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
26 #define UINT_MAX 0xffffffff
27 #define USHORT_MAX 0xffff
29 #define ALLOW_NULL_TERM_SCHEME 0x01
30 #define ALLOW_NULL_TERM_USER_NAME 0x02
31 #define ALLOW_NULL_TERM_PASSWORD 0x04
32 #define ALLOW_BRACKETLESS_IP_LITERAL 0x08
33 #define SKIP_IP_FUTURE_CHECK 0x10
34 #define IGNORE_PORT_DELIMITER 0x20
36 WINE_DEFAULT_DEBUG_CHANNEL(urlmon
);
38 static const IID IID_IUriObj
= {0x4b364760,0x9f51,0x11df,{0x98,0x1c,0x08,0x00,0x20,0x0c,0x9a,0x66}};
41 const IUriVtbl
*lpIUriVtbl
;
46 /* Information about the canonicalized URI's buffer. */
50 BOOL display_absolute
;
55 URL_SCHEME scheme_type
;
63 Uri_HOST_TYPE host_type
;
85 const IUriBuilderVtbl
*lpIUriBuilderVtbl
;
122 /* IPv6 addresses can hold up to 8 h16 components. */
126 /* An IPv6 can have 1 elision ("::"). */
127 const WCHAR
*elision
;
129 /* An IPv6 can contain 1 IPv4 address as the last 32bits of the address. */
142 BOOL has_implicit_scheme
;
143 BOOL has_implicit_ip
;
148 URL_SCHEME scheme_type
;
150 const WCHAR
*username
;
153 const WCHAR
*password
;
158 Uri_HOST_TYPE host_type
;
161 ipv6_address ipv6_address
;
174 const WCHAR
*fragment
;
178 static const CHAR hexDigits
[] = "0123456789ABCDEF";
180 /* List of scheme types/scheme names that are recognized by the IUri interface as of IE 7. */
181 static const struct {
183 WCHAR scheme_name
[16];
184 } recognized_schemes
[] = {
185 {URL_SCHEME_FTP
, {'f','t','p',0}},
186 {URL_SCHEME_HTTP
, {'h','t','t','p',0}},
187 {URL_SCHEME_GOPHER
, {'g','o','p','h','e','r',0}},
188 {URL_SCHEME_MAILTO
, {'m','a','i','l','t','o',0}},
189 {URL_SCHEME_NEWS
, {'n','e','w','s',0}},
190 {URL_SCHEME_NNTP
, {'n','n','t','p',0}},
191 {URL_SCHEME_TELNET
, {'t','e','l','n','e','t',0}},
192 {URL_SCHEME_WAIS
, {'w','a','i','s',0}},
193 {URL_SCHEME_FILE
, {'f','i','l','e',0}},
194 {URL_SCHEME_MK
, {'m','k',0}},
195 {URL_SCHEME_HTTPS
, {'h','t','t','p','s',0}},
196 {URL_SCHEME_SHELL
, {'s','h','e','l','l',0}},
197 {URL_SCHEME_SNEWS
, {'s','n','e','w','s',0}},
198 {URL_SCHEME_LOCAL
, {'l','o','c','a','l',0}},
199 {URL_SCHEME_JAVASCRIPT
, {'j','a','v','a','s','c','r','i','p','t',0}},
200 {URL_SCHEME_VBSCRIPT
, {'v','b','s','c','r','i','p','t',0}},
201 {URL_SCHEME_ABOUT
, {'a','b','o','u','t',0}},
202 {URL_SCHEME_RES
, {'r','e','s',0}},
203 {URL_SCHEME_MSSHELLROOTED
, {'m','s','-','s','h','e','l','l','-','r','o','o','t','e','d',0}},
204 {URL_SCHEME_MSSHELLIDLIST
, {'m','s','-','s','h','e','l','l','-','i','d','l','i','s','t',0}},
205 {URL_SCHEME_MSHELP
, {'h','c','p',0}},
206 {URL_SCHEME_WILDCARD
, {'*',0}}
209 /* List of default ports Windows recognizes. */
210 static const struct {
213 } default_ports
[] = {
214 {URL_SCHEME_FTP
, 21},
215 {URL_SCHEME_HTTP
, 80},
216 {URL_SCHEME_GOPHER
, 70},
217 {URL_SCHEME_NNTP
, 119},
218 {URL_SCHEME_TELNET
, 23},
219 {URL_SCHEME_WAIS
, 210},
220 {URL_SCHEME_HTTPS
, 443},
223 /* List of 3 character top level domain names Windows seems to recognize.
224 * There might be more, but, these are the only ones I've found so far.
226 static const struct {
228 } recognized_tlds
[] = {
238 static Uri
*get_uri_obj(IUri
*uri
)
243 hres
= IUri_QueryInterface(uri
, &IID_IUriObj
, (void**)&ret
);
244 return SUCCEEDED(hres
) ? ret
: NULL
;
247 static inline BOOL
is_alpha(WCHAR val
) {
248 return ((val
>= 'a' && val
<= 'z') || (val
>= 'A' && val
<= 'Z'));
251 static inline BOOL
is_num(WCHAR val
) {
252 return (val
>= '0' && val
<= '9');
255 static inline BOOL
is_drive_path(const WCHAR
*str
) {
256 return (is_alpha(str
[0]) && (str
[1] == ':' || str
[1] == '|'));
259 static inline BOOL
is_unc_path(const WCHAR
*str
) {
260 return (str
[0] == '\\' && str
[0] == '\\');
263 static inline BOOL
is_forbidden_dos_path_char(WCHAR val
) {
264 return (val
== '>' || val
== '<' || val
== '\"');
267 /* A URI is implicitly a file path if it begins with
268 * a drive letter (eg X:) or starts with "\\" (UNC path).
270 static inline BOOL
is_implicit_file_path(const WCHAR
*str
) {
271 return (is_unc_path(str
) || (is_alpha(str
[0]) && str
[1] == ':'));
274 /* Checks if the URI is a hierarchical URI. A hierarchical
275 * URI is one that has "//" after the scheme.
277 static BOOL
check_hierarchical(const WCHAR
**ptr
) {
278 const WCHAR
*start
= *ptr
;
293 /* unreserved = ALPHA / DIGIT / "-" / "." / "_" / "~" */
294 static inline BOOL
is_unreserved(WCHAR val
) {
295 return (is_alpha(val
) || is_num(val
) || val
== '-' || val
== '.' ||
296 val
== '_' || val
== '~');
299 /* sub-delims = "!" / "$" / "&" / "'" / "(" / ")"
300 * / "*" / "+" / "," / ";" / "="
302 static inline BOOL
is_subdelim(WCHAR val
) {
303 return (val
== '!' || val
== '$' || val
== '&' ||
304 val
== '\'' || val
== '(' || val
== ')' ||
305 val
== '*' || val
== '+' || val
== ',' ||
306 val
== ';' || val
== '=');
309 /* gen-delims = ":" / "/" / "?" / "#" / "[" / "]" / "@" */
310 static inline BOOL
is_gendelim(WCHAR val
) {
311 return (val
== ':' || val
== '/' || val
== '?' ||
312 val
== '#' || val
== '[' || val
== ']' ||
316 /* Characters that delimit the end of the authority
317 * section of a URI. Sometimes a '\\' is considered
318 * an authority delimeter.
320 static inline BOOL
is_auth_delim(WCHAR val
, BOOL acceptSlash
) {
321 return (val
== '#' || val
== '/' || val
== '?' ||
322 val
== '\0' || (acceptSlash
&& val
== '\\'));
325 /* reserved = gen-delims / sub-delims */
326 static inline BOOL
is_reserved(WCHAR val
) {
327 return (is_subdelim(val
) || is_gendelim(val
));
330 static inline BOOL
is_hexdigit(WCHAR val
) {
331 return ((val
>= 'a' && val
<= 'f') ||
332 (val
>= 'A' && val
<= 'F') ||
333 (val
>= '0' && val
<= '9'));
336 static inline BOOL
is_path_delim(WCHAR val
) {
337 return (!val
|| val
== '#' || val
== '?');
340 /* List of schemes types Windows seems to expect to be hierarchical. */
341 static inline BOOL
is_hierarchical_scheme(URL_SCHEME type
) {
342 return(type
== URL_SCHEME_HTTP
|| type
== URL_SCHEME_FTP
||
343 type
== URL_SCHEME_GOPHER
|| type
== URL_SCHEME_NNTP
||
344 type
== URL_SCHEME_TELNET
|| type
== URL_SCHEME_WAIS
||
345 type
== URL_SCHEME_FILE
|| type
== URL_SCHEME_HTTPS
||
346 type
== URL_SCHEME_RES
);
349 /* Checks if 'flags' contains an invalid combination of Uri_CREATE flags. */
350 static inline BOOL
has_invalid_flag_combination(DWORD flags
) {
351 return((flags
& Uri_CREATE_DECODE_EXTRA_INFO
&& flags
& Uri_CREATE_NO_DECODE_EXTRA_INFO
) ||
352 (flags
& Uri_CREATE_CANONICALIZE
&& flags
& Uri_CREATE_NO_CANONICALIZE
) ||
353 (flags
& Uri_CREATE_CRACK_UNKNOWN_SCHEMES
&& flags
& Uri_CREATE_NO_CRACK_UNKNOWN_SCHEMES
) ||
354 (flags
& Uri_CREATE_PRE_PROCESS_HTML_URI
&& flags
& Uri_CREATE_NO_PRE_PROCESS_HTML_URI
) ||
355 (flags
& Uri_CREATE_IE_SETTINGS
&& flags
& Uri_CREATE_NO_IE_SETTINGS
));
358 /* Applies each default Uri_CREATE flags to 'flags' if it
359 * doesn't cause a flag conflict.
361 static void apply_default_flags(DWORD
*flags
) {
362 if(!(*flags
& Uri_CREATE_NO_CANONICALIZE
))
363 *flags
|= Uri_CREATE_CANONICALIZE
;
364 if(!(*flags
& Uri_CREATE_NO_DECODE_EXTRA_INFO
))
365 *flags
|= Uri_CREATE_DECODE_EXTRA_INFO
;
366 if(!(*flags
& Uri_CREATE_NO_CRACK_UNKNOWN_SCHEMES
))
367 *flags
|= Uri_CREATE_CRACK_UNKNOWN_SCHEMES
;
368 if(!(*flags
& Uri_CREATE_NO_PRE_PROCESS_HTML_URI
))
369 *flags
|= Uri_CREATE_PRE_PROCESS_HTML_URI
;
370 if(!(*flags
& Uri_CREATE_IE_SETTINGS
))
371 *flags
|= Uri_CREATE_NO_IE_SETTINGS
;
374 /* Determines if the URI is hierarchical using the information already parsed into
375 * data and using the current location of parsing in the URI string.
377 * Windows considers a URI hierarchical if on of the following is true:
378 * A.) It's a wildcard scheme.
379 * B.) It's an implicit file scheme.
380 * C.) It's a known hierarchical scheme and it has two '\\' after the scheme name.
381 * (the '\\' will be converted into "//" during canonicalization).
382 * D.) It's not a relative URI and "//" appears after the scheme name.
384 static inline BOOL
is_hierarchical_uri(const WCHAR
**ptr
, const parse_data
*data
) {
385 const WCHAR
*start
= *ptr
;
387 if(data
->scheme_type
== URL_SCHEME_WILDCARD
)
389 else if(data
->scheme_type
== URL_SCHEME_FILE
&& data
->has_implicit_scheme
)
391 else if(is_hierarchical_scheme(data
->scheme_type
) && (*ptr
)[0] == '\\' && (*ptr
)[1] == '\\') {
394 } else if(!data
->is_relative
&& check_hierarchical(ptr
))
401 /* Checks if the two Uri's are logically equivalent. It's a simple
402 * comparison, since they are both of type Uri, and it can access
403 * the properties of each Uri directly without the need to go
404 * through the "IUri_Get*" interface calls.
406 static BOOL
are_equal_simple(const Uri
*a
, const Uri
*b
) {
407 if(a
->scheme_type
== b
->scheme_type
) {
408 const BOOL known_scheme
= a
->scheme_type
!= URL_SCHEME_UNKNOWN
;
409 const BOOL are_hierarchical
=
410 (a
->authority_start
> -1 && b
->authority_start
> -1);
412 if(a
->scheme_type
== URL_SCHEME_FILE
) {
413 if(a
->canon_len
== b
->canon_len
)
414 return !StrCmpIW(a
->canon_uri
, b
->canon_uri
);
417 /* Only compare the scheme names (if any) if their unknown scheme types. */
419 if((a
->scheme_start
> -1 && b
->scheme_start
> -1) &&
420 (a
->scheme_len
== b
->scheme_len
)) {
421 /* Make sure the schemes are the same. */
422 if(StrCmpNW(a
->canon_uri
+a
->scheme_start
, b
->canon_uri
+b
->scheme_start
, a
->scheme_len
))
424 } else if(a
->scheme_len
!= b
->scheme_len
)
425 /* One of the Uri's has a scheme name, while the other doesn't. */
429 /* If they have a userinfo component, perform case sensitive compare. */
430 if((a
->userinfo_start
> -1 && b
->userinfo_start
> -1) &&
431 (a
->userinfo_len
== b
->userinfo_len
)) {
432 if(StrCmpNW(a
->canon_uri
+a
->userinfo_start
, b
->canon_uri
+b
->userinfo_start
, a
->userinfo_len
))
434 } else if(a
->userinfo_len
!= b
->userinfo_len
)
435 /* One of the Uri's had a userinfo, while the other one doesn't. */
438 /* Check if they have a host name. */
439 if((a
->host_start
> -1 && b
->host_start
> -1) &&
440 (a
->host_len
== b
->host_len
)) {
441 /* Perform a case insensitive compare if they are a known scheme type. */
443 if(StrCmpNIW(a
->canon_uri
+a
->host_start
, b
->canon_uri
+b
->host_start
, a
->host_len
))
445 } else if(StrCmpNW(a
->canon_uri
+a
->host_start
, b
->canon_uri
+b
->host_start
, a
->host_len
))
447 } else if(a
->host_len
!= b
->host_len
)
448 /* One of the Uri's had a host, while the other one didn't. */
451 if(a
->has_port
&& b
->has_port
) {
452 if(a
->port
!= b
->port
)
454 } else if(a
->has_port
|| b
->has_port
)
455 /* One had a port, while the other one didn't. */
458 /* Windows is weird with how it handles paths. For example
459 * One URI could be "http://google.com" (after canonicalization)
460 * and one could be "http://google.com/" and the IsEqual function
461 * would still evaluate to TRUE, but, only if they are both hierarchical
464 if((a
->path_start
> -1 && b
->path_start
> -1) &&
465 (a
->path_len
== b
->path_len
)) {
466 if(StrCmpNW(a
->canon_uri
+a
->path_start
, b
->canon_uri
+b
->path_start
, a
->path_len
))
468 } else if(are_hierarchical
&& a
->path_len
== -1 && b
->path_len
== 0) {
469 if(*(a
->canon_uri
+a
->path_start
) != '/')
471 } else if(are_hierarchical
&& b
->path_len
== 1 && a
->path_len
== 0) {
472 if(*(b
->canon_uri
+b
->path_start
) != '/')
474 } else if(a
->path_len
!= b
->path_len
)
477 /* Compare the query strings of the two URIs. */
478 if((a
->query_start
> -1 && b
->query_start
> -1) &&
479 (a
->query_len
== b
->query_len
)) {
480 if(StrCmpNW(a
->canon_uri
+a
->query_start
, b
->canon_uri
+b
->query_start
, a
->query_len
))
482 } else if(a
->query_len
!= b
->query_len
)
485 if((a
->fragment_start
> -1 && b
->fragment_start
> -1) &&
486 (a
->fragment_len
== b
->fragment_len
)) {
487 if(StrCmpNW(a
->canon_uri
+a
->fragment_start
, b
->canon_uri
+b
->fragment_start
, a
->fragment_len
))
489 } else if(a
->fragment_len
!= b
->fragment_len
)
492 /* If we get here, the two URIs are equivalent. */
499 /* Computes the size of the given IPv6 address.
500 * Each h16 component is 16bits, if there is an IPv4 address, it's
501 * 32bits. If there's an elision it can be 16bits to 128bits, depending
502 * on the number of other components.
504 * Modeled after google-url's CheckIPv6ComponentsSize function
506 static void compute_ipv6_comps_size(ipv6_address
*address
) {
507 address
->components_size
= address
->h16_count
* 2;
510 /* IPv4 address is 4 bytes. */
511 address
->components_size
+= 4;
513 if(address
->elision
) {
514 /* An elision can be anywhere from 2 bytes up to 16 bytes.
515 * It size depends on the size of the h16 and IPv4 components.
517 address
->elision_size
= 16 - address
->components_size
;
518 if(address
->elision_size
< 2)
519 address
->elision_size
= 2;
521 address
->elision_size
= 0;
524 /* Taken from dlls/jscript/lex.c */
525 static int hex_to_int(WCHAR val
) {
526 if(val
>= '0' && val
<= '9')
528 else if(val
>= 'a' && val
<= 'f')
529 return val
- 'a' + 10;
530 else if(val
>= 'A' && val
<= 'F')
531 return val
- 'A' + 10;
536 /* Helper function for converting a percent encoded string
537 * representation of a WCHAR value into its actual WCHAR value. If
538 * the two characters following the '%' aren't valid hex values then
539 * this function returns the NULL character.
542 * "%2E" will result in '.' being returned by this function.
544 static WCHAR
decode_pct_val(const WCHAR
*ptr
) {
547 if(*ptr
== '%' && is_hexdigit(*(ptr
+ 1)) && is_hexdigit(*(ptr
+ 2))) {
548 INT a
= hex_to_int(*(ptr
+ 1));
549 INT b
= hex_to_int(*(ptr
+ 2));
558 /* Helper function for percent encoding a given character
559 * and storing the encoded value into a given buffer (dest).
561 * It's up to the calling function to ensure that there is
562 * at least enough space in 'dest' for the percent encoded
563 * value to be stored (so dest + 3 spaces available).
565 static inline void pct_encode_val(WCHAR val
, WCHAR
*dest
) {
567 dest
[1] = hexDigits
[(val
>> 4) & 0xf];
568 dest
[2] = hexDigits
[val
& 0xf];
571 /* Scans the range of characters [str, end] and returns the last occurrence
572 * of 'ch' or returns NULL.
574 static const WCHAR
*str_last_of(const WCHAR
*str
, const WCHAR
*end
, WCHAR ch
) {
575 const WCHAR
*ptr
= end
;
586 /* Attempts to parse the domain name from the host.
588 * This function also includes the Top-level Domain (TLD) name
589 * of the host when it tries to find the domain name. If it finds
590 * a valid domain name it will assign 'domain_start' the offset
591 * into 'host' where the domain name starts.
593 * It's implied that if a domain name its range is implied to be
594 * [host+domain_start, host+host_len).
596 static void find_domain_name(const WCHAR
*host
, DWORD host_len
,
598 const WCHAR
*last_tld
, *sec_last_tld
, *end
;
600 end
= host
+host_len
-1;
604 /* There has to be at least enough room for a '.' followed by a
605 * 3 character TLD for a domain to even exist in the host name.
610 last_tld
= str_last_of(host
, end
, '.');
612 /* http://hostname -> has no domain name. */
615 sec_last_tld
= str_last_of(host
, last_tld
-1, '.');
617 /* If the '.' is at the beginning of the host there
618 * has to be at least 3 characters in the TLD for it
620 * Ex: .com -> .com as the domain name.
621 * .co -> has no domain name.
623 if(last_tld
-host
== 0) {
624 if(end
-(last_tld
-1) < 3)
626 } else if(last_tld
-host
== 3) {
629 /* If there's three characters in front of last_tld and
630 * they are on the list of recognized TLDs, then this
631 * host doesn't have a domain (since the host only contains
633 * Ex: edu.uk -> has no domain name.
634 * foo.uk -> foo.uk as the domain name.
636 for(i
= 0; i
< sizeof(recognized_tlds
)/sizeof(recognized_tlds
[0]); ++i
) {
637 if(!StrCmpNIW(host
, recognized_tlds
[i
].tld_name
, 3))
640 } else if(last_tld
-host
< 3)
641 /* Anything less than 3 characters is considered part
643 * Ex: ak.uk -> Has no domain name.
647 /* Otherwise the domain name is the whole host name. */
649 } else if(end
+1-last_tld
> 3) {
650 /* If the last_tld has more than 3 characters, then it's automatically
651 * considered the TLD of the domain name.
652 * Ex: www.winehq.org.uk.test -> uk.test as the domain name.
654 *domain_start
= (sec_last_tld
+1)-host
;
655 } else if(last_tld
- (sec_last_tld
+1) < 4) {
657 /* If the sec_last_tld is 3 characters long it HAS to be on the list of
658 * recognized to still be considered part of the TLD name, otherwise
659 * its considered the domain name.
660 * Ex: www.google.com.uk -> google.com.uk as the domain name.
661 * www.google.foo.uk -> foo.uk as the domain name.
663 if(last_tld
- (sec_last_tld
+1) == 3) {
664 for(i
= 0; i
< sizeof(recognized_tlds
)/sizeof(recognized_tlds
[0]); ++i
) {
665 if(!StrCmpNIW(sec_last_tld
+1, recognized_tlds
[i
].tld_name
, 3)) {
666 const WCHAR
*domain
= str_last_of(host
, sec_last_tld
-1, '.');
671 *domain_start
= (domain
+1) - host
;
672 TRACE("Found domain name %s\n", debugstr_wn(host
+*domain_start
,
673 (host
+host_len
)-(host
+*domain_start
)));
678 *domain_start
= (sec_last_tld
+1)-host
;
680 /* Since the sec_last_tld is less than 3 characters it's considered
682 * Ex: www.google.fo.uk -> google.fo.uk as the domain name.
684 const WCHAR
*domain
= str_last_of(host
, sec_last_tld
-1, '.');
689 *domain_start
= (domain
+1) - host
;
692 /* The second to last TLD has more than 3 characters making it
694 * Ex: www.google.test.us -> test.us as the domain name.
696 *domain_start
= (sec_last_tld
+1)-host
;
699 TRACE("Found domain name %s\n", debugstr_wn(host
+*domain_start
,
700 (host
+host_len
)-(host
+*domain_start
)));
703 /* Removes the dot segments from a hierarchical URIs path component. This
704 * function performs the removal in place.
706 * This is a modified version of Qt's QUrl function "removeDotsFromPath".
708 * This function returns the new length of the path string.
710 static DWORD
remove_dot_segments(WCHAR
*path
, DWORD path_len
) {
712 const WCHAR
*in
= out
;
713 const WCHAR
*end
= out
+ path_len
;
717 /* A. if the input buffer begins with a prefix of "/./" or "/.",
718 * where "." is a complete path segment, then replace that
719 * prefix with "/" in the input buffer; otherwise,
721 if(in
<= end
- 3 && in
[0] == '/' && in
[1] == '.' && in
[2] == '/') {
724 } else if(in
== end
- 2 && in
[0] == '/' && in
[1] == '.') {
730 /* B. if the input buffer begins with a prefix of "/../" or "/..",
731 * where ".." is a complete path segment, then replace that
732 * prefix with "/" in the input buffer and remove the last
733 * segment and its preceding "/" (if any) from the output
736 if(in
<= end
- 4 && in
[0] == '/' && in
[1] == '.' && in
[2] == '.' && in
[3] == '/') {
737 while(out
> path
&& *(--out
) != '/');
741 } else if(in
== end
- 3 && in
[0] == '/' && in
[1] == '.' && in
[2] == '.') {
742 while(out
> path
&& *(--out
) != '/');
751 /* C. move the first path segment in the input buffer to the end of
752 * the output buffer, including the initial "/" character (if
753 * any) and any subsequent characters up to, but not including,
754 * the next "/" character or the end of the input buffer.
757 while(in
< end
&& *in
!= '/')
762 TRACE("(%p %d): Path after dot segments removed %s len=%d\n", path
, path_len
,
763 debugstr_wn(path
, len
), len
);
767 /* Attempts to find the file extension in a given path. */
768 static INT
find_file_extension(const WCHAR
*path
, DWORD path_len
) {
771 for(end
= path
+path_len
-1; end
>= path
&& *end
!= '/' && *end
!= '\\'; --end
) {
779 /* Computes the location where the elision should occur in the IPv6
780 * address using the numerical values of each component stored in
781 * 'values'. If the address shouldn't contain an elision then 'index'
782 * is assigned -1 as it's value. Otherwise 'index' will contain the
783 * starting index (into values) where the elision should be, and 'count'
784 * will contain the number of cells the elision covers.
787 * Windows will expand an elision if the elision only represents 1 h16
788 * component of the URI.
790 * Ex: [1::2:3:4:5:6:7] -> [1:0:2:3:4:5:6:7]
792 * If the IPv6 address contains an IPv4 address, the IPv4 address is also
793 * considered for being included as part of an elision if all it's components
796 * Ex: [1:2:3:4:5:6:0.0.0.0] -> [1:2:3:4:5:6::]
798 static void compute_elision_location(const ipv6_address
*address
, const USHORT values
[8],
799 INT
*index
, DWORD
*count
) {
800 DWORD i
, max_len
, cur_len
;
801 INT max_index
, cur_index
;
803 max_len
= cur_len
= 0;
804 max_index
= cur_index
= -1;
805 for(i
= 0; i
< 8; ++i
) {
806 BOOL check_ipv4
= (address
->ipv4
&& i
== 6);
807 BOOL is_end
= (check_ipv4
|| i
== 7);
810 /* Check if the IPv4 address contains only zeros. */
811 if(values
[i
] == 0 && values
[i
+1] == 0) {
818 } else if(values
[i
] == 0) {
825 if(is_end
|| values
[i
] != 0) {
826 /* We only consider it for an elision if it's
827 * more than 1 component long.
829 if(cur_len
> 1 && cur_len
> max_len
) {
830 /* Found the new elision location. */
832 max_index
= cur_index
;
835 /* Reset the current range for the next range of zeros. */
845 /* Removes all the leading and trailing white spaces or
846 * control characters from the URI and removes all control
847 * characters inside of the URI string.
849 static BSTR
pre_process_uri(LPCWSTR uri
) {
852 const WCHAR
*start
, *end
;
858 /* Skip leading controls and whitespace. */
859 while(iscntrlW(*start
) || isspaceW(*start
)) ++start
;
863 /* URI consisted only of control/whitespace. */
864 ret
= SysAllocStringLen(NULL
, 0);
866 while(iscntrlW(*end
) || isspaceW(*end
)) --end
;
868 buf
= heap_alloc(((end
+1)-start
)*sizeof(WCHAR
));
872 for(ptr
= buf
; start
< end
+1; ++start
) {
873 if(!iscntrlW(*start
))
877 ret
= SysAllocStringLen(buf
, ptr
-buf
);
884 /* Converts the specified IPv4 address into an uint value.
886 * This function assumes that the IPv4 address has already been validated.
888 static UINT
ipv4toui(const WCHAR
*ip
, DWORD len
) {
890 DWORD comp_value
= 0;
893 for(ptr
= ip
; ptr
< ip
+len
; ++ptr
) {
899 comp_value
= comp_value
*10 + (*ptr
-'0');
908 /* Converts an IPv4 address in numerical form into it's fully qualified
909 * string form. This function returns the number of characters written
910 * to 'dest'. If 'dest' is NULL this function will return the number of
911 * characters that would have been written.
913 * It's up to the caller to ensure there's enough space in 'dest' for the
916 static DWORD
ui2ipv4(WCHAR
*dest
, UINT address
) {
917 static const WCHAR formatW
[] =
918 {'%','u','.','%','u','.','%','u','.','%','u',0};
922 digits
[0] = (address
>> 24) & 0xff;
923 digits
[1] = (address
>> 16) & 0xff;
924 digits
[2] = (address
>> 8) & 0xff;
925 digits
[3] = address
& 0xff;
929 ret
= sprintfW(tmp
, formatW
, digits
[0], digits
[1], digits
[2], digits
[3]);
931 ret
= sprintfW(dest
, formatW
, digits
[0], digits
[1], digits
[2], digits
[3]);
936 /* Converts an h16 component (from an IPv6 address) into it's
939 * This function assumes that the h16 component has already been validated.
941 static USHORT
h16tous(h16 component
) {
945 for(i
= 0; i
< component
.len
; ++i
) {
947 ret
+= hex_to_int(component
.str
[i
]);
953 /* Converts an IPv6 address into it's 128 bits (16 bytes) numerical value.
955 * This function assumes that the ipv6_address has already been validated.
957 static BOOL
ipv6_to_number(const ipv6_address
*address
, USHORT number
[8]) {
958 DWORD i
, cur_component
= 0;
959 BOOL already_passed_elision
= FALSE
;
961 for(i
= 0; i
< address
->h16_count
; ++i
) {
962 if(address
->elision
) {
963 if(address
->components
[i
].str
> address
->elision
&& !already_passed_elision
) {
964 /* Means we just passed the elision and need to add it's values to
965 * 'number' before we do anything else.
968 for(j
= 0; j
< address
->elision_size
; j
+=2)
969 number
[cur_component
++] = 0;
971 already_passed_elision
= TRUE
;
975 number
[cur_component
++] = h16tous(address
->components
[i
]);
978 /* Case when the elision appears after the h16 components. */
979 if(!already_passed_elision
&& address
->elision
) {
980 for(i
= 0; i
< address
->elision_size
; i
+=2)
981 number
[cur_component
++] = 0;
982 already_passed_elision
= TRUE
;
986 UINT value
= ipv4toui(address
->ipv4
, address
->ipv4_len
);
988 if(cur_component
!= 6) {
989 ERR("(%p %p): Failed sanity check with %d\n", address
, number
, cur_component
);
993 number
[cur_component
++] = (value
>> 16) & 0xffff;
994 number
[cur_component
] = value
& 0xffff;
1000 /* Checks if the characters pointed to by 'ptr' are
1001 * a percent encoded data octet.
1003 * pct-encoded = "%" HEXDIG HEXDIG
1005 static BOOL
check_pct_encoded(const WCHAR
**ptr
) {
1006 const WCHAR
*start
= *ptr
;
1012 if(!is_hexdigit(**ptr
)) {
1018 if(!is_hexdigit(**ptr
)) {
1027 /* dec-octet = DIGIT ; 0-9
1028 * / %x31-39 DIGIT ; 10-99
1029 * / "1" 2DIGIT ; 100-199
1030 * / "2" %x30-34 DIGIT ; 200-249
1031 * / "25" %x30-35 ; 250-255
1033 static BOOL
check_dec_octet(const WCHAR
**ptr
) {
1034 const WCHAR
*c1
, *c2
, *c3
;
1037 /* A dec-octet must be at least 1 digit long. */
1038 if(*c1
< '0' || *c1
> '9')
1044 /* Since the 1 digit requirment was meet, it doesn't
1045 * matter if this is a DIGIT value, it's considered a
1048 if(*c2
< '0' || *c2
> '9')
1054 /* Same explanation as above. */
1055 if(*c3
< '0' || *c3
> '9')
1058 /* Anything > 255 isn't a valid IP dec-octet. */
1059 if(*c1
>= '2' && *c2
>= '5' && *c3
>= '5') {
1068 /* Checks if there is an implicit IPv4 address in the host component of the URI.
1069 * The max value of an implicit IPv4 address is UINT_MAX.
1072 * "234567" would be considered an implicit IPv4 address.
1074 static BOOL
check_implicit_ipv4(const WCHAR
**ptr
, UINT
*val
) {
1075 const WCHAR
*start
= *ptr
;
1079 while(is_num(**ptr
)) {
1080 ret
= ret
*10 + (**ptr
- '0');
1082 if(ret
> UINT_MAX
) {
1096 /* Checks if the string contains an IPv4 address.
1098 * This function has a strict mode or a non-strict mode of operation
1099 * When 'strict' is set to FALSE this function will return TRUE if
1100 * the string contains at least 'dec-octet "." dec-octet' since partial
1101 * IPv4 addresses will be normalized out into full IPv4 addresses. When
1102 * 'strict' is set this function expects there to be a full IPv4 address.
1104 * IPv4address = dec-octet "." dec-octet "." dec-octet "." dec-octet
1106 static BOOL
check_ipv4address(const WCHAR
**ptr
, BOOL strict
) {
1107 const WCHAR
*start
= *ptr
;
1109 if(!check_dec_octet(ptr
)) {
1120 if(!check_dec_octet(ptr
)) {
1134 if(!check_dec_octet(ptr
)) {
1148 if(!check_dec_octet(ptr
)) {
1153 /* Found a four digit ip address. */
1156 /* Tries to parse the scheme name of the URI.
1158 * scheme = ALPHA *(ALPHA | NUM | '+' | '-' | '.') as defined by RFC 3896.
1159 * NOTE: Windows accepts a number as the first character of a scheme.
1161 static BOOL
parse_scheme_name(const WCHAR
**ptr
, parse_data
*data
, DWORD extras
) {
1162 const WCHAR
*start
= *ptr
;
1164 data
->scheme
= NULL
;
1165 data
->scheme_len
= 0;
1168 if(**ptr
== '*' && *ptr
== start
) {
1169 /* Might have found a wildcard scheme. If it is the next
1170 * char has to be a ':' for it to be a valid URI
1174 } else if(!is_num(**ptr
) && !is_alpha(**ptr
) && **ptr
!= '+' &&
1175 **ptr
!= '-' && **ptr
!= '.')
1184 /* Schemes must end with a ':' */
1185 if(**ptr
!= ':' && !((extras
& ALLOW_NULL_TERM_SCHEME
) && !**ptr
)) {
1190 data
->scheme
= start
;
1191 data
->scheme_len
= *ptr
- start
;
1197 /* Tries to deduce the corresponding URL_SCHEME for the given URI. Stores
1198 * the deduced URL_SCHEME in data->scheme_type.
1200 static BOOL
parse_scheme_type(parse_data
*data
) {
1201 /* If there's scheme data then see if it's a recognized scheme. */
1202 if(data
->scheme
&& data
->scheme_len
) {
1205 for(i
= 0; i
< sizeof(recognized_schemes
)/sizeof(recognized_schemes
[0]); ++i
) {
1206 if(lstrlenW(recognized_schemes
[i
].scheme_name
) == data
->scheme_len
) {
1207 /* Has to be a case insensitive compare. */
1208 if(!StrCmpNIW(recognized_schemes
[i
].scheme_name
, data
->scheme
, data
->scheme_len
)) {
1209 data
->scheme_type
= recognized_schemes
[i
].scheme
;
1215 /* If we get here it means it's not a recognized scheme. */
1216 data
->scheme_type
= URL_SCHEME_UNKNOWN
;
1218 } else if(data
->is_relative
) {
1219 /* Relative URI's have no scheme. */
1220 data
->scheme_type
= URL_SCHEME_UNKNOWN
;
1223 /* Should never reach here! what happened... */
1224 FIXME("(%p): Unable to determine scheme type for URI %s\n", data
, debugstr_w(data
->uri
));
1229 /* Tries to parse (or deduce) the scheme_name of a URI. If it can't
1230 * parse a scheme from the URI it will try to deduce the scheme_name and scheme_type
1231 * using the flags specified in 'flags' (if any). Flags that affect how this function
1232 * operates are the Uri_CREATE_ALLOW_* flags.
1234 * All parsed/deduced information will be stored in 'data' when the function returns.
1236 * Returns TRUE if it was able to successfully parse the information.
1238 static BOOL
parse_scheme(const WCHAR
**ptr
, parse_data
*data
, DWORD flags
, DWORD extras
) {
1239 static const WCHAR fileW
[] = {'f','i','l','e',0};
1240 static const WCHAR wildcardW
[] = {'*',0};
1242 /* First check to see if the uri could implicitly be a file path. */
1243 if(is_implicit_file_path(*ptr
)) {
1244 if(flags
& Uri_CREATE_ALLOW_IMPLICIT_FILE_SCHEME
) {
1245 data
->scheme
= fileW
;
1246 data
->scheme_len
= lstrlenW(fileW
);
1247 data
->has_implicit_scheme
= TRUE
;
1249 TRACE("(%p %p %x): URI is an implicit file path.\n", ptr
, data
, flags
);
1251 /* Window's does not consider anything that can implicitly be a file
1252 * path to be a valid URI if the ALLOW_IMPLICIT_FILE_SCHEME flag is not set...
1254 TRACE("(%p %p %x): URI is implicitly a file path, but, the ALLOW_IMPLICIT_FILE_SCHEME flag wasn't set.\n",
1258 } else if(!parse_scheme_name(ptr
, data
, extras
)) {
1259 /* No Scheme was found, this means it could be:
1260 * a) an implicit Wildcard scheme
1264 if(flags
& Uri_CREATE_ALLOW_IMPLICIT_WILDCARD_SCHEME
) {
1265 data
->scheme
= wildcardW
;
1266 data
->scheme_len
= lstrlenW(wildcardW
);
1267 data
->has_implicit_scheme
= TRUE
;
1269 TRACE("(%p %p %x): URI is an implicit wildcard scheme.\n", ptr
, data
, flags
);
1270 } else if (flags
& Uri_CREATE_ALLOW_RELATIVE
) {
1271 data
->is_relative
= TRUE
;
1272 TRACE("(%p %p %x): URI is relative.\n", ptr
, data
, flags
);
1274 TRACE("(%p %p %x): Malformed URI found. Unable to deduce scheme name.\n", ptr
, data
, flags
);
1279 if(!data
->is_relative
)
1280 TRACE("(%p %p %x): Found scheme=%s scheme_len=%d\n", ptr
, data
, flags
,
1281 debugstr_wn(data
->scheme
, data
->scheme_len
), data
->scheme_len
);
1283 if(!parse_scheme_type(data
))
1286 TRACE("(%p %p %x): Assigned %d as the URL_SCHEME.\n", ptr
, data
, flags
, data
->scheme_type
);
1290 static BOOL
parse_username(const WCHAR
**ptr
, parse_data
*data
, DWORD flags
, DWORD extras
) {
1291 data
->username
= *ptr
;
1293 while(**ptr
!= ':' && **ptr
!= '@') {
1295 if(!check_pct_encoded(ptr
)) {
1296 if(data
->scheme_type
!= URL_SCHEME_UNKNOWN
) {
1297 *ptr
= data
->username
;
1298 data
->username
= NULL
;
1303 } else if(extras
& ALLOW_NULL_TERM_USER_NAME
&& !**ptr
)
1305 else if(is_auth_delim(**ptr
, data
->scheme_type
!= URL_SCHEME_UNKNOWN
)) {
1306 *ptr
= data
->username
;
1307 data
->username
= NULL
;
1314 data
->username_len
= *ptr
- data
->username
;
1318 static BOOL
parse_password(const WCHAR
**ptr
, parse_data
*data
, DWORD flags
, DWORD extras
) {
1319 const WCHAR
*start
= *ptr
;
1325 data
->password
= *ptr
;
1327 while(**ptr
!= '@') {
1329 if(!check_pct_encoded(ptr
)) {
1330 if(data
->scheme_type
!= URL_SCHEME_UNKNOWN
) {
1332 data
->password
= NULL
;
1337 } else if(extras
& ALLOW_NULL_TERM_PASSWORD
&& !**ptr
)
1339 else if(is_auth_delim(**ptr
, data
->scheme_type
!= URL_SCHEME_UNKNOWN
)) {
1341 data
->password
= NULL
;
1348 data
->password_len
= *ptr
- data
->password
;
1352 /* Parses the userinfo part of the URI (if it exists). The userinfo field of
1353 * a URI can consist of "username:password@", or just "username@".
1356 * userinfo = *( unreserved / pct-encoded / sub-delims / ":" )
1359 * 1) If there is more than one ':' in the userinfo part of the URI Windows
1360 * uses the first occurrence of ':' to delimit the username and password
1364 * ftp://user:pass:word@winehq.org
1366 * Would yield, "user" as the username and "pass:word" as the password.
1368 * 2) Windows allows any character to appear in the "userinfo" part of
1369 * a URI, as long as it's not an authority delimeter character set.
1371 static void parse_userinfo(const WCHAR
**ptr
, parse_data
*data
, DWORD flags
) {
1372 const WCHAR
*start
= *ptr
;
1374 if(!parse_username(ptr
, data
, flags
, 0)) {
1375 TRACE("(%p %p %x): URI contained no userinfo.\n", ptr
, data
, flags
);
1379 if(!parse_password(ptr
, data
, flags
, 0)) {
1381 data
->username
= NULL
;
1382 data
->username_len
= 0;
1383 TRACE("(%p %p %x): URI contained no userinfo.\n", ptr
, data
, flags
);
1389 data
->username
= NULL
;
1390 data
->username_len
= 0;
1391 data
->password
= NULL
;
1392 data
->password_len
= 0;
1394 TRACE("(%p %p %x): URI contained no userinfo.\n", ptr
, data
, flags
);
1399 TRACE("(%p %p %x): Found username %s len=%d.\n", ptr
, data
, flags
,
1400 debugstr_wn(data
->username
, data
->username_len
), data
->username_len
);
1403 TRACE("(%p %p %x): Found password %s len=%d.\n", ptr
, data
, flags
,
1404 debugstr_wn(data
->password
, data
->password_len
), data
->password_len
);
1409 /* Attempts to parse a port from the URI.
1412 * Windows seems to have a cap on what the maximum value
1413 * for a port can be. The max value is USHORT_MAX.
1417 static BOOL
parse_port(const WCHAR
**ptr
, parse_data
*data
, DWORD flags
) {
1421 while(!is_auth_delim(**ptr
, data
->scheme_type
!= URL_SCHEME_UNKNOWN
)) {
1422 if(!is_num(**ptr
)) {
1428 port
= port
*10 + (**ptr
-'0');
1430 if(port
> USHORT_MAX
) {
1439 data
->has_port
= TRUE
;
1440 data
->port_value
= port
;
1441 data
->port_len
= *ptr
- data
->port
;
1443 TRACE("(%p %p %x): Found port %s len=%d value=%u\n", ptr
, data
, flags
,
1444 debugstr_wn(data
->port
, data
->port_len
), data
->port_len
, data
->port_value
);
1448 /* Attempts to parse a IPv4 address from the URI.
1451 * Window's normalizes IPv4 addresses, This means there's three
1452 * possibilities for the URI to contain an IPv4 address.
1453 * 1) A well formed address (ex. 192.2.2.2).
1454 * 2) A partially formed address. For example "192.0" would
1455 * normalize to "192.0.0.0" during canonicalization.
1456 * 3) An implicit IPv4 address. For example "256" would
1457 * normalize to "0.0.1.0" during canonicalization. Also
1458 * note that the maximum value for an implicit IP address
1459 * is UINT_MAX, if the value in the URI exceeds this then
1460 * it is not considered an IPv4 address.
1462 static BOOL
parse_ipv4address(const WCHAR
**ptr
, parse_data
*data
, DWORD flags
) {
1463 const BOOL is_unknown
= data
->scheme_type
== URL_SCHEME_UNKNOWN
;
1466 if(!check_ipv4address(ptr
, FALSE
)) {
1467 if(!check_implicit_ipv4(ptr
, &data
->implicit_ipv4
)) {
1468 TRACE("(%p %p %x): URI didn't contain anything looking like an IPv4 address.\n",
1474 data
->has_implicit_ip
= TRUE
;
1477 /* Check if what we found is the only part of the host name (if it isn't
1478 * we don't have an IPv4 address).
1482 if(!parse_port(ptr
, data
, flags
)) {
1487 } else if(!is_auth_delim(**ptr
, !is_unknown
)) {
1488 /* Found more data which belongs the host, so this isn't an IPv4. */
1491 data
->has_implicit_ip
= FALSE
;
1495 data
->host_len
= *ptr
- data
->host
;
1496 data
->host_type
= Uri_HOST_IPV4
;
1498 TRACE("(%p %p %x): IPv4 address found. host=%s host_len=%d host_type=%d\n",
1499 ptr
, data
, flags
, debugstr_wn(data
->host
, data
->host_len
),
1500 data
->host_len
, data
->host_type
);
1504 /* Attempts to parse the reg-name from the URI.
1506 * Because of the way Windows handles ':' this function also
1507 * handles parsing the port.
1509 * reg-name = *( unreserved / pct-encoded / sub-delims )
1512 * Windows allows everything, but, the characters in "auth_delims" and ':'
1513 * to appear in a reg-name, unless it's an unknown scheme type then ':' is
1514 * allowed to appear (even if a valid port isn't after it).
1516 * Windows doesn't like host names which start with '[' and end with ']'
1517 * and don't contain a valid IP literal address in between them.
1519 * On Windows if an '[' is encountered in the host name the ':' no longer
1520 * counts as a delimiter until you reach the next ']' or an "authority delimeter".
1522 * A reg-name CAN be empty.
1524 static BOOL
parse_reg_name(const WCHAR
**ptr
, parse_data
*data
, DWORD flags
, DWORD extras
) {
1525 const BOOL has_start_bracket
= **ptr
== '[';
1526 const BOOL known_scheme
= data
->scheme_type
!= URL_SCHEME_UNKNOWN
;
1527 BOOL inside_brackets
= has_start_bracket
;
1528 BOOL ignore_col
= extras
& IGNORE_PORT_DELIMITER
;
1530 /* We have to be careful with file schemes. */
1531 if(data
->scheme_type
== URL_SCHEME_FILE
) {
1532 /* This is because an implicit file scheme could be "C:\\test" and it
1533 * would trick this function into thinking the host is "C", when after
1534 * canonicalization the host would end up being an empty string. A drive
1535 * path can also have a '|' instead of a ':' after the drive letter.
1537 if(is_drive_path(*ptr
)) {
1538 /* Regular old drive paths don't have a host type (or host name). */
1539 data
->host_type
= Uri_HOST_UNKNOWN
;
1543 } else if(is_unc_path(*ptr
))
1544 /* Skip past the "\\" of a UNC path. */
1550 while(!is_auth_delim(**ptr
, known_scheme
)) {
1551 if(**ptr
== ':' && !ignore_col
) {
1552 /* We can ignore ':' if were inside brackets.*/
1553 if(!inside_brackets
) {
1554 const WCHAR
*tmp
= (*ptr
)++;
1556 /* Attempt to parse the port. */
1557 if(!parse_port(ptr
, data
, flags
)) {
1558 /* Windows expects there to be a valid port for known scheme types. */
1559 if(data
->scheme_type
!= URL_SCHEME_UNKNOWN
) {
1562 TRACE("(%p %p %x %x): Expected valid port\n", ptr
, data
, flags
, extras
);
1565 /* Windows gives up on trying to parse a port when it
1566 * encounters 1 invalid port.
1570 data
->host_len
= tmp
- data
->host
;
1574 } else if(**ptr
== '%' && known_scheme
) {
1575 /* Has to be a legit % encoded value. */
1576 if(!check_pct_encoded(ptr
)) {
1582 } else if(**ptr
== ']')
1583 inside_brackets
= FALSE
;
1584 else if(**ptr
== '[')
1585 inside_brackets
= TRUE
;
1590 if(has_start_bracket
) {
1591 /* Make sure the last character of the host wasn't a ']'. */
1592 if(*(*ptr
-1) == ']') {
1593 TRACE("(%p %p %x %x): Expected an IP literal inside of the host\n",
1594 ptr
, data
, flags
, extras
);
1601 /* Don't overwrite our length if we found a port earlier. */
1603 data
->host_len
= *ptr
- data
->host
;
1605 /* If the host is empty, then it's an unknown host type. */
1606 if(data
->host_len
== 0)
1607 data
->host_type
= Uri_HOST_UNKNOWN
;
1609 data
->host_type
= Uri_HOST_DNS
;
1611 TRACE("(%p %p %x %x): Parsed reg-name. host=%s len=%d\n", ptr
, data
, flags
, extras
,
1612 debugstr_wn(data
->host
, data
->host_len
), data
->host_len
);
1616 /* Attempts to parse an IPv6 address out of the URI.
1618 * IPv6address = 6( h16 ":" ) ls32
1619 * / "::" 5( h16 ":" ) ls32
1620 * / [ h16 ] "::" 4( h16 ":" ) ls32
1621 * / [ *1( h16 ":" ) h16 ] "::" 3( h16 ":" ) ls32
1622 * / [ *2( h16 ":" ) h16 ] "::" 2( h16 ":" ) ls32
1623 * / [ *3( h16 ":" ) h16 ] "::" h16 ":" ls32
1624 * / [ *4( h16 ":" ) h16 ] "::" ls32
1625 * / [ *5( h16 ":" ) h16 ] "::" h16
1626 * / [ *6( h16 ":" ) h16 ] "::"
1628 * ls32 = ( h16 ":" h16 ) / IPv4address
1629 * ; least-significant 32 bits of address.
1632 * ; 16 bits of address represented in hexadecimal.
1634 * Modeled after google-url's 'DoParseIPv6' function.
1636 static BOOL
parse_ipv6address(const WCHAR
**ptr
, parse_data
*data
, DWORD flags
) {
1637 const WCHAR
*start
, *cur_start
;
1640 start
= cur_start
= *ptr
;
1641 memset(&ip
, 0, sizeof(ipv6_address
));
1644 /* Check if we're on the last character of the host. */
1645 BOOL is_end
= (is_auth_delim(**ptr
, data
->scheme_type
!= URL_SCHEME_UNKNOWN
)
1648 BOOL is_split
= (**ptr
== ':');
1649 BOOL is_elision
= (is_split
&& !is_end
&& *(*ptr
+1) == ':');
1651 /* Check if we're at the end of a component, or
1652 * if we're at the end of the IPv6 address.
1654 if(is_split
|| is_end
) {
1657 cur_len
= *ptr
- cur_start
;
1659 /* h16 can't have a length > 4. */
1663 TRACE("(%p %p %x): h16 component to long.\n",
1669 /* An h16 component can't have the length of 0 unless
1670 * the elision is at the beginning of the address, or
1671 * at the end of the address.
1673 if(!((*ptr
== start
&& is_elision
) ||
1674 (is_end
&& (*ptr
-2) == ip
.elision
))) {
1676 TRACE("(%p %p %x): IPv6 component cannot have a length of 0.\n",
1683 /* An IPv6 address can have no more than 8 h16 components. */
1684 if(ip
.h16_count
>= 8) {
1686 TRACE("(%p %p %x): Not a IPv6 address, to many h16 components.\n",
1691 ip
.components
[ip
.h16_count
].str
= cur_start
;
1692 ip
.components
[ip
.h16_count
].len
= cur_len
;
1694 TRACE("(%p %p %x): Found h16 component %s, len=%d, h16_count=%d\n",
1695 ptr
, data
, flags
, debugstr_wn(cur_start
, cur_len
), cur_len
,
1705 /* A IPv6 address can only have 1 elision ('::'). */
1709 TRACE("(%p %p %x): IPv6 address cannot have 2 elisions.\n",
1721 if(!check_ipv4address(ptr
, TRUE
)) {
1722 if(!is_hexdigit(**ptr
)) {
1723 /* Not a valid character for an IPv6 address. */
1728 /* Found an IPv4 address. */
1729 ip
.ipv4
= cur_start
;
1730 ip
.ipv4_len
= *ptr
- cur_start
;
1732 TRACE("(%p %p %x): Found an attached IPv4 address %s len=%d.\n",
1733 ptr
, data
, flags
, debugstr_wn(ip
.ipv4
, ip
.ipv4_len
),
1736 /* IPv4 addresses can only appear at the end of a IPv6. */
1742 compute_ipv6_comps_size(&ip
);
1744 /* Make sure the IPv6 address adds up to 16 bytes. */
1745 if(ip
.components_size
+ ip
.elision_size
!= 16) {
1747 TRACE("(%p %p %x): Invalid IPv6 address, did not add up to 16 bytes.\n",
1752 if(ip
.elision_size
== 2) {
1753 /* For some reason on Windows if an elision that represents
1754 * only 1 h16 component is encountered at the very begin or
1755 * end of an IPv6 address, Windows does not consider it a
1756 * valid IPv6 address.
1758 * Ex: [::2:3:4:5:6:7] is not valid, even though the sum
1759 * of all the components == 128bits.
1761 if(ip
.elision
< ip
.components
[0].str
||
1762 ip
.elision
> ip
.components
[ip
.h16_count
-1].str
) {
1764 TRACE("(%p %p %x): Invalid IPv6 address. Detected elision of 2 bytes at the beginning or end of the address.\n",
1770 data
->host_type
= Uri_HOST_IPV6
;
1771 data
->has_ipv6
= TRUE
;
1772 data
->ipv6_address
= ip
;
1774 TRACE("(%p %p %x): Found valid IPv6 literal %s len=%d\n",
1775 ptr
, data
, flags
, debugstr_wn(start
, *ptr
-start
),
1780 /* IPvFuture = "v" 1*HEXDIG "." 1*( unreserved / sub-delims / ":" ) */
1781 static BOOL
parse_ipvfuture(const WCHAR
**ptr
, parse_data
*data
, DWORD flags
) {
1782 const WCHAR
*start
= *ptr
;
1784 /* IPvFuture has to start with a 'v' or 'V'. */
1785 if(**ptr
!= 'v' && **ptr
!= 'V')
1788 /* Following the v there must be at least 1 hex digit. */
1790 if(!is_hexdigit(**ptr
)) {
1796 while(is_hexdigit(**ptr
))
1799 /* End of the hexdigit sequence must be a '.' */
1806 if(!is_unreserved(**ptr
) && !is_subdelim(**ptr
) && **ptr
!= ':') {
1812 while(is_unreserved(**ptr
) || is_subdelim(**ptr
) || **ptr
== ':')
1815 data
->host_type
= Uri_HOST_UNKNOWN
;
1817 TRACE("(%p %p %x): Parsed IPvFuture address %s len=%d\n", ptr
, data
, flags
,
1818 debugstr_wn(start
, *ptr
-start
), *ptr
-start
);
1823 /* IP-literal = "[" ( IPv6address / IPvFuture ) "]" */
1824 static BOOL
parse_ip_literal(const WCHAR
**ptr
, parse_data
*data
, DWORD flags
, DWORD extras
) {
1827 if(**ptr
!= '[' && !(extras
& ALLOW_BRACKETLESS_IP_LITERAL
)) {
1830 } else if(**ptr
== '[')
1833 if(!parse_ipv6address(ptr
, data
, flags
)) {
1834 if(extras
& SKIP_IP_FUTURE_CHECK
|| !parse_ipvfuture(ptr
, data
, flags
)) {
1841 if(**ptr
!= ']' && !(extras
& ALLOW_BRACKETLESS_IP_LITERAL
)) {
1845 } else if(!**ptr
&& extras
& ALLOW_BRACKETLESS_IP_LITERAL
) {
1846 /* The IP literal didn't contain brackets and was followed by
1847 * a NULL terminator, so no reason to even check the port.
1849 data
->host_len
= *ptr
- data
->host
;
1856 /* If a valid port is not found, then let it trickle down to
1859 if(!parse_port(ptr
, data
, flags
)) {
1865 data
->host_len
= *ptr
- data
->host
;
1870 /* Parses the host information from the URI.
1872 * host = IP-literal / IPv4address / reg-name
1874 static BOOL
parse_host(const WCHAR
**ptr
, parse_data
*data
, DWORD flags
, DWORD extras
) {
1875 if(!parse_ip_literal(ptr
, data
, flags
, extras
)) {
1876 if(!parse_ipv4address(ptr
, data
, flags
)) {
1877 if(!parse_reg_name(ptr
, data
, flags
, extras
)) {
1878 TRACE("(%p %p %x %x): Malformed URI, Unknown host type.\n",
1879 ptr
, data
, flags
, extras
);
1888 /* Parses the authority information from the URI.
1890 * authority = [ userinfo "@" ] host [ ":" port ]
1892 static BOOL
parse_authority(const WCHAR
**ptr
, parse_data
*data
, DWORD flags
) {
1893 parse_userinfo(ptr
, data
, flags
);
1895 /* Parsing the port will happen during one of the host parsing
1896 * routines (if the URI has a port).
1898 if(!parse_host(ptr
, data
, flags
, 0))
1904 /* Attempts to parse the path information of a hierarchical URI. */
1905 static BOOL
parse_path_hierarchical(const WCHAR
**ptr
, parse_data
*data
, DWORD flags
) {
1906 const WCHAR
*start
= *ptr
;
1907 static const WCHAR slash
[] = {'/',0};
1908 const BOOL is_file
= data
->scheme_type
== URL_SCHEME_FILE
;
1910 if(is_path_delim(**ptr
)) {
1911 if(data
->scheme_type
== URL_SCHEME_WILDCARD
) {
1912 /* Wildcard schemes don't get a '/' attached if their path is
1917 } else if(!(flags
& Uri_CREATE_NO_CANONICALIZE
)) {
1918 /* If the path component is empty, then a '/' is added. */
1923 while(!is_path_delim(**ptr
)) {
1924 if(**ptr
== '%' && data
->scheme_type
!= URL_SCHEME_UNKNOWN
&& !is_file
) {
1925 if(!check_pct_encoded(ptr
)) {
1930 } else if(is_forbidden_dos_path_char(**ptr
) && is_file
&&
1931 (flags
& Uri_CREATE_FILE_USE_DOS_PATH
)) {
1932 /* File schemes with USE_DOS_PATH set aren't allowed to have
1933 * a '<' or '>' or '\"' appear in them.
1937 } else if(**ptr
== '\\') {
1938 /* Not allowed to have a backslash if NO_CANONICALIZE is set
1939 * and the scheme is known type (but not a file scheme).
1941 if(flags
& Uri_CREATE_NO_CANONICALIZE
) {
1942 if(data
->scheme_type
!= URL_SCHEME_FILE
&&
1943 data
->scheme_type
!= URL_SCHEME_UNKNOWN
) {
1953 /* The only time a URI doesn't have a path is when
1954 * the NO_CANONICALIZE flag is set and the raw URI
1955 * didn't contain one.
1962 data
->path_len
= *ptr
- start
;
1967 TRACE("(%p %p %x): Parsed path %s len=%d\n", ptr
, data
, flags
,
1968 debugstr_wn(data
->path
, data
->path_len
), data
->path_len
);
1970 TRACE("(%p %p %x): The URI contained no path\n", ptr
, data
, flags
);
1975 /* Parses the path of a opaque URI (much less strict then the parser
1976 * for a hierarchical URI).
1979 * Windows allows invalid % encoded data to appear in opaque URI paths
1980 * for unknown scheme types.
1982 * File schemes with USE_DOS_PATH set aren't allowed to have '<', '>', or '\"'
1985 static BOOL
parse_path_opaque(const WCHAR
**ptr
, parse_data
*data
, DWORD flags
) {
1986 const BOOL known_scheme
= data
->scheme_type
!= URL_SCHEME_UNKNOWN
;
1987 const BOOL is_file
= data
->scheme_type
== URL_SCHEME_FILE
;
1991 while(!is_path_delim(**ptr
)) {
1992 if(**ptr
== '%' && known_scheme
) {
1993 if(!check_pct_encoded(ptr
)) {
1999 } else if(is_forbidden_dos_path_char(**ptr
) && is_file
&&
2000 (flags
& Uri_CREATE_FILE_USE_DOS_PATH
)) {
2009 data
->path_len
= *ptr
- data
->path
;
2010 TRACE("(%p %p %x): Parsed opaque URI path %s len=%d\n", ptr
, data
, flags
,
2011 debugstr_wn(data
->path
, data
->path_len
), data
->path_len
);
2015 /* Determines how the URI should be parsed after the scheme information.
2017 * If the scheme is followed, by "//" then, it is treated as an hierarchical URI
2018 * which then the authority and path information will be parsed out. Otherwise, the
2019 * URI will be treated as an opaque URI which the authority information is not parsed
2022 * RFC 3896 definition of hier-part:
2024 * hier-part = "//" authority path-abempty
2029 * MSDN opaque URI definition:
2030 * scheme ":" path [ "#" fragment ]
2033 * If the URI is of an unknown scheme type and has a "//" following the scheme then it
2034 * is treated as a hierarchical URI, but, if the CREATE_NO_CRACK_UNKNOWN_SCHEMES flag is
2035 * set then it is considered an opaque URI reguardless of what follows the scheme information
2036 * (per MSDN documentation).
2038 static BOOL
parse_hierpart(const WCHAR
**ptr
, parse_data
*data
, DWORD flags
) {
2039 const WCHAR
*start
= *ptr
;
2041 /* Checks if the authority information needs to be parsed. */
2042 if(is_hierarchical_uri(ptr
, data
)) {
2043 /* Only treat it as a hierarchical URI if the scheme_type is known or
2044 * the Uri_CREATE_NO_CRACK_UNKNOWN_SCHEMES flag is not set.
2046 if(data
->scheme_type
!= URL_SCHEME_UNKNOWN
||
2047 !(flags
& Uri_CREATE_NO_CRACK_UNKNOWN_SCHEMES
)) {
2048 TRACE("(%p %p %x): Treating URI as an hierarchical URI.\n", ptr
, data
, flags
);
2049 data
->is_opaque
= FALSE
;
2051 /* TODO: Handle hierarchical URI's, parse authority then parse the path. */
2052 if(!parse_authority(ptr
, data
, flags
))
2055 return parse_path_hierarchical(ptr
, data
, flags
);
2057 /* Reset ptr to it's starting position so opaque path parsing
2058 * begins at the correct location.
2063 /* If it reaches here, then the URI will be treated as an opaque
2067 TRACE("(%p %p %x): Treating URI as an opaque URI.\n", ptr
, data
, flags
);
2069 data
->is_opaque
= TRUE
;
2070 if(!parse_path_opaque(ptr
, data
, flags
))
2076 /* Attempts to parse the query string from the URI.
2079 * If NO_DECODE_EXTRA_INFO flag is set, then invalid percent encoded
2080 * data is allowed appear in the query string. For unknown scheme types
2081 * invalid percent encoded data is allowed to appear reguardless.
2083 static BOOL
parse_query(const WCHAR
**ptr
, parse_data
*data
, DWORD flags
) {
2084 const BOOL known_scheme
= data
->scheme_type
!= URL_SCHEME_UNKNOWN
;
2087 TRACE("(%p %p %x): URI didn't contain a query string.\n", ptr
, data
, flags
);
2094 while(**ptr
&& **ptr
!= '#') {
2095 if(**ptr
== '%' && known_scheme
&&
2096 !(flags
& Uri_CREATE_NO_DECODE_EXTRA_INFO
)) {
2097 if(!check_pct_encoded(ptr
)) {
2108 data
->query_len
= *ptr
- data
->query
;
2110 TRACE("(%p %p %x): Parsed query string %s len=%d\n", ptr
, data
, flags
,
2111 debugstr_wn(data
->query
, data
->query_len
), data
->query_len
);
2115 /* Attempts to parse the fragment from the URI.
2118 * If NO_DECODE_EXTRA_INFO flag is set, then invalid percent encoded
2119 * data is allowed appear in the query string. For unknown scheme types
2120 * invalid percent encoded data is allowed to appear reguardless.
2122 static BOOL
parse_fragment(const WCHAR
**ptr
, parse_data
*data
, DWORD flags
) {
2123 const BOOL known_scheme
= data
->scheme_type
!= URL_SCHEME_UNKNOWN
;
2126 TRACE("(%p %p %x): URI didn't contain a fragment.\n", ptr
, data
, flags
);
2130 data
->fragment
= *ptr
;
2134 if(**ptr
== '%' && known_scheme
&&
2135 !(flags
& Uri_CREATE_NO_DECODE_EXTRA_INFO
)) {
2136 if(!check_pct_encoded(ptr
)) {
2137 *ptr
= data
->fragment
;
2138 data
->fragment
= NULL
;
2147 data
->fragment_len
= *ptr
- data
->fragment
;
2149 TRACE("(%p %p %x): Parsed fragment %s len=%d\n", ptr
, data
, flags
,
2150 debugstr_wn(data
->fragment
, data
->fragment_len
), data
->fragment_len
);
2154 /* Parses and validates the components of the specified by data->uri
2155 * and stores the information it parses into 'data'.
2157 * Returns TRUE if it successfully parsed the URI. False otherwise.
2159 static BOOL
parse_uri(parse_data
*data
, DWORD flags
) {
2166 TRACE("(%p %x): BEGINNING TO PARSE URI %s.\n", data
, flags
, debugstr_w(data
->uri
));
2168 if(!parse_scheme(pptr
, data
, flags
, 0))
2171 if(!parse_hierpart(pptr
, data
, flags
))
2174 if(!parse_query(pptr
, data
, flags
))
2177 if(!parse_fragment(pptr
, data
, flags
))
2180 TRACE("(%p %x): FINISHED PARSING URI.\n", data
, flags
);
2184 static BOOL
canonicalize_username(const parse_data
*data
, Uri
*uri
, DWORD flags
, BOOL computeOnly
) {
2187 if(!data
->username
) {
2188 uri
->userinfo_start
= -1;
2192 uri
->userinfo_start
= uri
->canon_len
;
2193 for(ptr
= data
->username
; ptr
< data
->username
+data
->username_len
; ++ptr
) {
2195 /* Only decode % encoded values for known scheme types. */
2196 if(data
->scheme_type
!= URL_SCHEME_UNKNOWN
) {
2197 /* See if the value really needs decoded. */
2198 WCHAR val
= decode_pct_val(ptr
);
2199 if(is_unreserved(val
)) {
2201 uri
->canon_uri
[uri
->canon_len
] = val
;
2205 /* Move pass the hex characters. */
2210 } else if(!is_reserved(*ptr
) && !is_unreserved(*ptr
) && *ptr
!= '\\') {
2211 /* Only percent encode forbidden characters if the NO_ENCODE_FORBIDDEN_CHARACTERS flag
2214 if(!(flags
& Uri_CREATE_NO_ENCODE_FORBIDDEN_CHARACTERS
)) {
2216 pct_encode_val(*ptr
, uri
->canon_uri
+ uri
->canon_len
);
2218 uri
->canon_len
+= 3;
2224 /* Nothing special, so just copy the character over. */
2225 uri
->canon_uri
[uri
->canon_len
] = *ptr
;
2232 static BOOL
canonicalize_password(const parse_data
*data
, Uri
*uri
, DWORD flags
, BOOL computeOnly
) {
2235 if(!data
->password
) {
2236 uri
->userinfo_split
= -1;
2240 if(uri
->userinfo_start
== -1)
2241 /* Has a password, but, doesn't have a username. */
2242 uri
->userinfo_start
= uri
->canon_len
;
2244 uri
->userinfo_split
= uri
->canon_len
- uri
->userinfo_start
;
2246 /* Add the ':' to the userinfo component. */
2248 uri
->canon_uri
[uri
->canon_len
] = ':';
2251 for(ptr
= data
->password
; ptr
< data
->password
+data
->password_len
; ++ptr
) {
2253 /* Only decode % encoded values for known scheme types. */
2254 if(data
->scheme_type
!= URL_SCHEME_UNKNOWN
) {
2255 /* See if the value really needs decoded. */
2256 WCHAR val
= decode_pct_val(ptr
);
2257 if(is_unreserved(val
)) {
2259 uri
->canon_uri
[uri
->canon_len
] = val
;
2263 /* Move pass the hex characters. */
2268 } else if(!is_reserved(*ptr
) && !is_unreserved(*ptr
) && *ptr
!= '\\') {
2269 /* Only percent encode forbidden characters if the NO_ENCODE_FORBIDDEN_CHARACTERS flag
2272 if(!(flags
& Uri_CREATE_NO_ENCODE_FORBIDDEN_CHARACTERS
)) {
2274 pct_encode_val(*ptr
, uri
->canon_uri
+ uri
->canon_len
);
2276 uri
->canon_len
+= 3;
2282 /* Nothing special, so just copy the character over. */
2283 uri
->canon_uri
[uri
->canon_len
] = *ptr
;
2290 /* Canonicalizes the userinfo of the URI represented by the parse_data.
2292 * Canonicalization of the userinfo is a simple process. If there are any percent
2293 * encoded characters that fall in the "unreserved" character set, they are decoded
2294 * to their actual value. If a character is not in the "unreserved" or "reserved" sets
2295 * then it is percent encoded. Other than that the characters are copied over without
2298 static BOOL
canonicalize_userinfo(const parse_data
*data
, Uri
*uri
, DWORD flags
, BOOL computeOnly
) {
2299 uri
->userinfo_start
= uri
->userinfo_split
= -1;
2300 uri
->userinfo_len
= 0;
2302 if(!data
->username
&& !data
->password
)
2303 /* URI doesn't have userinfo, so nothing to do here. */
2306 if(!canonicalize_username(data
, uri
, flags
, computeOnly
))
2309 if(!canonicalize_password(data
, uri
, flags
, computeOnly
))
2312 uri
->userinfo_len
= uri
->canon_len
- uri
->userinfo_start
;
2314 TRACE("(%p %p %x %d): Canonicalized userinfo, userinfo_start=%d, userinfo=%s, userinfo_split=%d userinfo_len=%d.\n",
2315 data
, uri
, flags
, computeOnly
, uri
->userinfo_start
, debugstr_wn(uri
->canon_uri
+ uri
->userinfo_start
, uri
->userinfo_len
),
2316 uri
->userinfo_split
, uri
->userinfo_len
);
2318 /* Now insert the '@' after the userinfo. */
2320 uri
->canon_uri
[uri
->canon_len
] = '@';
2326 /* Attempts to canonicalize a reg_name.
2328 * Things that happen:
2329 * 1) If Uri_CREATE_NO_CANONICALIZE flag is not set, then the reg_name is
2330 * lower cased. Unless it's an unknown scheme type, which case it's
2331 * no lower cased reguardless.
2333 * 2) Unreserved % encoded characters are decoded for known
2336 * 3) Forbidden characters are % encoded as long as
2337 * Uri_CREATE_NO_ENCODE_FORBIDDEN_CHARACTERS flag is not set and
2338 * it isn't an unknown scheme type.
2340 * 4) If it's a file scheme and the host is "localhost" it's removed.
2342 static BOOL
canonicalize_reg_name(const parse_data
*data
, Uri
*uri
,
2343 DWORD flags
, BOOL computeOnly
) {
2344 static const WCHAR localhostW
[] =
2345 {'l','o','c','a','l','h','o','s','t',0};
2347 const BOOL known_scheme
= data
->scheme_type
!= URL_SCHEME_UNKNOWN
;
2349 uri
->host_start
= uri
->canon_len
;
2351 if(data
->scheme_type
== URL_SCHEME_FILE
&&
2352 data
->host_len
== lstrlenW(localhostW
)) {
2353 if(!StrCmpNIW(data
->host
, localhostW
, data
->host_len
)) {
2354 uri
->host_start
= -1;
2356 uri
->host_type
= Uri_HOST_UNKNOWN
;
2361 for(ptr
= data
->host
; ptr
< data
->host
+data
->host_len
; ++ptr
) {
2362 if(*ptr
== '%' && known_scheme
) {
2363 WCHAR val
= decode_pct_val(ptr
);
2364 if(is_unreserved(val
)) {
2365 /* If NO_CANONICALZE is not set, then windows lower cases the
2368 if(!(flags
& Uri_CREATE_NO_CANONICALIZE
) && isupperW(val
)) {
2370 uri
->canon_uri
[uri
->canon_len
] = tolowerW(val
);
2373 uri
->canon_uri
[uri
->canon_len
] = val
;
2377 /* Skip past the % encoded character. */
2381 /* Just copy the % over. */
2383 uri
->canon_uri
[uri
->canon_len
] = *ptr
;
2386 } else if(*ptr
== '\\') {
2387 /* Only unknown scheme types could have made it here with a '\\' in the host name. */
2389 uri
->canon_uri
[uri
->canon_len
] = *ptr
;
2391 } else if(!(flags
& Uri_CREATE_NO_ENCODE_FORBIDDEN_CHARACTERS
) &&
2392 !is_unreserved(*ptr
) && !is_reserved(*ptr
) && known_scheme
) {
2394 pct_encode_val(*ptr
, uri
->canon_uri
+uri
->canon_len
);
2396 /* The percent encoded value gets lower cased also. */
2397 if(!(flags
& Uri_CREATE_NO_CANONICALIZE
)) {
2398 uri
->canon_uri
[uri
->canon_len
+1] = tolowerW(uri
->canon_uri
[uri
->canon_len
+1]);
2399 uri
->canon_uri
[uri
->canon_len
+2] = tolowerW(uri
->canon_uri
[uri
->canon_len
+2]);
2403 uri
->canon_len
+= 3;
2406 if(!(flags
& Uri_CREATE_NO_CANONICALIZE
) && known_scheme
)
2407 uri
->canon_uri
[uri
->canon_len
] = tolowerW(*ptr
);
2409 uri
->canon_uri
[uri
->canon_len
] = *ptr
;
2416 uri
->host_len
= uri
->canon_len
- uri
->host_start
;
2419 TRACE("(%p %p %x %d): Canonicalize reg_name=%s len=%d\n", data
, uri
, flags
,
2420 computeOnly
, debugstr_wn(uri
->canon_uri
+uri
->host_start
, uri
->host_len
),
2424 find_domain_name(uri
->canon_uri
+uri
->host_start
, uri
->host_len
,
2425 &(uri
->domain_offset
));
2430 /* Attempts to canonicalize an implicit IPv4 address. */
2431 static BOOL
canonicalize_implicit_ipv4address(const parse_data
*data
, Uri
*uri
, DWORD flags
, BOOL computeOnly
) {
2432 uri
->host_start
= uri
->canon_len
;
2434 TRACE("%u\n", data
->implicit_ipv4
);
2435 /* For unknown scheme types Window's doesn't convert
2436 * the value into an IP address, but, it still considers
2437 * it an IPv4 address.
2439 if(data
->scheme_type
== URL_SCHEME_UNKNOWN
) {
2441 memcpy(uri
->canon_uri
+uri
->canon_len
, data
->host
, data
->host_len
*sizeof(WCHAR
));
2442 uri
->canon_len
+= data
->host_len
;
2445 uri
->canon_len
+= ui2ipv4(uri
->canon_uri
+uri
->canon_len
, data
->implicit_ipv4
);
2447 uri
->canon_len
+= ui2ipv4(NULL
, data
->implicit_ipv4
);
2450 uri
->host_len
= uri
->canon_len
- uri
->host_start
;
2451 uri
->host_type
= Uri_HOST_IPV4
;
2454 TRACE("%p %p %x %d): Canonicalized implicit IP address=%s len=%d\n",
2455 data
, uri
, flags
, computeOnly
,
2456 debugstr_wn(uri
->canon_uri
+uri
->host_start
, uri
->host_len
),
2462 /* Attempts to canonicalize an IPv4 address.
2464 * If the parse_data represents a URI that has an implicit IPv4 address
2465 * (ex. http://256/, this function will convert 256 into 0.0.1.0). If
2466 * the implicit IP address exceeds the value of UINT_MAX (maximum value
2467 * for an IPv4 address) it's canonicalized as if were a reg-name.
2469 * If the parse_data contains a partial or full IPv4 address it normalizes it.
2470 * A partial IPv4 address is something like "192.0" and would be normalized to
2471 * "192.0.0.0". With a full (or partial) IPv4 address like "192.002.01.003" would
2472 * be normalized to "192.2.1.3".
2475 * Window's ONLY normalizes IPv4 address for known scheme types (one that isn't
2476 * URL_SCHEME_UNKNOWN). For unknown scheme types, it simply copies the data from
2477 * the original URI into the canonicalized URI, but, it still recognizes URI's
2478 * host type as HOST_IPV4.
2480 static BOOL
canonicalize_ipv4address(const parse_data
*data
, Uri
*uri
, DWORD flags
, BOOL computeOnly
) {
2481 if(data
->has_implicit_ip
)
2482 return canonicalize_implicit_ipv4address(data
, uri
, flags
, computeOnly
);
2484 uri
->host_start
= uri
->canon_len
;
2486 /* Windows only normalizes for known scheme types. */
2487 if(data
->scheme_type
!= URL_SCHEME_UNKNOWN
) {
2488 /* parse_data contains a partial or full IPv4 address, so normalize it. */
2489 DWORD i
, octetDigitCount
= 0, octetCount
= 0;
2490 BOOL octetHasDigit
= FALSE
;
2492 for(i
= 0; i
< data
->host_len
; ++i
) {
2493 if(data
->host
[i
] == '0' && !octetHasDigit
) {
2494 /* Can ignore leading zeros if:
2495 * 1) It isn't the last digit of the octet.
2496 * 2) i+1 != data->host_len
2499 if(octetDigitCount
== 2 ||
2500 i
+1 == data
->host_len
||
2501 data
->host
[i
+1] == '.') {
2503 uri
->canon_uri
[uri
->canon_len
] = data
->host
[i
];
2505 TRACE("Adding zero\n");
2507 } else if(data
->host
[i
] == '.') {
2509 uri
->canon_uri
[uri
->canon_len
] = data
->host
[i
];
2512 octetDigitCount
= 0;
2513 octetHasDigit
= FALSE
;
2517 uri
->canon_uri
[uri
->canon_len
] = data
->host
[i
];
2521 octetHasDigit
= TRUE
;
2525 /* Make sure the canonicalized IP address has 4 dec-octets.
2526 * If doesn't add "0" ones until there is 4;
2528 for( ; octetCount
< 3; ++octetCount
) {
2530 uri
->canon_uri
[uri
->canon_len
] = '.';
2531 uri
->canon_uri
[uri
->canon_len
+1] = '0';
2534 uri
->canon_len
+= 2;
2537 /* Windows doesn't normalize addresses in unknown schemes. */
2539 memcpy(uri
->canon_uri
+uri
->canon_len
, data
->host
, data
->host_len
*sizeof(WCHAR
));
2540 uri
->canon_len
+= data
->host_len
;
2543 uri
->host_len
= uri
->canon_len
- uri
->host_start
;
2545 TRACE("(%p %p %x %d): Canonicalized IPv4 address, ip=%s len=%d\n",
2546 data
, uri
, flags
, computeOnly
,
2547 debugstr_wn(uri
->canon_uri
+uri
->host_start
, uri
->host_len
),
2554 /* Attempts to canonicalize the IPv6 address of the URI.
2556 * Multiple things happen during the canonicalization of an IPv6 address:
2557 * 1) Any leading zero's in an h16 component are removed.
2558 * Ex: [0001:0022::] -> [1:22::]
2560 * 2) The longest sequence of zero h16 components are compressed
2561 * into a "::" (elision). If there's a tie, the first is choosen.
2563 * Ex: [0:0:0:0:1:6:7:8] -> [::1:6:7:8]
2564 * [0:0:0:0:1:2::] -> [::1:2:0:0]
2565 * [0:0:1:2:0:0:7:8] -> [::1:2:0:0:7:8]
2567 * 3) If an IPv4 address is attached to the IPv6 address, it's
2569 * Ex: [::001.002.022.000] -> [::1.2.22.0]
2571 * 4) If an elision is present, but, only represents 1 h16 component
2574 * Ex: [1::2:3:4:5:6:7] -> [1:0:2:3:4:5:6:7]
2576 * 5) If the IPv6 address contains an IPv4 address and there exists
2577 * at least 1 non-zero h16 component the IPv4 address is converted
2578 * into two h16 components, otherwise it's normalized and kept as is.
2580 * Ex: [::192.200.003.4] -> [::192.200.3.4]
2581 * [ffff::192.200.003.4] -> [ffff::c0c8:3041]
2584 * For unknown scheme types Windows simply copies the address over without any
2587 * IPv4 address can be included in an elision if all its components are 0's.
2589 static BOOL
canonicalize_ipv6address(const parse_data
*data
, Uri
*uri
,
2590 DWORD flags
, BOOL computeOnly
) {
2591 uri
->host_start
= uri
->canon_len
;
2593 if(data
->scheme_type
== URL_SCHEME_UNKNOWN
) {
2595 memcpy(uri
->canon_uri
+uri
->canon_len
, data
->host
, data
->host_len
*sizeof(WCHAR
));
2596 uri
->canon_len
+= data
->host_len
;
2600 DWORD i
, elision_len
;
2602 if(!ipv6_to_number(&(data
->ipv6_address
), values
)) {
2603 TRACE("(%p %p %x %d): Failed to compute numerical value for IPv6 address.\n",
2604 data
, uri
, flags
, computeOnly
);
2609 uri
->canon_uri
[uri
->canon_len
] = '[';
2612 /* Find where the elision should occur (if any). */
2613 compute_elision_location(&(data
->ipv6_address
), values
, &elision_start
, &elision_len
);
2615 TRACE("%p %p %x %d): Elision starts at %d, len=%u\n", data
, uri
, flags
,
2616 computeOnly
, elision_start
, elision_len
);
2618 for(i
= 0; i
< 8; ++i
) {
2619 BOOL in_elision
= (elision_start
> -1 && i
>= elision_start
&&
2620 i
< elision_start
+elision_len
);
2621 BOOL do_ipv4
= (i
== 6 && data
->ipv6_address
.ipv4
&& !in_elision
&&
2622 data
->ipv6_address
.h16_count
== 0);
2624 if(i
== elision_start
) {
2626 uri
->canon_uri
[uri
->canon_len
] = ':';
2627 uri
->canon_uri
[uri
->canon_len
+1] = ':';
2629 uri
->canon_len
+= 2;
2632 /* We can ignore the current component if we're in the elision. */
2636 /* We only add a ':' if we're not at i == 0, or when we're at
2637 * the very end of elision range since the ':' colon was handled
2638 * earlier. Otherwise we would end up with ":::" after elision.
2640 if(i
!= 0 && !(elision_start
> -1 && i
== elision_start
+elision_len
)) {
2642 uri
->canon_uri
[uri
->canon_len
] = ':';
2650 /* Combine the two parts of the IPv4 address values. */
2656 len
= ui2ipv4(uri
->canon_uri
+uri
->canon_len
, val
);
2658 len
= ui2ipv4(NULL
, val
);
2660 uri
->canon_len
+= len
;
2663 /* Write a regular h16 component to the URI. */
2665 /* Short circuit for the trivial case. */
2666 if(values
[i
] == 0) {
2668 uri
->canon_uri
[uri
->canon_len
] = '0';
2671 static const WCHAR formatW
[] = {'%','x',0};
2674 uri
->canon_len
+= sprintfW(uri
->canon_uri
+uri
->canon_len
,
2675 formatW
, values
[i
]);
2678 uri
->canon_len
+= sprintfW(tmp
, formatW
, values
[i
]);
2684 /* Add the closing ']'. */
2686 uri
->canon_uri
[uri
->canon_len
] = ']';
2690 uri
->host_len
= uri
->canon_len
- uri
->host_start
;
2693 TRACE("(%p %p %x %d): Canonicalized IPv6 address %s, len=%d\n", data
, uri
, flags
,
2694 computeOnly
, debugstr_wn(uri
->canon_uri
+uri
->host_start
, uri
->host_len
),
2700 /* Attempts to canonicalize the host of the URI (if any). */
2701 static BOOL
canonicalize_host(const parse_data
*data
, Uri
*uri
, DWORD flags
, BOOL computeOnly
) {
2702 uri
->host_start
= -1;
2704 uri
->domain_offset
= -1;
2707 switch(data
->host_type
) {
2709 uri
->host_type
= Uri_HOST_DNS
;
2710 if(!canonicalize_reg_name(data
, uri
, flags
, computeOnly
))
2715 uri
->host_type
= Uri_HOST_IPV4
;
2716 if(!canonicalize_ipv4address(data
, uri
, flags
, computeOnly
))
2721 if(!canonicalize_ipv6address(data
, uri
, flags
, computeOnly
))
2724 uri
->host_type
= Uri_HOST_IPV6
;
2726 case Uri_HOST_UNKNOWN
:
2727 if(data
->host_len
> 0 || data
->scheme_type
!= URL_SCHEME_FILE
) {
2728 uri
->host_start
= uri
->canon_len
;
2730 /* Nothing happens to unknown host types. */
2732 memcpy(uri
->canon_uri
+uri
->canon_len
, data
->host
, data
->host_len
*sizeof(WCHAR
));
2733 uri
->canon_len
+= data
->host_len
;
2734 uri
->host_len
= data
->host_len
;
2737 uri
->host_type
= Uri_HOST_UNKNOWN
;
2740 FIXME("(%p %p %x %d): Canonicalization for host type %d not supported.\n", data
,
2741 uri
, flags
, computeOnly
, data
->host_type
);
2749 static BOOL
canonicalize_port(const parse_data
*data
, Uri
*uri
, DWORD flags
, BOOL computeOnly
) {
2750 BOOL has_default_port
= FALSE
;
2751 USHORT default_port
= 0;
2754 uri
->has_port
= FALSE
;
2756 /* Check if the scheme has a default port. */
2757 for(i
= 0; i
< sizeof(default_ports
)/sizeof(default_ports
[0]); ++i
) {
2758 if(default_ports
[i
].scheme
== data
->scheme_type
) {
2759 has_default_port
= TRUE
;
2760 default_port
= default_ports
[i
].port
;
2765 if(data
->port
|| has_default_port
)
2766 uri
->has_port
= TRUE
;
2769 * 1) Has a port which is the default port.
2770 * 2) Has a port (not the default).
2771 * 3) Doesn't have a port, but, scheme has a default port.
2774 if(has_default_port
&& data
->port
&& data
->port_value
== default_port
) {
2775 /* If it's the default port and this flag isn't set, don't do anything. */
2776 if(flags
& Uri_CREATE_NO_CANONICALIZE
) {
2777 /* Copy the original port over. */
2779 uri
->canon_uri
[uri
->canon_len
] = ':';
2780 memcpy(uri
->canon_uri
+uri
->canon_len
+1, data
->port
, data
->port_len
*sizeof(WCHAR
));
2782 uri
->canon_len
+= data
->port_len
+1;
2785 uri
->port
= default_port
;
2786 } else if(data
->port
) {
2788 uri
->canon_uri
[uri
->canon_len
] = ':';
2791 if(flags
& Uri_CREATE_NO_CANONICALIZE
) {
2792 /* Copy the original over without changes. */
2794 memcpy(uri
->canon_uri
+uri
->canon_len
, data
->port
, data
->port_len
*sizeof(WCHAR
));
2795 uri
->canon_len
+= data
->port_len
;
2797 const WCHAR formatW
[] = {'%','u',0};
2800 len
= sprintfW(uri
->canon_uri
+uri
->canon_len
, formatW
, data
->port_value
);
2803 len
= sprintfW(tmp
, formatW
, data
->port_value
);
2805 uri
->canon_len
+= len
;
2808 uri
->port
= data
->port_value
;
2809 } else if(has_default_port
)
2810 uri
->port
= default_port
;
2815 /* Canonicalizes the authority of the URI represented by the parse_data. */
2816 static BOOL
canonicalize_authority(const parse_data
*data
, Uri
*uri
, DWORD flags
, BOOL computeOnly
) {
2817 uri
->authority_start
= uri
->canon_len
;
2818 uri
->authority_len
= 0;
2820 if(!canonicalize_userinfo(data
, uri
, flags
, computeOnly
))
2823 if(!canonicalize_host(data
, uri
, flags
, computeOnly
))
2826 if(!canonicalize_port(data
, uri
, flags
, computeOnly
))
2829 if(uri
->host_start
!= -1)
2830 uri
->authority_len
= uri
->canon_len
- uri
->authority_start
;
2832 uri
->authority_start
= -1;
2837 /* Attempts to canonicalize the path of a hierarchical URI.
2839 * Things that happen:
2840 * 1). Forbidden characters are percent encoded, unless the NO_ENCODE_FORBIDDEN
2841 * flag is set or it's a file URI. Forbidden characters are always encoded
2842 * for file schemes reguardless and forbidden characters are never encoded
2843 * for unknown scheme types.
2845 * 2). For known scheme types '\\' are changed to '/'.
2847 * 3). Percent encoded, unreserved characters are decoded to their actual values.
2848 * Unless the scheme type is unknown. For file schemes any percent encoded
2849 * character in the unreserved or reserved set is decoded.
2851 * 4). For File schemes if the path is starts with a drive letter and doesn't
2852 * start with a '/' then one is appended.
2853 * Ex: file://c:/test.mp3 -> file:///c:/test.mp3
2855 * 5). Dot segments are removed from the path for all scheme types
2856 * unless NO_CANONICALIZE flag is set. Dot segments aren't removed
2857 * for wildcard scheme types.
2860 * file://c:/test%20test -> file:///c:/test%2520test
2861 * file://c:/test%3Etest -> file:///c:/test%253Etest
2862 * file:///c:/test%20test -> file:///c:/test%20test
2863 * file:///c:/test%test -> file:///c:/test%25test
2865 static BOOL
canonicalize_path_hierarchical(const parse_data
*data
, Uri
*uri
,
2866 DWORD flags
, BOOL computeOnly
) {
2868 const BOOL known_scheme
= data
->scheme_type
!= URL_SCHEME_UNKNOWN
;
2869 const BOOL is_file
= data
->scheme_type
== URL_SCHEME_FILE
;
2871 BOOL escape_pct
= FALSE
;
2874 uri
->path_start
= -1;
2879 uri
->path_start
= uri
->canon_len
;
2882 if(is_file
&& uri
->host_start
== -1) {
2883 /* Check if a '/' needs to be appended for the file scheme. */
2884 if(data
->path_len
> 1 && is_drive_path(ptr
) && !(flags
& Uri_CREATE_FILE_USE_DOS_PATH
)) {
2886 uri
->canon_uri
[uri
->canon_len
] = '/';
2889 } else if(*ptr
== '/') {
2890 if(!(flags
& Uri_CREATE_FILE_USE_DOS_PATH
)) {
2891 /* Copy the extra '/' over. */
2893 uri
->canon_uri
[uri
->canon_len
] = '/';
2899 if(is_drive_path(ptr
)) {
2901 uri
->canon_uri
[uri
->canon_len
] = *ptr
;
2902 /* If theres a '|' after the drive letter, convert it to a ':'. */
2903 uri
->canon_uri
[uri
->canon_len
+1] = ':';
2906 uri
->canon_len
+= 2;
2910 for(; ptr
< data
->path
+data
->path_len
; ++ptr
) {
2912 const WCHAR
*tmp
= ptr
;
2915 /* Check if the % represents a valid encoded char, or if it needs encoded. */
2916 BOOL force_encode
= !check_pct_encoded(&tmp
) && is_file
;
2917 val
= decode_pct_val(ptr
);
2919 if(force_encode
|| escape_pct
) {
2920 /* Escape the percent sign in the file URI. */
2922 pct_encode_val(*ptr
, uri
->canon_uri
+uri
->canon_len
);
2923 uri
->canon_len
+= 3;
2924 } else if((is_unreserved(val
) && known_scheme
) ||
2925 (is_file
&& (is_unreserved(val
) || is_reserved(val
)))) {
2927 uri
->canon_uri
[uri
->canon_len
] = val
;
2934 uri
->canon_uri
[uri
->canon_len
] = *ptr
;
2937 } else if(*ptr
== '/' && is_file
&& (flags
& Uri_CREATE_FILE_USE_DOS_PATH
)) {
2938 /* Convert the '/' back to a '\\'. */
2940 uri
->canon_uri
[uri
->canon_len
] = '\\';
2942 } else if(*ptr
== '\\' && known_scheme
) {
2943 if(is_file
&& (flags
& Uri_CREATE_FILE_USE_DOS_PATH
)) {
2944 /* Don't convert the '\\' to a '/'. */
2946 uri
->canon_uri
[uri
->canon_len
] = *ptr
;
2950 uri
->canon_uri
[uri
->canon_len
] = '/';
2953 } else if(known_scheme
&& !is_unreserved(*ptr
) && !is_reserved(*ptr
) &&
2954 (!(flags
& Uri_CREATE_NO_ENCODE_FORBIDDEN_CHARACTERS
) || is_file
)) {
2955 if(is_file
&& (flags
& Uri_CREATE_FILE_USE_DOS_PATH
)) {
2956 /* Don't escape the character. */
2958 uri
->canon_uri
[uri
->canon_len
] = *ptr
;
2961 /* Escape the forbidden character. */
2963 pct_encode_val(*ptr
, uri
->canon_uri
+uri
->canon_len
);
2964 uri
->canon_len
+= 3;
2968 uri
->canon_uri
[uri
->canon_len
] = *ptr
;
2973 uri
->path_len
= uri
->canon_len
- uri
->path_start
;
2975 /* Removing the dot segments only happens when it's not in
2976 * computeOnly mode and it's not a wildcard scheme. File schemes
2977 * with USE_DOS_PATH set don't get dot segments removed.
2979 if(!(is_file
&& (flags
& Uri_CREATE_FILE_USE_DOS_PATH
)) &&
2980 data
->scheme_type
!= URL_SCHEME_WILDCARD
) {
2981 if(!(flags
& Uri_CREATE_NO_CANONICALIZE
) && !computeOnly
) {
2982 /* Remove the dot segments (if any) and reset everything to the new
2985 DWORD new_len
= remove_dot_segments(uri
->canon_uri
+uri
->path_start
, uri
->path_len
);
2986 uri
->canon_len
-= uri
->path_len
-new_len
;
2987 uri
->path_len
= new_len
;
2992 TRACE("Canonicalized path %s len=%d\n",
2993 debugstr_wn(uri
->canon_uri
+uri
->path_start
, uri
->path_len
),
2999 /* Attempts to canonicalize the path for an opaque URI.
3001 * For known scheme types:
3002 * 1) forbidden characters are percent encoded if
3003 * NO_ENCODE_FORBIDDEN_CHARACTERS isn't set.
3005 * 2) Percent encoded, unreserved characters are decoded
3006 * to their actual values, for known scheme types.
3008 * 3) '\\' are changed to '/' for known scheme types
3009 * except for mailto schemes.
3011 * 4) For file schemes, if USE_DOS_PATH is set all '/'
3012 * are converted to backslashes.
3014 * 5) For file schemes, if USE_DOS_PATH isn't set all '\'
3015 * are converted to forward slashes.
3017 static BOOL
canonicalize_path_opaque(const parse_data
*data
, Uri
*uri
, DWORD flags
, BOOL computeOnly
) {
3019 const BOOL known_scheme
= data
->scheme_type
!= URL_SCHEME_UNKNOWN
;
3020 const BOOL is_file
= data
->scheme_type
== URL_SCHEME_FILE
;
3023 uri
->path_start
= -1;
3028 uri
->path_start
= uri
->canon_len
;
3030 /* Windows doesn't allow a "//" to appear after the scheme
3031 * of a URI, if it's an opaque URI.
3033 if(data
->scheme
&& *(data
->path
) == '/' && *(data
->path
+1) == '/') {
3034 /* So it inserts a "/." before the "//" if it exists. */
3036 uri
->canon_uri
[uri
->canon_len
] = '/';
3037 uri
->canon_uri
[uri
->canon_len
+1] = '.';
3040 uri
->canon_len
+= 2;
3043 for(ptr
= data
->path
; ptr
< data
->path
+data
->path_len
; ++ptr
) {
3044 if(*ptr
== '%' && known_scheme
) {
3045 WCHAR val
= decode_pct_val(ptr
);
3047 if(is_unreserved(val
)) {
3049 uri
->canon_uri
[uri
->canon_len
] = val
;
3056 uri
->canon_uri
[uri
->canon_len
] = *ptr
;
3059 } else if(*ptr
== '/' && is_file
&& (flags
& Uri_CREATE_FILE_USE_DOS_PATH
)) {
3061 uri
->canon_uri
[uri
->canon_len
] = '\\';
3063 } else if(*ptr
== '\\' && is_file
) {
3064 if(!(flags
& Uri_CREATE_FILE_USE_DOS_PATH
)) {
3065 /* Convert to a '/'. */
3067 uri
->canon_uri
[uri
->canon_len
] = '/';
3070 /* Just copy it over. */
3072 uri
->canon_uri
[uri
->canon_len
] = *ptr
;
3075 } else if(known_scheme
&& !is_unreserved(*ptr
) && !is_reserved(*ptr
) &&
3076 !(flags
& Uri_CREATE_NO_ENCODE_FORBIDDEN_CHARACTERS
)) {
3077 if(is_file
&& (flags
& Uri_CREATE_FILE_USE_DOS_PATH
)) {
3078 /* Forbidden characters aren't percent encoded for file schemes
3079 * with USE_DOS_PATH set.
3082 uri
->canon_uri
[uri
->canon_len
] = *ptr
;
3084 } else if(data
->scheme_type
== URL_SCHEME_MK
&& *ptr
== '\\') {
3085 /* MK URIs don't get '\\' percent encoded. */
3087 uri
->canon_uri
[uri
->canon_len
] = *ptr
;
3091 pct_encode_val(*ptr
, uri
->canon_uri
+uri
->canon_len
);
3092 uri
->canon_len
+= 3;
3096 uri
->canon_uri
[uri
->canon_len
] = *ptr
;
3101 uri
->path_len
= uri
->canon_len
- uri
->path_start
;
3103 TRACE("(%p %p %x %d): Canonicalized opaque URI path %s len=%d\n", data
, uri
, flags
, computeOnly
,
3104 debugstr_wn(uri
->canon_uri
+uri
->path_start
, uri
->path_len
), uri
->path_len
);
3108 /* Determines how the URI represented by the parse_data should be canonicalized.
3110 * Essentially, if the parse_data represents an hierarchical URI then it calls
3111 * canonicalize_authority and the canonicalization functions for the path. If the
3112 * URI is opaque it canonicalizes the path of the URI.
3114 static BOOL
canonicalize_hierpart(const parse_data
*data
, Uri
*uri
, DWORD flags
, BOOL computeOnly
) {
3115 uri
->display_absolute
= TRUE
;
3117 if(!data
->is_opaque
) {
3118 /* "//" is only added for non-wildcard scheme types. */
3119 if(data
->scheme_type
!= URL_SCHEME_WILDCARD
) {
3121 INT pos
= uri
->canon_len
;
3123 uri
->canon_uri
[pos
] = '/';
3124 uri
->canon_uri
[pos
+1] = '/';
3126 uri
->canon_len
+= 2;
3129 if(!canonicalize_authority(data
, uri
, flags
, computeOnly
))
3132 /* TODO: Canonicalize the path of the URI. */
3133 if(!canonicalize_path_hierarchical(data
, uri
, flags
, computeOnly
))
3137 /* Opaque URI's don't have an authority. */
3138 uri
->userinfo_start
= uri
->userinfo_split
= -1;
3139 uri
->userinfo_len
= 0;
3140 uri
->host_start
= -1;
3142 uri
->host_type
= Uri_HOST_UNKNOWN
;
3143 uri
->has_port
= FALSE
;
3144 uri
->authority_start
= -1;
3145 uri
->authority_len
= 0;
3146 uri
->domain_offset
= -1;
3148 if(is_hierarchical_scheme(data
->scheme_type
)) {
3151 /* Absolute URIs aren't displayed for known scheme types
3152 * which should be hierarchical URIs.
3154 uri
->display_absolute
= FALSE
;
3156 /* Windows also sets the port for these (if they have one). */
3157 for(i
= 0; i
< sizeof(default_ports
)/sizeof(default_ports
[0]); ++i
) {
3158 if(data
->scheme_type
== default_ports
[i
].scheme
) {
3159 uri
->has_port
= TRUE
;
3160 uri
->port
= default_ports
[i
].port
;
3166 if(!canonicalize_path_opaque(data
, uri
, flags
, computeOnly
))
3170 if(uri
->path_start
> -1 && !computeOnly
)
3171 /* Finding file extensions happens for both types of URIs. */
3172 uri
->extension_offset
= find_file_extension(uri
->canon_uri
+uri
->path_start
, uri
->path_len
);
3174 uri
->extension_offset
= -1;
3179 /* Attempts to canonicalize the query string of the URI.
3181 * Things that happen:
3182 * 1) For known scheme types forbidden characters
3183 * are percent encoded, unless the NO_DECODE_EXTRA_INFO flag is set
3184 * or NO_ENCODE_FORBIDDEN_CHARACTERS is set.
3186 * 2) For known scheme types, percent encoded, unreserved characters
3187 * are decoded as long as the NO_DECODE_EXTRA_INFO flag isn't set.
3189 static BOOL
canonicalize_query(const parse_data
*data
, Uri
*uri
, DWORD flags
, BOOL computeOnly
) {
3190 const WCHAR
*ptr
, *end
;
3191 const BOOL known_scheme
= data
->scheme_type
!= URL_SCHEME_UNKNOWN
;
3194 uri
->query_start
= -1;
3199 uri
->query_start
= uri
->canon_len
;
3201 end
= data
->query
+data
->query_len
;
3202 for(ptr
= data
->query
; ptr
< end
; ++ptr
) {
3204 if(known_scheme
&& !(flags
& Uri_CREATE_NO_DECODE_EXTRA_INFO
)) {
3205 WCHAR val
= decode_pct_val(ptr
);
3206 if(is_unreserved(val
)) {
3208 uri
->canon_uri
[uri
->canon_len
] = val
;
3215 } else if(known_scheme
&& !is_unreserved(*ptr
) && !is_reserved(*ptr
)) {
3216 if(!(flags
& Uri_CREATE_NO_ENCODE_FORBIDDEN_CHARACTERS
) &&
3217 !(flags
& Uri_CREATE_NO_DECODE_EXTRA_INFO
)) {
3219 pct_encode_val(*ptr
, uri
->canon_uri
+uri
->canon_len
);
3220 uri
->canon_len
+= 3;
3226 uri
->canon_uri
[uri
->canon_len
] = *ptr
;
3230 uri
->query_len
= uri
->canon_len
- uri
->query_start
;
3233 TRACE("(%p %p %x %d): Canonicalized query string %s len=%d\n", data
, uri
, flags
,
3234 computeOnly
, debugstr_wn(uri
->canon_uri
+uri
->query_start
, uri
->query_len
),
3239 static BOOL
canonicalize_fragment(const parse_data
*data
, Uri
*uri
, DWORD flags
, BOOL computeOnly
) {
3240 const WCHAR
*ptr
, *end
;
3241 const BOOL known_scheme
= data
->scheme_type
!= URL_SCHEME_UNKNOWN
;
3243 if(!data
->fragment
) {
3244 uri
->fragment_start
= -1;
3245 uri
->fragment_len
= 0;
3249 uri
->fragment_start
= uri
->canon_len
;
3251 end
= data
->fragment
+ data
->fragment_len
;
3252 for(ptr
= data
->fragment
; ptr
< end
; ++ptr
) {
3254 if(known_scheme
&& !(flags
& Uri_CREATE_NO_DECODE_EXTRA_INFO
)) {
3255 WCHAR val
= decode_pct_val(ptr
);
3256 if(is_unreserved(val
)) {
3258 uri
->canon_uri
[uri
->canon_len
] = val
;
3265 } else if(known_scheme
&& !is_unreserved(*ptr
) && !is_reserved(*ptr
)) {
3266 if(!(flags
& Uri_CREATE_NO_ENCODE_FORBIDDEN_CHARACTERS
) &&
3267 !(flags
& Uri_CREATE_NO_DECODE_EXTRA_INFO
)) {
3269 pct_encode_val(*ptr
, uri
->canon_uri
+uri
->canon_len
);
3270 uri
->canon_len
+= 3;
3276 uri
->canon_uri
[uri
->canon_len
] = *ptr
;
3280 uri
->fragment_len
= uri
->canon_len
- uri
->fragment_start
;
3283 TRACE("(%p %p %x %d): Canonicalized fragment %s len=%d\n", data
, uri
, flags
,
3284 computeOnly
, debugstr_wn(uri
->canon_uri
+uri
->fragment_start
, uri
->fragment_len
),
3289 /* Canonicalizes the scheme information specified in the parse_data using the specified flags. */
3290 static BOOL
canonicalize_scheme(const parse_data
*data
, Uri
*uri
, DWORD flags
, BOOL computeOnly
) {
3291 uri
->scheme_start
= -1;
3292 uri
->scheme_len
= 0;
3295 /* The only type of URI that doesn't have to have a scheme is a relative
3298 if(!data
->is_relative
) {
3299 FIXME("(%p %p %x): Unable to determine the scheme type of %s.\n", data
,
3300 uri
, flags
, debugstr_w(data
->uri
));
3306 INT pos
= uri
->canon_len
;
3308 for(i
= 0; i
< data
->scheme_len
; ++i
) {
3309 /* Scheme name must be lower case after canonicalization. */
3310 uri
->canon_uri
[i
+ pos
] = tolowerW(data
->scheme
[i
]);
3313 uri
->canon_uri
[i
+ pos
] = ':';
3314 uri
->scheme_start
= pos
;
3316 TRACE("(%p %p %x): Canonicalized scheme=%s, len=%d.\n", data
, uri
, flags
,
3317 debugstr_wn(uri
->canon_uri
, uri
->scheme_len
), data
->scheme_len
);
3320 /* This happens in both computation modes. */
3321 uri
->canon_len
+= data
->scheme_len
+ 1;
3322 uri
->scheme_len
= data
->scheme_len
;
3327 /* Compute's what the length of the URI specified by the parse_data will be
3328 * after canonicalization occurs using the specified flags.
3330 * This function will return a non-zero value indicating the length of the canonicalized
3331 * URI, or -1 on error.
3333 static int compute_canonicalized_length(const parse_data
*data
, DWORD flags
) {
3336 memset(&uri
, 0, sizeof(Uri
));
3338 TRACE("(%p %x): Beginning to compute canonicalized length for URI %s\n", data
, flags
,
3339 debugstr_w(data
->uri
));
3341 if(!canonicalize_scheme(data
, &uri
, flags
, TRUE
)) {
3342 ERR("(%p %x): Failed to compute URI scheme length.\n", data
, flags
);
3346 if(!canonicalize_hierpart(data
, &uri
, flags
, TRUE
)) {
3347 ERR("(%p %x): Failed to compute URI hierpart length.\n", data
, flags
);
3351 if(!canonicalize_query(data
, &uri
, flags
, TRUE
)) {
3352 ERR("(%p %x): Failed to compute query string length.\n", data
, flags
);
3356 if(!canonicalize_fragment(data
, &uri
, flags
, TRUE
)) {
3357 ERR("(%p %x): Failed to compute fragment length.\n", data
, flags
);
3361 TRACE("(%p %x): Finished computing canonicalized URI length. length=%d\n", data
, flags
, uri
.canon_len
);
3363 return uri
.canon_len
;
3366 /* Canonicalizes the URI data specified in the parse_data, using the given flags. If the
3367 * canonicalization succeededs it will store all the canonicalization information
3368 * in the pointer to the Uri.
3370 * To canonicalize a URI this function first computes what the length of the URI
3371 * specified by the parse_data will be. Once this is done it will then perfom the actual
3372 * canonicalization of the URI.
3374 static HRESULT
canonicalize_uri(const parse_data
*data
, Uri
*uri
, DWORD flags
) {
3377 uri
->canon_uri
= NULL
;
3378 len
= uri
->canon_size
= uri
->canon_len
= 0;
3380 TRACE("(%p %p %x): beginning to canonicalize URI %s.\n", data
, uri
, flags
, debugstr_w(data
->uri
));
3382 /* First try to compute the length of the URI. */
3383 len
= compute_canonicalized_length(data
, flags
);
3385 ERR("(%p %p %x): Could not compute the canonicalized length of %s.\n", data
, uri
, flags
,
3386 debugstr_w(data
->uri
));
3387 return E_INVALIDARG
;
3390 uri
->canon_uri
= heap_alloc((len
+1)*sizeof(WCHAR
));
3392 return E_OUTOFMEMORY
;
3394 uri
->canon_size
= len
;
3395 if(!canonicalize_scheme(data
, uri
, flags
, FALSE
)) {
3396 ERR("(%p %p %x): Unable to canonicalize the scheme of the URI.\n", data
, uri
, flags
);
3397 heap_free(uri
->canon_uri
);
3398 return E_INVALIDARG
;
3400 uri
->scheme_type
= data
->scheme_type
;
3402 if(!canonicalize_hierpart(data
, uri
, flags
, FALSE
)) {
3403 ERR("(%p %p %x): Unable to canonicalize the heirpart of the URI\n", data
, uri
, flags
);
3404 heap_free(uri
->canon_uri
);
3405 return E_INVALIDARG
;
3408 if(!canonicalize_query(data
, uri
, flags
, FALSE
)) {
3409 ERR("(%p %p %x): Unable to canonicalize query string of the URI.\n",
3411 return E_INVALIDARG
;
3414 if(!canonicalize_fragment(data
, uri
, flags
, FALSE
)) {
3415 ERR("(%p %p %x): Unable to canonicalize fragment of the URI.\n",
3417 return E_INVALIDARG
;
3420 /* There's a possibility we didn't use all the space we allocated
3423 if(uri
->canon_len
< uri
->canon_size
) {
3424 /* This happens if the URI is hierarchical and dot
3425 * segments were removed from it's path.
3427 WCHAR
*tmp
= heap_realloc(uri
->canon_uri
, (uri
->canon_len
+1)*sizeof(WCHAR
));
3429 return E_OUTOFMEMORY
;
3431 uri
->canon_uri
= tmp
;
3432 uri
->canon_size
= uri
->canon_len
;
3435 uri
->canon_uri
[uri
->canon_len
] = '\0';
3436 TRACE("(%p %p %x): finished canonicalizing the URI. uri=%s\n", data
, uri
, flags
, debugstr_w(uri
->canon_uri
));
3441 static HRESULT
get_builder_component(LPWSTR
*component
, DWORD
*component_len
,
3442 LPCWSTR source
, DWORD source_len
,
3443 LPCWSTR
*output
, DWORD
*output_len
)
3456 if(!(*component
) && source
) {
3457 /* Allocate 'component', and copy the contents from 'source'
3458 * into the new allocation.
3460 *component
= heap_alloc((source_len
+1)*sizeof(WCHAR
));
3462 return E_OUTOFMEMORY
;
3464 memcpy(*component
, source
, source_len
*sizeof(WCHAR
));
3465 (*component
)[source_len
] = '\0';
3466 *component_len
= source_len
;
3469 *output
= *component
;
3470 *output_len
= *component_len
;
3471 return *output
? S_OK
: S_FALSE
;
3474 /* Allocates 'component' and copies the string from 'new_value' into 'component'.
3475 * If 'prefix' is set and 'new_value' isn't NULL, then it checks if 'new_value'
3476 * starts with 'prefix'. If it doesn't then 'prefix' is prepended to 'component'.
3478 * If everything is successful, then will set 'success_flag' in 'flags'.
3480 static HRESULT
set_builder_component(LPWSTR
*component
, DWORD
*component_len
, LPCWSTR new_value
,
3481 WCHAR prefix
, DWORD
*flags
, DWORD success_flag
)
3483 heap_free(*component
);
3489 BOOL add_prefix
= FALSE
;
3490 DWORD len
= lstrlenW(new_value
);
3493 if(prefix
&& *new_value
!= prefix
) {
3495 *component
= heap_alloc((len
+2)*sizeof(WCHAR
));
3497 *component
= heap_alloc((len
+1)*sizeof(WCHAR
));
3500 return E_OUTOFMEMORY
;
3503 (*component
)[pos
++] = prefix
;
3505 memcpy(*component
+pos
, new_value
, (len
+1)*sizeof(WCHAR
));
3506 *component_len
= len
+pos
;
3509 *flags
|= success_flag
;
3513 #define URI(x) ((IUri*) &(x)->lpIUriVtbl)
3514 #define URIBUILDER(x) ((IUriBuilder*) &(x)->lpIUriBuilderVtbl)
3516 static void reset_builder(UriBuilder
*builder
) {
3518 IUri_Release(URI(builder
->uri
));
3519 builder
->uri
= NULL
;
3521 heap_free(builder
->fragment
);
3522 builder
->fragment
= NULL
;
3523 builder
->fragment_len
= 0;
3525 heap_free(builder
->host
);
3526 builder
->host
= NULL
;
3527 builder
->host_len
= 0;
3529 heap_free(builder
->password
);
3530 builder
->password
= NULL
;
3531 builder
->password_len
= 0;
3533 heap_free(builder
->path
);
3534 builder
->path
= NULL
;
3535 builder
->path_len
= 0;
3537 heap_free(builder
->query
);
3538 builder
->query
= NULL
;
3539 builder
->query_len
= 0;
3541 heap_free(builder
->scheme
);
3542 builder
->scheme
= NULL
;
3543 builder
->scheme_len
= 0;
3545 heap_free(builder
->username
);
3546 builder
->username
= NULL
;
3547 builder
->username_len
= 0;
3549 builder
->has_port
= FALSE
;
3551 builder
->modified_props
= 0;
3554 static HRESULT
validate_scheme_name(const UriBuilder
*builder
, parse_data
*data
, DWORD flags
) {
3555 const WCHAR
*component
;
3560 if(builder
->scheme
) {
3561 ptr
= builder
->scheme
;
3562 expected_len
= builder
->scheme_len
;
3563 } else if(builder
->uri
&& builder
->uri
->scheme_start
> -1) {
3564 ptr
= builder
->uri
->canon_uri
+builder
->uri
->scheme_start
;
3565 expected_len
= builder
->uri
->scheme_len
;
3567 static const WCHAR nullW
[] = {0};
3574 if(parse_scheme(pptr
, data
, flags
, ALLOW_NULL_TERM_SCHEME
) &&
3575 data
->scheme_len
== expected_len
) {
3577 TRACE("(%p %p %x): Found valid scheme component %s len=%d.\n", builder
, data
, flags
,
3578 debugstr_wn(data
->scheme
, data
->scheme_len
), data
->scheme_len
);
3580 TRACE("(%p %p %x): Invalid scheme component found %s.\n", builder
, data
, flags
,
3581 debugstr_wn(component
, expected_len
));
3582 return INET_E_INVALID_URL
;
3588 static HRESULT
validate_username(const UriBuilder
*builder
, parse_data
*data
, DWORD flags
) {
3593 if(builder
->username
) {
3594 ptr
= builder
->username
;
3595 expected_len
= builder
->username_len
;
3596 } else if(!(builder
->modified_props
& Uri_HAS_USER_NAME
) && builder
->uri
&&
3597 builder
->uri
->userinfo_start
> -1 && builder
->uri
->userinfo_split
!= 0) {
3598 /* Just use the username from the base Uri. */
3599 data
->username
= builder
->uri
->canon_uri
+builder
->uri
->userinfo_start
;
3600 data
->username_len
= (builder
->uri
->userinfo_split
> -1) ?
3601 builder
->uri
->userinfo_split
: builder
->uri
->userinfo_len
;
3609 const WCHAR
*component
= ptr
;
3611 if(parse_username(pptr
, data
, flags
, ALLOW_NULL_TERM_USER_NAME
) &&
3612 data
->username_len
== expected_len
)
3613 TRACE("(%p %p %x): Found valid username component %s len=%d.\n", builder
, data
, flags
,
3614 debugstr_wn(data
->username
, data
->username_len
), data
->username_len
);
3616 TRACE("(%p %p %x): Invalid username component found %s.\n", builder
, data
, flags
,
3617 debugstr_wn(component
, expected_len
));
3618 return INET_E_INVALID_URL
;
3625 static HRESULT
validate_password(const UriBuilder
*builder
, parse_data
*data
, DWORD flags
) {
3630 if(builder
->password
) {
3631 ptr
= builder
->password
;
3632 expected_len
= builder
->password_len
;
3633 } else if(!(builder
->modified_props
& Uri_HAS_PASSWORD
) && builder
->uri
&&
3634 builder
->uri
->userinfo_split
> -1) {
3635 data
->password
= builder
->uri
->canon_uri
+builder
->uri
->userinfo_start
+builder
->uri
->userinfo_split
+1;
3636 data
->password_len
= builder
->uri
->userinfo_len
-builder
->uri
->userinfo_split
-1;
3644 const WCHAR
*component
= ptr
;
3646 if(parse_password(pptr
, data
, flags
, ALLOW_NULL_TERM_PASSWORD
) &&
3647 data
->password_len
== expected_len
)
3648 TRACE("(%p %p %x): Found valid password component %s len=%d.\n", builder
, data
, flags
,
3649 debugstr_wn(data
->password
, data
->password_len
), data
->password_len
);
3651 TRACE("(%p %p %x): Invalid password component found %s.\n", builder
, data
, flags
,
3652 debugstr_wn(component
, expected_len
));
3653 return INET_E_INVALID_URL
;
3660 static HRESULT
validate_userinfo(const UriBuilder
*builder
, parse_data
*data
, DWORD flags
) {
3663 hr
= validate_username(builder
, data
, flags
);
3667 hr
= validate_password(builder
, data
, flags
);
3674 static HRESULT
validate_host(const UriBuilder
*builder
, parse_data
*data
, DWORD flags
) {
3680 ptr
= builder
->host
;
3681 expected_len
= builder
->host_len
;
3682 } else if(!(builder
->modified_props
& Uri_HAS_HOST
) && builder
->uri
&& builder
->uri
->host_start
> -1) {
3683 ptr
= builder
->uri
->canon_uri
+ builder
->uri
->host_start
;
3684 expected_len
= builder
->uri
->host_len
;
3689 const WCHAR
*component
= ptr
;
3690 DWORD extras
= ALLOW_BRACKETLESS_IP_LITERAL
|IGNORE_PORT_DELIMITER
|SKIP_IP_FUTURE_CHECK
;
3693 if(parse_host(pptr
, data
, flags
, extras
) && data
->host_len
== expected_len
)
3694 TRACE("(%p %p %x): Found valid host name %s len=%d type=%d.\n", builder
, data
, flags
,
3695 debugstr_wn(data
->host
, data
->host_len
), data
->host_len
, data
->host_type
);
3697 TRACE("(%p %p %x): Invalid host name found %s.\n", builder
, data
, flags
,
3698 debugstr_wn(component
, expected_len
));
3699 return INET_E_INVALID_URL
;
3706 static void setup_port(const UriBuilder
*builder
, parse_data
*data
, DWORD flags
) {
3707 if(builder
->modified_props
& Uri_HAS_PORT
) {
3708 if(builder
->has_port
) {
3709 data
->has_port
= TRUE
;
3710 data
->port_value
= builder
->port
;
3712 } else if(builder
->uri
&& builder
->uri
->has_port
) {
3713 data
->has_port
= TRUE
;
3714 data
->port_value
= builder
->uri
->port
;
3718 TRACE("(%p %p %x): Using %u as port for IUri.\n", builder
, data
, flags
, data
->port_value
);
3721 static HRESULT
validate_path(const UriBuilder
*builder
, parse_data
*data
, DWORD flags
) {
3722 const WCHAR
*ptr
= NULL
;
3727 ptr
= builder
->path
;
3728 expected_len
= builder
->path_len
;
3729 } else if(!(builder
->modified_props
& Uri_HAS_PATH
) &&
3730 builder
->uri
&& builder
->uri
->path_start
> -1) {
3731 ptr
= builder
->uri
->canon_uri
+builder
->uri
->path_start
;
3732 expected_len
= builder
->uri
->path_len
;
3737 const WCHAR
*component
= ptr
;
3740 /* How the path is validated depends on what type of
3743 valid
= data
->is_opaque
?
3744 parse_path_opaque(pptr
, data
, flags
) : parse_path_hierarchical(pptr
, data
, flags
);
3746 if(!valid
|| expected_len
!= data
->path_len
) {
3747 TRACE("(%p %p %x): Invalid path componet %s.\n", builder
, data
, flags
,
3748 debugstr_wn(component
, expected_len
));
3749 return INET_E_INVALID_URL
;
3752 TRACE("(%p %p %x): Valid path component %s len=%d.\n", builder
, data
, flags
,
3753 debugstr_wn(data
->path
, data
->path_len
), data
->path_len
);
3759 static HRESULT
validate_query(const UriBuilder
*builder
, parse_data
*data
, DWORD flags
) {
3760 const WCHAR
*ptr
= NULL
;
3764 if(builder
->query
) {
3765 ptr
= builder
->query
;
3766 expected_len
= builder
->query_len
;
3767 } else if(!(builder
->modified_props
& Uri_HAS_QUERY
) && builder
->uri
&&
3768 builder
->uri
->query_start
> -1) {
3769 ptr
= builder
->uri
->canon_uri
+builder
->uri
->query_start
;
3770 expected_len
= builder
->uri
->query_len
;
3774 const WCHAR
*component
= ptr
;
3777 if(parse_query(pptr
, data
, flags
) && expected_len
== data
->query_len
)
3778 TRACE("(%p %p %x): Valid query component %s len=%d.\n", builder
, data
, flags
,
3779 debugstr_wn(data
->query
, data
->query_len
), data
->query_len
);
3781 TRACE("(%p %p %x): Invalid query component %s.\n", builder
, data
, flags
,
3782 debugstr_wn(component
, expected_len
));
3783 return INET_E_INVALID_URL
;
3790 static HRESULT
validate_fragment(const UriBuilder
*builder
, parse_data
*data
, DWORD flags
) {
3791 const WCHAR
*ptr
= NULL
;
3795 if(builder
->fragment
) {
3796 ptr
= builder
->fragment
;
3797 expected_len
= builder
->fragment_len
;
3798 } else if(!(builder
->modified_props
& Uri_HAS_FRAGMENT
) && builder
->uri
&&
3799 builder
->uri
->fragment_start
> -1) {
3800 ptr
= builder
->uri
->canon_uri
+builder
->uri
->fragment_start
;
3801 expected_len
= builder
->uri
->fragment_len
;
3805 const WCHAR
*component
= ptr
;
3808 if(parse_query(pptr
, data
, flags
) && expected_len
== data
->fragment_len
)
3809 TRACE("(%p %p %x): Valid fragment component %s len=%d.\n", builder
, data
, flags
,
3810 debugstr_wn(data
->fragment
, data
->fragment_len
), data
->fragment_len
);
3812 TRACE("(%p %p %x): Invalid fragment component %s.\n", builder
, data
, flags
,
3813 debugstr_wn(component
, expected_len
));
3814 return INET_E_INVALID_URL
;
3821 static HRESULT
validate_components(const UriBuilder
*builder
, parse_data
*data
, DWORD flags
) {
3824 memset(data
, 0, sizeof(parse_data
));
3826 TRACE("(%p %p %x): Beginning to validate builder components.\n", builder
, data
, flags
);
3828 hr
= validate_scheme_name(builder
, data
, flags
);
3832 /* Extra validation for file schemes. */
3833 if(data
->scheme_type
== URL_SCHEME_FILE
) {
3834 if((builder
->password
|| (builder
->uri
&& builder
->uri
->userinfo_split
> -1)) ||
3835 (builder
->username
|| (builder
->uri
&& builder
->uri
->userinfo_start
> -1))) {
3836 TRACE("(%p %p %x): File schemes can't contain a username or password.\n",
3837 builder
, data
, flags
);
3838 return INET_E_INVALID_URL
;
3842 hr
= validate_userinfo(builder
, data
, flags
);
3846 hr
= validate_host(builder
, data
, flags
);
3850 /* The URI is opaque if it doesn't have an authority component. */
3851 data
->is_opaque
= !data
->username
&& !data
->password
&& !data
->host
;
3853 setup_port(builder
, data
, flags
);
3855 hr
= validate_path(builder
, data
, flags
);
3859 hr
= validate_query(builder
, data
, flags
);
3863 hr
= validate_fragment(builder
, data
, flags
);
3867 TRACE("(%p %p %x): Finished validating builder components.\n", builder
, data
, flags
);
3872 static HRESULT
build_uri(const UriBuilder
*builder
, IUri
**uri
, DWORD create_flags
,
3873 DWORD use_orig_flags
, DWORD encoding_mask
)
3881 if(encoding_mask
&& (!builder
->uri
|| builder
->modified_props
)) {
3886 /* Decide what flags should be used when creating the Uri. */
3887 if((use_orig_flags
& UriBuilder_USE_ORIGINAL_FLAGS
) && builder
->uri
)
3888 create_flags
= builder
->uri
->create_flags
;
3890 if(has_invalid_flag_combination(create_flags
)) {
3892 return E_INVALIDARG
;
3895 /* Set the default flags if they don't cause a conflict. */
3896 apply_default_flags(&create_flags
);
3899 /* Return the base IUri if no changes have been made and the create_flags match. */
3900 if(builder
->uri
&& !builder
->modified_props
&& builder
->uri
->create_flags
== create_flags
) {
3901 *uri
= URI(builder
->uri
);
3906 hr
= validate_components(builder
, &data
, create_flags
);
3915 #define URI_THIS(iface) DEFINE_THIS(Uri, IUri, iface)
3917 static HRESULT WINAPI
Uri_QueryInterface(IUri
*iface
, REFIID riid
, void **ppv
)
3919 Uri
*This
= URI_THIS(iface
);
3921 if(IsEqualGUID(&IID_IUnknown
, riid
)) {
3922 TRACE("(%p)->(IID_IUnknown %p)\n", This
, ppv
);
3924 }else if(IsEqualGUID(&IID_IUri
, riid
)) {
3925 TRACE("(%p)->(IID_IUri %p)\n", This
, ppv
);
3927 }else if(IsEqualGUID(&IID_IUriObj
, riid
)) {
3928 TRACE("(%p)->(IID_IUriObj %p)\n", This
, ppv
);
3932 TRACE("(%p)->(%s %p)\n", This
, debugstr_guid(riid
), ppv
);
3934 return E_NOINTERFACE
;
3937 IUnknown_AddRef((IUnknown
*)*ppv
);
3941 static ULONG WINAPI
Uri_AddRef(IUri
*iface
)
3943 Uri
*This
= URI_THIS(iface
);
3944 LONG ref
= InterlockedIncrement(&This
->ref
);
3946 TRACE("(%p) ref=%d\n", This
, ref
);
3951 static ULONG WINAPI
Uri_Release(IUri
*iface
)
3953 Uri
*This
= URI_THIS(iface
);
3954 LONG ref
= InterlockedDecrement(&This
->ref
);
3956 TRACE("(%p) ref=%d\n", This
, ref
);
3959 SysFreeString(This
->raw_uri
);
3960 heap_free(This
->canon_uri
);
3967 static HRESULT WINAPI
Uri_GetPropertyBSTR(IUri
*iface
, Uri_PROPERTY uriProp
, BSTR
*pbstrProperty
, DWORD dwFlags
)
3969 Uri
*This
= URI_THIS(iface
);
3971 TRACE("(%p)->(%d %p %x)\n", This
, uriProp
, pbstrProperty
, dwFlags
);
3976 if(uriProp
> Uri_PROPERTY_STRING_LAST
) {
3977 /* Windows allocates an empty BSTR for invalid Uri_PROPERTY's. */
3978 *pbstrProperty
= SysAllocStringLen(NULL
, 0);
3979 if(!(*pbstrProperty
))
3980 return E_OUTOFMEMORY
;
3982 /* It only returns S_FALSE for the ZONE property... */
3983 if(uriProp
== Uri_PROPERTY_ZONE
)
3989 /* Don't have support for flags yet. */
3991 FIXME("(%p)->(%d %p %x)\n", This
, uriProp
, pbstrProperty
, dwFlags
);
3996 case Uri_PROPERTY_ABSOLUTE_URI
:
3997 if(!This
->display_absolute
) {
3998 *pbstrProperty
= SysAllocStringLen(NULL
, 0);
4001 if(This
->scheme_type
!= URL_SCHEME_UNKNOWN
&& This
->userinfo_start
> -1) {
4002 if(This
->userinfo_len
== 0) {
4003 /* Don't include the '@' after the userinfo component. */
4004 *pbstrProperty
= SysAllocStringLen(NULL
, This
->canon_len
-1);
4006 if(*pbstrProperty
) {
4007 /* Copy everything before it. */
4008 memcpy(*pbstrProperty
, This
->canon_uri
, This
->userinfo_start
*sizeof(WCHAR
));
4010 /* And everything after it. */
4011 memcpy(*pbstrProperty
+This
->userinfo_start
, This
->canon_uri
+This
->userinfo_start
+1,
4012 (This
->canon_len
-This
->userinfo_start
-1)*sizeof(WCHAR
));
4014 } else if(This
->userinfo_split
== 0 && This
->userinfo_len
== 1) {
4015 /* Don't include the ":@" */
4016 *pbstrProperty
= SysAllocStringLen(NULL
, This
->canon_len
-2);
4018 if(*pbstrProperty
) {
4019 memcpy(*pbstrProperty
, This
->canon_uri
, This
->userinfo_start
*sizeof(WCHAR
));
4020 memcpy(*pbstrProperty
+This
->userinfo_start
, This
->canon_uri
+This
->userinfo_start
+2,
4021 (This
->canon_len
-This
->userinfo_start
-2)*sizeof(WCHAR
));
4024 *pbstrProperty
= SysAllocString(This
->canon_uri
);
4028 *pbstrProperty
= SysAllocString(This
->canon_uri
);
4033 if(!(*pbstrProperty
))
4034 hres
= E_OUTOFMEMORY
;
4037 case Uri_PROPERTY_AUTHORITY
:
4038 if(This
->authority_start
> -1) {
4039 *pbstrProperty
= SysAllocStringLen(This
->canon_uri
+This
->authority_start
, This
->authority_len
);
4042 *pbstrProperty
= SysAllocStringLen(NULL
, 0);
4046 if(!(*pbstrProperty
))
4047 hres
= E_OUTOFMEMORY
;
4050 case Uri_PROPERTY_DISPLAY_URI
:
4051 /* The Display URI contains everything except for the userinfo for known
4054 if(This
->scheme_type
!= URL_SCHEME_UNKNOWN
&& This
->userinfo_start
> -1) {
4055 *pbstrProperty
= SysAllocStringLen(NULL
, This
->canon_len
-This
->userinfo_len
);
4057 if(*pbstrProperty
) {
4058 /* Copy everything before the userinfo over. */
4059 memcpy(*pbstrProperty
, This
->canon_uri
, This
->userinfo_start
*sizeof(WCHAR
));
4060 /* Copy everything after the userinfo over. */
4061 memcpy(*pbstrProperty
+This
->userinfo_start
,
4062 This
->canon_uri
+This
->userinfo_start
+This
->userinfo_len
+1,
4063 (This
->canon_len
-(This
->userinfo_start
+This
->userinfo_len
+1))*sizeof(WCHAR
));
4066 *pbstrProperty
= SysAllocString(This
->canon_uri
);
4068 if(!(*pbstrProperty
))
4069 hres
= E_OUTOFMEMORY
;
4074 case Uri_PROPERTY_DOMAIN
:
4075 if(This
->domain_offset
> -1) {
4076 *pbstrProperty
= SysAllocStringLen(This
->canon_uri
+This
->host_start
+This
->domain_offset
,
4077 This
->host_len
-This
->domain_offset
);
4080 *pbstrProperty
= SysAllocStringLen(NULL
, 0);
4084 if(!(*pbstrProperty
))
4085 hres
= E_OUTOFMEMORY
;
4088 case Uri_PROPERTY_EXTENSION
:
4089 if(This
->extension_offset
> -1) {
4090 *pbstrProperty
= SysAllocStringLen(This
->canon_uri
+This
->path_start
+This
->extension_offset
,
4091 This
->path_len
-This
->extension_offset
);
4094 *pbstrProperty
= SysAllocStringLen(NULL
, 0);
4098 if(!(*pbstrProperty
))
4099 hres
= E_OUTOFMEMORY
;
4102 case Uri_PROPERTY_FRAGMENT
:
4103 if(This
->fragment_start
> -1) {
4104 *pbstrProperty
= SysAllocStringLen(This
->canon_uri
+This
->fragment_start
, This
->fragment_len
);
4107 *pbstrProperty
= SysAllocStringLen(NULL
, 0);
4111 if(!(*pbstrProperty
))
4112 hres
= E_OUTOFMEMORY
;
4115 case Uri_PROPERTY_HOST
:
4116 if(This
->host_start
> -1) {
4117 /* The '[' and ']' aren't included for IPv6 addresses. */
4118 if(This
->host_type
== Uri_HOST_IPV6
)
4119 *pbstrProperty
= SysAllocStringLen(This
->canon_uri
+This
->host_start
+1, This
->host_len
-2);
4121 *pbstrProperty
= SysAllocStringLen(This
->canon_uri
+This
->host_start
, This
->host_len
);
4125 *pbstrProperty
= SysAllocStringLen(NULL
, 0);
4129 if(!(*pbstrProperty
))
4130 hres
= E_OUTOFMEMORY
;
4133 case Uri_PROPERTY_PASSWORD
:
4134 if(This
->userinfo_split
> -1) {
4135 *pbstrProperty
= SysAllocStringLen(
4136 This
->canon_uri
+This
->userinfo_start
+This
->userinfo_split
+1,
4137 This
->userinfo_len
-This
->userinfo_split
-1);
4140 *pbstrProperty
= SysAllocStringLen(NULL
, 0);
4144 if(!(*pbstrProperty
))
4145 return E_OUTOFMEMORY
;
4148 case Uri_PROPERTY_PATH
:
4149 if(This
->path_start
> -1) {
4150 *pbstrProperty
= SysAllocStringLen(This
->canon_uri
+This
->path_start
, This
->path_len
);
4153 *pbstrProperty
= SysAllocStringLen(NULL
, 0);
4157 if(!(*pbstrProperty
))
4158 hres
= E_OUTOFMEMORY
;
4161 case Uri_PROPERTY_PATH_AND_QUERY
:
4162 if(This
->path_start
> -1) {
4163 *pbstrProperty
= SysAllocStringLen(This
->canon_uri
+This
->path_start
, This
->path_len
+This
->query_len
);
4165 } else if(This
->query_start
> -1) {
4166 *pbstrProperty
= SysAllocStringLen(This
->canon_uri
+This
->query_start
, This
->query_len
);
4169 *pbstrProperty
= SysAllocStringLen(NULL
, 0);
4173 if(!(*pbstrProperty
))
4174 hres
= E_OUTOFMEMORY
;
4177 case Uri_PROPERTY_QUERY
:
4178 if(This
->query_start
> -1) {
4179 *pbstrProperty
= SysAllocStringLen(This
->canon_uri
+This
->query_start
, This
->query_len
);
4182 *pbstrProperty
= SysAllocStringLen(NULL
, 0);
4186 if(!(*pbstrProperty
))
4187 hres
= E_OUTOFMEMORY
;
4190 case Uri_PROPERTY_RAW_URI
:
4191 *pbstrProperty
= SysAllocString(This
->raw_uri
);
4192 if(!(*pbstrProperty
))
4193 hres
= E_OUTOFMEMORY
;
4197 case Uri_PROPERTY_SCHEME_NAME
:
4198 if(This
->scheme_start
> -1) {
4199 *pbstrProperty
= SysAllocStringLen(This
->canon_uri
+ This
->scheme_start
, This
->scheme_len
);
4202 *pbstrProperty
= SysAllocStringLen(NULL
, 0);
4206 if(!(*pbstrProperty
))
4207 hres
= E_OUTOFMEMORY
;
4210 case Uri_PROPERTY_USER_INFO
:
4211 if(This
->userinfo_start
> -1) {
4212 *pbstrProperty
= SysAllocStringLen(This
->canon_uri
+This
->userinfo_start
, This
->userinfo_len
);
4215 *pbstrProperty
= SysAllocStringLen(NULL
, 0);
4219 if(!(*pbstrProperty
))
4220 hres
= E_OUTOFMEMORY
;
4223 case Uri_PROPERTY_USER_NAME
:
4224 if(This
->userinfo_start
> -1 && This
->userinfo_split
!= 0) {
4225 /* If userinfo_split is set, that means a password exists
4226 * so the username is only from userinfo_start to userinfo_split.
4228 if(This
->userinfo_split
> -1) {
4229 *pbstrProperty
= SysAllocStringLen(This
->canon_uri
+ This
->userinfo_start
, This
->userinfo_split
);
4232 *pbstrProperty
= SysAllocStringLen(This
->canon_uri
+ This
->userinfo_start
, This
->userinfo_len
);
4236 *pbstrProperty
= SysAllocStringLen(NULL
, 0);
4240 if(!(*pbstrProperty
))
4241 return E_OUTOFMEMORY
;
4245 FIXME("(%p)->(%d %p %x)\n", This
, uriProp
, pbstrProperty
, dwFlags
);
4252 static HRESULT WINAPI
Uri_GetPropertyLength(IUri
*iface
, Uri_PROPERTY uriProp
, DWORD
*pcchProperty
, DWORD dwFlags
)
4254 Uri
*This
= URI_THIS(iface
);
4256 TRACE("(%p)->(%d %p %x)\n", This
, uriProp
, pcchProperty
, dwFlags
);
4259 return E_INVALIDARG
;
4261 /* Can only return a length for a property if it's a string. */
4262 if(uriProp
> Uri_PROPERTY_STRING_LAST
)
4263 return E_INVALIDARG
;
4265 /* Don't have support for flags yet. */
4267 FIXME("(%p)->(%d %p %x)\n", This
, uriProp
, pcchProperty
, dwFlags
);
4272 case Uri_PROPERTY_ABSOLUTE_URI
:
4273 if(!This
->display_absolute
) {
4277 if(This
->scheme_type
!= URL_SCHEME_UNKNOWN
) {
4278 if(This
->userinfo_start
> -1 && This
->userinfo_len
== 0)
4279 /* Don't include the '@' in the length. */
4280 *pcchProperty
= This
->canon_len
-1;
4281 else if(This
->userinfo_start
> -1 && This
->userinfo_len
== 1 &&
4282 This
->userinfo_split
== 0)
4283 /* Don't include the ":@" in the length. */
4284 *pcchProperty
= This
->canon_len
-2;
4286 *pcchProperty
= This
->canon_len
;
4288 *pcchProperty
= This
->canon_len
;
4294 case Uri_PROPERTY_AUTHORITY
:
4295 *pcchProperty
= This
->authority_len
;
4296 hres
= (This
->authority_start
> -1) ? S_OK
: S_FALSE
;
4298 case Uri_PROPERTY_DISPLAY_URI
:
4299 if(This
->scheme_type
!= URL_SCHEME_UNKNOWN
&& This
->userinfo_start
> -1)
4300 *pcchProperty
= This
->canon_len
-This
->userinfo_len
-1;
4302 *pcchProperty
= This
->canon_len
;
4306 case Uri_PROPERTY_DOMAIN
:
4307 if(This
->domain_offset
> -1)
4308 *pcchProperty
= This
->host_len
- This
->domain_offset
;
4312 hres
= (This
->domain_offset
> -1) ? S_OK
: S_FALSE
;
4314 case Uri_PROPERTY_EXTENSION
:
4315 if(This
->extension_offset
> -1) {
4316 *pcchProperty
= This
->path_len
- This
->extension_offset
;
4324 case Uri_PROPERTY_FRAGMENT
:
4325 *pcchProperty
= This
->fragment_len
;
4326 hres
= (This
->fragment_start
> -1) ? S_OK
: S_FALSE
;
4328 case Uri_PROPERTY_HOST
:
4329 *pcchProperty
= This
->host_len
;
4331 /* '[' and ']' aren't included in the length. */
4332 if(This
->host_type
== Uri_HOST_IPV6
)
4335 hres
= (This
->host_start
> -1) ? S_OK
: S_FALSE
;
4337 case Uri_PROPERTY_PASSWORD
:
4338 *pcchProperty
= (This
->userinfo_split
> -1) ? This
->userinfo_len
-This
->userinfo_split
-1 : 0;
4339 hres
= (This
->userinfo_split
> -1) ? S_OK
: S_FALSE
;
4341 case Uri_PROPERTY_PATH
:
4342 *pcchProperty
= This
->path_len
;
4343 hres
= (This
->path_start
> -1) ? S_OK
: S_FALSE
;
4345 case Uri_PROPERTY_PATH_AND_QUERY
:
4346 *pcchProperty
= This
->path_len
+This
->query_len
;
4347 hres
= (This
->path_start
> -1 || This
->query_start
> -1) ? S_OK
: S_FALSE
;
4349 case Uri_PROPERTY_QUERY
:
4350 *pcchProperty
= This
->query_len
;
4351 hres
= (This
->query_start
> -1) ? S_OK
: S_FALSE
;
4353 case Uri_PROPERTY_RAW_URI
:
4354 *pcchProperty
= SysStringLen(This
->raw_uri
);
4357 case Uri_PROPERTY_SCHEME_NAME
:
4358 *pcchProperty
= This
->scheme_len
;
4359 hres
= (This
->scheme_start
> -1) ? S_OK
: S_FALSE
;
4361 case Uri_PROPERTY_USER_INFO
:
4362 *pcchProperty
= This
->userinfo_len
;
4363 hres
= (This
->userinfo_start
> -1) ? S_OK
: S_FALSE
;
4365 case Uri_PROPERTY_USER_NAME
:
4366 *pcchProperty
= (This
->userinfo_split
> -1) ? This
->userinfo_split
: This
->userinfo_len
;
4367 if(This
->userinfo_split
== 0)
4370 hres
= (This
->userinfo_start
> -1) ? S_OK
: S_FALSE
;
4373 FIXME("(%p)->(%d %p %x)\n", This
, uriProp
, pcchProperty
, dwFlags
);
4380 static HRESULT WINAPI
Uri_GetPropertyDWORD(IUri
*iface
, Uri_PROPERTY uriProp
, DWORD
*pcchProperty
, DWORD dwFlags
)
4382 Uri
*This
= URI_THIS(iface
);
4385 TRACE("(%p)->(%d %p %x)\n", This
, uriProp
, pcchProperty
, dwFlags
);
4388 return E_INVALIDARG
;
4390 /* Microsoft's implementation for the ZONE property of a URI seems to be lacking...
4391 * From what I can tell, instead of checking which URLZONE the URI belongs to it
4392 * simply assigns URLZONE_INVALID and returns E_NOTIMPL. This also applies to the GetZone
4395 if(uriProp
== Uri_PROPERTY_ZONE
) {
4396 *pcchProperty
= URLZONE_INVALID
;
4400 if(uriProp
< Uri_PROPERTY_DWORD_START
) {
4402 return E_INVALIDARG
;
4406 case Uri_PROPERTY_HOST_TYPE
:
4407 *pcchProperty
= This
->host_type
;
4410 case Uri_PROPERTY_PORT
:
4411 if(!This
->has_port
) {
4415 *pcchProperty
= This
->port
;
4420 case Uri_PROPERTY_SCHEME
:
4421 *pcchProperty
= This
->scheme_type
;
4425 FIXME("(%p)->(%d %p %x)\n", This
, uriProp
, pcchProperty
, dwFlags
);
4432 static HRESULT WINAPI
Uri_HasProperty(IUri
*iface
, Uri_PROPERTY uriProp
, BOOL
*pfHasProperty
)
4434 Uri
*This
= URI_THIS(iface
);
4435 TRACE("(%p)->(%d %p)\n", This
, uriProp
, pfHasProperty
);
4438 return E_INVALIDARG
;
4441 case Uri_PROPERTY_ABSOLUTE_URI
:
4442 *pfHasProperty
= This
->display_absolute
;
4444 case Uri_PROPERTY_AUTHORITY
:
4445 *pfHasProperty
= This
->authority_start
> -1;
4447 case Uri_PROPERTY_DISPLAY_URI
:
4448 *pfHasProperty
= TRUE
;
4450 case Uri_PROPERTY_DOMAIN
:
4451 *pfHasProperty
= This
->domain_offset
> -1;
4453 case Uri_PROPERTY_EXTENSION
:
4454 *pfHasProperty
= This
->extension_offset
> -1;
4456 case Uri_PROPERTY_FRAGMENT
:
4457 *pfHasProperty
= This
->fragment_start
> -1;
4459 case Uri_PROPERTY_HOST
:
4460 *pfHasProperty
= This
->host_start
> -1;
4462 case Uri_PROPERTY_PASSWORD
:
4463 *pfHasProperty
= This
->userinfo_split
> -1;
4465 case Uri_PROPERTY_PATH
:
4466 *pfHasProperty
= This
->path_start
> -1;
4468 case Uri_PROPERTY_PATH_AND_QUERY
:
4469 *pfHasProperty
= (This
->path_start
> -1 || This
->query_start
> -1);
4471 case Uri_PROPERTY_QUERY
:
4472 *pfHasProperty
= This
->query_start
> -1;
4474 case Uri_PROPERTY_RAW_URI
:
4475 *pfHasProperty
= TRUE
;
4477 case Uri_PROPERTY_SCHEME_NAME
:
4478 *pfHasProperty
= This
->scheme_start
> -1;
4480 case Uri_PROPERTY_USER_INFO
:
4481 *pfHasProperty
= This
->userinfo_start
> -1;
4483 case Uri_PROPERTY_USER_NAME
:
4484 if(This
->userinfo_split
== 0)
4485 *pfHasProperty
= FALSE
;
4487 *pfHasProperty
= This
->userinfo_start
> -1;
4489 case Uri_PROPERTY_HOST_TYPE
:
4490 *pfHasProperty
= TRUE
;
4492 case Uri_PROPERTY_PORT
:
4493 *pfHasProperty
= This
->has_port
;
4495 case Uri_PROPERTY_SCHEME
:
4496 *pfHasProperty
= TRUE
;
4498 case Uri_PROPERTY_ZONE
:
4499 *pfHasProperty
= FALSE
;
4502 FIXME("(%p)->(%d %p): Unsupported property type.\n", This
, uriProp
, pfHasProperty
);
4509 static HRESULT WINAPI
Uri_GetAbsoluteUri(IUri
*iface
, BSTR
*pstrAbsoluteUri
)
4511 TRACE("(%p)->(%p)\n", iface
, pstrAbsoluteUri
);
4512 return Uri_GetPropertyBSTR(iface
, Uri_PROPERTY_ABSOLUTE_URI
, pstrAbsoluteUri
, 0);
4515 static HRESULT WINAPI
Uri_GetAuthority(IUri
*iface
, BSTR
*pstrAuthority
)
4517 TRACE("(%p)->(%p)\n", iface
, pstrAuthority
);
4518 return Uri_GetPropertyBSTR(iface
, Uri_PROPERTY_AUTHORITY
, pstrAuthority
, 0);
4521 static HRESULT WINAPI
Uri_GetDisplayUri(IUri
*iface
, BSTR
*pstrDisplayUri
)
4523 TRACE("(%p)->(%p)\n", iface
, pstrDisplayUri
);
4524 return Uri_GetPropertyBSTR(iface
, Uri_PROPERTY_DISPLAY_URI
, pstrDisplayUri
, 0);
4527 static HRESULT WINAPI
Uri_GetDomain(IUri
*iface
, BSTR
*pstrDomain
)
4529 TRACE("(%p)->(%p)\n", iface
, pstrDomain
);
4530 return Uri_GetPropertyBSTR(iface
, Uri_PROPERTY_DOMAIN
, pstrDomain
, 0);
4533 static HRESULT WINAPI
Uri_GetExtension(IUri
*iface
, BSTR
*pstrExtension
)
4535 TRACE("(%p)->(%p)\n", iface
, pstrExtension
);
4536 return Uri_GetPropertyBSTR(iface
, Uri_PROPERTY_EXTENSION
, pstrExtension
, 0);
4539 static HRESULT WINAPI
Uri_GetFragment(IUri
*iface
, BSTR
*pstrFragment
)
4541 TRACE("(%p)->(%p)\n", iface
, pstrFragment
);
4542 return Uri_GetPropertyBSTR(iface
, Uri_PROPERTY_FRAGMENT
, pstrFragment
, 0);
4545 static HRESULT WINAPI
Uri_GetHost(IUri
*iface
, BSTR
*pstrHost
)
4547 TRACE("(%p)->(%p)\n", iface
, pstrHost
);
4548 return Uri_GetPropertyBSTR(iface
, Uri_PROPERTY_HOST
, pstrHost
, 0);
4551 static HRESULT WINAPI
Uri_GetPassword(IUri
*iface
, BSTR
*pstrPassword
)
4553 TRACE("(%p)->(%p)\n", iface
, pstrPassword
);
4554 return Uri_GetPropertyBSTR(iface
, Uri_PROPERTY_PASSWORD
, pstrPassword
, 0);
4557 static HRESULT WINAPI
Uri_GetPath(IUri
*iface
, BSTR
*pstrPath
)
4559 TRACE("(%p)->(%p)\n", iface
, pstrPath
);
4560 return Uri_GetPropertyBSTR(iface
, Uri_PROPERTY_PATH
, pstrPath
, 0);
4563 static HRESULT WINAPI
Uri_GetPathAndQuery(IUri
*iface
, BSTR
*pstrPathAndQuery
)
4565 TRACE("(%p)->(%p)\n", iface
, pstrPathAndQuery
);
4566 return Uri_GetPropertyBSTR(iface
, Uri_PROPERTY_PATH_AND_QUERY
, pstrPathAndQuery
, 0);
4569 static HRESULT WINAPI
Uri_GetQuery(IUri
*iface
, BSTR
*pstrQuery
)
4571 TRACE("(%p)->(%p)\n", iface
, pstrQuery
);
4572 return Uri_GetPropertyBSTR(iface
, Uri_PROPERTY_QUERY
, pstrQuery
, 0);
4575 static HRESULT WINAPI
Uri_GetRawUri(IUri
*iface
, BSTR
*pstrRawUri
)
4577 TRACE("(%p)->(%p)\n", iface
, pstrRawUri
);
4578 return Uri_GetPropertyBSTR(iface
, Uri_PROPERTY_RAW_URI
, pstrRawUri
, 0);
4581 static HRESULT WINAPI
Uri_GetSchemeName(IUri
*iface
, BSTR
*pstrSchemeName
)
4583 TRACE("(%p)->(%p)\n", iface
, pstrSchemeName
);
4584 return Uri_GetPropertyBSTR(iface
, Uri_PROPERTY_SCHEME_NAME
, pstrSchemeName
, 0);
4587 static HRESULT WINAPI
Uri_GetUserInfo(IUri
*iface
, BSTR
*pstrUserInfo
)
4589 TRACE("(%p)->(%p)\n", iface
, pstrUserInfo
);
4590 return Uri_GetPropertyBSTR(iface
, Uri_PROPERTY_USER_INFO
, pstrUserInfo
, 0);
4593 static HRESULT WINAPI
Uri_GetUserName(IUri
*iface
, BSTR
*pstrUserName
)
4595 TRACE("(%p)->(%p)\n", iface
, pstrUserName
);
4596 return Uri_GetPropertyBSTR(iface
, Uri_PROPERTY_USER_NAME
, pstrUserName
, 0);
4599 static HRESULT WINAPI
Uri_GetHostType(IUri
*iface
, DWORD
*pdwHostType
)
4601 TRACE("(%p)->(%p)\n", iface
, pdwHostType
);
4602 return Uri_GetPropertyDWORD(iface
, Uri_PROPERTY_HOST_TYPE
, pdwHostType
, 0);
4605 static HRESULT WINAPI
Uri_GetPort(IUri
*iface
, DWORD
*pdwPort
)
4607 TRACE("(%p)->(%p)\n", iface
, pdwPort
);
4608 return Uri_GetPropertyDWORD(iface
, Uri_PROPERTY_PORT
, pdwPort
, 0);
4611 static HRESULT WINAPI
Uri_GetScheme(IUri
*iface
, DWORD
*pdwScheme
)
4613 Uri
*This
= URI_THIS(iface
);
4614 TRACE("(%p)->(%p)\n", This
, pdwScheme
);
4615 return Uri_GetPropertyDWORD(iface
, Uri_PROPERTY_SCHEME
, pdwScheme
, 0);
4618 static HRESULT WINAPI
Uri_GetZone(IUri
*iface
, DWORD
*pdwZone
)
4620 TRACE("(%p)->(%p)\n", iface
, pdwZone
);
4621 return Uri_GetPropertyDWORD(iface
, Uri_PROPERTY_ZONE
,pdwZone
, 0);
4624 static HRESULT WINAPI
Uri_GetProperties(IUri
*iface
, DWORD
*pdwProperties
)
4626 Uri
*This
= URI_THIS(iface
);
4627 TRACE("(%p)->(%p)\n", This
, pdwProperties
);
4630 return E_INVALIDARG
;
4632 /* All URIs have these. */
4633 *pdwProperties
= Uri_HAS_DISPLAY_URI
|Uri_HAS_RAW_URI
|Uri_HAS_SCHEME
|Uri_HAS_HOST_TYPE
;
4635 if(This
->display_absolute
)
4636 *pdwProperties
|= Uri_HAS_ABSOLUTE_URI
;
4638 if(This
->scheme_start
> -1)
4639 *pdwProperties
|= Uri_HAS_SCHEME_NAME
;
4641 if(This
->authority_start
> -1) {
4642 *pdwProperties
|= Uri_HAS_AUTHORITY
;
4643 if(This
->userinfo_start
> -1) {
4644 *pdwProperties
|= Uri_HAS_USER_INFO
;
4645 if(This
->userinfo_split
!= 0)
4646 *pdwProperties
|= Uri_HAS_USER_NAME
;
4648 if(This
->userinfo_split
> -1)
4649 *pdwProperties
|= Uri_HAS_PASSWORD
;
4650 if(This
->host_start
> -1)
4651 *pdwProperties
|= Uri_HAS_HOST
;
4652 if(This
->domain_offset
> -1)
4653 *pdwProperties
|= Uri_HAS_DOMAIN
;
4657 *pdwProperties
|= Uri_HAS_PORT
;
4658 if(This
->path_start
> -1)
4659 *pdwProperties
|= Uri_HAS_PATH
|Uri_HAS_PATH_AND_QUERY
;
4660 if(This
->query_start
> -1)
4661 *pdwProperties
|= Uri_HAS_QUERY
|Uri_HAS_PATH_AND_QUERY
;
4663 if(This
->extension_offset
> -1)
4664 *pdwProperties
|= Uri_HAS_EXTENSION
;
4666 if(This
->fragment_start
> -1)
4667 *pdwProperties
|= Uri_HAS_FRAGMENT
;
4672 static HRESULT WINAPI
Uri_IsEqual(IUri
*iface
, IUri
*pUri
, BOOL
*pfEqual
)
4674 Uri
*This
= URI_THIS(iface
);
4677 TRACE("(%p)->(%p %p)\n", This
, pUri
, pfEqual
);
4685 /* For some reason Windows returns S_OK here... */
4689 /* Try to convert it to a Uri (allows for a more simple comparison). */
4690 if((other
= get_uri_obj(pUri
)))
4691 *pfEqual
= are_equal_simple(This
, other
);
4693 /* Do it the hard way. */
4694 FIXME("(%p)->(%p %p) No support for unknown IUri's yet.\n", iface
, pUri
, pfEqual
);
4703 static const IUriVtbl UriVtbl
= {
4707 Uri_GetPropertyBSTR
,
4708 Uri_GetPropertyLength
,
4709 Uri_GetPropertyDWORD
,
4720 Uri_GetPathAndQuery
,
4734 /***********************************************************************
4735 * CreateUri (urlmon.@)
4737 * Creates a new IUri object using the URI represented by pwzURI. This function
4738 * parses and validates the components of pwzURI and then canonicalizes the
4739 * parsed components.
4742 * pwzURI [I] The URI to parse, validate, and canonicalize.
4743 * dwFlags [I] Flags which can affect how the parsing/canonicalization is performed.
4744 * dwReserved [I] Reserved (not used).
4745 * ppURI [O] The resulting IUri after parsing/canonicalization occurs.
4748 * Success: Returns S_OK. ppURI contains the pointer to the newly allocated IUri.
4749 * Failure: E_INVALIDARG if there's invalid flag combinations in dwFlags, or an
4750 * invalid parameters, or pwzURI doesn't represnt a valid URI.
4751 * E_OUTOFMEMORY if any memory allocation fails.
4755 * Uri_CREATE_CANONICALIZE, Uri_CREATE_DECODE_EXTRA_INFO, Uri_CREATE_CRACK_UNKNOWN_SCHEMES,
4756 * Uri_CREATE_PRE_PROCESS_HTML_URI, Uri_CREATE_NO_IE_SETTINGS.
4758 HRESULT WINAPI
CreateUri(LPCWSTR pwzURI
, DWORD dwFlags
, DWORD_PTR dwReserved
, IUri
**ppURI
)
4760 const DWORD supported_flags
= Uri_CREATE_ALLOW_RELATIVE
|Uri_CREATE_ALLOW_IMPLICIT_WILDCARD_SCHEME
|
4761 Uri_CREATE_ALLOW_IMPLICIT_FILE_SCHEME
|Uri_CREATE_NO_CANONICALIZE
|Uri_CREATE_CANONICALIZE
|
4762 Uri_CREATE_DECODE_EXTRA_INFO
|Uri_CREATE_NO_DECODE_EXTRA_INFO
|Uri_CREATE_CRACK_UNKNOWN_SCHEMES
|
4763 Uri_CREATE_NO_CRACK_UNKNOWN_SCHEMES
|Uri_CREATE_PRE_PROCESS_HTML_URI
|Uri_CREATE_NO_PRE_PROCESS_HTML_URI
|
4764 Uri_CREATE_NO_IE_SETTINGS
|Uri_CREATE_NO_ENCODE_FORBIDDEN_CHARACTERS
|Uri_CREATE_FILE_USE_DOS_PATH
;
4769 TRACE("(%s %x %x %p)\n", debugstr_w(pwzURI
), dwFlags
, (DWORD
)dwReserved
, ppURI
);
4772 return E_INVALIDARG
;
4774 if(!pwzURI
|| !*pwzURI
) {
4776 return E_INVALIDARG
;
4779 /* Check for invalid flags. */
4780 if(has_invalid_flag_combination(dwFlags
)) {
4782 return E_INVALIDARG
;
4785 /* Currently unsupported. */
4786 if(dwFlags
& ~supported_flags
)
4787 FIXME("Ignoring unsupported flag(s) %x\n", dwFlags
& ~supported_flags
);
4789 ret
= heap_alloc(sizeof(Uri
));
4791 return E_OUTOFMEMORY
;
4793 ret
->lpIUriVtbl
= &UriVtbl
;
4796 /* Explicitly set the default flags if it doesn't cause a flag conflict. */
4797 apply_default_flags(&dwFlags
);
4799 /* Pre process the URI, unless told otherwise. */
4800 if(!(dwFlags
& Uri_CREATE_NO_PRE_PROCESS_HTML_URI
))
4801 ret
->raw_uri
= pre_process_uri(pwzURI
);
4803 ret
->raw_uri
= SysAllocString(pwzURI
);
4807 return E_OUTOFMEMORY
;
4810 memset(&data
, 0, sizeof(parse_data
));
4811 data
.uri
= ret
->raw_uri
;
4813 /* Validate and parse the URI into it's components. */
4814 if(!parse_uri(&data
, dwFlags
)) {
4815 /* Encountered an unsupported or invalid URI */
4816 SysFreeString(ret
->raw_uri
);
4819 return E_INVALIDARG
;
4822 /* Canonicalize the URI. */
4823 hr
= canonicalize_uri(&data
, ret
, dwFlags
);
4825 SysFreeString(ret
->raw_uri
);
4831 ret
->create_flags
= dwFlags
;
4837 /***********************************************************************
4838 * CreateUriWithFragment (urlmon.@)
4840 * Creates a new IUri object. This is almost the same as CreateUri, expect that
4841 * it allows you to explicitly specify a fragment (pwzFragment) for pwzURI.
4844 * pwzURI [I] The URI to parse and perform canonicalization on.
4845 * pwzFragment [I] The explict fragment string which should be added to pwzURI.
4846 * dwFlags [I] The flags which will be passed to CreateUri.
4847 * dwReserved [I] Reserved (not used).
4848 * ppURI [O] The resulting IUri after parsing/canonicalization.
4851 * Success: S_OK. ppURI contains the pointer to the newly allocated IUri.
4852 * Failure: E_INVALIDARG if pwzURI already contains a fragment and pwzFragment
4853 * isn't NULL. Will also return E_INVALIDARG for the same reasons as
4854 * CreateUri will. E_OUTOFMEMORY if any allocations fail.
4856 HRESULT WINAPI
CreateUriWithFragment(LPCWSTR pwzURI
, LPCWSTR pwzFragment
, DWORD dwFlags
,
4857 DWORD_PTR dwReserved
, IUri
**ppURI
)
4860 TRACE("(%s %s %x %x %p)\n", debugstr_w(pwzURI
), debugstr_w(pwzFragment
), dwFlags
, (DWORD
)dwReserved
, ppURI
);
4863 return E_INVALIDARG
;
4867 return E_INVALIDARG
;
4870 /* Check if a fragment should be appended to the URI string. */
4873 DWORD uri_len
, frag_len
;
4876 /* Check if the original URI already has a fragment component. */
4877 if(StrChrW(pwzURI
, '#')) {
4879 return E_INVALIDARG
;
4882 uri_len
= lstrlenW(pwzURI
);
4883 frag_len
= lstrlenW(pwzFragment
);
4885 /* If the fragment doesn't start with a '#', one will be added. */
4886 add_pound
= *pwzFragment
!= '#';
4889 uriW
= heap_alloc((uri_len
+frag_len
+2)*sizeof(WCHAR
));
4891 uriW
= heap_alloc((uri_len
+frag_len
+1)*sizeof(WCHAR
));
4894 return E_OUTOFMEMORY
;
4896 memcpy(uriW
, pwzURI
, uri_len
*sizeof(WCHAR
));
4898 uriW
[uri_len
++] = '#';
4899 memcpy(uriW
+uri_len
, pwzFragment
, (frag_len
+1)*sizeof(WCHAR
));
4901 hres
= CreateUri(uriW
, dwFlags
, 0, ppURI
);
4905 /* A fragment string wasn't specified, so just forward the call. */
4906 hres
= CreateUri(pwzURI
, dwFlags
, 0, ppURI
);
4911 #define URIBUILDER_THIS(iface) DEFINE_THIS(UriBuilder, IUriBuilder, iface)
4913 static HRESULT WINAPI
UriBuilder_QueryInterface(IUriBuilder
*iface
, REFIID riid
, void **ppv
)
4915 UriBuilder
*This
= URIBUILDER_THIS(iface
);
4917 if(IsEqualGUID(&IID_IUnknown
, riid
)) {
4918 TRACE("(%p)->(IID_IUnknown %p)\n", This
, ppv
);
4919 *ppv
= URIBUILDER(This
);
4920 }else if(IsEqualGUID(&IID_IUriBuilder
, riid
)) {
4921 TRACE("(%p)->(IID_IUri %p)\n", This
, ppv
);
4922 *ppv
= URIBUILDER(This
);
4924 TRACE("(%p)->(%s %p)\n", This
, debugstr_guid(riid
), ppv
);
4926 return E_NOINTERFACE
;
4929 IUnknown_AddRef((IUnknown
*)*ppv
);
4933 static ULONG WINAPI
UriBuilder_AddRef(IUriBuilder
*iface
)
4935 UriBuilder
*This
= URIBUILDER_THIS(iface
);
4936 LONG ref
= InterlockedIncrement(&This
->ref
);
4938 TRACE("(%p) ref=%d\n", This
, ref
);
4943 static ULONG WINAPI
UriBuilder_Release(IUriBuilder
*iface
)
4945 UriBuilder
*This
= URIBUILDER_THIS(iface
);
4946 LONG ref
= InterlockedDecrement(&This
->ref
);
4948 TRACE("(%p) ref=%d\n", This
, ref
);
4951 if(This
->uri
) IUri_Release(URI(This
->uri
));
4952 heap_free(This
->fragment
);
4953 heap_free(This
->host
);
4954 heap_free(This
->password
);
4955 heap_free(This
->path
);
4956 heap_free(This
->query
);
4957 heap_free(This
->scheme
);
4958 heap_free(This
->username
);
4965 static HRESULT WINAPI
UriBuilder_CreateUriSimple(IUriBuilder
*iface
,
4966 DWORD dwAllowEncodingPropertyMask
,
4967 DWORD_PTR dwReserved
,
4970 UriBuilder
*This
= URIBUILDER_THIS(iface
);
4972 TRACE("(%p)->(%d %d %p)\n", This
, dwAllowEncodingPropertyMask
, (DWORD
)dwReserved
, ppIUri
);
4974 hr
= build_uri(This
, ppIUri
, 0, UriBuilder_USE_ORIGINAL_FLAGS
, dwAllowEncodingPropertyMask
);
4976 FIXME("(%p)->(%d %d %p)\n", This
, dwAllowEncodingPropertyMask
, (DWORD
)dwReserved
, ppIUri
);
4980 static HRESULT WINAPI
UriBuilder_CreateUri(IUriBuilder
*iface
,
4981 DWORD dwCreateFlags
,
4982 DWORD dwAllowEncodingPropertyMask
,
4983 DWORD_PTR dwReserved
,
4986 UriBuilder
*This
= URIBUILDER_THIS(iface
);
4988 TRACE("(%p)->(0x%08x %d %d %p)\n", This
, dwCreateFlags
, dwAllowEncodingPropertyMask
, (DWORD
)dwReserved
, ppIUri
);
4990 if(dwCreateFlags
== -1)
4991 hr
= build_uri(This
, ppIUri
, 0, UriBuilder_USE_ORIGINAL_FLAGS
, dwAllowEncodingPropertyMask
);
4993 hr
= build_uri(This
, ppIUri
, dwCreateFlags
, 0, dwAllowEncodingPropertyMask
);
4996 FIXME("(%p)->(0x%08x %d %d %p)\n", This
, dwCreateFlags
, dwAllowEncodingPropertyMask
, (DWORD
)dwReserved
, ppIUri
);
5000 static HRESULT WINAPI
UriBuilder_CreateUriWithFlags(IUriBuilder
*iface
,
5001 DWORD dwCreateFlags
,
5002 DWORD dwUriBuilderFlags
,
5003 DWORD dwAllowEncodingPropertyMask
,
5004 DWORD_PTR dwReserved
,
5007 UriBuilder
*This
= URIBUILDER_THIS(iface
);
5009 TRACE("(%p)->(0x%08x 0x%08x %d %d %p)\n", This
, dwCreateFlags
, dwUriBuilderFlags
,
5010 dwAllowEncodingPropertyMask
, (DWORD
)dwReserved
, ppIUri
);
5012 hr
= build_uri(This
, ppIUri
, dwCreateFlags
, dwUriBuilderFlags
, dwAllowEncodingPropertyMask
);
5014 FIXME("(%p)->(0x%08x 0x%08x %d %d %p)\n", This
, dwCreateFlags
, dwUriBuilderFlags
,
5015 dwAllowEncodingPropertyMask
, (DWORD
)dwReserved
, ppIUri
);
5019 static HRESULT WINAPI
UriBuilder_GetIUri(IUriBuilder
*iface
, IUri
**ppIUri
)
5021 UriBuilder
*This
= URIBUILDER_THIS(iface
);
5022 TRACE("(%p)->(%p)\n", This
, ppIUri
);
5028 IUri
*uri
= URI(This
->uri
);
5037 static HRESULT WINAPI
UriBuilder_SetIUri(IUriBuilder
*iface
, IUri
*pIUri
)
5039 UriBuilder
*This
= URIBUILDER_THIS(iface
);
5040 TRACE("(%p)->(%p)\n", This
, pIUri
);
5045 if((uri
= get_uri_obj(pIUri
))) {
5046 /* Only reset the builder if it's Uri isn't the same as
5047 * the Uri passed to the function.
5049 if(This
->uri
!= uri
) {
5050 reset_builder(This
);
5054 This
->port
= uri
->port
;
5059 FIXME("(%p)->(%p) Unknown IUri types not supported yet.\n", This
, pIUri
);
5062 } else if(This
->uri
)
5063 /* Only reset the builder if it's Uri isn't NULL. */
5064 reset_builder(This
);
5069 static HRESULT WINAPI
UriBuilder_GetFragment(IUriBuilder
*iface
, DWORD
*pcchFragment
, LPCWSTR
*ppwzFragment
)
5071 UriBuilder
*This
= URIBUILDER_THIS(iface
);
5072 TRACE("(%p)->(%p %p)\n", This
, pcchFragment
, ppwzFragment
);
5074 if(!This
->uri
|| This
->uri
->fragment_start
== -1 || This
->modified_props
& Uri_HAS_FRAGMENT
)
5075 return get_builder_component(&This
->fragment
, &This
->fragment_len
, NULL
, 0, ppwzFragment
, pcchFragment
);
5077 return get_builder_component(&This
->fragment
, &This
->fragment_len
, This
->uri
->canon_uri
+This
->uri
->fragment_start
,
5078 This
->uri
->fragment_len
, ppwzFragment
, pcchFragment
);
5081 static HRESULT WINAPI
UriBuilder_GetHost(IUriBuilder
*iface
, DWORD
*pcchHost
, LPCWSTR
*ppwzHost
)
5083 UriBuilder
*This
= URIBUILDER_THIS(iface
);
5084 TRACE("(%p)->(%p %p)\n", This
, pcchHost
, ppwzHost
);
5086 if(!This
->uri
|| This
->uri
->host_start
== -1 || This
->modified_props
& Uri_HAS_HOST
)
5087 return get_builder_component(&This
->host
, &This
->host_len
, NULL
, 0, ppwzHost
, pcchHost
);
5089 if(This
->uri
->host_type
== Uri_HOST_IPV6
)
5090 /* Don't include the '[' and ']' around the address. */
5091 return get_builder_component(&This
->host
, &This
->host_len
, This
->uri
->canon_uri
+This
->uri
->host_start
+1,
5092 This
->uri
->host_len
-2, ppwzHost
, pcchHost
);
5094 return get_builder_component(&This
->host
, &This
->host_len
, This
->uri
->canon_uri
+This
->uri
->host_start
,
5095 This
->uri
->host_len
, ppwzHost
, pcchHost
);
5099 static HRESULT WINAPI
UriBuilder_GetPassword(IUriBuilder
*iface
, DWORD
*pcchPassword
, LPCWSTR
*ppwzPassword
)
5101 UriBuilder
*This
= URIBUILDER_THIS(iface
);
5102 TRACE("(%p)->(%p %p)\n", This
, pcchPassword
, ppwzPassword
);
5104 if(!This
->uri
|| This
->uri
->userinfo_split
== -1 || This
->modified_props
& Uri_HAS_PASSWORD
)
5105 return get_builder_component(&This
->password
, &This
->password_len
, NULL
, 0, ppwzPassword
, pcchPassword
);
5107 const WCHAR
*start
= This
->uri
->canon_uri
+This
->uri
->userinfo_start
+This
->uri
->userinfo_split
+1;
5108 DWORD len
= This
->uri
->userinfo_len
-This
->uri
->userinfo_split
-1;
5109 return get_builder_component(&This
->password
, &This
->password_len
, start
, len
, ppwzPassword
, pcchPassword
);
5113 static HRESULT WINAPI
UriBuilder_GetPath(IUriBuilder
*iface
, DWORD
*pcchPath
, LPCWSTR
*ppwzPath
)
5115 UriBuilder
*This
= URIBUILDER_THIS(iface
);
5116 TRACE("(%p)->(%p %p)\n", This
, pcchPath
, ppwzPath
);
5118 if(!This
->uri
|| This
->uri
->path_start
== -1 || This
->modified_props
& Uri_HAS_PATH
)
5119 return get_builder_component(&This
->path
, &This
->path_len
, NULL
, 0, ppwzPath
, pcchPath
);
5121 return get_builder_component(&This
->path
, &This
->path_len
, This
->uri
->canon_uri
+This
->uri
->path_start
,
5122 This
->uri
->path_len
, ppwzPath
, pcchPath
);
5125 static HRESULT WINAPI
UriBuilder_GetPort(IUriBuilder
*iface
, BOOL
*pfHasPort
, DWORD
*pdwPort
)
5127 UriBuilder
*This
= URIBUILDER_THIS(iface
);
5128 TRACE("(%p)->(%p %p)\n", This
, pfHasPort
, pdwPort
);
5141 *pfHasPort
= This
->has_port
;
5142 *pdwPort
= This
->port
;
5146 static HRESULT WINAPI
UriBuilder_GetQuery(IUriBuilder
*iface
, DWORD
*pcchQuery
, LPCWSTR
*ppwzQuery
)
5148 UriBuilder
*This
= URIBUILDER_THIS(iface
);
5149 TRACE("(%p)->(%p %p)\n", This
, pcchQuery
, ppwzQuery
);
5151 if(!This
->uri
|| This
->uri
->query_start
== -1 || This
->modified_props
& Uri_HAS_QUERY
)
5152 return get_builder_component(&This
->query
, &This
->query_len
, NULL
, 0, ppwzQuery
, pcchQuery
);
5154 return get_builder_component(&This
->query
, &This
->query_len
, This
->uri
->canon_uri
+This
->uri
->query_start
,
5155 This
->uri
->query_len
, ppwzQuery
, pcchQuery
);
5158 static HRESULT WINAPI
UriBuilder_GetSchemeName(IUriBuilder
*iface
, DWORD
*pcchSchemeName
, LPCWSTR
*ppwzSchemeName
)
5160 UriBuilder
*This
= URIBUILDER_THIS(iface
);
5161 TRACE("(%p)->(%p %p)\n", This
, pcchSchemeName
, ppwzSchemeName
);
5163 if(!This
->uri
|| This
->uri
->scheme_start
== -1 || This
->modified_props
& Uri_HAS_SCHEME_NAME
)
5164 return get_builder_component(&This
->scheme
, &This
->scheme_len
, NULL
, 0, ppwzSchemeName
, pcchSchemeName
);
5166 return get_builder_component(&This
->scheme
, &This
->scheme_len
, This
->uri
->canon_uri
+This
->uri
->scheme_start
,
5167 This
->uri
->scheme_len
, ppwzSchemeName
, pcchSchemeName
);
5170 static HRESULT WINAPI
UriBuilder_GetUserName(IUriBuilder
*iface
, DWORD
*pcchUserName
, LPCWSTR
*ppwzUserName
)
5172 UriBuilder
*This
= URIBUILDER_THIS(iface
);
5173 TRACE("(%p)->(%p %p)\n", This
, pcchUserName
, ppwzUserName
);
5175 if(!This
->uri
|| This
->uri
->userinfo_start
== -1 || This
->uri
->userinfo_split
== 0 ||
5176 This
->modified_props
& Uri_HAS_USER_NAME
)
5177 return get_builder_component(&This
->username
, &This
->username_len
, NULL
, 0, ppwzUserName
, pcchUserName
);
5179 const WCHAR
*start
= This
->uri
->canon_uri
+This
->uri
->userinfo_start
;
5181 /* Check if there's a password in the userinfo section. */
5182 if(This
->uri
->userinfo_split
> -1)
5183 /* Don't include the password. */
5184 return get_builder_component(&This
->username
, &This
->username_len
, start
,
5185 This
->uri
->userinfo_split
, ppwzUserName
, pcchUserName
);
5187 return get_builder_component(&This
->username
, &This
->username_len
, start
,
5188 This
->uri
->userinfo_len
, ppwzUserName
, pcchUserName
);
5192 static HRESULT WINAPI
UriBuilder_SetFragment(IUriBuilder
*iface
, LPCWSTR pwzNewValue
)
5194 UriBuilder
*This
= URIBUILDER_THIS(iface
);
5195 TRACE("(%p)->(%s)\n", This
, debugstr_w(pwzNewValue
));
5196 return set_builder_component(&This
->fragment
, &This
->fragment_len
, pwzNewValue
, '#',
5197 &This
->modified_props
, Uri_HAS_FRAGMENT
);
5200 static HRESULT WINAPI
UriBuilder_SetHost(IUriBuilder
*iface
, LPCWSTR pwzNewValue
)
5202 UriBuilder
*This
= URIBUILDER_THIS(iface
);
5203 TRACE("(%p)->(%s)\n", This
, debugstr_w(pwzNewValue
));
5205 /* Host name can't be set to NULL. */
5207 return E_INVALIDARG
;
5209 return set_builder_component(&This
->host
, &This
->host_len
, pwzNewValue
, 0,
5210 &This
->modified_props
, Uri_HAS_HOST
);
5213 static HRESULT WINAPI
UriBuilder_SetPassword(IUriBuilder
*iface
, LPCWSTR pwzNewValue
)
5215 UriBuilder
*This
= URIBUILDER_THIS(iface
);
5216 TRACE("(%p)->(%s)\n", This
, debugstr_w(pwzNewValue
));
5217 return set_builder_component(&This
->password
, &This
->password_len
, pwzNewValue
, 0,
5218 &This
->modified_props
, Uri_HAS_PASSWORD
);
5221 static HRESULT WINAPI
UriBuilder_SetPath(IUriBuilder
*iface
, LPCWSTR pwzNewValue
)
5223 UriBuilder
*This
= URIBUILDER_THIS(iface
);
5224 TRACE("(%p)->(%s)\n", This
, debugstr_w(pwzNewValue
));
5225 return set_builder_component(&This
->path
, &This
->path_len
, pwzNewValue
, 0,
5226 &This
->modified_props
, Uri_HAS_PATH
);
5229 static HRESULT WINAPI
UriBuilder_SetPort(IUriBuilder
*iface
, BOOL fHasPort
, DWORD dwNewValue
)
5231 UriBuilder
*This
= URIBUILDER_THIS(iface
);
5232 TRACE("(%p)->(%d %d)\n", This
, fHasPort
, dwNewValue
);
5234 This
->has_port
= fHasPort
;
5235 This
->port
= dwNewValue
;
5236 This
->modified_props
|= Uri_HAS_PORT
;
5240 static HRESULT WINAPI
UriBuilder_SetQuery(IUriBuilder
*iface
, LPCWSTR pwzNewValue
)
5242 UriBuilder
*This
= URIBUILDER_THIS(iface
);
5243 TRACE("(%p)->(%s)\n", This
, debugstr_w(pwzNewValue
));
5244 return set_builder_component(&This
->query
, &This
->query_len
, pwzNewValue
, '?',
5245 &This
->modified_props
, Uri_HAS_QUERY
);
5248 static HRESULT WINAPI
UriBuilder_SetSchemeName(IUriBuilder
*iface
, LPCWSTR pwzNewValue
)
5250 UriBuilder
*This
= URIBUILDER_THIS(iface
);
5251 TRACE("(%p)->(%s)\n", This
, debugstr_w(pwzNewValue
));
5253 /* Only set the scheme name if it's not NULL or empty. */
5254 if(!pwzNewValue
|| !*pwzNewValue
)
5255 return E_INVALIDARG
;
5257 return set_builder_component(&This
->scheme
, &This
->scheme_len
, pwzNewValue
, 0,
5258 &This
->modified_props
, Uri_HAS_SCHEME_NAME
);
5261 static HRESULT WINAPI
UriBuilder_SetUserName(IUriBuilder
*iface
, LPCWSTR pwzNewValue
)
5263 UriBuilder
*This
= URIBUILDER_THIS(iface
);
5264 TRACE("(%p)->(%s)\n", This
, debugstr_w(pwzNewValue
));
5265 return set_builder_component(&This
->username
, &This
->username_len
, pwzNewValue
, 0,
5266 &This
->modified_props
, Uri_HAS_USER_NAME
);
5269 static HRESULT WINAPI
UriBuilder_RemoveProperties(IUriBuilder
*iface
, DWORD dwPropertyMask
)
5271 const DWORD accepted_flags
= Uri_HAS_AUTHORITY
|Uri_HAS_DOMAIN
|Uri_HAS_EXTENSION
|Uri_HAS_FRAGMENT
|Uri_HAS_HOST
|
5272 Uri_HAS_PASSWORD
|Uri_HAS_PATH
|Uri_HAS_PATH_AND_QUERY
|Uri_HAS_QUERY
|
5273 Uri_HAS_USER_INFO
|Uri_HAS_USER_NAME
;
5275 UriBuilder
*This
= URIBUILDER_THIS(iface
);
5276 TRACE("(%p)->(0x%08x)\n", This
, dwPropertyMask
);
5278 if(dwPropertyMask
& ~accepted_flags
)
5279 return E_INVALIDARG
;
5281 if(dwPropertyMask
& Uri_HAS_FRAGMENT
)
5282 UriBuilder_SetFragment(iface
, NULL
);
5284 /* Even though you can't set the host name to NULL or an
5285 * empty string, you can still remove it... for some reason.
5287 if(dwPropertyMask
& Uri_HAS_HOST
)
5288 set_builder_component(&This
->host
, &This
->host_len
, NULL
, 0,
5289 &This
->modified_props
, Uri_HAS_HOST
);
5291 if(dwPropertyMask
& Uri_HAS_PASSWORD
)
5292 UriBuilder_SetPassword(iface
, NULL
);
5294 if(dwPropertyMask
& Uri_HAS_PATH
)
5295 UriBuilder_SetPath(iface
, NULL
);
5297 if(dwPropertyMask
& Uri_HAS_PORT
)
5298 UriBuilder_SetPort(iface
, FALSE
, 0);
5300 if(dwPropertyMask
& Uri_HAS_QUERY
)
5301 UriBuilder_SetQuery(iface
, NULL
);
5303 if(dwPropertyMask
& Uri_HAS_USER_NAME
)
5304 UriBuilder_SetUserName(iface
, NULL
);
5309 static HRESULT WINAPI
UriBuilder_HasBeenModified(IUriBuilder
*iface
, BOOL
*pfModified
)
5311 UriBuilder
*This
= URIBUILDER_THIS(iface
);
5312 TRACE("(%p)->(%p)\n", This
, pfModified
);
5317 *pfModified
= This
->modified_props
> 0;
5321 #undef URIBUILDER_THIS
5323 static const IUriBuilderVtbl UriBuilderVtbl
= {
5324 UriBuilder_QueryInterface
,
5327 UriBuilder_CreateUriSimple
,
5328 UriBuilder_CreateUri
,
5329 UriBuilder_CreateUriWithFlags
,
5332 UriBuilder_GetFragment
,
5334 UriBuilder_GetPassword
,
5337 UriBuilder_GetQuery
,
5338 UriBuilder_GetSchemeName
,
5339 UriBuilder_GetUserName
,
5340 UriBuilder_SetFragment
,
5342 UriBuilder_SetPassword
,
5345 UriBuilder_SetQuery
,
5346 UriBuilder_SetSchemeName
,
5347 UriBuilder_SetUserName
,
5348 UriBuilder_RemoveProperties
,
5349 UriBuilder_HasBeenModified
,
5352 /***********************************************************************
5353 * CreateIUriBuilder (urlmon.@)
5355 HRESULT WINAPI
CreateIUriBuilder(IUri
*pIUri
, DWORD dwFlags
, DWORD_PTR dwReserved
, IUriBuilder
**ppIUriBuilder
)
5359 TRACE("(%p %x %x %p)\n", pIUri
, dwFlags
, (DWORD
)dwReserved
, ppIUriBuilder
);
5364 ret
= heap_alloc_zero(sizeof(UriBuilder
));
5366 return E_OUTOFMEMORY
;
5368 ret
->lpIUriBuilderVtbl
= &UriBuilderVtbl
;
5374 if((uri
= get_uri_obj(pIUri
))) {
5379 /* Windows doesn't set 'has_port' to TRUE in this case. */
5380 ret
->port
= uri
->port
;
5384 *ppIUriBuilder
= NULL
;
5385 FIXME("(%p %x %x %p): Unknown IUri types not supported yet.\n", pIUri
, dwFlags
,
5386 (DWORD
)dwReserved
, ppIUriBuilder
);
5391 *ppIUriBuilder
= URIBUILDER(ret
);