quartz: COM cleanup for IAMFilterMiscFlags in VideoRenderer.
[wine/multimedia.git] / dlls / urlmon / uri.c
blob20e0582d1a21b99ac12ee721145367e82e810339
1 /*
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 <limits.h>
22 #include "urlmon_main.h"
23 #include "wine/debug.h"
25 #define NO_SHLWAPI_REG
26 #include "shlwapi.h"
28 #include "strsafe.h"
30 #define URI_DISPLAY_NO_ABSOLUTE_URI 0x1
31 #define URI_DISPLAY_NO_DEFAULT_PORT_AUTH 0x2
33 #define ALLOW_NULL_TERM_SCHEME 0x01
34 #define ALLOW_NULL_TERM_USER_NAME 0x02
35 #define ALLOW_NULL_TERM_PASSWORD 0x04
36 #define ALLOW_BRACKETLESS_IP_LITERAL 0x08
37 #define SKIP_IP_FUTURE_CHECK 0x10
38 #define IGNORE_PORT_DELIMITER 0x20
40 #define RAW_URI_FORCE_PORT_DISP 0x1
41 #define RAW_URI_CONVERT_TO_DOS_PATH 0x2
43 #define COMBINE_URI_FORCE_FLAG_USE 0x1
45 WINE_DEFAULT_DEBUG_CHANNEL(urlmon);
47 static const IID IID_IUriObj = {0x4b364760,0x9f51,0x11df,{0x98,0x1c,0x08,0x00,0x20,0x0c,0x9a,0x66}};
49 typedef struct {
50 IUri IUri_iface;
51 IUriBuilderFactory IUriBuilderFactory_iface;
52 IPersistStream IPersistStream_iface;
53 IMarshal IMarshal_iface;
55 LONG ref;
57 BSTR raw_uri;
59 /* Information about the canonicalized URI's buffer. */
60 WCHAR *canon_uri;
61 DWORD canon_size;
62 DWORD canon_len;
63 BOOL display_modifiers;
64 DWORD create_flags;
66 INT scheme_start;
67 DWORD scheme_len;
68 URL_SCHEME scheme_type;
70 INT userinfo_start;
71 DWORD userinfo_len;
72 INT userinfo_split;
74 INT host_start;
75 DWORD host_len;
76 Uri_HOST_TYPE host_type;
78 INT port_offset;
79 DWORD port;
80 BOOL has_port;
82 INT authority_start;
83 DWORD authority_len;
85 INT domain_offset;
87 INT path_start;
88 DWORD path_len;
89 INT extension_offset;
91 INT query_start;
92 DWORD query_len;
94 INT fragment_start;
95 DWORD fragment_len;
96 } Uri;
98 typedef struct {
99 IUriBuilder IUriBuilder_iface;
100 LONG ref;
102 Uri *uri;
103 DWORD modified_props;
105 WCHAR *fragment;
106 DWORD fragment_len;
108 WCHAR *host;
109 DWORD host_len;
111 WCHAR *password;
112 DWORD password_len;
114 WCHAR *path;
115 DWORD path_len;
117 BOOL has_port;
118 DWORD port;
120 WCHAR *query;
121 DWORD query_len;
123 WCHAR *scheme;
124 DWORD scheme_len;
126 WCHAR *username;
127 DWORD username_len;
128 } UriBuilder;
130 typedef struct {
131 const WCHAR *str;
132 DWORD len;
133 } h16;
135 typedef struct {
136 /* IPv6 addresses can hold up to 8 h16 components. */
137 h16 components[8];
138 DWORD h16_count;
140 /* An IPv6 can have 1 elision ("::"). */
141 const WCHAR *elision;
143 /* An IPv6 can contain 1 IPv4 address as the last 32bits of the address. */
144 const WCHAR *ipv4;
145 DWORD ipv4_len;
147 INT components_size;
148 INT elision_size;
149 } ipv6_address;
151 typedef struct {
152 BSTR uri;
154 BOOL is_relative;
155 BOOL is_opaque;
156 BOOL has_implicit_scheme;
157 BOOL has_implicit_ip;
158 UINT implicit_ipv4;
159 BOOL must_have_path;
161 const WCHAR *scheme;
162 DWORD scheme_len;
163 URL_SCHEME scheme_type;
165 const WCHAR *username;
166 DWORD username_len;
168 const WCHAR *password;
169 DWORD password_len;
171 const WCHAR *host;
172 DWORD host_len;
173 Uri_HOST_TYPE host_type;
175 BOOL has_ipv6;
176 ipv6_address ipv6_address;
178 BOOL has_port;
179 const WCHAR *port;
180 DWORD port_len;
181 DWORD port_value;
183 const WCHAR *path;
184 DWORD path_len;
186 const WCHAR *query;
187 DWORD query_len;
189 const WCHAR *fragment;
190 DWORD fragment_len;
191 } parse_data;
193 static const CHAR hexDigits[] = "0123456789ABCDEF";
195 /* List of scheme types/scheme names that are recognized by the IUri interface as of IE 7. */
196 static const struct {
197 URL_SCHEME scheme;
198 WCHAR scheme_name[16];
199 } recognized_schemes[] = {
200 {URL_SCHEME_FTP, {'f','t','p',0}},
201 {URL_SCHEME_HTTP, {'h','t','t','p',0}},
202 {URL_SCHEME_GOPHER, {'g','o','p','h','e','r',0}},
203 {URL_SCHEME_MAILTO, {'m','a','i','l','t','o',0}},
204 {URL_SCHEME_NEWS, {'n','e','w','s',0}},
205 {URL_SCHEME_NNTP, {'n','n','t','p',0}},
206 {URL_SCHEME_TELNET, {'t','e','l','n','e','t',0}},
207 {URL_SCHEME_WAIS, {'w','a','i','s',0}},
208 {URL_SCHEME_FILE, {'f','i','l','e',0}},
209 {URL_SCHEME_MK, {'m','k',0}},
210 {URL_SCHEME_HTTPS, {'h','t','t','p','s',0}},
211 {URL_SCHEME_SHELL, {'s','h','e','l','l',0}},
212 {URL_SCHEME_SNEWS, {'s','n','e','w','s',0}},
213 {URL_SCHEME_LOCAL, {'l','o','c','a','l',0}},
214 {URL_SCHEME_JAVASCRIPT, {'j','a','v','a','s','c','r','i','p','t',0}},
215 {URL_SCHEME_VBSCRIPT, {'v','b','s','c','r','i','p','t',0}},
216 {URL_SCHEME_ABOUT, {'a','b','o','u','t',0}},
217 {URL_SCHEME_RES, {'r','e','s',0}},
218 {URL_SCHEME_MSSHELLROOTED, {'m','s','-','s','h','e','l','l','-','r','o','o','t','e','d',0}},
219 {URL_SCHEME_MSSHELLIDLIST, {'m','s','-','s','h','e','l','l','-','i','d','l','i','s','t',0}},
220 {URL_SCHEME_MSHELP, {'h','c','p',0}},
221 {URL_SCHEME_WILDCARD, {'*',0}}
224 /* List of default ports Windows recognizes. */
225 static const struct {
226 URL_SCHEME scheme;
227 USHORT port;
228 } default_ports[] = {
229 {URL_SCHEME_FTP, 21},
230 {URL_SCHEME_HTTP, 80},
231 {URL_SCHEME_GOPHER, 70},
232 {URL_SCHEME_NNTP, 119},
233 {URL_SCHEME_TELNET, 23},
234 {URL_SCHEME_WAIS, 210},
235 {URL_SCHEME_HTTPS, 443},
238 /* List of 3-character top level domain names Windows seems to recognize.
239 * There might be more, but, these are the only ones I've found so far.
241 static const struct {
242 WCHAR tld_name[4];
243 } recognized_tlds[] = {
244 {{'c','o','m',0}},
245 {{'e','d','u',0}},
246 {{'g','o','v',0}},
247 {{'i','n','t',0}},
248 {{'m','i','l',0}},
249 {{'n','e','t',0}},
250 {{'o','r','g',0}}
253 static Uri *get_uri_obj(IUri *uri)
255 Uri *ret;
256 HRESULT hres;
258 hres = IUri_QueryInterface(uri, &IID_IUriObj, (void**)&ret);
259 return SUCCEEDED(hres) ? ret : NULL;
262 static inline BOOL is_alpha(WCHAR val) {
263 return ((val >= 'a' && val <= 'z') || (val >= 'A' && val <= 'Z'));
266 static inline BOOL is_num(WCHAR val) {
267 return (val >= '0' && val <= '9');
270 static inline BOOL is_drive_path(const WCHAR *str) {
271 return (is_alpha(str[0]) && (str[1] == ':' || str[1] == '|'));
274 static inline BOOL is_unc_path(const WCHAR *str) {
275 return (str[0] == '\\' && str[0] == '\\');
278 static inline BOOL is_forbidden_dos_path_char(WCHAR val) {
279 return (val == '>' || val == '<' || val == '\"');
282 /* A URI is implicitly a file path if it begins with
283 * a drive letter (e.g. X:) or starts with "\\" (UNC path).
285 static inline BOOL is_implicit_file_path(const WCHAR *str) {
286 return (is_unc_path(str) || (is_alpha(str[0]) && str[1] == ':'));
289 /* Checks if the URI is a hierarchical URI. A hierarchical
290 * URI is one that has "//" after the scheme.
292 static BOOL check_hierarchical(const WCHAR **ptr) {
293 const WCHAR *start = *ptr;
295 if(**ptr != '/')
296 return FALSE;
298 ++(*ptr);
299 if(**ptr != '/') {
300 *ptr = start;
301 return FALSE;
304 ++(*ptr);
305 return TRUE;
308 /* unreserved = ALPHA / DIGIT / "-" / "." / "_" / "~" */
309 static inline BOOL is_unreserved(WCHAR val) {
310 return (is_alpha(val) || is_num(val) || val == '-' || val == '.' ||
311 val == '_' || val == '~');
314 /* sub-delims = "!" / "$" / "&" / "'" / "(" / ")"
315 * / "*" / "+" / "," / ";" / "="
317 static inline BOOL is_subdelim(WCHAR val) {
318 return (val == '!' || val == '$' || val == '&' ||
319 val == '\'' || val == '(' || val == ')' ||
320 val == '*' || val == '+' || val == ',' ||
321 val == ';' || val == '=');
324 /* gen-delims = ":" / "/" / "?" / "#" / "[" / "]" / "@" */
325 static inline BOOL is_gendelim(WCHAR val) {
326 return (val == ':' || val == '/' || val == '?' ||
327 val == '#' || val == '[' || val == ']' ||
328 val == '@');
331 /* Characters that delimit the end of the authority
332 * section of a URI. Sometimes a '\\' is considered
333 * an authority delimiter.
335 static inline BOOL is_auth_delim(WCHAR val, BOOL acceptSlash) {
336 return (val == '#' || val == '/' || val == '?' ||
337 val == '\0' || (acceptSlash && val == '\\'));
340 /* reserved = gen-delims / sub-delims */
341 static inline BOOL is_reserved(WCHAR val) {
342 return (is_subdelim(val) || is_gendelim(val));
345 static inline BOOL is_hexdigit(WCHAR val) {
346 return ((val >= 'a' && val <= 'f') ||
347 (val >= 'A' && val <= 'F') ||
348 (val >= '0' && val <= '9'));
351 static inline BOOL is_path_delim(WCHAR val) {
352 return (!val || val == '#' || val == '?');
355 static inline BOOL is_slash(WCHAR c)
357 return c == '/' || c == '\\';
360 static BOOL is_default_port(URL_SCHEME scheme, DWORD port) {
361 DWORD i;
363 for(i = 0; i < sizeof(default_ports)/sizeof(default_ports[0]); ++i) {
364 if(default_ports[i].scheme == scheme && default_ports[i].port)
365 return TRUE;
368 return FALSE;
371 /* List of schemes types Windows seems to expect to be hierarchical. */
372 static inline BOOL is_hierarchical_scheme(URL_SCHEME type) {
373 return(type == URL_SCHEME_HTTP || type == URL_SCHEME_FTP ||
374 type == URL_SCHEME_GOPHER || type == URL_SCHEME_NNTP ||
375 type == URL_SCHEME_TELNET || type == URL_SCHEME_WAIS ||
376 type == URL_SCHEME_FILE || type == URL_SCHEME_HTTPS ||
377 type == URL_SCHEME_RES);
380 /* Checks if 'flags' contains an invalid combination of Uri_CREATE flags. */
381 static inline BOOL has_invalid_flag_combination(DWORD flags) {
382 return((flags & Uri_CREATE_DECODE_EXTRA_INFO && flags & Uri_CREATE_NO_DECODE_EXTRA_INFO) ||
383 (flags & Uri_CREATE_CANONICALIZE && flags & Uri_CREATE_NO_CANONICALIZE) ||
384 (flags & Uri_CREATE_CRACK_UNKNOWN_SCHEMES && flags & Uri_CREATE_NO_CRACK_UNKNOWN_SCHEMES) ||
385 (flags & Uri_CREATE_PRE_PROCESS_HTML_URI && flags & Uri_CREATE_NO_PRE_PROCESS_HTML_URI) ||
386 (flags & Uri_CREATE_IE_SETTINGS && flags & Uri_CREATE_NO_IE_SETTINGS));
389 /* Applies each default Uri_CREATE flags to 'flags' if it
390 * doesn't cause a flag conflict.
392 static void apply_default_flags(DWORD *flags) {
393 if(!(*flags & Uri_CREATE_NO_CANONICALIZE))
394 *flags |= Uri_CREATE_CANONICALIZE;
395 if(!(*flags & Uri_CREATE_NO_DECODE_EXTRA_INFO))
396 *flags |= Uri_CREATE_DECODE_EXTRA_INFO;
397 if(!(*flags & Uri_CREATE_NO_CRACK_UNKNOWN_SCHEMES))
398 *flags |= Uri_CREATE_CRACK_UNKNOWN_SCHEMES;
399 if(!(*flags & Uri_CREATE_NO_PRE_PROCESS_HTML_URI))
400 *flags |= Uri_CREATE_PRE_PROCESS_HTML_URI;
401 if(!(*flags & Uri_CREATE_IE_SETTINGS))
402 *flags |= Uri_CREATE_NO_IE_SETTINGS;
405 /* Determines if the URI is hierarchical using the information already parsed into
406 * data and using the current location of parsing in the URI string.
408 * Windows considers a URI hierarchical if one of the following is true:
409 * A.) It's a wildcard scheme.
410 * B.) It's an implicit file scheme.
411 * C.) It's a known hierarchical scheme and it has two '\\' after the scheme name.
412 * (the '\\' will be converted into "//" during canonicalization).
413 * D.) "//" appears after the scheme name (or at the beginning if no scheme is given).
415 static inline BOOL is_hierarchical_uri(const WCHAR **ptr, const parse_data *data) {
416 const WCHAR *start = *ptr;
418 if(data->scheme_type == URL_SCHEME_WILDCARD)
419 return TRUE;
420 else if(data->scheme_type == URL_SCHEME_FILE && data->has_implicit_scheme)
421 return TRUE;
422 else if(is_hierarchical_scheme(data->scheme_type) && (*ptr)[0] == '\\' && (*ptr)[1] == '\\') {
423 *ptr += 2;
424 return TRUE;
425 } else if(check_hierarchical(ptr))
426 return TRUE;
428 *ptr = start;
429 return FALSE;
432 /* Computes the size of the given IPv6 address.
433 * Each h16 component is 16 bits. If there is an IPv4 address, it's
434 * 32 bits. If there's an elision it can be 16 to 128 bits, depending
435 * on the number of other components.
437 * Modeled after google-url's CheckIPv6ComponentsSize function
439 static void compute_ipv6_comps_size(ipv6_address *address) {
440 address->components_size = address->h16_count * 2;
442 if(address->ipv4)
443 /* IPv4 address is 4 bytes. */
444 address->components_size += 4;
446 if(address->elision) {
447 /* An elision can be anywhere from 2 bytes up to 16 bytes.
448 * Its size depends on the size of the h16 and IPv4 components.
450 address->elision_size = 16 - address->components_size;
451 if(address->elision_size < 2)
452 address->elision_size = 2;
453 } else
454 address->elision_size = 0;
457 /* Taken from dlls/jscript/lex.c */
458 static int hex_to_int(WCHAR val) {
459 if(val >= '0' && val <= '9')
460 return val - '0';
461 else if(val >= 'a' && val <= 'f')
462 return val - 'a' + 10;
463 else if(val >= 'A' && val <= 'F')
464 return val - 'A' + 10;
466 return -1;
469 /* Helper function for converting a percent encoded string
470 * representation of a WCHAR value into its actual WCHAR value. If
471 * the two characters following the '%' aren't valid hex values then
472 * this function returns the NULL character.
474 * E.g.
475 * "%2E" will result in '.' being returned by this function.
477 static WCHAR decode_pct_val(const WCHAR *ptr) {
478 WCHAR ret = '\0';
480 if(*ptr == '%' && is_hexdigit(*(ptr + 1)) && is_hexdigit(*(ptr + 2))) {
481 INT a = hex_to_int(*(ptr + 1));
482 INT b = hex_to_int(*(ptr + 2));
484 ret = a << 4;
485 ret += b;
488 return ret;
491 /* Helper function for percent encoding a given character
492 * and storing the encoded value into a given buffer (dest).
494 * It's up to the calling function to ensure that there is
495 * at least enough space in 'dest' for the percent encoded
496 * value to be stored (so dest + 3 spaces available).
498 static inline void pct_encode_val(WCHAR val, WCHAR *dest) {
499 dest[0] = '%';
500 dest[1] = hexDigits[(val >> 4) & 0xf];
501 dest[2] = hexDigits[val & 0xf];
504 /* Attempts to parse the domain name from the host.
506 * This function also includes the Top-level Domain (TLD) name
507 * of the host when it tries to find the domain name. If it finds
508 * a valid domain name it will assign 'domain_start' the offset
509 * into 'host' where the domain name starts.
511 * It's implied that if there is a domain name its range is:
512 * [host+domain_start, host+host_len).
514 void find_domain_name(const WCHAR *host, DWORD host_len,
515 INT *domain_start) {
516 const WCHAR *last_tld, *sec_last_tld, *end;
518 end = host+host_len-1;
520 *domain_start = -1;
522 /* There has to be at least enough room for a '.' followed by a
523 * 3-character TLD for a domain to even exist in the host name.
525 if(host_len < 4)
526 return;
528 last_tld = memrchrW(host, '.', host_len);
529 if(!last_tld)
530 /* http://hostname -> has no domain name. */
531 return;
533 sec_last_tld = memrchrW(host, '.', last_tld-host);
534 if(!sec_last_tld) {
535 /* If the '.' is at the beginning of the host there
536 * has to be at least 3 characters in the TLD for it
537 * to be valid.
538 * Ex: .com -> .com as the domain name.
539 * .co -> has no domain name.
541 if(last_tld-host == 0) {
542 if(end-(last_tld-1) < 3)
543 return;
544 } else if(last_tld-host == 3) {
545 DWORD i;
547 /* If there are three characters in front of last_tld and
548 * they are on the list of recognized TLDs, then this
549 * host doesn't have a domain (since the host only contains
550 * a TLD name.
551 * Ex: edu.uk -> has no domain name.
552 * foo.uk -> foo.uk as the domain name.
554 for(i = 0; i < sizeof(recognized_tlds)/sizeof(recognized_tlds[0]); ++i) {
555 if(!StrCmpNIW(host, recognized_tlds[i].tld_name, 3))
556 return;
558 } else if(last_tld-host < 3)
559 /* Anything less than 3 characters is considered part
560 * of the TLD name.
561 * Ex: ak.uk -> Has no domain name.
563 return;
565 /* Otherwise the domain name is the whole host name. */
566 *domain_start = 0;
567 } else if(end+1-last_tld > 3) {
568 /* If the last_tld has more than 3 characters, then it's automatically
569 * considered the TLD of the domain name.
570 * Ex: www.winehq.org.uk.test -> uk.test as the domain name.
572 *domain_start = (sec_last_tld+1)-host;
573 } else if(last_tld - (sec_last_tld+1) < 4) {
574 DWORD i;
575 /* If the sec_last_tld is 3 characters long it HAS to be on the list of
576 * recognized to still be considered part of the TLD name, otherwise
577 * its considered the domain name.
578 * Ex: www.google.com.uk -> google.com.uk as the domain name.
579 * www.google.foo.uk -> foo.uk as the domain name.
581 if(last_tld - (sec_last_tld+1) == 3) {
582 for(i = 0; i < sizeof(recognized_tlds)/sizeof(recognized_tlds[0]); ++i) {
583 if(!StrCmpNIW(sec_last_tld+1, recognized_tlds[i].tld_name, 3)) {
584 const WCHAR *domain = memrchrW(host, '.', sec_last_tld-host);
586 if(!domain)
587 *domain_start = 0;
588 else
589 *domain_start = (domain+1) - host;
590 TRACE("Found domain name %s\n", debugstr_wn(host+*domain_start,
591 (host+host_len)-(host+*domain_start)));
592 return;
596 *domain_start = (sec_last_tld+1)-host;
597 } else {
598 /* Since the sec_last_tld is less than 3 characters it's considered
599 * part of the TLD.
600 * Ex: www.google.fo.uk -> google.fo.uk as the domain name.
602 const WCHAR *domain = memrchrW(host, '.', sec_last_tld-host);
604 if(!domain)
605 *domain_start = 0;
606 else
607 *domain_start = (domain+1) - host;
609 } else {
610 /* The second to last TLD has more than 3 characters making it
611 * the domain name.
612 * Ex: www.google.test.us -> test.us as the domain name.
614 *domain_start = (sec_last_tld+1)-host;
617 TRACE("Found domain name %s\n", debugstr_wn(host+*domain_start,
618 (host+host_len)-(host+*domain_start)));
621 /* Removes the dot segments from a hierarchical URIs path component. This
622 * function performs the removal in place.
624 * This function returns the new length of the path string.
626 static DWORD remove_dot_segments(WCHAR *path, DWORD path_len) {
627 WCHAR *out = path;
628 const WCHAR *in = out;
629 const WCHAR *end = out + path_len;
630 DWORD len;
632 while(in < end) {
633 /* Move the first path segment in the input buffer to the end of
634 * the output buffer, and any subsequent characters up to, including
635 * the next "/" character (if any) or the end of the input buffer.
637 while(in < end && !is_slash(*in))
638 *out++ = *in++;
639 if(in == end)
640 break;
641 *out++ = *in++;
643 while(in < end) {
644 if(*in != '.')
645 break;
647 /* Handle ending "/." */
648 if(in + 1 == end) {
649 ++in;
650 break;
653 /* Handle "/./" */
654 if(is_slash(in[1])) {
655 in += 2;
656 continue;
659 /* If we don't have "/../" or ending "/.." */
660 if(in[1] != '.' || (in + 2 != end && !is_slash(in[2])))
661 break;
663 /* Find the slash preceding out pointer and move out pointer to it */
664 if(out > path+1 && is_slash(*--out))
665 --out;
666 while(out > path && !is_slash(*(--out)));
667 if(is_slash(*out))
668 ++out;
669 in += 2;
670 if(in != end)
671 ++in;
675 len = out - path;
676 TRACE("(%p %d): Path after dot segments removed %s len=%d\n", path, path_len,
677 debugstr_wn(path, len), len);
678 return len;
681 /* Attempts to find the file extension in a given path. */
682 static INT find_file_extension(const WCHAR *path, DWORD path_len) {
683 const WCHAR *end;
685 for(end = path+path_len-1; end >= path && *end != '/' && *end != '\\'; --end) {
686 if(*end == '.')
687 return end-path;
690 return -1;
693 /* Computes the location where the elision should occur in the IPv6
694 * address using the numerical values of each component stored in
695 * 'values'. If the address shouldn't contain an elision then 'index'
696 * is assigned -1 as its value. Otherwise 'index' will contain the
697 * starting index (into values) where the elision should be, and 'count'
698 * will contain the number of cells the elision covers.
700 * NOTES:
701 * Windows will expand an elision if the elision only represents one h16
702 * component of the address.
704 * Ex: [1::2:3:4:5:6:7] -> [1:0:2:3:4:5:6:7]
706 * If the IPv6 address contains an IPv4 address, the IPv4 address is also
707 * considered for being included as part of an elision if all its components
708 * are zeros.
710 * Ex: [1:2:3:4:5:6:0.0.0.0] -> [1:2:3:4:5:6::]
712 static void compute_elision_location(const ipv6_address *address, const USHORT values[8],
713 INT *index, DWORD *count) {
714 DWORD i, max_len, cur_len;
715 INT max_index, cur_index;
717 max_len = cur_len = 0;
718 max_index = cur_index = -1;
719 for(i = 0; i < 8; ++i) {
720 BOOL check_ipv4 = (address->ipv4 && i == 6);
721 BOOL is_end = (check_ipv4 || i == 7);
723 if(check_ipv4) {
724 /* Check if the IPv4 address contains only zeros. */
725 if(values[i] == 0 && values[i+1] == 0) {
726 if(cur_index == -1)
727 cur_index = i;
729 cur_len += 2;
730 ++i;
732 } else if(values[i] == 0) {
733 if(cur_index == -1)
734 cur_index = i;
736 ++cur_len;
739 if(is_end || values[i] != 0) {
740 /* We only consider it for an elision if it's
741 * more than 1 component long.
743 if(cur_len > 1 && cur_len > max_len) {
744 /* Found the new elision location. */
745 max_len = cur_len;
746 max_index = cur_index;
749 /* Reset the current range for the next range of zeros. */
750 cur_index = -1;
751 cur_len = 0;
755 *index = max_index;
756 *count = max_len;
759 /* Removes all the leading and trailing white spaces or
760 * control characters from the URI and removes all control
761 * characters inside of the URI string.
763 static BSTR pre_process_uri(LPCWSTR uri) {
764 const WCHAR *start, *end, *ptr;
765 WCHAR *ptr2;
766 DWORD len;
767 BSTR ret;
769 start = uri;
770 /* Skip leading controls and whitespace. */
771 while(*start && (iscntrlW(*start) || isspaceW(*start))) ++start;
773 /* URI consisted only of control/whitespace. */
774 if(!*start)
775 return SysAllocStringLen(NULL, 0);
777 end = start + strlenW(start);
778 while(--end > start && (iscntrlW(*end) || isspaceW(*end)));
780 len = ++end - start;
781 for(ptr = start; ptr < end; ptr++) {
782 if(iscntrlW(*ptr))
783 len--;
786 ret = SysAllocStringLen(NULL, len);
787 if(!ret)
788 return NULL;
790 for(ptr = start, ptr2=ret; ptr < end; ptr++) {
791 if(!iscntrlW(*ptr))
792 *ptr2++ = *ptr;
795 return ret;
798 /* Converts the specified IPv4 address into an uint value.
800 * This function assumes that the IPv4 address has already been validated.
802 static UINT ipv4toui(const WCHAR *ip, DWORD len) {
803 UINT ret = 0;
804 DWORD comp_value = 0;
805 const WCHAR *ptr;
807 for(ptr = ip; ptr < ip+len; ++ptr) {
808 if(*ptr == '.') {
809 ret <<= 8;
810 ret += comp_value;
811 comp_value = 0;
812 } else
813 comp_value = comp_value*10 + (*ptr-'0');
816 ret <<= 8;
817 ret += comp_value;
819 return ret;
822 /* Converts an IPv4 address in numerical form into its fully qualified
823 * string form. This function returns the number of characters written
824 * to 'dest'. If 'dest' is NULL this function will return the number of
825 * characters that would have been written.
827 * It's up to the caller to ensure there's enough space in 'dest' for the
828 * address.
830 static DWORD ui2ipv4(WCHAR *dest, UINT address) {
831 static const WCHAR formatW[] =
832 {'%','u','.','%','u','.','%','u','.','%','u',0};
833 DWORD ret = 0;
834 UCHAR digits[4];
836 digits[0] = (address >> 24) & 0xff;
837 digits[1] = (address >> 16) & 0xff;
838 digits[2] = (address >> 8) & 0xff;
839 digits[3] = address & 0xff;
841 if(!dest) {
842 WCHAR tmp[16];
843 ret = sprintfW(tmp, formatW, digits[0], digits[1], digits[2], digits[3]);
844 } else
845 ret = sprintfW(dest, formatW, digits[0], digits[1], digits[2], digits[3]);
847 return ret;
850 static DWORD ui2str(WCHAR *dest, UINT value) {
851 static const WCHAR formatW[] = {'%','u',0};
852 DWORD ret = 0;
854 if(!dest) {
855 WCHAR tmp[11];
856 ret = sprintfW(tmp, formatW, value);
857 } else
858 ret = sprintfW(dest, formatW, value);
860 return ret;
863 /* Converts a h16 component (from an IPv6 address) into its
864 * numerical value.
866 * This function assumes that the h16 component has already been validated.
868 static USHORT h16tous(h16 component) {
869 DWORD i;
870 USHORT ret = 0;
872 for(i = 0; i < component.len; ++i) {
873 ret <<= 4;
874 ret += hex_to_int(component.str[i]);
877 return ret;
880 /* Converts an IPv6 address into its 128 bits (16 bytes) numerical value.
882 * This function assumes that the ipv6_address has already been validated.
884 static BOOL ipv6_to_number(const ipv6_address *address, USHORT number[8]) {
885 DWORD i, cur_component = 0;
886 BOOL already_passed_elision = FALSE;
888 for(i = 0; i < address->h16_count; ++i) {
889 if(address->elision) {
890 if(address->components[i].str > address->elision && !already_passed_elision) {
891 /* Means we just passed the elision and need to add its values to
892 * 'number' before we do anything else.
894 DWORD j = 0;
895 for(j = 0; j < address->elision_size; j+=2)
896 number[cur_component++] = 0;
898 already_passed_elision = TRUE;
902 number[cur_component++] = h16tous(address->components[i]);
905 /* Case when the elision appears after the h16 components. */
906 if(!already_passed_elision && address->elision) {
907 for(i = 0; i < address->elision_size; i+=2)
908 number[cur_component++] = 0;
911 if(address->ipv4) {
912 UINT value = ipv4toui(address->ipv4, address->ipv4_len);
914 if(cur_component != 6) {
915 ERR("(%p %p): Failed sanity check with %d\n", address, number, cur_component);
916 return FALSE;
919 number[cur_component++] = (value >> 16) & 0xffff;
920 number[cur_component] = value & 0xffff;
923 return TRUE;
926 /* Checks if the characters pointed to by 'ptr' are
927 * a percent encoded data octet.
929 * pct-encoded = "%" HEXDIG HEXDIG
931 static BOOL check_pct_encoded(const WCHAR **ptr) {
932 const WCHAR *start = *ptr;
934 if(**ptr != '%')
935 return FALSE;
937 ++(*ptr);
938 if(!is_hexdigit(**ptr)) {
939 *ptr = start;
940 return FALSE;
943 ++(*ptr);
944 if(!is_hexdigit(**ptr)) {
945 *ptr = start;
946 return FALSE;
949 ++(*ptr);
950 return TRUE;
953 /* dec-octet = DIGIT ; 0-9
954 * / %x31-39 DIGIT ; 10-99
955 * / "1" 2DIGIT ; 100-199
956 * / "2" %x30-34 DIGIT ; 200-249
957 * / "25" %x30-35 ; 250-255
959 static BOOL check_dec_octet(const WCHAR **ptr) {
960 const WCHAR *c1, *c2, *c3;
962 c1 = *ptr;
963 /* A dec-octet must be at least 1 digit long. */
964 if(*c1 < '0' || *c1 > '9')
965 return FALSE;
967 ++(*ptr);
969 c2 = *ptr;
970 /* Since the 1-digit requirement was met, it doesn't
971 * matter if this is a DIGIT value, it's considered a
972 * dec-octet.
974 if(*c2 < '0' || *c2 > '9')
975 return TRUE;
977 ++(*ptr);
979 c3 = *ptr;
980 /* Same explanation as above. */
981 if(*c3 < '0' || *c3 > '9')
982 return TRUE;
984 /* Anything > 255 isn't a valid IP dec-octet. */
985 if(*c1 >= '2' && *c2 >= '5' && *c3 >= '5') {
986 *ptr = c1;
987 return FALSE;
990 ++(*ptr);
991 return TRUE;
994 /* Checks if there is an implicit IPv4 address in the host component of the URI.
995 * The max value of an implicit IPv4 address is UINT_MAX.
997 * Ex:
998 * "234567" would be considered an implicit IPv4 address.
1000 static BOOL check_implicit_ipv4(const WCHAR **ptr, UINT *val) {
1001 const WCHAR *start = *ptr;
1002 ULONGLONG ret = 0;
1003 *val = 0;
1005 while(is_num(**ptr)) {
1006 ret = ret*10 + (**ptr - '0');
1008 if(ret > UINT_MAX) {
1009 *ptr = start;
1010 return FALSE;
1012 ++(*ptr);
1015 if(*ptr == start)
1016 return FALSE;
1018 *val = ret;
1019 return TRUE;
1022 /* Checks if the string contains an IPv4 address.
1024 * This function has a strict mode or a non-strict mode of operation
1025 * When 'strict' is set to FALSE this function will return TRUE if
1026 * the string contains at least 'dec-octet "." dec-octet' since partial
1027 * IPv4 addresses will be normalized out into full IPv4 addresses. When
1028 * 'strict' is set this function expects there to be a full IPv4 address.
1030 * IPv4address = dec-octet "." dec-octet "." dec-octet "." dec-octet
1032 static BOOL check_ipv4address(const WCHAR **ptr, BOOL strict) {
1033 const WCHAR *start = *ptr;
1035 if(!check_dec_octet(ptr)) {
1036 *ptr = start;
1037 return FALSE;
1040 if(**ptr != '.') {
1041 *ptr = start;
1042 return FALSE;
1045 ++(*ptr);
1046 if(!check_dec_octet(ptr)) {
1047 *ptr = start;
1048 return FALSE;
1051 if(**ptr != '.') {
1052 if(strict) {
1053 *ptr = start;
1054 return FALSE;
1055 } else
1056 return TRUE;
1059 ++(*ptr);
1060 if(!check_dec_octet(ptr)) {
1061 *ptr = start;
1062 return FALSE;
1065 if(**ptr != '.') {
1066 if(strict) {
1067 *ptr = start;
1068 return FALSE;
1069 } else
1070 return TRUE;
1073 ++(*ptr);
1074 if(!check_dec_octet(ptr)) {
1075 *ptr = start;
1076 return FALSE;
1079 /* Found a four digit ip address. */
1080 return TRUE;
1082 /* Tries to parse the scheme name of the URI.
1084 * scheme = ALPHA *(ALPHA | NUM | '+' | '-' | '.') as defined by RFC 3896.
1085 * NOTE: Windows accepts a number as the first character of a scheme.
1087 static BOOL parse_scheme_name(const WCHAR **ptr, parse_data *data, DWORD extras) {
1088 const WCHAR *start = *ptr;
1090 data->scheme = NULL;
1091 data->scheme_len = 0;
1093 while(**ptr) {
1094 if(**ptr == '*' && *ptr == start) {
1095 /* Might have found a wildcard scheme. If it is the next
1096 * char has to be a ':' for it to be a valid URI
1098 ++(*ptr);
1099 break;
1100 } else if(!is_num(**ptr) && !is_alpha(**ptr) && **ptr != '+' &&
1101 **ptr != '-' && **ptr != '.')
1102 break;
1104 (*ptr)++;
1107 if(*ptr == start)
1108 return FALSE;
1110 /* Schemes must end with a ':' */
1111 if(**ptr != ':' && !((extras & ALLOW_NULL_TERM_SCHEME) && !**ptr)) {
1112 *ptr = start;
1113 return FALSE;
1116 data->scheme = start;
1117 data->scheme_len = *ptr - start;
1119 ++(*ptr);
1120 return TRUE;
1123 /* Tries to deduce the corresponding URL_SCHEME for the given URI. Stores
1124 * the deduced URL_SCHEME in data->scheme_type.
1126 static BOOL parse_scheme_type(parse_data *data) {
1127 /* If there's scheme data then see if it's a recognized scheme. */
1128 if(data->scheme && data->scheme_len) {
1129 DWORD i;
1131 for(i = 0; i < sizeof(recognized_schemes)/sizeof(recognized_schemes[0]); ++i) {
1132 if(lstrlenW(recognized_schemes[i].scheme_name) == data->scheme_len) {
1133 /* Has to be a case insensitive compare. */
1134 if(!StrCmpNIW(recognized_schemes[i].scheme_name, data->scheme, data->scheme_len)) {
1135 data->scheme_type = recognized_schemes[i].scheme;
1136 return TRUE;
1141 /* If we get here it means it's not a recognized scheme. */
1142 data->scheme_type = URL_SCHEME_UNKNOWN;
1143 return TRUE;
1144 } else if(data->is_relative) {
1145 /* Relative URI's have no scheme. */
1146 data->scheme_type = URL_SCHEME_UNKNOWN;
1147 return TRUE;
1148 } else {
1149 /* Should never reach here! what happened... */
1150 FIXME("(%p): Unable to determine scheme type for URI %s\n", data, debugstr_w(data->uri));
1151 return FALSE;
1155 /* Tries to parse (or deduce) the scheme_name of a URI. If it can't
1156 * parse a scheme from the URI it will try to deduce the scheme_name and scheme_type
1157 * using the flags specified in 'flags' (if any). Flags that affect how this function
1158 * operates are the Uri_CREATE_ALLOW_* flags.
1160 * All parsed/deduced information will be stored in 'data' when the function returns.
1162 * Returns TRUE if it was able to successfully parse the information.
1164 static BOOL parse_scheme(const WCHAR **ptr, parse_data *data, DWORD flags, DWORD extras) {
1165 static const WCHAR fileW[] = {'f','i','l','e',0};
1166 static const WCHAR wildcardW[] = {'*',0};
1168 /* First check to see if the uri could implicitly be a file path. */
1169 if(is_implicit_file_path(*ptr)) {
1170 if(flags & Uri_CREATE_ALLOW_IMPLICIT_FILE_SCHEME) {
1171 data->scheme = fileW;
1172 data->scheme_len = lstrlenW(fileW);
1173 data->has_implicit_scheme = TRUE;
1175 TRACE("(%p %p %x): URI is an implicit file path.\n", ptr, data, flags);
1176 } else {
1177 /* Windows does not consider anything that can implicitly be a file
1178 * path to be a valid URI if the ALLOW_IMPLICIT_FILE_SCHEME flag is not set...
1180 TRACE("(%p %p %x): URI is implicitly a file path, but, the ALLOW_IMPLICIT_FILE_SCHEME flag wasn't set.\n",
1181 ptr, data, flags);
1182 return FALSE;
1184 } else if(!parse_scheme_name(ptr, data, extras)) {
1185 /* No scheme was found, this means it could be:
1186 * a) an implicit Wildcard scheme
1187 * b) a relative URI
1188 * c) an invalid URI.
1190 if(flags & Uri_CREATE_ALLOW_IMPLICIT_WILDCARD_SCHEME) {
1191 data->scheme = wildcardW;
1192 data->scheme_len = lstrlenW(wildcardW);
1193 data->has_implicit_scheme = TRUE;
1195 TRACE("(%p %p %x): URI is an implicit wildcard scheme.\n", ptr, data, flags);
1196 } else if (flags & Uri_CREATE_ALLOW_RELATIVE) {
1197 data->is_relative = TRUE;
1198 TRACE("(%p %p %x): URI is relative.\n", ptr, data, flags);
1199 } else {
1200 TRACE("(%p %p %x): Malformed URI found. Unable to deduce scheme name.\n", ptr, data, flags);
1201 return FALSE;
1205 if(!data->is_relative)
1206 TRACE("(%p %p %x): Found scheme=%s scheme_len=%d\n", ptr, data, flags,
1207 debugstr_wn(data->scheme, data->scheme_len), data->scheme_len);
1209 if(!parse_scheme_type(data))
1210 return FALSE;
1212 TRACE("(%p %p %x): Assigned %d as the URL_SCHEME.\n", ptr, data, flags, data->scheme_type);
1213 return TRUE;
1216 static BOOL parse_username(const WCHAR **ptr, parse_data *data, DWORD flags, DWORD extras) {
1217 data->username = *ptr;
1219 while(**ptr != ':' && **ptr != '@') {
1220 if(**ptr == '%') {
1221 if(!check_pct_encoded(ptr)) {
1222 if(data->scheme_type != URL_SCHEME_UNKNOWN) {
1223 *ptr = data->username;
1224 data->username = NULL;
1225 return FALSE;
1227 } else
1228 continue;
1229 } else if(extras & ALLOW_NULL_TERM_USER_NAME && !**ptr)
1230 break;
1231 else if(is_auth_delim(**ptr, data->scheme_type != URL_SCHEME_UNKNOWN)) {
1232 *ptr = data->username;
1233 data->username = NULL;
1234 return FALSE;
1237 ++(*ptr);
1240 data->username_len = *ptr - data->username;
1241 return TRUE;
1244 static BOOL parse_password(const WCHAR **ptr, parse_data *data, DWORD flags, DWORD extras) {
1245 data->password = *ptr;
1247 while(**ptr != '@') {
1248 if(**ptr == '%') {
1249 if(!check_pct_encoded(ptr)) {
1250 if(data->scheme_type != URL_SCHEME_UNKNOWN) {
1251 *ptr = data->password;
1252 data->password = NULL;
1253 return FALSE;
1255 } else
1256 continue;
1257 } else if(extras & ALLOW_NULL_TERM_PASSWORD && !**ptr)
1258 break;
1259 else if(is_auth_delim(**ptr, data->scheme_type != URL_SCHEME_UNKNOWN)) {
1260 *ptr = data->password;
1261 data->password = NULL;
1262 return FALSE;
1265 ++(*ptr);
1268 data->password_len = *ptr - data->password;
1269 return TRUE;
1272 /* Parses the userinfo part of the URI (if it exists). The userinfo field of
1273 * a URI can consist of "username:password@", or just "username@".
1275 * RFC def:
1276 * userinfo = *( unreserved / pct-encoded / sub-delims / ":" )
1278 * NOTES:
1279 * 1) If there is more than one ':' in the userinfo part of the URI Windows
1280 * uses the first occurrence of ':' to delimit the username and password
1281 * components.
1283 * ex:
1284 * ftp://user:pass:word@winehq.org
1286 * would yield "user" as the username and "pass:word" as the password.
1288 * 2) Windows allows any character to appear in the "userinfo" part of
1289 * a URI, as long as it's not an authority delimiter character set.
1291 static void parse_userinfo(const WCHAR **ptr, parse_data *data, DWORD flags) {
1292 const WCHAR *start = *ptr;
1294 if(!parse_username(ptr, data, flags, 0)) {
1295 TRACE("(%p %p %x): URI contained no userinfo.\n", ptr, data, flags);
1296 return;
1299 if(**ptr == ':') {
1300 ++(*ptr);
1301 if(!parse_password(ptr, data, flags, 0)) {
1302 *ptr = start;
1303 data->username = NULL;
1304 data->username_len = 0;
1305 TRACE("(%p %p %x): URI contained no userinfo.\n", ptr, data, flags);
1306 return;
1310 if(**ptr != '@') {
1311 *ptr = start;
1312 data->username = NULL;
1313 data->username_len = 0;
1314 data->password = NULL;
1315 data->password_len = 0;
1317 TRACE("(%p %p %x): URI contained no userinfo.\n", ptr, data, flags);
1318 return;
1321 if(data->username)
1322 TRACE("(%p %p %x): Found username %s len=%d.\n", ptr, data, flags,
1323 debugstr_wn(data->username, data->username_len), data->username_len);
1325 if(data->password)
1326 TRACE("(%p %p %x): Found password %s len=%d.\n", ptr, data, flags,
1327 debugstr_wn(data->password, data->password_len), data->password_len);
1329 ++(*ptr);
1332 /* Attempts to parse a port from the URI.
1334 * NOTES:
1335 * Windows seems to have a cap on what the maximum value
1336 * for a port can be. The max value is USHORT_MAX.
1338 * port = *DIGIT
1340 static BOOL parse_port(const WCHAR **ptr, parse_data *data, DWORD flags) {
1341 UINT port = 0;
1342 data->port = *ptr;
1344 while(!is_auth_delim(**ptr, data->scheme_type != URL_SCHEME_UNKNOWN)) {
1345 if(!is_num(**ptr)) {
1346 *ptr = data->port;
1347 data->port = NULL;
1348 return FALSE;
1351 port = port*10 + (**ptr-'0');
1353 if(port > USHRT_MAX) {
1354 *ptr = data->port;
1355 data->port = NULL;
1356 return FALSE;
1359 ++(*ptr);
1362 data->has_port = TRUE;
1363 data->port_value = port;
1364 data->port_len = *ptr - data->port;
1366 TRACE("(%p %p %x): Found port %s len=%d value=%u\n", ptr, data, flags,
1367 debugstr_wn(data->port, data->port_len), data->port_len, data->port_value);
1368 return TRUE;
1371 /* Attempts to parse a IPv4 address from the URI.
1373 * NOTES:
1374 * Windows normalizes IPv4 addresses, This means there are three
1375 * possibilities for the URI to contain an IPv4 address.
1376 * 1) A well formed address (ex. 192.2.2.2).
1377 * 2) A partially formed address. For example "192.0" would
1378 * normalize to "192.0.0.0" during canonicalization.
1379 * 3) An implicit IPv4 address. For example "256" would
1380 * normalize to "0.0.1.0" during canonicalization. Also
1381 * note that the maximum value for an implicit IP address
1382 * is UINT_MAX, if the value in the URI exceeds this then
1383 * it is not considered an IPv4 address.
1385 static BOOL parse_ipv4address(const WCHAR **ptr, parse_data *data, DWORD flags) {
1386 const BOOL is_unknown = data->scheme_type == URL_SCHEME_UNKNOWN;
1387 data->host = *ptr;
1389 if(!check_ipv4address(ptr, FALSE)) {
1390 if(!check_implicit_ipv4(ptr, &data->implicit_ipv4)) {
1391 TRACE("(%p %p %x): URI didn't contain anything looking like an IPv4 address.\n",
1392 ptr, data, flags);
1393 *ptr = data->host;
1394 data->host = NULL;
1395 return FALSE;
1396 } else
1397 data->has_implicit_ip = TRUE;
1400 data->host_len = *ptr - data->host;
1401 data->host_type = Uri_HOST_IPV4;
1403 /* Check if what we found is the only part of the host name (if it isn't
1404 * we don't have an IPv4 address).
1406 if(**ptr == ':') {
1407 ++(*ptr);
1408 if(!parse_port(ptr, data, flags)) {
1409 *ptr = data->host;
1410 data->host = NULL;
1411 return FALSE;
1413 } else if(!is_auth_delim(**ptr, !is_unknown)) {
1414 /* Found more data which belongs to the host, so this isn't an IPv4. */
1415 *ptr = data->host;
1416 data->host = NULL;
1417 data->has_implicit_ip = FALSE;
1418 return FALSE;
1421 TRACE("(%p %p %x): IPv4 address found. host=%s host_len=%d host_type=%d\n",
1422 ptr, data, flags, debugstr_wn(data->host, data->host_len),
1423 data->host_len, data->host_type);
1424 return TRUE;
1427 /* Attempts to parse the reg-name from the URI.
1429 * Because of the way Windows handles ':' this function also
1430 * handles parsing the port.
1432 * reg-name = *( unreserved / pct-encoded / sub-delims )
1434 * NOTE:
1435 * Windows allows everything, but, the characters in "auth_delims" and ':'
1436 * to appear in a reg-name, unless it's an unknown scheme type then ':' is
1437 * allowed to appear (even if a valid port isn't after it).
1439 * Windows doesn't like host names which start with '[' and end with ']'
1440 * and don't contain a valid IP literal address in between them.
1442 * On Windows if a '[' is encountered in the host name the ':' no longer
1443 * counts as a delimiter until you reach the next ']' or an "authority delimiter".
1445 * A reg-name CAN be empty.
1447 static BOOL parse_reg_name(const WCHAR **ptr, parse_data *data, DWORD flags, DWORD extras) {
1448 const BOOL has_start_bracket = **ptr == '[';
1449 const BOOL known_scheme = data->scheme_type != URL_SCHEME_UNKNOWN;
1450 const BOOL is_res = data->scheme_type == URL_SCHEME_RES;
1451 BOOL inside_brackets = has_start_bracket;
1453 /* res URIs don't have ports. */
1454 BOOL ignore_col = (extras & IGNORE_PORT_DELIMITER) || is_res;
1456 /* We have to be careful with file schemes. */
1457 if(data->scheme_type == URL_SCHEME_FILE) {
1458 /* This is because an implicit file scheme could be "C:\\test" and it
1459 * would trick this function into thinking the host is "C", when after
1460 * canonicalization the host would end up being an empty string. A drive
1461 * path can also have a '|' instead of a ':' after the drive letter.
1463 if(is_drive_path(*ptr)) {
1464 /* Regular old drive paths have no host type (or host name). */
1465 data->host_type = Uri_HOST_UNKNOWN;
1466 data->host = *ptr;
1467 data->host_len = 0;
1468 return TRUE;
1469 } else if(is_unc_path(*ptr))
1470 /* Skip past the "\\" of a UNC path. */
1471 *ptr += 2;
1474 data->host = *ptr;
1476 /* For res URIs, everything before the first '/' is
1477 * considered the host.
1479 while((!is_res && !is_auth_delim(**ptr, known_scheme)) ||
1480 (is_res && **ptr && **ptr != '/')) {
1481 if(**ptr == ':' && !ignore_col) {
1482 /* We can ignore ':' if were inside brackets.*/
1483 if(!inside_brackets) {
1484 const WCHAR *tmp = (*ptr)++;
1486 /* Attempt to parse the port. */
1487 if(!parse_port(ptr, data, flags)) {
1488 /* Windows expects there to be a valid port for known scheme types. */
1489 if(data->scheme_type != URL_SCHEME_UNKNOWN) {
1490 *ptr = data->host;
1491 data->host = NULL;
1492 TRACE("(%p %p %x %x): Expected valid port\n", ptr, data, flags, extras);
1493 return FALSE;
1494 } else
1495 /* Windows gives up on trying to parse a port when it
1496 * encounters an invalid port.
1498 ignore_col = TRUE;
1499 } else {
1500 data->host_len = tmp - data->host;
1501 break;
1504 } else if(**ptr == '%' && (known_scheme && !is_res)) {
1505 /* Has to be a legit % encoded value. */
1506 if(!check_pct_encoded(ptr)) {
1507 *ptr = data->host;
1508 data->host = NULL;
1509 return FALSE;
1510 } else
1511 continue;
1512 } else if(is_res && is_forbidden_dos_path_char(**ptr)) {
1513 *ptr = data->host;
1514 data->host = NULL;
1515 return FALSE;
1516 } else if(**ptr == ']')
1517 inside_brackets = FALSE;
1518 else if(**ptr == '[')
1519 inside_brackets = TRUE;
1521 ++(*ptr);
1524 if(has_start_bracket) {
1525 /* Make sure the last character of the host wasn't a ']'. */
1526 if(*(*ptr-1) == ']') {
1527 TRACE("(%p %p %x %x): Expected an IP literal inside of the host\n",
1528 ptr, data, flags, extras);
1529 *ptr = data->host;
1530 data->host = NULL;
1531 return FALSE;
1535 /* Don't overwrite our length if we found a port earlier. */
1536 if(!data->port)
1537 data->host_len = *ptr - data->host;
1539 /* If the host is empty, then it's an unknown host type. */
1540 if(data->host_len == 0 || is_res)
1541 data->host_type = Uri_HOST_UNKNOWN;
1542 else
1543 data->host_type = Uri_HOST_DNS;
1545 TRACE("(%p %p %x %x): Parsed reg-name. host=%s len=%d\n", ptr, data, flags, extras,
1546 debugstr_wn(data->host, data->host_len), data->host_len);
1547 return TRUE;
1550 /* Attempts to parse an IPv6 address out of the URI.
1552 * IPv6address = 6( h16 ":" ) ls32
1553 * / "::" 5( h16 ":" ) ls32
1554 * / [ h16 ] "::" 4( h16 ":" ) ls32
1555 * / [ *1( h16 ":" ) h16 ] "::" 3( h16 ":" ) ls32
1556 * / [ *2( h16 ":" ) h16 ] "::" 2( h16 ":" ) ls32
1557 * / [ *3( h16 ":" ) h16 ] "::" h16 ":" ls32
1558 * / [ *4( h16 ":" ) h16 ] "::" ls32
1559 * / [ *5( h16 ":" ) h16 ] "::" h16
1560 * / [ *6( h16 ":" ) h16 ] "::"
1562 * ls32 = ( h16 ":" h16 ) / IPv4address
1563 * ; least-significant 32 bits of address.
1565 * h16 = 1*4HEXDIG
1566 * ; 16 bits of address represented in hexadecimal.
1568 * Modeled after google-url's 'DoParseIPv6' function.
1570 static BOOL parse_ipv6address(const WCHAR **ptr, parse_data *data, DWORD flags) {
1571 const WCHAR *start, *cur_start;
1572 ipv6_address ip;
1574 start = cur_start = *ptr;
1575 memset(&ip, 0, sizeof(ipv6_address));
1577 for(;; ++(*ptr)) {
1578 /* Check if we're on the last character of the host. */
1579 BOOL is_end = (is_auth_delim(**ptr, data->scheme_type != URL_SCHEME_UNKNOWN)
1580 || **ptr == ']');
1582 BOOL is_split = (**ptr == ':');
1583 BOOL is_elision = (is_split && !is_end && *(*ptr+1) == ':');
1585 /* Check if we're at the end of a component, or
1586 * if we're at the end of the IPv6 address.
1588 if(is_split || is_end) {
1589 DWORD cur_len = 0;
1591 cur_len = *ptr - cur_start;
1593 /* h16 can't have a length > 4. */
1594 if(cur_len > 4) {
1595 *ptr = start;
1597 TRACE("(%p %p %x): h16 component to long.\n",
1598 ptr, data, flags);
1599 return FALSE;
1602 if(cur_len == 0) {
1603 /* An h16 component can't have the length of 0 unless
1604 * the elision is at the beginning of the address, or
1605 * at the end of the address.
1607 if(!((*ptr == start && is_elision) ||
1608 (is_end && (*ptr-2) == ip.elision))) {
1609 *ptr = start;
1610 TRACE("(%p %p %x): IPv6 component cannot have a length of 0.\n",
1611 ptr, data, flags);
1612 return FALSE;
1616 if(cur_len > 0) {
1617 /* An IPv6 address can have no more than 8 h16 components. */
1618 if(ip.h16_count >= 8) {
1619 *ptr = start;
1620 TRACE("(%p %p %x): Not a IPv6 address, to many h16 components.\n",
1621 ptr, data, flags);
1622 return FALSE;
1625 ip.components[ip.h16_count].str = cur_start;
1626 ip.components[ip.h16_count].len = cur_len;
1628 TRACE("(%p %p %x): Found h16 component %s, len=%d, h16_count=%d\n",
1629 ptr, data, flags, debugstr_wn(cur_start, cur_len), cur_len,
1630 ip.h16_count);
1631 ++ip.h16_count;
1635 if(is_end)
1636 break;
1638 if(is_elision) {
1639 /* A IPv6 address can only have 1 elision ('::'). */
1640 if(ip.elision) {
1641 *ptr = start;
1643 TRACE("(%p %p %x): IPv6 address cannot have 2 elisions.\n",
1644 ptr, data, flags);
1645 return FALSE;
1648 ip.elision = *ptr;
1649 ++(*ptr);
1652 if(is_split)
1653 cur_start = *ptr+1;
1654 else {
1655 if(!check_ipv4address(ptr, TRUE)) {
1656 if(!is_hexdigit(**ptr)) {
1657 /* Not a valid character for an IPv6 address. */
1658 *ptr = start;
1659 return FALSE;
1661 } else {
1662 /* Found an IPv4 address. */
1663 ip.ipv4 = cur_start;
1664 ip.ipv4_len = *ptr - cur_start;
1666 TRACE("(%p %p %x): Found an attached IPv4 address %s len=%d.\n",
1667 ptr, data, flags, debugstr_wn(ip.ipv4, ip.ipv4_len),
1668 ip.ipv4_len);
1670 /* IPv4 addresses can only appear at the end of a IPv6. */
1671 break;
1676 compute_ipv6_comps_size(&ip);
1678 /* Make sure the IPv6 address adds up to 16 bytes. */
1679 if(ip.components_size + ip.elision_size != 16) {
1680 *ptr = start;
1681 TRACE("(%p %p %x): Invalid IPv6 address, did not add up to 16 bytes.\n",
1682 ptr, data, flags);
1683 return FALSE;
1686 if(ip.elision_size == 2) {
1687 /* For some reason on Windows if an elision that represents
1688 * only one h16 component is encountered at the very begin or
1689 * end of an IPv6 address, Windows does not consider it a
1690 * valid IPv6 address.
1692 * Ex: [::2:3:4:5:6:7] is not valid, even though the sum
1693 * of all the components == 128bits.
1695 if(ip.elision < ip.components[0].str ||
1696 ip.elision > ip.components[ip.h16_count-1].str) {
1697 *ptr = start;
1698 TRACE("(%p %p %x): Invalid IPv6 address. Detected elision of 2 bytes at the beginning or end of the address.\n",
1699 ptr, data, flags);
1700 return FALSE;
1704 data->host_type = Uri_HOST_IPV6;
1705 data->has_ipv6 = TRUE;
1706 data->ipv6_address = ip;
1708 TRACE("(%p %p %x): Found valid IPv6 literal %s len=%d\n",
1709 ptr, data, flags, debugstr_wn(start, *ptr-start),
1710 (int)(*ptr-start));
1711 return TRUE;
1714 /* IPvFuture = "v" 1*HEXDIG "." 1*( unreserved / sub-delims / ":" ) */
1715 static BOOL parse_ipvfuture(const WCHAR **ptr, parse_data *data, DWORD flags) {
1716 const WCHAR *start = *ptr;
1718 /* IPvFuture has to start with a 'v' or 'V'. */
1719 if(**ptr != 'v' && **ptr != 'V')
1720 return FALSE;
1722 /* Following the v there must be at least 1 hex digit. */
1723 ++(*ptr);
1724 if(!is_hexdigit(**ptr)) {
1725 *ptr = start;
1726 return FALSE;
1729 ++(*ptr);
1730 while(is_hexdigit(**ptr))
1731 ++(*ptr);
1733 /* End of the hexdigit sequence must be a '.' */
1734 if(**ptr != '.') {
1735 *ptr = start;
1736 return FALSE;
1739 ++(*ptr);
1740 if(!is_unreserved(**ptr) && !is_subdelim(**ptr) && **ptr != ':') {
1741 *ptr = start;
1742 return FALSE;
1745 ++(*ptr);
1746 while(is_unreserved(**ptr) || is_subdelim(**ptr) || **ptr == ':')
1747 ++(*ptr);
1749 data->host_type = Uri_HOST_UNKNOWN;
1751 TRACE("(%p %p %x): Parsed IPvFuture address %s len=%d\n", ptr, data, flags,
1752 debugstr_wn(start, *ptr-start), (int)(*ptr-start));
1754 return TRUE;
1757 /* IP-literal = "[" ( IPv6address / IPvFuture ) "]" */
1758 static BOOL parse_ip_literal(const WCHAR **ptr, parse_data *data, DWORD flags, DWORD extras) {
1759 data->host = *ptr;
1761 if(**ptr != '[' && !(extras & ALLOW_BRACKETLESS_IP_LITERAL)) {
1762 data->host = NULL;
1763 return FALSE;
1764 } else if(**ptr == '[')
1765 ++(*ptr);
1767 if(!parse_ipv6address(ptr, data, flags)) {
1768 if(extras & SKIP_IP_FUTURE_CHECK || !parse_ipvfuture(ptr, data, flags)) {
1769 *ptr = data->host;
1770 data->host = NULL;
1771 return FALSE;
1775 if(**ptr != ']' && !(extras & ALLOW_BRACKETLESS_IP_LITERAL)) {
1776 *ptr = data->host;
1777 data->host = NULL;
1778 return FALSE;
1779 } else if(!**ptr && extras & ALLOW_BRACKETLESS_IP_LITERAL) {
1780 /* The IP literal didn't contain brackets and was followed by
1781 * a NULL terminator, so no reason to even check the port.
1783 data->host_len = *ptr - data->host;
1784 return TRUE;
1787 ++(*ptr);
1788 if(**ptr == ':') {
1789 ++(*ptr);
1790 /* If a valid port is not found, then let it trickle down to
1791 * parse_reg_name.
1793 if(!parse_port(ptr, data, flags)) {
1794 *ptr = data->host;
1795 data->host = NULL;
1796 return FALSE;
1798 } else
1799 data->host_len = *ptr - data->host;
1801 return TRUE;
1804 /* Parses the host information from the URI.
1806 * host = IP-literal / IPv4address / reg-name
1808 static BOOL parse_host(const WCHAR **ptr, parse_data *data, DWORD flags, DWORD extras) {
1809 if(!parse_ip_literal(ptr, data, flags, extras)) {
1810 if(!parse_ipv4address(ptr, data, flags)) {
1811 if(!parse_reg_name(ptr, data, flags, extras)) {
1812 TRACE("(%p %p %x %x): Malformed URI, Unknown host type.\n",
1813 ptr, data, flags, extras);
1814 return FALSE;
1819 return TRUE;
1822 /* Parses the authority information from the URI.
1824 * authority = [ userinfo "@" ] host [ ":" port ]
1826 static BOOL parse_authority(const WCHAR **ptr, parse_data *data, DWORD flags) {
1827 parse_userinfo(ptr, data, flags);
1829 /* Parsing the port will happen during one of the host parsing
1830 * routines (if the URI has a port).
1832 if(!parse_host(ptr, data, flags, 0))
1833 return FALSE;
1835 return TRUE;
1838 /* Attempts to parse the path information of a hierarchical URI. */
1839 static BOOL parse_path_hierarchical(const WCHAR **ptr, parse_data *data, DWORD flags) {
1840 const WCHAR *start = *ptr;
1841 static const WCHAR slash[] = {'/',0};
1842 const BOOL is_file = data->scheme_type == URL_SCHEME_FILE;
1844 if(is_path_delim(**ptr)) {
1845 if(data->scheme_type == URL_SCHEME_WILDCARD && !data->must_have_path) {
1846 data->path = NULL;
1847 data->path_len = 0;
1848 } else if(!(flags & Uri_CREATE_NO_CANONICALIZE)) {
1849 /* If the path component is empty, then a '/' is added. */
1850 data->path = slash;
1851 data->path_len = 1;
1853 } else {
1854 while(!is_path_delim(**ptr)) {
1855 if(**ptr == '%' && data->scheme_type != URL_SCHEME_UNKNOWN && !is_file) {
1856 if(!check_pct_encoded(ptr)) {
1857 *ptr = start;
1858 return FALSE;
1859 } else
1860 continue;
1861 } else if(is_forbidden_dos_path_char(**ptr) && is_file &&
1862 (flags & Uri_CREATE_FILE_USE_DOS_PATH)) {
1863 /* File schemes with USE_DOS_PATH set aren't allowed to have
1864 * a '<' or '>' or '\"' appear in them.
1866 *ptr = start;
1867 return FALSE;
1868 } else if(**ptr == '\\') {
1869 /* Not allowed to have a backslash if NO_CANONICALIZE is set
1870 * and the scheme is known type (but not a file scheme).
1872 if(flags & Uri_CREATE_NO_CANONICALIZE) {
1873 if(data->scheme_type != URL_SCHEME_FILE &&
1874 data->scheme_type != URL_SCHEME_UNKNOWN) {
1875 *ptr = start;
1876 return FALSE;
1881 ++(*ptr);
1884 /* The only time a URI doesn't have a path is when
1885 * the NO_CANONICALIZE flag is set and the raw URI
1886 * didn't contain one.
1888 if(*ptr == start) {
1889 data->path = NULL;
1890 data->path_len = 0;
1891 } else {
1892 data->path = start;
1893 data->path_len = *ptr - start;
1897 if(data->path)
1898 TRACE("(%p %p %x): Parsed path %s len=%d\n", ptr, data, flags,
1899 debugstr_wn(data->path, data->path_len), data->path_len);
1900 else
1901 TRACE("(%p %p %x): The URI contained no path\n", ptr, data, flags);
1903 return TRUE;
1906 /* Parses the path of an opaque URI (much less strict then the parser
1907 * for a hierarchical URI).
1909 * NOTE:
1910 * Windows allows invalid % encoded data to appear in opaque URI paths
1911 * for unknown scheme types.
1913 * File schemes with USE_DOS_PATH set aren't allowed to have '<', '>', or '\"'
1914 * appear in them.
1916 static BOOL parse_path_opaque(const WCHAR **ptr, parse_data *data, DWORD flags) {
1917 const BOOL known_scheme = data->scheme_type != URL_SCHEME_UNKNOWN;
1918 const BOOL is_file = data->scheme_type == URL_SCHEME_FILE;
1920 data->path = *ptr;
1922 while(!is_path_delim(**ptr)) {
1923 if(**ptr == '%' && known_scheme) {
1924 if(!check_pct_encoded(ptr)) {
1925 *ptr = data->path;
1926 data->path = NULL;
1927 return FALSE;
1928 } else
1929 continue;
1930 } else if(is_forbidden_dos_path_char(**ptr) && is_file &&
1931 (flags & Uri_CREATE_FILE_USE_DOS_PATH)) {
1932 *ptr = data->path;
1933 data->path = NULL;
1934 return FALSE;
1937 ++(*ptr);
1940 data->path_len = *ptr - data->path;
1941 TRACE("(%p %p %x): Parsed opaque URI path %s len=%d\n", ptr, data, flags,
1942 debugstr_wn(data->path, data->path_len), data->path_len);
1943 return TRUE;
1946 /* Determines how the URI should be parsed after the scheme information.
1948 * If the scheme is followed by "//", then it is treated as a hierarchical URI
1949 * which then the authority and path information will be parsed out. Otherwise, the
1950 * URI will be treated as an opaque URI which the authority information is not parsed
1951 * out.
1953 * RFC 3896 definition of hier-part:
1955 * hier-part = "//" authority path-abempty
1956 * / path-absolute
1957 * / path-rootless
1958 * / path-empty
1960 * MSDN opaque URI definition:
1961 * scheme ":" path [ "#" fragment ]
1963 * NOTES:
1964 * If the URI is of an unknown scheme type and has a "//" following the scheme then it
1965 * is treated as a hierarchical URI, but, if the CREATE_NO_CRACK_UNKNOWN_SCHEMES flag is
1966 * set then it is considered an opaque URI regardless of what follows the scheme information
1967 * (per MSDN documentation).
1969 static BOOL parse_hierpart(const WCHAR **ptr, parse_data *data, DWORD flags) {
1970 const WCHAR *start = *ptr;
1972 data->must_have_path = FALSE;
1974 /* For javascript: URIs, simply set everything as a path */
1975 if(data->scheme_type == URL_SCHEME_JAVASCRIPT) {
1976 data->path = *ptr;
1977 data->path_len = strlenW(*ptr);
1978 data->is_opaque = TRUE;
1979 *ptr += data->path_len;
1980 return TRUE;
1983 /* Checks if the authority information needs to be parsed. */
1984 if(is_hierarchical_uri(ptr, data)) {
1985 /* Only treat it as a hierarchical URI if the scheme_type is known or
1986 * the Uri_CREATE_NO_CRACK_UNKNOWN_SCHEMES flag is not set.
1988 if(data->scheme_type != URL_SCHEME_UNKNOWN ||
1989 !(flags & Uri_CREATE_NO_CRACK_UNKNOWN_SCHEMES)) {
1990 TRACE("(%p %p %x): Treating URI as an hierarchical URI.\n", ptr, data, flags);
1991 data->is_opaque = FALSE;
1993 if(data->scheme_type == URL_SCHEME_WILDCARD && !data->has_implicit_scheme) {
1994 if(**ptr == '/' && *(*ptr+1) == '/') {
1995 data->must_have_path = TRUE;
1996 *ptr += 2;
2000 /* TODO: Handle hierarchical URI's, parse authority then parse the path. */
2001 if(!parse_authority(ptr, data, flags))
2002 return FALSE;
2004 return parse_path_hierarchical(ptr, data, flags);
2005 } else
2006 /* Reset ptr to its starting position so opaque path parsing
2007 * begins at the correct location.
2009 *ptr = start;
2012 /* If it reaches here, then the URI will be treated as an opaque
2013 * URI.
2016 TRACE("(%p %p %x): Treating URI as an opaque URI.\n", ptr, data, flags);
2018 data->is_opaque = TRUE;
2019 if(!parse_path_opaque(ptr, data, flags))
2020 return FALSE;
2022 return TRUE;
2025 /* Attempts to parse the query string from the URI.
2027 * NOTES:
2028 * If NO_DECODE_EXTRA_INFO flag is set, then invalid percent encoded
2029 * data is allowed to appear in the query string. For unknown scheme types
2030 * invalid percent encoded data is allowed to appear regardless.
2032 static BOOL parse_query(const WCHAR **ptr, parse_data *data, DWORD flags) {
2033 const BOOL known_scheme = data->scheme_type != URL_SCHEME_UNKNOWN;
2035 if(**ptr != '?') {
2036 TRACE("(%p %p %x): URI didn't contain a query string.\n", ptr, data, flags);
2037 return TRUE;
2040 data->query = *ptr;
2042 ++(*ptr);
2043 while(**ptr && **ptr != '#') {
2044 if(**ptr == '%' && known_scheme &&
2045 !(flags & Uri_CREATE_NO_DECODE_EXTRA_INFO)) {
2046 if(!check_pct_encoded(ptr)) {
2047 *ptr = data->query;
2048 data->query = NULL;
2049 return FALSE;
2050 } else
2051 continue;
2054 ++(*ptr);
2057 data->query_len = *ptr - data->query;
2059 TRACE("(%p %p %x): Parsed query string %s len=%d\n", ptr, data, flags,
2060 debugstr_wn(data->query, data->query_len), data->query_len);
2061 return TRUE;
2064 /* Attempts to parse the fragment from the URI.
2066 * NOTES:
2067 * If NO_DECODE_EXTRA_INFO flag is set, then invalid percent encoded
2068 * data is allowed to appear in the query string. For unknown scheme types
2069 * invalid percent encoded data is allowed to appear regardless.
2071 static BOOL parse_fragment(const WCHAR **ptr, parse_data *data, DWORD flags) {
2072 const BOOL known_scheme = data->scheme_type != URL_SCHEME_UNKNOWN;
2074 if(**ptr != '#') {
2075 TRACE("(%p %p %x): URI didn't contain a fragment.\n", ptr, data, flags);
2076 return TRUE;
2079 data->fragment = *ptr;
2081 ++(*ptr);
2082 while(**ptr) {
2083 if(**ptr == '%' && known_scheme &&
2084 !(flags & Uri_CREATE_NO_DECODE_EXTRA_INFO)) {
2085 if(!check_pct_encoded(ptr)) {
2086 *ptr = data->fragment;
2087 data->fragment = NULL;
2088 return FALSE;
2089 } else
2090 continue;
2093 ++(*ptr);
2096 data->fragment_len = *ptr - data->fragment;
2098 TRACE("(%p %p %x): Parsed fragment %s len=%d\n", ptr, data, flags,
2099 debugstr_wn(data->fragment, data->fragment_len), data->fragment_len);
2100 return TRUE;
2103 /* Parses and validates the components of the specified by data->uri
2104 * and stores the information it parses into 'data'.
2106 * Returns TRUE if it successfully parsed the URI. False otherwise.
2108 static BOOL parse_uri(parse_data *data, DWORD flags) {
2109 const WCHAR *ptr;
2110 const WCHAR **pptr;
2112 ptr = data->uri;
2113 pptr = &ptr;
2115 TRACE("(%p %x): BEGINNING TO PARSE URI %s.\n", data, flags, debugstr_w(data->uri));
2117 if(!parse_scheme(pptr, data, flags, 0))
2118 return FALSE;
2120 if(!parse_hierpart(pptr, data, flags))
2121 return FALSE;
2123 if(!parse_query(pptr, data, flags))
2124 return FALSE;
2126 if(!parse_fragment(pptr, data, flags))
2127 return FALSE;
2129 TRACE("(%p %x): FINISHED PARSING URI.\n", data, flags);
2130 return TRUE;
2133 static BOOL canonicalize_username(const parse_data *data, Uri *uri, DWORD flags, BOOL computeOnly) {
2134 const WCHAR *ptr;
2136 if(!data->username) {
2137 uri->userinfo_start = -1;
2138 return TRUE;
2141 uri->userinfo_start = uri->canon_len;
2142 for(ptr = data->username; ptr < data->username+data->username_len; ++ptr) {
2143 if(*ptr == '%') {
2144 /* Only decode % encoded values for known scheme types. */
2145 if(data->scheme_type != URL_SCHEME_UNKNOWN) {
2146 /* See if the value really needs decoding. */
2147 WCHAR val = decode_pct_val(ptr);
2148 if(is_unreserved(val)) {
2149 if(!computeOnly)
2150 uri->canon_uri[uri->canon_len] = val;
2152 ++uri->canon_len;
2154 /* Move pass the hex characters. */
2155 ptr += 2;
2156 continue;
2159 } else if(!is_reserved(*ptr) && !is_unreserved(*ptr) && *ptr != '\\') {
2160 /* Only percent encode forbidden characters if the NO_ENCODE_FORBIDDEN_CHARACTERS flag
2161 * is NOT set.
2163 if(!(flags & Uri_CREATE_NO_ENCODE_FORBIDDEN_CHARACTERS)) {
2164 if(!computeOnly)
2165 pct_encode_val(*ptr, uri->canon_uri + uri->canon_len);
2167 uri->canon_len += 3;
2168 continue;
2172 if(!computeOnly)
2173 /* Nothing special, so just copy the character over. */
2174 uri->canon_uri[uri->canon_len] = *ptr;
2175 ++uri->canon_len;
2178 return TRUE;
2181 static BOOL canonicalize_password(const parse_data *data, Uri *uri, DWORD flags, BOOL computeOnly) {
2182 const WCHAR *ptr;
2184 if(!data->password) {
2185 uri->userinfo_split = -1;
2186 return TRUE;
2189 if(uri->userinfo_start == -1)
2190 /* Has a password, but, doesn't have a username. */
2191 uri->userinfo_start = uri->canon_len;
2193 uri->userinfo_split = uri->canon_len - uri->userinfo_start;
2195 /* Add the ':' to the userinfo component. */
2196 if(!computeOnly)
2197 uri->canon_uri[uri->canon_len] = ':';
2198 ++uri->canon_len;
2200 for(ptr = data->password; ptr < data->password+data->password_len; ++ptr) {
2201 if(*ptr == '%') {
2202 /* Only decode % encoded values for known scheme types. */
2203 if(data->scheme_type != URL_SCHEME_UNKNOWN) {
2204 /* See if the value really needs decoding. */
2205 WCHAR val = decode_pct_val(ptr);
2206 if(is_unreserved(val)) {
2207 if(!computeOnly)
2208 uri->canon_uri[uri->canon_len] = val;
2210 ++uri->canon_len;
2212 /* Move pass the hex characters. */
2213 ptr += 2;
2214 continue;
2217 } else if(!is_reserved(*ptr) && !is_unreserved(*ptr) && *ptr != '\\') {
2218 /* Only percent encode forbidden characters if the NO_ENCODE_FORBIDDEN_CHARACTERS flag
2219 * is NOT set.
2221 if(!(flags & Uri_CREATE_NO_ENCODE_FORBIDDEN_CHARACTERS)) {
2222 if(!computeOnly)
2223 pct_encode_val(*ptr, uri->canon_uri + uri->canon_len);
2225 uri->canon_len += 3;
2226 continue;
2230 if(!computeOnly)
2231 /* Nothing special, so just copy the character over. */
2232 uri->canon_uri[uri->canon_len] = *ptr;
2233 ++uri->canon_len;
2236 return TRUE;
2239 /* Canonicalizes the userinfo of the URI represented by the parse_data.
2241 * Canonicalization of the userinfo is a simple process. If there are any percent
2242 * encoded characters that fall in the "unreserved" character set, they are decoded
2243 * to their actual value. If a character is not in the "unreserved" or "reserved" sets
2244 * then it is percent encoded. Other than that the characters are copied over without
2245 * change.
2247 static BOOL canonicalize_userinfo(const parse_data *data, Uri *uri, DWORD flags, BOOL computeOnly) {
2248 uri->userinfo_start = uri->userinfo_split = -1;
2249 uri->userinfo_len = 0;
2251 if(!data->username && !data->password)
2252 /* URI doesn't have userinfo, so nothing to do here. */
2253 return TRUE;
2255 if(!canonicalize_username(data, uri, flags, computeOnly))
2256 return FALSE;
2258 if(!canonicalize_password(data, uri, flags, computeOnly))
2259 return FALSE;
2261 uri->userinfo_len = uri->canon_len - uri->userinfo_start;
2262 if(!computeOnly)
2263 TRACE("(%p %p %x %d): Canonicalized userinfo, userinfo_start=%d, userinfo=%s, userinfo_split=%d userinfo_len=%d.\n",
2264 data, uri, flags, computeOnly, uri->userinfo_start, debugstr_wn(uri->canon_uri + uri->userinfo_start, uri->userinfo_len),
2265 uri->userinfo_split, uri->userinfo_len);
2267 /* Now insert the '@' after the userinfo. */
2268 if(!computeOnly)
2269 uri->canon_uri[uri->canon_len] = '@';
2270 ++uri->canon_len;
2272 return TRUE;
2275 /* Attempts to canonicalize a reg_name.
2277 * Things that happen:
2278 * 1) If Uri_CREATE_NO_CANONICALIZE flag is not set, then the reg_name is
2279 * lower cased. Unless it's an unknown scheme type, which case it's
2280 * no lower cased regardless.
2282 * 2) Unreserved % encoded characters are decoded for known
2283 * scheme types.
2285 * 3) Forbidden characters are % encoded as long as
2286 * Uri_CREATE_NO_ENCODE_FORBIDDEN_CHARACTERS flag is not set and
2287 * it isn't an unknown scheme type.
2289 * 4) If it's a file scheme and the host is "localhost" it's removed.
2291 * 5) If it's a file scheme and Uri_CREATE_FILE_USE_DOS_PATH is set,
2292 * then the UNC path characters are added before the host name.
2294 static BOOL canonicalize_reg_name(const parse_data *data, Uri *uri,
2295 DWORD flags, BOOL computeOnly) {
2296 static const WCHAR localhostW[] =
2297 {'l','o','c','a','l','h','o','s','t',0};
2298 const WCHAR *ptr;
2299 const BOOL known_scheme = data->scheme_type != URL_SCHEME_UNKNOWN;
2301 if(data->scheme_type == URL_SCHEME_FILE &&
2302 data->host_len == lstrlenW(localhostW)) {
2303 if(!StrCmpNIW(data->host, localhostW, data->host_len)) {
2304 uri->host_start = -1;
2305 uri->host_len = 0;
2306 uri->host_type = Uri_HOST_UNKNOWN;
2307 return TRUE;
2311 if(data->scheme_type == URL_SCHEME_FILE && flags & Uri_CREATE_FILE_USE_DOS_PATH) {
2312 if(!computeOnly) {
2313 uri->canon_uri[uri->canon_len] = '\\';
2314 uri->canon_uri[uri->canon_len+1] = '\\';
2316 uri->canon_len += 2;
2317 uri->authority_start = uri->canon_len;
2320 uri->host_start = uri->canon_len;
2322 for(ptr = data->host; ptr < data->host+data->host_len; ++ptr) {
2323 if(*ptr == '%' && known_scheme) {
2324 WCHAR val = decode_pct_val(ptr);
2325 if(is_unreserved(val)) {
2326 /* If NO_CANONICALIZE is not set, then windows lower cases the
2327 * decoded value.
2329 if(!(flags & Uri_CREATE_NO_CANONICALIZE) && isupperW(val)) {
2330 if(!computeOnly)
2331 uri->canon_uri[uri->canon_len] = tolowerW(val);
2332 } else {
2333 if(!computeOnly)
2334 uri->canon_uri[uri->canon_len] = val;
2336 ++uri->canon_len;
2338 /* Skip past the % encoded character. */
2339 ptr += 2;
2340 continue;
2341 } else {
2342 /* Just copy the % over. */
2343 if(!computeOnly)
2344 uri->canon_uri[uri->canon_len] = *ptr;
2345 ++uri->canon_len;
2347 } else if(*ptr == '\\') {
2348 /* Only unknown scheme types could have made it here with a '\\' in the host name. */
2349 if(!computeOnly)
2350 uri->canon_uri[uri->canon_len] = *ptr;
2351 ++uri->canon_len;
2352 } else if(!(flags & Uri_CREATE_NO_ENCODE_FORBIDDEN_CHARACTERS) &&
2353 !is_unreserved(*ptr) && !is_reserved(*ptr) && known_scheme) {
2354 if(!computeOnly) {
2355 pct_encode_val(*ptr, uri->canon_uri+uri->canon_len);
2357 /* The percent encoded value gets lower cased also. */
2358 if(!(flags & Uri_CREATE_NO_CANONICALIZE)) {
2359 uri->canon_uri[uri->canon_len+1] = tolowerW(uri->canon_uri[uri->canon_len+1]);
2360 uri->canon_uri[uri->canon_len+2] = tolowerW(uri->canon_uri[uri->canon_len+2]);
2364 uri->canon_len += 3;
2365 } else {
2366 if(!computeOnly) {
2367 if(!(flags & Uri_CREATE_NO_CANONICALIZE) && known_scheme)
2368 uri->canon_uri[uri->canon_len] = tolowerW(*ptr);
2369 else
2370 uri->canon_uri[uri->canon_len] = *ptr;
2373 ++uri->canon_len;
2377 uri->host_len = uri->canon_len - uri->host_start;
2379 if(!computeOnly)
2380 TRACE("(%p %p %x %d): Canonicalize reg_name=%s len=%d\n", data, uri, flags,
2381 computeOnly, debugstr_wn(uri->canon_uri+uri->host_start, uri->host_len),
2382 uri->host_len);
2384 if(!computeOnly)
2385 find_domain_name(uri->canon_uri+uri->host_start, uri->host_len,
2386 &(uri->domain_offset));
2388 return TRUE;
2391 /* Attempts to canonicalize an implicit IPv4 address. */
2392 static BOOL canonicalize_implicit_ipv4address(const parse_data *data, Uri *uri, DWORD flags, BOOL computeOnly) {
2393 uri->host_start = uri->canon_len;
2395 TRACE("%u\n", data->implicit_ipv4);
2396 /* For unknown scheme types Windows doesn't convert
2397 * the value into an IP address, but it still considers
2398 * it an IPv4 address.
2400 if(data->scheme_type == URL_SCHEME_UNKNOWN) {
2401 if(!computeOnly)
2402 memcpy(uri->canon_uri+uri->canon_len, data->host, data->host_len*sizeof(WCHAR));
2403 uri->canon_len += data->host_len;
2404 } else {
2405 if(!computeOnly)
2406 uri->canon_len += ui2ipv4(uri->canon_uri+uri->canon_len, data->implicit_ipv4);
2407 else
2408 uri->canon_len += ui2ipv4(NULL, data->implicit_ipv4);
2411 uri->host_len = uri->canon_len - uri->host_start;
2412 uri->host_type = Uri_HOST_IPV4;
2414 if(!computeOnly)
2415 TRACE("%p %p %x %d): Canonicalized implicit IP address=%s len=%d\n",
2416 data, uri, flags, computeOnly,
2417 debugstr_wn(uri->canon_uri+uri->host_start, uri->host_len),
2418 uri->host_len);
2420 return TRUE;
2423 /* Attempts to canonicalize an IPv4 address.
2425 * If the parse_data represents a URI that has an implicit IPv4 address
2426 * (ex. http://256/, this function will convert 256 into 0.0.1.0). If
2427 * the implicit IP address exceeds the value of UINT_MAX (maximum value
2428 * for an IPv4 address) it's canonicalized as if it were a reg-name.
2430 * If the parse_data contains a partial or full IPv4 address it normalizes it.
2431 * A partial IPv4 address is something like "192.0" and would be normalized to
2432 * "192.0.0.0". With a full (or partial) IPv4 address like "192.002.01.003" would
2433 * be normalized to "192.2.1.3".
2435 * NOTES:
2436 * Windows ONLY normalizes IPv4 address for known scheme types (one that isn't
2437 * URL_SCHEME_UNKNOWN). For unknown scheme types, it simply copies the data from
2438 * the original URI into the canonicalized URI, but, it still recognizes URI's
2439 * host type as HOST_IPV4.
2441 static BOOL canonicalize_ipv4address(const parse_data *data, Uri *uri, DWORD flags, BOOL computeOnly) {
2442 if(data->has_implicit_ip)
2443 return canonicalize_implicit_ipv4address(data, uri, flags, computeOnly);
2444 else {
2445 uri->host_start = uri->canon_len;
2447 /* Windows only normalizes for known scheme types. */
2448 if(data->scheme_type != URL_SCHEME_UNKNOWN) {
2449 /* parse_data contains a partial or full IPv4 address, so normalize it. */
2450 DWORD i, octetDigitCount = 0, octetCount = 0;
2451 BOOL octetHasDigit = FALSE;
2453 for(i = 0; i < data->host_len; ++i) {
2454 if(data->host[i] == '0' && !octetHasDigit) {
2455 /* Can ignore leading zeros if:
2456 * 1) It isn't the last digit of the octet.
2457 * 2) i+1 != data->host_len
2458 * 3) i+1 != '.'
2460 if(octetDigitCount == 2 ||
2461 i+1 == data->host_len ||
2462 data->host[i+1] == '.') {
2463 if(!computeOnly)
2464 uri->canon_uri[uri->canon_len] = data->host[i];
2465 ++uri->canon_len;
2466 TRACE("Adding zero\n");
2468 } else if(data->host[i] == '.') {
2469 if(!computeOnly)
2470 uri->canon_uri[uri->canon_len] = data->host[i];
2471 ++uri->canon_len;
2473 octetDigitCount = 0;
2474 octetHasDigit = FALSE;
2475 ++octetCount;
2476 } else {
2477 if(!computeOnly)
2478 uri->canon_uri[uri->canon_len] = data->host[i];
2479 ++uri->canon_len;
2481 ++octetDigitCount;
2482 octetHasDigit = TRUE;
2486 /* Make sure the canonicalized IP address has 4 dec-octets.
2487 * If doesn't add "0" ones until there is 4;
2489 for( ; octetCount < 3; ++octetCount) {
2490 if(!computeOnly) {
2491 uri->canon_uri[uri->canon_len] = '.';
2492 uri->canon_uri[uri->canon_len+1] = '0';
2495 uri->canon_len += 2;
2497 } else {
2498 /* Windows doesn't normalize addresses in unknown schemes. */
2499 if(!computeOnly)
2500 memcpy(uri->canon_uri+uri->canon_len, data->host, data->host_len*sizeof(WCHAR));
2501 uri->canon_len += data->host_len;
2504 uri->host_len = uri->canon_len - uri->host_start;
2505 if(!computeOnly)
2506 TRACE("(%p %p %x %d): Canonicalized IPv4 address, ip=%s len=%d\n",
2507 data, uri, flags, computeOnly,
2508 debugstr_wn(uri->canon_uri+uri->host_start, uri->host_len),
2509 uri->host_len);
2512 return TRUE;
2515 /* Attempts to canonicalize the IPv6 address of the URI.
2517 * Multiple things happen during the canonicalization of an IPv6 address:
2518 * 1) Any leading zero's in a h16 component are removed.
2519 * Ex: [0001:0022::] -> [1:22::]
2521 * 2) The longest sequence of zero h16 components are compressed
2522 * into a "::" (elision). If there's a tie, the first is chosen.
2524 * Ex: [0:0:0:0:1:6:7:8] -> [::1:6:7:8]
2525 * [0:0:0:0:1:2::] -> [::1:2:0:0]
2526 * [0:0:1:2:0:0:7:8] -> [::1:2:0:0:7:8]
2528 * 3) If an IPv4 address is attached to the IPv6 address, it's
2529 * also normalized.
2530 * Ex: [::001.002.022.000] -> [::1.2.22.0]
2532 * 4) If an elision is present, but, only represents one h16 component
2533 * it's expanded.
2535 * Ex: [1::2:3:4:5:6:7] -> [1:0:2:3:4:5:6:7]
2537 * 5) If the IPv6 address contains an IPv4 address and there exists
2538 * at least 1 non-zero h16 component the IPv4 address is converted
2539 * into two h16 components, otherwise it's normalized and kept as is.
2541 * Ex: [::192.200.003.4] -> [::192.200.3.4]
2542 * [ffff::192.200.003.4] -> [ffff::c0c8:3041]
2544 * NOTE:
2545 * For unknown scheme types Windows simply copies the address over without any
2546 * changes.
2548 * IPv4 address can be included in an elision if all its components are 0's.
2550 static BOOL canonicalize_ipv6address(const parse_data *data, Uri *uri,
2551 DWORD flags, BOOL computeOnly) {
2552 uri->host_start = uri->canon_len;
2554 if(data->scheme_type == URL_SCHEME_UNKNOWN) {
2555 if(!computeOnly)
2556 memcpy(uri->canon_uri+uri->canon_len, data->host, data->host_len*sizeof(WCHAR));
2557 uri->canon_len += data->host_len;
2558 } else {
2559 USHORT values[8];
2560 INT elision_start;
2561 DWORD i, elision_len;
2563 if(!ipv6_to_number(&(data->ipv6_address), values)) {
2564 TRACE("(%p %p %x %d): Failed to compute numerical value for IPv6 address.\n",
2565 data, uri, flags, computeOnly);
2566 return FALSE;
2569 if(!computeOnly)
2570 uri->canon_uri[uri->canon_len] = '[';
2571 ++uri->canon_len;
2573 /* Find where the elision should occur (if any). */
2574 compute_elision_location(&(data->ipv6_address), values, &elision_start, &elision_len);
2576 TRACE("%p %p %x %d): Elision starts at %d, len=%u\n", data, uri, flags,
2577 computeOnly, elision_start, elision_len);
2579 for(i = 0; i < 8; ++i) {
2580 BOOL in_elision = (elision_start > -1 && i >= elision_start &&
2581 i < elision_start+elision_len);
2582 BOOL do_ipv4 = (i == 6 && data->ipv6_address.ipv4 && !in_elision &&
2583 data->ipv6_address.h16_count == 0);
2585 if(i == elision_start) {
2586 if(!computeOnly) {
2587 uri->canon_uri[uri->canon_len] = ':';
2588 uri->canon_uri[uri->canon_len+1] = ':';
2590 uri->canon_len += 2;
2593 /* We can ignore the current component if we're in the elision. */
2594 if(in_elision)
2595 continue;
2597 /* We only add a ':' if we're not at i == 0, or when we're at
2598 * the very end of elision range since the ':' colon was handled
2599 * earlier. Otherwise we would end up with ":::" after elision.
2601 if(i != 0 && !(elision_start > -1 && i == elision_start+elision_len)) {
2602 if(!computeOnly)
2603 uri->canon_uri[uri->canon_len] = ':';
2604 ++uri->canon_len;
2607 if(do_ipv4) {
2608 UINT val;
2609 DWORD len;
2611 /* Combine the two parts of the IPv4 address values. */
2612 val = values[i];
2613 val <<= 16;
2614 val += values[i+1];
2616 if(!computeOnly)
2617 len = ui2ipv4(uri->canon_uri+uri->canon_len, val);
2618 else
2619 len = ui2ipv4(NULL, val);
2621 uri->canon_len += len;
2622 ++i;
2623 } else {
2624 /* Write a regular h16 component to the URI. */
2626 /* Short circuit for the trivial case. */
2627 if(values[i] == 0) {
2628 if(!computeOnly)
2629 uri->canon_uri[uri->canon_len] = '0';
2630 ++uri->canon_len;
2631 } else {
2632 static const WCHAR formatW[] = {'%','x',0};
2634 if(!computeOnly)
2635 uri->canon_len += sprintfW(uri->canon_uri+uri->canon_len,
2636 formatW, values[i]);
2637 else {
2638 WCHAR tmp[5];
2639 uri->canon_len += sprintfW(tmp, formatW, values[i]);
2645 /* Add the closing ']'. */
2646 if(!computeOnly)
2647 uri->canon_uri[uri->canon_len] = ']';
2648 ++uri->canon_len;
2651 uri->host_len = uri->canon_len - uri->host_start;
2653 if(!computeOnly)
2654 TRACE("(%p %p %x %d): Canonicalized IPv6 address %s, len=%d\n", data, uri, flags,
2655 computeOnly, debugstr_wn(uri->canon_uri+uri->host_start, uri->host_len),
2656 uri->host_len);
2658 return TRUE;
2661 /* Attempts to canonicalize the host of the URI (if any). */
2662 static BOOL canonicalize_host(const parse_data *data, Uri *uri, DWORD flags, BOOL computeOnly) {
2663 uri->host_start = -1;
2664 uri->host_len = 0;
2665 uri->domain_offset = -1;
2667 if(data->host) {
2668 switch(data->host_type) {
2669 case Uri_HOST_DNS:
2670 uri->host_type = Uri_HOST_DNS;
2671 if(!canonicalize_reg_name(data, uri, flags, computeOnly))
2672 return FALSE;
2674 break;
2675 case Uri_HOST_IPV4:
2676 uri->host_type = Uri_HOST_IPV4;
2677 if(!canonicalize_ipv4address(data, uri, flags, computeOnly))
2678 return FALSE;
2680 break;
2681 case Uri_HOST_IPV6:
2682 if(!canonicalize_ipv6address(data, uri, flags, computeOnly))
2683 return FALSE;
2685 uri->host_type = Uri_HOST_IPV6;
2686 break;
2687 case Uri_HOST_UNKNOWN:
2688 if(data->host_len > 0 || data->scheme_type != URL_SCHEME_FILE) {
2689 uri->host_start = uri->canon_len;
2691 /* Nothing happens to unknown host types. */
2692 if(!computeOnly)
2693 memcpy(uri->canon_uri+uri->canon_len, data->host, data->host_len*sizeof(WCHAR));
2694 uri->canon_len += data->host_len;
2695 uri->host_len = data->host_len;
2698 uri->host_type = Uri_HOST_UNKNOWN;
2699 break;
2700 default:
2701 FIXME("(%p %p %x %d): Canonicalization for host type %d not supported.\n", data,
2702 uri, flags, computeOnly, data->host_type);
2703 return FALSE;
2707 return TRUE;
2710 static BOOL canonicalize_port(const parse_data *data, Uri *uri, DWORD flags, BOOL computeOnly) {
2711 BOOL has_default_port = FALSE;
2712 USHORT default_port = 0;
2713 DWORD i;
2715 uri->port_offset = -1;
2717 /* Check if the scheme has a default port. */
2718 for(i = 0; i < sizeof(default_ports)/sizeof(default_ports[0]); ++i) {
2719 if(default_ports[i].scheme == data->scheme_type) {
2720 has_default_port = TRUE;
2721 default_port = default_ports[i].port;
2722 break;
2726 uri->has_port = data->has_port || has_default_port;
2728 /* Possible cases:
2729 * 1) Has a port which is the default port.
2730 * 2) Has a port (not the default).
2731 * 3) Doesn't have a port, but, scheme has a default port.
2732 * 4) No port.
2734 if(has_default_port && data->has_port && data->port_value == default_port) {
2735 /* If it's the default port and this flag isn't set, don't do anything. */
2736 if(flags & Uri_CREATE_NO_CANONICALIZE) {
2737 uri->port_offset = uri->canon_len-uri->authority_start;
2738 if(!computeOnly)
2739 uri->canon_uri[uri->canon_len] = ':';
2740 ++uri->canon_len;
2742 if(data->port) {
2743 /* Copy the original port over. */
2744 if(!computeOnly)
2745 memcpy(uri->canon_uri+uri->canon_len, data->port, data->port_len*sizeof(WCHAR));
2746 uri->canon_len += data->port_len;
2747 } else {
2748 if(!computeOnly)
2749 uri->canon_len += ui2str(uri->canon_uri+uri->canon_len, data->port_value);
2750 else
2751 uri->canon_len += ui2str(NULL, data->port_value);
2755 uri->port = default_port;
2756 } else if(data->has_port) {
2757 uri->port_offset = uri->canon_len-uri->authority_start;
2758 if(!computeOnly)
2759 uri->canon_uri[uri->canon_len] = ':';
2760 ++uri->canon_len;
2762 if(flags & Uri_CREATE_NO_CANONICALIZE && data->port) {
2763 /* Copy the original over without changes. */
2764 if(!computeOnly)
2765 memcpy(uri->canon_uri+uri->canon_len, data->port, data->port_len*sizeof(WCHAR));
2766 uri->canon_len += data->port_len;
2767 } else {
2768 if(!computeOnly)
2769 uri->canon_len += ui2str(uri->canon_uri+uri->canon_len, data->port_value);
2770 else
2771 uri->canon_len += ui2str(NULL, data->port_value);
2774 uri->port = data->port_value;
2775 } else if(has_default_port)
2776 uri->port = default_port;
2778 return TRUE;
2781 /* Canonicalizes the authority of the URI represented by the parse_data. */
2782 static BOOL canonicalize_authority(const parse_data *data, Uri *uri, DWORD flags, BOOL computeOnly) {
2783 uri->authority_start = uri->canon_len;
2784 uri->authority_len = 0;
2786 if(!canonicalize_userinfo(data, uri, flags, computeOnly))
2787 return FALSE;
2789 if(!canonicalize_host(data, uri, flags, computeOnly))
2790 return FALSE;
2792 if(!canonicalize_port(data, uri, flags, computeOnly))
2793 return FALSE;
2795 if(uri->host_start != -1 || (data->is_relative && (data->password || data->username)))
2796 uri->authority_len = uri->canon_len - uri->authority_start;
2797 else
2798 uri->authority_start = -1;
2800 return TRUE;
2803 /* Attempts to canonicalize the path of a hierarchical URI.
2805 * Things that happen:
2806 * 1). Forbidden characters are percent encoded, unless the NO_ENCODE_FORBIDDEN
2807 * flag is set or it's a file URI. Forbidden characters are always encoded
2808 * for file schemes regardless and forbidden characters are never encoded
2809 * for unknown scheme types.
2811 * 2). For known scheme types '\\' are changed to '/'.
2813 * 3). Percent encoded, unreserved characters are decoded to their actual values.
2814 * Unless the scheme type is unknown. For file schemes any percent encoded
2815 * character in the unreserved or reserved set is decoded.
2817 * 4). For File schemes if the path is starts with a drive letter and doesn't
2818 * start with a '/' then one is appended.
2819 * Ex: file://c:/test.mp3 -> file:///c:/test.mp3
2821 * 5). Dot segments are removed from the path for all scheme types
2822 * unless NO_CANONICALIZE flag is set. Dot segments aren't removed
2823 * for wildcard scheme types.
2825 * NOTES:
2826 * file://c:/test%20test -> file:///c:/test%2520test
2827 * file://c:/test%3Etest -> file:///c:/test%253Etest
2828 * if Uri_CREATE_FILE_USE_DOS_PATH is not set:
2829 * file:///c:/test%20test -> file:///c:/test%20test
2830 * file:///c:/test%test -> file:///c:/test%25test
2832 static DWORD canonicalize_path_hierarchical(const WCHAR *path, DWORD path_len, URL_SCHEME scheme_type, BOOL has_host, DWORD flags,
2833 WCHAR *ret_path) {
2834 const BOOL known_scheme = scheme_type != URL_SCHEME_UNKNOWN;
2835 const BOOL is_file = scheme_type == URL_SCHEME_FILE;
2836 const BOOL is_res = scheme_type == URL_SCHEME_RES;
2837 const WCHAR *ptr;
2838 BOOL escape_pct = FALSE;
2839 DWORD len = 0;
2841 if(!path)
2842 return 0;
2844 ptr = path;
2846 if(is_file && !has_host) {
2847 /* Check if a '/' needs to be appended for the file scheme. */
2848 if(path_len > 1 && is_drive_path(ptr) && !(flags & Uri_CREATE_FILE_USE_DOS_PATH)) {
2849 if(ret_path)
2850 ret_path[len] = '/';
2851 len++;
2852 escape_pct = TRUE;
2853 } else if(*ptr == '/') {
2854 if(!(flags & Uri_CREATE_FILE_USE_DOS_PATH)) {
2855 /* Copy the extra '/' over. */
2856 if(ret_path)
2857 ret_path[len] = '/';
2858 len++;
2860 ++ptr;
2863 if(is_drive_path(ptr)) {
2864 if(ret_path) {
2865 ret_path[len] = *ptr;
2866 /* If there's a '|' after the drive letter, convert it to a ':'. */
2867 ret_path[len+1] = ':';
2869 ptr += 2;
2870 len += 2;
2874 if(!is_file && *path && *path != '/') {
2875 /* Prepend a '/' to the path if it doesn't have one. */
2876 if(ret_path)
2877 ret_path[len] = '/';
2878 len++;
2881 for(; ptr < path+path_len; ++ptr) {
2882 BOOL do_default_action = TRUE;
2884 if(*ptr == '%' && !is_res) {
2885 const WCHAR *tmp = ptr;
2886 WCHAR val;
2888 /* Check if the % represents a valid encoded char, or if it needs encoding. */
2889 BOOL force_encode = !check_pct_encoded(&tmp) && is_file && !(flags&Uri_CREATE_FILE_USE_DOS_PATH);
2890 val = decode_pct_val(ptr);
2892 if(force_encode || escape_pct) {
2893 /* Escape the percent sign in the file URI. */
2894 if(ret_path)
2895 pct_encode_val(*ptr, ret_path+len);
2896 len += 3;
2897 do_default_action = FALSE;
2898 } else if((is_unreserved(val) && known_scheme) ||
2899 (is_file && (is_unreserved(val) || is_reserved(val) ||
2900 (val && flags&Uri_CREATE_FILE_USE_DOS_PATH && !is_forbidden_dos_path_char(val))))) {
2901 if(ret_path)
2902 ret_path[len] = val;
2903 len++;
2905 ptr += 2;
2906 continue;
2908 } else if(*ptr == '/' && is_file && (flags & Uri_CREATE_FILE_USE_DOS_PATH)) {
2909 /* Convert the '/' back to a '\\'. */
2910 if(ret_path)
2911 ret_path[len] = '\\';
2912 len++;
2913 do_default_action = FALSE;
2914 } else if(*ptr == '\\' && known_scheme) {
2915 if(!(is_file && (flags & Uri_CREATE_FILE_USE_DOS_PATH))) {
2916 /* Convert '\\' into a '/'. */
2917 if(ret_path)
2918 ret_path[len] = '/';
2919 len++;
2920 do_default_action = FALSE;
2922 } else if(known_scheme && !is_res && !is_unreserved(*ptr) && !is_reserved(*ptr) &&
2923 (!(flags & Uri_CREATE_NO_ENCODE_FORBIDDEN_CHARACTERS) || is_file)) {
2924 if(!(is_file && (flags & Uri_CREATE_FILE_USE_DOS_PATH))) {
2925 /* Escape the forbidden character. */
2926 if(ret_path)
2927 pct_encode_val(*ptr, ret_path+len);
2928 len += 3;
2929 do_default_action = FALSE;
2933 if(do_default_action) {
2934 if(ret_path)
2935 ret_path[len] = *ptr;
2936 len++;
2940 /* Removing the dot segments only happens when it's not in
2941 * computeOnly mode and it's not a wildcard scheme. File schemes
2942 * with USE_DOS_PATH set don't get dot segments removed.
2944 if(!(is_file && (flags & Uri_CREATE_FILE_USE_DOS_PATH)) &&
2945 scheme_type != URL_SCHEME_WILDCARD) {
2946 if(!(flags & Uri_CREATE_NO_CANONICALIZE) && ret_path) {
2947 /* Remove the dot segments (if any) and reset everything to the new
2948 * correct length.
2950 len = remove_dot_segments(ret_path, len);
2954 if(ret_path)
2955 TRACE("Canonicalized path %s len=%d\n", debugstr_wn(ret_path, len), len);
2956 return len;
2959 /* Attempts to canonicalize the path for an opaque URI.
2961 * For known scheme types:
2962 * 1) forbidden characters are percent encoded if
2963 * NO_ENCODE_FORBIDDEN_CHARACTERS isn't set.
2965 * 2) Percent encoded, unreserved characters are decoded
2966 * to their actual values, for known scheme types.
2968 * 3) '\\' are changed to '/' for known scheme types
2969 * except for mailto schemes.
2971 * 4) For file schemes, if USE_DOS_PATH is set all '/'
2972 * are converted to backslashes.
2974 * 5) For file schemes, if USE_DOS_PATH isn't set all '\'
2975 * are converted to forward slashes.
2977 static BOOL canonicalize_path_opaque(const parse_data *data, Uri *uri, DWORD flags, BOOL computeOnly) {
2978 const WCHAR *ptr;
2979 const BOOL known_scheme = data->scheme_type != URL_SCHEME_UNKNOWN;
2980 const BOOL is_file = data->scheme_type == URL_SCHEME_FILE;
2981 const BOOL is_mk = data->scheme_type == URL_SCHEME_MK;
2983 if(!data->path) {
2984 uri->path_start = -1;
2985 uri->path_len = 0;
2986 return TRUE;
2989 uri->path_start = uri->canon_len;
2991 if(is_mk){
2992 /* hijack this flag for SCHEME_MK to tell the function when to start
2993 * converting slashes */
2994 flags |= Uri_CREATE_FILE_USE_DOS_PATH;
2997 /* For javascript: URIs, simply copy path part without any canonicalization */
2998 if(data->scheme_type == URL_SCHEME_JAVASCRIPT) {
2999 if(!computeOnly)
3000 memcpy(uri->canon_uri+uri->canon_len, data->path, data->path_len*sizeof(WCHAR));
3001 uri->path_len = data->path_len;
3002 uri->canon_len += data->path_len;
3003 return TRUE;
3006 /* Windows doesn't allow a "//" to appear after the scheme
3007 * of a URI, if it's an opaque URI.
3009 if(data->scheme && *(data->path) == '/' && *(data->path+1) == '/') {
3010 /* So it inserts a "/." before the "//" if it exists. */
3011 if(!computeOnly) {
3012 uri->canon_uri[uri->canon_len] = '/';
3013 uri->canon_uri[uri->canon_len+1] = '.';
3016 uri->canon_len += 2;
3019 for(ptr = data->path; ptr < data->path+data->path_len; ++ptr) {
3020 BOOL do_default_action = TRUE;
3022 if(*ptr == '%' && known_scheme) {
3023 WCHAR val = decode_pct_val(ptr);
3025 if(is_unreserved(val)) {
3026 if(!computeOnly)
3027 uri->canon_uri[uri->canon_len] = val;
3028 ++uri->canon_len;
3030 ptr += 2;
3031 continue;
3033 } else if(*ptr == '/' && is_file && (flags & Uri_CREATE_FILE_USE_DOS_PATH)) {
3034 if(!computeOnly)
3035 uri->canon_uri[uri->canon_len] = '\\';
3036 ++uri->canon_len;
3037 do_default_action = FALSE;
3038 } else if(*ptr == '\\') {
3039 if((data->is_relative || is_mk || is_file) && !(flags & Uri_CREATE_FILE_USE_DOS_PATH)) {
3040 /* Convert to a '/'. */
3041 if(!computeOnly)
3042 uri->canon_uri[uri->canon_len] = '/';
3043 ++uri->canon_len;
3044 do_default_action = FALSE;
3046 } else if(is_mk && *ptr == ':' && ptr + 1 < data->path + data->path_len && *(ptr + 1) == ':') {
3047 flags &= ~Uri_CREATE_FILE_USE_DOS_PATH;
3048 } else if(known_scheme && !is_unreserved(*ptr) && !is_reserved(*ptr) &&
3049 !(flags & Uri_CREATE_NO_ENCODE_FORBIDDEN_CHARACTERS)) {
3050 if(!(is_file && (flags & Uri_CREATE_FILE_USE_DOS_PATH))) {
3051 if(!computeOnly)
3052 pct_encode_val(*ptr, uri->canon_uri+uri->canon_len);
3053 uri->canon_len += 3;
3054 do_default_action = FALSE;
3058 if(do_default_action) {
3059 if(!computeOnly)
3060 uri->canon_uri[uri->canon_len] = *ptr;
3061 ++uri->canon_len;
3065 if(is_mk && !computeOnly && !(flags & Uri_CREATE_NO_CANONICALIZE)) {
3066 DWORD new_len = remove_dot_segments(uri->canon_uri + uri->path_start,
3067 uri->canon_len - uri->path_start);
3068 uri->canon_len = uri->path_start + new_len;
3071 uri->path_len = uri->canon_len - uri->path_start;
3073 if(!computeOnly)
3074 TRACE("(%p %p %x %d): Canonicalized opaque URI path %s len=%d\n", data, uri, flags, computeOnly,
3075 debugstr_wn(uri->canon_uri+uri->path_start, uri->path_len), uri->path_len);
3076 return TRUE;
3079 /* Determines how the URI represented by the parse_data should be canonicalized.
3081 * Essentially, if the parse_data represents an hierarchical URI then it calls
3082 * canonicalize_authority and the canonicalization functions for the path. If the
3083 * URI is opaque it canonicalizes the path of the URI.
3085 static BOOL canonicalize_hierpart(const parse_data *data, Uri *uri, DWORD flags, BOOL computeOnly) {
3086 if(!data->is_opaque || (data->is_relative && (data->password || data->username))) {
3087 /* "//" is only added for non-wildcard scheme types.
3089 * A "//" is only added to a relative URI if it has a
3090 * host or port component (this only happens if a IUriBuilder
3091 * is generating an IUri).
3093 if((data->is_relative && (data->host || data->has_port)) ||
3094 (!data->is_relative && data->scheme_type != URL_SCHEME_WILDCARD)) {
3095 if(data->scheme_type == URL_SCHEME_WILDCARD)
3096 FIXME("Here\n");
3098 if(!computeOnly) {
3099 INT pos = uri->canon_len;
3101 uri->canon_uri[pos] = '/';
3102 uri->canon_uri[pos+1] = '/';
3104 uri->canon_len += 2;
3107 if(!canonicalize_authority(data, uri, flags, computeOnly))
3108 return FALSE;
3110 if(data->is_relative && (data->password || data->username)) {
3111 if(!canonicalize_path_opaque(data, uri, flags, computeOnly))
3112 return FALSE;
3113 } else {
3114 if(!computeOnly)
3115 uri->path_start = uri->canon_len;
3116 uri->path_len = canonicalize_path_hierarchical(data->path, data->path_len, data->scheme_type, data->host_len != 0,
3117 flags, computeOnly ? NULL : uri->canon_uri+uri->canon_len);
3118 uri->canon_len += uri->path_len;
3119 if(!computeOnly && !uri->path_len)
3120 uri->path_start = -1;
3122 } else {
3123 /* Opaque URI's don't have an authority. */
3124 uri->userinfo_start = uri->userinfo_split = -1;
3125 uri->userinfo_len = 0;
3126 uri->host_start = -1;
3127 uri->host_len = 0;
3128 uri->host_type = Uri_HOST_UNKNOWN;
3129 uri->has_port = FALSE;
3130 uri->authority_start = -1;
3131 uri->authority_len = 0;
3132 uri->domain_offset = -1;
3133 uri->port_offset = -1;
3135 if(is_hierarchical_scheme(data->scheme_type)) {
3136 DWORD i;
3138 /* Absolute URIs aren't displayed for known scheme types
3139 * which should be hierarchical URIs.
3141 uri->display_modifiers |= URI_DISPLAY_NO_ABSOLUTE_URI;
3143 /* Windows also sets the port for these (if they have one). */
3144 for(i = 0; i < sizeof(default_ports)/sizeof(default_ports[0]); ++i) {
3145 if(data->scheme_type == default_ports[i].scheme) {
3146 uri->has_port = TRUE;
3147 uri->port = default_ports[i].port;
3148 break;
3153 if(!canonicalize_path_opaque(data, uri, flags, computeOnly))
3154 return FALSE;
3157 if(uri->path_start > -1 && !computeOnly)
3158 /* Finding file extensions happens for both types of URIs. */
3159 uri->extension_offset = find_file_extension(uri->canon_uri+uri->path_start, uri->path_len);
3160 else
3161 uri->extension_offset = -1;
3163 return TRUE;
3166 /* Attempts to canonicalize the query string of the URI.
3168 * Things that happen:
3169 * 1) For known scheme types forbidden characters
3170 * are percent encoded, unless the NO_DECODE_EXTRA_INFO flag is set
3171 * or NO_ENCODE_FORBIDDEN_CHARACTERS is set.
3173 * 2) For known scheme types, percent encoded, unreserved characters
3174 * are decoded as long as the NO_DECODE_EXTRA_INFO flag isn't set.
3176 static BOOL canonicalize_query(const parse_data *data, Uri *uri, DWORD flags, BOOL computeOnly) {
3177 const WCHAR *ptr, *end;
3178 const BOOL known_scheme = data->scheme_type != URL_SCHEME_UNKNOWN;
3180 if(!data->query) {
3181 uri->query_start = -1;
3182 uri->query_len = 0;
3183 return TRUE;
3186 uri->query_start = uri->canon_len;
3188 end = data->query+data->query_len;
3189 for(ptr = data->query; ptr < end; ++ptr) {
3190 if(*ptr == '%') {
3191 if(known_scheme && !(flags & Uri_CREATE_NO_DECODE_EXTRA_INFO)) {
3192 WCHAR val = decode_pct_val(ptr);
3193 if(is_unreserved(val)) {
3194 if(!computeOnly)
3195 uri->canon_uri[uri->canon_len] = val;
3196 ++uri->canon_len;
3198 ptr += 2;
3199 continue;
3202 } else if(known_scheme && !is_unreserved(*ptr) && !is_reserved(*ptr)) {
3203 if(!(flags & Uri_CREATE_NO_ENCODE_FORBIDDEN_CHARACTERS) &&
3204 !(flags & Uri_CREATE_NO_DECODE_EXTRA_INFO)) {
3205 if(!computeOnly)
3206 pct_encode_val(*ptr, uri->canon_uri+uri->canon_len);
3207 uri->canon_len += 3;
3208 continue;
3212 if(!computeOnly)
3213 uri->canon_uri[uri->canon_len] = *ptr;
3214 ++uri->canon_len;
3217 uri->query_len = uri->canon_len - uri->query_start;
3219 if(!computeOnly)
3220 TRACE("(%p %p %x %d): Canonicalized query string %s len=%d\n", data, uri, flags,
3221 computeOnly, debugstr_wn(uri->canon_uri+uri->query_start, uri->query_len),
3222 uri->query_len);
3223 return TRUE;
3226 static BOOL canonicalize_fragment(const parse_data *data, Uri *uri, DWORD flags, BOOL computeOnly) {
3227 const WCHAR *ptr, *end;
3228 const BOOL known_scheme = data->scheme_type != URL_SCHEME_UNKNOWN;
3230 if(!data->fragment) {
3231 uri->fragment_start = -1;
3232 uri->fragment_len = 0;
3233 return TRUE;
3236 uri->fragment_start = uri->canon_len;
3238 end = data->fragment + data->fragment_len;
3239 for(ptr = data->fragment; ptr < end; ++ptr) {
3240 if(*ptr == '%') {
3241 if(known_scheme && !(flags & Uri_CREATE_NO_DECODE_EXTRA_INFO)) {
3242 WCHAR val = decode_pct_val(ptr);
3243 if(is_unreserved(val)) {
3244 if(!computeOnly)
3245 uri->canon_uri[uri->canon_len] = val;
3246 ++uri->canon_len;
3248 ptr += 2;
3249 continue;
3252 } else if(known_scheme && !is_unreserved(*ptr) && !is_reserved(*ptr)) {
3253 if(!(flags & Uri_CREATE_NO_ENCODE_FORBIDDEN_CHARACTERS) &&
3254 !(flags & Uri_CREATE_NO_DECODE_EXTRA_INFO)) {
3255 if(!computeOnly)
3256 pct_encode_val(*ptr, uri->canon_uri+uri->canon_len);
3257 uri->canon_len += 3;
3258 continue;
3262 if(!computeOnly)
3263 uri->canon_uri[uri->canon_len] = *ptr;
3264 ++uri->canon_len;
3267 uri->fragment_len = uri->canon_len - uri->fragment_start;
3269 if(!computeOnly)
3270 TRACE("(%p %p %x %d): Canonicalized fragment %s len=%d\n", data, uri, flags,
3271 computeOnly, debugstr_wn(uri->canon_uri+uri->fragment_start, uri->fragment_len),
3272 uri->fragment_len);
3273 return TRUE;
3276 /* Canonicalizes the scheme information specified in the parse_data using the specified flags. */
3277 static BOOL canonicalize_scheme(const parse_data *data, Uri *uri, DWORD flags, BOOL computeOnly) {
3278 uri->scheme_start = -1;
3279 uri->scheme_len = 0;
3281 if(!data->scheme) {
3282 /* The only type of URI that doesn't have to have a scheme is a relative
3283 * URI.
3285 if(!data->is_relative) {
3286 FIXME("(%p %p %x): Unable to determine the scheme type of %s.\n", data,
3287 uri, flags, debugstr_w(data->uri));
3288 return FALSE;
3290 } else {
3291 if(!computeOnly) {
3292 DWORD i;
3293 INT pos = uri->canon_len;
3295 for(i = 0; i < data->scheme_len; ++i) {
3296 /* Scheme name must be lower case after canonicalization. */
3297 uri->canon_uri[i + pos] = tolowerW(data->scheme[i]);
3300 uri->canon_uri[i + pos] = ':';
3301 uri->scheme_start = pos;
3303 TRACE("(%p %p %x): Canonicalized scheme=%s, len=%d.\n", data, uri, flags,
3304 debugstr_wn(uri->canon_uri+uri->scheme_start, data->scheme_len), data->scheme_len);
3307 /* This happens in both computation modes. */
3308 uri->canon_len += data->scheme_len + 1;
3309 uri->scheme_len = data->scheme_len;
3311 return TRUE;
3314 /* Computes what the length of the URI specified by the parse_data will be
3315 * after canonicalization occurs using the specified flags.
3317 * This function will return a non-zero value indicating the length of the canonicalized
3318 * URI, or -1 on error.
3320 static int compute_canonicalized_length(const parse_data *data, DWORD flags) {
3321 Uri uri;
3323 memset(&uri, 0, sizeof(Uri));
3325 TRACE("(%p %x): Beginning to compute canonicalized length for URI %s\n", data, flags,
3326 debugstr_w(data->uri));
3328 if(!canonicalize_scheme(data, &uri, flags, TRUE)) {
3329 ERR("(%p %x): Failed to compute URI scheme length.\n", data, flags);
3330 return -1;
3333 if(!canonicalize_hierpart(data, &uri, flags, TRUE)) {
3334 ERR("(%p %x): Failed to compute URI hierpart length.\n", data, flags);
3335 return -1;
3338 if(!canonicalize_query(data, &uri, flags, TRUE)) {
3339 ERR("(%p %x): Failed to compute query string length.\n", data, flags);
3340 return -1;
3343 if(!canonicalize_fragment(data, &uri, flags, TRUE)) {
3344 ERR("(%p %x): Failed to compute fragment length.\n", data, flags);
3345 return -1;
3348 TRACE("(%p %x): Finished computing canonicalized URI length. length=%d\n", data, flags, uri.canon_len);
3350 return uri.canon_len;
3353 /* Canonicalizes the URI data specified in the parse_data, using the given flags. If the
3354 * canonicalization succeeds it will store all the canonicalization information
3355 * in the pointer to the Uri.
3357 * To canonicalize a URI this function first computes what the length of the URI
3358 * specified by the parse_data will be. Once this is done it will then perform the actual
3359 * canonicalization of the URI.
3361 static HRESULT canonicalize_uri(const parse_data *data, Uri *uri, DWORD flags) {
3362 INT len;
3364 uri->canon_uri = NULL;
3365 uri->canon_size = uri->canon_len = 0;
3367 TRACE("(%p %p %x): beginning to canonicalize URI %s.\n", data, uri, flags, debugstr_w(data->uri));
3369 /* First try to compute the length of the URI. */
3370 len = compute_canonicalized_length(data, flags);
3371 if(len == -1) {
3372 ERR("(%p %p %x): Could not compute the canonicalized length of %s.\n", data, uri, flags,
3373 debugstr_w(data->uri));
3374 return E_INVALIDARG;
3377 uri->canon_uri = heap_alloc((len+1)*sizeof(WCHAR));
3378 if(!uri->canon_uri)
3379 return E_OUTOFMEMORY;
3381 uri->canon_size = len;
3382 if(!canonicalize_scheme(data, uri, flags, FALSE)) {
3383 ERR("(%p %p %x): Unable to canonicalize the scheme of the URI.\n", data, uri, flags);
3384 return E_INVALIDARG;
3386 uri->scheme_type = data->scheme_type;
3388 if(!canonicalize_hierpart(data, uri, flags, FALSE)) {
3389 ERR("(%p %p %x): Unable to canonicalize the heirpart of the URI\n", data, uri, flags);
3390 return E_INVALIDARG;
3393 if(!canonicalize_query(data, uri, flags, FALSE)) {
3394 ERR("(%p %p %x): Unable to canonicalize query string of the URI.\n",
3395 data, uri, flags);
3396 return E_INVALIDARG;
3399 if(!canonicalize_fragment(data, uri, flags, FALSE)) {
3400 ERR("(%p %p %x): Unable to canonicalize fragment of the URI.\n",
3401 data, uri, flags);
3402 return E_INVALIDARG;
3405 /* There's a possibility we didn't use all the space we allocated
3406 * earlier.
3408 if(uri->canon_len < uri->canon_size) {
3409 /* This happens if the URI is hierarchical and dot
3410 * segments were removed from its path.
3412 WCHAR *tmp = heap_realloc(uri->canon_uri, (uri->canon_len+1)*sizeof(WCHAR));
3413 if(!tmp)
3414 return E_OUTOFMEMORY;
3416 uri->canon_uri = tmp;
3417 uri->canon_size = uri->canon_len;
3420 uri->canon_uri[uri->canon_len] = '\0';
3421 TRACE("(%p %p %x): finished canonicalizing the URI. uri=%s\n", data, uri, flags, debugstr_w(uri->canon_uri));
3423 return S_OK;
3426 static HRESULT get_builder_component(LPWSTR *component, DWORD *component_len,
3427 LPCWSTR source, DWORD source_len,
3428 LPCWSTR *output, DWORD *output_len)
3430 if(!output_len) {
3431 if(output)
3432 *output = NULL;
3433 return E_POINTER;
3436 if(!output) {
3437 *output_len = 0;
3438 return E_POINTER;
3441 if(!(*component) && source) {
3442 /* Allocate 'component', and copy the contents from 'source'
3443 * into the new allocation.
3445 *component = heap_alloc((source_len+1)*sizeof(WCHAR));
3446 if(!(*component))
3447 return E_OUTOFMEMORY;
3449 memcpy(*component, source, source_len*sizeof(WCHAR));
3450 (*component)[source_len] = '\0';
3451 *component_len = source_len;
3454 *output = *component;
3455 *output_len = *component_len;
3456 return *output ? S_OK : S_FALSE;
3459 /* Allocates 'component' and copies the string from 'new_value' into 'component'.
3460 * If 'prefix' is set and 'new_value' isn't NULL, then it checks if 'new_value'
3461 * starts with 'prefix'. If it doesn't then 'prefix' is prepended to 'component'.
3463 * If everything is successful, then will set 'success_flag' in 'flags'.
3465 static HRESULT set_builder_component(LPWSTR *component, DWORD *component_len, LPCWSTR new_value,
3466 WCHAR prefix, DWORD *flags, DWORD success_flag)
3468 heap_free(*component);
3470 if(!new_value) {
3471 *component = NULL;
3472 *component_len = 0;
3473 } else {
3474 BOOL add_prefix = FALSE;
3475 DWORD len = lstrlenW(new_value);
3476 DWORD pos = 0;
3478 if(prefix && *new_value != prefix) {
3479 add_prefix = TRUE;
3480 *component = heap_alloc((len+2)*sizeof(WCHAR));
3481 } else
3482 *component = heap_alloc((len+1)*sizeof(WCHAR));
3484 if(!(*component))
3485 return E_OUTOFMEMORY;
3487 if(add_prefix)
3488 (*component)[pos++] = prefix;
3490 memcpy(*component+pos, new_value, (len+1)*sizeof(WCHAR));
3491 *component_len = len+pos;
3494 *flags |= success_flag;
3495 return S_OK;
3498 static void reset_builder(UriBuilder *builder) {
3499 if(builder->uri)
3500 IUri_Release(&builder->uri->IUri_iface);
3501 builder->uri = NULL;
3503 heap_free(builder->fragment);
3504 builder->fragment = NULL;
3505 builder->fragment_len = 0;
3507 heap_free(builder->host);
3508 builder->host = NULL;
3509 builder->host_len = 0;
3511 heap_free(builder->password);
3512 builder->password = NULL;
3513 builder->password_len = 0;
3515 heap_free(builder->path);
3516 builder->path = NULL;
3517 builder->path_len = 0;
3519 heap_free(builder->query);
3520 builder->query = NULL;
3521 builder->query_len = 0;
3523 heap_free(builder->scheme);
3524 builder->scheme = NULL;
3525 builder->scheme_len = 0;
3527 heap_free(builder->username);
3528 builder->username = NULL;
3529 builder->username_len = 0;
3531 builder->has_port = FALSE;
3532 builder->port = 0;
3533 builder->modified_props = 0;
3536 static HRESULT validate_scheme_name(const UriBuilder *builder, parse_data *data, DWORD flags) {
3537 const WCHAR *component;
3538 const WCHAR *ptr;
3539 const WCHAR **pptr;
3540 DWORD expected_len;
3542 if(builder->scheme) {
3543 ptr = builder->scheme;
3544 expected_len = builder->scheme_len;
3545 } else if(builder->uri && builder->uri->scheme_start > -1) {
3546 ptr = builder->uri->canon_uri+builder->uri->scheme_start;
3547 expected_len = builder->uri->scheme_len;
3548 } else {
3549 static const WCHAR nullW[] = {0};
3550 ptr = nullW;
3551 expected_len = 0;
3554 component = ptr;
3555 pptr = &ptr;
3556 if(parse_scheme(pptr, data, flags, ALLOW_NULL_TERM_SCHEME) &&
3557 data->scheme_len == expected_len) {
3558 if(data->scheme)
3559 TRACE("(%p %p %x): Found valid scheme component %s len=%d.\n", builder, data, flags,
3560 debugstr_wn(data->scheme, data->scheme_len), data->scheme_len);
3561 } else {
3562 TRACE("(%p %p %x): Invalid scheme component found %s.\n", builder, data, flags,
3563 debugstr_wn(component, expected_len));
3564 return INET_E_INVALID_URL;
3567 return S_OK;
3570 static HRESULT validate_username(const UriBuilder *builder, parse_data *data, DWORD flags) {
3571 const WCHAR *ptr;
3572 const WCHAR **pptr;
3573 DWORD expected_len;
3575 if(builder->username) {
3576 ptr = builder->username;
3577 expected_len = builder->username_len;
3578 } else if(!(builder->modified_props & Uri_HAS_USER_NAME) && builder->uri &&
3579 builder->uri->userinfo_start > -1 && builder->uri->userinfo_split != 0) {
3580 /* Just use the username from the base Uri. */
3581 data->username = builder->uri->canon_uri+builder->uri->userinfo_start;
3582 data->username_len = (builder->uri->userinfo_split > -1) ?
3583 builder->uri->userinfo_split : builder->uri->userinfo_len;
3584 ptr = NULL;
3585 } else {
3586 ptr = NULL;
3587 expected_len = 0;
3590 if(ptr) {
3591 const WCHAR *component = ptr;
3592 pptr = &ptr;
3593 if(parse_username(pptr, data, flags, ALLOW_NULL_TERM_USER_NAME) &&
3594 data->username_len == expected_len)
3595 TRACE("(%p %p %x): Found valid username component %s len=%d.\n", builder, data, flags,
3596 debugstr_wn(data->username, data->username_len), data->username_len);
3597 else {
3598 TRACE("(%p %p %x): Invalid username component found %s.\n", builder, data, flags,
3599 debugstr_wn(component, expected_len));
3600 return INET_E_INVALID_URL;
3604 return S_OK;
3607 static HRESULT validate_password(const UriBuilder *builder, parse_data *data, DWORD flags) {
3608 const WCHAR *ptr;
3609 const WCHAR **pptr;
3610 DWORD expected_len;
3612 if(builder->password) {
3613 ptr = builder->password;
3614 expected_len = builder->password_len;
3615 } else if(!(builder->modified_props & Uri_HAS_PASSWORD) && builder->uri &&
3616 builder->uri->userinfo_split > -1) {
3617 data->password = builder->uri->canon_uri+builder->uri->userinfo_start+builder->uri->userinfo_split+1;
3618 data->password_len = builder->uri->userinfo_len-builder->uri->userinfo_split-1;
3619 ptr = NULL;
3620 } else {
3621 ptr = NULL;
3622 expected_len = 0;
3625 if(ptr) {
3626 const WCHAR *component = ptr;
3627 pptr = &ptr;
3628 if(parse_password(pptr, data, flags, ALLOW_NULL_TERM_PASSWORD) &&
3629 data->password_len == expected_len)
3630 TRACE("(%p %p %x): Found valid password component %s len=%d.\n", builder, data, flags,
3631 debugstr_wn(data->password, data->password_len), data->password_len);
3632 else {
3633 TRACE("(%p %p %x): Invalid password component found %s.\n", builder, data, flags,
3634 debugstr_wn(component, expected_len));
3635 return INET_E_INVALID_URL;
3639 return S_OK;
3642 static HRESULT validate_userinfo(const UriBuilder *builder, parse_data *data, DWORD flags) {
3643 HRESULT hr;
3645 hr = validate_username(builder, data, flags);
3646 if(FAILED(hr))
3647 return hr;
3649 hr = validate_password(builder, data, flags);
3650 if(FAILED(hr))
3651 return hr;
3653 return S_OK;
3656 static HRESULT validate_host(const UriBuilder *builder, parse_data *data, DWORD flags) {
3657 const WCHAR *ptr;
3658 const WCHAR **pptr;
3659 DWORD expected_len;
3661 if(builder->host) {
3662 ptr = builder->host;
3663 expected_len = builder->host_len;
3664 } else if(!(builder->modified_props & Uri_HAS_HOST) && builder->uri && builder->uri->host_start > -1) {
3665 ptr = builder->uri->canon_uri + builder->uri->host_start;
3666 expected_len = builder->uri->host_len;
3667 } else
3668 ptr = NULL;
3670 if(ptr) {
3671 const WCHAR *component = ptr;
3672 DWORD extras = ALLOW_BRACKETLESS_IP_LITERAL|IGNORE_PORT_DELIMITER|SKIP_IP_FUTURE_CHECK;
3673 pptr = &ptr;
3675 if(parse_host(pptr, data, flags, extras) && data->host_len == expected_len)
3676 TRACE("(%p %p %x): Found valid host name %s len=%d type=%d.\n", builder, data, flags,
3677 debugstr_wn(data->host, data->host_len), data->host_len, data->host_type);
3678 else {
3679 TRACE("(%p %p %x): Invalid host name found %s.\n", builder, data, flags,
3680 debugstr_wn(component, expected_len));
3681 return INET_E_INVALID_URL;
3685 return S_OK;
3688 static void setup_port(const UriBuilder *builder, parse_data *data, DWORD flags) {
3689 if(builder->modified_props & Uri_HAS_PORT) {
3690 if(builder->has_port) {
3691 data->has_port = TRUE;
3692 data->port_value = builder->port;
3694 } else if(builder->uri && builder->uri->has_port) {
3695 data->has_port = TRUE;
3696 data->port_value = builder->uri->port;
3699 if(data->has_port)
3700 TRACE("(%p %p %x): Using %u as port for IUri.\n", builder, data, flags, data->port_value);
3703 static HRESULT validate_path(const UriBuilder *builder, parse_data *data, DWORD flags) {
3704 const WCHAR *ptr = NULL;
3705 const WCHAR *component;
3706 const WCHAR **pptr;
3707 DWORD expected_len;
3708 BOOL check_len = TRUE;
3709 BOOL valid = FALSE;
3711 if(builder->path) {
3712 ptr = builder->path;
3713 expected_len = builder->path_len;
3714 } else if(!(builder->modified_props & Uri_HAS_PATH) &&
3715 builder->uri && builder->uri->path_start > -1) {
3716 ptr = builder->uri->canon_uri+builder->uri->path_start;
3717 expected_len = builder->uri->path_len;
3718 } else {
3719 static const WCHAR nullW[] = {0};
3720 ptr = nullW;
3721 check_len = FALSE;
3722 expected_len = -1;
3725 component = ptr;
3726 pptr = &ptr;
3728 /* How the path is validated depends on what type of
3729 * URI it is.
3731 valid = data->is_opaque ?
3732 parse_path_opaque(pptr, data, flags) : parse_path_hierarchical(pptr, data, flags);
3734 if(!valid || (check_len && expected_len != data->path_len)) {
3735 TRACE("(%p %p %x): Invalid path component %s.\n", builder, data, flags,
3736 debugstr_wn(component, expected_len) );
3737 return INET_E_INVALID_URL;
3740 TRACE("(%p %p %x): Valid path component %s len=%d.\n", builder, data, flags,
3741 debugstr_wn(data->path, data->path_len), data->path_len);
3743 return S_OK;
3746 static HRESULT validate_query(const UriBuilder *builder, parse_data *data, DWORD flags) {
3747 const WCHAR *ptr = NULL;
3748 const WCHAR **pptr;
3749 DWORD expected_len;
3751 if(builder->query) {
3752 ptr = builder->query;
3753 expected_len = builder->query_len;
3754 } else if(!(builder->modified_props & Uri_HAS_QUERY) && builder->uri &&
3755 builder->uri->query_start > -1) {
3756 ptr = builder->uri->canon_uri+builder->uri->query_start;
3757 expected_len = builder->uri->query_len;
3760 if(ptr) {
3761 const WCHAR *component = ptr;
3762 pptr = &ptr;
3764 if(parse_query(pptr, data, flags) && expected_len == data->query_len)
3765 TRACE("(%p %p %x): Valid query component %s len=%d.\n", builder, data, flags,
3766 debugstr_wn(data->query, data->query_len), data->query_len);
3767 else {
3768 TRACE("(%p %p %x): Invalid query component %s.\n", builder, data, flags,
3769 debugstr_wn(component, expected_len));
3770 return INET_E_INVALID_URL;
3774 return S_OK;
3777 static HRESULT validate_fragment(const UriBuilder *builder, parse_data *data, DWORD flags) {
3778 const WCHAR *ptr = NULL;
3779 const WCHAR **pptr;
3780 DWORD expected_len;
3782 if(builder->fragment) {
3783 ptr = builder->fragment;
3784 expected_len = builder->fragment_len;
3785 } else if(!(builder->modified_props & Uri_HAS_FRAGMENT) && builder->uri &&
3786 builder->uri->fragment_start > -1) {
3787 ptr = builder->uri->canon_uri+builder->uri->fragment_start;
3788 expected_len = builder->uri->fragment_len;
3791 if(ptr) {
3792 const WCHAR *component = ptr;
3793 pptr = &ptr;
3795 if(parse_fragment(pptr, data, flags) && expected_len == data->fragment_len)
3796 TRACE("(%p %p %x): Valid fragment component %s len=%d.\n", builder, data, flags,
3797 debugstr_wn(data->fragment, data->fragment_len), data->fragment_len);
3798 else {
3799 TRACE("(%p %p %x): Invalid fragment component %s.\n", builder, data, flags,
3800 debugstr_wn(component, expected_len));
3801 return INET_E_INVALID_URL;
3805 return S_OK;
3808 static HRESULT validate_components(const UriBuilder *builder, parse_data *data, DWORD flags) {
3809 HRESULT hr;
3811 memset(data, 0, sizeof(parse_data));
3813 TRACE("(%p %p %x): Beginning to validate builder components.\n", builder, data, flags);
3815 hr = validate_scheme_name(builder, data, flags);
3816 if(FAILED(hr))
3817 return hr;
3819 /* Extra validation for file schemes. */
3820 if(data->scheme_type == URL_SCHEME_FILE) {
3821 if((builder->password || (builder->uri && builder->uri->userinfo_split > -1)) ||
3822 (builder->username || (builder->uri && builder->uri->userinfo_start > -1))) {
3823 TRACE("(%p %p %x): File schemes can't contain a username or password.\n",
3824 builder, data, flags);
3825 return INET_E_INVALID_URL;
3829 hr = validate_userinfo(builder, data, flags);
3830 if(FAILED(hr))
3831 return hr;
3833 hr = validate_host(builder, data, flags);
3834 if(FAILED(hr))
3835 return hr;
3837 setup_port(builder, data, flags);
3839 /* The URI is opaque if it doesn't have an authority component. */
3840 if(!data->is_relative)
3841 data->is_opaque = !data->username && !data->password && !data->host && !data->has_port
3842 && data->scheme_type != URL_SCHEME_FILE;
3843 else
3844 data->is_opaque = !data->host && !data->has_port;
3846 hr = validate_path(builder, data, flags);
3847 if(FAILED(hr))
3848 return hr;
3850 hr = validate_query(builder, data, flags);
3851 if(FAILED(hr))
3852 return hr;
3854 hr = validate_fragment(builder, data, flags);
3855 if(FAILED(hr))
3856 return hr;
3858 TRACE("(%p %p %x): Finished validating builder components.\n", builder, data, flags);
3860 return S_OK;
3863 static HRESULT compare_file_paths(const Uri *a, const Uri *b, BOOL *ret)
3865 WCHAR *canon_path_a, *canon_path_b;
3866 DWORD len_a, len_b;
3868 if(!a->path_len) {
3869 *ret = !b->path_len;
3870 return S_OK;
3873 if(!b->path_len) {
3874 *ret = FALSE;
3875 return S_OK;
3878 /* Fast path */
3879 if(a->path_len == b->path_len && !memicmpW(a->canon_uri+a->path_start, b->canon_uri+b->path_start, a->path_len)) {
3880 *ret = TRUE;
3881 return S_OK;
3884 len_a = canonicalize_path_hierarchical(a->canon_uri+a->path_start, a->path_len, a->scheme_type, FALSE, 0, NULL);
3885 len_b = canonicalize_path_hierarchical(b->canon_uri+b->path_start, b->path_len, b->scheme_type, FALSE, 0, NULL);
3887 canon_path_a = heap_alloc(len_a*sizeof(WCHAR));
3888 if(!canon_path_a)
3889 return E_OUTOFMEMORY;
3890 canon_path_b = heap_alloc(len_b*sizeof(WCHAR));
3891 if(!canon_path_b) {
3892 heap_free(canon_path_a);
3893 return E_OUTOFMEMORY;
3896 len_a = canonicalize_path_hierarchical(a->canon_uri+a->path_start, a->path_len, a->scheme_type, FALSE, 0, canon_path_a);
3897 len_b = canonicalize_path_hierarchical(b->canon_uri+b->path_start, b->path_len, b->scheme_type, FALSE, 0, canon_path_b);
3899 *ret = len_a == len_b && !memicmpW(canon_path_a, canon_path_b, len_a);
3901 heap_free(canon_path_a);
3902 heap_free(canon_path_b);
3903 return S_OK;
3906 /* Checks if the two Uri's are logically equivalent. It's a simple
3907 * comparison, since they are both of type Uri, and it can access
3908 * the properties of each Uri directly without the need to go
3909 * through the "IUri_Get*" interface calls.
3911 static HRESULT compare_uris(const Uri *a, const Uri *b, BOOL *ret) {
3912 const BOOL known_scheme = a->scheme_type != URL_SCHEME_UNKNOWN;
3913 const BOOL are_hierarchical = a->authority_start > -1 && b->authority_start > -1;
3914 HRESULT hres;
3916 *ret = FALSE;
3918 if(a->scheme_type != b->scheme_type)
3919 return S_OK;
3921 /* Only compare the scheme names (if any) if their unknown scheme types. */
3922 if(!known_scheme) {
3923 if((a->scheme_start > -1 && b->scheme_start > -1) &&
3924 (a->scheme_len == b->scheme_len)) {
3925 /* Make sure the schemes are the same. */
3926 if(StrCmpNW(a->canon_uri+a->scheme_start, b->canon_uri+b->scheme_start, a->scheme_len))
3927 return S_OK;
3928 } else if(a->scheme_len != b->scheme_len)
3929 /* One of the Uri's has a scheme name, while the other doesn't. */
3930 return S_OK;
3933 /* If they have a userinfo component, perform case sensitive compare. */
3934 if((a->userinfo_start > -1 && b->userinfo_start > -1) &&
3935 (a->userinfo_len == b->userinfo_len)) {
3936 if(StrCmpNW(a->canon_uri+a->userinfo_start, b->canon_uri+b->userinfo_start, a->userinfo_len))
3937 return S_OK;
3938 } else if(a->userinfo_len != b->userinfo_len)
3939 /* One of the Uri's had a userinfo, while the other one doesn't. */
3940 return S_OK;
3942 /* Check if they have a host name. */
3943 if((a->host_start > -1 && b->host_start > -1) &&
3944 (a->host_len == b->host_len)) {
3945 /* Perform a case insensitive compare if they are a known scheme type. */
3946 if(known_scheme) {
3947 if(StrCmpNIW(a->canon_uri+a->host_start, b->canon_uri+b->host_start, a->host_len))
3948 return S_OK;
3949 } else if(StrCmpNW(a->canon_uri+a->host_start, b->canon_uri+b->host_start, a->host_len))
3950 return S_OK;
3951 } else if(a->host_len != b->host_len)
3952 /* One of the Uri's had a host, while the other one didn't. */
3953 return S_OK;
3955 if(a->has_port && b->has_port) {
3956 if(a->port != b->port)
3957 return S_OK;
3958 } else if(a->has_port || b->has_port)
3959 /* One had a port, while the other one didn't. */
3960 return S_OK;
3962 /* Windows is weird with how it handles paths. For example
3963 * One URI could be "http://google.com" (after canonicalization)
3964 * and one could be "http://google.com/" and the IsEqual function
3965 * would still evaluate to TRUE, but, only if they are both hierarchical
3966 * URIs.
3968 if(a->scheme_type == URL_SCHEME_FILE) {
3969 BOOL cmp;
3971 hres = compare_file_paths(a, b, &cmp);
3972 if(FAILED(hres) || !cmp)
3973 return hres;
3974 } else if((a->path_start > -1 && b->path_start > -1) &&
3975 (a->path_len == b->path_len)) {
3976 if(StrCmpNW(a->canon_uri+a->path_start, b->canon_uri+b->path_start, a->path_len))
3977 return S_OK;
3978 } else if(are_hierarchical && a->path_len == -1 && b->path_len == 0) {
3979 if(*(a->canon_uri+a->path_start) != '/')
3980 return S_OK;
3981 } else if(are_hierarchical && b->path_len == 1 && a->path_len == 0) {
3982 if(*(b->canon_uri+b->path_start) != '/')
3983 return S_OK;
3984 } else if(a->path_len != b->path_len)
3985 return S_OK;
3987 /* Compare the query strings of the two URIs. */
3988 if((a->query_start > -1 && b->query_start > -1) &&
3989 (a->query_len == b->query_len)) {
3990 if(StrCmpNW(a->canon_uri+a->query_start, b->canon_uri+b->query_start, a->query_len))
3991 return S_OK;
3992 } else if(a->query_len != b->query_len)
3993 return S_OK;
3995 if((a->fragment_start > -1 && b->fragment_start > -1) &&
3996 (a->fragment_len == b->fragment_len)) {
3997 if(StrCmpNW(a->canon_uri+a->fragment_start, b->canon_uri+b->fragment_start, a->fragment_len))
3998 return S_OK;
3999 } else if(a->fragment_len != b->fragment_len)
4000 return S_OK;
4002 /* If we get here, the two URIs are equivalent. */
4003 *ret = TRUE;
4004 return S_OK;
4007 static void convert_to_dos_path(const WCHAR *path, DWORD path_len,
4008 WCHAR *output, DWORD *output_len)
4010 const WCHAR *ptr = path;
4012 if(path_len > 3 && *ptr == '/' && is_drive_path(path+1))
4013 /* Skip over the leading / before the drive path. */
4014 ++ptr;
4016 for(; ptr < path+path_len; ++ptr) {
4017 if(*ptr == '/') {
4018 if(output)
4019 *output++ = '\\';
4020 (*output_len)++;
4021 } else {
4022 if(output)
4023 *output++ = *ptr;
4024 (*output_len)++;
4029 /* Generates a raw uri string using the parse_data. */
4030 static DWORD generate_raw_uri(const parse_data *data, BSTR uri, DWORD flags) {
4031 DWORD length = 0;
4033 if(data->scheme) {
4034 if(uri) {
4035 memcpy(uri, data->scheme, data->scheme_len*sizeof(WCHAR));
4036 uri[data->scheme_len] = ':';
4038 length += data->scheme_len+1;
4041 if(!data->is_opaque) {
4042 /* For the "//" which appears before the authority component. */
4043 if(uri) {
4044 uri[length] = '/';
4045 uri[length+1] = '/';
4047 length += 2;
4049 /* Check if we need to add the "\\" before the host name
4050 * of a UNC server name in a DOS path.
4052 if(flags & RAW_URI_CONVERT_TO_DOS_PATH &&
4053 data->scheme_type == URL_SCHEME_FILE && data->host) {
4054 if(uri) {
4055 uri[length] = '\\';
4056 uri[length+1] = '\\';
4058 length += 2;
4062 if(data->username) {
4063 if(uri)
4064 memcpy(uri+length, data->username, data->username_len*sizeof(WCHAR));
4065 length += data->username_len;
4068 if(data->password) {
4069 if(uri) {
4070 uri[length] = ':';
4071 memcpy(uri+length+1, data->password, data->password_len*sizeof(WCHAR));
4073 length += data->password_len+1;
4076 if(data->password || data->username) {
4077 if(uri)
4078 uri[length] = '@';
4079 ++length;
4082 if(data->host) {
4083 /* IPv6 addresses get the brackets added around them if they don't already
4084 * have them.
4086 const BOOL add_brackets = data->host_type == Uri_HOST_IPV6 && *(data->host) != '[';
4087 if(add_brackets) {
4088 if(uri)
4089 uri[length] = '[';
4090 ++length;
4093 if(uri)
4094 memcpy(uri+length, data->host, data->host_len*sizeof(WCHAR));
4095 length += data->host_len;
4097 if(add_brackets) {
4098 if(uri)
4099 uri[length] = ']';
4100 length++;
4104 if(data->has_port) {
4105 /* The port isn't included in the raw uri if it's the default
4106 * port for the scheme type.
4108 DWORD i;
4109 BOOL is_default = FALSE;
4111 for(i = 0; i < sizeof(default_ports)/sizeof(default_ports[0]); ++i) {
4112 if(data->scheme_type == default_ports[i].scheme &&
4113 data->port_value == default_ports[i].port)
4114 is_default = TRUE;
4117 if(!is_default || flags & RAW_URI_FORCE_PORT_DISP) {
4118 if(uri)
4119 uri[length] = ':';
4120 ++length;
4122 if(uri)
4123 length += ui2str(uri+length, data->port_value);
4124 else
4125 length += ui2str(NULL, data->port_value);
4129 /* Check if a '/' should be added before the path for hierarchical URIs. */
4130 if(!data->is_opaque && data->path && *(data->path) != '/') {
4131 if(uri)
4132 uri[length] = '/';
4133 ++length;
4136 if(data->path) {
4137 if(!data->is_opaque && data->scheme_type == URL_SCHEME_FILE &&
4138 flags & RAW_URI_CONVERT_TO_DOS_PATH) {
4139 DWORD len = 0;
4141 if(uri)
4142 convert_to_dos_path(data->path, data->path_len, uri+length, &len);
4143 else
4144 convert_to_dos_path(data->path, data->path_len, NULL, &len);
4146 length += len;
4147 } else {
4148 if(uri)
4149 memcpy(uri+length, data->path, data->path_len*sizeof(WCHAR));
4150 length += data->path_len;
4154 if(data->query) {
4155 if(uri)
4156 memcpy(uri+length, data->query, data->query_len*sizeof(WCHAR));
4157 length += data->query_len;
4160 if(data->fragment) {
4161 if(uri)
4162 memcpy(uri+length, data->fragment, data->fragment_len*sizeof(WCHAR));
4163 length += data->fragment_len;
4166 if(uri)
4167 TRACE("(%p %p): Generated raw uri=%s len=%d\n", data, uri, debugstr_wn(uri, length), length);
4168 else
4169 TRACE("(%p %p): Computed raw uri len=%d\n", data, uri, length);
4171 return length;
4174 static HRESULT generate_uri(const UriBuilder *builder, const parse_data *data, Uri *uri, DWORD flags) {
4175 HRESULT hr;
4176 DWORD length = generate_raw_uri(data, NULL, 0);
4177 uri->raw_uri = SysAllocStringLen(NULL, length);
4178 if(!uri->raw_uri)
4179 return E_OUTOFMEMORY;
4181 generate_raw_uri(data, uri->raw_uri, 0);
4183 hr = canonicalize_uri(data, uri, flags);
4184 if(FAILED(hr)) {
4185 if(hr == E_INVALIDARG)
4186 return INET_E_INVALID_URL;
4187 return hr;
4190 uri->create_flags = flags;
4191 return S_OK;
4194 static inline Uri* impl_from_IUri(IUri *iface)
4196 return CONTAINING_RECORD(iface, Uri, IUri_iface);
4199 static inline void destroy_uri_obj(Uri *This)
4201 SysFreeString(This->raw_uri);
4202 heap_free(This->canon_uri);
4203 heap_free(This);
4206 static HRESULT WINAPI Uri_QueryInterface(IUri *iface, REFIID riid, void **ppv)
4208 Uri *This = impl_from_IUri(iface);
4210 if(IsEqualGUID(&IID_IUnknown, riid)) {
4211 TRACE("(%p)->(IID_IUnknown %p)\n", This, ppv);
4212 *ppv = &This->IUri_iface;
4213 }else if(IsEqualGUID(&IID_IUri, riid)) {
4214 TRACE("(%p)->(IID_IUri %p)\n", This, ppv);
4215 *ppv = &This->IUri_iface;
4216 }else if(IsEqualGUID(&IID_IUriBuilderFactory, riid)) {
4217 TRACE("(%p)->(IID_IUriBuilderFactory %p)\n", This, ppv);
4218 *ppv = &This->IUriBuilderFactory_iface;
4219 }else if(IsEqualGUID(&IID_IPersistStream, riid)) {
4220 TRACE("(%p)->(IID_IPersistStream %p)\n", This, ppv);
4221 *ppv = &This->IPersistStream_iface;
4222 }else if(IsEqualGUID(&IID_IMarshal, riid)) {
4223 TRACE("(%p)->(IID_IMarshal %p)\n", This, ppv);
4224 *ppv = &This->IMarshal_iface;
4225 }else if(IsEqualGUID(&IID_IUriObj, riid)) {
4226 TRACE("(%p)->(IID_IUriObj %p)\n", This, ppv);
4227 *ppv = This;
4228 return S_OK;
4229 }else {
4230 TRACE("(%p)->(%s %p)\n", This, debugstr_guid(riid), ppv);
4231 *ppv = NULL;
4232 return E_NOINTERFACE;
4235 IUnknown_AddRef((IUnknown*)*ppv);
4236 return S_OK;
4239 static ULONG WINAPI Uri_AddRef(IUri *iface)
4241 Uri *This = impl_from_IUri(iface);
4242 LONG ref = InterlockedIncrement(&This->ref);
4244 TRACE("(%p) ref=%d\n", This, ref);
4246 return ref;
4249 static ULONG WINAPI Uri_Release(IUri *iface)
4251 Uri *This = impl_from_IUri(iface);
4252 LONG ref = InterlockedDecrement(&This->ref);
4254 TRACE("(%p) ref=%d\n", This, ref);
4256 if(!ref)
4257 destroy_uri_obj(This);
4259 return ref;
4262 static HRESULT WINAPI Uri_GetPropertyBSTR(IUri *iface, Uri_PROPERTY uriProp, BSTR *pbstrProperty, DWORD dwFlags)
4264 Uri *This = impl_from_IUri(iface);
4265 HRESULT hres;
4266 TRACE("(%p %s)->(%d %p %x)\n", This, debugstr_w(This->canon_uri), uriProp, pbstrProperty, dwFlags);
4268 if(!This->create_flags)
4269 return E_UNEXPECTED;
4270 if(!pbstrProperty)
4271 return E_POINTER;
4273 if(uriProp > Uri_PROPERTY_STRING_LAST) {
4274 /* Windows allocates an empty BSTR for invalid Uri_PROPERTY's. */
4275 *pbstrProperty = SysAllocStringLen(NULL, 0);
4276 if(!(*pbstrProperty))
4277 return E_OUTOFMEMORY;
4279 /* It only returns S_FALSE for the ZONE property... */
4280 if(uriProp == Uri_PROPERTY_ZONE)
4281 return S_FALSE;
4282 else
4283 return S_OK;
4286 /* Don't have support for flags yet. */
4287 if(dwFlags) {
4288 FIXME("(%p)->(%d %p %x)\n", This, uriProp, pbstrProperty, dwFlags);
4289 return E_NOTIMPL;
4292 switch(uriProp) {
4293 case Uri_PROPERTY_ABSOLUTE_URI:
4294 if(This->display_modifiers & URI_DISPLAY_NO_ABSOLUTE_URI) {
4295 *pbstrProperty = SysAllocStringLen(NULL, 0);
4296 hres = S_FALSE;
4297 } else {
4298 if(This->scheme_type != URL_SCHEME_UNKNOWN && This->userinfo_start > -1) {
4299 if(This->userinfo_len == 0) {
4300 /* Don't include the '@' after the userinfo component. */
4301 *pbstrProperty = SysAllocStringLen(NULL, This->canon_len-1);
4302 hres = S_OK;
4303 if(*pbstrProperty) {
4304 /* Copy everything before it. */
4305 memcpy(*pbstrProperty, This->canon_uri, This->userinfo_start*sizeof(WCHAR));
4307 /* And everything after it. */
4308 memcpy(*pbstrProperty+This->userinfo_start, This->canon_uri+This->userinfo_start+1,
4309 (This->canon_len-This->userinfo_start-1)*sizeof(WCHAR));
4311 } else if(This->userinfo_split == 0 && This->userinfo_len == 1) {
4312 /* Don't include the ":@" */
4313 *pbstrProperty = SysAllocStringLen(NULL, This->canon_len-2);
4314 hres = S_OK;
4315 if(*pbstrProperty) {
4316 memcpy(*pbstrProperty, This->canon_uri, This->userinfo_start*sizeof(WCHAR));
4317 memcpy(*pbstrProperty+This->userinfo_start, This->canon_uri+This->userinfo_start+2,
4318 (This->canon_len-This->userinfo_start-2)*sizeof(WCHAR));
4320 } else {
4321 *pbstrProperty = SysAllocString(This->canon_uri);
4322 hres = S_OK;
4324 } else {
4325 *pbstrProperty = SysAllocString(This->canon_uri);
4326 hres = S_OK;
4330 if(!(*pbstrProperty))
4331 hres = E_OUTOFMEMORY;
4333 break;
4334 case Uri_PROPERTY_AUTHORITY:
4335 if(This->authority_start > -1) {
4336 if(This->port_offset > -1 && is_default_port(This->scheme_type, This->port) &&
4337 This->display_modifiers & URI_DISPLAY_NO_DEFAULT_PORT_AUTH)
4338 /* Don't include the port in the authority component. */
4339 *pbstrProperty = SysAllocStringLen(This->canon_uri+This->authority_start, This->port_offset);
4340 else
4341 *pbstrProperty = SysAllocStringLen(This->canon_uri+This->authority_start, This->authority_len);
4342 hres = S_OK;
4343 } else {
4344 *pbstrProperty = SysAllocStringLen(NULL, 0);
4345 hres = S_FALSE;
4348 if(!(*pbstrProperty))
4349 hres = E_OUTOFMEMORY;
4351 break;
4352 case Uri_PROPERTY_DISPLAY_URI:
4353 /* The Display URI contains everything except for the userinfo for known
4354 * scheme types.
4356 if(This->scheme_type != URL_SCHEME_UNKNOWN && This->userinfo_start > -1) {
4357 *pbstrProperty = SysAllocStringLen(NULL, This->canon_len-This->userinfo_len);
4359 if(*pbstrProperty) {
4360 /* Copy everything before the userinfo over. */
4361 memcpy(*pbstrProperty, This->canon_uri, This->userinfo_start*sizeof(WCHAR));
4362 /* Copy everything after the userinfo over. */
4363 memcpy(*pbstrProperty+This->userinfo_start,
4364 This->canon_uri+This->userinfo_start+This->userinfo_len+1,
4365 (This->canon_len-(This->userinfo_start+This->userinfo_len+1))*sizeof(WCHAR));
4367 } else
4368 *pbstrProperty = SysAllocString(This->canon_uri);
4370 if(!(*pbstrProperty))
4371 hres = E_OUTOFMEMORY;
4372 else
4373 hres = S_OK;
4375 break;
4376 case Uri_PROPERTY_DOMAIN:
4377 if(This->domain_offset > -1) {
4378 *pbstrProperty = SysAllocStringLen(This->canon_uri+This->host_start+This->domain_offset,
4379 This->host_len-This->domain_offset);
4380 hres = S_OK;
4381 } else {
4382 *pbstrProperty = SysAllocStringLen(NULL, 0);
4383 hres = S_FALSE;
4386 if(!(*pbstrProperty))
4387 hres = E_OUTOFMEMORY;
4389 break;
4390 case Uri_PROPERTY_EXTENSION:
4391 if(This->extension_offset > -1) {
4392 *pbstrProperty = SysAllocStringLen(This->canon_uri+This->path_start+This->extension_offset,
4393 This->path_len-This->extension_offset);
4394 hres = S_OK;
4395 } else {
4396 *pbstrProperty = SysAllocStringLen(NULL, 0);
4397 hres = S_FALSE;
4400 if(!(*pbstrProperty))
4401 hres = E_OUTOFMEMORY;
4403 break;
4404 case Uri_PROPERTY_FRAGMENT:
4405 if(This->fragment_start > -1) {
4406 *pbstrProperty = SysAllocStringLen(This->canon_uri+This->fragment_start, This->fragment_len);
4407 hres = S_OK;
4408 } else {
4409 *pbstrProperty = SysAllocStringLen(NULL, 0);
4410 hres = S_FALSE;
4413 if(!(*pbstrProperty))
4414 hres = E_OUTOFMEMORY;
4416 break;
4417 case Uri_PROPERTY_HOST:
4418 if(This->host_start > -1) {
4419 /* The '[' and ']' aren't included for IPv6 addresses. */
4420 if(This->host_type == Uri_HOST_IPV6)
4421 *pbstrProperty = SysAllocStringLen(This->canon_uri+This->host_start+1, This->host_len-2);
4422 else
4423 *pbstrProperty = SysAllocStringLen(This->canon_uri+This->host_start, This->host_len);
4425 hres = S_OK;
4426 } else {
4427 *pbstrProperty = SysAllocStringLen(NULL, 0);
4428 hres = S_FALSE;
4431 if(!(*pbstrProperty))
4432 hres = E_OUTOFMEMORY;
4434 break;
4435 case Uri_PROPERTY_PASSWORD:
4436 if(This->userinfo_split > -1) {
4437 *pbstrProperty = SysAllocStringLen(
4438 This->canon_uri+This->userinfo_start+This->userinfo_split+1,
4439 This->userinfo_len-This->userinfo_split-1);
4440 hres = S_OK;
4441 } else {
4442 *pbstrProperty = SysAllocStringLen(NULL, 0);
4443 hres = S_FALSE;
4446 if(!(*pbstrProperty))
4447 return E_OUTOFMEMORY;
4449 break;
4450 case Uri_PROPERTY_PATH:
4451 if(This->path_start > -1) {
4452 *pbstrProperty = SysAllocStringLen(This->canon_uri+This->path_start, This->path_len);
4453 hres = S_OK;
4454 } else {
4455 *pbstrProperty = SysAllocStringLen(NULL, 0);
4456 hres = S_FALSE;
4459 if(!(*pbstrProperty))
4460 hres = E_OUTOFMEMORY;
4462 break;
4463 case Uri_PROPERTY_PATH_AND_QUERY:
4464 if(This->path_start > -1) {
4465 *pbstrProperty = SysAllocStringLen(This->canon_uri+This->path_start, This->path_len+This->query_len);
4466 hres = S_OK;
4467 } else if(This->query_start > -1) {
4468 *pbstrProperty = SysAllocStringLen(This->canon_uri+This->query_start, This->query_len);
4469 hres = S_OK;
4470 } else {
4471 *pbstrProperty = SysAllocStringLen(NULL, 0);
4472 hres = S_FALSE;
4475 if(!(*pbstrProperty))
4476 hres = E_OUTOFMEMORY;
4478 break;
4479 case Uri_PROPERTY_QUERY:
4480 if(This->query_start > -1) {
4481 *pbstrProperty = SysAllocStringLen(This->canon_uri+This->query_start, This->query_len);
4482 hres = S_OK;
4483 } else {
4484 *pbstrProperty = SysAllocStringLen(NULL, 0);
4485 hres = S_FALSE;
4488 if(!(*pbstrProperty))
4489 hres = E_OUTOFMEMORY;
4491 break;
4492 case Uri_PROPERTY_RAW_URI:
4493 *pbstrProperty = SysAllocString(This->raw_uri);
4494 if(!(*pbstrProperty))
4495 hres = E_OUTOFMEMORY;
4496 else
4497 hres = S_OK;
4498 break;
4499 case Uri_PROPERTY_SCHEME_NAME:
4500 if(This->scheme_start > -1) {
4501 *pbstrProperty = SysAllocStringLen(This->canon_uri + This->scheme_start, This->scheme_len);
4502 hres = S_OK;
4503 } else {
4504 *pbstrProperty = SysAllocStringLen(NULL, 0);
4505 hres = S_FALSE;
4508 if(!(*pbstrProperty))
4509 hres = E_OUTOFMEMORY;
4511 break;
4512 case Uri_PROPERTY_USER_INFO:
4513 if(This->userinfo_start > -1) {
4514 *pbstrProperty = SysAllocStringLen(This->canon_uri+This->userinfo_start, This->userinfo_len);
4515 hres = S_OK;
4516 } else {
4517 *pbstrProperty = SysAllocStringLen(NULL, 0);
4518 hres = S_FALSE;
4521 if(!(*pbstrProperty))
4522 hres = E_OUTOFMEMORY;
4524 break;
4525 case Uri_PROPERTY_USER_NAME:
4526 if(This->userinfo_start > -1 && This->userinfo_split != 0) {
4527 /* If userinfo_split is set, that means a password exists
4528 * so the username is only from userinfo_start to userinfo_split.
4530 if(This->userinfo_split > -1) {
4531 *pbstrProperty = SysAllocStringLen(This->canon_uri + This->userinfo_start, This->userinfo_split);
4532 hres = S_OK;
4533 } else {
4534 *pbstrProperty = SysAllocStringLen(This->canon_uri + This->userinfo_start, This->userinfo_len);
4535 hres = S_OK;
4537 } else {
4538 *pbstrProperty = SysAllocStringLen(NULL, 0);
4539 hres = S_FALSE;
4542 if(!(*pbstrProperty))
4543 return E_OUTOFMEMORY;
4545 break;
4546 default:
4547 FIXME("(%p)->(%d %p %x)\n", This, uriProp, pbstrProperty, dwFlags);
4548 hres = E_NOTIMPL;
4551 return hres;
4554 static HRESULT WINAPI Uri_GetPropertyLength(IUri *iface, Uri_PROPERTY uriProp, DWORD *pcchProperty, DWORD dwFlags)
4556 Uri *This = impl_from_IUri(iface);
4557 HRESULT hres;
4558 TRACE("(%p %s)->(%d %p %x)\n", This, debugstr_w(This->canon_uri), uriProp, pcchProperty, dwFlags);
4560 if(!This->create_flags)
4561 return E_UNEXPECTED;
4562 if(!pcchProperty)
4563 return E_INVALIDARG;
4565 /* Can only return a length for a property if it's a string. */
4566 if(uriProp > Uri_PROPERTY_STRING_LAST)
4567 return E_INVALIDARG;
4569 /* Don't have support for flags yet. */
4570 if(dwFlags) {
4571 FIXME("(%p)->(%d %p %x)\n", This, uriProp, pcchProperty, dwFlags);
4572 return E_NOTIMPL;
4575 switch(uriProp) {
4576 case Uri_PROPERTY_ABSOLUTE_URI:
4577 if(This->display_modifiers & URI_DISPLAY_NO_ABSOLUTE_URI) {
4578 *pcchProperty = 0;
4579 hres = S_FALSE;
4580 } else {
4581 if(This->scheme_type != URL_SCHEME_UNKNOWN) {
4582 if(This->userinfo_start > -1 && This->userinfo_len == 0)
4583 /* Don't include the '@' in the length. */
4584 *pcchProperty = This->canon_len-1;
4585 else if(This->userinfo_start > -1 && This->userinfo_len == 1 &&
4586 This->userinfo_split == 0)
4587 /* Don't include the ":@" in the length. */
4588 *pcchProperty = This->canon_len-2;
4589 else
4590 *pcchProperty = This->canon_len;
4591 } else
4592 *pcchProperty = This->canon_len;
4594 hres = S_OK;
4597 break;
4598 case Uri_PROPERTY_AUTHORITY:
4599 if(This->port_offset > -1 &&
4600 This->display_modifiers & URI_DISPLAY_NO_DEFAULT_PORT_AUTH &&
4601 is_default_port(This->scheme_type, This->port))
4602 /* Only count up until the port in the authority. */
4603 *pcchProperty = This->port_offset;
4604 else
4605 *pcchProperty = This->authority_len;
4606 hres = (This->authority_start > -1) ? S_OK : S_FALSE;
4607 break;
4608 case Uri_PROPERTY_DISPLAY_URI:
4609 if(This->scheme_type != URL_SCHEME_UNKNOWN && This->userinfo_start > -1)
4610 *pcchProperty = This->canon_len-This->userinfo_len-1;
4611 else
4612 *pcchProperty = This->canon_len;
4614 hres = S_OK;
4615 break;
4616 case Uri_PROPERTY_DOMAIN:
4617 if(This->domain_offset > -1)
4618 *pcchProperty = This->host_len - This->domain_offset;
4619 else
4620 *pcchProperty = 0;
4622 hres = (This->domain_offset > -1) ? S_OK : S_FALSE;
4623 break;
4624 case Uri_PROPERTY_EXTENSION:
4625 if(This->extension_offset > -1) {
4626 *pcchProperty = This->path_len - This->extension_offset;
4627 hres = S_OK;
4628 } else {
4629 *pcchProperty = 0;
4630 hres = S_FALSE;
4633 break;
4634 case Uri_PROPERTY_FRAGMENT:
4635 *pcchProperty = This->fragment_len;
4636 hres = (This->fragment_start > -1) ? S_OK : S_FALSE;
4637 break;
4638 case Uri_PROPERTY_HOST:
4639 *pcchProperty = This->host_len;
4641 /* '[' and ']' aren't included in the length. */
4642 if(This->host_type == Uri_HOST_IPV6)
4643 *pcchProperty -= 2;
4645 hres = (This->host_start > -1) ? S_OK : S_FALSE;
4646 break;
4647 case Uri_PROPERTY_PASSWORD:
4648 *pcchProperty = (This->userinfo_split > -1) ? This->userinfo_len-This->userinfo_split-1 : 0;
4649 hres = (This->userinfo_split > -1) ? S_OK : S_FALSE;
4650 break;
4651 case Uri_PROPERTY_PATH:
4652 *pcchProperty = This->path_len;
4653 hres = (This->path_start > -1) ? S_OK : S_FALSE;
4654 break;
4655 case Uri_PROPERTY_PATH_AND_QUERY:
4656 *pcchProperty = This->path_len+This->query_len;
4657 hres = (This->path_start > -1 || This->query_start > -1) ? S_OK : S_FALSE;
4658 break;
4659 case Uri_PROPERTY_QUERY:
4660 *pcchProperty = This->query_len;
4661 hres = (This->query_start > -1) ? S_OK : S_FALSE;
4662 break;
4663 case Uri_PROPERTY_RAW_URI:
4664 *pcchProperty = SysStringLen(This->raw_uri);
4665 hres = S_OK;
4666 break;
4667 case Uri_PROPERTY_SCHEME_NAME:
4668 *pcchProperty = This->scheme_len;
4669 hres = (This->scheme_start > -1) ? S_OK : S_FALSE;
4670 break;
4671 case Uri_PROPERTY_USER_INFO:
4672 *pcchProperty = This->userinfo_len;
4673 hres = (This->userinfo_start > -1) ? S_OK : S_FALSE;
4674 break;
4675 case Uri_PROPERTY_USER_NAME:
4676 *pcchProperty = (This->userinfo_split > -1) ? This->userinfo_split : This->userinfo_len;
4677 if(This->userinfo_split == 0)
4678 hres = S_FALSE;
4679 else
4680 hres = (This->userinfo_start > -1) ? S_OK : S_FALSE;
4681 break;
4682 default:
4683 FIXME("(%p)->(%d %p %x)\n", This, uriProp, pcchProperty, dwFlags);
4684 hres = E_NOTIMPL;
4687 return hres;
4690 static HRESULT WINAPI Uri_GetPropertyDWORD(IUri *iface, Uri_PROPERTY uriProp, DWORD *pcchProperty, DWORD dwFlags)
4692 Uri *This = impl_from_IUri(iface);
4693 HRESULT hres;
4695 TRACE("(%p %s)->(%d %p %x)\n", This, debugstr_w(This->canon_uri), uriProp, pcchProperty, dwFlags);
4697 if(!This->create_flags)
4698 return E_UNEXPECTED;
4699 if(!pcchProperty)
4700 return E_INVALIDARG;
4702 /* Microsoft's implementation for the ZONE property of a URI seems to be lacking...
4703 * From what I can tell, instead of checking which URLZONE the URI belongs to it
4704 * simply assigns URLZONE_INVALID and returns E_NOTIMPL. This also applies to the GetZone
4705 * function.
4707 if(uriProp == Uri_PROPERTY_ZONE) {
4708 *pcchProperty = URLZONE_INVALID;
4709 return E_NOTIMPL;
4712 if(uriProp < Uri_PROPERTY_DWORD_START) {
4713 *pcchProperty = 0;
4714 return E_INVALIDARG;
4717 switch(uriProp) {
4718 case Uri_PROPERTY_HOST_TYPE:
4719 *pcchProperty = This->host_type;
4720 hres = S_OK;
4721 break;
4722 case Uri_PROPERTY_PORT:
4723 if(!This->has_port) {
4724 *pcchProperty = 0;
4725 hres = S_FALSE;
4726 } else {
4727 *pcchProperty = This->port;
4728 hres = S_OK;
4731 break;
4732 case Uri_PROPERTY_SCHEME:
4733 *pcchProperty = This->scheme_type;
4734 hres = S_OK;
4735 break;
4736 default:
4737 FIXME("(%p)->(%d %p %x)\n", This, uriProp, pcchProperty, dwFlags);
4738 hres = E_NOTIMPL;
4741 return hres;
4744 static HRESULT WINAPI Uri_HasProperty(IUri *iface, Uri_PROPERTY uriProp, BOOL *pfHasProperty)
4746 Uri *This = impl_from_IUri(iface);
4748 TRACE("(%p %s)->(%d %p)\n", This, debugstr_w(This->canon_uri), uriProp, pfHasProperty);
4750 if(!pfHasProperty)
4751 return E_INVALIDARG;
4753 switch(uriProp) {
4754 case Uri_PROPERTY_ABSOLUTE_URI:
4755 *pfHasProperty = !(This->display_modifiers & URI_DISPLAY_NO_ABSOLUTE_URI);
4756 break;
4757 case Uri_PROPERTY_AUTHORITY:
4758 *pfHasProperty = This->authority_start > -1;
4759 break;
4760 case Uri_PROPERTY_DISPLAY_URI:
4761 *pfHasProperty = TRUE;
4762 break;
4763 case Uri_PROPERTY_DOMAIN:
4764 *pfHasProperty = This->domain_offset > -1;
4765 break;
4766 case Uri_PROPERTY_EXTENSION:
4767 *pfHasProperty = This->extension_offset > -1;
4768 break;
4769 case Uri_PROPERTY_FRAGMENT:
4770 *pfHasProperty = This->fragment_start > -1;
4771 break;
4772 case Uri_PROPERTY_HOST:
4773 *pfHasProperty = This->host_start > -1;
4774 break;
4775 case Uri_PROPERTY_PASSWORD:
4776 *pfHasProperty = This->userinfo_split > -1;
4777 break;
4778 case Uri_PROPERTY_PATH:
4779 *pfHasProperty = This->path_start > -1;
4780 break;
4781 case Uri_PROPERTY_PATH_AND_QUERY:
4782 *pfHasProperty = (This->path_start > -1 || This->query_start > -1);
4783 break;
4784 case Uri_PROPERTY_QUERY:
4785 *pfHasProperty = This->query_start > -1;
4786 break;
4787 case Uri_PROPERTY_RAW_URI:
4788 *pfHasProperty = TRUE;
4789 break;
4790 case Uri_PROPERTY_SCHEME_NAME:
4791 *pfHasProperty = This->scheme_start > -1;
4792 break;
4793 case Uri_PROPERTY_USER_INFO:
4794 *pfHasProperty = This->userinfo_start > -1;
4795 break;
4796 case Uri_PROPERTY_USER_NAME:
4797 if(This->userinfo_split == 0)
4798 *pfHasProperty = FALSE;
4799 else
4800 *pfHasProperty = This->userinfo_start > -1;
4801 break;
4802 case Uri_PROPERTY_HOST_TYPE:
4803 *pfHasProperty = TRUE;
4804 break;
4805 case Uri_PROPERTY_PORT:
4806 *pfHasProperty = This->has_port;
4807 break;
4808 case Uri_PROPERTY_SCHEME:
4809 *pfHasProperty = TRUE;
4810 break;
4811 case Uri_PROPERTY_ZONE:
4812 *pfHasProperty = FALSE;
4813 break;
4814 default:
4815 FIXME("(%p)->(%d %p): Unsupported property type.\n", This, uriProp, pfHasProperty);
4816 return E_NOTIMPL;
4819 return S_OK;
4822 static HRESULT WINAPI Uri_GetAbsoluteUri(IUri *iface, BSTR *pstrAbsoluteUri)
4824 TRACE("(%p)->(%p)\n", iface, pstrAbsoluteUri);
4825 return IUri_GetPropertyBSTR(iface, Uri_PROPERTY_ABSOLUTE_URI, pstrAbsoluteUri, 0);
4828 static HRESULT WINAPI Uri_GetAuthority(IUri *iface, BSTR *pstrAuthority)
4830 TRACE("(%p)->(%p)\n", iface, pstrAuthority);
4831 return IUri_GetPropertyBSTR(iface, Uri_PROPERTY_AUTHORITY, pstrAuthority, 0);
4834 static HRESULT WINAPI Uri_GetDisplayUri(IUri *iface, BSTR *pstrDisplayUri)
4836 TRACE("(%p)->(%p)\n", iface, pstrDisplayUri);
4837 return IUri_GetPropertyBSTR(iface, Uri_PROPERTY_DISPLAY_URI, pstrDisplayUri, 0);
4840 static HRESULT WINAPI Uri_GetDomain(IUri *iface, BSTR *pstrDomain)
4842 TRACE("(%p)->(%p)\n", iface, pstrDomain);
4843 return IUri_GetPropertyBSTR(iface, Uri_PROPERTY_DOMAIN, pstrDomain, 0);
4846 static HRESULT WINAPI Uri_GetExtension(IUri *iface, BSTR *pstrExtension)
4848 TRACE("(%p)->(%p)\n", iface, pstrExtension);
4849 return IUri_GetPropertyBSTR(iface, Uri_PROPERTY_EXTENSION, pstrExtension, 0);
4852 static HRESULT WINAPI Uri_GetFragment(IUri *iface, BSTR *pstrFragment)
4854 TRACE("(%p)->(%p)\n", iface, pstrFragment);
4855 return IUri_GetPropertyBSTR(iface, Uri_PROPERTY_FRAGMENT, pstrFragment, 0);
4858 static HRESULT WINAPI Uri_GetHost(IUri *iface, BSTR *pstrHost)
4860 TRACE("(%p)->(%p)\n", iface, pstrHost);
4861 return IUri_GetPropertyBSTR(iface, Uri_PROPERTY_HOST, pstrHost, 0);
4864 static HRESULT WINAPI Uri_GetPassword(IUri *iface, BSTR *pstrPassword)
4866 TRACE("(%p)->(%p)\n", iface, pstrPassword);
4867 return IUri_GetPropertyBSTR(iface, Uri_PROPERTY_PASSWORD, pstrPassword, 0);
4870 static HRESULT WINAPI Uri_GetPath(IUri *iface, BSTR *pstrPath)
4872 TRACE("(%p)->(%p)\n", iface, pstrPath);
4873 return IUri_GetPropertyBSTR(iface, Uri_PROPERTY_PATH, pstrPath, 0);
4876 static HRESULT WINAPI Uri_GetPathAndQuery(IUri *iface, BSTR *pstrPathAndQuery)
4878 TRACE("(%p)->(%p)\n", iface, pstrPathAndQuery);
4879 return IUri_GetPropertyBSTR(iface, Uri_PROPERTY_PATH_AND_QUERY, pstrPathAndQuery, 0);
4882 static HRESULT WINAPI Uri_GetQuery(IUri *iface, BSTR *pstrQuery)
4884 TRACE("(%p)->(%p)\n", iface, pstrQuery);
4885 return IUri_GetPropertyBSTR(iface, Uri_PROPERTY_QUERY, pstrQuery, 0);
4888 static HRESULT WINAPI Uri_GetRawUri(IUri *iface, BSTR *pstrRawUri)
4890 TRACE("(%p)->(%p)\n", iface, pstrRawUri);
4891 return IUri_GetPropertyBSTR(iface, Uri_PROPERTY_RAW_URI, pstrRawUri, 0);
4894 static HRESULT WINAPI Uri_GetSchemeName(IUri *iface, BSTR *pstrSchemeName)
4896 TRACE("(%p)->(%p)\n", iface, pstrSchemeName);
4897 return IUri_GetPropertyBSTR(iface, Uri_PROPERTY_SCHEME_NAME, pstrSchemeName, 0);
4900 static HRESULT WINAPI Uri_GetUserInfo(IUri *iface, BSTR *pstrUserInfo)
4902 TRACE("(%p)->(%p)\n", iface, pstrUserInfo);
4903 return IUri_GetPropertyBSTR(iface, Uri_PROPERTY_USER_INFO, pstrUserInfo, 0);
4906 static HRESULT WINAPI Uri_GetUserName(IUri *iface, BSTR *pstrUserName)
4908 TRACE("(%p)->(%p)\n", iface, pstrUserName);
4909 return IUri_GetPropertyBSTR(iface, Uri_PROPERTY_USER_NAME, pstrUserName, 0);
4912 static HRESULT WINAPI Uri_GetHostType(IUri *iface, DWORD *pdwHostType)
4914 TRACE("(%p)->(%p)\n", iface, pdwHostType);
4915 return IUri_GetPropertyDWORD(iface, Uri_PROPERTY_HOST_TYPE, pdwHostType, 0);
4918 static HRESULT WINAPI Uri_GetPort(IUri *iface, DWORD *pdwPort)
4920 TRACE("(%p)->(%p)\n", iface, pdwPort);
4921 return IUri_GetPropertyDWORD(iface, Uri_PROPERTY_PORT, pdwPort, 0);
4924 static HRESULT WINAPI Uri_GetScheme(IUri *iface, DWORD *pdwScheme)
4926 TRACE("(%p)->(%p)\n", iface, pdwScheme);
4927 return IUri_GetPropertyDWORD(iface, Uri_PROPERTY_SCHEME, pdwScheme, 0);
4930 static HRESULT WINAPI Uri_GetZone(IUri *iface, DWORD *pdwZone)
4932 TRACE("(%p)->(%p)\n", iface, pdwZone);
4933 return IUri_GetPropertyDWORD(iface, Uri_PROPERTY_ZONE,pdwZone, 0);
4936 static HRESULT WINAPI Uri_GetProperties(IUri *iface, DWORD *pdwProperties)
4938 Uri *This = impl_from_IUri(iface);
4939 TRACE("(%p %s)->(%p)\n", This, debugstr_w(This->canon_uri), pdwProperties);
4941 if(!This->create_flags)
4942 return E_UNEXPECTED;
4943 if(!pdwProperties)
4944 return E_INVALIDARG;
4946 /* All URIs have these. */
4947 *pdwProperties = Uri_HAS_DISPLAY_URI|Uri_HAS_RAW_URI|Uri_HAS_SCHEME|Uri_HAS_HOST_TYPE;
4949 if(!(This->display_modifiers & URI_DISPLAY_NO_ABSOLUTE_URI))
4950 *pdwProperties |= Uri_HAS_ABSOLUTE_URI;
4952 if(This->scheme_start > -1)
4953 *pdwProperties |= Uri_HAS_SCHEME_NAME;
4955 if(This->authority_start > -1) {
4956 *pdwProperties |= Uri_HAS_AUTHORITY;
4957 if(This->userinfo_start > -1) {
4958 *pdwProperties |= Uri_HAS_USER_INFO;
4959 if(This->userinfo_split != 0)
4960 *pdwProperties |= Uri_HAS_USER_NAME;
4962 if(This->userinfo_split > -1)
4963 *pdwProperties |= Uri_HAS_PASSWORD;
4964 if(This->host_start > -1)
4965 *pdwProperties |= Uri_HAS_HOST;
4966 if(This->domain_offset > -1)
4967 *pdwProperties |= Uri_HAS_DOMAIN;
4970 if(This->has_port)
4971 *pdwProperties |= Uri_HAS_PORT;
4972 if(This->path_start > -1)
4973 *pdwProperties |= Uri_HAS_PATH|Uri_HAS_PATH_AND_QUERY;
4974 if(This->query_start > -1)
4975 *pdwProperties |= Uri_HAS_QUERY|Uri_HAS_PATH_AND_QUERY;
4977 if(This->extension_offset > -1)
4978 *pdwProperties |= Uri_HAS_EXTENSION;
4980 if(This->fragment_start > -1)
4981 *pdwProperties |= Uri_HAS_FRAGMENT;
4983 return S_OK;
4986 static HRESULT WINAPI Uri_IsEqual(IUri *iface, IUri *pUri, BOOL *pfEqual)
4988 Uri *This = impl_from_IUri(iface);
4989 Uri *other;
4991 TRACE("(%p %s)->(%p %p)\n", This, debugstr_w(This->canon_uri), pUri, pfEqual);
4993 if(!This->create_flags)
4994 return E_UNEXPECTED;
4995 if(!pfEqual)
4996 return E_POINTER;
4998 if(!pUri) {
4999 *pfEqual = FALSE;
5001 /* For some reason Windows returns S_OK here... */
5002 return S_OK;
5005 /* Try to convert it to a Uri (allows for a more simple comparison). */
5006 if(!(other = get_uri_obj(pUri))) {
5007 FIXME("(%p)->(%p %p) No support for unknown IUri's yet.\n", iface, pUri, pfEqual);
5008 return E_NOTIMPL;
5011 TRACE("comparing to %s\n", debugstr_w(other->canon_uri));
5012 return compare_uris(This, other, pfEqual);
5015 static const IUriVtbl UriVtbl = {
5016 Uri_QueryInterface,
5017 Uri_AddRef,
5018 Uri_Release,
5019 Uri_GetPropertyBSTR,
5020 Uri_GetPropertyLength,
5021 Uri_GetPropertyDWORD,
5022 Uri_HasProperty,
5023 Uri_GetAbsoluteUri,
5024 Uri_GetAuthority,
5025 Uri_GetDisplayUri,
5026 Uri_GetDomain,
5027 Uri_GetExtension,
5028 Uri_GetFragment,
5029 Uri_GetHost,
5030 Uri_GetPassword,
5031 Uri_GetPath,
5032 Uri_GetPathAndQuery,
5033 Uri_GetQuery,
5034 Uri_GetRawUri,
5035 Uri_GetSchemeName,
5036 Uri_GetUserInfo,
5037 Uri_GetUserName,
5038 Uri_GetHostType,
5039 Uri_GetPort,
5040 Uri_GetScheme,
5041 Uri_GetZone,
5042 Uri_GetProperties,
5043 Uri_IsEqual
5046 static inline Uri* impl_from_IUriBuilderFactory(IUriBuilderFactory *iface)
5048 return CONTAINING_RECORD(iface, Uri, IUriBuilderFactory_iface);
5051 static HRESULT WINAPI UriBuilderFactory_QueryInterface(IUriBuilderFactory *iface, REFIID riid, void **ppv)
5053 Uri *This = impl_from_IUriBuilderFactory(iface);
5054 return IUri_QueryInterface(&This->IUri_iface, riid, ppv);
5057 static ULONG WINAPI UriBuilderFactory_AddRef(IUriBuilderFactory *iface)
5059 Uri *This = impl_from_IUriBuilderFactory(iface);
5060 return IUri_AddRef(&This->IUri_iface);
5063 static ULONG WINAPI UriBuilderFactory_Release(IUriBuilderFactory *iface)
5065 Uri *This = impl_from_IUriBuilderFactory(iface);
5066 return IUri_Release(&This->IUri_iface);
5069 static HRESULT WINAPI UriBuilderFactory_CreateIUriBuilder(IUriBuilderFactory *iface,
5070 DWORD dwFlags,
5071 DWORD_PTR dwReserved,
5072 IUriBuilder **ppIUriBuilder)
5074 Uri *This = impl_from_IUriBuilderFactory(iface);
5075 TRACE("(%p)->(%08x %08x %p)\n", This, dwFlags, (DWORD)dwReserved, ppIUriBuilder);
5077 if(!ppIUriBuilder)
5078 return E_POINTER;
5080 if(dwFlags || dwReserved) {
5081 *ppIUriBuilder = NULL;
5082 return E_INVALIDARG;
5085 return CreateIUriBuilder(NULL, 0, 0, ppIUriBuilder);
5088 static HRESULT WINAPI UriBuilderFactory_CreateInitializedIUriBuilder(IUriBuilderFactory *iface,
5089 DWORD dwFlags,
5090 DWORD_PTR dwReserved,
5091 IUriBuilder **ppIUriBuilder)
5093 Uri *This = impl_from_IUriBuilderFactory(iface);
5094 TRACE("(%p)->(%08x %08x %p)\n", This, dwFlags, (DWORD)dwReserved, ppIUriBuilder);
5096 if(!ppIUriBuilder)
5097 return E_POINTER;
5099 if(dwFlags || dwReserved) {
5100 *ppIUriBuilder = NULL;
5101 return E_INVALIDARG;
5104 return CreateIUriBuilder(&This->IUri_iface, 0, 0, ppIUriBuilder);
5107 static const IUriBuilderFactoryVtbl UriBuilderFactoryVtbl = {
5108 UriBuilderFactory_QueryInterface,
5109 UriBuilderFactory_AddRef,
5110 UriBuilderFactory_Release,
5111 UriBuilderFactory_CreateIUriBuilder,
5112 UriBuilderFactory_CreateInitializedIUriBuilder
5115 static inline Uri* impl_from_IPersistStream(IPersistStream *iface)
5117 return CONTAINING_RECORD(iface, Uri, IPersistStream_iface);
5120 static HRESULT WINAPI PersistStream_QueryInterface(IPersistStream *iface, REFIID riid, void **ppvObject)
5122 Uri *This = impl_from_IPersistStream(iface);
5123 return IUri_QueryInterface(&This->IUri_iface, riid, ppvObject);
5126 static ULONG WINAPI PersistStream_AddRef(IPersistStream *iface)
5128 Uri *This = impl_from_IPersistStream(iface);
5129 return IUri_AddRef(&This->IUri_iface);
5132 static ULONG WINAPI PersistStream_Release(IPersistStream *iface)
5134 Uri *This = impl_from_IPersistStream(iface);
5135 return IUri_Release(&This->IUri_iface);
5138 static HRESULT WINAPI PersistStream_GetClassID(IPersistStream *iface, CLSID *pClassID)
5140 Uri *This = impl_from_IPersistStream(iface);
5141 TRACE("(%p)->(%p)\n", This, pClassID);
5143 if(!pClassID)
5144 return E_INVALIDARG;
5146 *pClassID = CLSID_CUri;
5147 return S_OK;
5150 static HRESULT WINAPI PersistStream_IsDirty(IPersistStream *iface)
5152 Uri *This = impl_from_IPersistStream(iface);
5153 TRACE("(%p)\n", This);
5154 return S_FALSE;
5157 struct persist_uri {
5158 DWORD size;
5159 DWORD unk1[2];
5160 DWORD create_flags;
5161 DWORD unk2[3];
5162 DWORD fields_no;
5163 BYTE data[1];
5166 static HRESULT WINAPI PersistStream_Load(IPersistStream *iface, IStream *pStm)
5168 Uri *This = impl_from_IPersistStream(iface);
5169 struct persist_uri *data;
5170 parse_data parse;
5171 DWORD size;
5172 HRESULT hr;
5174 TRACE("(%p)->(%p)\n", This, pStm);
5176 if(This->create_flags)
5177 return E_UNEXPECTED;
5178 if(!pStm)
5179 return E_INVALIDARG;
5181 hr = IStream_Read(pStm, &size, sizeof(DWORD), NULL);
5182 if(FAILED(hr))
5183 return hr;
5184 data = heap_alloc(size);
5185 if(!data)
5186 return E_OUTOFMEMORY;
5187 hr = IStream_Read(pStm, &data->unk1, size-sizeof(DWORD)-2, NULL);
5188 if(FAILED(hr))
5189 return hr;
5191 if(size < sizeof(struct persist_uri)) {
5192 heap_free(data);
5193 return S_OK;
5196 if(*(DWORD*)data->data != Uri_PROPERTY_RAW_URI) {
5197 heap_free(data);
5198 ERR("Can't find raw_uri\n");
5199 return E_UNEXPECTED;
5202 This->raw_uri = SysAllocString((WCHAR*)(data->data+sizeof(DWORD)*2));
5203 if(!This->raw_uri) {
5204 heap_free(data);
5205 return E_OUTOFMEMORY;
5207 This->create_flags = data->create_flags;
5208 heap_free(data);
5209 TRACE("%x %s\n", This->create_flags, debugstr_w(This->raw_uri));
5211 memset(&parse, 0, sizeof(parse_data));
5212 parse.uri = This->raw_uri;
5213 if(!parse_uri(&parse, This->create_flags)) {
5214 SysFreeString(This->raw_uri);
5215 This->create_flags = 0;
5216 return E_UNEXPECTED;
5219 hr = canonicalize_uri(&parse, This, This->create_flags);
5220 if(FAILED(hr)) {
5221 SysFreeString(This->raw_uri);
5222 This->create_flags = 0;
5223 return hr;
5226 return S_OK;
5229 static inline BYTE* persist_stream_add_strprop(Uri *This, BYTE *p, DWORD type, DWORD len, WCHAR *data)
5231 len *= sizeof(WCHAR);
5232 *(DWORD*)p = type;
5233 p += sizeof(DWORD);
5234 *(DWORD*)p = len+sizeof(WCHAR);
5235 p += sizeof(DWORD);
5236 memcpy(p, data, len);
5237 p += len;
5238 *(WCHAR*)p = 0;
5239 return p+sizeof(WCHAR);
5242 static inline void persist_stream_save(Uri *This, IStream *pStm, BOOL marshal, struct persist_uri *data)
5244 BYTE *p = NULL;
5246 data->create_flags = This->create_flags;
5248 if(This->create_flags) {
5249 data->fields_no = 1;
5250 p = persist_stream_add_strprop(This, data->data, Uri_PROPERTY_RAW_URI,
5251 SysStringLen(This->raw_uri), This->raw_uri);
5253 if(This->scheme_type!=URL_SCHEME_HTTP && This->scheme_type!=URL_SCHEME_HTTPS
5254 && This->scheme_type!=URL_SCHEME_FTP)
5255 return;
5257 if(This->fragment_len) {
5258 data->fields_no++;
5259 p = persist_stream_add_strprop(This, p, Uri_PROPERTY_FRAGMENT,
5260 This->fragment_len, This->canon_uri+This->fragment_start);
5263 if(This->host_len) {
5264 data->fields_no++;
5265 if(This->host_type == Uri_HOST_IPV6)
5266 p = persist_stream_add_strprop(This, p, Uri_PROPERTY_HOST,
5267 This->host_len-2, This->canon_uri+This->host_start+1);
5268 else
5269 p = persist_stream_add_strprop(This, p, Uri_PROPERTY_HOST,
5270 This->host_len, This->canon_uri+This->host_start);
5273 if(This->userinfo_split > -1) {
5274 data->fields_no++;
5275 p = persist_stream_add_strprop(This, p, Uri_PROPERTY_PASSWORD,
5276 This->userinfo_len-This->userinfo_split-1,
5277 This->canon_uri+This->userinfo_start+This->userinfo_split+1);
5280 if(This->path_len) {
5281 data->fields_no++;
5282 p = persist_stream_add_strprop(This, p, Uri_PROPERTY_PATH,
5283 This->path_len, This->canon_uri+This->path_start);
5284 } else if(marshal) {
5285 WCHAR no_path = '/';
5286 data->fields_no++;
5287 p = persist_stream_add_strprop(This, p, Uri_PROPERTY_PATH, 1, &no_path);
5290 if(This->has_port) {
5291 data->fields_no++;
5292 *(DWORD*)p = Uri_PROPERTY_PORT;
5293 p += sizeof(DWORD);
5294 *(DWORD*)p = sizeof(DWORD);
5295 p += sizeof(DWORD);
5296 *(DWORD*)p = This->port;
5297 p += sizeof(DWORD);
5300 if(This->query_len) {
5301 data->fields_no++;
5302 p = persist_stream_add_strprop(This, p, Uri_PROPERTY_QUERY,
5303 This->query_len, This->canon_uri+This->query_start);
5306 if(This->scheme_len) {
5307 data->fields_no++;
5308 p = persist_stream_add_strprop(This, p, Uri_PROPERTY_SCHEME_NAME,
5309 This->scheme_len, This->canon_uri+This->scheme_start);
5312 if(This->userinfo_start>-1 && This->userinfo_split!=0) {
5313 data->fields_no++;
5314 if(This->userinfo_split > -1)
5315 p = persist_stream_add_strprop(This, p, Uri_PROPERTY_USER_NAME,
5316 This->userinfo_split, This->canon_uri+This->userinfo_start);
5317 else
5318 p = persist_stream_add_strprop(This, p, Uri_PROPERTY_USER_NAME,
5319 This->userinfo_len, This->canon_uri+This->userinfo_start);
5323 static HRESULT WINAPI PersistStream_Save(IPersistStream *iface, IStream *pStm, BOOL fClearDirty)
5325 Uri *This = impl_from_IPersistStream(iface);
5326 struct persist_uri *data;
5327 ULARGE_INTEGER size;
5328 HRESULT hres;
5330 TRACE("(%p)->(%p %x)\n", This, pStm, fClearDirty);
5332 if(!pStm)
5333 return E_INVALIDARG;
5335 hres = IPersistStream_GetSizeMax(&This->IPersistStream_iface, &size);
5336 if(FAILED(hres))
5337 return hres;
5339 data = heap_alloc_zero(size.u.LowPart);
5340 if(!data)
5341 return E_OUTOFMEMORY;
5342 data->size = size.u.LowPart;
5343 persist_stream_save(This, pStm, FALSE, data);
5345 hres = IStream_Write(pStm, data, data->size-2, NULL);
5346 heap_free(data);
5347 return hres;
5350 static HRESULT WINAPI PersistStream_GetSizeMax(IPersistStream *iface, ULARGE_INTEGER *pcbSize)
5352 Uri *This = impl_from_IPersistStream(iface);
5353 TRACE("(%p)->(%p)\n", This, pcbSize);
5355 if(!pcbSize)
5356 return E_INVALIDARG;
5358 pcbSize->u.LowPart = 2+sizeof(struct persist_uri);
5359 pcbSize->u.HighPart = 0;
5360 if(This->create_flags)
5361 pcbSize->u.LowPart += (SysStringLen(This->raw_uri)+1)*sizeof(WCHAR) + 2*sizeof(DWORD);
5362 else /* there's no place for fields no */
5363 pcbSize->u.LowPart -= sizeof(DWORD);
5364 if(This->scheme_type!=URL_SCHEME_HTTP && This->scheme_type!=URL_SCHEME_HTTPS
5365 && This->scheme_type!=URL_SCHEME_FTP)
5366 return S_OK;
5368 if(This->fragment_len)
5369 pcbSize->u.LowPart += (This->fragment_len+1)*sizeof(WCHAR) + 2*sizeof(DWORD);
5370 if(This->host_len) {
5371 if(This->host_type == Uri_HOST_IPV6)
5372 pcbSize->u.LowPart += (This->host_len-1)*sizeof(WCHAR) + 2*sizeof(DWORD);
5373 else
5374 pcbSize->u.LowPart += (This->host_len+1)*sizeof(WCHAR) + 2*sizeof(DWORD);
5376 if(This->userinfo_split > -1)
5377 pcbSize->u.LowPart += (This->userinfo_len-This->userinfo_split)*sizeof(WCHAR) + 2*sizeof(DWORD);
5378 if(This->path_len)
5379 pcbSize->u.LowPart += (This->path_len+1)*sizeof(WCHAR) + 2*sizeof(DWORD);
5380 if(This->has_port)
5381 pcbSize->u.LowPart += 3*sizeof(DWORD);
5382 if(This->query_len)
5383 pcbSize->u.LowPart += (This->query_len+1)*sizeof(WCHAR) + 2*sizeof(DWORD);
5384 if(This->scheme_len)
5385 pcbSize->u.LowPart += (This->scheme_len+1)*sizeof(WCHAR) + 2*sizeof(DWORD);
5386 if(This->userinfo_start>-1 && This->userinfo_split!=0) {
5387 if(This->userinfo_split > -1)
5388 pcbSize->u.LowPart += (This->userinfo_split+1)*sizeof(WCHAR) + 2*sizeof(DWORD);
5389 else
5390 pcbSize->u.LowPart += (This->userinfo_len+1)*sizeof(WCHAR) + 2*sizeof(DWORD);
5392 return S_OK;
5395 static const IPersistStreamVtbl PersistStreamVtbl = {
5396 PersistStream_QueryInterface,
5397 PersistStream_AddRef,
5398 PersistStream_Release,
5399 PersistStream_GetClassID,
5400 PersistStream_IsDirty,
5401 PersistStream_Load,
5402 PersistStream_Save,
5403 PersistStream_GetSizeMax
5406 static inline Uri* impl_from_IMarshal(IMarshal *iface)
5408 return CONTAINING_RECORD(iface, Uri, IMarshal_iface);
5411 static HRESULT WINAPI Marshal_QueryInterface(IMarshal *iface, REFIID riid, void **ppvObject)
5413 Uri *This = impl_from_IMarshal(iface);
5414 return IUri_QueryInterface(&This->IUri_iface, riid, ppvObject);
5417 static ULONG WINAPI Marshal_AddRef(IMarshal *iface)
5419 Uri *This = impl_from_IMarshal(iface);
5420 return IUri_AddRef(&This->IUri_iface);
5423 static ULONG WINAPI Marshal_Release(IMarshal *iface)
5425 Uri *This = impl_from_IMarshal(iface);
5426 return IUri_Release(&This->IUri_iface);
5429 static HRESULT WINAPI Marshal_GetUnmarshalClass(IMarshal *iface, REFIID riid, void *pv,
5430 DWORD dwDestContext, void *pvDestContext, DWORD mshlflags, CLSID *pCid)
5432 Uri *This = impl_from_IMarshal(iface);
5433 TRACE("(%p)->(%s %p %x %p %x %p)\n", This, debugstr_guid(riid), pv,
5434 dwDestContext, pvDestContext, mshlflags, pCid);
5436 if(!pCid || (dwDestContext!=MSHCTX_LOCAL && dwDestContext!=MSHCTX_NOSHAREDMEM
5437 && dwDestContext!=MSHCTX_INPROC))
5438 return E_INVALIDARG;
5440 *pCid = CLSID_CUri;
5441 return S_OK;
5444 struct inproc_marshal_uri {
5445 DWORD size;
5446 DWORD mshlflags;
5447 DWORD unk[4]; /* process identifier? */
5448 Uri *uri;
5451 static HRESULT WINAPI Marshal_GetMarshalSizeMax(IMarshal *iface, REFIID riid, void *pv,
5452 DWORD dwDestContext, void *pvDestContext, DWORD mshlflags, DWORD *pSize)
5454 Uri *This = impl_from_IMarshal(iface);
5455 ULARGE_INTEGER size;
5456 HRESULT hres;
5457 TRACE("(%p)->(%s %p %x %p %x %p)\n", This, debugstr_guid(riid), pv,
5458 dwDestContext, pvDestContext, mshlflags, pSize);
5460 if(!pSize || (dwDestContext!=MSHCTX_LOCAL && dwDestContext!=MSHCTX_NOSHAREDMEM
5461 && dwDestContext!=MSHCTX_INPROC))
5462 return E_INVALIDARG;
5464 if(dwDestContext == MSHCTX_INPROC) {
5465 *pSize = sizeof(struct inproc_marshal_uri);
5466 return S_OK;
5469 hres = IPersistStream_GetSizeMax(&This->IPersistStream_iface, &size);
5470 if(FAILED(hres))
5471 return hres;
5472 if(!This->path_len && (This->scheme_type==URL_SCHEME_HTTP
5473 || This->scheme_type==URL_SCHEME_HTTPS
5474 || This->scheme_type==URL_SCHEME_FTP))
5475 size.u.LowPart += 3*sizeof(DWORD);
5476 *pSize = size.u.LowPart+2*sizeof(DWORD);
5477 return S_OK;
5480 static HRESULT WINAPI Marshal_MarshalInterface(IMarshal *iface, IStream *pStm, REFIID riid,
5481 void *pv, DWORD dwDestContext, void *pvDestContext, DWORD mshlflags)
5483 Uri *This = impl_from_IMarshal(iface);
5484 DWORD *data;
5485 DWORD size;
5486 HRESULT hres;
5488 TRACE("(%p)->(%p %s %p %x %p %x)\n", This, pStm, debugstr_guid(riid), pv,
5489 dwDestContext, pvDestContext, mshlflags);
5491 if(!pStm || mshlflags!=MSHLFLAGS_NORMAL || (dwDestContext!=MSHCTX_LOCAL
5492 && dwDestContext!=MSHCTX_NOSHAREDMEM && dwDestContext!=MSHCTX_INPROC))
5493 return E_INVALIDARG;
5495 if(dwDestContext == MSHCTX_INPROC) {
5496 struct inproc_marshal_uri data;
5498 data.size = sizeof(data);
5499 data.mshlflags = MSHCTX_INPROC;
5500 data.unk[0] = 0;
5501 data.unk[1] = 0;
5502 data.unk[2] = 0;
5503 data.unk[3] = 0;
5504 data.uri = This;
5506 hres = IStream_Write(pStm, &data, data.size, NULL);
5507 if(FAILED(hres))
5508 return hres;
5510 IUri_AddRef(&This->IUri_iface);
5511 return S_OK;
5514 hres = IMarshal_GetMarshalSizeMax(iface, riid, pv, dwDestContext,
5515 pvDestContext, mshlflags, &size);
5516 if(FAILED(hres))
5517 return hres;
5519 data = heap_alloc_zero(size);
5520 if(!data)
5521 return E_OUTOFMEMORY;
5523 data[0] = size;
5524 data[1] = dwDestContext;
5525 data[2] = size-2*sizeof(DWORD);
5526 persist_stream_save(This, pStm, TRUE, (struct persist_uri*)(data+2));
5528 hres = IStream_Write(pStm, data, data[0]-2, NULL);
5529 heap_free(data);
5530 return hres;
5533 static HRESULT WINAPI Marshal_UnmarshalInterface(IMarshal *iface,
5534 IStream *pStm, REFIID riid, void **ppv)
5536 Uri *This = impl_from_IMarshal(iface);
5537 DWORD header[2];
5538 HRESULT hres;
5540 TRACE("(%p)->(%p %s %p)\n", This, pStm, debugstr_guid(riid), ppv);
5542 if(This->create_flags)
5543 return E_UNEXPECTED;
5544 if(!pStm || !riid || !ppv)
5545 return E_INVALIDARG;
5547 hres = IStream_Read(pStm, header, sizeof(header), NULL);
5548 if(FAILED(hres))
5549 return hres;
5551 if(header[1]!=MSHCTX_LOCAL && header[1]!=MSHCTX_NOSHAREDMEM
5552 && header[1]!=MSHCTX_INPROC)
5553 return E_UNEXPECTED;
5555 if(header[1] == MSHCTX_INPROC) {
5556 struct inproc_marshal_uri data;
5557 parse_data parse;
5559 hres = IStream_Read(pStm, data.unk, sizeof(data)-2*sizeof(DWORD), NULL);
5560 if(FAILED(hres))
5561 return hres;
5563 This->raw_uri = SysAllocString(data.uri->raw_uri);
5564 if(!This->raw_uri) {
5565 return E_OUTOFMEMORY;
5568 memset(&parse, 0, sizeof(parse_data));
5569 parse.uri = This->raw_uri;
5571 if(!parse_uri(&parse, data.uri->create_flags))
5572 return E_INVALIDARG;
5574 hres = canonicalize_uri(&parse, This, data.uri->create_flags);
5575 if(FAILED(hres))
5576 return hres;
5578 This->create_flags = data.uri->create_flags;
5579 IUri_Release(&data.uri->IUri_iface);
5581 return IUri_QueryInterface(&This->IUri_iface, riid, ppv);
5584 hres = IPersistStream_Load(&This->IPersistStream_iface, pStm);
5585 if(FAILED(hres))
5586 return hres;
5588 return IUri_QueryInterface(&This->IUri_iface, riid, ppv);
5591 static HRESULT WINAPI Marshal_ReleaseMarshalData(IMarshal *iface, IStream *pStm)
5593 Uri *This = impl_from_IMarshal(iface);
5594 LARGE_INTEGER off;
5595 DWORD header[2];
5596 HRESULT hres;
5598 TRACE("(%p)->(%p)\n", This, pStm);
5600 if(!pStm)
5601 return E_INVALIDARG;
5603 hres = IStream_Read(pStm, header, 2*sizeof(DWORD), NULL);
5604 if(FAILED(hres))
5605 return hres;
5607 if(header[1] == MSHCTX_INPROC) {
5608 struct inproc_marshal_uri data;
5610 hres = IStream_Read(pStm, data.unk, sizeof(data)-2*sizeof(DWORD), NULL);
5611 if(FAILED(hres))
5612 return hres;
5614 IUri_Release(&data.uri->IUri_iface);
5615 return S_OK;
5618 off.u.LowPart = header[0]-sizeof(header)-2;
5619 off.u.HighPart = 0;
5620 return IStream_Seek(pStm, off, STREAM_SEEK_CUR, NULL);
5623 static HRESULT WINAPI Marshal_DisconnectObject(IMarshal *iface, DWORD dwReserved)
5625 Uri *This = impl_from_IMarshal(iface);
5626 TRACE("(%p)->(%x)\n", This, dwReserved);
5627 return S_OK;
5630 static const IMarshalVtbl MarshalVtbl = {
5631 Marshal_QueryInterface,
5632 Marshal_AddRef,
5633 Marshal_Release,
5634 Marshal_GetUnmarshalClass,
5635 Marshal_GetMarshalSizeMax,
5636 Marshal_MarshalInterface,
5637 Marshal_UnmarshalInterface,
5638 Marshal_ReleaseMarshalData,
5639 Marshal_DisconnectObject
5642 HRESULT Uri_Construct(IUnknown *pUnkOuter, LPVOID *ppobj)
5644 Uri *ret = heap_alloc_zero(sizeof(Uri));
5646 TRACE("(%p %p)\n", pUnkOuter, ppobj);
5648 *ppobj = ret;
5649 if(!ret)
5650 return E_OUTOFMEMORY;
5652 ret->IUri_iface.lpVtbl = &UriVtbl;
5653 ret->IUriBuilderFactory_iface.lpVtbl = &UriBuilderFactoryVtbl;
5654 ret->IPersistStream_iface.lpVtbl = &PersistStreamVtbl;
5655 ret->IMarshal_iface.lpVtbl = &MarshalVtbl;
5656 ret->ref = 1;
5658 *ppobj = &ret->IUri_iface;
5659 return S_OK;
5662 /***********************************************************************
5663 * CreateUri (urlmon.@)
5665 * Creates a new IUri object using the URI represented by pwzURI. This function
5666 * parses and validates the components of pwzURI and then canonicalizes the
5667 * parsed components.
5669 * PARAMS
5670 * pwzURI [I] The URI to parse, validate, and canonicalize.
5671 * dwFlags [I] Flags which can affect how the parsing/canonicalization is performed.
5672 * dwReserved [I] Reserved (not used).
5673 * ppURI [O] The resulting IUri after parsing/canonicalization occurs.
5675 * RETURNS
5676 * Success: Returns S_OK. ppURI contains the pointer to the newly allocated IUri.
5677 * Failure: E_INVALIDARG if there are invalid flag combinations in dwFlags, or an
5678 * invalid parameter, or pwzURI doesn't represent a valid URI.
5679 * E_OUTOFMEMORY if any memory allocation fails.
5681 * NOTES
5682 * Default flags:
5683 * Uri_CREATE_CANONICALIZE, Uri_CREATE_DECODE_EXTRA_INFO, Uri_CREATE_CRACK_UNKNOWN_SCHEMES,
5684 * Uri_CREATE_PRE_PROCESS_HTML_URI, Uri_CREATE_NO_IE_SETTINGS.
5686 HRESULT WINAPI CreateUri(LPCWSTR pwzURI, DWORD dwFlags, DWORD_PTR dwReserved, IUri **ppURI)
5688 const DWORD supported_flags = Uri_CREATE_ALLOW_RELATIVE|Uri_CREATE_ALLOW_IMPLICIT_WILDCARD_SCHEME|
5689 Uri_CREATE_ALLOW_IMPLICIT_FILE_SCHEME|Uri_CREATE_NO_CANONICALIZE|Uri_CREATE_CANONICALIZE|
5690 Uri_CREATE_DECODE_EXTRA_INFO|Uri_CREATE_NO_DECODE_EXTRA_INFO|Uri_CREATE_CRACK_UNKNOWN_SCHEMES|
5691 Uri_CREATE_NO_CRACK_UNKNOWN_SCHEMES|Uri_CREATE_PRE_PROCESS_HTML_URI|Uri_CREATE_NO_PRE_PROCESS_HTML_URI|
5692 Uri_CREATE_NO_IE_SETTINGS|Uri_CREATE_NO_ENCODE_FORBIDDEN_CHARACTERS|Uri_CREATE_FILE_USE_DOS_PATH;
5693 Uri *ret;
5694 HRESULT hr;
5695 parse_data data;
5697 TRACE("(%s %x %x %p)\n", debugstr_w(pwzURI), dwFlags, (DWORD)dwReserved, ppURI);
5699 if(!ppURI)
5700 return E_INVALIDARG;
5702 if(!pwzURI) {
5703 *ppURI = NULL;
5704 return E_INVALIDARG;
5707 /* Check for invalid flags. */
5708 if(has_invalid_flag_combination(dwFlags)) {
5709 *ppURI = NULL;
5710 return E_INVALIDARG;
5713 /* Currently unsupported. */
5714 if(dwFlags & ~supported_flags)
5715 FIXME("Ignoring unsupported flag(s) %x\n", dwFlags & ~supported_flags);
5717 hr = Uri_Construct(NULL, (void**)&ret);
5718 if(FAILED(hr)) {
5719 *ppURI = NULL;
5720 return hr;
5723 /* Explicitly set the default flags if it doesn't cause a flag conflict. */
5724 apply_default_flags(&dwFlags);
5726 /* Pre process the URI, unless told otherwise. */
5727 if(!(dwFlags & Uri_CREATE_NO_PRE_PROCESS_HTML_URI))
5728 ret->raw_uri = pre_process_uri(pwzURI);
5729 else
5730 ret->raw_uri = SysAllocString(pwzURI);
5732 if(!ret->raw_uri) {
5733 heap_free(ret);
5734 return E_OUTOFMEMORY;
5737 memset(&data, 0, sizeof(parse_data));
5738 data.uri = ret->raw_uri;
5740 /* Validate and parse the URI into its components. */
5741 if(!parse_uri(&data, dwFlags)) {
5742 /* Encountered an unsupported or invalid URI */
5743 IUri_Release(&ret->IUri_iface);
5744 *ppURI = NULL;
5745 return E_INVALIDARG;
5748 /* Canonicalize the URI. */
5749 hr = canonicalize_uri(&data, ret, dwFlags);
5750 if(FAILED(hr)) {
5751 IUri_Release(&ret->IUri_iface);
5752 *ppURI = NULL;
5753 return hr;
5756 ret->create_flags = dwFlags;
5758 *ppURI = &ret->IUri_iface;
5759 return S_OK;
5762 /***********************************************************************
5763 * CreateUriWithFragment (urlmon.@)
5765 * Creates a new IUri object. This is almost the same as CreateUri, expect that
5766 * it allows you to explicitly specify a fragment (pwzFragment) for pwzURI.
5768 * PARAMS
5769 * pwzURI [I] The URI to parse and perform canonicalization on.
5770 * pwzFragment [I] The explicit fragment string which should be added to pwzURI.
5771 * dwFlags [I] The flags which will be passed to CreateUri.
5772 * dwReserved [I] Reserved (not used).
5773 * ppURI [O] The resulting IUri after parsing/canonicalization.
5775 * RETURNS
5776 * Success: S_OK. ppURI contains the pointer to the newly allocated IUri.
5777 * Failure: E_INVALIDARG if pwzURI already contains a fragment and pwzFragment
5778 * isn't NULL. Will also return E_INVALIDARG for the same reasons as
5779 * CreateUri will. E_OUTOFMEMORY if any allocation fails.
5781 HRESULT WINAPI CreateUriWithFragment(LPCWSTR pwzURI, LPCWSTR pwzFragment, DWORD dwFlags,
5782 DWORD_PTR dwReserved, IUri **ppURI)
5784 HRESULT hres;
5785 TRACE("(%s %s %x %x %p)\n", debugstr_w(pwzURI), debugstr_w(pwzFragment), dwFlags, (DWORD)dwReserved, ppURI);
5787 if(!ppURI)
5788 return E_INVALIDARG;
5790 if(!pwzURI) {
5791 *ppURI = NULL;
5792 return E_INVALIDARG;
5795 /* Check if a fragment should be appended to the URI string. */
5796 if(pwzFragment) {
5797 WCHAR *uriW;
5798 DWORD uri_len, frag_len;
5799 BOOL add_pound;
5801 /* Check if the original URI already has a fragment component. */
5802 if(StrChrW(pwzURI, '#')) {
5803 *ppURI = NULL;
5804 return E_INVALIDARG;
5807 uri_len = lstrlenW(pwzURI);
5808 frag_len = lstrlenW(pwzFragment);
5810 /* If the fragment doesn't start with a '#', one will be added. */
5811 add_pound = *pwzFragment != '#';
5813 if(add_pound)
5814 uriW = heap_alloc((uri_len+frag_len+2)*sizeof(WCHAR));
5815 else
5816 uriW = heap_alloc((uri_len+frag_len+1)*sizeof(WCHAR));
5818 if(!uriW)
5819 return E_OUTOFMEMORY;
5821 memcpy(uriW, pwzURI, uri_len*sizeof(WCHAR));
5822 if(add_pound)
5823 uriW[uri_len++] = '#';
5824 memcpy(uriW+uri_len, pwzFragment, (frag_len+1)*sizeof(WCHAR));
5826 hres = CreateUri(uriW, dwFlags, 0, ppURI);
5828 heap_free(uriW);
5829 } else
5830 /* A fragment string wasn't specified, so just forward the call. */
5831 hres = CreateUri(pwzURI, dwFlags, 0, ppURI);
5833 return hres;
5836 static HRESULT build_uri(const UriBuilder *builder, IUri **uri, DWORD create_flags,
5837 DWORD use_orig_flags, DWORD encoding_mask)
5839 HRESULT hr;
5840 parse_data data;
5841 Uri *ret;
5843 if(!uri)
5844 return E_POINTER;
5846 if(encoding_mask && (!builder->uri || builder->modified_props)) {
5847 *uri = NULL;
5848 return E_NOTIMPL;
5851 /* Decide what flags should be used when creating the Uri. */
5852 if((use_orig_flags & UriBuilder_USE_ORIGINAL_FLAGS) && builder->uri)
5853 create_flags = builder->uri->create_flags;
5854 else {
5855 if(has_invalid_flag_combination(create_flags)) {
5856 *uri = NULL;
5857 return E_INVALIDARG;
5860 /* Set the default flags if they don't cause a conflict. */
5861 apply_default_flags(&create_flags);
5864 /* Return the base IUri if no changes have been made and the create_flags match. */
5865 if(builder->uri && !builder->modified_props && builder->uri->create_flags == create_flags) {
5866 *uri = &builder->uri->IUri_iface;
5867 IUri_AddRef(*uri);
5868 return S_OK;
5871 hr = validate_components(builder, &data, create_flags);
5872 if(FAILED(hr)) {
5873 *uri = NULL;
5874 return hr;
5877 hr = Uri_Construct(NULL, (void**)&ret);
5878 if(FAILED(hr)) {
5879 *uri = NULL;
5880 return hr;
5883 hr = generate_uri(builder, &data, ret, create_flags);
5884 if(FAILED(hr)) {
5885 IUri_Release(&ret->IUri_iface);
5886 *uri = NULL;
5887 return hr;
5890 *uri = &ret->IUri_iface;
5891 return S_OK;
5894 static inline UriBuilder* impl_from_IUriBuilder(IUriBuilder *iface)
5896 return CONTAINING_RECORD(iface, UriBuilder, IUriBuilder_iface);
5899 static HRESULT WINAPI UriBuilder_QueryInterface(IUriBuilder *iface, REFIID riid, void **ppv)
5901 UriBuilder *This = impl_from_IUriBuilder(iface);
5903 if(IsEqualGUID(&IID_IUnknown, riid)) {
5904 TRACE("(%p)->(IID_IUnknown %p)\n", This, ppv);
5905 *ppv = &This->IUriBuilder_iface;
5906 }else if(IsEqualGUID(&IID_IUriBuilder, riid)) {
5907 TRACE("(%p)->(IID_IUriBuilder %p)\n", This, ppv);
5908 *ppv = &This->IUriBuilder_iface;
5909 }else {
5910 TRACE("(%p)->(%s %p)\n", This, debugstr_guid(riid), ppv);
5911 *ppv = NULL;
5912 return E_NOINTERFACE;
5915 IUnknown_AddRef((IUnknown*)*ppv);
5916 return S_OK;
5919 static ULONG WINAPI UriBuilder_AddRef(IUriBuilder *iface)
5921 UriBuilder *This = impl_from_IUriBuilder(iface);
5922 LONG ref = InterlockedIncrement(&This->ref);
5924 TRACE("(%p) ref=%d\n", This, ref);
5926 return ref;
5929 static ULONG WINAPI UriBuilder_Release(IUriBuilder *iface)
5931 UriBuilder *This = impl_from_IUriBuilder(iface);
5932 LONG ref = InterlockedDecrement(&This->ref);
5934 TRACE("(%p) ref=%d\n", This, ref);
5936 if(!ref) {
5937 if(This->uri) IUri_Release(&This->uri->IUri_iface);
5938 heap_free(This->fragment);
5939 heap_free(This->host);
5940 heap_free(This->password);
5941 heap_free(This->path);
5942 heap_free(This->query);
5943 heap_free(This->scheme);
5944 heap_free(This->username);
5945 heap_free(This);
5948 return ref;
5951 static HRESULT WINAPI UriBuilder_CreateUriSimple(IUriBuilder *iface,
5952 DWORD dwAllowEncodingPropertyMask,
5953 DWORD_PTR dwReserved,
5954 IUri **ppIUri)
5956 UriBuilder *This = impl_from_IUriBuilder(iface);
5957 HRESULT hr;
5958 TRACE("(%p)->(%d %d %p)\n", This, dwAllowEncodingPropertyMask, (DWORD)dwReserved, ppIUri);
5960 hr = build_uri(This, ppIUri, 0, UriBuilder_USE_ORIGINAL_FLAGS, dwAllowEncodingPropertyMask);
5961 if(hr == E_NOTIMPL)
5962 FIXME("(%p)->(%d %d %p)\n", This, dwAllowEncodingPropertyMask, (DWORD)dwReserved, ppIUri);
5963 return hr;
5966 static HRESULT WINAPI UriBuilder_CreateUri(IUriBuilder *iface,
5967 DWORD dwCreateFlags,
5968 DWORD dwAllowEncodingPropertyMask,
5969 DWORD_PTR dwReserved,
5970 IUri **ppIUri)
5972 UriBuilder *This = impl_from_IUriBuilder(iface);
5973 HRESULT hr;
5974 TRACE("(%p)->(0x%08x %d %d %p)\n", This, dwCreateFlags, dwAllowEncodingPropertyMask, (DWORD)dwReserved, ppIUri);
5976 if(dwCreateFlags == -1)
5977 hr = build_uri(This, ppIUri, 0, UriBuilder_USE_ORIGINAL_FLAGS, dwAllowEncodingPropertyMask);
5978 else
5979 hr = build_uri(This, ppIUri, dwCreateFlags, 0, dwAllowEncodingPropertyMask);
5981 if(hr == E_NOTIMPL)
5982 FIXME("(%p)->(0x%08x %d %d %p)\n", This, dwCreateFlags, dwAllowEncodingPropertyMask, (DWORD)dwReserved, ppIUri);
5983 return hr;
5986 static HRESULT WINAPI UriBuilder_CreateUriWithFlags(IUriBuilder *iface,
5987 DWORD dwCreateFlags,
5988 DWORD dwUriBuilderFlags,
5989 DWORD dwAllowEncodingPropertyMask,
5990 DWORD_PTR dwReserved,
5991 IUri **ppIUri)
5993 UriBuilder *This = impl_from_IUriBuilder(iface);
5994 HRESULT hr;
5995 TRACE("(%p)->(0x%08x 0x%08x %d %d %p)\n", This, dwCreateFlags, dwUriBuilderFlags,
5996 dwAllowEncodingPropertyMask, (DWORD)dwReserved, ppIUri);
5998 hr = build_uri(This, ppIUri, dwCreateFlags, dwUriBuilderFlags, dwAllowEncodingPropertyMask);
5999 if(hr == E_NOTIMPL)
6000 FIXME("(%p)->(0x%08x 0x%08x %d %d %p)\n", This, dwCreateFlags, dwUriBuilderFlags,
6001 dwAllowEncodingPropertyMask, (DWORD)dwReserved, ppIUri);
6002 return hr;
6005 static HRESULT WINAPI UriBuilder_GetIUri(IUriBuilder *iface, IUri **ppIUri)
6007 UriBuilder *This = impl_from_IUriBuilder(iface);
6008 TRACE("(%p)->(%p)\n", This, ppIUri);
6010 if(!ppIUri)
6011 return E_POINTER;
6013 if(This->uri) {
6014 IUri *uri = &This->uri->IUri_iface;
6015 IUri_AddRef(uri);
6016 *ppIUri = uri;
6017 } else
6018 *ppIUri = NULL;
6020 return S_OK;
6023 static HRESULT WINAPI UriBuilder_SetIUri(IUriBuilder *iface, IUri *pIUri)
6025 UriBuilder *This = impl_from_IUriBuilder(iface);
6026 TRACE("(%p)->(%p)\n", This, pIUri);
6028 if(pIUri) {
6029 Uri *uri;
6031 if((uri = get_uri_obj(pIUri))) {
6032 /* Only reset the builder if it's Uri isn't the same as
6033 * the Uri passed to the function.
6035 if(This->uri != uri) {
6036 reset_builder(This);
6038 This->uri = uri;
6039 if(uri->has_port)
6040 This->port = uri->port;
6042 IUri_AddRef(pIUri);
6044 } else {
6045 FIXME("(%p)->(%p) Unknown IUri types not supported yet.\n", This, pIUri);
6046 return E_NOTIMPL;
6048 } else if(This->uri)
6049 /* Only reset the builder if it's Uri isn't NULL. */
6050 reset_builder(This);
6052 return S_OK;
6055 static HRESULT WINAPI UriBuilder_GetFragment(IUriBuilder *iface, DWORD *pcchFragment, LPCWSTR *ppwzFragment)
6057 UriBuilder *This = impl_from_IUriBuilder(iface);
6058 TRACE("(%p)->(%p %p)\n", This, pcchFragment, ppwzFragment);
6060 if(!This->uri || This->uri->fragment_start == -1 || This->modified_props & Uri_HAS_FRAGMENT)
6061 return get_builder_component(&This->fragment, &This->fragment_len, NULL, 0, ppwzFragment, pcchFragment);
6062 else
6063 return get_builder_component(&This->fragment, &This->fragment_len, This->uri->canon_uri+This->uri->fragment_start,
6064 This->uri->fragment_len, ppwzFragment, pcchFragment);
6067 static HRESULT WINAPI UriBuilder_GetHost(IUriBuilder *iface, DWORD *pcchHost, LPCWSTR *ppwzHost)
6069 UriBuilder *This = impl_from_IUriBuilder(iface);
6070 TRACE("(%p)->(%p %p)\n", This, pcchHost, ppwzHost);
6072 if(!This->uri || This->uri->host_start == -1 || This->modified_props & Uri_HAS_HOST)
6073 return get_builder_component(&This->host, &This->host_len, NULL, 0, ppwzHost, pcchHost);
6074 else {
6075 if(This->uri->host_type == Uri_HOST_IPV6)
6076 /* Don't include the '[' and ']' around the address. */
6077 return get_builder_component(&This->host, &This->host_len, This->uri->canon_uri+This->uri->host_start+1,
6078 This->uri->host_len-2, ppwzHost, pcchHost);
6079 else
6080 return get_builder_component(&This->host, &This->host_len, This->uri->canon_uri+This->uri->host_start,
6081 This->uri->host_len, ppwzHost, pcchHost);
6085 static HRESULT WINAPI UriBuilder_GetPassword(IUriBuilder *iface, DWORD *pcchPassword, LPCWSTR *ppwzPassword)
6087 UriBuilder *This = impl_from_IUriBuilder(iface);
6088 TRACE("(%p)->(%p %p)\n", This, pcchPassword, ppwzPassword);
6090 if(!This->uri || This->uri->userinfo_split == -1 || This->modified_props & Uri_HAS_PASSWORD)
6091 return get_builder_component(&This->password, &This->password_len, NULL, 0, ppwzPassword, pcchPassword);
6092 else {
6093 const WCHAR *start = This->uri->canon_uri+This->uri->userinfo_start+This->uri->userinfo_split+1;
6094 DWORD len = This->uri->userinfo_len-This->uri->userinfo_split-1;
6095 return get_builder_component(&This->password, &This->password_len, start, len, ppwzPassword, pcchPassword);
6099 static HRESULT WINAPI UriBuilder_GetPath(IUriBuilder *iface, DWORD *pcchPath, LPCWSTR *ppwzPath)
6101 UriBuilder *This = impl_from_IUriBuilder(iface);
6102 TRACE("(%p)->(%p %p)\n", This, pcchPath, ppwzPath);
6104 if(!This->uri || This->uri->path_start == -1 || This->modified_props & Uri_HAS_PATH)
6105 return get_builder_component(&This->path, &This->path_len, NULL, 0, ppwzPath, pcchPath);
6106 else
6107 return get_builder_component(&This->path, &This->path_len, This->uri->canon_uri+This->uri->path_start,
6108 This->uri->path_len, ppwzPath, pcchPath);
6111 static HRESULT WINAPI UriBuilder_GetPort(IUriBuilder *iface, BOOL *pfHasPort, DWORD *pdwPort)
6113 UriBuilder *This = impl_from_IUriBuilder(iface);
6114 TRACE("(%p)->(%p %p)\n", This, pfHasPort, pdwPort);
6116 if(!pfHasPort) {
6117 if(pdwPort)
6118 *pdwPort = 0;
6119 return E_POINTER;
6122 if(!pdwPort) {
6123 *pfHasPort = FALSE;
6124 return E_POINTER;
6127 *pfHasPort = This->has_port;
6128 *pdwPort = This->port;
6129 return S_OK;
6132 static HRESULT WINAPI UriBuilder_GetQuery(IUriBuilder *iface, DWORD *pcchQuery, LPCWSTR *ppwzQuery)
6134 UriBuilder *This = impl_from_IUriBuilder(iface);
6135 TRACE("(%p)->(%p %p)\n", This, pcchQuery, ppwzQuery);
6137 if(!This->uri || This->uri->query_start == -1 || This->modified_props & Uri_HAS_QUERY)
6138 return get_builder_component(&This->query, &This->query_len, NULL, 0, ppwzQuery, pcchQuery);
6139 else
6140 return get_builder_component(&This->query, &This->query_len, This->uri->canon_uri+This->uri->query_start,
6141 This->uri->query_len, ppwzQuery, pcchQuery);
6144 static HRESULT WINAPI UriBuilder_GetSchemeName(IUriBuilder *iface, DWORD *pcchSchemeName, LPCWSTR *ppwzSchemeName)
6146 UriBuilder *This = impl_from_IUriBuilder(iface);
6147 TRACE("(%p)->(%p %p)\n", This, pcchSchemeName, ppwzSchemeName);
6149 if(!This->uri || This->uri->scheme_start == -1 || This->modified_props & Uri_HAS_SCHEME_NAME)
6150 return get_builder_component(&This->scheme, &This->scheme_len, NULL, 0, ppwzSchemeName, pcchSchemeName);
6151 else
6152 return get_builder_component(&This->scheme, &This->scheme_len, This->uri->canon_uri+This->uri->scheme_start,
6153 This->uri->scheme_len, ppwzSchemeName, pcchSchemeName);
6156 static HRESULT WINAPI UriBuilder_GetUserName(IUriBuilder *iface, DWORD *pcchUserName, LPCWSTR *ppwzUserName)
6158 UriBuilder *This = impl_from_IUriBuilder(iface);
6159 TRACE("(%p)->(%p %p)\n", This, pcchUserName, ppwzUserName);
6161 if(!This->uri || This->uri->userinfo_start == -1 || This->uri->userinfo_split == 0 ||
6162 This->modified_props & Uri_HAS_USER_NAME)
6163 return get_builder_component(&This->username, &This->username_len, NULL, 0, ppwzUserName, pcchUserName);
6164 else {
6165 const WCHAR *start = This->uri->canon_uri+This->uri->userinfo_start;
6167 /* Check if there's a password in the userinfo section. */
6168 if(This->uri->userinfo_split > -1)
6169 /* Don't include the password. */
6170 return get_builder_component(&This->username, &This->username_len, start,
6171 This->uri->userinfo_split, ppwzUserName, pcchUserName);
6172 else
6173 return get_builder_component(&This->username, &This->username_len, start,
6174 This->uri->userinfo_len, ppwzUserName, pcchUserName);
6178 static HRESULT WINAPI UriBuilder_SetFragment(IUriBuilder *iface, LPCWSTR pwzNewValue)
6180 UriBuilder *This = impl_from_IUriBuilder(iface);
6181 TRACE("(%p)->(%s)\n", This, debugstr_w(pwzNewValue));
6182 return set_builder_component(&This->fragment, &This->fragment_len, pwzNewValue, '#',
6183 &This->modified_props, Uri_HAS_FRAGMENT);
6186 static HRESULT WINAPI UriBuilder_SetHost(IUriBuilder *iface, LPCWSTR pwzNewValue)
6188 UriBuilder *This = impl_from_IUriBuilder(iface);
6189 TRACE("(%p)->(%s)\n", This, debugstr_w(pwzNewValue));
6191 /* Host name can't be set to NULL. */
6192 if(!pwzNewValue)
6193 return E_INVALIDARG;
6195 return set_builder_component(&This->host, &This->host_len, pwzNewValue, 0,
6196 &This->modified_props, Uri_HAS_HOST);
6199 static HRESULT WINAPI UriBuilder_SetPassword(IUriBuilder *iface, LPCWSTR pwzNewValue)
6201 UriBuilder *This = impl_from_IUriBuilder(iface);
6202 TRACE("(%p)->(%s)\n", This, debugstr_w(pwzNewValue));
6203 return set_builder_component(&This->password, &This->password_len, pwzNewValue, 0,
6204 &This->modified_props, Uri_HAS_PASSWORD);
6207 static HRESULT WINAPI UriBuilder_SetPath(IUriBuilder *iface, LPCWSTR pwzNewValue)
6209 UriBuilder *This = impl_from_IUriBuilder(iface);
6210 TRACE("(%p)->(%s)\n", This, debugstr_w(pwzNewValue));
6211 return set_builder_component(&This->path, &This->path_len, pwzNewValue, 0,
6212 &This->modified_props, Uri_HAS_PATH);
6215 static HRESULT WINAPI UriBuilder_SetPort(IUriBuilder *iface, BOOL fHasPort, DWORD dwNewValue)
6217 UriBuilder *This = impl_from_IUriBuilder(iface);
6218 TRACE("(%p)->(%d %d)\n", This, fHasPort, dwNewValue);
6220 This->has_port = fHasPort;
6221 This->port = dwNewValue;
6222 This->modified_props |= Uri_HAS_PORT;
6223 return S_OK;
6226 static HRESULT WINAPI UriBuilder_SetQuery(IUriBuilder *iface, LPCWSTR pwzNewValue)
6228 UriBuilder *This = impl_from_IUriBuilder(iface);
6229 TRACE("(%p)->(%s)\n", This, debugstr_w(pwzNewValue));
6230 return set_builder_component(&This->query, &This->query_len, pwzNewValue, '?',
6231 &This->modified_props, Uri_HAS_QUERY);
6234 static HRESULT WINAPI UriBuilder_SetSchemeName(IUriBuilder *iface, LPCWSTR pwzNewValue)
6236 UriBuilder *This = impl_from_IUriBuilder(iface);
6237 TRACE("(%p)->(%s)\n", This, debugstr_w(pwzNewValue));
6239 /* Only set the scheme name if it's not NULL or empty. */
6240 if(!pwzNewValue || !*pwzNewValue)
6241 return E_INVALIDARG;
6243 return set_builder_component(&This->scheme, &This->scheme_len, pwzNewValue, 0,
6244 &This->modified_props, Uri_HAS_SCHEME_NAME);
6247 static HRESULT WINAPI UriBuilder_SetUserName(IUriBuilder *iface, LPCWSTR pwzNewValue)
6249 UriBuilder *This = impl_from_IUriBuilder(iface);
6250 TRACE("(%p)->(%s)\n", This, debugstr_w(pwzNewValue));
6251 return set_builder_component(&This->username, &This->username_len, pwzNewValue, 0,
6252 &This->modified_props, Uri_HAS_USER_NAME);
6255 static HRESULT WINAPI UriBuilder_RemoveProperties(IUriBuilder *iface, DWORD dwPropertyMask)
6257 const DWORD accepted_flags = Uri_HAS_AUTHORITY|Uri_HAS_DOMAIN|Uri_HAS_EXTENSION|Uri_HAS_FRAGMENT|Uri_HAS_HOST|
6258 Uri_HAS_PASSWORD|Uri_HAS_PATH|Uri_HAS_PATH_AND_QUERY|Uri_HAS_QUERY|
6259 Uri_HAS_USER_INFO|Uri_HAS_USER_NAME;
6261 UriBuilder *This = impl_from_IUriBuilder(iface);
6262 TRACE("(%p)->(0x%08x)\n", This, dwPropertyMask);
6264 if(dwPropertyMask & ~accepted_flags)
6265 return E_INVALIDARG;
6267 if(dwPropertyMask & Uri_HAS_FRAGMENT)
6268 UriBuilder_SetFragment(iface, NULL);
6270 /* Even though you can't set the host name to NULL or an
6271 * empty string, you can still remove it... for some reason.
6273 if(dwPropertyMask & Uri_HAS_HOST)
6274 set_builder_component(&This->host, &This->host_len, NULL, 0,
6275 &This->modified_props, Uri_HAS_HOST);
6277 if(dwPropertyMask & Uri_HAS_PASSWORD)
6278 UriBuilder_SetPassword(iface, NULL);
6280 if(dwPropertyMask & Uri_HAS_PATH)
6281 UriBuilder_SetPath(iface, NULL);
6283 if(dwPropertyMask & Uri_HAS_PORT)
6284 UriBuilder_SetPort(iface, FALSE, 0);
6286 if(dwPropertyMask & Uri_HAS_QUERY)
6287 UriBuilder_SetQuery(iface, NULL);
6289 if(dwPropertyMask & Uri_HAS_USER_NAME)
6290 UriBuilder_SetUserName(iface, NULL);
6292 return S_OK;
6295 static HRESULT WINAPI UriBuilder_HasBeenModified(IUriBuilder *iface, BOOL *pfModified)
6297 UriBuilder *This = impl_from_IUriBuilder(iface);
6298 TRACE("(%p)->(%p)\n", This, pfModified);
6300 if(!pfModified)
6301 return E_POINTER;
6303 *pfModified = This->modified_props > 0;
6304 return S_OK;
6307 static const IUriBuilderVtbl UriBuilderVtbl = {
6308 UriBuilder_QueryInterface,
6309 UriBuilder_AddRef,
6310 UriBuilder_Release,
6311 UriBuilder_CreateUriSimple,
6312 UriBuilder_CreateUri,
6313 UriBuilder_CreateUriWithFlags,
6314 UriBuilder_GetIUri,
6315 UriBuilder_SetIUri,
6316 UriBuilder_GetFragment,
6317 UriBuilder_GetHost,
6318 UriBuilder_GetPassword,
6319 UriBuilder_GetPath,
6320 UriBuilder_GetPort,
6321 UriBuilder_GetQuery,
6322 UriBuilder_GetSchemeName,
6323 UriBuilder_GetUserName,
6324 UriBuilder_SetFragment,
6325 UriBuilder_SetHost,
6326 UriBuilder_SetPassword,
6327 UriBuilder_SetPath,
6328 UriBuilder_SetPort,
6329 UriBuilder_SetQuery,
6330 UriBuilder_SetSchemeName,
6331 UriBuilder_SetUserName,
6332 UriBuilder_RemoveProperties,
6333 UriBuilder_HasBeenModified,
6336 /***********************************************************************
6337 * CreateIUriBuilder (urlmon.@)
6339 HRESULT WINAPI CreateIUriBuilder(IUri *pIUri, DWORD dwFlags, DWORD_PTR dwReserved, IUriBuilder **ppIUriBuilder)
6341 UriBuilder *ret;
6343 TRACE("(%p %x %x %p)\n", pIUri, dwFlags, (DWORD)dwReserved, ppIUriBuilder);
6345 if(!ppIUriBuilder)
6346 return E_POINTER;
6348 ret = heap_alloc_zero(sizeof(UriBuilder));
6349 if(!ret)
6350 return E_OUTOFMEMORY;
6352 ret->IUriBuilder_iface.lpVtbl = &UriBuilderVtbl;
6353 ret->ref = 1;
6355 if(pIUri) {
6356 Uri *uri;
6358 if((uri = get_uri_obj(pIUri))) {
6359 if(!uri->create_flags)
6360 return E_UNEXPECTED;
6361 IUri_AddRef(pIUri);
6362 ret->uri = uri;
6364 if(uri->has_port)
6365 /* Windows doesn't set 'has_port' to TRUE in this case. */
6366 ret->port = uri->port;
6368 } else {
6369 heap_free(ret);
6370 *ppIUriBuilder = NULL;
6371 FIXME("(%p %x %x %p): Unknown IUri types not supported yet.\n", pIUri, dwFlags,
6372 (DWORD)dwReserved, ppIUriBuilder);
6373 return E_NOTIMPL;
6377 *ppIUriBuilder = &ret->IUriBuilder_iface;
6378 return S_OK;
6381 /* Merges the base path with the relative path and stores the resulting path
6382 * and path len in 'result' and 'result_len'.
6384 static HRESULT merge_paths(parse_data *data, const WCHAR *base, DWORD base_len, const WCHAR *relative,
6385 DWORD relative_len, WCHAR **result, DWORD *result_len, DWORD flags)
6387 const WCHAR *end = NULL;
6388 DWORD base_copy_len = 0;
6389 WCHAR *ptr;
6391 if(base_len) {
6392 if(data->scheme_type == URL_SCHEME_MK && *relative == '/') {
6393 /* Find '::' segment */
6394 for(end = base; end < base+base_len-1; end++) {
6395 if(end[0] == ':' && end[1] == ':') {
6396 end++;
6397 break;
6401 /* If not found, try finding the end of @xxx: */
6402 if(end == base+base_len-1)
6403 end = *base == '@' ? memchr(base, ':', base_len) : NULL;
6404 }else {
6405 /* Find the characters that will be copied over from the base path. */
6406 end = memrchrW(base, '/', base_len);
6407 if(!end && data->scheme_type == URL_SCHEME_FILE)
6408 /* Try looking for a '\\'. */
6409 end = memrchrW(base, '\\', base_len);
6413 if(end) {
6414 base_copy_len = (end+1)-base;
6415 *result = heap_alloc((base_copy_len+relative_len+1)*sizeof(WCHAR));
6416 } else
6417 *result = heap_alloc((relative_len+1)*sizeof(WCHAR));
6419 if(!(*result)) {
6420 *result_len = 0;
6421 return E_OUTOFMEMORY;
6424 ptr = *result;
6425 if(end) {
6426 memcpy(ptr, base, base_copy_len*sizeof(WCHAR));
6427 ptr += base_copy_len;
6430 memcpy(ptr, relative, relative_len*sizeof(WCHAR));
6431 ptr += relative_len;
6432 *ptr = '\0';
6434 *result_len = (ptr-*result);
6435 TRACE("ret %s\n", debugstr_wn(*result, *result_len));
6436 return S_OK;
6439 static HRESULT combine_uri(Uri *base, Uri *relative, DWORD flags, IUri **result, DWORD extras) {
6440 Uri *ret;
6441 HRESULT hr;
6442 parse_data data;
6443 Uri *proc_uri = base;
6444 DWORD create_flags = 0, len = 0;
6446 memset(&data, 0, sizeof(parse_data));
6448 /* Base case is when the relative Uri has a scheme name,
6449 * if it does, then 'result' will contain the same data
6450 * as the relative Uri.
6452 if(relative->scheme_start > -1) {
6453 data.uri = SysAllocString(relative->raw_uri);
6454 if(!data.uri) {
6455 *result = NULL;
6456 return E_OUTOFMEMORY;
6459 parse_uri(&data, 0);
6461 hr = Uri_Construct(NULL, (void**)&ret);
6462 if(FAILED(hr)) {
6463 *result = NULL;
6464 return hr;
6467 if(extras & COMBINE_URI_FORCE_FLAG_USE) {
6468 if(flags & URL_DONT_SIMPLIFY)
6469 create_flags |= Uri_CREATE_NO_CANONICALIZE;
6470 if(flags & URL_DONT_UNESCAPE_EXTRA_INFO)
6471 create_flags |= Uri_CREATE_NO_DECODE_EXTRA_INFO;
6474 ret->raw_uri = data.uri;
6475 hr = canonicalize_uri(&data, ret, create_flags);
6476 if(FAILED(hr)) {
6477 IUri_Release(&ret->IUri_iface);
6478 *result = NULL;
6479 return hr;
6482 apply_default_flags(&create_flags);
6483 ret->create_flags = create_flags;
6485 *result = &ret->IUri_iface;
6486 } else {
6487 WCHAR *path = NULL;
6488 DWORD raw_flags = 0;
6490 if(base->scheme_start > -1) {
6491 data.scheme = base->canon_uri+base->scheme_start;
6492 data.scheme_len = base->scheme_len;
6493 data.scheme_type = base->scheme_type;
6494 } else {
6495 data.is_relative = TRUE;
6496 data.scheme_type = URL_SCHEME_UNKNOWN;
6497 create_flags |= Uri_CREATE_ALLOW_RELATIVE;
6500 if(relative->authority_start > -1)
6501 proc_uri = relative;
6503 if(proc_uri->authority_start > -1) {
6504 if(proc_uri->userinfo_start > -1 && proc_uri->userinfo_split != 0) {
6505 data.username = proc_uri->canon_uri+proc_uri->userinfo_start;
6506 data.username_len = (proc_uri->userinfo_split > -1) ? proc_uri->userinfo_split : proc_uri->userinfo_len;
6509 if(proc_uri->userinfo_split > -1) {
6510 data.password = proc_uri->canon_uri+proc_uri->userinfo_start+proc_uri->userinfo_split+1;
6511 data.password_len = proc_uri->userinfo_len-proc_uri->userinfo_split-1;
6514 if(proc_uri->host_start > -1) {
6515 data.host = proc_uri->canon_uri+proc_uri->host_start;
6516 data.host_len = proc_uri->host_len;
6517 data.host_type = proc_uri->host_type;
6520 if(proc_uri->has_port) {
6521 data.has_port = TRUE;
6522 data.port_value = proc_uri->port;
6524 } else if(base->scheme_type != URL_SCHEME_FILE)
6525 data.is_opaque = TRUE;
6527 if(proc_uri == relative || relative->path_start == -1 || !relative->path_len) {
6528 if(proc_uri->path_start > -1) {
6529 data.path = proc_uri->canon_uri+proc_uri->path_start;
6530 data.path_len = proc_uri->path_len;
6531 } else if(!data.is_opaque) {
6532 /* Just set the path as a '/' if the base didn't have
6533 * one and if it's an hierarchical URI.
6535 static const WCHAR slashW[] = {'/',0};
6536 data.path = slashW;
6537 data.path_len = 1;
6540 if(relative->query_start > -1)
6541 proc_uri = relative;
6543 if(proc_uri->query_start > -1) {
6544 data.query = proc_uri->canon_uri+proc_uri->query_start;
6545 data.query_len = proc_uri->query_len;
6547 } else {
6548 const WCHAR *ptr, **pptr;
6549 DWORD path_offset = 0, path_len = 0;
6551 /* There's two possibilities on what will happen to the path component
6552 * of the result IUri. First, if the relative path begins with a '/'
6553 * then the resulting path will just be the relative path. Second, if
6554 * relative path doesn't begin with a '/' then the base path and relative
6555 * path are merged together.
6557 if(relative->path_len && *(relative->canon_uri+relative->path_start) == '/' && data.scheme_type != URL_SCHEME_MK) {
6558 WCHAR *tmp = NULL;
6559 BOOL copy_drive_path = FALSE;
6561 /* If the relative IUri's path starts with a '/', then we
6562 * don't use the base IUri's path. Unless the base IUri
6563 * is a file URI, in which case it uses the drive path of
6564 * the base IUri (if it has any) in the new path.
6566 if(base->scheme_type == URL_SCHEME_FILE) {
6567 if(base->path_len > 3 && *(base->canon_uri+base->path_start) == '/' &&
6568 is_drive_path(base->canon_uri+base->path_start+1)) {
6569 path_len += 3;
6570 copy_drive_path = TRUE;
6574 path_len += relative->path_len;
6576 path = heap_alloc((path_len+1)*sizeof(WCHAR));
6577 if(!path) {
6578 *result = NULL;
6579 return E_OUTOFMEMORY;
6582 tmp = path;
6584 /* Copy the base paths, drive path over. */
6585 if(copy_drive_path) {
6586 memcpy(tmp, base->canon_uri+base->path_start, 3*sizeof(WCHAR));
6587 tmp += 3;
6590 memcpy(tmp, relative->canon_uri+relative->path_start, relative->path_len*sizeof(WCHAR));
6591 path[path_len] = '\0';
6592 } else {
6593 /* Merge the base path with the relative path. */
6594 hr = merge_paths(&data, base->canon_uri+base->path_start, base->path_len,
6595 relative->canon_uri+relative->path_start, relative->path_len,
6596 &path, &path_len, flags);
6597 if(FAILED(hr)) {
6598 *result = NULL;
6599 return hr;
6602 /* If the resulting IUri is a file URI, the drive path isn't
6603 * reduced out when the dot segments are removed.
6605 if(path_len >= 3 && data.scheme_type == URL_SCHEME_FILE && !data.host) {
6606 if(*path == '/' && is_drive_path(path+1))
6607 path_offset = 2;
6608 else if(is_drive_path(path))
6609 path_offset = 1;
6613 /* Check if the dot segments need to be removed from the path. */
6614 if(!(flags & URL_DONT_SIMPLIFY) && !data.is_opaque) {
6615 DWORD offset = (path_offset > 0) ? path_offset+1 : 0;
6616 DWORD new_len = remove_dot_segments(path+offset,path_len-offset);
6618 if(new_len != path_len) {
6619 WCHAR *tmp = heap_realloc(path, (offset+new_len+1)*sizeof(WCHAR));
6620 if(!tmp) {
6621 heap_free(path);
6622 *result = NULL;
6623 return E_OUTOFMEMORY;
6626 tmp[new_len+offset] = '\0';
6627 path = tmp;
6628 path_len = new_len+offset;
6632 if(relative->query_start > -1) {
6633 data.query = relative->canon_uri+relative->query_start;
6634 data.query_len = relative->query_len;
6637 /* Make sure the path component is valid. */
6638 ptr = path;
6639 pptr = &ptr;
6640 if((data.is_opaque && !parse_path_opaque(pptr, &data, 0)) ||
6641 (!data.is_opaque && !parse_path_hierarchical(pptr, &data, 0))) {
6642 heap_free(path);
6643 *result = NULL;
6644 return E_INVALIDARG;
6648 if(relative->fragment_start > -1) {
6649 data.fragment = relative->canon_uri+relative->fragment_start;
6650 data.fragment_len = relative->fragment_len;
6653 if(flags & URL_DONT_SIMPLIFY)
6654 raw_flags |= RAW_URI_FORCE_PORT_DISP;
6655 if(flags & URL_FILE_USE_PATHURL)
6656 raw_flags |= RAW_URI_CONVERT_TO_DOS_PATH;
6658 len = generate_raw_uri(&data, data.uri, raw_flags);
6659 data.uri = SysAllocStringLen(NULL, len);
6660 if(!data.uri) {
6661 heap_free(path);
6662 *result = NULL;
6663 return E_OUTOFMEMORY;
6666 generate_raw_uri(&data, data.uri, raw_flags);
6668 hr = Uri_Construct(NULL, (void**)&ret);
6669 if(FAILED(hr)) {
6670 SysFreeString(data.uri);
6671 heap_free(path);
6672 *result = NULL;
6673 return hr;
6676 if(flags & URL_DONT_SIMPLIFY)
6677 create_flags |= Uri_CREATE_NO_CANONICALIZE;
6678 if(flags & URL_FILE_USE_PATHURL)
6679 create_flags |= Uri_CREATE_FILE_USE_DOS_PATH;
6681 ret->raw_uri = data.uri;
6682 hr = canonicalize_uri(&data, ret, create_flags);
6683 if(FAILED(hr)) {
6684 IUri_Release(&ret->IUri_iface);
6685 *result = NULL;
6686 return hr;
6689 if(flags & URL_DONT_SIMPLIFY)
6690 ret->display_modifiers |= URI_DISPLAY_NO_DEFAULT_PORT_AUTH;
6692 apply_default_flags(&create_flags);
6693 ret->create_flags = create_flags;
6694 *result = &ret->IUri_iface;
6696 heap_free(path);
6699 return S_OK;
6702 /***********************************************************************
6703 * CoInternetCombineIUri (urlmon.@)
6705 HRESULT WINAPI CoInternetCombineIUri(IUri *pBaseUri, IUri *pRelativeUri, DWORD dwCombineFlags,
6706 IUri **ppCombinedUri, DWORD_PTR dwReserved)
6708 HRESULT hr;
6709 IInternetProtocolInfo *info;
6710 Uri *relative, *base;
6711 TRACE("(%p %p %x %p %x)\n", pBaseUri, pRelativeUri, dwCombineFlags, ppCombinedUri, (DWORD)dwReserved);
6713 if(!ppCombinedUri)
6714 return E_INVALIDARG;
6716 if(!pBaseUri || !pRelativeUri) {
6717 *ppCombinedUri = NULL;
6718 return E_INVALIDARG;
6721 relative = get_uri_obj(pRelativeUri);
6722 base = get_uri_obj(pBaseUri);
6723 if(!relative || !base) {
6724 *ppCombinedUri = NULL;
6725 FIXME("(%p %p %x %p %x) Unknown IUri types not supported yet.\n",
6726 pBaseUri, pRelativeUri, dwCombineFlags, ppCombinedUri, (DWORD)dwReserved);
6727 return E_NOTIMPL;
6730 info = get_protocol_info(base->canon_uri);
6731 if(info) {
6732 WCHAR result[INTERNET_MAX_URL_LENGTH+1];
6733 DWORD result_len = 0;
6735 hr = IInternetProtocolInfo_CombineUrl(info, base->canon_uri, relative->canon_uri, dwCombineFlags,
6736 result, INTERNET_MAX_URL_LENGTH+1, &result_len, 0);
6737 IInternetProtocolInfo_Release(info);
6738 if(SUCCEEDED(hr)) {
6739 hr = CreateUri(result, Uri_CREATE_ALLOW_RELATIVE, 0, ppCombinedUri);
6740 if(SUCCEEDED(hr))
6741 return hr;
6745 return combine_uri(base, relative, dwCombineFlags, ppCombinedUri, 0);
6748 /***********************************************************************
6749 * CoInternetCombineUrlEx (urlmon.@)
6751 HRESULT WINAPI CoInternetCombineUrlEx(IUri *pBaseUri, LPCWSTR pwzRelativeUrl, DWORD dwCombineFlags,
6752 IUri **ppCombinedUri, DWORD_PTR dwReserved)
6754 IUri *relative;
6755 Uri *base;
6756 HRESULT hr;
6757 IInternetProtocolInfo *info;
6759 TRACE("(%p %s %x %p %x) stub\n", pBaseUri, debugstr_w(pwzRelativeUrl), dwCombineFlags,
6760 ppCombinedUri, (DWORD)dwReserved);
6762 if(!ppCombinedUri)
6763 return E_POINTER;
6765 if(!pwzRelativeUrl) {
6766 *ppCombinedUri = NULL;
6767 return E_UNEXPECTED;
6770 if(!pBaseUri) {
6771 *ppCombinedUri = NULL;
6772 return E_INVALIDARG;
6775 base = get_uri_obj(pBaseUri);
6776 if(!base) {
6777 *ppCombinedUri = NULL;
6778 FIXME("(%p %s %x %p %x) Unknown IUri's not supported yet.\n", pBaseUri, debugstr_w(pwzRelativeUrl),
6779 dwCombineFlags, ppCombinedUri, (DWORD)dwReserved);
6780 return E_NOTIMPL;
6783 info = get_protocol_info(base->canon_uri);
6784 if(info) {
6785 WCHAR result[INTERNET_MAX_URL_LENGTH+1];
6786 DWORD result_len = 0;
6788 hr = IInternetProtocolInfo_CombineUrl(info, base->canon_uri, pwzRelativeUrl, dwCombineFlags,
6789 result, INTERNET_MAX_URL_LENGTH+1, &result_len, 0);
6790 IInternetProtocolInfo_Release(info);
6791 if(SUCCEEDED(hr)) {
6792 hr = CreateUri(result, Uri_CREATE_ALLOW_RELATIVE, 0, ppCombinedUri);
6793 if(SUCCEEDED(hr))
6794 return hr;
6798 hr = CreateUri(pwzRelativeUrl, Uri_CREATE_ALLOW_RELATIVE, 0, &relative);
6799 if(FAILED(hr)) {
6800 *ppCombinedUri = NULL;
6801 return hr;
6804 hr = combine_uri(base, get_uri_obj(relative), dwCombineFlags, ppCombinedUri, COMBINE_URI_FORCE_FLAG_USE);
6806 IUri_Release(relative);
6807 return hr;
6810 static HRESULT parse_canonicalize(const Uri *uri, DWORD flags, LPWSTR output,
6811 DWORD output_len, DWORD *result_len)
6813 const WCHAR *ptr = NULL;
6814 WCHAR *path = NULL;
6815 const WCHAR **pptr;
6816 WCHAR buffer[INTERNET_MAX_URL_LENGTH+1];
6817 DWORD len = 0;
6818 BOOL reduce_path;
6820 /* URL_UNESCAPE only has effect if none of the URL_ESCAPE flags are set. */
6821 const BOOL allow_unescape = !(flags & URL_ESCAPE_UNSAFE) &&
6822 !(flags & URL_ESCAPE_SPACES_ONLY) &&
6823 !(flags & URL_ESCAPE_PERCENT);
6826 /* Check if the dot segments need to be removed from the
6827 * path component.
6829 if(uri->scheme_start > -1 && uri->path_start > -1) {
6830 ptr = uri->canon_uri+uri->scheme_start+uri->scheme_len+1;
6831 pptr = &ptr;
6833 reduce_path = !(flags & URL_NO_META) &&
6834 !(flags & URL_DONT_SIMPLIFY) &&
6835 ptr && check_hierarchical(pptr);
6837 for(ptr = uri->canon_uri; ptr < uri->canon_uri+uri->canon_len; ++ptr) {
6838 BOOL do_default_action = TRUE;
6840 /* Keep track of the path if we need to remove dot segments from
6841 * it later.
6843 if(reduce_path && !path && ptr == uri->canon_uri+uri->path_start)
6844 path = buffer+len;
6846 /* Check if it's time to reduce the path. */
6847 if(reduce_path && ptr == uri->canon_uri+uri->path_start+uri->path_len) {
6848 DWORD current_path_len = (buffer+len) - path;
6849 DWORD new_path_len = remove_dot_segments(path, current_path_len);
6851 /* Update the current length. */
6852 len -= (current_path_len-new_path_len);
6853 reduce_path = FALSE;
6856 if(*ptr == '%') {
6857 const WCHAR decoded = decode_pct_val(ptr);
6858 if(decoded) {
6859 if(allow_unescape && (flags & URL_UNESCAPE)) {
6860 buffer[len++] = decoded;
6861 ptr += 2;
6862 do_default_action = FALSE;
6866 /* See if %'s needed to encoded. */
6867 if(do_default_action && (flags & URL_ESCAPE_PERCENT)) {
6868 pct_encode_val(*ptr, buffer+len);
6869 len += 3;
6870 do_default_action = FALSE;
6872 } else if(*ptr == ' ') {
6873 if((flags & URL_ESCAPE_SPACES_ONLY) &&
6874 !(flags & URL_ESCAPE_UNSAFE)) {
6875 pct_encode_val(*ptr, buffer+len);
6876 len += 3;
6877 do_default_action = FALSE;
6879 } else if(!is_reserved(*ptr) && !is_unreserved(*ptr)) {
6880 if(flags & URL_ESCAPE_UNSAFE) {
6881 pct_encode_val(*ptr, buffer+len);
6882 len += 3;
6883 do_default_action = FALSE;
6887 if(do_default_action)
6888 buffer[len++] = *ptr;
6891 /* Sometimes the path is the very last component of the IUri, so
6892 * see if the dot segments need to be reduced now.
6894 if(reduce_path && path) {
6895 DWORD current_path_len = (buffer+len) - path;
6896 DWORD new_path_len = remove_dot_segments(path, current_path_len);
6898 /* Update the current length. */
6899 len -= (current_path_len-new_path_len);
6902 buffer[len++] = 0;
6904 /* The null terminator isn't included in the length. */
6905 *result_len = len-1;
6906 if(len > output_len)
6907 return STRSAFE_E_INSUFFICIENT_BUFFER;
6908 else
6909 memcpy(output, buffer, len*sizeof(WCHAR));
6911 return S_OK;
6914 static HRESULT parse_friendly(IUri *uri, LPWSTR output, DWORD output_len,
6915 DWORD *result_len)
6917 HRESULT hr;
6918 DWORD display_len;
6919 BSTR display;
6921 hr = IUri_GetPropertyLength(uri, Uri_PROPERTY_DISPLAY_URI, &display_len, 0);
6922 if(FAILED(hr)) {
6923 *result_len = 0;
6924 return hr;
6927 *result_len = display_len;
6928 if(display_len+1 > output_len)
6929 return STRSAFE_E_INSUFFICIENT_BUFFER;
6931 hr = IUri_GetDisplayUri(uri, &display);
6932 if(FAILED(hr)) {
6933 *result_len = 0;
6934 return hr;
6937 memcpy(output, display, (display_len+1)*sizeof(WCHAR));
6938 SysFreeString(display);
6939 return S_OK;
6942 static HRESULT parse_rootdocument(const Uri *uri, LPWSTR output, DWORD output_len,
6943 DWORD *result_len)
6945 static const WCHAR colon_slashesW[] = {':','/','/'};
6947 WCHAR *ptr;
6948 DWORD len = 0;
6950 /* Windows only returns the root document if the URI has an authority
6951 * and it's not an unknown scheme type or a file scheme type.
6953 if(uri->authority_start == -1 ||
6954 uri->scheme_type == URL_SCHEME_UNKNOWN ||
6955 uri->scheme_type == URL_SCHEME_FILE) {
6956 *result_len = 0;
6957 if(!output_len)
6958 return STRSAFE_E_INSUFFICIENT_BUFFER;
6960 output[0] = 0;
6961 return S_OK;
6964 len = uri->scheme_len+uri->authority_len;
6965 /* For the "://" and '/' which will be added. */
6966 len += 4;
6968 if(len+1 > output_len) {
6969 *result_len = len;
6970 return STRSAFE_E_INSUFFICIENT_BUFFER;
6973 ptr = output;
6974 memcpy(ptr, uri->canon_uri+uri->scheme_start, uri->scheme_len*sizeof(WCHAR));
6976 /* Add the "://". */
6977 ptr += uri->scheme_len;
6978 memcpy(ptr, colon_slashesW, sizeof(colon_slashesW));
6980 /* Add the authority. */
6981 ptr += sizeof(colon_slashesW)/sizeof(WCHAR);
6982 memcpy(ptr, uri->canon_uri+uri->authority_start, uri->authority_len*sizeof(WCHAR));
6984 /* Add the '/' after the authority. */
6985 ptr += uri->authority_len;
6986 *ptr = '/';
6987 ptr[1] = 0;
6989 *result_len = len;
6990 return S_OK;
6993 static HRESULT parse_document(const Uri *uri, LPWSTR output, DWORD output_len,
6994 DWORD *result_len)
6996 DWORD len = 0;
6998 /* It has to be a known scheme type, but, it can't be a file
6999 * scheme. It also has to hierarchical.
7001 if(uri->scheme_type == URL_SCHEME_UNKNOWN ||
7002 uri->scheme_type == URL_SCHEME_FILE ||
7003 uri->authority_start == -1) {
7004 *result_len = 0;
7005 if(output_len < 1)
7006 return STRSAFE_E_INSUFFICIENT_BUFFER;
7008 output[0] = 0;
7009 return S_OK;
7012 if(uri->fragment_start > -1)
7013 len = uri->fragment_start;
7014 else
7015 len = uri->canon_len;
7017 *result_len = len;
7018 if(len+1 > output_len)
7019 return STRSAFE_E_INSUFFICIENT_BUFFER;
7021 memcpy(output, uri->canon_uri, len*sizeof(WCHAR));
7022 output[len] = 0;
7023 return S_OK;
7026 static HRESULT parse_path_from_url(const Uri *uri, LPWSTR output, DWORD output_len,
7027 DWORD *result_len)
7029 const WCHAR *path_ptr;
7030 WCHAR buffer[INTERNET_MAX_URL_LENGTH+1];
7031 WCHAR *ptr;
7033 if(uri->scheme_type != URL_SCHEME_FILE) {
7034 *result_len = 0;
7035 if(output_len > 0)
7036 output[0] = 0;
7037 return E_INVALIDARG;
7040 ptr = buffer;
7041 if(uri->host_start > -1) {
7042 static const WCHAR slash_slashW[] = {'\\','\\'};
7044 memcpy(ptr, slash_slashW, sizeof(slash_slashW));
7045 ptr += sizeof(slash_slashW)/sizeof(WCHAR);
7046 memcpy(ptr, uri->canon_uri+uri->host_start, uri->host_len*sizeof(WCHAR));
7047 ptr += uri->host_len;
7050 path_ptr = uri->canon_uri+uri->path_start;
7051 if(uri->path_len > 3 && *path_ptr == '/' && is_drive_path(path_ptr+1))
7052 /* Skip past the '/' in front of the drive path. */
7053 ++path_ptr;
7055 for(; path_ptr < uri->canon_uri+uri->path_start+uri->path_len; ++path_ptr, ++ptr) {
7056 BOOL do_default_action = TRUE;
7058 if(*path_ptr == '%') {
7059 const WCHAR decoded = decode_pct_val(path_ptr);
7060 if(decoded) {
7061 *ptr = decoded;
7062 path_ptr += 2;
7063 do_default_action = FALSE;
7065 } else if(*path_ptr == '/') {
7066 *ptr = '\\';
7067 do_default_action = FALSE;
7070 if(do_default_action)
7071 *ptr = *path_ptr;
7074 *ptr = 0;
7076 *result_len = ptr-buffer;
7077 if(*result_len+1 > output_len)
7078 return STRSAFE_E_INSUFFICIENT_BUFFER;
7080 memcpy(output, buffer, (*result_len+1)*sizeof(WCHAR));
7081 return S_OK;
7084 static HRESULT parse_url_from_path(IUri *uri, LPWSTR output, DWORD output_len,
7085 DWORD *result_len)
7087 HRESULT hr;
7088 BSTR received;
7089 DWORD len = 0;
7091 hr = IUri_GetPropertyLength(uri, Uri_PROPERTY_ABSOLUTE_URI, &len, 0);
7092 if(FAILED(hr)) {
7093 *result_len = 0;
7094 return hr;
7097 *result_len = len;
7098 if(len+1 > output_len)
7099 return STRSAFE_E_INSUFFICIENT_BUFFER;
7101 hr = IUri_GetAbsoluteUri(uri, &received);
7102 if(FAILED(hr)) {
7103 *result_len = 0;
7104 return hr;
7107 memcpy(output, received, (len+1)*sizeof(WCHAR));
7108 SysFreeString(received);
7110 return S_OK;
7113 static HRESULT parse_schema(IUri *uri, LPWSTR output, DWORD output_len,
7114 DWORD *result_len)
7116 HRESULT hr;
7117 DWORD len;
7118 BSTR received;
7120 hr = IUri_GetPropertyLength(uri, Uri_PROPERTY_SCHEME_NAME, &len, 0);
7121 if(FAILED(hr)) {
7122 *result_len = 0;
7123 return hr;
7126 *result_len = len;
7127 if(len+1 > output_len)
7128 return STRSAFE_E_INSUFFICIENT_BUFFER;
7130 hr = IUri_GetSchemeName(uri, &received);
7131 if(FAILED(hr)) {
7132 *result_len = 0;
7133 return hr;
7136 memcpy(output, received, (len+1)*sizeof(WCHAR));
7137 SysFreeString(received);
7139 return S_OK;
7142 static HRESULT parse_site(IUri *uri, LPWSTR output, DWORD output_len, DWORD *result_len)
7144 HRESULT hr;
7145 DWORD len;
7146 BSTR received;
7148 hr = IUri_GetPropertyLength(uri, Uri_PROPERTY_HOST, &len, 0);
7149 if(FAILED(hr)) {
7150 *result_len = 0;
7151 return hr;
7154 *result_len = len;
7155 if(len+1 > output_len)
7156 return STRSAFE_E_INSUFFICIENT_BUFFER;
7158 hr = IUri_GetHost(uri, &received);
7159 if(FAILED(hr)) {
7160 *result_len = 0;
7161 return hr;
7164 memcpy(output, received, (len+1)*sizeof(WCHAR));
7165 SysFreeString(received);
7167 return S_OK;
7170 static HRESULT parse_domain(IUri *uri, LPWSTR output, DWORD output_len, DWORD *result_len)
7172 HRESULT hr;
7173 DWORD len;
7174 BSTR received;
7176 hr = IUri_GetPropertyLength(uri, Uri_PROPERTY_DOMAIN, &len, 0);
7177 if(FAILED(hr)) {
7178 *result_len = 0;
7179 return hr;
7182 *result_len = len;
7183 if(len+1 > output_len)
7184 return STRSAFE_E_INSUFFICIENT_BUFFER;
7186 hr = IUri_GetDomain(uri, &received);
7187 if(FAILED(hr)) {
7188 *result_len = 0;
7189 return hr;
7192 memcpy(output, received, (len+1)*sizeof(WCHAR));
7193 SysFreeString(received);
7195 return S_OK;
7198 static HRESULT parse_anchor(IUri *uri, LPWSTR output, DWORD output_len, DWORD *result_len)
7200 HRESULT hr;
7201 DWORD len;
7202 BSTR received;
7204 hr = IUri_GetPropertyLength(uri, Uri_PROPERTY_FRAGMENT, &len, 0);
7205 if(FAILED(hr)) {
7206 *result_len = 0;
7207 return hr;
7210 *result_len = len;
7211 if(len+1 > output_len)
7212 return STRSAFE_E_INSUFFICIENT_BUFFER;
7214 hr = IUri_GetFragment(uri, &received);
7215 if(FAILED(hr)) {
7216 *result_len = 0;
7217 return hr;
7220 memcpy(output, received, (len+1)*sizeof(WCHAR));
7221 SysFreeString(received);
7223 return S_OK;
7226 /***********************************************************************
7227 * CoInternetParseIUri (urlmon.@)
7229 HRESULT WINAPI CoInternetParseIUri(IUri *pIUri, PARSEACTION ParseAction, DWORD dwFlags,
7230 LPWSTR pwzResult, DWORD cchResult, DWORD *pcchResult,
7231 DWORD_PTR dwReserved)
7233 HRESULT hr;
7234 Uri *uri;
7235 IInternetProtocolInfo *info;
7237 TRACE("(%p %d %x %p %d %p %x)\n", pIUri, ParseAction, dwFlags, pwzResult,
7238 cchResult, pcchResult, (DWORD)dwReserved);
7240 if(!pcchResult)
7241 return E_POINTER;
7243 if(!pwzResult || !pIUri) {
7244 *pcchResult = 0;
7245 return E_INVALIDARG;
7248 if(!(uri = get_uri_obj(pIUri))) {
7249 *pcchResult = 0;
7250 FIXME("(%p %d %x %p %d %p %x) Unknown IUri's not supported for this action.\n",
7251 pIUri, ParseAction, dwFlags, pwzResult, cchResult, pcchResult, (DWORD)dwReserved);
7252 return E_NOTIMPL;
7255 info = get_protocol_info(uri->canon_uri);
7256 if(info) {
7257 hr = IInternetProtocolInfo_ParseUrl(info, uri->canon_uri, ParseAction, dwFlags,
7258 pwzResult, cchResult, pcchResult, 0);
7259 IInternetProtocolInfo_Release(info);
7260 if(SUCCEEDED(hr)) return hr;
7263 switch(ParseAction) {
7264 case PARSE_CANONICALIZE:
7265 hr = parse_canonicalize(uri, dwFlags, pwzResult, cchResult, pcchResult);
7266 break;
7267 case PARSE_FRIENDLY:
7268 hr = parse_friendly(pIUri, pwzResult, cchResult, pcchResult);
7269 break;
7270 case PARSE_ROOTDOCUMENT:
7271 hr = parse_rootdocument(uri, pwzResult, cchResult, pcchResult);
7272 break;
7273 case PARSE_DOCUMENT:
7274 hr = parse_document(uri, pwzResult, cchResult, pcchResult);
7275 break;
7276 case PARSE_PATH_FROM_URL:
7277 hr = parse_path_from_url(uri, pwzResult, cchResult, pcchResult);
7278 break;
7279 case PARSE_URL_FROM_PATH:
7280 hr = parse_url_from_path(pIUri, pwzResult, cchResult, pcchResult);
7281 break;
7282 case PARSE_SCHEMA:
7283 hr = parse_schema(pIUri, pwzResult, cchResult, pcchResult);
7284 break;
7285 case PARSE_SITE:
7286 hr = parse_site(pIUri, pwzResult, cchResult, pcchResult);
7287 break;
7288 case PARSE_DOMAIN:
7289 hr = parse_domain(pIUri, pwzResult, cchResult, pcchResult);
7290 break;
7291 case PARSE_LOCATION:
7292 case PARSE_ANCHOR:
7293 hr = parse_anchor(pIUri, pwzResult, cchResult, pcchResult);
7294 break;
7295 case PARSE_SECURITY_URL:
7296 case PARSE_MIME:
7297 case PARSE_SERVER:
7298 case PARSE_SECURITY_DOMAIN:
7299 *pcchResult = 0;
7300 hr = E_FAIL;
7301 break;
7302 default:
7303 *pcchResult = 0;
7304 hr = E_NOTIMPL;
7305 FIXME("(%p %d %x %p %d %p %x) Partial stub.\n", pIUri, ParseAction, dwFlags,
7306 pwzResult, cchResult, pcchResult, (DWORD)dwReserved);
7309 return hr;