->host_len*sizeof(WCHAR));
6456 ptr += uri->host_len;
6459 path_ptr = uri->canon_uri+uri->path_start;
6460 if(uri->path_len > 3 && *path_ptr == '/' && is_drive_path(path_ptr+1))
6461 /* Skip past the '/' in front of the drive path. */
6464 for(; path_ptr < uri->canon_uri+uri->path_start+uri->path_len; ++path_ptr, ++ptr) {
6465 BOOL do_default_action = TRUE;
6467 if(*path_ptr == '%') {
6468 const WCHAR decoded = decode_pct_val(path_ptr);
6472 do_default_action = FALSE;
6474 } else if(*path_ptr == '/') {
6476 do_default_action = FALSE;
6479 if(do_default_action)
6485 *result_len = ptr-buffer;
6486 if(*result_len+1 > output_len)
6487 return STRSAFE_E_INSUFFICIENT_BUFFER;
6489 memcpy(output, buffer, (*result_len+1)*sizeof(WCHAR));
6493 static HRESULT parse_url_from_path(IUri *uri, LPWSTR output, DWORD output_len,
6500 hr = IUri_GetPropertyLength(uri, Uri_PROPERTY_ABSOLUTE_URI, &len, 0);
6507 if(len+1 > output_len)
6508 return STRSAFE_E_INSUFFICIENT_BUFFER;
6510 hr = IUri_GetAbsoluteUri(uri, &received);
6516 memcpy(output, received, (len+1)*sizeof(WCHAR));
6517 SysFreeString(received);
6522 static HRESULT parse_schema(IUri *uri, LPWSTR output, DWORD output_len,
6529 hr = IUri_GetPropertyLength(uri, Uri_PROPERTY_SCHEME_NAME, &len, 0);
6536 if(len+1 > output_len)
6537 return STRSAFE_E_INSUFFICIENT_BUFFER;
6539 hr = IUri_GetSchemeName(uri, &received);
6545 memcpy(output, received, (len+1)*sizeof(WCHAR));
6546 SysFreeString(received);
6551 static HRESULT parse_site(IUri *uri, LPWSTR output, DWORD output_len, DWORD *result_len)
6557 hr = IUri_GetPropertyLength(uri, Uri_PROPERTY_HOST, &len, 0);
6564 if(len+1 > output_len)
6565 return STRSAFE_E_INSUFFICIENT_BUFFER;
6567 hr = IUri_GetHost(uri, &received);
6573 memcpy(output, received, (len+1)*sizeof(WCHAR));
6574 SysFreeString(received);
6579 static HRESULT parse_domain(IUri *uri, LPWSTR output, DWORD output_len, DWORD *result_len)
6585 hr = IUri_GetPropertyLength(uri, Uri_PROPERTY_DOMAIN, &len, 0);
6592 if(len+1 > output_len)
6593 return STRSAFE_E_INSUFFICIENT_BUFFER;
6595 hr = IUri_GetDomain(uri, &received);
6601 memcpy(output, received, (len+1)*sizeof(WCHAR));
6602 SysFreeString(received);
6607 static HRESULT parse_anchor(IUri *uri, LPWSTR output, DWORD output_len, DWORD *result_len)
6613 hr = IUri_GetPropertyLength(uri, Uri_PROPERTY_FRAGMENT, &len, 0);
6620 if(len+1 > output_len)
6621 return STRSAFE_E_INSUFFICIENT_BUFFER;
6623 hr = IUri_GetFragment(uri, &received);
6629 memcpy(output, received, (len+1)*sizeof(WCHAR));
6630 SysFreeString(received);
6635 /***********************************************************************
6636 * CoInternetParseIUri (urlmon.@)
6638 HRESULT WINAPI CoInternetParseIUri(IUri *pIUri, PARSEACTION ParseAction, DWORD dwFlags,
6639 LPWSTR pwzResult, DWORD cchResult, DWORD *pcchResult,
6640 DWORD_PTR dwReserved)
6644 IInternetProtocolInfo *info;
6646 TRACE("(%p %d %x %p %d %p %x)\n", pIUri, ParseAction, dwFlags, pwzResult,
6647 cchResult, pcchResult, (DWORD)dwReserved);
6652 if(!pwzResult || !pIUri) {
6654 return E_INVALIDARG;
6657 if(!(uri = get_uri_obj(pIUri))) {
6659 FIXME("(%p %d %x %p %d %p %x) Unknown IUri's not supported for this action.\n",
6660 pIUri, ParseAction, dwFlags, pwzResult, cchResult, pcchResult, (DWORD)dwReserved);
6664 info = get_protocol_info(uri->canon_uri);
6666 hr = IInternetProtocolInfo_ParseUrl(info, uri->canon_uri, ParseAction, dwFlags,
6667 pwzResult, cchResult, pcchResult, 0);
6668 IInternetProtocolInfo_Release(info);
6669 if(SUCCEEDED(hr)) return hr;
6672 switch(ParseAction) {
6673 case PARSE_CANONICALIZE:
6674 hr = parse_canonicalize(uri, dwFlags, pwzResult, cchResult, pcchResult);
6676 case PARSE_FRIENDLY:
6677 hr = parse_friendly(pIUri, pwzResult, cchResult, pcchResult);
6679 case PARSE_ROOTDOCUMENT:
6680 hr = parse_rootdocument(uri, pwzResult, cchResult, pcchResult);
6682 case PARSE_DOCUMENT:
6683 hr = parse_document(uri, pwzResult, cchResult, pcchResult);
6685 case PARSE_PATH_FROM_URL:
6686 hr = parse_path_from_url(uri, pwzResult, cchResult, pcchResult);
6688 case PARSE_URL_FROM_PATH:
6689 hr = parse_url_from_path(pIUri, pwzResult, cchResult, pcchResult);
6692 hr = parse_schema(pIUri, pwzResult, cchResult, pcchResult);
6695 hr = parse_site(pIUri, pwzResult, cchResult, pcchResult);
6698 hr = parse_domain(pIUri, pwzResult, cchResult, pcchResult);
6700 case PARSE_LOCATION:
6702 hr = parse_anchor(pIUri, pwzResult, cchResult, pcchResult);
6704 case PARSE_SECURITY_URL:
6707 case PARSE_SECURITY_DOMAIN:
6714 FIXME("(%p %d %x %p %d %p %x) Partial stub.\n", pIUri, ParseAction, dwFlags,
6715 pwzResult, cchResult, pcchResult, (DWORD)dwReserved);