crypt32: Fix adding trusted roots to store.
[wine/zf.git] / dlls / urlmon / uri.c
blob7f5dc1550270adb083135bfd6a115cbe23c245e1
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>
21 #include <wchar.h>
23 #include "urlmon_main.h"
24 #include "wine/debug.h"
26 #define NO_SHLWAPI_REG
27 #include "shlwapi.h"
29 #include "strsafe.h"
31 #define URI_DISPLAY_NO_ABSOLUTE_URI 0x1
32 #define URI_DISPLAY_NO_DEFAULT_PORT_AUTH 0x2
34 #define ALLOW_NULL_TERM_SCHEME 0x01
35 #define ALLOW_NULL_TERM_USER_NAME 0x02
36 #define ALLOW_NULL_TERM_PASSWORD 0x04
37 #define ALLOW_BRACKETLESS_IP_LITERAL 0x08
38 #define SKIP_IP_FUTURE_CHECK 0x10
39 #define IGNORE_PORT_DELIMITER 0x20
41 #define RAW_URI_FORCE_PORT_DISP 0x1
42 #define RAW_URI_CONVERT_TO_DOS_PATH 0x2
44 #define COMBINE_URI_FORCE_FLAG_USE 0x1
46 WINE_DEFAULT_DEBUG_CHANNEL(urlmon);
48 static const IID IID_IUriObj = {0x4b364760,0x9f51,0x11df,{0x98,0x1c,0x08,0x00,0x20,0x0c,0x9a,0x66}};
50 typedef struct {
51 IUri IUri_iface;
52 IUriBuilderFactory IUriBuilderFactory_iface;
53 IPersistStream IPersistStream_iface;
54 IMarshal IMarshal_iface;
56 LONG ref;
58 BSTR raw_uri;
60 /* Information about the canonicalized URI's buffer. */
61 WCHAR *canon_uri;
62 DWORD canon_size;
63 DWORD canon_len;
64 BOOL display_modifiers;
65 DWORD create_flags;
67 INT scheme_start;
68 DWORD scheme_len;
69 URL_SCHEME scheme_type;
71 INT userinfo_start;
72 DWORD userinfo_len;
73 INT userinfo_split;
75 INT host_start;
76 DWORD host_len;
77 Uri_HOST_TYPE host_type;
79 INT port_offset;
80 DWORD port;
81 BOOL has_port;
83 INT authority_start;
84 DWORD authority_len;
86 INT domain_offset;
88 INT path_start;
89 DWORD path_len;
90 INT extension_offset;
92 INT query_start;
93 DWORD query_len;
95 INT fragment_start;
96 DWORD fragment_len;
97 } Uri;
99 typedef struct {
100 IUriBuilder IUriBuilder_iface;
101 LONG ref;
103 Uri *uri;
104 DWORD modified_props;
106 WCHAR *fragment;
107 DWORD fragment_len;
109 WCHAR *host;
110 DWORD host_len;
112 WCHAR *password;
113 DWORD password_len;
115 WCHAR *path;
116 DWORD path_len;
118 BOOL has_port;
119 DWORD port;
121 WCHAR *query;
122 DWORD query_len;
124 WCHAR *scheme;
125 DWORD scheme_len;
127 WCHAR *username;
128 DWORD username_len;
129 } UriBuilder;
131 typedef struct {
132 const WCHAR *str;
133 DWORD len;
134 } h16;
136 typedef struct {
137 /* IPv6 addresses can hold up to 8 h16 components. */
138 h16 components[8];
139 DWORD h16_count;
141 /* An IPv6 can have 1 elision ("::"). */
142 const WCHAR *elision;
144 /* An IPv6 can contain 1 IPv4 address as the last 32bits of the address. */
145 const WCHAR *ipv4;
146 DWORD ipv4_len;
148 INT components_size;
149 INT elision_size;
150 } ipv6_address;
152 typedef struct {
153 BSTR uri;
155 BOOL is_relative;
156 BOOL is_opaque;
157 BOOL has_implicit_scheme;
158 BOOL has_implicit_ip;
159 UINT implicit_ipv4;
160 BOOL must_have_path;
162 const WCHAR *scheme;
163 DWORD scheme_len;
164 URL_SCHEME scheme_type;
166 const WCHAR *username;
167 DWORD username_len;
169 const WCHAR *password;
170 DWORD password_len;
172 const WCHAR *host;
173 DWORD host_len;
174 Uri_HOST_TYPE host_type;
176 BOOL has_ipv6;
177 ipv6_address ipv6_address;
179 BOOL has_port;
180 const WCHAR *port;
181 DWORD port_len;
182 DWORD port_value;
184 const WCHAR *path;
185 DWORD path_len;
187 const WCHAR *query;
188 DWORD query_len;
190 const WCHAR *fragment;
191 DWORD fragment_len;
192 } parse_data;
194 static const CHAR hexDigits[] = "0123456789ABCDEF";
196 /* List of scheme types/scheme names that are recognized by the IUri interface as of IE 7. */
197 static const struct {
198 URL_SCHEME scheme;
199 WCHAR scheme_name[16];
200 } recognized_schemes[] = {
201 {URL_SCHEME_FTP, {'f','t','p',0}},
202 {URL_SCHEME_HTTP, {'h','t','t','p',0}},
203 {URL_SCHEME_GOPHER, {'g','o','p','h','e','r',0}},
204 {URL_SCHEME_MAILTO, {'m','a','i','l','t','o',0}},
205 {URL_SCHEME_NEWS, {'n','e','w','s',0}},
206 {URL_SCHEME_NNTP, {'n','n','t','p',0}},
207 {URL_SCHEME_TELNET, {'t','e','l','n','e','t',0}},
208 {URL_SCHEME_WAIS, {'w','a','i','s',0}},
209 {URL_SCHEME_FILE, {'f','i','l','e',0}},
210 {URL_SCHEME_MK, {'m','k',0}},
211 {URL_SCHEME_HTTPS, {'h','t','t','p','s',0}},
212 {URL_SCHEME_SHELL, {'s','h','e','l','l',0}},
213 {URL_SCHEME_SNEWS, {'s','n','e','w','s',0}},
214 {URL_SCHEME_LOCAL, {'l','o','c','a','l',0}},
215 {URL_SCHEME_JAVASCRIPT, {'j','a','v','a','s','c','r','i','p','t',0}},
216 {URL_SCHEME_VBSCRIPT, {'v','b','s','c','r','i','p','t',0}},
217 {URL_SCHEME_ABOUT, {'a','b','o','u','t',0}},
218 {URL_SCHEME_RES, {'r','e','s',0}},
219 {URL_SCHEME_MSSHELLROOTED, {'m','s','-','s','h','e','l','l','-','r','o','o','t','e','d',0}},
220 {URL_SCHEME_MSSHELLIDLIST, {'m','s','-','s','h','e','l','l','-','i','d','l','i','s','t',0}},
221 {URL_SCHEME_MSHELP, {'h','c','p',0}},
222 {URL_SCHEME_WILDCARD, {'*',0}}
225 /* List of default ports Windows recognizes. */
226 static const struct {
227 URL_SCHEME scheme;
228 USHORT port;
229 } default_ports[] = {
230 {URL_SCHEME_FTP, 21},
231 {URL_SCHEME_HTTP, 80},
232 {URL_SCHEME_GOPHER, 70},
233 {URL_SCHEME_NNTP, 119},
234 {URL_SCHEME_TELNET, 23},
235 {URL_SCHEME_WAIS, 210},
236 {URL_SCHEME_HTTPS, 443},
239 /* List of 3-character top level domain names Windows seems to recognize.
240 * There might be more, but, these are the only ones I've found so far.
242 static const struct {
243 WCHAR tld_name[4];
244 } recognized_tlds[] = {
245 {{'c','o','m',0}},
246 {{'e','d','u',0}},
247 {{'g','o','v',0}},
248 {{'i','n','t',0}},
249 {{'m','i','l',0}},
250 {{'n','e','t',0}},
251 {{'o','r','g',0}}
254 static Uri *get_uri_obj(IUri *uri)
256 Uri *ret;
257 HRESULT hres;
259 hres = IUri_QueryInterface(uri, &IID_IUriObj, (void**)&ret);
260 return SUCCEEDED(hres) ? ret : NULL;
263 static inline BOOL is_alpha(WCHAR val) {
264 return ((val >= 'a' && val <= 'z') || (val >= 'A' && val <= 'Z'));
267 static inline BOOL is_num(WCHAR val) {
268 return (val >= '0' && val <= '9');
271 static inline BOOL is_drive_path(const WCHAR *str) {
272 return (is_alpha(str[0]) && (str[1] == ':' || str[1] == '|'));
275 static inline BOOL is_unc_path(const WCHAR *str) {
276 return (str[0] == '\\' && str[1] == '\\');
279 static inline BOOL is_forbidden_dos_path_char(WCHAR val) {
280 return (val == '>' || val == '<' || val == '\"');
283 /* A URI is implicitly a file path if it begins with
284 * a drive letter (e.g. X:) or starts with "\\" (UNC path).
286 static inline BOOL is_implicit_file_path(const WCHAR *str) {
287 return (is_unc_path(str) || (is_alpha(str[0]) && str[1] == ':'));
290 /* Checks if the URI is a hierarchical URI. A hierarchical
291 * URI is one that has "//" after the scheme.
293 static BOOL check_hierarchical(const WCHAR **ptr) {
294 const WCHAR *start = *ptr;
296 if(**ptr != '/')
297 return FALSE;
299 ++(*ptr);
300 if(**ptr != '/') {
301 *ptr = start;
302 return FALSE;
305 ++(*ptr);
306 return TRUE;
309 /* unreserved = ALPHA / DIGIT / "-" / "." / "_" / "~" */
310 static inline BOOL is_unreserved(WCHAR val) {
311 return (is_alpha(val) || is_num(val) || val == '-' || val == '.' ||
312 val == '_' || val == '~');
315 /* sub-delims = "!" / "$" / "&" / "'" / "(" / ")"
316 * / "*" / "+" / "," / ";" / "="
318 static inline BOOL is_subdelim(WCHAR val) {
319 return (val == '!' || val == '$' || val == '&' ||
320 val == '\'' || val == '(' || val == ')' ||
321 val == '*' || val == '+' || val == ',' ||
322 val == ';' || val == '=');
325 /* gen-delims = ":" / "/" / "?" / "#" / "[" / "]" / "@" */
326 static inline BOOL is_gendelim(WCHAR val) {
327 return (val == ':' || val == '/' || val == '?' ||
328 val == '#' || val == '[' || val == ']' ||
329 val == '@');
332 /* Characters that delimit the end of the authority
333 * section of a URI. Sometimes a '\\' is considered
334 * an authority delimiter.
336 static inline BOOL is_auth_delim(WCHAR val, BOOL acceptSlash) {
337 return (val == '#' || val == '/' || val == '?' ||
338 val == '\0' || (acceptSlash && val == '\\'));
341 /* reserved = gen-delims / sub-delims */
342 static inline BOOL is_reserved(WCHAR val) {
343 return (is_subdelim(val) || is_gendelim(val));
346 static inline BOOL is_hexdigit(WCHAR val) {
347 return ((val >= 'a' && val <= 'f') ||
348 (val >= 'A' && val <= 'F') ||
349 (val >= '0' && val <= '9'));
352 static inline BOOL is_path_delim(URL_SCHEME scheme, WCHAR val) {
353 return (!val || (val == '#' && scheme != URL_SCHEME_FILE) || val == '?');
356 static inline BOOL is_slash(WCHAR c)
358 return c == '/' || c == '\\';
361 static inline BOOL is_ascii(WCHAR c)
363 return c < 0x80;
366 static BOOL is_default_port(URL_SCHEME scheme, DWORD port) {
367 DWORD i;
369 for(i = 0; i < ARRAY_SIZE(default_ports); ++i) {
370 if(default_ports[i].scheme == scheme && default_ports[i].port)
371 return TRUE;
374 return FALSE;
377 /* List of schemes types Windows seems to expect to be hierarchical. */
378 static inline BOOL is_hierarchical_scheme(URL_SCHEME type) {
379 return(type == URL_SCHEME_HTTP || type == URL_SCHEME_FTP ||
380 type == URL_SCHEME_GOPHER || type == URL_SCHEME_NNTP ||
381 type == URL_SCHEME_TELNET || type == URL_SCHEME_WAIS ||
382 type == URL_SCHEME_FILE || type == URL_SCHEME_HTTPS ||
383 type == URL_SCHEME_RES);
386 /* Checks if 'flags' contains an invalid combination of Uri_CREATE flags. */
387 static inline BOOL has_invalid_flag_combination(DWORD flags) {
388 return((flags & Uri_CREATE_DECODE_EXTRA_INFO && flags & Uri_CREATE_NO_DECODE_EXTRA_INFO) ||
389 (flags & Uri_CREATE_CANONICALIZE && flags & Uri_CREATE_NO_CANONICALIZE) ||
390 (flags & Uri_CREATE_CRACK_UNKNOWN_SCHEMES && flags & Uri_CREATE_NO_CRACK_UNKNOWN_SCHEMES) ||
391 (flags & Uri_CREATE_PRE_PROCESS_HTML_URI && flags & Uri_CREATE_NO_PRE_PROCESS_HTML_URI) ||
392 (flags & Uri_CREATE_IE_SETTINGS && flags & Uri_CREATE_NO_IE_SETTINGS));
395 /* Applies each default Uri_CREATE flags to 'flags' if it
396 * doesn't cause a flag conflict.
398 static void apply_default_flags(DWORD *flags) {
399 if(!(*flags & Uri_CREATE_NO_CANONICALIZE))
400 *flags |= Uri_CREATE_CANONICALIZE;
401 if(!(*flags & Uri_CREATE_NO_DECODE_EXTRA_INFO))
402 *flags |= Uri_CREATE_DECODE_EXTRA_INFO;
403 if(!(*flags & Uri_CREATE_NO_CRACK_UNKNOWN_SCHEMES))
404 *flags |= Uri_CREATE_CRACK_UNKNOWN_SCHEMES;
405 if(!(*flags & Uri_CREATE_NO_PRE_PROCESS_HTML_URI))
406 *flags |= Uri_CREATE_PRE_PROCESS_HTML_URI;
407 if(!(*flags & Uri_CREATE_IE_SETTINGS))
408 *flags |= Uri_CREATE_NO_IE_SETTINGS;
411 /* Determines if the URI is hierarchical using the information already parsed into
412 * data and using the current location of parsing in the URI string.
414 * Windows considers a URI hierarchical if one of the following is true:
415 * A.) It's a wildcard scheme.
416 * B.) It's an implicit file scheme.
417 * C.) It's a known hierarchical scheme and it has two '\\' after the scheme name.
418 * (the '\\' will be converted into "//" during canonicalization).
419 * D.) "//" appears after the scheme name (or at the beginning if no scheme is given).
421 static inline BOOL is_hierarchical_uri(const WCHAR **ptr, const parse_data *data) {
422 const WCHAR *start = *ptr;
424 if(data->scheme_type == URL_SCHEME_WILDCARD)
425 return TRUE;
426 else if(data->scheme_type == URL_SCHEME_FILE && data->has_implicit_scheme)
427 return TRUE;
428 else if(is_hierarchical_scheme(data->scheme_type) && (*ptr)[0] == '\\' && (*ptr)[1] == '\\') {
429 *ptr += 2;
430 return TRUE;
431 } else if(data->scheme_type != URL_SCHEME_MAILTO && check_hierarchical(ptr))
432 return TRUE;
434 *ptr = start;
435 return FALSE;
438 /* Computes the size of the given IPv6 address.
439 * Each h16 component is 16 bits. If there is an IPv4 address, it's
440 * 32 bits. If there's an elision it can be 16 to 128 bits, depending
441 * on the number of other components.
443 * Modeled after google-url's CheckIPv6ComponentsSize function
445 static void compute_ipv6_comps_size(ipv6_address *address) {
446 address->components_size = address->h16_count * 2;
448 if(address->ipv4)
449 /* IPv4 address is 4 bytes. */
450 address->components_size += 4;
452 if(address->elision) {
453 /* An elision can be anywhere from 2 bytes up to 16 bytes.
454 * Its size depends on the size of the h16 and IPv4 components.
456 address->elision_size = 16 - address->components_size;
457 if(address->elision_size < 2)
458 address->elision_size = 2;
459 } else
460 address->elision_size = 0;
463 /* Taken from dlls/jscript/lex.c */
464 static int hex_to_int(WCHAR val) {
465 if(val >= '0' && val <= '9')
466 return val - '0';
467 else if(val >= 'a' && val <= 'f')
468 return val - 'a' + 10;
469 else if(val >= 'A' && val <= 'F')
470 return val - 'A' + 10;
472 return -1;
475 /* Helper function for converting a percent encoded string
476 * representation of a WCHAR value into its actual WCHAR value. If
477 * the two characters following the '%' aren't valid hex values then
478 * this function returns the NULL character.
480 * E.g.
481 * "%2E" will result in '.' being returned by this function.
483 static WCHAR decode_pct_val(const WCHAR *ptr) {
484 WCHAR ret = '\0';
486 if(*ptr == '%' && is_hexdigit(*(ptr + 1)) && is_hexdigit(*(ptr + 2))) {
487 INT a = hex_to_int(*(ptr + 1));
488 INT b = hex_to_int(*(ptr + 2));
490 ret = a << 4;
491 ret += b;
494 return ret;
497 /* Helper function for percent encoding a given character
498 * and storing the encoded value into a given buffer (dest).
500 * It's up to the calling function to ensure that there is
501 * at least enough space in 'dest' for the percent encoded
502 * value to be stored (so dest + 3 spaces available).
504 static inline void pct_encode_val(WCHAR val, WCHAR *dest) {
505 dest[0] = '%';
506 dest[1] = hexDigits[(val >> 4) & 0xf];
507 dest[2] = hexDigits[val & 0xf];
510 /* Attempts to parse the domain name from the host.
512 * This function also includes the Top-level Domain (TLD) name
513 * of the host when it tries to find the domain name. If it finds
514 * a valid domain name it will assign 'domain_start' the offset
515 * into 'host' where the domain name starts.
517 * It's implied that if there is a domain name its range is:
518 * [host+domain_start, host+host_len).
520 void find_domain_name(const WCHAR *host, DWORD host_len,
521 INT *domain_start) {
522 const WCHAR *last_tld, *sec_last_tld, *end, *p;
524 end = host+host_len-1;
526 *domain_start = -1;
528 /* There has to be at least enough room for a '.' followed by a
529 * 3-character TLD for a domain to even exist in the host name.
531 if(host_len < 4)
532 return;
534 for (last_tld = sec_last_tld = NULL, p = host; p <= end; p++)
536 if (*p == '.')
538 sec_last_tld = last_tld;
539 last_tld = p;
542 if(!last_tld)
543 /* http://hostname -> has no domain name. */
544 return;
546 if(!sec_last_tld) {
547 /* If the '.' is at the beginning of the host there
548 * has to be at least 3 characters in the TLD for it
549 * to be valid.
550 * Ex: .com -> .com as the domain name.
551 * .co -> has no domain name.
553 if(last_tld-host == 0) {
554 if(end-(last_tld-1) < 3)
555 return;
556 } else if(last_tld-host == 3) {
557 DWORD i;
559 /* If there are three characters in front of last_tld and
560 * they are on the list of recognized TLDs, then this
561 * host doesn't have a domain (since the host only contains
562 * a TLD name.
563 * Ex: edu.uk -> has no domain name.
564 * foo.uk -> foo.uk as the domain name.
566 for(i = 0; i < ARRAY_SIZE(recognized_tlds); ++i) {
567 if(!StrCmpNIW(host, recognized_tlds[i].tld_name, 3))
568 return;
570 } else if(last_tld-host < 3)
571 /* Anything less than 3 characters is considered part
572 * of the TLD name.
573 * Ex: ak.uk -> Has no domain name.
575 return;
577 /* Otherwise the domain name is the whole host name. */
578 *domain_start = 0;
579 } else if(end+1-last_tld > 3) {
580 /* If the last_tld has more than 3 characters, then it's automatically
581 * considered the TLD of the domain name.
582 * Ex: www.winehq.org.uk.test -> uk.test as the domain name.
584 *domain_start = (sec_last_tld+1)-host;
585 } else if(last_tld - (sec_last_tld+1) < 4) {
586 DWORD i;
587 /* If the sec_last_tld is 3 characters long it HAS to be on the list of
588 * recognized to still be considered part of the TLD name, otherwise
589 * it's considered the domain name.
590 * Ex: www.google.com.uk -> google.com.uk as the domain name.
591 * www.google.foo.uk -> foo.uk as the domain name.
593 if(last_tld - (sec_last_tld+1) == 3) {
594 for(i = 0; i < ARRAY_SIZE(recognized_tlds); ++i) {
595 if(!StrCmpNIW(sec_last_tld+1, recognized_tlds[i].tld_name, 3)) {
596 for (p = sec_last_tld; p > host; p--) if (p[-1] == '.') break;
597 *domain_start = p - host;
598 TRACE("Found domain name %s\n", debugstr_wn(host+*domain_start,
599 (host+host_len)-(host+*domain_start)));
600 return;
604 *domain_start = (sec_last_tld+1)-host;
605 } else {
606 /* Since the sec_last_tld is less than 3 characters it's considered
607 * part of the TLD.
608 * Ex: www.google.fo.uk -> google.fo.uk as the domain name.
610 for (p = sec_last_tld; p > host; p--) if (p[-1] == '.') break;
611 *domain_start = p - host;
613 } else {
614 /* The second to last TLD has more than 3 characters making it
615 * the domain name.
616 * Ex: www.google.test.us -> test.us as the domain name.
618 *domain_start = (sec_last_tld+1)-host;
621 TRACE("Found domain name %s\n", debugstr_wn(host+*domain_start,
622 (host+host_len)-(host+*domain_start)));
625 /* Removes the dot segments from a hierarchical URIs path component. This
626 * function performs the removal in place.
628 * This function returns the new length of the path string.
630 static DWORD remove_dot_segments(WCHAR *path, DWORD path_len) {
631 WCHAR *out = path;
632 const WCHAR *in = out;
633 const WCHAR *end = out + path_len;
634 DWORD len;
636 while(in < end) {
637 /* Move the first path segment in the input buffer to the end of
638 * the output buffer, and any subsequent characters up to, including
639 * the next "/" character (if any) or the end of the input buffer.
641 while(in < end && !is_slash(*in))
642 *out++ = *in++;
643 if(in == end)
644 break;
645 *out++ = *in++;
647 while(in < end) {
648 if(*in != '.')
649 break;
651 /* Handle ending "/." */
652 if(in + 1 == end) {
653 ++in;
654 break;
657 /* Handle "/./" */
658 if(is_slash(in[1])) {
659 in += 2;
660 continue;
663 /* If we don't have "/../" or ending "/.." */
664 if(in[1] != '.' || (in + 2 != end && !is_slash(in[2])))
665 break;
667 /* Find the slash preceding out pointer and move out pointer to it */
668 if(out > path+1 && is_slash(*--out))
669 --out;
670 while(out > path && !is_slash(*(--out)));
671 if(is_slash(*out))
672 ++out;
673 in += 2;
674 if(in != end)
675 ++in;
679 len = out - path;
680 TRACE("(%p %d): Path after dot segments removed %s len=%d\n", path, path_len,
681 debugstr_wn(path, len), len);
682 return len;
685 /* Attempts to find the file extension in a given path. */
686 static INT find_file_extension(const WCHAR *path, DWORD path_len) {
687 const WCHAR *end;
689 for(end = path+path_len-1; end >= path && *end != '/' && *end != '\\'; --end) {
690 if(*end == '.')
691 return end-path;
694 return -1;
697 /* Computes the location where the elision should occur in the IPv6
698 * address using the numerical values of each component stored in
699 * 'values'. If the address shouldn't contain an elision then 'index'
700 * is assigned -1 as its value. Otherwise 'index' will contain the
701 * starting index (into values) where the elision should be, and 'count'
702 * will contain the number of cells the elision covers.
704 * NOTES:
705 * Windows will expand an elision if the elision only represents one h16
706 * component of the address.
708 * Ex: [1::2:3:4:5:6:7] -> [1:0:2:3:4:5:6:7]
710 * If the IPv6 address contains an IPv4 address, the IPv4 address is also
711 * considered for being included as part of an elision if all its components
712 * are zeros.
714 * Ex: [1:2:3:4:5:6:0.0.0.0] -> [1:2:3:4:5:6::]
716 static void compute_elision_location(const ipv6_address *address, const USHORT values[8],
717 INT *index, DWORD *count) {
718 DWORD i, max_len, cur_len;
719 INT max_index, cur_index;
721 max_len = cur_len = 0;
722 max_index = cur_index = -1;
723 for(i = 0; i < 8; ++i) {
724 BOOL check_ipv4 = (address->ipv4 && i == 6);
725 BOOL is_end = (check_ipv4 || i == 7);
727 if(check_ipv4) {
728 /* Check if the IPv4 address contains only zeros. */
729 if(values[i] == 0 && values[i+1] == 0) {
730 if(cur_index == -1)
731 cur_index = i;
733 cur_len += 2;
734 ++i;
736 } else if(values[i] == 0) {
737 if(cur_index == -1)
738 cur_index = i;
740 ++cur_len;
743 if(is_end || values[i] != 0) {
744 /* We only consider it for an elision if it's
745 * more than 1 component long.
747 if(cur_len > 1 && cur_len > max_len) {
748 /* Found the new elision location. */
749 max_len = cur_len;
750 max_index = cur_index;
753 /* Reset the current range for the next range of zeros. */
754 cur_index = -1;
755 cur_len = 0;
759 *index = max_index;
760 *count = max_len;
763 /* Removes all the leading and trailing white spaces or
764 * control characters from the URI and removes all control
765 * characters inside of the URI string.
767 static BSTR pre_process_uri(LPCWSTR uri) {
768 const WCHAR *start, *end, *ptr;
769 WCHAR *ptr2;
770 DWORD len;
771 BSTR ret;
773 start = uri;
774 /* Skip leading controls and whitespace. */
775 while(*start && (iswcntrl(*start) || iswspace(*start))) ++start;
777 /* URI consisted only of control/whitespace. */
778 if(!*start)
779 return SysAllocStringLen(NULL, 0);
781 end = start + lstrlenW(start);
782 while(--end > start && (iswcntrl(*end) || iswspace(*end)));
784 len = ++end - start;
785 for(ptr = start; ptr < end; ptr++) {
786 if(iswcntrl(*ptr))
787 len--;
790 ret = SysAllocStringLen(NULL, len);
791 if(!ret)
792 return NULL;
794 for(ptr = start, ptr2=ret; ptr < end; ptr++) {
795 if(!iswcntrl(*ptr))
796 *ptr2++ = *ptr;
799 return ret;
802 /* Converts the specified IPv4 address into an uint value.
804 * This function assumes that the IPv4 address has already been validated.
806 static UINT ipv4toui(const WCHAR *ip, DWORD len) {
807 UINT ret = 0;
808 DWORD comp_value = 0;
809 const WCHAR *ptr;
811 for(ptr = ip; ptr < ip+len; ++ptr) {
812 if(*ptr == '.') {
813 ret <<= 8;
814 ret += comp_value;
815 comp_value = 0;
816 } else
817 comp_value = comp_value*10 + (*ptr-'0');
820 ret <<= 8;
821 ret += comp_value;
823 return ret;
826 /* Converts an IPv4 address in numerical form into its fully qualified
827 * string form. This function returns the number of characters written
828 * to 'dest'. If 'dest' is NULL this function will return the number of
829 * characters that would have been written.
831 * It's up to the caller to ensure there's enough space in 'dest' for the
832 * address.
834 static DWORD ui2ipv4(WCHAR *dest, UINT address) {
835 static const WCHAR formatW[] =
836 {'%','u','.','%','u','.','%','u','.','%','u',0};
837 DWORD ret = 0;
838 UCHAR digits[4];
840 digits[0] = (address >> 24) & 0xff;
841 digits[1] = (address >> 16) & 0xff;
842 digits[2] = (address >> 8) & 0xff;
843 digits[3] = address & 0xff;
845 if(!dest) {
846 WCHAR tmp[16];
847 ret = swprintf(tmp, ARRAY_SIZE(tmp), formatW, digits[0], digits[1], digits[2], digits[3]);
848 } else
849 ret = swprintf(dest, 16, formatW, digits[0], digits[1], digits[2], digits[3]);
851 return ret;
854 static DWORD ui2str(WCHAR *dest, UINT value) {
855 static const WCHAR formatW[] = {'%','u',0};
856 DWORD ret = 0;
858 if(!dest) {
859 WCHAR tmp[11];
860 ret = swprintf(tmp, ARRAY_SIZE(tmp), formatW, value);
861 } else
862 ret = swprintf(dest, 11, formatW, value);
864 return ret;
867 /* Converts a h16 component (from an IPv6 address) into its
868 * numerical value.
870 * This function assumes that the h16 component has already been validated.
872 static USHORT h16tous(h16 component) {
873 DWORD i;
874 USHORT ret = 0;
876 for(i = 0; i < component.len; ++i) {
877 ret <<= 4;
878 ret += hex_to_int(component.str[i]);
881 return ret;
884 /* Converts an IPv6 address into its 128 bits (16 bytes) numerical value.
886 * This function assumes that the ipv6_address has already been validated.
888 static BOOL ipv6_to_number(const ipv6_address *address, USHORT number[8]) {
889 DWORD i, cur_component = 0;
890 BOOL already_passed_elision = FALSE;
892 for(i = 0; i < address->h16_count; ++i) {
893 if(address->elision) {
894 if(address->components[i].str > address->elision && !already_passed_elision) {
895 /* Means we just passed the elision and need to add its values to
896 * 'number' before we do anything else.
898 INT j;
899 for(j = 0; j < address->elision_size; j+=2)
900 number[cur_component++] = 0;
902 already_passed_elision = TRUE;
906 number[cur_component++] = h16tous(address->components[i]);
909 /* Case when the elision appears after the h16 components. */
910 if(!already_passed_elision && address->elision) {
911 INT j;
912 for(j = 0; j < address->elision_size; j+=2)
913 number[cur_component++] = 0;
916 if(address->ipv4) {
917 UINT value = ipv4toui(address->ipv4, address->ipv4_len);
919 if(cur_component != 6) {
920 ERR("(%p %p): Failed sanity check with %d\n", address, number, cur_component);
921 return FALSE;
924 number[cur_component++] = (value >> 16) & 0xffff;
925 number[cur_component] = value & 0xffff;
928 return TRUE;
931 /* Checks if the characters pointed to by 'ptr' are
932 * a percent encoded data octet.
934 * pct-encoded = "%" HEXDIG HEXDIG
936 static BOOL check_pct_encoded(const WCHAR **ptr) {
937 const WCHAR *start = *ptr;
939 if(**ptr != '%')
940 return FALSE;
942 ++(*ptr);
943 if(!is_hexdigit(**ptr)) {
944 *ptr = start;
945 return FALSE;
948 ++(*ptr);
949 if(!is_hexdigit(**ptr)) {
950 *ptr = start;
951 return FALSE;
954 ++(*ptr);
955 return TRUE;
958 /* dec-octet = DIGIT ; 0-9
959 * / %x31-39 DIGIT ; 10-99
960 * / "1" 2DIGIT ; 100-199
961 * / "2" %x30-34 DIGIT ; 200-249
962 * / "25" %x30-35 ; 250-255
964 static BOOL check_dec_octet(const WCHAR **ptr) {
965 const WCHAR *c1, *c2, *c3;
967 c1 = *ptr;
968 /* A dec-octet must be at least 1 digit long. */
969 if(*c1 < '0' || *c1 > '9')
970 return FALSE;
972 ++(*ptr);
974 c2 = *ptr;
975 /* Since the 1-digit requirement was met, it doesn't
976 * matter if this is a DIGIT value, it's considered a
977 * dec-octet.
979 if(*c2 < '0' || *c2 > '9')
980 return TRUE;
982 ++(*ptr);
984 c3 = *ptr;
985 /* Same explanation as above. */
986 if(*c3 < '0' || *c3 > '9')
987 return TRUE;
989 /* Anything > 255 isn't a valid IP dec-octet. */
990 if(*c1 >= '2' && *c2 >= '5' && *c3 >= '5') {
991 *ptr = c1;
992 return FALSE;
995 ++(*ptr);
996 return TRUE;
999 /* Checks if there is an implicit IPv4 address in the host component of the URI.
1000 * The max value of an implicit IPv4 address is UINT_MAX.
1002 * Ex:
1003 * "234567" would be considered an implicit IPv4 address.
1005 static BOOL check_implicit_ipv4(const WCHAR **ptr, UINT *val) {
1006 const WCHAR *start = *ptr;
1007 ULONGLONG ret = 0;
1008 *val = 0;
1010 while(is_num(**ptr)) {
1011 ret = ret*10 + (**ptr - '0');
1013 if(ret > UINT_MAX) {
1014 *ptr = start;
1015 return FALSE;
1017 ++(*ptr);
1020 if(*ptr == start)
1021 return FALSE;
1023 *val = ret;
1024 return TRUE;
1027 /* Checks if the string contains an IPv4 address.
1029 * This function has a strict mode or a non-strict mode of operation
1030 * When 'strict' is set to FALSE this function will return TRUE if
1031 * the string contains at least 'dec-octet "." dec-octet' since partial
1032 * IPv4 addresses will be normalized out into full IPv4 addresses. When
1033 * 'strict' is set this function expects there to be a full IPv4 address.
1035 * IPv4address = dec-octet "." dec-octet "." dec-octet "." dec-octet
1037 static BOOL check_ipv4address(const WCHAR **ptr, BOOL strict) {
1038 const WCHAR *start = *ptr;
1040 if(!check_dec_octet(ptr)) {
1041 *ptr = start;
1042 return FALSE;
1045 if(**ptr != '.') {
1046 *ptr = start;
1047 return FALSE;
1050 ++(*ptr);
1051 if(!check_dec_octet(ptr)) {
1052 *ptr = start;
1053 return FALSE;
1056 if(**ptr != '.') {
1057 if(strict) {
1058 *ptr = start;
1059 return FALSE;
1060 } else
1061 return TRUE;
1064 ++(*ptr);
1065 if(!check_dec_octet(ptr)) {
1066 *ptr = start;
1067 return FALSE;
1070 if(**ptr != '.') {
1071 if(strict) {
1072 *ptr = start;
1073 return FALSE;
1074 } else
1075 return TRUE;
1078 ++(*ptr);
1079 if(!check_dec_octet(ptr)) {
1080 *ptr = start;
1081 return FALSE;
1084 /* Found a four digit ip address. */
1085 return TRUE;
1087 /* Tries to parse the scheme name of the URI.
1089 * scheme = ALPHA *(ALPHA | NUM | '+' | '-' | '.') as defined by RFC 3896.
1090 * NOTE: Windows accepts a number as the first character of a scheme.
1092 static BOOL parse_scheme_name(const WCHAR **ptr, parse_data *data, DWORD extras) {
1093 const WCHAR *start = *ptr;
1095 data->scheme = NULL;
1096 data->scheme_len = 0;
1098 while(**ptr) {
1099 if(**ptr == '*' && *ptr == start) {
1100 /* Might have found a wildcard scheme. If it is the next
1101 * char has to be a ':' for it to be a valid URI
1103 ++(*ptr);
1104 break;
1105 } else if(!is_num(**ptr) && !is_alpha(**ptr) && **ptr != '+' &&
1106 **ptr != '-' && **ptr != '.')
1107 break;
1109 (*ptr)++;
1112 if(*ptr == start)
1113 return FALSE;
1115 /* Schemes must end with a ':' */
1116 if(**ptr != ':' && !((extras & ALLOW_NULL_TERM_SCHEME) && !**ptr)) {
1117 *ptr = start;
1118 return FALSE;
1121 data->scheme = start;
1122 data->scheme_len = *ptr - start;
1124 ++(*ptr);
1125 return TRUE;
1128 /* Tries to deduce the corresponding URL_SCHEME for the given URI. Stores
1129 * the deduced URL_SCHEME in data->scheme_type.
1131 static BOOL parse_scheme_type(parse_data *data) {
1132 /* If there's scheme data then see if it's a recognized scheme. */
1133 if(data->scheme && data->scheme_len) {
1134 DWORD i;
1136 for(i = 0; i < ARRAY_SIZE(recognized_schemes); ++i) {
1137 if(lstrlenW(recognized_schemes[i].scheme_name) == data->scheme_len) {
1138 /* Has to be a case insensitive compare. */
1139 if(!StrCmpNIW(recognized_schemes[i].scheme_name, data->scheme, data->scheme_len)) {
1140 data->scheme_type = recognized_schemes[i].scheme;
1141 return TRUE;
1146 /* If we get here it means it's not a recognized scheme. */
1147 data->scheme_type = URL_SCHEME_UNKNOWN;
1148 return TRUE;
1149 } else if(data->is_relative) {
1150 /* Relative URI's have no scheme. */
1151 data->scheme_type = URL_SCHEME_UNKNOWN;
1152 return TRUE;
1153 } else {
1154 /* Should never reach here! what happened... */
1155 FIXME("(%p): Unable to determine scheme type for URI %s\n", data, debugstr_w(data->uri));
1156 return FALSE;
1160 /* Tries to parse (or deduce) the scheme_name of a URI. If it can't
1161 * parse a scheme from the URI it will try to deduce the scheme_name and scheme_type
1162 * using the flags specified in 'flags' (if any). Flags that affect how this function
1163 * operates are the Uri_CREATE_ALLOW_* flags.
1165 * All parsed/deduced information will be stored in 'data' when the function returns.
1167 * Returns TRUE if it was able to successfully parse the information.
1169 static BOOL parse_scheme(const WCHAR **ptr, parse_data *data, DWORD flags, DWORD extras) {
1170 static const WCHAR fileW[] = {'f','i','l','e',0};
1171 static const WCHAR wildcardW[] = {'*',0};
1173 /* First check to see if the uri could implicitly be a file path. */
1174 if(is_implicit_file_path(*ptr)) {
1175 if(flags & Uri_CREATE_ALLOW_IMPLICIT_FILE_SCHEME) {
1176 data->scheme = fileW;
1177 data->scheme_len = lstrlenW(fileW);
1178 data->has_implicit_scheme = TRUE;
1180 TRACE("(%p %p %x): URI is an implicit file path.\n", ptr, data, flags);
1181 } else {
1182 /* Windows does not consider anything that can implicitly be a file
1183 * path to be a valid URI if the ALLOW_IMPLICIT_FILE_SCHEME flag is not set...
1185 TRACE("(%p %p %x): URI is implicitly a file path, but, the ALLOW_IMPLICIT_FILE_SCHEME flag wasn't set.\n",
1186 ptr, data, flags);
1187 return FALSE;
1189 } else if(!parse_scheme_name(ptr, data, extras)) {
1190 /* No scheme was found, this means it could be:
1191 * a) an implicit Wildcard scheme
1192 * b) a relative URI
1193 * c) an invalid URI.
1195 if(flags & Uri_CREATE_ALLOW_IMPLICIT_WILDCARD_SCHEME) {
1196 data->scheme = wildcardW;
1197 data->scheme_len = lstrlenW(wildcardW);
1198 data->has_implicit_scheme = TRUE;
1200 TRACE("(%p %p %x): URI is an implicit wildcard scheme.\n", ptr, data, flags);
1201 } else if (flags & Uri_CREATE_ALLOW_RELATIVE) {
1202 data->is_relative = TRUE;
1203 TRACE("(%p %p %x): URI is relative.\n", ptr, data, flags);
1204 } else {
1205 TRACE("(%p %p %x): Malformed URI found. Unable to deduce scheme name.\n", ptr, data, flags);
1206 return FALSE;
1210 if(!data->is_relative)
1211 TRACE("(%p %p %x): Found scheme=%s scheme_len=%d\n", ptr, data, flags,
1212 debugstr_wn(data->scheme, data->scheme_len), data->scheme_len);
1214 if(!parse_scheme_type(data))
1215 return FALSE;
1217 TRACE("(%p %p %x): Assigned %d as the URL_SCHEME.\n", ptr, data, flags, data->scheme_type);
1218 return TRUE;
1221 static BOOL parse_username(const WCHAR **ptr, parse_data *data, DWORD flags, DWORD extras) {
1222 data->username = *ptr;
1224 while(**ptr != ':' && **ptr != '@') {
1225 if(**ptr == '%') {
1226 if(!check_pct_encoded(ptr)) {
1227 if(data->scheme_type != URL_SCHEME_UNKNOWN) {
1228 *ptr = data->username;
1229 data->username = NULL;
1230 return FALSE;
1232 } else
1233 continue;
1234 } else if(extras & ALLOW_NULL_TERM_USER_NAME && !**ptr)
1235 break;
1236 else if(is_auth_delim(**ptr, data->scheme_type != URL_SCHEME_UNKNOWN)) {
1237 *ptr = data->username;
1238 data->username = NULL;
1239 return FALSE;
1242 ++(*ptr);
1245 data->username_len = *ptr - data->username;
1246 return TRUE;
1249 static BOOL parse_password(const WCHAR **ptr, parse_data *data, DWORD flags, DWORD extras) {
1250 data->password = *ptr;
1252 while(**ptr != '@') {
1253 if(**ptr == '%') {
1254 if(!check_pct_encoded(ptr)) {
1255 if(data->scheme_type != URL_SCHEME_UNKNOWN) {
1256 *ptr = data->password;
1257 data->password = NULL;
1258 return FALSE;
1260 } else
1261 continue;
1262 } else if(extras & ALLOW_NULL_TERM_PASSWORD && !**ptr)
1263 break;
1264 else if(is_auth_delim(**ptr, data->scheme_type != URL_SCHEME_UNKNOWN)) {
1265 *ptr = data->password;
1266 data->password = NULL;
1267 return FALSE;
1270 ++(*ptr);
1273 data->password_len = *ptr - data->password;
1274 return TRUE;
1277 /* Parses the userinfo part of the URI (if it exists). The userinfo field of
1278 * a URI can consist of "username:password@", or just "username@".
1280 * RFC def:
1281 * userinfo = *( unreserved / pct-encoded / sub-delims / ":" )
1283 * NOTES:
1284 * 1) If there is more than one ':' in the userinfo part of the URI Windows
1285 * uses the first occurrence of ':' to delimit the username and password
1286 * components.
1288 * ex:
1289 * ftp://user:pass:word@winehq.org
1291 * would yield "user" as the username and "pass:word" as the password.
1293 * 2) Windows allows any character to appear in the "userinfo" part of
1294 * a URI, as long as it's not an authority delimiter character set.
1296 static void parse_userinfo(const WCHAR **ptr, parse_data *data, DWORD flags) {
1297 const WCHAR *start = *ptr;
1299 if(!parse_username(ptr, data, flags, 0)) {
1300 TRACE("(%p %p %x): URI contained no userinfo.\n", ptr, data, flags);
1301 return;
1304 if(**ptr == ':') {
1305 ++(*ptr);
1306 if(!parse_password(ptr, data, flags, 0)) {
1307 *ptr = start;
1308 data->username = NULL;
1309 data->username_len = 0;
1310 TRACE("(%p %p %x): URI contained no userinfo.\n", ptr, data, flags);
1311 return;
1315 if(**ptr != '@') {
1316 *ptr = start;
1317 data->username = NULL;
1318 data->username_len = 0;
1319 data->password = NULL;
1320 data->password_len = 0;
1322 TRACE("(%p %p %x): URI contained no userinfo.\n", ptr, data, flags);
1323 return;
1326 if(data->username)
1327 TRACE("(%p %p %x): Found username %s len=%d.\n", ptr, data, flags,
1328 debugstr_wn(data->username, data->username_len), data->username_len);
1330 if(data->password)
1331 TRACE("(%p %p %x): Found password %s len=%d.\n", ptr, data, flags,
1332 debugstr_wn(data->password, data->password_len), data->password_len);
1334 ++(*ptr);
1337 /* Attempts to parse a port from the URI.
1339 * NOTES:
1340 * Windows seems to have a cap on what the maximum value
1341 * for a port can be. The max value is USHORT_MAX.
1343 * port = *DIGIT
1345 static BOOL parse_port(const WCHAR **ptr, parse_data *data) {
1346 UINT port = 0;
1347 data->port = *ptr;
1349 while(!is_auth_delim(**ptr, data->scheme_type != URL_SCHEME_UNKNOWN)) {
1350 if(!is_num(**ptr)) {
1351 *ptr = data->port;
1352 data->port = NULL;
1353 return FALSE;
1356 port = port*10 + (**ptr-'0');
1358 if(port > USHRT_MAX) {
1359 *ptr = data->port;
1360 data->port = NULL;
1361 return FALSE;
1364 ++(*ptr);
1367 data->has_port = TRUE;
1368 data->port_value = port;
1369 data->port_len = *ptr - data->port;
1371 TRACE("(%p %p): Found port %s len=%d value=%u\n", ptr, data,
1372 debugstr_wn(data->port, data->port_len), data->port_len, data->port_value);
1373 return TRUE;
1376 /* Attempts to parse a IPv4 address from the URI.
1378 * NOTES:
1379 * Windows normalizes IPv4 addresses, This means there are three
1380 * possibilities for the URI to contain an IPv4 address.
1381 * 1) A well formed address (ex. 192.2.2.2).
1382 * 2) A partially formed address. For example "192.0" would
1383 * normalize to "192.0.0.0" during canonicalization.
1384 * 3) An implicit IPv4 address. For example "256" would
1385 * normalize to "0.0.1.0" during canonicalization. Also
1386 * note that the maximum value for an implicit IP address
1387 * is UINT_MAX, if the value in the URI exceeds this then
1388 * it is not considered an IPv4 address.
1390 static BOOL parse_ipv4address(const WCHAR **ptr, parse_data *data) {
1391 const BOOL is_unknown = data->scheme_type == URL_SCHEME_UNKNOWN;
1392 data->host = *ptr;
1394 if(!check_ipv4address(ptr, FALSE)) {
1395 if(!check_implicit_ipv4(ptr, &data->implicit_ipv4)) {
1396 TRACE("(%p %p): URI didn't contain anything looking like an IPv4 address.\n", ptr, data);
1397 *ptr = data->host;
1398 data->host = NULL;
1399 return FALSE;
1400 } else
1401 data->has_implicit_ip = TRUE;
1404 data->host_len = *ptr - data->host;
1405 data->host_type = Uri_HOST_IPV4;
1407 /* Check if what we found is the only part of the host name (if it isn't
1408 * we don't have an IPv4 address).
1410 if(**ptr == ':') {
1411 ++(*ptr);
1412 if(!parse_port(ptr, data)) {
1413 *ptr = data->host;
1414 data->host = NULL;
1415 return FALSE;
1417 } else if(!is_auth_delim(**ptr, !is_unknown)) {
1418 /* Found more data which belongs to the host, so this isn't an IPv4. */
1419 *ptr = data->host;
1420 data->host = NULL;
1421 data->has_implicit_ip = FALSE;
1422 return FALSE;
1425 TRACE("(%p %p): IPv4 address found. host=%s host_len=%d host_type=%d\n",
1426 ptr, data, debugstr_wn(data->host, data->host_len),
1427 data->host_len, data->host_type);
1428 return TRUE;
1431 /* Attempts to parse the reg-name from the URI.
1433 * Because of the way Windows handles ':' this function also
1434 * handles parsing the port.
1436 * reg-name = *( unreserved / pct-encoded / sub-delims )
1438 * NOTE:
1439 * Windows allows everything, but, the characters in "auth_delims" and ':'
1440 * to appear in a reg-name, unless it's an unknown scheme type then ':' is
1441 * allowed to appear (even if a valid port isn't after it).
1443 * Windows doesn't like host names which start with '[' and end with ']'
1444 * and don't contain a valid IP literal address in between them.
1446 * On Windows if a '[' is encountered in the host name the ':' no longer
1447 * counts as a delimiter until you reach the next ']' or an "authority delimiter".
1449 * A reg-name CAN be empty.
1451 static BOOL parse_reg_name(const WCHAR **ptr, parse_data *data, DWORD extras) {
1452 const BOOL has_start_bracket = **ptr == '[';
1453 const BOOL known_scheme = data->scheme_type != URL_SCHEME_UNKNOWN;
1454 const BOOL is_res = data->scheme_type == URL_SCHEME_RES;
1455 BOOL inside_brackets = has_start_bracket;
1457 /* res URIs don't have ports. */
1458 BOOL ignore_col = (extras & IGNORE_PORT_DELIMITER) || is_res;
1460 /* We have to be careful with file schemes. */
1461 if(data->scheme_type == URL_SCHEME_FILE) {
1462 /* This is because an implicit file scheme could be "C:\\test" and it
1463 * would trick this function into thinking the host is "C", when after
1464 * canonicalization the host would end up being an empty string. A drive
1465 * path can also have a '|' instead of a ':' after the drive letter.
1467 if(is_drive_path(*ptr)) {
1468 /* Regular old drive paths have no host type (or host name). */
1469 data->host_type = Uri_HOST_UNKNOWN;
1470 data->host = *ptr;
1471 data->host_len = 0;
1472 return TRUE;
1473 } else if(is_unc_path(*ptr))
1474 /* Skip past the "\\" of a UNC path. */
1475 *ptr += 2;
1478 data->host = *ptr;
1480 /* For res URIs, everything before the first '/' is
1481 * considered the host.
1483 while((!is_res && !is_auth_delim(**ptr, known_scheme)) ||
1484 (is_res && **ptr && **ptr != '/')) {
1485 if(**ptr == ':' && !ignore_col) {
1486 /* We can ignore ':' if we are inside brackets.*/
1487 if(!inside_brackets) {
1488 const WCHAR *tmp = (*ptr)++;
1490 /* Attempt to parse the port. */
1491 if(!parse_port(ptr, data)) {
1492 /* Windows expects there to be a valid port for known scheme types. */
1493 if(data->scheme_type != URL_SCHEME_UNKNOWN) {
1494 *ptr = data->host;
1495 data->host = NULL;
1496 TRACE("(%p %p %x): Expected valid port\n", ptr, data, extras);
1497 return FALSE;
1498 } else
1499 /* Windows gives up on trying to parse a port when it
1500 * encounters an invalid port.
1502 ignore_col = TRUE;
1503 } else {
1504 data->host_len = tmp - data->host;
1505 break;
1508 } else if(**ptr == '%' && (known_scheme && !is_res)) {
1509 /* Has to be a legit % encoded value. */
1510 if(!check_pct_encoded(ptr)) {
1511 *ptr = data->host;
1512 data->host = NULL;
1513 return FALSE;
1514 } else
1515 continue;
1516 } else if(is_res && is_forbidden_dos_path_char(**ptr)) {
1517 *ptr = data->host;
1518 data->host = NULL;
1519 return FALSE;
1520 } else if(**ptr == ']')
1521 inside_brackets = FALSE;
1522 else if(**ptr == '[')
1523 inside_brackets = TRUE;
1525 ++(*ptr);
1528 if(has_start_bracket) {
1529 /* Make sure the last character of the host wasn't a ']'. */
1530 if(*(*ptr-1) == ']') {
1531 TRACE("(%p %p %x): Expected an IP literal inside of the host\n", ptr, data, extras);
1532 *ptr = data->host;
1533 data->host = NULL;
1534 return FALSE;
1538 /* Don't overwrite our length if we found a port earlier. */
1539 if(!data->port)
1540 data->host_len = *ptr - data->host;
1542 /* If the host is empty, then it's an unknown host type. */
1543 if(data->host_len == 0 || is_res)
1544 data->host_type = Uri_HOST_UNKNOWN;
1545 else
1546 data->host_type = Uri_HOST_DNS;
1548 TRACE("(%p %p %x): Parsed reg-name. host=%s len=%d\n", ptr, data, extras,
1549 debugstr_wn(data->host, data->host_len), data->host_len);
1550 return TRUE;
1553 /* Attempts to parse an IPv6 address out of the URI.
1555 * IPv6address = 6( h16 ":" ) ls32
1556 * / "::" 5( h16 ":" ) ls32
1557 * / [ h16 ] "::" 4( h16 ":" ) ls32
1558 * / [ *1( h16 ":" ) h16 ] "::" 3( h16 ":" ) ls32
1559 * / [ *2( h16 ":" ) h16 ] "::" 2( h16 ":" ) ls32
1560 * / [ *3( h16 ":" ) h16 ] "::" h16 ":" ls32
1561 * / [ *4( h16 ":" ) h16 ] "::" ls32
1562 * / [ *5( h16 ":" ) h16 ] "::" h16
1563 * / [ *6( h16 ":" ) h16 ] "::"
1565 * ls32 = ( h16 ":" h16 ) / IPv4address
1566 * ; least-significant 32 bits of address.
1568 * h16 = 1*4HEXDIG
1569 * ; 16 bits of address represented in hexadecimal.
1571 * Modeled after google-url's 'DoParseIPv6' function.
1573 static BOOL parse_ipv6address(const WCHAR **ptr, parse_data *data) {
1574 const WCHAR *start, *cur_start;
1575 ipv6_address ip;
1577 start = cur_start = *ptr;
1578 memset(&ip, 0, sizeof(ipv6_address));
1580 for(;; ++(*ptr)) {
1581 /* Check if we're on the last character of the host. */
1582 BOOL is_end = (is_auth_delim(**ptr, data->scheme_type != URL_SCHEME_UNKNOWN)
1583 || **ptr == ']');
1585 BOOL is_split = (**ptr == ':');
1586 BOOL is_elision = (is_split && !is_end && *(*ptr+1) == ':');
1588 /* Check if we're at the end of a component, or
1589 * if we're at the end of the IPv6 address.
1591 if(is_split || is_end) {
1592 DWORD cur_len = 0;
1594 cur_len = *ptr - cur_start;
1596 /* h16 can't have a length > 4. */
1597 if(cur_len > 4) {
1598 *ptr = start;
1600 TRACE("(%p %p): h16 component to long.\n", ptr, data);
1601 return FALSE;
1604 if(cur_len == 0) {
1605 /* An h16 component can't have the length of 0 unless
1606 * the elision is at the beginning of the address, or
1607 * at the end of the address.
1609 if(!((*ptr == start && is_elision) ||
1610 (is_end && (*ptr-2) == ip.elision))) {
1611 *ptr = start;
1612 TRACE("(%p %p): IPv6 component cannot have a length of 0.\n", ptr, data);
1613 return FALSE;
1617 if(cur_len > 0) {
1618 /* An IPv6 address can have no more than 8 h16 components. */
1619 if(ip.h16_count >= 8) {
1620 *ptr = start;
1621 TRACE("(%p %p): Not a IPv6 address, too many h16 components.\n", ptr, data);
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): Found h16 component %s, len=%d, h16_count=%d\n",
1629 ptr, data, 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): IPv6 address cannot have 2 elisions.\n", ptr, data);
1644 return FALSE;
1647 ip.elision = *ptr;
1648 ++(*ptr);
1651 if(is_split)
1652 cur_start = *ptr+1;
1653 else {
1654 if(!check_ipv4address(ptr, TRUE)) {
1655 if(!is_hexdigit(**ptr)) {
1656 /* Not a valid character for an IPv6 address. */
1657 *ptr = start;
1658 return FALSE;
1660 } else {
1661 /* Found an IPv4 address. */
1662 ip.ipv4 = cur_start;
1663 ip.ipv4_len = *ptr - cur_start;
1665 TRACE("(%p %p): Found an attached IPv4 address %s len=%d.\n",
1666 ptr, data, debugstr_wn(ip.ipv4, ip.ipv4_len), ip.ipv4_len);
1668 /* IPv4 addresses can only appear at the end of a IPv6. */
1669 break;
1674 compute_ipv6_comps_size(&ip);
1676 /* Make sure the IPv6 address adds up to 16 bytes. */
1677 if(ip.components_size + ip.elision_size != 16) {
1678 *ptr = start;
1679 TRACE("(%p %p): Invalid IPv6 address, did not add up to 16 bytes.\n", ptr, data);
1680 return FALSE;
1683 if(ip.elision_size == 2) {
1684 /* For some reason on Windows if an elision that represents
1685 * only one h16 component is encountered at the very begin or
1686 * end of an IPv6 address, Windows does not consider it a
1687 * valid IPv6 address.
1689 * Ex: [::2:3:4:5:6:7] is not valid, even though the sum
1690 * of all the components == 128bits.
1692 if(ip.elision < ip.components[0].str ||
1693 ip.elision > ip.components[ip.h16_count-1].str) {
1694 *ptr = start;
1695 TRACE("(%p %p): Invalid IPv6 address. Detected elision of 2 bytes at the beginning or end of the address.\n",
1696 ptr, data);
1697 return FALSE;
1701 data->host_type = Uri_HOST_IPV6;
1702 data->has_ipv6 = TRUE;
1703 data->ipv6_address = ip;
1705 TRACE("(%p %p): Found valid IPv6 literal %s len=%d\n", ptr, data, debugstr_wn(start, *ptr-start), (int)(*ptr-start));
1706 return TRUE;
1709 /* IPvFuture = "v" 1*HEXDIG "." 1*( unreserved / sub-delims / ":" ) */
1710 static BOOL parse_ipvfuture(const WCHAR **ptr, parse_data *data) {
1711 const WCHAR *start = *ptr;
1713 /* IPvFuture has to start with a 'v' or 'V'. */
1714 if(**ptr != 'v' && **ptr != 'V')
1715 return FALSE;
1717 /* Following the v there must be at least 1 hex digit. */
1718 ++(*ptr);
1719 if(!is_hexdigit(**ptr)) {
1720 *ptr = start;
1721 return FALSE;
1724 ++(*ptr);
1725 while(is_hexdigit(**ptr))
1726 ++(*ptr);
1728 /* End of the hexdigit sequence must be a '.' */
1729 if(**ptr != '.') {
1730 *ptr = start;
1731 return FALSE;
1734 ++(*ptr);
1735 if(!is_unreserved(**ptr) && !is_subdelim(**ptr) && **ptr != ':') {
1736 *ptr = start;
1737 return FALSE;
1740 ++(*ptr);
1741 while(is_unreserved(**ptr) || is_subdelim(**ptr) || **ptr == ':')
1742 ++(*ptr);
1744 data->host_type = Uri_HOST_UNKNOWN;
1746 TRACE("(%p %p): Parsed IPvFuture address %s len=%d\n", ptr, data,
1747 debugstr_wn(start, *ptr-start), (int)(*ptr-start));
1749 return TRUE;
1752 /* IP-literal = "[" ( IPv6address / IPvFuture ) "]" */
1753 static BOOL parse_ip_literal(const WCHAR **ptr, parse_data *data, DWORD extras) {
1754 data->host = *ptr;
1756 if(**ptr != '[' && !(extras & ALLOW_BRACKETLESS_IP_LITERAL)) {
1757 data->host = NULL;
1758 return FALSE;
1759 } else if(**ptr == '[')
1760 ++(*ptr);
1762 if(!parse_ipv6address(ptr, data)) {
1763 if(extras & SKIP_IP_FUTURE_CHECK || !parse_ipvfuture(ptr, data)) {
1764 *ptr = data->host;
1765 data->host = NULL;
1766 return FALSE;
1770 if(**ptr != ']' && !(extras & ALLOW_BRACKETLESS_IP_LITERAL)) {
1771 *ptr = data->host;
1772 data->host = NULL;
1773 return FALSE;
1774 } else if(!**ptr && extras & ALLOW_BRACKETLESS_IP_LITERAL) {
1775 /* The IP literal didn't contain brackets and was followed by
1776 * a NULL terminator, so no reason to even check the port.
1778 data->host_len = *ptr - data->host;
1779 return TRUE;
1782 ++(*ptr);
1783 if(**ptr == ':') {
1784 ++(*ptr);
1785 /* If a valid port is not found, then let it trickle down to
1786 * parse_reg_name.
1788 if(!parse_port(ptr, data)) {
1789 *ptr = data->host;
1790 data->host = NULL;
1791 return FALSE;
1793 } else
1794 data->host_len = *ptr - data->host;
1796 return TRUE;
1799 /* Parses the host information from the URI.
1801 * host = IP-literal / IPv4address / reg-name
1803 static BOOL parse_host(const WCHAR **ptr, parse_data *data, DWORD extras) {
1804 if(!parse_ip_literal(ptr, data, extras)) {
1805 if(!parse_ipv4address(ptr, data)) {
1806 if(!parse_reg_name(ptr, data, extras)) {
1807 TRACE("(%p %p %x): Malformed URI, Unknown host type.\n", ptr, data, extras);
1808 return FALSE;
1813 return TRUE;
1816 /* Parses the authority information from the URI.
1818 * authority = [ userinfo "@" ] host [ ":" port ]
1820 static BOOL parse_authority(const WCHAR **ptr, parse_data *data, DWORD flags) {
1821 parse_userinfo(ptr, data, flags);
1823 /* Parsing the port will happen during one of the host parsing
1824 * routines (if the URI has a port).
1826 if(!parse_host(ptr, data, 0))
1827 return FALSE;
1829 return TRUE;
1832 /* Attempts to parse the path information of a hierarchical URI. */
1833 static BOOL parse_path_hierarchical(const WCHAR **ptr, parse_data *data, DWORD flags) {
1834 const WCHAR *start = *ptr;
1835 static const WCHAR slash[] = {'/',0};
1836 const BOOL is_file = data->scheme_type == URL_SCHEME_FILE;
1838 if(is_path_delim(data->scheme_type, **ptr)) {
1839 if(data->scheme_type == URL_SCHEME_WILDCARD && !data->must_have_path) {
1840 data->path = NULL;
1841 data->path_len = 0;
1842 } else if(!(flags & Uri_CREATE_NO_CANONICALIZE)) {
1843 /* If the path component is empty, then a '/' is added. */
1844 data->path = slash;
1845 data->path_len = 1;
1847 } else {
1848 while(!is_path_delim(data->scheme_type, **ptr)) {
1849 if(**ptr == '%' && data->scheme_type != URL_SCHEME_UNKNOWN && !is_file) {
1850 if(!check_pct_encoded(ptr)) {
1851 *ptr = start;
1852 return FALSE;
1853 } else
1854 continue;
1855 } else if(is_forbidden_dos_path_char(**ptr) && is_file &&
1856 (flags & Uri_CREATE_FILE_USE_DOS_PATH)) {
1857 /* File schemes with USE_DOS_PATH set aren't allowed to have
1858 * a '<' or '>' or '\"' appear in them.
1860 *ptr = start;
1861 return FALSE;
1862 } else if(**ptr == '\\') {
1863 /* Not allowed to have a backslash if NO_CANONICALIZE is set
1864 * and the scheme is known type (but not a file scheme).
1866 if(flags & Uri_CREATE_NO_CANONICALIZE) {
1867 if(data->scheme_type != URL_SCHEME_FILE &&
1868 data->scheme_type != URL_SCHEME_UNKNOWN) {
1869 *ptr = start;
1870 return FALSE;
1875 ++(*ptr);
1878 /* The only time a URI doesn't have a path is when
1879 * the NO_CANONICALIZE flag is set and the raw URI
1880 * didn't contain one.
1882 if(*ptr == start) {
1883 data->path = NULL;
1884 data->path_len = 0;
1885 } else {
1886 data->path = start;
1887 data->path_len = *ptr - start;
1891 if(data->path)
1892 TRACE("(%p %p %x): Parsed path %s len=%d\n", ptr, data, flags,
1893 debugstr_wn(data->path, data->path_len), data->path_len);
1894 else
1895 TRACE("(%p %p %x): The URI contained no path\n", ptr, data, flags);
1897 return TRUE;
1900 /* Parses the path of an opaque URI (much less strict than the parser
1901 * for a hierarchical URI).
1903 * NOTE:
1904 * Windows allows invalid % encoded data to appear in opaque URI paths
1905 * for unknown scheme types.
1907 * File schemes with USE_DOS_PATH set aren't allowed to have '<', '>', or '\"'
1908 * appear in them.
1910 static BOOL parse_path_opaque(const WCHAR **ptr, parse_data *data, DWORD flags) {
1911 const BOOL known_scheme = data->scheme_type != URL_SCHEME_UNKNOWN;
1912 const BOOL is_file = data->scheme_type == URL_SCHEME_FILE;
1913 const BOOL is_mailto = data->scheme_type == URL_SCHEME_MAILTO;
1915 if (is_mailto && (*ptr)[0] == '/' && (*ptr)[1] == '/')
1917 if ((*ptr)[2]) data->path = *ptr + 2;
1918 else data->path = NULL;
1920 else
1921 data->path = *ptr;
1923 while(!is_path_delim(data->scheme_type, **ptr)) {
1924 if(**ptr == '%' && known_scheme) {
1925 if(!check_pct_encoded(ptr)) {
1926 *ptr = data->path;
1927 data->path = NULL;
1928 return FALSE;
1929 } else
1930 continue;
1931 } else if(is_forbidden_dos_path_char(**ptr) && is_file &&
1932 (flags & Uri_CREATE_FILE_USE_DOS_PATH)) {
1933 *ptr = data->path;
1934 data->path = NULL;
1935 return FALSE;
1938 ++(*ptr);
1941 if (data->path) data->path_len = *ptr - data->path;
1942 TRACE("(%p %p %x): Parsed opaque URI path %s len=%d\n", ptr, data, flags,
1943 debugstr_wn(data->path, data->path_len), data->path_len);
1944 return TRUE;
1947 /* Determines how the URI should be parsed after the scheme information.
1949 * If the scheme is followed by "//", then it is treated as a hierarchical URI
1950 * which then the authority and path information will be parsed out. Otherwise, the
1951 * URI will be treated as an opaque URI which the authority information is not parsed
1952 * out.
1954 * RFC 3896 definition of hier-part:
1956 * hier-part = "//" authority path-abempty
1957 * / path-absolute
1958 * / path-rootless
1959 * / path-empty
1961 * MSDN opaque URI definition:
1962 * scheme ":" path [ "#" fragment ]
1964 * NOTES:
1965 * If the URI is of an unknown scheme type and has a "//" following the scheme then it
1966 * is treated as a hierarchical URI, but, if the CREATE_NO_CRACK_UNKNOWN_SCHEMES flag is
1967 * set then it is considered an opaque URI regardless of what follows the scheme information
1968 * (per MSDN documentation).
1970 static BOOL parse_hierpart(const WCHAR **ptr, parse_data *data, DWORD flags) {
1971 const WCHAR *start = *ptr;
1973 data->must_have_path = FALSE;
1975 /* For javascript: URIs, simply set everything as a path */
1976 if(data->scheme_type == URL_SCHEME_JAVASCRIPT) {
1977 data->path = *ptr;
1978 data->path_len = lstrlenW(*ptr);
1979 data->is_opaque = TRUE;
1980 *ptr += data->path_len;
1981 return TRUE;
1984 /* Checks if the authority information needs to be parsed. */
1985 if(is_hierarchical_uri(ptr, data)) {
1986 /* Only treat it as a hierarchical URI if the scheme_type is known or
1987 * the Uri_CREATE_NO_CRACK_UNKNOWN_SCHEMES flag is not set.
1989 if(data->scheme_type != URL_SCHEME_UNKNOWN ||
1990 !(flags & Uri_CREATE_NO_CRACK_UNKNOWN_SCHEMES)) {
1991 TRACE("(%p %p %x): Treating URI as an hierarchical URI.\n", ptr, data, flags);
1992 data->is_opaque = FALSE;
1994 if(data->scheme_type == URL_SCHEME_WILDCARD && !data->has_implicit_scheme) {
1995 if(**ptr == '/' && *(*ptr+1) == '/') {
1996 data->must_have_path = TRUE;
1997 *ptr += 2;
2001 /* TODO: Handle hierarchical URI's, parse authority then parse the path. */
2002 if(!parse_authority(ptr, data, flags))
2003 return FALSE;
2005 return parse_path_hierarchical(ptr, data, flags);
2006 } else
2007 /* Reset ptr to its starting position so opaque path parsing
2008 * begins at the correct location.
2010 *ptr = start;
2013 /* If it reaches here, then the URI will be treated as an opaque
2014 * URI.
2017 TRACE("(%p %p %x): Treating URI as an opaque URI.\n", ptr, data, flags);
2019 data->is_opaque = TRUE;
2020 if(!parse_path_opaque(ptr, data, flags))
2021 return FALSE;
2023 return TRUE;
2026 /* Attempts to parse the query string from the URI.
2028 * NOTES:
2029 * If NO_DECODE_EXTRA_INFO flag is set, then invalid percent encoded
2030 * data is allowed to appear in the query string. For unknown scheme types
2031 * invalid percent encoded data is allowed to appear regardless.
2033 static BOOL parse_query(const WCHAR **ptr, parse_data *data, DWORD flags) {
2034 const BOOL known_scheme = data->scheme_type != URL_SCHEME_UNKNOWN;
2036 if(**ptr != '?') {
2037 TRACE("(%p %p %x): URI didn't contain a query string.\n", ptr, data, flags);
2038 return TRUE;
2041 data->query = *ptr;
2043 ++(*ptr);
2044 while(**ptr && **ptr != '#') {
2045 if(**ptr == '%' && known_scheme &&
2046 !(flags & Uri_CREATE_NO_DECODE_EXTRA_INFO)) {
2047 if(!check_pct_encoded(ptr)) {
2048 *ptr = data->query;
2049 data->query = NULL;
2050 return FALSE;
2051 } else
2052 continue;
2055 ++(*ptr);
2058 data->query_len = *ptr - data->query;
2060 TRACE("(%p %p %x): Parsed query string %s len=%d\n", ptr, data, flags,
2061 debugstr_wn(data->query, data->query_len), data->query_len);
2062 return TRUE;
2065 /* Attempts to parse the fragment from the URI.
2067 * NOTES:
2068 * If NO_DECODE_EXTRA_INFO flag is set, then invalid percent encoded
2069 * data is allowed to appear in the query string. For unknown scheme types
2070 * invalid percent encoded data is allowed to appear regardless.
2072 static BOOL parse_fragment(const WCHAR **ptr, parse_data *data, DWORD flags) {
2073 const BOOL known_scheme = data->scheme_type != URL_SCHEME_UNKNOWN;
2075 if(**ptr != '#') {
2076 TRACE("(%p %p %x): URI didn't contain a fragment.\n", ptr, data, flags);
2077 return TRUE;
2080 data->fragment = *ptr;
2082 ++(*ptr);
2083 while(**ptr) {
2084 if(**ptr == '%' && known_scheme &&
2085 !(flags & Uri_CREATE_NO_DECODE_EXTRA_INFO)) {
2086 if(!check_pct_encoded(ptr)) {
2087 *ptr = data->fragment;
2088 data->fragment = NULL;
2089 return FALSE;
2090 } else
2091 continue;
2094 ++(*ptr);
2097 data->fragment_len = *ptr - data->fragment;
2099 TRACE("(%p %p %x): Parsed fragment %s len=%d\n", ptr, data, flags,
2100 debugstr_wn(data->fragment, data->fragment_len), data->fragment_len);
2101 return TRUE;
2104 /* Parses and validates the components of the specified by data->uri
2105 * and stores the information it parses into 'data'.
2107 * Returns TRUE if it successfully parsed the URI. False otherwise.
2109 static BOOL parse_uri(parse_data *data, DWORD flags) {
2110 const WCHAR *ptr;
2111 const WCHAR **pptr;
2113 ptr = data->uri;
2114 pptr = &ptr;
2116 TRACE("(%p %x): BEGINNING TO PARSE URI %s.\n", data, flags, debugstr_w(data->uri));
2118 if(!parse_scheme(pptr, data, flags, 0))
2119 return FALSE;
2121 if(!parse_hierpart(pptr, data, flags))
2122 return FALSE;
2124 if(!parse_query(pptr, data, flags))
2125 return FALSE;
2127 if(!parse_fragment(pptr, data, flags))
2128 return FALSE;
2130 TRACE("(%p %x): FINISHED PARSING URI.\n", data, flags);
2131 return TRUE;
2134 static BOOL canonicalize_username(const parse_data *data, Uri *uri, DWORD flags, BOOL computeOnly) {
2135 const WCHAR *ptr;
2137 if(!data->username) {
2138 uri->userinfo_start = -1;
2139 return TRUE;
2142 uri->userinfo_start = uri->canon_len;
2143 for(ptr = data->username; ptr < data->username+data->username_len; ++ptr) {
2144 if(*ptr == '%') {
2145 /* Only decode % encoded values for known scheme types. */
2146 if(data->scheme_type != URL_SCHEME_UNKNOWN) {
2147 /* See if the value really needs decoding. */
2148 WCHAR val = decode_pct_val(ptr);
2149 if(is_unreserved(val)) {
2150 if(!computeOnly)
2151 uri->canon_uri[uri->canon_len] = val;
2153 ++uri->canon_len;
2155 /* Move pass the hex characters. */
2156 ptr += 2;
2157 continue;
2160 } else if(is_ascii(*ptr) && !is_reserved(*ptr) && !is_unreserved(*ptr) && *ptr != '\\') {
2161 /* Only percent encode forbidden characters if the NO_ENCODE_FORBIDDEN_CHARACTERS flag
2162 * is NOT set.
2164 if(!(flags & Uri_CREATE_NO_ENCODE_FORBIDDEN_CHARACTERS)) {
2165 if(!computeOnly)
2166 pct_encode_val(*ptr, uri->canon_uri + uri->canon_len);
2168 uri->canon_len += 3;
2169 continue;
2173 if(!computeOnly)
2174 /* Nothing special, so just copy the character over. */
2175 uri->canon_uri[uri->canon_len] = *ptr;
2176 ++uri->canon_len;
2179 return TRUE;
2182 static BOOL canonicalize_password(const parse_data *data, Uri *uri, DWORD flags, BOOL computeOnly) {
2183 const WCHAR *ptr;
2185 if(!data->password) {
2186 uri->userinfo_split = -1;
2187 return TRUE;
2190 if(uri->userinfo_start == -1)
2191 /* Has a password, but, doesn't have a username. */
2192 uri->userinfo_start = uri->canon_len;
2194 uri->userinfo_split = uri->canon_len - uri->userinfo_start;
2196 /* Add the ':' to the userinfo component. */
2197 if(!computeOnly)
2198 uri->canon_uri[uri->canon_len] = ':';
2199 ++uri->canon_len;
2201 for(ptr = data->password; ptr < data->password+data->password_len; ++ptr) {
2202 if(*ptr == '%') {
2203 /* Only decode % encoded values for known scheme types. */
2204 if(data->scheme_type != URL_SCHEME_UNKNOWN) {
2205 /* See if the value really needs decoding. */
2206 WCHAR val = decode_pct_val(ptr);
2207 if(is_unreserved(val)) {
2208 if(!computeOnly)
2209 uri->canon_uri[uri->canon_len] = val;
2211 ++uri->canon_len;
2213 /* Move pass the hex characters. */
2214 ptr += 2;
2215 continue;
2218 } else if(is_ascii(*ptr) && !is_reserved(*ptr) && !is_unreserved(*ptr) && *ptr != '\\') {
2219 /* Only percent encode forbidden characters if the NO_ENCODE_FORBIDDEN_CHARACTERS flag
2220 * is NOT set.
2222 if(!(flags & Uri_CREATE_NO_ENCODE_FORBIDDEN_CHARACTERS)) {
2223 if(!computeOnly)
2224 pct_encode_val(*ptr, uri->canon_uri + uri->canon_len);
2226 uri->canon_len += 3;
2227 continue;
2231 if(!computeOnly)
2232 /* Nothing special, so just copy the character over. */
2233 uri->canon_uri[uri->canon_len] = *ptr;
2234 ++uri->canon_len;
2237 return TRUE;
2240 /* Canonicalizes the userinfo of the URI represented by the parse_data.
2242 * Canonicalization of the userinfo is a simple process. If there are any percent
2243 * encoded characters that fall in the "unreserved" character set, they are decoded
2244 * to their actual value. If a character is not in the "unreserved" or "reserved" sets
2245 * then it is percent encoded. Other than that the characters are copied over without
2246 * change.
2248 static BOOL canonicalize_userinfo(const parse_data *data, Uri *uri, DWORD flags, BOOL computeOnly) {
2249 uri->userinfo_start = uri->userinfo_split = -1;
2250 uri->userinfo_len = 0;
2252 if(!data->username && !data->password)
2253 /* URI doesn't have userinfo, so nothing to do here. */
2254 return TRUE;
2256 if(!canonicalize_username(data, uri, flags, computeOnly))
2257 return FALSE;
2259 if(!canonicalize_password(data, uri, flags, computeOnly))
2260 return FALSE;
2262 uri->userinfo_len = uri->canon_len - uri->userinfo_start;
2263 if(!computeOnly)
2264 TRACE("(%p %p %x %d): Canonicalized userinfo, userinfo_start=%d, userinfo=%s, userinfo_split=%d userinfo_len=%d.\n",
2265 data, uri, flags, computeOnly, uri->userinfo_start, debugstr_wn(uri->canon_uri + uri->userinfo_start, uri->userinfo_len),
2266 uri->userinfo_split, uri->userinfo_len);
2268 /* Now insert the '@' after the userinfo. */
2269 if(!computeOnly)
2270 uri->canon_uri[uri->canon_len] = '@';
2271 ++uri->canon_len;
2273 return TRUE;
2276 /* Attempts to canonicalize a reg_name.
2278 * Things that happen:
2279 * 1) If Uri_CREATE_NO_CANONICALIZE flag is not set, then the reg_name is
2280 * lower cased. Unless it's an unknown scheme type, which case it's
2281 * no lower cased regardless.
2283 * 2) Unreserved % encoded characters are decoded for known
2284 * scheme types.
2286 * 3) Forbidden characters are % encoded as long as
2287 * Uri_CREATE_NO_ENCODE_FORBIDDEN_CHARACTERS flag is not set and
2288 * it isn't an unknown scheme type.
2290 * 4) If it's a file scheme and the host is "localhost" it's removed.
2292 * 5) If it's a file scheme and Uri_CREATE_FILE_USE_DOS_PATH is set,
2293 * then the UNC path characters are added before the host name.
2295 static BOOL canonicalize_reg_name(const parse_data *data, Uri *uri,
2296 DWORD flags, BOOL computeOnly) {
2297 static const WCHAR localhostW[] =
2298 {'l','o','c','a','l','h','o','s','t',0};
2299 const WCHAR *ptr;
2300 const BOOL known_scheme = data->scheme_type != URL_SCHEME_UNKNOWN;
2302 if(data->scheme_type == URL_SCHEME_FILE &&
2303 data->host_len == lstrlenW(localhostW)) {
2304 if(!StrCmpNIW(data->host, localhostW, data->host_len)) {
2305 uri->host_start = -1;
2306 uri->host_len = 0;
2307 uri->host_type = Uri_HOST_UNKNOWN;
2308 return TRUE;
2312 if(data->scheme_type == URL_SCHEME_FILE && flags & Uri_CREATE_FILE_USE_DOS_PATH) {
2313 if(!computeOnly) {
2314 uri->canon_uri[uri->canon_len] = '\\';
2315 uri->canon_uri[uri->canon_len+1] = '\\';
2317 uri->canon_len += 2;
2318 uri->authority_start = uri->canon_len;
2321 uri->host_start = uri->canon_len;
2323 for(ptr = data->host; ptr < data->host+data->host_len; ++ptr) {
2324 if(*ptr == '%' && known_scheme) {
2325 WCHAR val = decode_pct_val(ptr);
2326 if(is_unreserved(val)) {
2327 /* If NO_CANONICALIZE is not set, then windows lower cases the
2328 * decoded value.
2330 if(!(flags & Uri_CREATE_NO_CANONICALIZE) && iswupper(val)) {
2331 if(!computeOnly)
2332 uri->canon_uri[uri->canon_len] = towlower(val);
2333 } else {
2334 if(!computeOnly)
2335 uri->canon_uri[uri->canon_len] = val;
2337 ++uri->canon_len;
2339 /* Skip past the % encoded character. */
2340 ptr += 2;
2341 continue;
2342 } else {
2343 /* Just copy the % over. */
2344 if(!computeOnly)
2345 uri->canon_uri[uri->canon_len] = *ptr;
2346 ++uri->canon_len;
2348 } else if(*ptr == '\\') {
2349 /* Only unknown scheme types could have made it here with a '\\' in the host name. */
2350 if(!computeOnly)
2351 uri->canon_uri[uri->canon_len] = *ptr;
2352 ++uri->canon_len;
2353 } else if(!(flags & Uri_CREATE_NO_ENCODE_FORBIDDEN_CHARACTERS) && is_ascii(*ptr) &&
2354 !is_unreserved(*ptr) && !is_reserved(*ptr) && known_scheme) {
2355 if(!computeOnly) {
2356 pct_encode_val(*ptr, uri->canon_uri+uri->canon_len);
2358 /* The percent encoded value gets lower cased also. */
2359 if(!(flags & Uri_CREATE_NO_CANONICALIZE)) {
2360 uri->canon_uri[uri->canon_len+1] = towlower(uri->canon_uri[uri->canon_len+1]);
2361 uri->canon_uri[uri->canon_len+2] = towlower(uri->canon_uri[uri->canon_len+2]);
2365 uri->canon_len += 3;
2366 } else {
2367 if(!computeOnly) {
2368 if(!(flags & Uri_CREATE_NO_CANONICALIZE) && known_scheme)
2369 uri->canon_uri[uri->canon_len] = towlower(*ptr);
2370 else
2371 uri->canon_uri[uri->canon_len] = *ptr;
2374 ++uri->canon_len;
2378 uri->host_len = uri->canon_len - uri->host_start;
2380 if(!computeOnly)
2381 TRACE("(%p %p %x %d): Canonicalize reg_name=%s len=%d\n", data, uri, flags,
2382 computeOnly, debugstr_wn(uri->canon_uri+uri->host_start, uri->host_len),
2383 uri->host_len);
2385 if(!computeOnly)
2386 find_domain_name(uri->canon_uri+uri->host_start, uri->host_len,
2387 &(uri->domain_offset));
2389 return TRUE;
2392 /* Attempts to canonicalize an implicit IPv4 address. */
2393 static BOOL canonicalize_implicit_ipv4address(const parse_data *data, Uri *uri, DWORD flags, BOOL computeOnly) {
2394 uri->host_start = uri->canon_len;
2396 TRACE("%u\n", data->implicit_ipv4);
2397 /* For unknown scheme types Windows doesn't convert
2398 * the value into an IP address, but it still considers
2399 * it an IPv4 address.
2401 if(data->scheme_type == URL_SCHEME_UNKNOWN) {
2402 if(!computeOnly)
2403 memcpy(uri->canon_uri+uri->canon_len, data->host, data->host_len*sizeof(WCHAR));
2404 uri->canon_len += data->host_len;
2405 } else {
2406 if(!computeOnly)
2407 uri->canon_len += ui2ipv4(uri->canon_uri+uri->canon_len, data->implicit_ipv4);
2408 else
2409 uri->canon_len += ui2ipv4(NULL, data->implicit_ipv4);
2412 uri->host_len = uri->canon_len - uri->host_start;
2413 uri->host_type = Uri_HOST_IPV4;
2415 if(!computeOnly)
2416 TRACE("%p %p %x %d): Canonicalized implicit IP address=%s len=%d\n",
2417 data, uri, flags, computeOnly,
2418 debugstr_wn(uri->canon_uri+uri->host_start, uri->host_len),
2419 uri->host_len);
2421 return TRUE;
2424 /* Attempts to canonicalize an IPv4 address.
2426 * If the parse_data represents a URI that has an implicit IPv4 address
2427 * (ex. http://256/, this function will convert 256 into 0.0.1.0). If
2428 * the implicit IP address exceeds the value of UINT_MAX (maximum value
2429 * for an IPv4 address) it's canonicalized as if it were a reg-name.
2431 * If the parse_data contains a partial or full IPv4 address it normalizes it.
2432 * A partial IPv4 address is something like "192.0" and would be normalized to
2433 * "192.0.0.0". With a full (or partial) IPv4 address like "192.002.01.003" would
2434 * be normalized to "192.2.1.3".
2436 * NOTES:
2437 * Windows ONLY normalizes IPv4 address for known scheme types (one that isn't
2438 * URL_SCHEME_UNKNOWN). For unknown scheme types, it simply copies the data from
2439 * the original URI into the canonicalized URI, but, it still recognizes URI's
2440 * host type as HOST_IPV4.
2442 static BOOL canonicalize_ipv4address(const parse_data *data, Uri *uri, DWORD flags, BOOL computeOnly) {
2443 if(data->has_implicit_ip)
2444 return canonicalize_implicit_ipv4address(data, uri, flags, computeOnly);
2445 else {
2446 uri->host_start = uri->canon_len;
2448 /* Windows only normalizes for known scheme types. */
2449 if(data->scheme_type != URL_SCHEME_UNKNOWN) {
2450 /* parse_data contains a partial or full IPv4 address, so normalize it. */
2451 DWORD i, octetDigitCount = 0, octetCount = 0;
2452 BOOL octetHasDigit = FALSE;
2454 for(i = 0; i < data->host_len; ++i) {
2455 if(data->host[i] == '0' && !octetHasDigit) {
2456 /* Can ignore leading zeros if:
2457 * 1) It isn't the last digit of the octet.
2458 * 2) i+1 != data->host_len
2459 * 3) i+1 != '.'
2461 if(octetDigitCount == 2 ||
2462 i+1 == data->host_len ||
2463 data->host[i+1] == '.') {
2464 if(!computeOnly)
2465 uri->canon_uri[uri->canon_len] = data->host[i];
2466 ++uri->canon_len;
2467 TRACE("Adding zero\n");
2469 } else if(data->host[i] == '.') {
2470 if(!computeOnly)
2471 uri->canon_uri[uri->canon_len] = data->host[i];
2472 ++uri->canon_len;
2474 octetDigitCount = 0;
2475 octetHasDigit = FALSE;
2476 ++octetCount;
2477 } else {
2478 if(!computeOnly)
2479 uri->canon_uri[uri->canon_len] = data->host[i];
2480 ++uri->canon_len;
2482 ++octetDigitCount;
2483 octetHasDigit = TRUE;
2487 /* Make sure the canonicalized IP address has 4 dec-octets.
2488 * If doesn't add "0" ones until there is 4;
2490 for( ; octetCount < 3; ++octetCount) {
2491 if(!computeOnly) {
2492 uri->canon_uri[uri->canon_len] = '.';
2493 uri->canon_uri[uri->canon_len+1] = '0';
2496 uri->canon_len += 2;
2498 } else {
2499 /* Windows doesn't normalize addresses in unknown schemes. */
2500 if(!computeOnly)
2501 memcpy(uri->canon_uri+uri->canon_len, data->host, data->host_len*sizeof(WCHAR));
2502 uri->canon_len += data->host_len;
2505 uri->host_len = uri->canon_len - uri->host_start;
2506 if(!computeOnly)
2507 TRACE("(%p %p %x %d): Canonicalized IPv4 address, ip=%s len=%d\n",
2508 data, uri, flags, computeOnly,
2509 debugstr_wn(uri->canon_uri+uri->host_start, uri->host_len),
2510 uri->host_len);
2513 return TRUE;
2516 /* Attempts to canonicalize the IPv6 address of the URI.
2518 * Multiple things happen during the canonicalization of an IPv6 address:
2519 * 1) Any leading zero's in a h16 component are removed.
2520 * Ex: [0001:0022::] -> [1:22::]
2522 * 2) The longest sequence of zero h16 components are compressed
2523 * into a "::" (elision). If there's a tie, the first is chosen.
2525 * Ex: [0:0:0:0:1:6:7:8] -> [::1:6:7:8]
2526 * [0:0:0:0:1:2::] -> [::1:2:0:0]
2527 * [0:0:1:2:0:0:7:8] -> [::1:2:0:0:7:8]
2529 * 3) If an IPv4 address is attached to the IPv6 address, it's
2530 * also normalized.
2531 * Ex: [::001.002.022.000] -> [::1.2.22.0]
2533 * 4) If an elision is present, but, only represents one h16 component
2534 * it's expanded.
2536 * Ex: [1::2:3:4:5:6:7] -> [1:0:2:3:4:5:6:7]
2538 * 5) If the IPv6 address contains an IPv4 address and there exists
2539 * at least 1 non-zero h16 component the IPv4 address is converted
2540 * into two h16 components, otherwise it's normalized and kept as is.
2542 * Ex: [::192.200.003.4] -> [::192.200.3.4]
2543 * [ffff::192.200.003.4] -> [ffff::c0c8:3041]
2545 * NOTE:
2546 * For unknown scheme types Windows simply copies the address over without any
2547 * changes.
2549 * IPv4 address can be included in an elision if all its components are 0's.
2551 static BOOL canonicalize_ipv6address(const parse_data *data, Uri *uri,
2552 DWORD flags, BOOL computeOnly) {
2553 uri->host_start = uri->canon_len;
2555 if(data->scheme_type == URL_SCHEME_UNKNOWN) {
2556 if(!computeOnly)
2557 memcpy(uri->canon_uri+uri->canon_len, data->host, data->host_len*sizeof(WCHAR));
2558 uri->canon_len += data->host_len;
2559 } else {
2560 USHORT values[8];
2561 INT elision_start;
2562 DWORD i, elision_len;
2564 if(!ipv6_to_number(&(data->ipv6_address), values)) {
2565 TRACE("(%p %p %x %d): Failed to compute numerical value for IPv6 address.\n",
2566 data, uri, flags, computeOnly);
2567 return FALSE;
2570 if(!computeOnly)
2571 uri->canon_uri[uri->canon_len] = '[';
2572 ++uri->canon_len;
2574 /* Find where the elision should occur (if any). */
2575 compute_elision_location(&(data->ipv6_address), values, &elision_start, &elision_len);
2577 TRACE("%p %p %x %d): Elision starts at %d, len=%u\n", data, uri, flags,
2578 computeOnly, elision_start, elision_len);
2580 for(i = 0; i < 8; ++i) {
2581 BOOL in_elision = (elision_start > -1 && i >= elision_start &&
2582 i < elision_start+elision_len);
2583 BOOL do_ipv4 = (i == 6 && data->ipv6_address.ipv4 && !in_elision &&
2584 data->ipv6_address.h16_count == 0);
2586 if(i == elision_start) {
2587 if(!computeOnly) {
2588 uri->canon_uri[uri->canon_len] = ':';
2589 uri->canon_uri[uri->canon_len+1] = ':';
2591 uri->canon_len += 2;
2594 /* We can ignore the current component if we're in the elision. */
2595 if(in_elision)
2596 continue;
2598 /* We only add a ':' if we're not at i == 0, or when we're at
2599 * the very end of elision range since the ':' colon was handled
2600 * earlier. Otherwise we would end up with ":::" after elision.
2602 if(i != 0 && !(elision_start > -1 && i == elision_start+elision_len)) {
2603 if(!computeOnly)
2604 uri->canon_uri[uri->canon_len] = ':';
2605 ++uri->canon_len;
2608 if(do_ipv4) {
2609 UINT val;
2610 DWORD len;
2612 /* Combine the two parts of the IPv4 address values. */
2613 val = values[i];
2614 val <<= 16;
2615 val += values[i+1];
2617 if(!computeOnly)
2618 len = ui2ipv4(uri->canon_uri+uri->canon_len, val);
2619 else
2620 len = ui2ipv4(NULL, val);
2622 uri->canon_len += len;
2623 ++i;
2624 } else {
2625 /* Write a regular h16 component to the URI. */
2627 /* Short circuit for the trivial case. */
2628 if(values[i] == 0) {
2629 if(!computeOnly)
2630 uri->canon_uri[uri->canon_len] = '0';
2631 ++uri->canon_len;
2632 } else {
2633 static const WCHAR formatW[] = {'%','x',0};
2635 if(!computeOnly)
2636 uri->canon_len += swprintf(uri->canon_uri+uri->canon_len, 5,
2637 formatW, values[i]);
2638 else {
2639 WCHAR tmp[5];
2640 uri->canon_len += swprintf(tmp, ARRAY_SIZE(tmp), formatW, values[i]);
2646 /* Add the closing ']'. */
2647 if(!computeOnly)
2648 uri->canon_uri[uri->canon_len] = ']';
2649 ++uri->canon_len;
2652 uri->host_len = uri->canon_len - uri->host_start;
2654 if(!computeOnly)
2655 TRACE("(%p %p %x %d): Canonicalized IPv6 address %s, len=%d\n", data, uri, flags,
2656 computeOnly, debugstr_wn(uri->canon_uri+uri->host_start, uri->host_len),
2657 uri->host_len);
2659 return TRUE;
2662 /* Attempts to canonicalize the host of the URI (if any). */
2663 static BOOL canonicalize_host(const parse_data *data, Uri *uri, DWORD flags, BOOL computeOnly) {
2664 uri->host_start = -1;
2665 uri->host_len = 0;
2666 uri->domain_offset = -1;
2668 if(data->host) {
2669 switch(data->host_type) {
2670 case Uri_HOST_DNS:
2671 uri->host_type = Uri_HOST_DNS;
2672 if(!canonicalize_reg_name(data, uri, flags, computeOnly))
2673 return FALSE;
2675 break;
2676 case Uri_HOST_IPV4:
2677 uri->host_type = Uri_HOST_IPV4;
2678 if(!canonicalize_ipv4address(data, uri, flags, computeOnly))
2679 return FALSE;
2681 break;
2682 case Uri_HOST_IPV6:
2683 if(!canonicalize_ipv6address(data, uri, flags, computeOnly))
2684 return FALSE;
2686 uri->host_type = Uri_HOST_IPV6;
2687 break;
2688 case Uri_HOST_UNKNOWN:
2689 if(data->host_len > 0 || data->scheme_type != URL_SCHEME_FILE) {
2690 uri->host_start = uri->canon_len;
2692 /* Nothing happens to unknown host types. */
2693 if(!computeOnly)
2694 memcpy(uri->canon_uri+uri->canon_len, data->host, data->host_len*sizeof(WCHAR));
2695 uri->canon_len += data->host_len;
2696 uri->host_len = data->host_len;
2699 uri->host_type = Uri_HOST_UNKNOWN;
2700 break;
2701 default:
2702 FIXME("(%p %p %x %d): Canonicalization for host type %d not supported.\n", data,
2703 uri, flags, computeOnly, data->host_type);
2704 return FALSE;
2708 return TRUE;
2711 static BOOL canonicalize_port(const parse_data *data, Uri *uri, DWORD flags, BOOL computeOnly) {
2712 BOOL has_default_port = FALSE;
2713 USHORT default_port = 0;
2714 DWORD i;
2716 uri->port_offset = -1;
2718 /* Check if the scheme has a default port. */
2719 for(i = 0; i < ARRAY_SIZE(default_ports); ++i) {
2720 if(default_ports[i].scheme == data->scheme_type) {
2721 has_default_port = TRUE;
2722 default_port = default_ports[i].port;
2723 break;
2727 uri->has_port = data->has_port || has_default_port;
2729 /* Possible cases:
2730 * 1) Has a port which is the default port.
2731 * 2) Has a port (not the default).
2732 * 3) Doesn't have a port, but, scheme has a default port.
2733 * 4) No port.
2735 if(has_default_port && data->has_port && data->port_value == default_port) {
2736 /* If it's the default port and this flag isn't set, don't do anything. */
2737 if(flags & Uri_CREATE_NO_CANONICALIZE) {
2738 uri->port_offset = uri->canon_len-uri->authority_start;
2739 if(!computeOnly)
2740 uri->canon_uri[uri->canon_len] = ':';
2741 ++uri->canon_len;
2743 if(data->port) {
2744 /* Copy the original port over. */
2745 if(!computeOnly)
2746 memcpy(uri->canon_uri+uri->canon_len, data->port, data->port_len*sizeof(WCHAR));
2747 uri->canon_len += data->port_len;
2748 } else {
2749 if(!computeOnly)
2750 uri->canon_len += ui2str(uri->canon_uri+uri->canon_len, data->port_value);
2751 else
2752 uri->canon_len += ui2str(NULL, data->port_value);
2756 uri->port = default_port;
2757 } else if(data->has_port) {
2758 uri->port_offset = uri->canon_len-uri->authority_start;
2759 if(!computeOnly)
2760 uri->canon_uri[uri->canon_len] = ':';
2761 ++uri->canon_len;
2763 if(flags & Uri_CREATE_NO_CANONICALIZE && data->port) {
2764 /* Copy the original over without changes. */
2765 if(!computeOnly)
2766 memcpy(uri->canon_uri+uri->canon_len, data->port, data->port_len*sizeof(WCHAR));
2767 uri->canon_len += data->port_len;
2768 } else {
2769 if(!computeOnly)
2770 uri->canon_len += ui2str(uri->canon_uri+uri->canon_len, data->port_value);
2771 else
2772 uri->canon_len += ui2str(NULL, data->port_value);
2775 uri->port = data->port_value;
2776 } else if(has_default_port)
2777 uri->port = default_port;
2779 return TRUE;
2782 /* Canonicalizes the authority of the URI represented by the parse_data. */
2783 static BOOL canonicalize_authority(const parse_data *data, Uri *uri, DWORD flags, BOOL computeOnly) {
2784 uri->authority_start = uri->canon_len;
2785 uri->authority_len = 0;
2787 if(!canonicalize_userinfo(data, uri, flags, computeOnly))
2788 return FALSE;
2790 if(!canonicalize_host(data, uri, flags, computeOnly))
2791 return FALSE;
2793 if(!canonicalize_port(data, uri, flags, computeOnly))
2794 return FALSE;
2796 if(uri->host_start != -1 || (data->is_relative && (data->password || data->username)))
2797 uri->authority_len = uri->canon_len - uri->authority_start;
2798 else
2799 uri->authority_start = -1;
2801 return TRUE;
2804 /* Attempts to canonicalize the path of a hierarchical URI.
2806 * Things that happen:
2807 * 1). Forbidden characters are percent encoded, unless the NO_ENCODE_FORBIDDEN
2808 * flag is set or it's a file URI. Forbidden characters are always encoded
2809 * for file schemes regardless and forbidden characters are never encoded
2810 * for unknown scheme types.
2812 * 2). For known scheme types '\\' are changed to '/'.
2814 * 3). Percent encoded, unreserved characters are decoded to their actual values.
2815 * Unless the scheme type is unknown. For file schemes any percent encoded
2816 * character in the unreserved or reserved set is decoded.
2818 * 4). For File schemes if the path is starts with a drive letter and doesn't
2819 * start with a '/' then one is appended.
2820 * Ex: file://c:/test.mp3 -> file:///c:/test.mp3
2822 * 5). Dot segments are removed from the path for all scheme types
2823 * unless NO_CANONICALIZE flag is set. Dot segments aren't removed
2824 * for wildcard scheme types.
2826 * NOTES:
2827 * file://c:/test%20test -> file:///c:/test%2520test
2828 * file://c:/test%3Etest -> file:///c:/test%253Etest
2829 * if Uri_CREATE_FILE_USE_DOS_PATH is not set:
2830 * file:///c:/test%20test -> file:///c:/test%20test
2831 * file:///c:/test%test -> file:///c:/test%25test
2833 static DWORD canonicalize_path_hierarchical(const WCHAR *path, DWORD path_len, URL_SCHEME scheme_type, BOOL has_host, DWORD flags,
2834 BOOL is_implicit_scheme, WCHAR *ret_path) {
2835 const BOOL known_scheme = scheme_type != URL_SCHEME_UNKNOWN;
2836 const BOOL is_file = scheme_type == URL_SCHEME_FILE;
2837 const BOOL is_res = scheme_type == URL_SCHEME_RES;
2838 const WCHAR *ptr;
2839 BOOL escape_pct = FALSE;
2840 DWORD len = 0;
2842 if(!path)
2843 return 0;
2845 ptr = path;
2847 if(is_file && !has_host) {
2848 /* Check if a '/' needs to be appended for the file scheme. */
2849 if(path_len > 1 && is_drive_path(ptr) && !(flags & Uri_CREATE_FILE_USE_DOS_PATH)) {
2850 if(ret_path)
2851 ret_path[len] = '/';
2852 len++;
2853 escape_pct = TRUE;
2854 } else if(*ptr == '/') {
2855 if(!(flags & Uri_CREATE_FILE_USE_DOS_PATH)) {
2856 /* Copy the extra '/' over. */
2857 if(ret_path)
2858 ret_path[len] = '/';
2859 len++;
2861 ++ptr;
2864 if(is_drive_path(ptr)) {
2865 if(ret_path) {
2866 ret_path[len] = *ptr;
2867 /* If there's a '|' after the drive letter, convert it to a ':'. */
2868 ret_path[len+1] = ':';
2870 ptr += 2;
2871 len += 2;
2875 if(!is_file && *path && *path != '/') {
2876 /* Prepend a '/' to the path if it doesn't have one. */
2877 if(ret_path)
2878 ret_path[len] = '/';
2879 len++;
2882 for(; ptr < path+path_len; ++ptr) {
2883 BOOL do_default_action = TRUE;
2885 if(*ptr == '%' && !is_res) {
2886 const WCHAR *tmp = ptr;
2887 WCHAR val;
2889 /* Check if the % represents a valid encoded char, or if it needs encoding. */
2890 BOOL force_encode = !check_pct_encoded(&tmp) && is_file && !(flags&Uri_CREATE_FILE_USE_DOS_PATH);
2891 val = decode_pct_val(ptr);
2893 if(force_encode || escape_pct) {
2894 /* Escape the percent sign in the file URI. */
2895 if(ret_path)
2896 pct_encode_val(*ptr, ret_path+len);
2897 len += 3;
2898 do_default_action = FALSE;
2899 } else if((is_unreserved(val) && known_scheme) ||
2900 (is_file && !is_implicit_scheme && (is_unreserved(val) || is_reserved(val) ||
2901 (val && flags&Uri_CREATE_FILE_USE_DOS_PATH && !is_forbidden_dos_path_char(val))))) {
2902 if(ret_path)
2903 ret_path[len] = val;
2904 len++;
2906 ptr += 2;
2907 continue;
2909 } else if(*ptr == '/' && is_file && (flags & Uri_CREATE_FILE_USE_DOS_PATH)) {
2910 /* Convert the '/' back to a '\\'. */
2911 if(ret_path)
2912 ret_path[len] = '\\';
2913 len++;
2914 do_default_action = FALSE;
2915 } else if(*ptr == '\\' && known_scheme) {
2916 if(!(is_file && (flags & Uri_CREATE_FILE_USE_DOS_PATH))) {
2917 /* Convert '\\' into a '/'. */
2918 if(ret_path)
2919 ret_path[len] = '/';
2920 len++;
2921 do_default_action = FALSE;
2923 } else if(known_scheme && !is_res && is_ascii(*ptr) && !is_unreserved(*ptr) && !is_reserved(*ptr) &&
2924 (!(flags & Uri_CREATE_NO_ENCODE_FORBIDDEN_CHARACTERS) || is_file)) {
2925 if(!is_file || !(flags & Uri_CREATE_FILE_USE_DOS_PATH)) {
2926 /* Escape the forbidden character. */
2927 if(ret_path)
2928 pct_encode_val(*ptr, ret_path+len);
2929 len += 3;
2930 do_default_action = FALSE;
2934 if(do_default_action) {
2935 if(ret_path)
2936 ret_path[len] = *ptr;
2937 len++;
2941 /* Removing the dot segments only happens when it's not in
2942 * computeOnly mode and it's not a wildcard scheme. File schemes
2943 * with USE_DOS_PATH set don't get dot segments removed.
2945 if(!(is_file && (flags & Uri_CREATE_FILE_USE_DOS_PATH)) &&
2946 scheme_type != URL_SCHEME_WILDCARD) {
2947 if(!(flags & Uri_CREATE_NO_CANONICALIZE) && ret_path) {
2948 /* Remove the dot segments (if any) and reset everything to the new
2949 * correct length.
2951 len = remove_dot_segments(ret_path, len);
2955 if(ret_path)
2956 TRACE("Canonicalized path %s len=%d\n", debugstr_wn(ret_path, len), len);
2957 return len;
2960 /* Attempts to canonicalize the path for an opaque URI.
2962 * For known scheme types:
2963 * 1) forbidden characters are percent encoded if
2964 * NO_ENCODE_FORBIDDEN_CHARACTERS isn't set.
2966 * 2) Percent encoded, unreserved characters are decoded
2967 * to their actual values, for known scheme types.
2969 * 3) '\\' are changed to '/' for known scheme types
2970 * except for mailto schemes.
2972 * 4) For file schemes, if USE_DOS_PATH is set all '/'
2973 * are converted to backslashes.
2975 * 5) For file schemes, if USE_DOS_PATH isn't set all '\'
2976 * are converted to forward slashes.
2978 static BOOL canonicalize_path_opaque(const parse_data *data, Uri *uri, DWORD flags, BOOL computeOnly) {
2979 const WCHAR *ptr;
2980 const BOOL known_scheme = data->scheme_type != URL_SCHEME_UNKNOWN;
2981 const BOOL is_file = data->scheme_type == URL_SCHEME_FILE;
2982 const BOOL is_mk = data->scheme_type == URL_SCHEME_MK;
2984 if(!data->path) {
2985 uri->path_start = -1;
2986 uri->path_len = 0;
2987 return TRUE;
2990 uri->path_start = uri->canon_len;
2992 if(is_mk){
2993 /* hijack this flag for SCHEME_MK to tell the function when to start
2994 * converting slashes */
2995 flags |= Uri_CREATE_FILE_USE_DOS_PATH;
2998 /* For javascript: URIs, simply copy path part without any canonicalization */
2999 if(data->scheme_type == URL_SCHEME_JAVASCRIPT) {
3000 if(!computeOnly)
3001 memcpy(uri->canon_uri+uri->canon_len, data->path, data->path_len*sizeof(WCHAR));
3002 uri->path_len = data->path_len;
3003 uri->canon_len += data->path_len;
3004 return TRUE;
3007 /* Windows doesn't allow a "//" to appear after the scheme
3008 * of a URI, if it's an opaque URI.
3010 if(data->scheme && *(data->path) == '/' && *(data->path+1) == '/') {
3011 /* So it inserts a "/." before the "//" if it exists. */
3012 if(!computeOnly) {
3013 uri->canon_uri[uri->canon_len] = '/';
3014 uri->canon_uri[uri->canon_len+1] = '.';
3017 uri->canon_len += 2;
3020 for(ptr = data->path; ptr < data->path+data->path_len; ++ptr) {
3021 BOOL do_default_action = TRUE;
3023 if(*ptr == '%' && known_scheme) {
3024 WCHAR val = decode_pct_val(ptr);
3026 if(is_unreserved(val)) {
3027 if(!computeOnly)
3028 uri->canon_uri[uri->canon_len] = val;
3029 ++uri->canon_len;
3031 ptr += 2;
3032 continue;
3034 } else if(*ptr == '/' && is_file && (flags & Uri_CREATE_FILE_USE_DOS_PATH)) {
3035 if(!computeOnly)
3036 uri->canon_uri[uri->canon_len] = '\\';
3037 ++uri->canon_len;
3038 do_default_action = FALSE;
3039 } else if(*ptr == '\\') {
3040 if((data->is_relative || is_mk || is_file) && !(flags & Uri_CREATE_FILE_USE_DOS_PATH)) {
3041 /* Convert to a '/'. */
3042 if(!computeOnly)
3043 uri->canon_uri[uri->canon_len] = '/';
3044 ++uri->canon_len;
3045 do_default_action = FALSE;
3047 } else if(is_mk && *ptr == ':' && ptr + 1 < data->path + data->path_len && *(ptr + 1) == ':') {
3048 flags &= ~Uri_CREATE_FILE_USE_DOS_PATH;
3049 } else if(known_scheme && is_ascii(*ptr) && !is_unreserved(*ptr) && !is_reserved(*ptr) &&
3050 !(flags & Uri_CREATE_NO_ENCODE_FORBIDDEN_CHARACTERS)) {
3051 if(!(is_file && (flags & Uri_CREATE_FILE_USE_DOS_PATH))) {
3052 if(!computeOnly)
3053 pct_encode_val(*ptr, uri->canon_uri+uri->canon_len);
3054 uri->canon_len += 3;
3055 do_default_action = FALSE;
3059 if(do_default_action) {
3060 if(!computeOnly)
3061 uri->canon_uri[uri->canon_len] = *ptr;
3062 ++uri->canon_len;
3066 if(is_mk && !computeOnly && !(flags & Uri_CREATE_NO_CANONICALIZE)) {
3067 DWORD new_len = remove_dot_segments(uri->canon_uri + uri->path_start,
3068 uri->canon_len - uri->path_start);
3069 uri->canon_len = uri->path_start + new_len;
3072 uri->path_len = uri->canon_len - uri->path_start;
3074 if(!computeOnly)
3075 TRACE("(%p %p %x %d): Canonicalized opaque URI path %s len=%d\n", data, uri, flags, computeOnly,
3076 debugstr_wn(uri->canon_uri+uri->path_start, uri->path_len), uri->path_len);
3077 return TRUE;
3080 /* Determines how the URI represented by the parse_data should be canonicalized.
3082 * Essentially, if the parse_data represents an hierarchical URI then it calls
3083 * canonicalize_authority and the canonicalization functions for the path. If the
3084 * URI is opaque it canonicalizes the path of the URI.
3086 static BOOL canonicalize_hierpart(const parse_data *data, Uri *uri, DWORD flags, BOOL computeOnly) {
3087 if(!data->is_opaque || (data->is_relative && (data->password || data->username))) {
3088 /* "//" is only added for non-wildcard scheme types.
3090 * A "//" is only added to a relative URI if it has a
3091 * host or port component (this only happens if a IUriBuilder
3092 * is generating an IUri).
3094 if((data->is_relative && (data->host || data->has_port)) ||
3095 (!data->is_relative && data->scheme_type != URL_SCHEME_WILDCARD)) {
3096 if(data->scheme_type == URL_SCHEME_WILDCARD)
3097 FIXME("Here\n");
3099 if(!computeOnly) {
3100 INT pos = uri->canon_len;
3102 uri->canon_uri[pos] = '/';
3103 uri->canon_uri[pos+1] = '/';
3105 uri->canon_len += 2;
3108 if(!canonicalize_authority(data, uri, flags, computeOnly))
3109 return FALSE;
3111 if(data->is_relative && (data->password || data->username)) {
3112 if(!canonicalize_path_opaque(data, uri, flags, computeOnly))
3113 return FALSE;
3114 } else {
3115 if(!computeOnly)
3116 uri->path_start = uri->canon_len;
3117 uri->path_len = canonicalize_path_hierarchical(data->path, data->path_len, data->scheme_type, data->host_len != 0,
3118 flags, data->has_implicit_scheme, computeOnly ? NULL : uri->canon_uri+uri->canon_len);
3119 uri->canon_len += uri->path_len;
3120 if(!computeOnly && !uri->path_len)
3121 uri->path_start = -1;
3123 } else {
3124 /* Opaque URI's don't have an authority. */
3125 uri->userinfo_start = uri->userinfo_split = -1;
3126 uri->userinfo_len = 0;
3127 uri->host_start = -1;
3128 uri->host_len = 0;
3129 uri->host_type = Uri_HOST_UNKNOWN;
3130 uri->has_port = FALSE;
3131 uri->authority_start = -1;
3132 uri->authority_len = 0;
3133 uri->domain_offset = -1;
3134 uri->port_offset = -1;
3136 if(is_hierarchical_scheme(data->scheme_type)) {
3137 DWORD i;
3139 /* Absolute URIs aren't displayed for known scheme types
3140 * which should be hierarchical URIs.
3142 uri->display_modifiers |= URI_DISPLAY_NO_ABSOLUTE_URI;
3144 /* Windows also sets the port for these (if they have one). */
3145 for(i = 0; i < ARRAY_SIZE(default_ports); ++i) {
3146 if(data->scheme_type == default_ports[i].scheme) {
3147 uri->has_port = TRUE;
3148 uri->port = default_ports[i].port;
3149 break;
3154 if(!canonicalize_path_opaque(data, uri, flags, computeOnly))
3155 return FALSE;
3158 if(uri->path_start > -1 && !computeOnly)
3159 /* Finding file extensions happens for both types of URIs. */
3160 uri->extension_offset = find_file_extension(uri->canon_uri+uri->path_start, uri->path_len);
3161 else
3162 uri->extension_offset = -1;
3164 return TRUE;
3167 /* Attempts to canonicalize the query string of the URI.
3169 * Things that happen:
3170 * 1) For known scheme types forbidden characters
3171 * are percent encoded, unless the NO_DECODE_EXTRA_INFO flag is set
3172 * or NO_ENCODE_FORBIDDEN_CHARACTERS is set.
3174 * 2) For known scheme types, percent encoded, unreserved characters
3175 * are decoded as long as the NO_DECODE_EXTRA_INFO flag isn't set.
3177 static BOOL canonicalize_query(const parse_data *data, Uri *uri, DWORD flags, BOOL computeOnly) {
3178 const WCHAR *ptr, *end;
3179 const BOOL known_scheme = data->scheme_type != URL_SCHEME_UNKNOWN;
3181 if(!data->query) {
3182 uri->query_start = -1;
3183 uri->query_len = 0;
3184 return TRUE;
3187 uri->query_start = uri->canon_len;
3189 end = data->query+data->query_len;
3190 for(ptr = data->query; ptr < end; ++ptr) {
3191 if(*ptr == '%') {
3192 if(known_scheme && !(flags & Uri_CREATE_NO_DECODE_EXTRA_INFO)) {
3193 WCHAR val = decode_pct_val(ptr);
3194 if(is_unreserved(val)) {
3195 if(!computeOnly)
3196 uri->canon_uri[uri->canon_len] = val;
3197 ++uri->canon_len;
3199 ptr += 2;
3200 continue;
3203 } else if(known_scheme && is_ascii(*ptr) && !is_unreserved(*ptr) && !is_reserved(*ptr)) {
3204 if(!(flags & Uri_CREATE_NO_ENCODE_FORBIDDEN_CHARACTERS) &&
3205 !(flags & Uri_CREATE_NO_DECODE_EXTRA_INFO)) {
3206 if(!computeOnly)
3207 pct_encode_val(*ptr, uri->canon_uri+uri->canon_len);
3208 uri->canon_len += 3;
3209 continue;
3213 if(!computeOnly)
3214 uri->canon_uri[uri->canon_len] = *ptr;
3215 ++uri->canon_len;
3218 uri->query_len = uri->canon_len - uri->query_start;
3220 if(!computeOnly)
3221 TRACE("(%p %p %x %d): Canonicalized query string %s len=%d\n", data, uri, flags,
3222 computeOnly, debugstr_wn(uri->canon_uri+uri->query_start, uri->query_len),
3223 uri->query_len);
3224 return TRUE;
3227 static BOOL canonicalize_fragment(const parse_data *data, Uri *uri, DWORD flags, BOOL computeOnly) {
3228 const WCHAR *ptr, *end;
3229 const BOOL known_scheme = data->scheme_type != URL_SCHEME_UNKNOWN;
3231 if(!data->fragment) {
3232 uri->fragment_start = -1;
3233 uri->fragment_len = 0;
3234 return TRUE;
3237 uri->fragment_start = uri->canon_len;
3239 end = data->fragment + data->fragment_len;
3240 for(ptr = data->fragment; ptr < end; ++ptr) {
3241 if(*ptr == '%') {
3242 if(known_scheme && !(flags & Uri_CREATE_NO_DECODE_EXTRA_INFO)) {
3243 WCHAR val = decode_pct_val(ptr);
3244 if(is_unreserved(val)) {
3245 if(!computeOnly)
3246 uri->canon_uri[uri->canon_len] = val;
3247 ++uri->canon_len;
3249 ptr += 2;
3250 continue;
3253 } else if(known_scheme && is_ascii(*ptr) && !is_unreserved(*ptr) && !is_reserved(*ptr)) {
3254 if(!(flags & Uri_CREATE_NO_ENCODE_FORBIDDEN_CHARACTERS) &&
3255 !(flags & Uri_CREATE_NO_DECODE_EXTRA_INFO)) {
3256 if(!computeOnly)
3257 pct_encode_val(*ptr, uri->canon_uri+uri->canon_len);
3258 uri->canon_len += 3;
3259 continue;
3263 if(!computeOnly)
3264 uri->canon_uri[uri->canon_len] = *ptr;
3265 ++uri->canon_len;
3268 uri->fragment_len = uri->canon_len - uri->fragment_start;
3270 if(!computeOnly)
3271 TRACE("(%p %p %x %d): Canonicalized fragment %s len=%d\n", data, uri, flags,
3272 computeOnly, debugstr_wn(uri->canon_uri+uri->fragment_start, uri->fragment_len),
3273 uri->fragment_len);
3274 return TRUE;
3277 /* Canonicalizes the scheme information specified in the parse_data using the specified flags. */
3278 static BOOL canonicalize_scheme(const parse_data *data, Uri *uri, DWORD flags, BOOL computeOnly) {
3279 uri->scheme_start = -1;
3280 uri->scheme_len = 0;
3282 if(!data->scheme) {
3283 /* The only type of URI that doesn't have to have a scheme is a relative
3284 * URI.
3286 if(!data->is_relative) {
3287 FIXME("(%p %p %x): Unable to determine the scheme type of %s.\n", data,
3288 uri, flags, debugstr_w(data->uri));
3289 return FALSE;
3291 } else {
3292 if(!computeOnly) {
3293 DWORD i;
3294 INT pos = uri->canon_len;
3296 for(i = 0; i < data->scheme_len; ++i) {
3297 /* Scheme name must be lower case after canonicalization. */
3298 uri->canon_uri[i + pos] = towlower(data->scheme[i]);
3301 uri->canon_uri[i + pos] = ':';
3302 uri->scheme_start = pos;
3304 TRACE("(%p %p %x): Canonicalized scheme=%s, len=%d.\n", data, uri, flags,
3305 debugstr_wn(uri->canon_uri+uri->scheme_start, data->scheme_len), data->scheme_len);
3308 /* This happens in both computation modes. */
3309 uri->canon_len += data->scheme_len + 1;
3310 uri->scheme_len = data->scheme_len;
3312 return TRUE;
3315 /* Computes what the length of the URI specified by the parse_data will be
3316 * after canonicalization occurs using the specified flags.
3318 * This function will return a non-zero value indicating the length of the canonicalized
3319 * URI, or -1 on error.
3321 static int compute_canonicalized_length(const parse_data *data, DWORD flags) {
3322 Uri uri;
3324 memset(&uri, 0, sizeof(Uri));
3326 TRACE("(%p %x): Beginning to compute canonicalized length for URI %s\n", data, flags,
3327 debugstr_w(data->uri));
3329 if(!canonicalize_scheme(data, &uri, flags, TRUE)) {
3330 ERR("(%p %x): Failed to compute URI scheme length.\n", data, flags);
3331 return -1;
3334 if(!canonicalize_hierpart(data, &uri, flags, TRUE)) {
3335 ERR("(%p %x): Failed to compute URI hierpart length.\n", data, flags);
3336 return -1;
3339 if(!canonicalize_query(data, &uri, flags, TRUE)) {
3340 ERR("(%p %x): Failed to compute query string length.\n", data, flags);
3341 return -1;
3344 if(!canonicalize_fragment(data, &uri, flags, TRUE)) {
3345 ERR("(%p %x): Failed to compute fragment length.\n", data, flags);
3346 return -1;
3349 TRACE("(%p %x): Finished computing canonicalized URI length. length=%d\n", data, flags, uri.canon_len);
3351 return uri.canon_len;
3354 /* Canonicalizes the URI data specified in the parse_data, using the given flags. If the
3355 * canonicalization succeeds it will store all the canonicalization information
3356 * in the pointer to the Uri.
3358 * To canonicalize a URI this function first computes what the length of the URI
3359 * specified by the parse_data will be. Once this is done it will then perform the actual
3360 * canonicalization of the URI.
3362 static HRESULT canonicalize_uri(const parse_data *data, Uri *uri, DWORD flags) {
3363 INT len;
3365 uri->canon_uri = NULL;
3366 uri->canon_size = uri->canon_len = 0;
3368 TRACE("(%p %p %x): beginning to canonicalize URI %s.\n", data, uri, flags, debugstr_w(data->uri));
3370 /* First try to compute the length of the URI. */
3371 len = compute_canonicalized_length(data, flags);
3372 if(len == -1) {
3373 ERR("(%p %p %x): Could not compute the canonicalized length of %s.\n", data, uri, flags,
3374 debugstr_w(data->uri));
3375 return E_INVALIDARG;
3378 uri->canon_uri = heap_alloc((len+1)*sizeof(WCHAR));
3379 if(!uri->canon_uri)
3380 return E_OUTOFMEMORY;
3382 uri->canon_size = len;
3383 if(!canonicalize_scheme(data, uri, flags, FALSE)) {
3384 ERR("(%p %p %x): Unable to canonicalize the scheme of the URI.\n", data, uri, flags);
3385 return E_INVALIDARG;
3387 uri->scheme_type = data->scheme_type;
3389 if(!canonicalize_hierpart(data, uri, flags, FALSE)) {
3390 ERR("(%p %p %x): Unable to canonicalize the hierpart of the URI\n", data, uri, flags);
3391 return E_INVALIDARG;
3394 if(!canonicalize_query(data, uri, flags, FALSE)) {
3395 ERR("(%p %p %x): Unable to canonicalize query string of the URI.\n",
3396 data, uri, flags);
3397 return E_INVALIDARG;
3400 if(!canonicalize_fragment(data, uri, flags, FALSE)) {
3401 ERR("(%p %p %x): Unable to canonicalize fragment of the URI.\n",
3402 data, uri, flags);
3403 return E_INVALIDARG;
3406 /* There's a possibility we didn't use all the space we allocated
3407 * earlier.
3409 if(uri->canon_len < uri->canon_size) {
3410 /* This happens if the URI is hierarchical and dot
3411 * segments were removed from its path.
3413 WCHAR *tmp = heap_realloc(uri->canon_uri, (uri->canon_len+1)*sizeof(WCHAR));
3414 if(!tmp)
3415 return E_OUTOFMEMORY;
3417 uri->canon_uri = tmp;
3418 uri->canon_size = uri->canon_len;
3421 uri->canon_uri[uri->canon_len] = '\0';
3422 TRACE("(%p %p %x): finished canonicalizing the URI. uri=%s\n", data, uri, flags, debugstr_w(uri->canon_uri));
3424 return S_OK;
3427 static HRESULT get_builder_component(LPWSTR *component, DWORD *component_len,
3428 LPCWSTR source, DWORD source_len,
3429 LPCWSTR *output, DWORD *output_len)
3431 if(!output_len) {
3432 if(output)
3433 *output = NULL;
3434 return E_POINTER;
3437 if(!output) {
3438 *output_len = 0;
3439 return E_POINTER;
3442 if(!(*component) && source) {
3443 /* Allocate 'component', and copy the contents from 'source'
3444 * into the new allocation.
3446 *component = heap_alloc((source_len+1)*sizeof(WCHAR));
3447 if(!(*component))
3448 return E_OUTOFMEMORY;
3450 memcpy(*component, source, source_len*sizeof(WCHAR));
3451 (*component)[source_len] = '\0';
3452 *component_len = source_len;
3455 *output = *component;
3456 *output_len = *component_len;
3457 return *output ? S_OK : S_FALSE;
3460 /* Allocates 'component' and copies the string from 'new_value' into 'component'.
3461 * If 'prefix' is set and 'new_value' isn't NULL, then it checks if 'new_value'
3462 * starts with 'prefix'. If it doesn't then 'prefix' is prepended to 'component'.
3464 * If everything is successful, then will set 'success_flag' in 'flags'.
3466 static HRESULT set_builder_component(LPWSTR *component, DWORD *component_len, LPCWSTR new_value,
3467 WCHAR prefix, DWORD *flags, DWORD success_flag)
3469 heap_free(*component);
3471 if(!new_value) {
3472 *component = NULL;
3473 *component_len = 0;
3474 } else {
3475 BOOL add_prefix = FALSE;
3476 DWORD len = lstrlenW(new_value);
3477 DWORD pos = 0;
3479 if(prefix && *new_value != prefix) {
3480 add_prefix = TRUE;
3481 *component = heap_alloc((len+2)*sizeof(WCHAR));
3482 } else
3483 *component = heap_alloc((len+1)*sizeof(WCHAR));
3485 if(!(*component))
3486 return E_OUTOFMEMORY;
3488 if(add_prefix)
3489 (*component)[pos++] = prefix;
3491 memcpy(*component+pos, new_value, (len+1)*sizeof(WCHAR));
3492 *component_len = len+pos;
3495 *flags |= success_flag;
3496 return S_OK;
3499 static void reset_builder(UriBuilder *builder) {
3500 if(builder->uri)
3501 IUri_Release(&builder->uri->IUri_iface);
3502 builder->uri = NULL;
3504 heap_free(builder->fragment);
3505 builder->fragment = NULL;
3506 builder->fragment_len = 0;
3508 heap_free(builder->host);
3509 builder->host = NULL;
3510 builder->host_len = 0;
3512 heap_free(builder->password);
3513 builder->password = NULL;
3514 builder->password_len = 0;
3516 heap_free(builder->path);
3517 builder->path = NULL;
3518 builder->path_len = 0;
3520 heap_free(builder->query);
3521 builder->query = NULL;
3522 builder->query_len = 0;
3524 heap_free(builder->scheme);
3525 builder->scheme = NULL;
3526 builder->scheme_len = 0;
3528 heap_free(builder->username);
3529 builder->username = NULL;
3530 builder->username_len = 0;
3532 builder->has_port = FALSE;
3533 builder->port = 0;
3534 builder->modified_props = 0;
3537 static HRESULT validate_scheme_name(const UriBuilder *builder, parse_data *data, DWORD flags) {
3538 const WCHAR *component;
3539 const WCHAR *ptr;
3540 const WCHAR **pptr;
3541 DWORD expected_len;
3543 if(builder->scheme) {
3544 ptr = builder->scheme;
3545 expected_len = builder->scheme_len;
3546 } else if(builder->uri && builder->uri->scheme_start > -1) {
3547 ptr = builder->uri->canon_uri+builder->uri->scheme_start;
3548 expected_len = builder->uri->scheme_len;
3549 } else {
3550 static const WCHAR nullW[] = {0};
3551 ptr = nullW;
3552 expected_len = 0;
3555 component = ptr;
3556 pptr = &ptr;
3557 if(parse_scheme(pptr, data, flags, ALLOW_NULL_TERM_SCHEME) &&
3558 data->scheme_len == expected_len) {
3559 if(data->scheme)
3560 TRACE("(%p %p %x): Found valid scheme component %s len=%d.\n", builder, data, flags,
3561 debugstr_wn(data->scheme, data->scheme_len), data->scheme_len);
3562 } else {
3563 TRACE("(%p %p %x): Invalid scheme component found %s.\n", builder, data, flags,
3564 debugstr_wn(component, expected_len));
3565 return INET_E_INVALID_URL;
3568 return S_OK;
3571 static HRESULT validate_username(const UriBuilder *builder, parse_data *data, DWORD flags) {
3572 const WCHAR *ptr;
3573 const WCHAR **pptr;
3574 DWORD expected_len;
3576 if(builder->username) {
3577 ptr = builder->username;
3578 expected_len = builder->username_len;
3579 } else if(!(builder->modified_props & Uri_HAS_USER_NAME) && builder->uri &&
3580 builder->uri->userinfo_start > -1 && builder->uri->userinfo_split != 0) {
3581 /* Just use the username from the base Uri. */
3582 data->username = builder->uri->canon_uri+builder->uri->userinfo_start;
3583 data->username_len = (builder->uri->userinfo_split > -1) ?
3584 builder->uri->userinfo_split : builder->uri->userinfo_len;
3585 ptr = NULL;
3586 } else {
3587 ptr = NULL;
3588 expected_len = 0;
3591 if(ptr) {
3592 const WCHAR *component = ptr;
3593 pptr = &ptr;
3594 if(parse_username(pptr, data, flags, ALLOW_NULL_TERM_USER_NAME) &&
3595 data->username_len == expected_len)
3596 TRACE("(%p %p %x): Found valid username component %s len=%d.\n", builder, data, flags,
3597 debugstr_wn(data->username, data->username_len), data->username_len);
3598 else {
3599 TRACE("(%p %p %x): Invalid username component found %s.\n", builder, data, flags,
3600 debugstr_wn(component, expected_len));
3601 return INET_E_INVALID_URL;
3605 return S_OK;
3608 static HRESULT validate_password(const UriBuilder *builder, parse_data *data, DWORD flags) {
3609 const WCHAR *ptr;
3610 const WCHAR **pptr;
3611 DWORD expected_len;
3613 if(builder->password) {
3614 ptr = builder->password;
3615 expected_len = builder->password_len;
3616 } else if(!(builder->modified_props & Uri_HAS_PASSWORD) && builder->uri &&
3617 builder->uri->userinfo_split > -1) {
3618 data->password = builder->uri->canon_uri+builder->uri->userinfo_start+builder->uri->userinfo_split+1;
3619 data->password_len = builder->uri->userinfo_len-builder->uri->userinfo_split-1;
3620 ptr = NULL;
3621 } else {
3622 ptr = NULL;
3623 expected_len = 0;
3626 if(ptr) {
3627 const WCHAR *component = ptr;
3628 pptr = &ptr;
3629 if(parse_password(pptr, data, flags, ALLOW_NULL_TERM_PASSWORD) &&
3630 data->password_len == expected_len)
3631 TRACE("(%p %p %x): Found valid password component %s len=%d.\n", builder, data, flags,
3632 debugstr_wn(data->password, data->password_len), data->password_len);
3633 else {
3634 TRACE("(%p %p %x): Invalid password component found %s.\n", builder, data, flags,
3635 debugstr_wn(component, expected_len));
3636 return INET_E_INVALID_URL;
3640 return S_OK;
3643 static HRESULT validate_userinfo(const UriBuilder *builder, parse_data *data, DWORD flags) {
3644 HRESULT hr;
3646 hr = validate_username(builder, data, flags);
3647 if(FAILED(hr))
3648 return hr;
3650 hr = validate_password(builder, data, flags);
3651 if(FAILED(hr))
3652 return hr;
3654 return S_OK;
3657 static HRESULT validate_host(const UriBuilder *builder, parse_data *data) {
3658 const WCHAR *ptr;
3659 const WCHAR **pptr;
3660 DWORD expected_len;
3662 if(builder->host) {
3663 ptr = builder->host;
3664 expected_len = builder->host_len;
3665 } else if(!(builder->modified_props & Uri_HAS_HOST) && builder->uri && builder->uri->host_start > -1) {
3666 ptr = builder->uri->canon_uri + builder->uri->host_start;
3667 expected_len = builder->uri->host_len;
3668 } else
3669 ptr = NULL;
3671 if(ptr) {
3672 const WCHAR *component = ptr;
3673 DWORD extras = ALLOW_BRACKETLESS_IP_LITERAL|IGNORE_PORT_DELIMITER|SKIP_IP_FUTURE_CHECK;
3674 pptr = &ptr;
3676 if(parse_host(pptr, data, extras) && data->host_len == expected_len)
3677 TRACE("(%p %p): Found valid host name %s len=%d type=%d.\n", builder, data,
3678 debugstr_wn(data->host, data->host_len), data->host_len, data->host_type);
3679 else {
3680 TRACE("(%p %p): Invalid host name found %s.\n", builder, data,
3681 debugstr_wn(component, expected_len));
3682 return INET_E_INVALID_URL;
3686 return S_OK;
3689 static void setup_port(const UriBuilder *builder, parse_data *data, DWORD flags) {
3690 if(builder->modified_props & Uri_HAS_PORT) {
3691 if(builder->has_port) {
3692 data->has_port = TRUE;
3693 data->port_value = builder->port;
3695 } else if(builder->uri && builder->uri->has_port) {
3696 data->has_port = TRUE;
3697 data->port_value = builder->uri->port;
3700 if(data->has_port)
3701 TRACE("(%p %p %x): Using %u as port for IUri.\n", builder, data, flags, data->port_value);
3704 static HRESULT validate_path(const UriBuilder *builder, parse_data *data, DWORD flags) {
3705 const WCHAR *ptr = NULL;
3706 const WCHAR *component;
3707 const WCHAR **pptr;
3708 DWORD expected_len;
3709 BOOL check_len = TRUE;
3710 BOOL valid = FALSE;
3712 if(builder->path) {
3713 ptr = builder->path;
3714 expected_len = builder->path_len;
3715 } else if(!(builder->modified_props & Uri_HAS_PATH) &&
3716 builder->uri && builder->uri->path_start > -1) {
3717 ptr = builder->uri->canon_uri+builder->uri->path_start;
3718 expected_len = builder->uri->path_len;
3719 } else {
3720 static const WCHAR nullW[] = {0};
3721 ptr = nullW;
3722 check_len = FALSE;
3723 expected_len = -1;
3726 component = ptr;
3727 pptr = &ptr;
3729 /* How the path is validated depends on what type of
3730 * URI it is.
3732 valid = data->is_opaque ?
3733 parse_path_opaque(pptr, data, flags) : parse_path_hierarchical(pptr, data, flags);
3735 if(!valid || (check_len && expected_len != data->path_len)) {
3736 TRACE("(%p %p %x): Invalid path component %s.\n", builder, data, flags,
3737 debugstr_wn(component, expected_len) );
3738 return INET_E_INVALID_URL;
3741 TRACE("(%p %p %x): Valid path component %s len=%d.\n", builder, data, flags,
3742 debugstr_wn(data->path, data->path_len), data->path_len);
3744 return S_OK;
3747 static HRESULT validate_query(const UriBuilder *builder, parse_data *data, DWORD flags) {
3748 const WCHAR *ptr = NULL;
3749 const WCHAR **pptr;
3750 DWORD expected_len;
3752 if(builder->query) {
3753 ptr = builder->query;
3754 expected_len = builder->query_len;
3755 } else if(!(builder->modified_props & Uri_HAS_QUERY) && builder->uri &&
3756 builder->uri->query_start > -1) {
3757 ptr = builder->uri->canon_uri+builder->uri->query_start;
3758 expected_len = builder->uri->query_len;
3761 if(ptr) {
3762 const WCHAR *component = ptr;
3763 pptr = &ptr;
3765 if(parse_query(pptr, data, flags) && expected_len == data->query_len)
3766 TRACE("(%p %p %x): Valid query component %s len=%d.\n", builder, data, flags,
3767 debugstr_wn(data->query, data->query_len), data->query_len);
3768 else {
3769 TRACE("(%p %p %x): Invalid query component %s.\n", builder, data, flags,
3770 debugstr_wn(component, expected_len));
3771 return INET_E_INVALID_URL;
3775 return S_OK;
3778 static HRESULT validate_fragment(const UriBuilder *builder, parse_data *data, DWORD flags) {
3779 const WCHAR *ptr = NULL;
3780 const WCHAR **pptr;
3781 DWORD expected_len;
3783 if(builder->fragment) {
3784 ptr = builder->fragment;
3785 expected_len = builder->fragment_len;
3786 } else if(!(builder->modified_props & Uri_HAS_FRAGMENT) && builder->uri &&
3787 builder->uri->fragment_start > -1) {
3788 ptr = builder->uri->canon_uri+builder->uri->fragment_start;
3789 expected_len = builder->uri->fragment_len;
3792 if(ptr) {
3793 const WCHAR *component = ptr;
3794 pptr = &ptr;
3796 if(parse_fragment(pptr, data, flags) && expected_len == data->fragment_len)
3797 TRACE("(%p %p %x): Valid fragment component %s len=%d.\n", builder, data, flags,
3798 debugstr_wn(data->fragment, data->fragment_len), data->fragment_len);
3799 else {
3800 TRACE("(%p %p %x): Invalid fragment component %s.\n", builder, data, flags,
3801 debugstr_wn(component, expected_len));
3802 return INET_E_INVALID_URL;
3806 return S_OK;
3809 static HRESULT validate_components(const UriBuilder *builder, parse_data *data, DWORD flags) {
3810 HRESULT hr;
3812 memset(data, 0, sizeof(parse_data));
3814 TRACE("(%p %p %x): Beginning to validate builder components.\n", builder, data, flags);
3816 hr = validate_scheme_name(builder, data, flags);
3817 if(FAILED(hr))
3818 return hr;
3820 /* Extra validation for file schemes. */
3821 if(data->scheme_type == URL_SCHEME_FILE) {
3822 if((builder->password || (builder->uri && builder->uri->userinfo_split > -1)) ||
3823 (builder->username || (builder->uri && builder->uri->userinfo_start > -1))) {
3824 TRACE("(%p %p %x): File schemes can't contain a username or password.\n",
3825 builder, data, flags);
3826 return INET_E_INVALID_URL;
3830 hr = validate_userinfo(builder, data, flags);
3831 if(FAILED(hr))
3832 return hr;
3834 hr = validate_host(builder, data);
3835 if(FAILED(hr))
3836 return hr;
3838 setup_port(builder, data, flags);
3840 /* The URI is opaque if it doesn't have an authority component. */
3841 if(!data->is_relative)
3842 data->is_opaque = !data->username && !data->password && !data->host && !data->has_port
3843 && data->scheme_type != URL_SCHEME_FILE;
3844 else
3845 data->is_opaque = !data->host && !data->has_port;
3847 hr = validate_path(builder, data, flags);
3848 if(FAILED(hr))
3849 return hr;
3851 hr = validate_query(builder, data, flags);
3852 if(FAILED(hr))
3853 return hr;
3855 hr = validate_fragment(builder, data, flags);
3856 if(FAILED(hr))
3857 return hr;
3859 TRACE("(%p %p %x): Finished validating builder components.\n", builder, data, flags);
3861 return S_OK;
3864 static HRESULT compare_file_paths(const Uri *a, const Uri *b, BOOL *ret)
3866 WCHAR *canon_path_a, *canon_path_b;
3867 DWORD len_a, len_b;
3869 if(!a->path_len) {
3870 *ret = !b->path_len;
3871 return S_OK;
3874 if(!b->path_len) {
3875 *ret = FALSE;
3876 return S_OK;
3879 /* Fast path */
3880 if(a->path_len == b->path_len && !wcsnicmp(a->canon_uri+a->path_start, b->canon_uri+b->path_start, a->path_len)) {
3881 *ret = TRUE;
3882 return S_OK;
3885 len_a = canonicalize_path_hierarchical(a->canon_uri+a->path_start, a->path_len, a->scheme_type, FALSE, 0, FALSE, NULL);
3886 len_b = canonicalize_path_hierarchical(b->canon_uri+b->path_start, b->path_len, b->scheme_type, FALSE, 0, FALSE, NULL);
3888 canon_path_a = heap_alloc(len_a*sizeof(WCHAR));
3889 if(!canon_path_a)
3890 return E_OUTOFMEMORY;
3891 canon_path_b = heap_alloc(len_b*sizeof(WCHAR));
3892 if(!canon_path_b) {
3893 heap_free(canon_path_a);
3894 return E_OUTOFMEMORY;
3897 len_a = canonicalize_path_hierarchical(a->canon_uri+a->path_start, a->path_len, a->scheme_type, FALSE, 0, FALSE, canon_path_a);
3898 len_b = canonicalize_path_hierarchical(b->canon_uri+b->path_start, b->path_len, b->scheme_type, FALSE, 0, FALSE, canon_path_b);
3900 *ret = len_a == len_b && !wcsnicmp(canon_path_a, canon_path_b, len_a);
3902 heap_free(canon_path_a);
3903 heap_free(canon_path_b);
3904 return S_OK;
3907 /* Checks if the two Uri's are logically equivalent. It's a simple
3908 * comparison, since they are both of type Uri, and it can access
3909 * the properties of each Uri directly without the need to go
3910 * through the "IUri_Get*" interface calls.
3912 static HRESULT compare_uris(const Uri *a, const Uri *b, BOOL *ret) {
3913 const BOOL known_scheme = a->scheme_type != URL_SCHEME_UNKNOWN;
3914 const BOOL are_hierarchical = a->authority_start > -1 && b->authority_start > -1;
3915 HRESULT hres;
3917 *ret = FALSE;
3919 if(a->scheme_type != b->scheme_type)
3920 return S_OK;
3922 /* Only compare the scheme names (if any) if their unknown scheme types. */
3923 if(!known_scheme) {
3924 if((a->scheme_start > -1 && b->scheme_start > -1) &&
3925 (a->scheme_len == b->scheme_len)) {
3926 /* Make sure the schemes are the same. */
3927 if(StrCmpNW(a->canon_uri+a->scheme_start, b->canon_uri+b->scheme_start, a->scheme_len))
3928 return S_OK;
3929 } else if(a->scheme_len != b->scheme_len)
3930 /* One of the Uri's has a scheme name, while the other doesn't. */
3931 return S_OK;
3934 /* If they have a userinfo component, perform case sensitive compare. */
3935 if((a->userinfo_start > -1 && b->userinfo_start > -1) &&
3936 (a->userinfo_len == b->userinfo_len)) {
3937 if(StrCmpNW(a->canon_uri+a->userinfo_start, b->canon_uri+b->userinfo_start, a->userinfo_len))
3938 return S_OK;
3939 } else if(a->userinfo_len != b->userinfo_len)
3940 /* One of the Uri's had a userinfo, while the other one doesn't. */
3941 return S_OK;
3943 /* Check if they have a host name. */
3944 if((a->host_start > -1 && b->host_start > -1) &&
3945 (a->host_len == b->host_len)) {
3946 /* Perform a case insensitive compare if they are a known scheme type. */
3947 if(known_scheme) {
3948 if(StrCmpNIW(a->canon_uri+a->host_start, b->canon_uri+b->host_start, a->host_len))
3949 return S_OK;
3950 } else if(StrCmpNW(a->canon_uri+a->host_start, b->canon_uri+b->host_start, a->host_len))
3951 return S_OK;
3952 } else if(a->host_len != b->host_len)
3953 /* One of the Uri's had a host, while the other one didn't. */
3954 return S_OK;
3956 if(a->has_port && b->has_port) {
3957 if(a->port != b->port)
3958 return S_OK;
3959 } else if(a->has_port || b->has_port)
3960 /* One had a port, while the other one didn't. */
3961 return S_OK;
3963 /* Windows is weird with how it handles paths. For example
3964 * One URI could be "http://google.com" (after canonicalization)
3965 * and one could be "http://google.com/" and the IsEqual function
3966 * would still evaluate to TRUE, but, only if they are both hierarchical
3967 * URIs.
3969 if(a->scheme_type == URL_SCHEME_FILE) {
3970 BOOL cmp;
3972 hres = compare_file_paths(a, b, &cmp);
3973 if(FAILED(hres) || !cmp)
3974 return hres;
3975 } else if((a->path_start > -1 && b->path_start > -1) &&
3976 (a->path_len == b->path_len)) {
3977 if(StrCmpNW(a->canon_uri+a->path_start, b->canon_uri+b->path_start, a->path_len))
3978 return S_OK;
3979 } else if(are_hierarchical && a->path_len == -1 && b->path_len == 0) {
3980 if(*(a->canon_uri+a->path_start) != '/')
3981 return S_OK;
3982 } else if(are_hierarchical && b->path_len == 1 && a->path_len == 0) {
3983 if(*(b->canon_uri+b->path_start) != '/')
3984 return S_OK;
3985 } else if(a->path_len != b->path_len)
3986 return S_OK;
3988 /* Compare the query strings of the two URIs. */
3989 if((a->query_start > -1 && b->query_start > -1) &&
3990 (a->query_len == b->query_len)) {
3991 if(StrCmpNW(a->canon_uri+a->query_start, b->canon_uri+b->query_start, a->query_len))
3992 return S_OK;
3993 } else if(a->query_len != b->query_len)
3994 return S_OK;
3996 if((a->fragment_start > -1 && b->fragment_start > -1) &&
3997 (a->fragment_len == b->fragment_len)) {
3998 if(StrCmpNW(a->canon_uri+a->fragment_start, b->canon_uri+b->fragment_start, a->fragment_len))
3999 return S_OK;
4000 } else if(a->fragment_len != b->fragment_len)
4001 return S_OK;
4003 /* If we get here, the two URIs are equivalent. */
4004 *ret = TRUE;
4005 return S_OK;
4008 static void convert_to_dos_path(const WCHAR *path, DWORD path_len,
4009 WCHAR *output, DWORD *output_len)
4011 const WCHAR *ptr = path;
4013 if(path_len > 3 && *ptr == '/' && is_drive_path(path+1))
4014 /* Skip over the leading / before the drive path. */
4015 ++ptr;
4017 for(; ptr < path+path_len; ++ptr) {
4018 if(*ptr == '/') {
4019 if(output)
4020 *output++ = '\\';
4021 (*output_len)++;
4022 } else {
4023 if(output)
4024 *output++ = *ptr;
4025 (*output_len)++;
4030 /* Generates a raw uri string using the parse_data. */
4031 static DWORD generate_raw_uri(const parse_data *data, BSTR uri, DWORD flags) {
4032 DWORD length = 0;
4034 if(data->scheme) {
4035 if(uri) {
4036 memcpy(uri, data->scheme, data->scheme_len*sizeof(WCHAR));
4037 uri[data->scheme_len] = ':';
4039 length += data->scheme_len+1;
4042 if(!data->is_opaque) {
4043 /* For the "//" which appears before the authority component. */
4044 if(uri) {
4045 uri[length] = '/';
4046 uri[length+1] = '/';
4048 length += 2;
4050 /* Check if we need to add the "\\" before the host name
4051 * of a UNC server name in a DOS path.
4053 if(flags & RAW_URI_CONVERT_TO_DOS_PATH &&
4054 data->scheme_type == URL_SCHEME_FILE && data->host) {
4055 if(uri) {
4056 uri[length] = '\\';
4057 uri[length+1] = '\\';
4059 length += 2;
4063 if(data->username) {
4064 if(uri)
4065 memcpy(uri+length, data->username, data->username_len*sizeof(WCHAR));
4066 length += data->username_len;
4069 if(data->password) {
4070 if(uri) {
4071 uri[length] = ':';
4072 memcpy(uri+length+1, data->password, data->password_len*sizeof(WCHAR));
4074 length += data->password_len+1;
4077 if(data->password || data->username) {
4078 if(uri)
4079 uri[length] = '@';
4080 ++length;
4083 if(data->host) {
4084 /* IPv6 addresses get the brackets added around them if they don't already
4085 * have them.
4087 const BOOL add_brackets = data->host_type == Uri_HOST_IPV6 && *(data->host) != '[';
4088 if(add_brackets) {
4089 if(uri)
4090 uri[length] = '[';
4091 ++length;
4094 if(uri)
4095 memcpy(uri+length, data->host, data->host_len*sizeof(WCHAR));
4096 length += data->host_len;
4098 if(add_brackets) {
4099 if(uri)
4100 uri[length] = ']';
4101 length++;
4105 if(data->has_port) {
4106 /* The port isn't included in the raw uri if it's the default
4107 * port for the scheme type.
4109 DWORD i;
4110 BOOL is_default = FALSE;
4112 for(i = 0; i < ARRAY_SIZE(default_ports); ++i) {
4113 if(data->scheme_type == default_ports[i].scheme &&
4114 data->port_value == default_ports[i].port)
4115 is_default = TRUE;
4118 if(!is_default || flags & RAW_URI_FORCE_PORT_DISP) {
4119 if(uri)
4120 uri[length] = ':';
4121 ++length;
4123 if(uri)
4124 length += ui2str(uri+length, data->port_value);
4125 else
4126 length += ui2str(NULL, data->port_value);
4130 /* Check if a '/' should be added before the path for hierarchical URIs. */
4131 if(!data->is_opaque && data->path && *(data->path) != '/') {
4132 if(uri)
4133 uri[length] = '/';
4134 ++length;
4137 if(data->path) {
4138 if(!data->is_opaque && data->scheme_type == URL_SCHEME_FILE &&
4139 flags & RAW_URI_CONVERT_TO_DOS_PATH) {
4140 DWORD len = 0;
4142 if(uri)
4143 convert_to_dos_path(data->path, data->path_len, uri+length, &len);
4144 else
4145 convert_to_dos_path(data->path, data->path_len, NULL, &len);
4147 length += len;
4148 } else {
4149 if(uri)
4150 memcpy(uri+length, data->path, data->path_len*sizeof(WCHAR));
4151 length += data->path_len;
4155 if(data->query) {
4156 if(uri)
4157 memcpy(uri+length, data->query, data->query_len*sizeof(WCHAR));
4158 length += data->query_len;
4161 if(data->fragment) {
4162 if(uri)
4163 memcpy(uri+length, data->fragment, data->fragment_len*sizeof(WCHAR));
4164 length += data->fragment_len;
4167 if(uri)
4168 TRACE("(%p %p): Generated raw uri=%s len=%d\n", data, uri, debugstr_wn(uri, length), length);
4169 else
4170 TRACE("(%p %p): Computed raw uri len=%d\n", data, uri, length);
4172 return length;
4175 static HRESULT generate_uri(const UriBuilder *builder, const parse_data *data, Uri *uri, DWORD flags) {
4176 HRESULT hr;
4177 DWORD length = generate_raw_uri(data, NULL, 0);
4178 uri->raw_uri = SysAllocStringLen(NULL, length);
4179 if(!uri->raw_uri)
4180 return E_OUTOFMEMORY;
4182 generate_raw_uri(data, uri->raw_uri, 0);
4184 hr = canonicalize_uri(data, uri, flags);
4185 if(FAILED(hr)) {
4186 if(hr == E_INVALIDARG)
4187 return INET_E_INVALID_URL;
4188 return hr;
4191 uri->create_flags = flags;
4192 return S_OK;
4195 static inline Uri* impl_from_IUri(IUri *iface)
4197 return CONTAINING_RECORD(iface, Uri, IUri_iface);
4200 static inline void destroy_uri_obj(Uri *This)
4202 SysFreeString(This->raw_uri);
4203 heap_free(This->canon_uri);
4204 heap_free(This);
4207 static HRESULT WINAPI Uri_QueryInterface(IUri *iface, REFIID riid, void **ppv)
4209 Uri *This = impl_from_IUri(iface);
4211 if(IsEqualGUID(&IID_IUnknown, riid)) {
4212 TRACE("(%p)->(IID_IUnknown %p)\n", This, ppv);
4213 *ppv = &This->IUri_iface;
4214 }else if(IsEqualGUID(&IID_IUri, riid)) {
4215 TRACE("(%p)->(IID_IUri %p)\n", This, ppv);
4216 *ppv = &This->IUri_iface;
4217 }else if(IsEqualGUID(&IID_IUriBuilderFactory, riid)) {
4218 TRACE("(%p)->(IID_IUriBuilderFactory %p)\n", This, ppv);
4219 *ppv = &This->IUriBuilderFactory_iface;
4220 }else if(IsEqualGUID(&IID_IPersistStream, riid)) {
4221 TRACE("(%p)->(IID_IPersistStream %p)\n", This, ppv);
4222 *ppv = &This->IPersistStream_iface;
4223 }else if(IsEqualGUID(&IID_IMarshal, riid)) {
4224 TRACE("(%p)->(IID_IMarshal %p)\n", This, ppv);
4225 *ppv = &This->IMarshal_iface;
4226 }else if(IsEqualGUID(&IID_IUriObj, riid)) {
4227 TRACE("(%p)->(IID_IUriObj %p)\n", This, ppv);
4228 *ppv = This;
4229 return S_OK;
4230 }else {
4231 TRACE("(%p)->(%s %p)\n", This, debugstr_guid(riid), ppv);
4232 *ppv = NULL;
4233 return E_NOINTERFACE;
4236 IUnknown_AddRef((IUnknown*)*ppv);
4237 return S_OK;
4240 static ULONG WINAPI Uri_AddRef(IUri *iface)
4242 Uri *This = impl_from_IUri(iface);
4243 LONG ref = InterlockedIncrement(&This->ref);
4245 TRACE("(%p) ref=%d\n", This, ref);
4247 return ref;
4250 static ULONG WINAPI Uri_Release(IUri *iface)
4252 Uri *This = impl_from_IUri(iface);
4253 LONG ref = InterlockedDecrement(&This->ref);
4255 TRACE("(%p) ref=%d\n", This, ref);
4257 if(!ref)
4258 destroy_uri_obj(This);
4260 return ref;
4263 static HRESULT WINAPI Uri_GetPropertyBSTR(IUri *iface, Uri_PROPERTY uriProp, BSTR *pbstrProperty, DWORD dwFlags)
4265 Uri *This = impl_from_IUri(iface);
4266 HRESULT hres;
4267 TRACE("(%p %s)->(%d %p %x)\n", This, debugstr_w(This->canon_uri), uriProp, pbstrProperty, dwFlags);
4269 if(!This->create_flags)
4270 return E_UNEXPECTED;
4271 if(!pbstrProperty)
4272 return E_POINTER;
4274 if(uriProp > Uri_PROPERTY_STRING_LAST) {
4275 /* It only returns S_FALSE for the ZONE property... */
4276 if(uriProp == Uri_PROPERTY_ZONE) {
4277 *pbstrProperty = SysAllocStringLen(NULL, 0);
4278 if(!(*pbstrProperty))
4279 return E_OUTOFMEMORY;
4280 return S_FALSE;
4283 *pbstrProperty = NULL;
4284 return E_INVALIDARG;
4287 /* Don't have support for flags yet. */
4288 if(dwFlags) {
4289 FIXME("(%p)->(%d %p %x)\n", This, uriProp, pbstrProperty, dwFlags);
4290 return E_NOTIMPL;
4293 switch(uriProp) {
4294 case Uri_PROPERTY_ABSOLUTE_URI:
4295 if(This->display_modifiers & URI_DISPLAY_NO_ABSOLUTE_URI) {
4296 *pbstrProperty = SysAllocStringLen(NULL, 0);
4297 hres = S_FALSE;
4298 } else {
4299 if(This->scheme_type != URL_SCHEME_UNKNOWN && This->userinfo_start > -1) {
4300 if(This->userinfo_len == 0) {
4301 /* Don't include the '@' after the userinfo component. */
4302 *pbstrProperty = SysAllocStringLen(NULL, This->canon_len-1);
4303 hres = S_OK;
4304 if(*pbstrProperty) {
4305 /* Copy everything before it. */
4306 memcpy(*pbstrProperty, This->canon_uri, This->userinfo_start*sizeof(WCHAR));
4308 /* And everything after it. */
4309 memcpy(*pbstrProperty+This->userinfo_start, This->canon_uri+This->userinfo_start+1,
4310 (This->canon_len-This->userinfo_start-1)*sizeof(WCHAR));
4312 } else if(This->userinfo_split == 0 && This->userinfo_len == 1) {
4313 /* Don't include the ":@" */
4314 *pbstrProperty = SysAllocStringLen(NULL, This->canon_len-2);
4315 hres = S_OK;
4316 if(*pbstrProperty) {
4317 memcpy(*pbstrProperty, This->canon_uri, This->userinfo_start*sizeof(WCHAR));
4318 memcpy(*pbstrProperty+This->userinfo_start, This->canon_uri+This->userinfo_start+2,
4319 (This->canon_len-This->userinfo_start-2)*sizeof(WCHAR));
4321 } else {
4322 *pbstrProperty = SysAllocString(This->canon_uri);
4323 hres = S_OK;
4325 } else {
4326 *pbstrProperty = SysAllocString(This->canon_uri);
4327 hres = S_OK;
4331 if(!(*pbstrProperty))
4332 hres = E_OUTOFMEMORY;
4334 break;
4335 case Uri_PROPERTY_AUTHORITY:
4336 if(This->authority_start > -1) {
4337 if(This->port_offset > -1 && is_default_port(This->scheme_type, This->port) &&
4338 This->display_modifiers & URI_DISPLAY_NO_DEFAULT_PORT_AUTH)
4339 /* Don't include the port in the authority component. */
4340 *pbstrProperty = SysAllocStringLen(This->canon_uri+This->authority_start, This->port_offset);
4341 else
4342 *pbstrProperty = SysAllocStringLen(This->canon_uri+This->authority_start, This->authority_len);
4343 hres = S_OK;
4344 } else {
4345 *pbstrProperty = SysAllocStringLen(NULL, 0);
4346 hres = S_FALSE;
4349 if(!(*pbstrProperty))
4350 hres = E_OUTOFMEMORY;
4352 break;
4353 case Uri_PROPERTY_DISPLAY_URI:
4354 /* The Display URI contains everything except for the userinfo for known
4355 * scheme types.
4357 if(This->scheme_type != URL_SCHEME_UNKNOWN && This->userinfo_start > -1) {
4358 *pbstrProperty = SysAllocStringLen(NULL, This->canon_len-This->userinfo_len);
4360 if(*pbstrProperty) {
4361 /* Copy everything before the userinfo over. */
4362 memcpy(*pbstrProperty, This->canon_uri, This->userinfo_start*sizeof(WCHAR));
4363 /* Copy everything after the userinfo over. */
4364 memcpy(*pbstrProperty+This->userinfo_start,
4365 This->canon_uri+This->userinfo_start+This->userinfo_len+1,
4366 (This->canon_len-(This->userinfo_start+This->userinfo_len+1))*sizeof(WCHAR));
4368 } else
4369 *pbstrProperty = SysAllocString(This->canon_uri);
4371 if(!(*pbstrProperty))
4372 hres = E_OUTOFMEMORY;
4373 else
4374 hres = S_OK;
4376 break;
4377 case Uri_PROPERTY_DOMAIN:
4378 if(This->domain_offset > -1) {
4379 *pbstrProperty = SysAllocStringLen(This->canon_uri+This->host_start+This->domain_offset,
4380 This->host_len-This->domain_offset);
4381 hres = S_OK;
4382 } else {
4383 *pbstrProperty = SysAllocStringLen(NULL, 0);
4384 hres = S_FALSE;
4387 if(!(*pbstrProperty))
4388 hres = E_OUTOFMEMORY;
4390 break;
4391 case Uri_PROPERTY_EXTENSION:
4392 if(This->extension_offset > -1) {
4393 *pbstrProperty = SysAllocStringLen(This->canon_uri+This->path_start+This->extension_offset,
4394 This->path_len-This->extension_offset);
4395 hres = S_OK;
4396 } else {
4397 *pbstrProperty = SysAllocStringLen(NULL, 0);
4398 hres = S_FALSE;
4401 if(!(*pbstrProperty))
4402 hres = E_OUTOFMEMORY;
4404 break;
4405 case Uri_PROPERTY_FRAGMENT:
4406 if(This->fragment_start > -1) {
4407 *pbstrProperty = SysAllocStringLen(This->canon_uri+This->fragment_start, This->fragment_len);
4408 hres = S_OK;
4409 } else {
4410 *pbstrProperty = SysAllocStringLen(NULL, 0);
4411 hres = S_FALSE;
4414 if(!(*pbstrProperty))
4415 hres = E_OUTOFMEMORY;
4417 break;
4418 case Uri_PROPERTY_HOST:
4419 if(This->host_start > -1) {
4420 /* The '[' and ']' aren't included for IPv6 addresses. */
4421 if(This->host_type == Uri_HOST_IPV6)
4422 *pbstrProperty = SysAllocStringLen(This->canon_uri+This->host_start+1, This->host_len-2);
4423 else
4424 *pbstrProperty = SysAllocStringLen(This->canon_uri+This->host_start, This->host_len);
4426 hres = S_OK;
4427 } else {
4428 *pbstrProperty = SysAllocStringLen(NULL, 0);
4429 hres = S_FALSE;
4432 if(!(*pbstrProperty))
4433 hres = E_OUTOFMEMORY;
4435 break;
4436 case Uri_PROPERTY_PASSWORD:
4437 if(This->userinfo_split > -1) {
4438 *pbstrProperty = SysAllocStringLen(
4439 This->canon_uri+This->userinfo_start+This->userinfo_split+1,
4440 This->userinfo_len-This->userinfo_split-1);
4441 hres = S_OK;
4442 } else {
4443 *pbstrProperty = SysAllocStringLen(NULL, 0);
4444 hres = S_FALSE;
4447 if(!(*pbstrProperty))
4448 return E_OUTOFMEMORY;
4450 break;
4451 case Uri_PROPERTY_PATH:
4452 if(This->path_start > -1) {
4453 *pbstrProperty = SysAllocStringLen(This->canon_uri+This->path_start, This->path_len);
4454 hres = S_OK;
4455 } else {
4456 *pbstrProperty = SysAllocStringLen(NULL, 0);
4457 hres = S_FALSE;
4460 if(!(*pbstrProperty))
4461 hres = E_OUTOFMEMORY;
4463 break;
4464 case Uri_PROPERTY_PATH_AND_QUERY:
4465 if(This->path_start > -1) {
4466 *pbstrProperty = SysAllocStringLen(This->canon_uri+This->path_start, This->path_len+This->query_len);
4467 hres = S_OK;
4468 } else if(This->query_start > -1) {
4469 *pbstrProperty = SysAllocStringLen(This->canon_uri+This->query_start, This->query_len);
4470 hres = S_OK;
4471 } else {
4472 *pbstrProperty = SysAllocStringLen(NULL, 0);
4473 hres = S_FALSE;
4476 if(!(*pbstrProperty))
4477 hres = E_OUTOFMEMORY;
4479 break;
4480 case Uri_PROPERTY_QUERY:
4481 if(This->query_start > -1) {
4482 *pbstrProperty = SysAllocStringLen(This->canon_uri+This->query_start, This->query_len);
4483 hres = S_OK;
4484 } else {
4485 *pbstrProperty = SysAllocStringLen(NULL, 0);
4486 hres = S_FALSE;
4489 if(!(*pbstrProperty))
4490 hres = E_OUTOFMEMORY;
4492 break;
4493 case Uri_PROPERTY_RAW_URI:
4494 *pbstrProperty = SysAllocString(This->raw_uri);
4495 if(!(*pbstrProperty))
4496 hres = E_OUTOFMEMORY;
4497 else
4498 hres = S_OK;
4499 break;
4500 case Uri_PROPERTY_SCHEME_NAME:
4501 if(This->scheme_start > -1) {
4502 *pbstrProperty = SysAllocStringLen(This->canon_uri + This->scheme_start, This->scheme_len);
4503 hres = S_OK;
4504 } else {
4505 *pbstrProperty = SysAllocStringLen(NULL, 0);
4506 hres = S_FALSE;
4509 if(!(*pbstrProperty))
4510 hres = E_OUTOFMEMORY;
4512 break;
4513 case Uri_PROPERTY_USER_INFO:
4514 if(This->userinfo_start > -1) {
4515 *pbstrProperty = SysAllocStringLen(This->canon_uri+This->userinfo_start, This->userinfo_len);
4516 hres = S_OK;
4517 } else {
4518 *pbstrProperty = SysAllocStringLen(NULL, 0);
4519 hres = S_FALSE;
4522 if(!(*pbstrProperty))
4523 hres = E_OUTOFMEMORY;
4525 break;
4526 case Uri_PROPERTY_USER_NAME:
4527 if(This->userinfo_start > -1 && This->userinfo_split != 0) {
4528 /* If userinfo_split is set, that means a password exists
4529 * so the username is only from userinfo_start to userinfo_split.
4531 if(This->userinfo_split > -1) {
4532 *pbstrProperty = SysAllocStringLen(This->canon_uri + This->userinfo_start, This->userinfo_split);
4533 hres = S_OK;
4534 } else {
4535 *pbstrProperty = SysAllocStringLen(This->canon_uri + This->userinfo_start, This->userinfo_len);
4536 hres = S_OK;
4538 } else {
4539 *pbstrProperty = SysAllocStringLen(NULL, 0);
4540 hres = S_FALSE;
4543 if(!(*pbstrProperty))
4544 return E_OUTOFMEMORY;
4546 break;
4547 default:
4548 FIXME("(%p)->(%d %p %x)\n", This, uriProp, pbstrProperty, dwFlags);
4549 hres = E_NOTIMPL;
4552 return hres;
4555 static HRESULT WINAPI Uri_GetPropertyLength(IUri *iface, Uri_PROPERTY uriProp, DWORD *pcchProperty, DWORD dwFlags)
4557 Uri *This = impl_from_IUri(iface);
4558 HRESULT hres;
4559 TRACE("(%p %s)->(%d %p %x)\n", This, debugstr_w(This->canon_uri), uriProp, pcchProperty, dwFlags);
4561 if(!This->create_flags)
4562 return E_UNEXPECTED;
4563 if(!pcchProperty)
4564 return E_INVALIDARG;
4566 /* Can only return a length for a property if it's a string. */
4567 if(uriProp > Uri_PROPERTY_STRING_LAST)
4568 return E_INVALIDARG;
4570 /* Don't have support for flags yet. */
4571 if(dwFlags) {
4572 FIXME("(%p)->(%d %p %x)\n", This, uriProp, pcchProperty, dwFlags);
4573 return E_NOTIMPL;
4576 switch(uriProp) {
4577 case Uri_PROPERTY_ABSOLUTE_URI:
4578 if(This->display_modifiers & URI_DISPLAY_NO_ABSOLUTE_URI) {
4579 *pcchProperty = 0;
4580 hres = S_FALSE;
4581 } else {
4582 if(This->scheme_type != URL_SCHEME_UNKNOWN) {
4583 if(This->userinfo_start > -1 && This->userinfo_len == 0)
4584 /* Don't include the '@' in the length. */
4585 *pcchProperty = This->canon_len-1;
4586 else if(This->userinfo_start > -1 && This->userinfo_len == 1 &&
4587 This->userinfo_split == 0)
4588 /* Don't include the ":@" in the length. */
4589 *pcchProperty = This->canon_len-2;
4590 else
4591 *pcchProperty = This->canon_len;
4592 } else
4593 *pcchProperty = This->canon_len;
4595 hres = S_OK;
4598 break;
4599 case Uri_PROPERTY_AUTHORITY:
4600 if(This->port_offset > -1 &&
4601 This->display_modifiers & URI_DISPLAY_NO_DEFAULT_PORT_AUTH &&
4602 is_default_port(This->scheme_type, This->port))
4603 /* Only count up until the port in the authority. */
4604 *pcchProperty = This->port_offset;
4605 else
4606 *pcchProperty = This->authority_len;
4607 hres = (This->authority_start > -1) ? S_OK : S_FALSE;
4608 break;
4609 case Uri_PROPERTY_DISPLAY_URI:
4610 if(This->scheme_type != URL_SCHEME_UNKNOWN && This->userinfo_start > -1)
4611 *pcchProperty = This->canon_len-This->userinfo_len-1;
4612 else
4613 *pcchProperty = This->canon_len;
4615 hres = S_OK;
4616 break;
4617 case Uri_PROPERTY_DOMAIN:
4618 if(This->domain_offset > -1)
4619 *pcchProperty = This->host_len - This->domain_offset;
4620 else
4621 *pcchProperty = 0;
4623 hres = (This->domain_offset > -1) ? S_OK : S_FALSE;
4624 break;
4625 case Uri_PROPERTY_EXTENSION:
4626 if(This->extension_offset > -1) {
4627 *pcchProperty = This->path_len - This->extension_offset;
4628 hres = S_OK;
4629 } else {
4630 *pcchProperty = 0;
4631 hres = S_FALSE;
4634 break;
4635 case Uri_PROPERTY_FRAGMENT:
4636 *pcchProperty = This->fragment_len;
4637 hres = (This->fragment_start > -1) ? S_OK : S_FALSE;
4638 break;
4639 case Uri_PROPERTY_HOST:
4640 *pcchProperty = This->host_len;
4642 /* '[' and ']' aren't included in the length. */
4643 if(This->host_type == Uri_HOST_IPV6)
4644 *pcchProperty -= 2;
4646 hres = (This->host_start > -1) ? S_OK : S_FALSE;
4647 break;
4648 case Uri_PROPERTY_PASSWORD:
4649 *pcchProperty = (This->userinfo_split > -1) ? This->userinfo_len-This->userinfo_split-1 : 0;
4650 hres = (This->userinfo_split > -1) ? S_OK : S_FALSE;
4651 break;
4652 case Uri_PROPERTY_PATH:
4653 *pcchProperty = This->path_len;
4654 hres = (This->path_start > -1) ? S_OK : S_FALSE;
4655 break;
4656 case Uri_PROPERTY_PATH_AND_QUERY:
4657 *pcchProperty = This->path_len+This->query_len;
4658 hres = (This->path_start > -1 || This->query_start > -1) ? S_OK : S_FALSE;
4659 break;
4660 case Uri_PROPERTY_QUERY:
4661 *pcchProperty = This->query_len;
4662 hres = (This->query_start > -1) ? S_OK : S_FALSE;
4663 break;
4664 case Uri_PROPERTY_RAW_URI:
4665 *pcchProperty = SysStringLen(This->raw_uri);
4666 hres = S_OK;
4667 break;
4668 case Uri_PROPERTY_SCHEME_NAME:
4669 *pcchProperty = This->scheme_len;
4670 hres = (This->scheme_start > -1) ? S_OK : S_FALSE;
4671 break;
4672 case Uri_PROPERTY_USER_INFO:
4673 *pcchProperty = This->userinfo_len;
4674 hres = (This->userinfo_start > -1) ? S_OK : S_FALSE;
4675 break;
4676 case Uri_PROPERTY_USER_NAME:
4677 *pcchProperty = (This->userinfo_split > -1) ? This->userinfo_split : This->userinfo_len;
4678 if(This->userinfo_split == 0)
4679 hres = S_FALSE;
4680 else
4681 hres = (This->userinfo_start > -1) ? S_OK : S_FALSE;
4682 break;
4683 default:
4684 FIXME("(%p)->(%d %p %x)\n", This, uriProp, pcchProperty, dwFlags);
4685 hres = E_NOTIMPL;
4688 return hres;
4691 static HRESULT WINAPI Uri_GetPropertyDWORD(IUri *iface, Uri_PROPERTY uriProp, DWORD *pcchProperty, DWORD dwFlags)
4693 Uri *This = impl_from_IUri(iface);
4694 HRESULT hres;
4696 TRACE("(%p %s)->(%d %p %x)\n", This, debugstr_w(This->canon_uri), uriProp, pcchProperty, dwFlags);
4698 if(!This->create_flags)
4699 return E_UNEXPECTED;
4700 if(!pcchProperty)
4701 return E_INVALIDARG;
4703 /* Microsoft's implementation for the ZONE property of a URI seems to be lacking...
4704 * From what I can tell, instead of checking which URLZONE the URI belongs to it
4705 * simply assigns URLZONE_INVALID and returns E_NOTIMPL. This also applies to the GetZone
4706 * function.
4708 if(uriProp == Uri_PROPERTY_ZONE) {
4709 *pcchProperty = URLZONE_INVALID;
4710 return E_NOTIMPL;
4713 if(uriProp < Uri_PROPERTY_DWORD_START) {
4714 *pcchProperty = 0;
4715 return E_INVALIDARG;
4718 switch(uriProp) {
4719 case Uri_PROPERTY_HOST_TYPE:
4720 *pcchProperty = This->host_type;
4721 hres = S_OK;
4722 break;
4723 case Uri_PROPERTY_PORT:
4724 if(!This->has_port) {
4725 *pcchProperty = 0;
4726 hres = S_FALSE;
4727 } else {
4728 *pcchProperty = This->port;
4729 hres = S_OK;
4732 break;
4733 case Uri_PROPERTY_SCHEME:
4734 *pcchProperty = This->scheme_type;
4735 hres = S_OK;
4736 break;
4737 default:
4738 FIXME("(%p)->(%d %p %x)\n", This, uriProp, pcchProperty, dwFlags);
4739 hres = E_NOTIMPL;
4742 return hres;
4745 static HRESULT WINAPI Uri_HasProperty(IUri *iface, Uri_PROPERTY uriProp, BOOL *pfHasProperty)
4747 Uri *This = impl_from_IUri(iface);
4749 TRACE("(%p %s)->(%d %p)\n", This, debugstr_w(This->canon_uri), uriProp, pfHasProperty);
4751 if(!pfHasProperty)
4752 return E_INVALIDARG;
4754 switch(uriProp) {
4755 case Uri_PROPERTY_ABSOLUTE_URI:
4756 *pfHasProperty = !(This->display_modifiers & URI_DISPLAY_NO_ABSOLUTE_URI);
4757 break;
4758 case Uri_PROPERTY_AUTHORITY:
4759 *pfHasProperty = This->authority_start > -1;
4760 break;
4761 case Uri_PROPERTY_DISPLAY_URI:
4762 *pfHasProperty = TRUE;
4763 break;
4764 case Uri_PROPERTY_DOMAIN:
4765 *pfHasProperty = This->domain_offset > -1;
4766 break;
4767 case Uri_PROPERTY_EXTENSION:
4768 *pfHasProperty = This->extension_offset > -1;
4769 break;
4770 case Uri_PROPERTY_FRAGMENT:
4771 *pfHasProperty = This->fragment_start > -1;
4772 break;
4773 case Uri_PROPERTY_HOST:
4774 *pfHasProperty = This->host_start > -1;
4775 break;
4776 case Uri_PROPERTY_PASSWORD:
4777 *pfHasProperty = This->userinfo_split > -1;
4778 break;
4779 case Uri_PROPERTY_PATH:
4780 *pfHasProperty = This->path_start > -1;
4781 break;
4782 case Uri_PROPERTY_PATH_AND_QUERY:
4783 *pfHasProperty = (This->path_start > -1 || This->query_start > -1);
4784 break;
4785 case Uri_PROPERTY_QUERY:
4786 *pfHasProperty = This->query_start > -1;
4787 break;
4788 case Uri_PROPERTY_RAW_URI:
4789 *pfHasProperty = TRUE;
4790 break;
4791 case Uri_PROPERTY_SCHEME_NAME:
4792 *pfHasProperty = This->scheme_start > -1;
4793 break;
4794 case Uri_PROPERTY_USER_INFO:
4795 *pfHasProperty = This->userinfo_start > -1;
4796 break;
4797 case Uri_PROPERTY_USER_NAME:
4798 if(This->userinfo_split == 0)
4799 *pfHasProperty = FALSE;
4800 else
4801 *pfHasProperty = This->userinfo_start > -1;
4802 break;
4803 case Uri_PROPERTY_HOST_TYPE:
4804 *pfHasProperty = TRUE;
4805 break;
4806 case Uri_PROPERTY_PORT:
4807 *pfHasProperty = This->has_port;
4808 break;
4809 case Uri_PROPERTY_SCHEME:
4810 *pfHasProperty = TRUE;
4811 break;
4812 case Uri_PROPERTY_ZONE:
4813 *pfHasProperty = FALSE;
4814 break;
4815 default:
4816 FIXME("(%p)->(%d %p): Unsupported property type.\n", This, uriProp, pfHasProperty);
4817 return E_NOTIMPL;
4820 return S_OK;
4823 static HRESULT WINAPI Uri_GetAbsoluteUri(IUri *iface, BSTR *pstrAbsoluteUri)
4825 TRACE("(%p)->(%p)\n", iface, pstrAbsoluteUri);
4826 return IUri_GetPropertyBSTR(iface, Uri_PROPERTY_ABSOLUTE_URI, pstrAbsoluteUri, 0);
4829 static HRESULT WINAPI Uri_GetAuthority(IUri *iface, BSTR *pstrAuthority)
4831 TRACE("(%p)->(%p)\n", iface, pstrAuthority);
4832 return IUri_GetPropertyBSTR(iface, Uri_PROPERTY_AUTHORITY, pstrAuthority, 0);
4835 static HRESULT WINAPI Uri_GetDisplayUri(IUri *iface, BSTR *pstrDisplayUri)
4837 TRACE("(%p)->(%p)\n", iface, pstrDisplayUri);
4838 return IUri_GetPropertyBSTR(iface, Uri_PROPERTY_DISPLAY_URI, pstrDisplayUri, 0);
4841 static HRESULT WINAPI Uri_GetDomain(IUri *iface, BSTR *pstrDomain)
4843 TRACE("(%p)->(%p)\n", iface, pstrDomain);
4844 return IUri_GetPropertyBSTR(iface, Uri_PROPERTY_DOMAIN, pstrDomain, 0);
4847 static HRESULT WINAPI Uri_GetExtension(IUri *iface, BSTR *pstrExtension)
4849 TRACE("(%p)->(%p)\n", iface, pstrExtension);
4850 return IUri_GetPropertyBSTR(iface, Uri_PROPERTY_EXTENSION, pstrExtension, 0);
4853 static HRESULT WINAPI Uri_GetFragment(IUri *iface, BSTR *pstrFragment)
4855 TRACE("(%p)->(%p)\n", iface, pstrFragment);
4856 return IUri_GetPropertyBSTR(iface, Uri_PROPERTY_FRAGMENT, pstrFragment, 0);
4859 static HRESULT WINAPI Uri_GetHost(IUri *iface, BSTR *pstrHost)
4861 TRACE("(%p)->(%p)\n", iface, pstrHost);
4862 return IUri_GetPropertyBSTR(iface, Uri_PROPERTY_HOST, pstrHost, 0);
4865 static HRESULT WINAPI Uri_GetPassword(IUri *iface, BSTR *pstrPassword)
4867 TRACE("(%p)->(%p)\n", iface, pstrPassword);
4868 return IUri_GetPropertyBSTR(iface, Uri_PROPERTY_PASSWORD, pstrPassword, 0);
4871 static HRESULT WINAPI Uri_GetPath(IUri *iface, BSTR *pstrPath)
4873 TRACE("(%p)->(%p)\n", iface, pstrPath);
4874 return IUri_GetPropertyBSTR(iface, Uri_PROPERTY_PATH, pstrPath, 0);
4877 static HRESULT WINAPI Uri_GetPathAndQuery(IUri *iface, BSTR *pstrPathAndQuery)
4879 TRACE("(%p)->(%p)\n", iface, pstrPathAndQuery);
4880 return IUri_GetPropertyBSTR(iface, Uri_PROPERTY_PATH_AND_QUERY, pstrPathAndQuery, 0);
4883 static HRESULT WINAPI Uri_GetQuery(IUri *iface, BSTR *pstrQuery)
4885 TRACE("(%p)->(%p)\n", iface, pstrQuery);
4886 return IUri_GetPropertyBSTR(iface, Uri_PROPERTY_QUERY, pstrQuery, 0);
4889 static HRESULT WINAPI Uri_GetRawUri(IUri *iface, BSTR *pstrRawUri)
4891 TRACE("(%p)->(%p)\n", iface, pstrRawUri);
4892 return IUri_GetPropertyBSTR(iface, Uri_PROPERTY_RAW_URI, pstrRawUri, 0);
4895 static HRESULT WINAPI Uri_GetSchemeName(IUri *iface, BSTR *pstrSchemeName)
4897 TRACE("(%p)->(%p)\n", iface, pstrSchemeName);
4898 return IUri_GetPropertyBSTR(iface, Uri_PROPERTY_SCHEME_NAME, pstrSchemeName, 0);
4901 static HRESULT WINAPI Uri_GetUserInfo(IUri *iface, BSTR *pstrUserInfo)
4903 TRACE("(%p)->(%p)\n", iface, pstrUserInfo);
4904 return IUri_GetPropertyBSTR(iface, Uri_PROPERTY_USER_INFO, pstrUserInfo, 0);
4907 static HRESULT WINAPI Uri_GetUserName(IUri *iface, BSTR *pstrUserName)
4909 TRACE("(%p)->(%p)\n", iface, pstrUserName);
4910 return IUri_GetPropertyBSTR(iface, Uri_PROPERTY_USER_NAME, pstrUserName, 0);
4913 static HRESULT WINAPI Uri_GetHostType(IUri *iface, DWORD *pdwHostType)
4915 TRACE("(%p)->(%p)\n", iface, pdwHostType);
4916 return IUri_GetPropertyDWORD(iface, Uri_PROPERTY_HOST_TYPE, pdwHostType, 0);
4919 static HRESULT WINAPI Uri_GetPort(IUri *iface, DWORD *pdwPort)
4921 TRACE("(%p)->(%p)\n", iface, pdwPort);
4922 return IUri_GetPropertyDWORD(iface, Uri_PROPERTY_PORT, pdwPort, 0);
4925 static HRESULT WINAPI Uri_GetScheme(IUri *iface, DWORD *pdwScheme)
4927 TRACE("(%p)->(%p)\n", iface, pdwScheme);
4928 return IUri_GetPropertyDWORD(iface, Uri_PROPERTY_SCHEME, pdwScheme, 0);
4931 static HRESULT WINAPI Uri_GetZone(IUri *iface, DWORD *pdwZone)
4933 TRACE("(%p)->(%p)\n", iface, pdwZone);
4934 return IUri_GetPropertyDWORD(iface, Uri_PROPERTY_ZONE,pdwZone, 0);
4937 static HRESULT WINAPI Uri_GetProperties(IUri *iface, DWORD *pdwProperties)
4939 Uri *This = impl_from_IUri(iface);
4940 TRACE("(%p %s)->(%p)\n", This, debugstr_w(This->canon_uri), pdwProperties);
4942 if(!This->create_flags)
4943 return E_UNEXPECTED;
4944 if(!pdwProperties)
4945 return E_INVALIDARG;
4947 /* All URIs have these. */
4948 *pdwProperties = Uri_HAS_DISPLAY_URI|Uri_HAS_RAW_URI|Uri_HAS_SCHEME|Uri_HAS_HOST_TYPE;
4950 if(!(This->display_modifiers & URI_DISPLAY_NO_ABSOLUTE_URI))
4951 *pdwProperties |= Uri_HAS_ABSOLUTE_URI;
4953 if(This->scheme_start > -1)
4954 *pdwProperties |= Uri_HAS_SCHEME_NAME;
4956 if(This->authority_start > -1) {
4957 *pdwProperties |= Uri_HAS_AUTHORITY;
4958 if(This->userinfo_start > -1) {
4959 *pdwProperties |= Uri_HAS_USER_INFO;
4960 if(This->userinfo_split != 0)
4961 *pdwProperties |= Uri_HAS_USER_NAME;
4963 if(This->userinfo_split > -1)
4964 *pdwProperties |= Uri_HAS_PASSWORD;
4965 if(This->host_start > -1)
4966 *pdwProperties |= Uri_HAS_HOST;
4967 if(This->domain_offset > -1)
4968 *pdwProperties |= Uri_HAS_DOMAIN;
4971 if(This->has_port)
4972 *pdwProperties |= Uri_HAS_PORT;
4973 if(This->path_start > -1)
4974 *pdwProperties |= Uri_HAS_PATH|Uri_HAS_PATH_AND_QUERY;
4975 if(This->query_start > -1)
4976 *pdwProperties |= Uri_HAS_QUERY|Uri_HAS_PATH_AND_QUERY;
4978 if(This->extension_offset > -1)
4979 *pdwProperties |= Uri_HAS_EXTENSION;
4981 if(This->fragment_start > -1)
4982 *pdwProperties |= Uri_HAS_FRAGMENT;
4984 return S_OK;
4987 static HRESULT WINAPI Uri_IsEqual(IUri *iface, IUri *pUri, BOOL *pfEqual)
4989 Uri *This = impl_from_IUri(iface);
4990 Uri *other;
4992 TRACE("(%p %s)->(%p %p)\n", This, debugstr_w(This->canon_uri), pUri, pfEqual);
4994 if(!This->create_flags)
4995 return E_UNEXPECTED;
4996 if(!pfEqual)
4997 return E_POINTER;
4999 if(!pUri) {
5000 *pfEqual = FALSE;
5002 /* For some reason Windows returns S_OK here... */
5003 return S_OK;
5006 /* Try to convert it to a Uri (allows for a more simple comparison). */
5007 if(!(other = get_uri_obj(pUri))) {
5008 FIXME("(%p)->(%p %p) No support for unknown IUri's yet.\n", iface, pUri, pfEqual);
5009 return E_NOTIMPL;
5012 TRACE("comparing to %s\n", debugstr_w(other->canon_uri));
5013 return compare_uris(This, other, pfEqual);
5016 static const IUriVtbl UriVtbl = {
5017 Uri_QueryInterface,
5018 Uri_AddRef,
5019 Uri_Release,
5020 Uri_GetPropertyBSTR,
5021 Uri_GetPropertyLength,
5022 Uri_GetPropertyDWORD,
5023 Uri_HasProperty,
5024 Uri_GetAbsoluteUri,
5025 Uri_GetAuthority,
5026 Uri_GetDisplayUri,
5027 Uri_GetDomain,
5028 Uri_GetExtension,
5029 Uri_GetFragment,
5030 Uri_GetHost,
5031 Uri_GetPassword,
5032 Uri_GetPath,
5033 Uri_GetPathAndQuery,
5034 Uri_GetQuery,
5035 Uri_GetRawUri,
5036 Uri_GetSchemeName,
5037 Uri_GetUserInfo,
5038 Uri_GetUserName,
5039 Uri_GetHostType,
5040 Uri_GetPort,
5041 Uri_GetScheme,
5042 Uri_GetZone,
5043 Uri_GetProperties,
5044 Uri_IsEqual
5047 static inline Uri* impl_from_IUriBuilderFactory(IUriBuilderFactory *iface)
5049 return CONTAINING_RECORD(iface, Uri, IUriBuilderFactory_iface);
5052 static HRESULT WINAPI UriBuilderFactory_QueryInterface(IUriBuilderFactory *iface, REFIID riid, void **ppv)
5054 Uri *This = impl_from_IUriBuilderFactory(iface);
5055 return IUri_QueryInterface(&This->IUri_iface, riid, ppv);
5058 static ULONG WINAPI UriBuilderFactory_AddRef(IUriBuilderFactory *iface)
5060 Uri *This = impl_from_IUriBuilderFactory(iface);
5061 return IUri_AddRef(&This->IUri_iface);
5064 static ULONG WINAPI UriBuilderFactory_Release(IUriBuilderFactory *iface)
5066 Uri *This = impl_from_IUriBuilderFactory(iface);
5067 return IUri_Release(&This->IUri_iface);
5070 static HRESULT WINAPI UriBuilderFactory_CreateIUriBuilder(IUriBuilderFactory *iface,
5071 DWORD dwFlags,
5072 DWORD_PTR dwReserved,
5073 IUriBuilder **ppIUriBuilder)
5075 Uri *This = impl_from_IUriBuilderFactory(iface);
5076 TRACE("(%p)->(%08x %08x %p)\n", This, dwFlags, (DWORD)dwReserved, ppIUriBuilder);
5078 if(!ppIUriBuilder)
5079 return E_POINTER;
5081 if(dwFlags || dwReserved) {
5082 *ppIUriBuilder = NULL;
5083 return E_INVALIDARG;
5086 return CreateIUriBuilder(NULL, 0, 0, ppIUriBuilder);
5089 static HRESULT WINAPI UriBuilderFactory_CreateInitializedIUriBuilder(IUriBuilderFactory *iface,
5090 DWORD dwFlags,
5091 DWORD_PTR dwReserved,
5092 IUriBuilder **ppIUriBuilder)
5094 Uri *This = impl_from_IUriBuilderFactory(iface);
5095 TRACE("(%p)->(%08x %08x %p)\n", This, dwFlags, (DWORD)dwReserved, ppIUriBuilder);
5097 if(!ppIUriBuilder)
5098 return E_POINTER;
5100 if(dwFlags || dwReserved) {
5101 *ppIUriBuilder = NULL;
5102 return E_INVALIDARG;
5105 return CreateIUriBuilder(&This->IUri_iface, 0, 0, ppIUriBuilder);
5108 static const IUriBuilderFactoryVtbl UriBuilderFactoryVtbl = {
5109 UriBuilderFactory_QueryInterface,
5110 UriBuilderFactory_AddRef,
5111 UriBuilderFactory_Release,
5112 UriBuilderFactory_CreateIUriBuilder,
5113 UriBuilderFactory_CreateInitializedIUriBuilder
5116 static inline Uri* impl_from_IPersistStream(IPersistStream *iface)
5118 return CONTAINING_RECORD(iface, Uri, IPersistStream_iface);
5121 static HRESULT WINAPI PersistStream_QueryInterface(IPersistStream *iface, REFIID riid, void **ppvObject)
5123 Uri *This = impl_from_IPersistStream(iface);
5124 return IUri_QueryInterface(&This->IUri_iface, riid, ppvObject);
5127 static ULONG WINAPI PersistStream_AddRef(IPersistStream *iface)
5129 Uri *This = impl_from_IPersistStream(iface);
5130 return IUri_AddRef(&This->IUri_iface);
5133 static ULONG WINAPI PersistStream_Release(IPersistStream *iface)
5135 Uri *This = impl_from_IPersistStream(iface);
5136 return IUri_Release(&This->IUri_iface);
5139 static HRESULT WINAPI PersistStream_GetClassID(IPersistStream *iface, CLSID *pClassID)
5141 Uri *This = impl_from_IPersistStream(iface);
5142 TRACE("(%p)->(%p)\n", This, pClassID);
5144 if(!pClassID)
5145 return E_INVALIDARG;
5147 *pClassID = CLSID_CUri;
5148 return S_OK;
5151 static HRESULT WINAPI PersistStream_IsDirty(IPersistStream *iface)
5153 Uri *This = impl_from_IPersistStream(iface);
5154 TRACE("(%p)\n", This);
5155 return S_FALSE;
5158 struct persist_uri {
5159 DWORD size;
5160 DWORD unk1[2];
5161 DWORD create_flags;
5162 DWORD unk2[3];
5163 DWORD fields_no;
5164 BYTE data[1];
5167 static HRESULT WINAPI PersistStream_Load(IPersistStream *iface, IStream *pStm)
5169 Uri *This = impl_from_IPersistStream(iface);
5170 struct persist_uri *data;
5171 parse_data parse;
5172 DWORD size;
5173 HRESULT hr;
5175 TRACE("(%p)->(%p)\n", This, pStm);
5177 if(This->create_flags)
5178 return E_UNEXPECTED;
5179 if(!pStm)
5180 return E_INVALIDARG;
5182 hr = IStream_Read(pStm, &size, sizeof(DWORD), NULL);
5183 if(FAILED(hr))
5184 return hr;
5185 data = heap_alloc(size);
5186 if(!data)
5187 return E_OUTOFMEMORY;
5188 hr = IStream_Read(pStm, data->unk1, size-sizeof(DWORD)-2, NULL);
5189 if(FAILED(hr)) {
5190 heap_free(data);
5191 return hr;
5194 if(size < sizeof(struct persist_uri)) {
5195 heap_free(data);
5196 return S_OK;
5199 if(*(DWORD*)data->data != Uri_PROPERTY_RAW_URI) {
5200 heap_free(data);
5201 ERR("Can't find raw_uri\n");
5202 return E_UNEXPECTED;
5205 This->raw_uri = SysAllocString((WCHAR*)(data->data+sizeof(DWORD)*2));
5206 if(!This->raw_uri) {
5207 heap_free(data);
5208 return E_OUTOFMEMORY;
5210 This->create_flags = data->create_flags;
5211 heap_free(data);
5212 TRACE("%x %s\n", This->create_flags, debugstr_w(This->raw_uri));
5214 memset(&parse, 0, sizeof(parse_data));
5215 parse.uri = This->raw_uri;
5216 if(!parse_uri(&parse, This->create_flags)) {
5217 SysFreeString(This->raw_uri);
5218 This->create_flags = 0;
5219 return E_UNEXPECTED;
5222 hr = canonicalize_uri(&parse, This, This->create_flags);
5223 if(FAILED(hr)) {
5224 SysFreeString(This->raw_uri);
5225 This->create_flags = 0;
5226 return hr;
5229 return S_OK;
5232 static inline BYTE* persist_stream_add_strprop(Uri *This, BYTE *p, DWORD type, DWORD len, WCHAR *data)
5234 len *= sizeof(WCHAR);
5235 *(DWORD*)p = type;
5236 p += sizeof(DWORD);
5237 *(DWORD*)p = len+sizeof(WCHAR);
5238 p += sizeof(DWORD);
5239 memcpy(p, data, len);
5240 p += len;
5241 *(WCHAR*)p = 0;
5242 return p+sizeof(WCHAR);
5245 static inline void persist_stream_save(Uri *This, IStream *pStm, BOOL marshal, struct persist_uri *data)
5247 BYTE *p = NULL;
5249 data->create_flags = This->create_flags;
5251 if(This->create_flags) {
5252 data->fields_no = 1;
5253 p = persist_stream_add_strprop(This, data->data, Uri_PROPERTY_RAW_URI,
5254 SysStringLen(This->raw_uri), This->raw_uri);
5256 if(This->scheme_type!=URL_SCHEME_HTTP && This->scheme_type!=URL_SCHEME_HTTPS
5257 && This->scheme_type!=URL_SCHEME_FTP)
5258 return;
5260 if(This->fragment_len) {
5261 data->fields_no++;
5262 p = persist_stream_add_strprop(This, p, Uri_PROPERTY_FRAGMENT,
5263 This->fragment_len, This->canon_uri+This->fragment_start);
5266 if(This->host_len) {
5267 data->fields_no++;
5268 if(This->host_type == Uri_HOST_IPV6)
5269 p = persist_stream_add_strprop(This, p, Uri_PROPERTY_HOST,
5270 This->host_len-2, This->canon_uri+This->host_start+1);
5271 else
5272 p = persist_stream_add_strprop(This, p, Uri_PROPERTY_HOST,
5273 This->host_len, This->canon_uri+This->host_start);
5276 if(This->userinfo_split > -1) {
5277 data->fields_no++;
5278 p = persist_stream_add_strprop(This, p, Uri_PROPERTY_PASSWORD,
5279 This->userinfo_len-This->userinfo_split-1,
5280 This->canon_uri+This->userinfo_start+This->userinfo_split+1);
5283 if(This->path_len) {
5284 data->fields_no++;
5285 p = persist_stream_add_strprop(This, p, Uri_PROPERTY_PATH,
5286 This->path_len, This->canon_uri+This->path_start);
5287 } else if(marshal) {
5288 WCHAR no_path = '/';
5289 data->fields_no++;
5290 p = persist_stream_add_strprop(This, p, Uri_PROPERTY_PATH, 1, &no_path);
5293 if(This->has_port) {
5294 data->fields_no++;
5295 *(DWORD*)p = Uri_PROPERTY_PORT;
5296 p += sizeof(DWORD);
5297 *(DWORD*)p = sizeof(DWORD);
5298 p += sizeof(DWORD);
5299 *(DWORD*)p = This->port;
5300 p += sizeof(DWORD);
5303 if(This->query_len) {
5304 data->fields_no++;
5305 p = persist_stream_add_strprop(This, p, Uri_PROPERTY_QUERY,
5306 This->query_len, This->canon_uri+This->query_start);
5309 if(This->scheme_len) {
5310 data->fields_no++;
5311 p = persist_stream_add_strprop(This, p, Uri_PROPERTY_SCHEME_NAME,
5312 This->scheme_len, This->canon_uri+This->scheme_start);
5315 if(This->userinfo_start>-1 && This->userinfo_split!=0) {
5316 data->fields_no++;
5317 if(This->userinfo_split > -1)
5318 p = persist_stream_add_strprop(This, p, Uri_PROPERTY_USER_NAME,
5319 This->userinfo_split, This->canon_uri+This->userinfo_start);
5320 else
5321 p = persist_stream_add_strprop(This, p, Uri_PROPERTY_USER_NAME,
5322 This->userinfo_len, This->canon_uri+This->userinfo_start);
5326 static HRESULT WINAPI PersistStream_Save(IPersistStream *iface, IStream *pStm, BOOL fClearDirty)
5328 Uri *This = impl_from_IPersistStream(iface);
5329 struct persist_uri *data;
5330 ULARGE_INTEGER size;
5331 HRESULT hres;
5333 TRACE("(%p)->(%p %x)\n", This, pStm, fClearDirty);
5335 if(!pStm)
5336 return E_INVALIDARG;
5338 hres = IPersistStream_GetSizeMax(&This->IPersistStream_iface, &size);
5339 if(FAILED(hres))
5340 return hres;
5342 data = heap_alloc_zero(size.u.LowPart);
5343 if(!data)
5344 return E_OUTOFMEMORY;
5345 data->size = size.u.LowPart;
5346 persist_stream_save(This, pStm, FALSE, data);
5348 hres = IStream_Write(pStm, data, data->size-2, NULL);
5349 heap_free(data);
5350 return hres;
5353 static HRESULT WINAPI PersistStream_GetSizeMax(IPersistStream *iface, ULARGE_INTEGER *pcbSize)
5355 Uri *This = impl_from_IPersistStream(iface);
5356 TRACE("(%p)->(%p)\n", This, pcbSize);
5358 if(!pcbSize)
5359 return E_INVALIDARG;
5361 pcbSize->u.LowPart = 2+sizeof(struct persist_uri);
5362 pcbSize->u.HighPart = 0;
5363 if(This->create_flags)
5364 pcbSize->u.LowPart += (SysStringLen(This->raw_uri)+1)*sizeof(WCHAR) + 2*sizeof(DWORD);
5365 else /* there's no place for fields no */
5366 pcbSize->u.LowPart -= sizeof(DWORD);
5367 if(This->scheme_type!=URL_SCHEME_HTTP && This->scheme_type!=URL_SCHEME_HTTPS
5368 && This->scheme_type!=URL_SCHEME_FTP)
5369 return S_OK;
5371 if(This->fragment_len)
5372 pcbSize->u.LowPart += (This->fragment_len+1)*sizeof(WCHAR) + 2*sizeof(DWORD);
5373 if(This->host_len) {
5374 if(This->host_type == Uri_HOST_IPV6)
5375 pcbSize->u.LowPart += (This->host_len-1)*sizeof(WCHAR) + 2*sizeof(DWORD);
5376 else
5377 pcbSize->u.LowPart += (This->host_len+1)*sizeof(WCHAR) + 2*sizeof(DWORD);
5379 if(This->userinfo_split > -1)
5380 pcbSize->u.LowPart += (This->userinfo_len-This->userinfo_split)*sizeof(WCHAR) + 2*sizeof(DWORD);
5381 if(This->path_len)
5382 pcbSize->u.LowPart += (This->path_len+1)*sizeof(WCHAR) + 2*sizeof(DWORD);
5383 if(This->has_port)
5384 pcbSize->u.LowPart += 3*sizeof(DWORD);
5385 if(This->query_len)
5386 pcbSize->u.LowPart += (This->query_len+1)*sizeof(WCHAR) + 2*sizeof(DWORD);
5387 if(This->scheme_len)
5388 pcbSize->u.LowPart += (This->scheme_len+1)*sizeof(WCHAR) + 2*sizeof(DWORD);
5389 if(This->userinfo_start>-1 && This->userinfo_split!=0) {
5390 if(This->userinfo_split > -1)
5391 pcbSize->u.LowPart += (This->userinfo_split+1)*sizeof(WCHAR) + 2*sizeof(DWORD);
5392 else
5393 pcbSize->u.LowPart += (This->userinfo_len+1)*sizeof(WCHAR) + 2*sizeof(DWORD);
5395 return S_OK;
5398 static const IPersistStreamVtbl PersistStreamVtbl = {
5399 PersistStream_QueryInterface,
5400 PersistStream_AddRef,
5401 PersistStream_Release,
5402 PersistStream_GetClassID,
5403 PersistStream_IsDirty,
5404 PersistStream_Load,
5405 PersistStream_Save,
5406 PersistStream_GetSizeMax
5409 static inline Uri* impl_from_IMarshal(IMarshal *iface)
5411 return CONTAINING_RECORD(iface, Uri, IMarshal_iface);
5414 static HRESULT WINAPI Marshal_QueryInterface(IMarshal *iface, REFIID riid, void **ppvObject)
5416 Uri *This = impl_from_IMarshal(iface);
5417 return IUri_QueryInterface(&This->IUri_iface, riid, ppvObject);
5420 static ULONG WINAPI Marshal_AddRef(IMarshal *iface)
5422 Uri *This = impl_from_IMarshal(iface);
5423 return IUri_AddRef(&This->IUri_iface);
5426 static ULONG WINAPI Marshal_Release(IMarshal *iface)
5428 Uri *This = impl_from_IMarshal(iface);
5429 return IUri_Release(&This->IUri_iface);
5432 static HRESULT WINAPI Marshal_GetUnmarshalClass(IMarshal *iface, REFIID riid, void *pv,
5433 DWORD dwDestContext, void *pvDestContext, DWORD mshlflags, CLSID *pCid)
5435 Uri *This = impl_from_IMarshal(iface);
5436 TRACE("(%p)->(%s %p %x %p %x %p)\n", This, debugstr_guid(riid), pv,
5437 dwDestContext, pvDestContext, mshlflags, pCid);
5439 if(!pCid || (dwDestContext!=MSHCTX_LOCAL && dwDestContext!=MSHCTX_NOSHAREDMEM
5440 && dwDestContext!=MSHCTX_INPROC))
5441 return E_INVALIDARG;
5443 *pCid = CLSID_CUri;
5444 return S_OK;
5447 struct inproc_marshal_uri {
5448 DWORD size;
5449 DWORD mshlflags;
5450 DWORD unk[4]; /* process identifier? */
5451 Uri *uri;
5454 static HRESULT WINAPI Marshal_GetMarshalSizeMax(IMarshal *iface, REFIID riid, void *pv,
5455 DWORD dwDestContext, void *pvDestContext, DWORD mshlflags, DWORD *pSize)
5457 Uri *This = impl_from_IMarshal(iface);
5458 ULARGE_INTEGER size;
5459 HRESULT hres;
5460 TRACE("(%p)->(%s %p %x %p %x %p)\n", This, debugstr_guid(riid), pv,
5461 dwDestContext, pvDestContext, mshlflags, pSize);
5463 if(!pSize || (dwDestContext!=MSHCTX_LOCAL && dwDestContext!=MSHCTX_NOSHAREDMEM
5464 && dwDestContext!=MSHCTX_INPROC))
5465 return E_INVALIDARG;
5467 if(dwDestContext == MSHCTX_INPROC) {
5468 *pSize = sizeof(struct inproc_marshal_uri);
5469 return S_OK;
5472 hres = IPersistStream_GetSizeMax(&This->IPersistStream_iface, &size);
5473 if(FAILED(hres))
5474 return hres;
5475 if(!This->path_len && (This->scheme_type==URL_SCHEME_HTTP
5476 || This->scheme_type==URL_SCHEME_HTTPS
5477 || This->scheme_type==URL_SCHEME_FTP))
5478 size.u.LowPart += 3*sizeof(DWORD);
5479 *pSize = size.u.LowPart+2*sizeof(DWORD);
5480 return S_OK;
5483 static HRESULT WINAPI Marshal_MarshalInterface(IMarshal *iface, IStream *pStm, REFIID riid,
5484 void *pv, DWORD dwDestContext, void *pvDestContext, DWORD mshlflags)
5486 Uri *This = impl_from_IMarshal(iface);
5487 DWORD *data;
5488 DWORD size;
5489 HRESULT hres;
5491 TRACE("(%p)->(%p %s %p %x %p %x)\n", This, pStm, debugstr_guid(riid), pv,
5492 dwDestContext, pvDestContext, mshlflags);
5494 if(!pStm || mshlflags!=MSHLFLAGS_NORMAL || (dwDestContext!=MSHCTX_LOCAL
5495 && dwDestContext!=MSHCTX_NOSHAREDMEM && dwDestContext!=MSHCTX_INPROC))
5496 return E_INVALIDARG;
5498 if(dwDestContext == MSHCTX_INPROC) {
5499 struct inproc_marshal_uri data;
5501 data.size = sizeof(data);
5502 data.mshlflags = MSHCTX_INPROC;
5503 data.unk[0] = 0;
5504 data.unk[1] = 0;
5505 data.unk[2] = 0;
5506 data.unk[3] = 0;
5507 data.uri = This;
5509 hres = IStream_Write(pStm, &data, data.size, NULL);
5510 if(FAILED(hres))
5511 return hres;
5513 IUri_AddRef(&This->IUri_iface);
5514 return S_OK;
5517 hres = IMarshal_GetMarshalSizeMax(iface, riid, pv, dwDestContext,
5518 pvDestContext, mshlflags, &size);
5519 if(FAILED(hres))
5520 return hres;
5522 data = heap_alloc_zero(size);
5523 if(!data)
5524 return E_OUTOFMEMORY;
5526 data[0] = size;
5527 data[1] = dwDestContext;
5528 data[2] = size-2*sizeof(DWORD);
5529 persist_stream_save(This, pStm, TRUE, (struct persist_uri*)(data+2));
5531 hres = IStream_Write(pStm, data, data[0]-2, NULL);
5532 heap_free(data);
5533 return hres;
5536 static HRESULT WINAPI Marshal_UnmarshalInterface(IMarshal *iface,
5537 IStream *pStm, REFIID riid, void **ppv)
5539 Uri *This = impl_from_IMarshal(iface);
5540 DWORD header[2];
5541 HRESULT hres;
5543 TRACE("(%p)->(%p %s %p)\n", This, pStm, debugstr_guid(riid), ppv);
5545 if(This->create_flags)
5546 return E_UNEXPECTED;
5547 if(!pStm || !riid || !ppv)
5548 return E_INVALIDARG;
5550 hres = IStream_Read(pStm, header, sizeof(header), NULL);
5551 if(FAILED(hres))
5552 return hres;
5554 if(header[1]!=MSHCTX_LOCAL && header[1]!=MSHCTX_NOSHAREDMEM
5555 && header[1]!=MSHCTX_INPROC)
5556 return E_UNEXPECTED;
5558 if(header[1] == MSHCTX_INPROC) {
5559 struct inproc_marshal_uri data;
5560 parse_data parse;
5562 hres = IStream_Read(pStm, data.unk, sizeof(data)-2*sizeof(DWORD), NULL);
5563 if(FAILED(hres))
5564 return hres;
5566 This->raw_uri = SysAllocString(data.uri->raw_uri);
5567 if(!This->raw_uri) {
5568 return E_OUTOFMEMORY;
5571 memset(&parse, 0, sizeof(parse_data));
5572 parse.uri = This->raw_uri;
5574 if(!parse_uri(&parse, data.uri->create_flags))
5575 return E_INVALIDARG;
5577 hres = canonicalize_uri(&parse, This, data.uri->create_flags);
5578 if(FAILED(hres))
5579 return hres;
5581 This->create_flags = data.uri->create_flags;
5582 IUri_Release(&data.uri->IUri_iface);
5584 return IUri_QueryInterface(&This->IUri_iface, riid, ppv);
5587 hres = IPersistStream_Load(&This->IPersistStream_iface, pStm);
5588 if(FAILED(hres))
5589 return hres;
5591 return IUri_QueryInterface(&This->IUri_iface, riid, ppv);
5594 static HRESULT WINAPI Marshal_ReleaseMarshalData(IMarshal *iface, IStream *pStm)
5596 Uri *This = impl_from_IMarshal(iface);
5597 LARGE_INTEGER off;
5598 DWORD header[2];
5599 HRESULT hres;
5601 TRACE("(%p)->(%p)\n", This, pStm);
5603 if(!pStm)
5604 return E_INVALIDARG;
5606 hres = IStream_Read(pStm, header, 2*sizeof(DWORD), NULL);
5607 if(FAILED(hres))
5608 return hres;
5610 if(header[1] == MSHCTX_INPROC) {
5611 struct inproc_marshal_uri data;
5613 hres = IStream_Read(pStm, data.unk, sizeof(data)-2*sizeof(DWORD), NULL);
5614 if(FAILED(hres))
5615 return hres;
5617 IUri_Release(&data.uri->IUri_iface);
5618 return S_OK;
5621 off.u.LowPart = header[0]-sizeof(header)-2;
5622 off.u.HighPart = 0;
5623 return IStream_Seek(pStm, off, STREAM_SEEK_CUR, NULL);
5626 static HRESULT WINAPI Marshal_DisconnectObject(IMarshal *iface, DWORD dwReserved)
5628 Uri *This = impl_from_IMarshal(iface);
5629 TRACE("(%p)->(%x)\n", This, dwReserved);
5630 return S_OK;
5633 static const IMarshalVtbl MarshalVtbl = {
5634 Marshal_QueryInterface,
5635 Marshal_AddRef,
5636 Marshal_Release,
5637 Marshal_GetUnmarshalClass,
5638 Marshal_GetMarshalSizeMax,
5639 Marshal_MarshalInterface,
5640 Marshal_UnmarshalInterface,
5641 Marshal_ReleaseMarshalData,
5642 Marshal_DisconnectObject
5645 HRESULT Uri_Construct(IUnknown *pUnkOuter, LPVOID *ppobj)
5647 Uri *ret = heap_alloc_zero(sizeof(Uri));
5649 TRACE("(%p %p)\n", pUnkOuter, ppobj);
5651 *ppobj = ret;
5652 if(!ret)
5653 return E_OUTOFMEMORY;
5655 ret->IUri_iface.lpVtbl = &UriVtbl;
5656 ret->IUriBuilderFactory_iface.lpVtbl = &UriBuilderFactoryVtbl;
5657 ret->IPersistStream_iface.lpVtbl = &PersistStreamVtbl;
5658 ret->IMarshal_iface.lpVtbl = &MarshalVtbl;
5659 ret->ref = 1;
5661 *ppobj = &ret->IUri_iface;
5662 return S_OK;
5665 /***********************************************************************
5666 * CreateUri (urlmon.@)
5668 * Creates a new IUri object using the URI represented by pwzURI. This function
5669 * parses and validates the components of pwzURI and then canonicalizes the
5670 * parsed components.
5672 * PARAMS
5673 * pwzURI [I] The URI to parse, validate, and canonicalize.
5674 * dwFlags [I] Flags which can affect how the parsing/canonicalization is performed.
5675 * dwReserved [I] Reserved (not used).
5676 * ppURI [O] The resulting IUri after parsing/canonicalization occurs.
5678 * RETURNS
5679 * Success: Returns S_OK. ppURI contains the pointer to the newly allocated IUri.
5680 * Failure: E_INVALIDARG if there are invalid flag combinations in dwFlags, or an
5681 * invalid parameter, or pwzURI doesn't represent a valid URI.
5682 * E_OUTOFMEMORY if any memory allocation fails.
5684 * NOTES
5685 * Default flags:
5686 * Uri_CREATE_CANONICALIZE, Uri_CREATE_DECODE_EXTRA_INFO, Uri_CREATE_CRACK_UNKNOWN_SCHEMES,
5687 * Uri_CREATE_PRE_PROCESS_HTML_URI, Uri_CREATE_NO_IE_SETTINGS.
5689 HRESULT WINAPI CreateUri(LPCWSTR pwzURI, DWORD dwFlags, DWORD_PTR dwReserved, IUri **ppURI)
5691 const DWORD supported_flags = Uri_CREATE_ALLOW_RELATIVE|Uri_CREATE_ALLOW_IMPLICIT_WILDCARD_SCHEME|
5692 Uri_CREATE_ALLOW_IMPLICIT_FILE_SCHEME|Uri_CREATE_NO_CANONICALIZE|Uri_CREATE_CANONICALIZE|
5693 Uri_CREATE_DECODE_EXTRA_INFO|Uri_CREATE_NO_DECODE_EXTRA_INFO|Uri_CREATE_CRACK_UNKNOWN_SCHEMES|
5694 Uri_CREATE_NO_CRACK_UNKNOWN_SCHEMES|Uri_CREATE_PRE_PROCESS_HTML_URI|Uri_CREATE_NO_PRE_PROCESS_HTML_URI|
5695 Uri_CREATE_NO_IE_SETTINGS|Uri_CREATE_NO_ENCODE_FORBIDDEN_CHARACTERS|Uri_CREATE_FILE_USE_DOS_PATH;
5696 Uri *ret;
5697 HRESULT hr;
5698 parse_data data;
5700 TRACE("(%s %x %x %p)\n", debugstr_w(pwzURI), dwFlags, (DWORD)dwReserved, ppURI);
5702 if(!ppURI)
5703 return E_INVALIDARG;
5705 if(!pwzURI) {
5706 *ppURI = NULL;
5707 return E_INVALIDARG;
5710 /* Check for invalid flags. */
5711 if(has_invalid_flag_combination(dwFlags)) {
5712 *ppURI = NULL;
5713 return E_INVALIDARG;
5716 /* Currently unsupported. */
5717 if(dwFlags & ~supported_flags)
5718 FIXME("Ignoring unsupported flag(s) %x\n", dwFlags & ~supported_flags);
5720 hr = Uri_Construct(NULL, (void**)&ret);
5721 if(FAILED(hr)) {
5722 *ppURI = NULL;
5723 return hr;
5726 /* Explicitly set the default flags if it doesn't cause a flag conflict. */
5727 apply_default_flags(&dwFlags);
5729 /* Pre process the URI, unless told otherwise. */
5730 if(!(dwFlags & Uri_CREATE_NO_PRE_PROCESS_HTML_URI))
5731 ret->raw_uri = pre_process_uri(pwzURI);
5732 else
5733 ret->raw_uri = SysAllocString(pwzURI);
5735 if(!ret->raw_uri) {
5736 heap_free(ret);
5737 return E_OUTOFMEMORY;
5740 memset(&data, 0, sizeof(parse_data));
5741 data.uri = ret->raw_uri;
5743 /* Validate and parse the URI into its components. */
5744 if(!parse_uri(&data, dwFlags)) {
5745 /* Encountered an unsupported or invalid URI */
5746 IUri_Release(&ret->IUri_iface);
5747 *ppURI = NULL;
5748 return E_INVALIDARG;
5751 /* Canonicalize the URI. */
5752 hr = canonicalize_uri(&data, ret, dwFlags);
5753 if(FAILED(hr)) {
5754 IUri_Release(&ret->IUri_iface);
5755 *ppURI = NULL;
5756 return hr;
5759 ret->create_flags = dwFlags;
5761 *ppURI = &ret->IUri_iface;
5762 return S_OK;
5765 /***********************************************************************
5766 * CreateUriWithFragment (urlmon.@)
5768 * Creates a new IUri object. This is almost the same as CreateUri, expect that
5769 * it allows you to explicitly specify a fragment (pwzFragment) for pwzURI.
5771 * PARAMS
5772 * pwzURI [I] The URI to parse and perform canonicalization on.
5773 * pwzFragment [I] The explicit fragment string which should be added to pwzURI.
5774 * dwFlags [I] The flags which will be passed to CreateUri.
5775 * dwReserved [I] Reserved (not used).
5776 * ppURI [O] The resulting IUri after parsing/canonicalization.
5778 * RETURNS
5779 * Success: S_OK. ppURI contains the pointer to the newly allocated IUri.
5780 * Failure: E_INVALIDARG if pwzURI already contains a fragment and pwzFragment
5781 * isn't NULL. Will also return E_INVALIDARG for the same reasons as
5782 * CreateUri will. E_OUTOFMEMORY if any allocation fails.
5784 HRESULT WINAPI CreateUriWithFragment(LPCWSTR pwzURI, LPCWSTR pwzFragment, DWORD dwFlags,
5785 DWORD_PTR dwReserved, IUri **ppURI)
5787 HRESULT hres;
5788 TRACE("(%s %s %x %x %p)\n", debugstr_w(pwzURI), debugstr_w(pwzFragment), dwFlags, (DWORD)dwReserved, ppURI);
5790 if(!ppURI)
5791 return E_INVALIDARG;
5793 if(!pwzURI) {
5794 *ppURI = NULL;
5795 return E_INVALIDARG;
5798 /* Check if a fragment should be appended to the URI string. */
5799 if(pwzFragment) {
5800 WCHAR *uriW;
5801 DWORD uri_len, frag_len;
5802 BOOL add_pound;
5804 /* Check if the original URI already has a fragment component. */
5805 if(StrChrW(pwzURI, '#')) {
5806 *ppURI = NULL;
5807 return E_INVALIDARG;
5810 uri_len = lstrlenW(pwzURI);
5811 frag_len = lstrlenW(pwzFragment);
5813 /* If the fragment doesn't start with a '#', one will be added. */
5814 add_pound = *pwzFragment != '#';
5816 if(add_pound)
5817 uriW = heap_alloc((uri_len+frag_len+2)*sizeof(WCHAR));
5818 else
5819 uriW = heap_alloc((uri_len+frag_len+1)*sizeof(WCHAR));
5821 if(!uriW)
5822 return E_OUTOFMEMORY;
5824 memcpy(uriW, pwzURI, uri_len*sizeof(WCHAR));
5825 if(add_pound)
5826 uriW[uri_len++] = '#';
5827 memcpy(uriW+uri_len, pwzFragment, (frag_len+1)*sizeof(WCHAR));
5829 hres = CreateUri(uriW, dwFlags, 0, ppURI);
5831 heap_free(uriW);
5832 } else
5833 /* A fragment string wasn't specified, so just forward the call. */
5834 hres = CreateUri(pwzURI, dwFlags, 0, ppURI);
5836 return hres;
5839 static HRESULT build_uri(const UriBuilder *builder, IUri **uri, DWORD create_flags,
5840 DWORD use_orig_flags, DWORD encoding_mask)
5842 HRESULT hr;
5843 parse_data data;
5844 Uri *ret;
5846 if(!uri)
5847 return E_POINTER;
5849 if(encoding_mask && (!builder->uri || builder->modified_props)) {
5850 *uri = NULL;
5851 return E_NOTIMPL;
5854 /* Decide what flags should be used when creating the Uri. */
5855 if((use_orig_flags & UriBuilder_USE_ORIGINAL_FLAGS) && builder->uri)
5856 create_flags = builder->uri->create_flags;
5857 else {
5858 if(has_invalid_flag_combination(create_flags)) {
5859 *uri = NULL;
5860 return E_INVALIDARG;
5863 /* Set the default flags if they don't cause a conflict. */
5864 apply_default_flags(&create_flags);
5867 /* Return the base IUri if no changes have been made and the create_flags match. */
5868 if(builder->uri && !builder->modified_props && builder->uri->create_flags == create_flags) {
5869 *uri = &builder->uri->IUri_iface;
5870 IUri_AddRef(*uri);
5871 return S_OK;
5874 hr = validate_components(builder, &data, create_flags);
5875 if(FAILED(hr)) {
5876 *uri = NULL;
5877 return hr;
5880 hr = Uri_Construct(NULL, (void**)&ret);
5881 if(FAILED(hr)) {
5882 *uri = NULL;
5883 return hr;
5886 hr = generate_uri(builder, &data, ret, create_flags);
5887 if(FAILED(hr)) {
5888 IUri_Release(&ret->IUri_iface);
5889 *uri = NULL;
5890 return hr;
5893 *uri = &ret->IUri_iface;
5894 return S_OK;
5897 static inline UriBuilder* impl_from_IUriBuilder(IUriBuilder *iface)
5899 return CONTAINING_RECORD(iface, UriBuilder, IUriBuilder_iface);
5902 static HRESULT WINAPI UriBuilder_QueryInterface(IUriBuilder *iface, REFIID riid, void **ppv)
5904 UriBuilder *This = impl_from_IUriBuilder(iface);
5906 if(IsEqualGUID(&IID_IUnknown, riid)) {
5907 TRACE("(%p)->(IID_IUnknown %p)\n", This, ppv);
5908 *ppv = &This->IUriBuilder_iface;
5909 }else if(IsEqualGUID(&IID_IUriBuilder, riid)) {
5910 TRACE("(%p)->(IID_IUriBuilder %p)\n", This, ppv);
5911 *ppv = &This->IUriBuilder_iface;
5912 }else {
5913 TRACE("(%p)->(%s %p)\n", This, debugstr_guid(riid), ppv);
5914 *ppv = NULL;
5915 return E_NOINTERFACE;
5918 IUnknown_AddRef((IUnknown*)*ppv);
5919 return S_OK;
5922 static ULONG WINAPI UriBuilder_AddRef(IUriBuilder *iface)
5924 UriBuilder *This = impl_from_IUriBuilder(iface);
5925 LONG ref = InterlockedIncrement(&This->ref);
5927 TRACE("(%p) ref=%d\n", This, ref);
5929 return ref;
5932 static ULONG WINAPI UriBuilder_Release(IUriBuilder *iface)
5934 UriBuilder *This = impl_from_IUriBuilder(iface);
5935 LONG ref = InterlockedDecrement(&This->ref);
5937 TRACE("(%p) ref=%d\n", This, ref);
5939 if(!ref) {
5940 if(This->uri) IUri_Release(&This->uri->IUri_iface);
5941 heap_free(This->fragment);
5942 heap_free(This->host);
5943 heap_free(This->password);
5944 heap_free(This->path);
5945 heap_free(This->query);
5946 heap_free(This->scheme);
5947 heap_free(This->username);
5948 heap_free(This);
5951 return ref;
5954 static HRESULT WINAPI UriBuilder_CreateUriSimple(IUriBuilder *iface,
5955 DWORD dwAllowEncodingPropertyMask,
5956 DWORD_PTR dwReserved,
5957 IUri **ppIUri)
5959 UriBuilder *This = impl_from_IUriBuilder(iface);
5960 HRESULT hr;
5961 TRACE("(%p)->(%d %d %p)\n", This, dwAllowEncodingPropertyMask, (DWORD)dwReserved, ppIUri);
5963 hr = build_uri(This, ppIUri, 0, UriBuilder_USE_ORIGINAL_FLAGS, dwAllowEncodingPropertyMask);
5964 if(hr == E_NOTIMPL)
5965 FIXME("(%p)->(%d %d %p)\n", This, dwAllowEncodingPropertyMask, (DWORD)dwReserved, ppIUri);
5966 return hr;
5969 static HRESULT WINAPI UriBuilder_CreateUri(IUriBuilder *iface,
5970 DWORD dwCreateFlags,
5971 DWORD dwAllowEncodingPropertyMask,
5972 DWORD_PTR dwReserved,
5973 IUri **ppIUri)
5975 UriBuilder *This = impl_from_IUriBuilder(iface);
5976 HRESULT hr;
5977 TRACE("(%p)->(0x%08x %d %d %p)\n", This, dwCreateFlags, dwAllowEncodingPropertyMask, (DWORD)dwReserved, ppIUri);
5979 if(dwCreateFlags == -1)
5980 hr = build_uri(This, ppIUri, 0, UriBuilder_USE_ORIGINAL_FLAGS, dwAllowEncodingPropertyMask);
5981 else
5982 hr = build_uri(This, ppIUri, dwCreateFlags, 0, dwAllowEncodingPropertyMask);
5984 if(hr == E_NOTIMPL)
5985 FIXME("(%p)->(0x%08x %d %d %p)\n", This, dwCreateFlags, dwAllowEncodingPropertyMask, (DWORD)dwReserved, ppIUri);
5986 return hr;
5989 static HRESULT WINAPI UriBuilder_CreateUriWithFlags(IUriBuilder *iface,
5990 DWORD dwCreateFlags,
5991 DWORD dwUriBuilderFlags,
5992 DWORD dwAllowEncodingPropertyMask,
5993 DWORD_PTR dwReserved,
5994 IUri **ppIUri)
5996 UriBuilder *This = impl_from_IUriBuilder(iface);
5997 HRESULT hr;
5998 TRACE("(%p)->(0x%08x 0x%08x %d %d %p)\n", This, dwCreateFlags, dwUriBuilderFlags,
5999 dwAllowEncodingPropertyMask, (DWORD)dwReserved, ppIUri);
6001 hr = build_uri(This, ppIUri, dwCreateFlags, dwUriBuilderFlags, dwAllowEncodingPropertyMask);
6002 if(hr == E_NOTIMPL)
6003 FIXME("(%p)->(0x%08x 0x%08x %d %d %p)\n", This, dwCreateFlags, dwUriBuilderFlags,
6004 dwAllowEncodingPropertyMask, (DWORD)dwReserved, ppIUri);
6005 return hr;
6008 static HRESULT WINAPI UriBuilder_GetIUri(IUriBuilder *iface, IUri **ppIUri)
6010 UriBuilder *This = impl_from_IUriBuilder(iface);
6011 TRACE("(%p)->(%p)\n", This, ppIUri);
6013 if(!ppIUri)
6014 return E_POINTER;
6016 if(This->uri) {
6017 IUri *uri = &This->uri->IUri_iface;
6018 IUri_AddRef(uri);
6019 *ppIUri = uri;
6020 } else
6021 *ppIUri = NULL;
6023 return S_OK;
6026 static HRESULT WINAPI UriBuilder_SetIUri(IUriBuilder *iface, IUri *pIUri)
6028 UriBuilder *This = impl_from_IUriBuilder(iface);
6029 TRACE("(%p)->(%p)\n", This, pIUri);
6031 if(pIUri) {
6032 Uri *uri;
6034 if((uri = get_uri_obj(pIUri))) {
6035 /* Only reset the builder if its Uri isn't the same as
6036 * the Uri passed to the function.
6038 if(This->uri != uri) {
6039 reset_builder(This);
6041 This->uri = uri;
6042 if(uri->has_port)
6043 This->port = uri->port;
6045 IUri_AddRef(pIUri);
6047 } else {
6048 FIXME("(%p)->(%p) Unknown IUri types not supported yet.\n", This, pIUri);
6049 return E_NOTIMPL;
6051 } else if(This->uri)
6052 /* Only reset the builder if its Uri isn't NULL. */
6053 reset_builder(This);
6055 return S_OK;
6058 static HRESULT WINAPI UriBuilder_GetFragment(IUriBuilder *iface, DWORD *pcchFragment, LPCWSTR *ppwzFragment)
6060 UriBuilder *This = impl_from_IUriBuilder(iface);
6061 TRACE("(%p)->(%p %p)\n", This, pcchFragment, ppwzFragment);
6063 if(!This->uri || This->uri->fragment_start == -1 || This->modified_props & Uri_HAS_FRAGMENT)
6064 return get_builder_component(&This->fragment, &This->fragment_len, NULL, 0, ppwzFragment, pcchFragment);
6065 else
6066 return get_builder_component(&This->fragment, &This->fragment_len, This->uri->canon_uri+This->uri->fragment_start,
6067 This->uri->fragment_len, ppwzFragment, pcchFragment);
6070 static HRESULT WINAPI UriBuilder_GetHost(IUriBuilder *iface, DWORD *pcchHost, LPCWSTR *ppwzHost)
6072 UriBuilder *This = impl_from_IUriBuilder(iface);
6073 TRACE("(%p)->(%p %p)\n", This, pcchHost, ppwzHost);
6075 if(!This->uri || This->uri->host_start == -1 || This->modified_props & Uri_HAS_HOST)
6076 return get_builder_component(&This->host, &This->host_len, NULL, 0, ppwzHost, pcchHost);
6077 else {
6078 if(This->uri->host_type == Uri_HOST_IPV6)
6079 /* Don't include the '[' and ']' around the address. */
6080 return get_builder_component(&This->host, &This->host_len, This->uri->canon_uri+This->uri->host_start+1,
6081 This->uri->host_len-2, ppwzHost, pcchHost);
6082 else
6083 return get_builder_component(&This->host, &This->host_len, This->uri->canon_uri+This->uri->host_start,
6084 This->uri->host_len, ppwzHost, pcchHost);
6088 static HRESULT WINAPI UriBuilder_GetPassword(IUriBuilder *iface, DWORD *pcchPassword, LPCWSTR *ppwzPassword)
6090 UriBuilder *This = impl_from_IUriBuilder(iface);
6091 TRACE("(%p)->(%p %p)\n", This, pcchPassword, ppwzPassword);
6093 if(!This->uri || This->uri->userinfo_split == -1 || This->modified_props & Uri_HAS_PASSWORD)
6094 return get_builder_component(&This->password, &This->password_len, NULL, 0, ppwzPassword, pcchPassword);
6095 else {
6096 const WCHAR *start = This->uri->canon_uri+This->uri->userinfo_start+This->uri->userinfo_split+1;
6097 DWORD len = This->uri->userinfo_len-This->uri->userinfo_split-1;
6098 return get_builder_component(&This->password, &This->password_len, start, len, ppwzPassword, pcchPassword);
6102 static HRESULT WINAPI UriBuilder_GetPath(IUriBuilder *iface, DWORD *pcchPath, LPCWSTR *ppwzPath)
6104 UriBuilder *This = impl_from_IUriBuilder(iface);
6105 TRACE("(%p)->(%p %p)\n", This, pcchPath, ppwzPath);
6107 if(!This->uri || This->uri->path_start == -1 || This->modified_props & Uri_HAS_PATH)
6108 return get_builder_component(&This->path, &This->path_len, NULL, 0, ppwzPath, pcchPath);
6109 else
6110 return get_builder_component(&This->path, &This->path_len, This->uri->canon_uri+This->uri->path_start,
6111 This->uri->path_len, ppwzPath, pcchPath);
6114 static HRESULT WINAPI UriBuilder_GetPort(IUriBuilder *iface, BOOL *pfHasPort, DWORD *pdwPort)
6116 UriBuilder *This = impl_from_IUriBuilder(iface);
6117 TRACE("(%p)->(%p %p)\n", This, pfHasPort, pdwPort);
6119 if(!pfHasPort) {
6120 if(pdwPort)
6121 *pdwPort = 0;
6122 return E_POINTER;
6125 if(!pdwPort) {
6126 *pfHasPort = FALSE;
6127 return E_POINTER;
6130 *pfHasPort = This->has_port;
6131 *pdwPort = This->port;
6132 return S_OK;
6135 static HRESULT WINAPI UriBuilder_GetQuery(IUriBuilder *iface, DWORD *pcchQuery, LPCWSTR *ppwzQuery)
6137 UriBuilder *This = impl_from_IUriBuilder(iface);
6138 TRACE("(%p)->(%p %p)\n", This, pcchQuery, ppwzQuery);
6140 if(!This->uri || This->uri->query_start == -1 || This->modified_props & Uri_HAS_QUERY)
6141 return get_builder_component(&This->query, &This->query_len, NULL, 0, ppwzQuery, pcchQuery);
6142 else
6143 return get_builder_component(&This->query, &This->query_len, This->uri->canon_uri+This->uri->query_start,
6144 This->uri->query_len, ppwzQuery, pcchQuery);
6147 static HRESULT WINAPI UriBuilder_GetSchemeName(IUriBuilder *iface, DWORD *pcchSchemeName, LPCWSTR *ppwzSchemeName)
6149 UriBuilder *This = impl_from_IUriBuilder(iface);
6150 TRACE("(%p)->(%p %p)\n", This, pcchSchemeName, ppwzSchemeName);
6152 if(!This->uri || This->uri->scheme_start == -1 || This->modified_props & Uri_HAS_SCHEME_NAME)
6153 return get_builder_component(&This->scheme, &This->scheme_len, NULL, 0, ppwzSchemeName, pcchSchemeName);
6154 else
6155 return get_builder_component(&This->scheme, &This->scheme_len, This->uri->canon_uri+This->uri->scheme_start,
6156 This->uri->scheme_len, ppwzSchemeName, pcchSchemeName);
6159 static HRESULT WINAPI UriBuilder_GetUserName(IUriBuilder *iface, DWORD *pcchUserName, LPCWSTR *ppwzUserName)
6161 UriBuilder *This = impl_from_IUriBuilder(iface);
6162 TRACE("(%p)->(%p %p)\n", This, pcchUserName, ppwzUserName);
6164 if(!This->uri || This->uri->userinfo_start == -1 || This->uri->userinfo_split == 0 ||
6165 This->modified_props & Uri_HAS_USER_NAME)
6166 return get_builder_component(&This->username, &This->username_len, NULL, 0, ppwzUserName, pcchUserName);
6167 else {
6168 const WCHAR *start = This->uri->canon_uri+This->uri->userinfo_start;
6170 /* Check if there's a password in the userinfo section. */
6171 if(This->uri->userinfo_split > -1)
6172 /* Don't include the password. */
6173 return get_builder_component(&This->username, &This->username_len, start,
6174 This->uri->userinfo_split, ppwzUserName, pcchUserName);
6175 else
6176 return get_builder_component(&This->username, &This->username_len, start,
6177 This->uri->userinfo_len, ppwzUserName, pcchUserName);
6181 static HRESULT WINAPI UriBuilder_SetFragment(IUriBuilder *iface, LPCWSTR pwzNewValue)
6183 UriBuilder *This = impl_from_IUriBuilder(iface);
6184 TRACE("(%p)->(%s)\n", This, debugstr_w(pwzNewValue));
6185 return set_builder_component(&This->fragment, &This->fragment_len, pwzNewValue, '#',
6186 &This->modified_props, Uri_HAS_FRAGMENT);
6189 static HRESULT WINAPI UriBuilder_SetHost(IUriBuilder *iface, LPCWSTR pwzNewValue)
6191 UriBuilder *This = impl_from_IUriBuilder(iface);
6192 TRACE("(%p)->(%s)\n", This, debugstr_w(pwzNewValue));
6194 /* Host name can't be set to NULL. */
6195 if(!pwzNewValue)
6196 return E_INVALIDARG;
6198 return set_builder_component(&This->host, &This->host_len, pwzNewValue, 0,
6199 &This->modified_props, Uri_HAS_HOST);
6202 static HRESULT WINAPI UriBuilder_SetPassword(IUriBuilder *iface, LPCWSTR pwzNewValue)
6204 UriBuilder *This = impl_from_IUriBuilder(iface);
6205 TRACE("(%p)->(%s)\n", This, debugstr_w(pwzNewValue));
6206 return set_builder_component(&This->password, &This->password_len, pwzNewValue, 0,
6207 &This->modified_props, Uri_HAS_PASSWORD);
6210 static HRESULT WINAPI UriBuilder_SetPath(IUriBuilder *iface, LPCWSTR pwzNewValue)
6212 UriBuilder *This = impl_from_IUriBuilder(iface);
6213 TRACE("(%p)->(%s)\n", This, debugstr_w(pwzNewValue));
6214 return set_builder_component(&This->path, &This->path_len, pwzNewValue, 0,
6215 &This->modified_props, Uri_HAS_PATH);
6218 static HRESULT WINAPI UriBuilder_SetPort(IUriBuilder *iface, BOOL fHasPort, DWORD dwNewValue)
6220 UriBuilder *This = impl_from_IUriBuilder(iface);
6221 TRACE("(%p)->(%d %d)\n", This, fHasPort, dwNewValue);
6223 This->has_port = fHasPort;
6224 This->port = dwNewValue;
6225 This->modified_props |= Uri_HAS_PORT;
6226 return S_OK;
6229 static HRESULT WINAPI UriBuilder_SetQuery(IUriBuilder *iface, LPCWSTR pwzNewValue)
6231 UriBuilder *This = impl_from_IUriBuilder(iface);
6232 TRACE("(%p)->(%s)\n", This, debugstr_w(pwzNewValue));
6233 return set_builder_component(&This->query, &This->query_len, pwzNewValue, '?',
6234 &This->modified_props, Uri_HAS_QUERY);
6237 static HRESULT WINAPI UriBuilder_SetSchemeName(IUriBuilder *iface, LPCWSTR pwzNewValue)
6239 UriBuilder *This = impl_from_IUriBuilder(iface);
6240 TRACE("(%p)->(%s)\n", This, debugstr_w(pwzNewValue));
6242 /* Only set the scheme name if it's not NULL or empty. */
6243 if(!pwzNewValue || !*pwzNewValue)
6244 return E_INVALIDARG;
6246 return set_builder_component(&This->scheme, &This->scheme_len, pwzNewValue, 0,
6247 &This->modified_props, Uri_HAS_SCHEME_NAME);
6250 static HRESULT WINAPI UriBuilder_SetUserName(IUriBuilder *iface, LPCWSTR pwzNewValue)
6252 UriBuilder *This = impl_from_IUriBuilder(iface);
6253 TRACE("(%p)->(%s)\n", This, debugstr_w(pwzNewValue));
6254 return set_builder_component(&This->username, &This->username_len, pwzNewValue, 0,
6255 &This->modified_props, Uri_HAS_USER_NAME);
6258 static HRESULT WINAPI UriBuilder_RemoveProperties(IUriBuilder *iface, DWORD dwPropertyMask)
6260 const DWORD accepted_flags = Uri_HAS_AUTHORITY|Uri_HAS_DOMAIN|Uri_HAS_EXTENSION|Uri_HAS_FRAGMENT|Uri_HAS_HOST|
6261 Uri_HAS_PASSWORD|Uri_HAS_PATH|Uri_HAS_PATH_AND_QUERY|Uri_HAS_QUERY|
6262 Uri_HAS_USER_INFO|Uri_HAS_USER_NAME;
6264 UriBuilder *This = impl_from_IUriBuilder(iface);
6265 TRACE("(%p)->(0x%08x)\n", This, dwPropertyMask);
6267 if(dwPropertyMask & ~accepted_flags)
6268 return E_INVALIDARG;
6270 if(dwPropertyMask & Uri_HAS_FRAGMENT)
6271 UriBuilder_SetFragment(iface, NULL);
6273 /* Even though you can't set the host name to NULL or an
6274 * empty string, you can still remove it... for some reason.
6276 if(dwPropertyMask & Uri_HAS_HOST)
6277 set_builder_component(&This->host, &This->host_len, NULL, 0,
6278 &This->modified_props, Uri_HAS_HOST);
6280 if(dwPropertyMask & Uri_HAS_PASSWORD)
6281 UriBuilder_SetPassword(iface, NULL);
6283 if(dwPropertyMask & Uri_HAS_PATH)
6284 UriBuilder_SetPath(iface, NULL);
6286 if(dwPropertyMask & Uri_HAS_PORT)
6287 UriBuilder_SetPort(iface, FALSE, 0);
6289 if(dwPropertyMask & Uri_HAS_QUERY)
6290 UriBuilder_SetQuery(iface, NULL);
6292 if(dwPropertyMask & Uri_HAS_USER_NAME)
6293 UriBuilder_SetUserName(iface, NULL);
6295 return S_OK;
6298 static HRESULT WINAPI UriBuilder_HasBeenModified(IUriBuilder *iface, BOOL *pfModified)
6300 UriBuilder *This = impl_from_IUriBuilder(iface);
6301 TRACE("(%p)->(%p)\n", This, pfModified);
6303 if(!pfModified)
6304 return E_POINTER;
6306 *pfModified = This->modified_props > 0;
6307 return S_OK;
6310 static const IUriBuilderVtbl UriBuilderVtbl = {
6311 UriBuilder_QueryInterface,
6312 UriBuilder_AddRef,
6313 UriBuilder_Release,
6314 UriBuilder_CreateUriSimple,
6315 UriBuilder_CreateUri,
6316 UriBuilder_CreateUriWithFlags,
6317 UriBuilder_GetIUri,
6318 UriBuilder_SetIUri,
6319 UriBuilder_GetFragment,
6320 UriBuilder_GetHost,
6321 UriBuilder_GetPassword,
6322 UriBuilder_GetPath,
6323 UriBuilder_GetPort,
6324 UriBuilder_GetQuery,
6325 UriBuilder_GetSchemeName,
6326 UriBuilder_GetUserName,
6327 UriBuilder_SetFragment,
6328 UriBuilder_SetHost,
6329 UriBuilder_SetPassword,
6330 UriBuilder_SetPath,
6331 UriBuilder_SetPort,
6332 UriBuilder_SetQuery,
6333 UriBuilder_SetSchemeName,
6334 UriBuilder_SetUserName,
6335 UriBuilder_RemoveProperties,
6336 UriBuilder_HasBeenModified,
6339 /***********************************************************************
6340 * CreateIUriBuilder (urlmon.@)
6342 HRESULT WINAPI CreateIUriBuilder(IUri *pIUri, DWORD dwFlags, DWORD_PTR dwReserved, IUriBuilder **ppIUriBuilder)
6344 UriBuilder *ret;
6346 TRACE("(%p %x %x %p)\n", pIUri, dwFlags, (DWORD)dwReserved, ppIUriBuilder);
6348 if(!ppIUriBuilder)
6349 return E_POINTER;
6351 ret = heap_alloc_zero(sizeof(UriBuilder));
6352 if(!ret)
6353 return E_OUTOFMEMORY;
6355 ret->IUriBuilder_iface.lpVtbl = &UriBuilderVtbl;
6356 ret->ref = 1;
6358 if(pIUri) {
6359 Uri *uri;
6361 if((uri = get_uri_obj(pIUri))) {
6362 if(!uri->create_flags) {
6363 heap_free(ret);
6364 return E_UNEXPECTED;
6366 IUri_AddRef(pIUri);
6367 ret->uri = uri;
6369 if(uri->has_port)
6370 /* Windows doesn't set 'has_port' to TRUE in this case. */
6371 ret->port = uri->port;
6373 } else {
6374 heap_free(ret);
6375 *ppIUriBuilder = NULL;
6376 FIXME("(%p %x %x %p): Unknown IUri types not supported yet.\n", pIUri, dwFlags,
6377 (DWORD)dwReserved, ppIUriBuilder);
6378 return E_NOTIMPL;
6382 *ppIUriBuilder = &ret->IUriBuilder_iface;
6383 return S_OK;
6386 /* Merges the base path with the relative path and stores the resulting path
6387 * and path len in 'result' and 'result_len'.
6389 static HRESULT merge_paths(parse_data *data, const WCHAR *base, DWORD base_len, const WCHAR *relative,
6390 DWORD relative_len, WCHAR **result, DWORD *result_len, DWORD flags)
6392 const WCHAR *end = NULL;
6393 DWORD base_copy_len = 0;
6394 WCHAR *ptr;
6396 if(base_len) {
6397 if(data->scheme_type == URL_SCHEME_MK && *relative == '/') {
6398 /* Find '::' segment */
6399 for(end = base; end < base+base_len-1; end++) {
6400 if(end[0] == ':' && end[1] == ':') {
6401 end++;
6402 break;
6406 /* If not found, try finding the end of @xxx: */
6407 if(end == base+base_len-1)
6408 end = *base == '@' ? wmemchr(base, ':', base_len) : NULL;
6409 }else {
6410 /* Find the characters that will be copied over from the base path. */
6411 for (end = base + base_len - 1; end >= base; end--) if (*end == '/') break;
6412 if(end < base && data->scheme_type == URL_SCHEME_FILE)
6413 /* Try looking for a '\\'. */
6414 for (end = base + base_len - 1; end >= base; end--) if (*end == '\\') break;
6418 if (end) base_copy_len = (end+1)-base;
6419 *result = heap_alloc((base_copy_len+relative_len+1)*sizeof(WCHAR));
6421 if(!(*result)) {
6422 *result_len = 0;
6423 return E_OUTOFMEMORY;
6426 ptr = *result;
6427 memcpy(ptr, base, base_copy_len*sizeof(WCHAR));
6428 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, Uri_CREATE_ALLOW_IMPLICIT_FILE_SCHEME);
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 const WCHAR *host = proc_uri->canon_uri+proc_uri->host_start;
6516 parse_host(&host, &data, 0);
6519 if(proc_uri->has_port) {
6520 data.has_port = TRUE;
6521 data.port_value = proc_uri->port;
6523 } else if(base->scheme_type != URL_SCHEME_FILE)
6524 data.is_opaque = TRUE;
6526 if(proc_uri == relative || relative->path_start == -1 || !relative->path_len) {
6527 if(proc_uri->path_start > -1) {
6528 data.path = proc_uri->canon_uri+proc_uri->path_start;
6529 data.path_len = proc_uri->path_len;
6530 } else if(!data.is_opaque) {
6531 /* Just set the path as a '/' if the base didn't have
6532 * one and if it's a hierarchical URI.
6534 static const WCHAR slashW[] = {'/',0};
6535 data.path = slashW;
6536 data.path_len = 1;
6539 if(relative->query_start > -1)
6540 proc_uri = relative;
6542 if(proc_uri->query_start > -1) {
6543 data.query = proc_uri->canon_uri+proc_uri->query_start;
6544 data.query_len = proc_uri->query_len;
6546 } else {
6547 const WCHAR *ptr, **pptr;
6548 DWORD path_offset = 0, path_len = 0;
6550 /* There's two possibilities on what will happen to the path component
6551 * of the result IUri. First, if the relative path begins with a '/'
6552 * then the resulting path will just be the relative path. Second, if
6553 * relative path doesn't begin with a '/' then the base path and relative
6554 * path are merged together.
6556 if(relative->path_len && *(relative->canon_uri+relative->path_start) == '/' && data.scheme_type != URL_SCHEME_MK) {
6557 WCHAR *tmp = NULL;
6558 BOOL copy_drive_path = FALSE;
6560 /* If the relative IUri's path starts with a '/', then we
6561 * don't use the base IUri's path. Unless the base IUri
6562 * is a file URI, in which case it uses the drive path of
6563 * the base IUri (if it has any) in the new path.
6565 if(base->scheme_type == URL_SCHEME_FILE) {
6566 if(base->path_len > 3 && *(base->canon_uri+base->path_start) == '/' &&
6567 is_drive_path(base->canon_uri+base->path_start+1)) {
6568 path_len += 3;
6569 copy_drive_path = TRUE;
6573 path_len += relative->path_len;
6575 path = heap_alloc((path_len+1)*sizeof(WCHAR));
6576 if(!path) {
6577 *result = NULL;
6578 return E_OUTOFMEMORY;
6581 tmp = path;
6583 /* Copy the base paths, drive path over. */
6584 if(copy_drive_path) {
6585 memcpy(tmp, base->canon_uri+base->path_start, 3*sizeof(WCHAR));
6586 tmp += 3;
6589 memcpy(tmp, relative->canon_uri+relative->path_start, relative->path_len*sizeof(WCHAR));
6590 path[path_len] = '\0';
6591 } else {
6592 /* Merge the base path with the relative path. */
6593 hr = merge_paths(&data, base->canon_uri+base->path_start, base->path_len,
6594 relative->canon_uri+relative->path_start, relative->path_len,
6595 &path, &path_len, flags);
6596 if(FAILED(hr)) {
6597 *result = NULL;
6598 return hr;
6601 /* If the resulting IUri is a file URI, the drive path isn't
6602 * reduced out when the dot segments are removed.
6604 if(path_len >= 3 && data.scheme_type == URL_SCHEME_FILE && !data.host) {
6605 if(*path == '/' && is_drive_path(path+1))
6606 path_offset = 2;
6607 else if(is_drive_path(path))
6608 path_offset = 1;
6612 /* Check if the dot segments need to be removed from the path. */
6613 if(!(flags & URL_DONT_SIMPLIFY) && !data.is_opaque) {
6614 DWORD offset = (path_offset > 0) ? path_offset+1 : 0;
6615 DWORD new_len = remove_dot_segments(path+offset,path_len-offset);
6617 if(new_len != path_len) {
6618 WCHAR *tmp = heap_realloc(path, (offset+new_len+1)*sizeof(WCHAR));
6619 if(!tmp) {
6620 heap_free(path);
6621 *result = NULL;
6622 return E_OUTOFMEMORY;
6625 tmp[new_len+offset] = '\0';
6626 path = tmp;
6627 path_len = new_len+offset;
6631 if(relative->query_start > -1) {
6632 data.query = relative->canon_uri+relative->query_start;
6633 data.query_len = relative->query_len;
6636 /* Make sure the path component is valid. */
6637 ptr = path;
6638 pptr = &ptr;
6639 if((data.is_opaque && !parse_path_opaque(pptr, &data, 0)) ||
6640 (!data.is_opaque && !parse_path_hierarchical(pptr, &data, 0))) {
6641 heap_free(path);
6642 *result = NULL;
6643 return E_INVALIDARG;
6647 if(relative->fragment_start > -1) {
6648 data.fragment = relative->canon_uri+relative->fragment_start;
6649 data.fragment_len = relative->fragment_len;
6652 if(flags & URL_DONT_SIMPLIFY)
6653 raw_flags |= RAW_URI_FORCE_PORT_DISP;
6654 if(flags & URL_FILE_USE_PATHURL)
6655 raw_flags |= RAW_URI_CONVERT_TO_DOS_PATH;
6657 len = generate_raw_uri(&data, data.uri, raw_flags);
6658 data.uri = SysAllocStringLen(NULL, len);
6659 if(!data.uri) {
6660 heap_free(path);
6661 *result = NULL;
6662 return E_OUTOFMEMORY;
6665 generate_raw_uri(&data, data.uri, raw_flags);
6667 hr = Uri_Construct(NULL, (void**)&ret);
6668 if(FAILED(hr)) {
6669 SysFreeString(data.uri);
6670 heap_free(path);
6671 *result = NULL;
6672 return hr;
6675 if(flags & URL_DONT_SIMPLIFY)
6676 create_flags |= Uri_CREATE_NO_CANONICALIZE;
6677 if(flags & URL_FILE_USE_PATHURL)
6678 create_flags |= Uri_CREATE_FILE_USE_DOS_PATH;
6680 ret->raw_uri = data.uri;
6681 hr = canonicalize_uri(&data, ret, create_flags);
6682 if(FAILED(hr)) {
6683 IUri_Release(&ret->IUri_iface);
6684 *result = NULL;
6685 return hr;
6688 if(flags & URL_DONT_SIMPLIFY)
6689 ret->display_modifiers |= URI_DISPLAY_NO_DEFAULT_PORT_AUTH;
6691 apply_default_flags(&create_flags);
6692 ret->create_flags = create_flags;
6693 *result = &ret->IUri_iface;
6695 heap_free(path);
6698 return S_OK;
6701 /***********************************************************************
6702 * CoInternetCombineIUri (urlmon.@)
6704 HRESULT WINAPI CoInternetCombineIUri(IUri *pBaseUri, IUri *pRelativeUri, DWORD dwCombineFlags,
6705 IUri **ppCombinedUri, DWORD_PTR dwReserved)
6707 HRESULT hr;
6708 IInternetProtocolInfo *info;
6709 Uri *relative, *base;
6710 TRACE("(%p %p %x %p %x)\n", pBaseUri, pRelativeUri, dwCombineFlags, ppCombinedUri, (DWORD)dwReserved);
6712 if(!ppCombinedUri)
6713 return E_INVALIDARG;
6715 if(!pBaseUri || !pRelativeUri) {
6716 *ppCombinedUri = NULL;
6717 return E_INVALIDARG;
6720 relative = get_uri_obj(pRelativeUri);
6721 base = get_uri_obj(pBaseUri);
6722 if(!relative || !base) {
6723 *ppCombinedUri = NULL;
6724 FIXME("(%p %p %x %p %x) Unknown IUri types not supported yet.\n",
6725 pBaseUri, pRelativeUri, dwCombineFlags, ppCombinedUri, (DWORD)dwReserved);
6726 return E_NOTIMPL;
6729 info = get_protocol_info(base->canon_uri);
6730 if(info) {
6731 WCHAR result[INTERNET_MAX_URL_LENGTH+1];
6732 DWORD result_len = 0;
6734 hr = IInternetProtocolInfo_CombineUrl(info, base->canon_uri, relative->canon_uri, dwCombineFlags,
6735 result, INTERNET_MAX_URL_LENGTH+1, &result_len, 0);
6736 IInternetProtocolInfo_Release(info);
6737 if(SUCCEEDED(hr)) {
6738 hr = CreateUri(result, Uri_CREATE_ALLOW_RELATIVE, 0, ppCombinedUri);
6739 if(SUCCEEDED(hr))
6740 return hr;
6744 return combine_uri(base, relative, dwCombineFlags, ppCombinedUri, 0);
6747 /***********************************************************************
6748 * CoInternetCombineUrlEx (urlmon.@)
6750 HRESULT WINAPI CoInternetCombineUrlEx(IUri *pBaseUri, LPCWSTR pwzRelativeUrl, DWORD dwCombineFlags,
6751 IUri **ppCombinedUri, DWORD_PTR dwReserved)
6753 IUri *relative;
6754 Uri *base;
6755 HRESULT hr;
6756 IInternetProtocolInfo *info;
6758 TRACE("(%p %s %x %p %x)\n", pBaseUri, debugstr_w(pwzRelativeUrl), dwCombineFlags,
6759 ppCombinedUri, (DWORD)dwReserved);
6761 if(!ppCombinedUri)
6762 return E_POINTER;
6764 if(!pwzRelativeUrl) {
6765 *ppCombinedUri = NULL;
6766 return E_UNEXPECTED;
6769 if(!pBaseUri) {
6770 *ppCombinedUri = NULL;
6771 return E_INVALIDARG;
6774 base = get_uri_obj(pBaseUri);
6775 if(!base) {
6776 *ppCombinedUri = NULL;
6777 FIXME("(%p %s %x %p %x) Unknown IUri's not supported yet.\n", pBaseUri, debugstr_w(pwzRelativeUrl),
6778 dwCombineFlags, ppCombinedUri, (DWORD)dwReserved);
6779 return E_NOTIMPL;
6782 info = get_protocol_info(base->canon_uri);
6783 if(info) {
6784 WCHAR result[INTERNET_MAX_URL_LENGTH+1];
6785 DWORD result_len = 0;
6787 hr = IInternetProtocolInfo_CombineUrl(info, base->canon_uri, pwzRelativeUrl, dwCombineFlags,
6788 result, INTERNET_MAX_URL_LENGTH+1, &result_len, 0);
6789 IInternetProtocolInfo_Release(info);
6790 if(SUCCEEDED(hr)) {
6791 hr = CreateUri(result, Uri_CREATE_ALLOW_RELATIVE, 0, ppCombinedUri);
6792 if(SUCCEEDED(hr))
6793 return hr;
6797 hr = CreateUri(pwzRelativeUrl, Uri_CREATE_ALLOW_RELATIVE|Uri_CREATE_ALLOW_IMPLICIT_FILE_SCHEME, 0, &relative);
6798 if(FAILED(hr)) {
6799 *ppCombinedUri = NULL;
6800 return hr;
6803 hr = combine_uri(base, get_uri_obj(relative), dwCombineFlags, ppCombinedUri, COMBINE_URI_FORCE_FLAG_USE);
6805 IUri_Release(relative);
6806 return hr;
6809 static HRESULT parse_canonicalize(const Uri *uri, DWORD flags, LPWSTR output,
6810 DWORD output_len, DWORD *result_len)
6812 const WCHAR *ptr = NULL;
6813 WCHAR *path = NULL;
6814 const WCHAR **pptr;
6815 DWORD len = 0;
6816 BOOL reduce_path;
6818 /* URL_UNESCAPE only has effect if none of the URL_ESCAPE flags are set. */
6819 const BOOL allow_unescape = !(flags & URL_ESCAPE_UNSAFE) &&
6820 !(flags & URL_ESCAPE_SPACES_ONLY) &&
6821 !(flags & URL_ESCAPE_PERCENT);
6824 /* Check if the dot segments need to be removed from the
6825 * path component.
6827 if(uri->scheme_start > -1 && uri->path_start > -1) {
6828 ptr = uri->canon_uri+uri->scheme_start+uri->scheme_len+1;
6829 pptr = &ptr;
6831 reduce_path = !(flags & URL_DONT_SIMPLIFY) &&
6832 ptr && check_hierarchical(pptr);
6834 for(ptr = uri->canon_uri; ptr < uri->canon_uri+uri->canon_len; ++ptr) {
6835 BOOL do_default_action = TRUE;
6837 /* Keep track of the path if we need to remove dot segments from
6838 * it later.
6840 if(reduce_path && !path && ptr == uri->canon_uri+uri->path_start)
6841 path = output+len;
6843 /* Check if it's time to reduce the path. */
6844 if(reduce_path && ptr == uri->canon_uri+uri->path_start+uri->path_len) {
6845 DWORD current_path_len = (output+len) - path;
6846 DWORD new_path_len = remove_dot_segments(path, current_path_len);
6848 /* Update the current length. */
6849 len -= (current_path_len-new_path_len);
6850 reduce_path = FALSE;
6853 if(*ptr == '%') {
6854 const WCHAR decoded = decode_pct_val(ptr);
6855 if(decoded) {
6856 if(allow_unescape && (flags & URL_UNESCAPE)) {
6857 if(len < output_len)
6858 output[len] = decoded;
6859 len++;
6860 ptr += 2;
6861 do_default_action = FALSE;
6865 /* See if %'s needed to encoded. */
6866 if(do_default_action && (flags & URL_ESCAPE_PERCENT)) {
6867 if(len + 3 < output_len)
6868 pct_encode_val(*ptr, output+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 if(len + 3 < output_len)
6876 pct_encode_val(*ptr, output+len);
6877 len += 3;
6878 do_default_action = FALSE;
6880 } else if(is_ascii(*ptr) && !is_reserved(*ptr) && !is_unreserved(*ptr)) {
6881 if(flags & URL_ESCAPE_UNSAFE) {
6882 if(len + 3 < output_len)
6883 pct_encode_val(*ptr, output+len);
6884 len += 3;
6885 do_default_action = FALSE;
6889 if(do_default_action) {
6890 if(len < output_len)
6891 output[len] = *ptr;
6892 len++;
6896 /* Sometimes the path is the very last component of the IUri, so
6897 * see if the dot segments need to be reduced now.
6899 if(reduce_path && path) {
6900 DWORD current_path_len = (output+len) - path;
6901 DWORD new_path_len = remove_dot_segments(path, current_path_len);
6903 /* Update the current length. */
6904 len -= (current_path_len-new_path_len);
6907 if(len < output_len)
6908 output[len] = 0;
6909 else
6910 output[output_len-1] = 0;
6912 /* The null terminator isn't included in the length. */
6913 *result_len = len;
6914 if(len >= output_len)
6915 return STRSAFE_E_INSUFFICIENT_BUFFER;
6917 return S_OK;
6920 static HRESULT parse_friendly(IUri *uri, LPWSTR output, DWORD output_len,
6921 DWORD *result_len)
6923 HRESULT hr;
6924 DWORD display_len;
6925 BSTR display;
6927 hr = IUri_GetPropertyLength(uri, Uri_PROPERTY_DISPLAY_URI, &display_len, 0);
6928 if(FAILED(hr)) {
6929 *result_len = 0;
6930 return hr;
6933 *result_len = display_len;
6934 if(display_len+1 > output_len)
6935 return STRSAFE_E_INSUFFICIENT_BUFFER;
6937 hr = IUri_GetDisplayUri(uri, &display);
6938 if(FAILED(hr)) {
6939 *result_len = 0;
6940 return hr;
6943 memcpy(output, display, (display_len+1)*sizeof(WCHAR));
6944 SysFreeString(display);
6945 return S_OK;
6948 static HRESULT parse_rootdocument(const Uri *uri, LPWSTR output, DWORD output_len,
6949 DWORD *result_len)
6951 static const WCHAR colon_slashesW[] = {':','/','/'};
6953 WCHAR *ptr;
6954 DWORD len = 0;
6956 /* Windows only returns the root document if the URI has an authority
6957 * and it's not an unknown scheme type or a file scheme type.
6959 if(uri->authority_start == -1 ||
6960 uri->scheme_type == URL_SCHEME_UNKNOWN ||
6961 uri->scheme_type == URL_SCHEME_FILE) {
6962 *result_len = 0;
6963 if(!output_len)
6964 return STRSAFE_E_INSUFFICIENT_BUFFER;
6966 output[0] = 0;
6967 return S_OK;
6970 len = uri->scheme_len+uri->authority_len;
6971 /* For the "://" and '/' which will be added. */
6972 len += 4;
6974 if(len+1 > output_len) {
6975 *result_len = len;
6976 return STRSAFE_E_INSUFFICIENT_BUFFER;
6979 ptr = output;
6980 memcpy(ptr, uri->canon_uri+uri->scheme_start, uri->scheme_len*sizeof(WCHAR));
6982 /* Add the "://". */
6983 ptr += uri->scheme_len;
6984 memcpy(ptr, colon_slashesW, sizeof(colon_slashesW));
6986 /* Add the authority. */
6987 ptr += ARRAY_SIZE(colon_slashesW);
6988 memcpy(ptr, uri->canon_uri+uri->authority_start, uri->authority_len*sizeof(WCHAR));
6990 /* Add the '/' after the authority. */
6991 ptr += uri->authority_len;
6992 *ptr = '/';
6993 ptr[1] = 0;
6995 *result_len = len;
6996 return S_OK;
6999 static HRESULT parse_document(const Uri *uri, LPWSTR output, DWORD output_len,
7000 DWORD *result_len)
7002 DWORD len = 0;
7004 /* It has to be a known scheme type, but, it can't be a file
7005 * scheme. It also has to hierarchical.
7007 if(uri->scheme_type == URL_SCHEME_UNKNOWN ||
7008 uri->scheme_type == URL_SCHEME_FILE ||
7009 uri->authority_start == -1) {
7010 *result_len = 0;
7011 if(output_len < 1)
7012 return STRSAFE_E_INSUFFICIENT_BUFFER;
7014 output[0] = 0;
7015 return S_OK;
7018 if(uri->fragment_start > -1)
7019 len = uri->fragment_start;
7020 else
7021 len = uri->canon_len;
7023 *result_len = len;
7024 if(len+1 > output_len)
7025 return STRSAFE_E_INSUFFICIENT_BUFFER;
7027 memcpy(output, uri->canon_uri, len*sizeof(WCHAR));
7028 output[len] = 0;
7029 return S_OK;
7032 static HRESULT parse_path_from_url(const Uri *uri, LPWSTR output, DWORD output_len,
7033 DWORD *result_len)
7035 const WCHAR *path_ptr;
7036 WCHAR buffer[INTERNET_MAX_URL_LENGTH+1];
7037 WCHAR *ptr;
7039 if(uri->scheme_type != URL_SCHEME_FILE) {
7040 *result_len = 0;
7041 if(output_len > 0)
7042 output[0] = 0;
7043 return E_INVALIDARG;
7046 ptr = buffer;
7047 if(uri->host_start > -1) {
7048 static const WCHAR slash_slashW[] = {'\\','\\'};
7050 memcpy(ptr, slash_slashW, sizeof(slash_slashW));
7051 ptr += ARRAY_SIZE(slash_slashW);
7052 memcpy(ptr, uri->canon_uri+uri->host_start, uri->host_len*sizeof(WCHAR));
7053 ptr += uri->host_len;
7056 path_ptr = uri->canon_uri+uri->path_start;
7057 if(uri->path_len > 3 && *path_ptr == '/' && is_drive_path(path_ptr+1))
7058 /* Skip past the '/' in front of the drive path. */
7059 ++path_ptr;
7061 for(; path_ptr < uri->canon_uri+uri->path_start+uri->path_len; ++path_ptr, ++ptr) {
7062 BOOL do_default_action = TRUE;
7064 if(*path_ptr == '%') {
7065 const WCHAR decoded = decode_pct_val(path_ptr);
7066 if(decoded) {
7067 *ptr = decoded;
7068 path_ptr += 2;
7069 do_default_action = FALSE;
7071 } else if(*path_ptr == '/') {
7072 *ptr = '\\';
7073 do_default_action = FALSE;
7076 if(do_default_action)
7077 *ptr = *path_ptr;
7080 *ptr = 0;
7082 *result_len = ptr-buffer;
7083 if(*result_len+1 > output_len)
7084 return STRSAFE_E_INSUFFICIENT_BUFFER;
7086 memcpy(output, buffer, (*result_len+1)*sizeof(WCHAR));
7087 return S_OK;
7090 static HRESULT parse_url_from_path(IUri *uri, LPWSTR output, DWORD output_len,
7091 DWORD *result_len)
7093 HRESULT hr;
7094 BSTR received;
7095 DWORD len = 0;
7097 hr = IUri_GetPropertyLength(uri, Uri_PROPERTY_ABSOLUTE_URI, &len, 0);
7098 if(FAILED(hr)) {
7099 *result_len = 0;
7100 return hr;
7103 *result_len = len;
7104 if(len+1 > output_len)
7105 return STRSAFE_E_INSUFFICIENT_BUFFER;
7107 hr = IUri_GetAbsoluteUri(uri, &received);
7108 if(FAILED(hr)) {
7109 *result_len = 0;
7110 return hr;
7113 memcpy(output, received, (len+1)*sizeof(WCHAR));
7114 SysFreeString(received);
7116 return S_OK;
7119 static HRESULT parse_schema(IUri *uri, LPWSTR output, DWORD output_len,
7120 DWORD *result_len)
7122 HRESULT hr;
7123 DWORD len;
7124 BSTR received;
7126 hr = IUri_GetPropertyLength(uri, Uri_PROPERTY_SCHEME_NAME, &len, 0);
7127 if(FAILED(hr)) {
7128 *result_len = 0;
7129 return hr;
7132 *result_len = len;
7133 if(len+1 > output_len)
7134 return STRSAFE_E_INSUFFICIENT_BUFFER;
7136 hr = IUri_GetSchemeName(uri, &received);
7137 if(FAILED(hr)) {
7138 *result_len = 0;
7139 return hr;
7142 memcpy(output, received, (len+1)*sizeof(WCHAR));
7143 SysFreeString(received);
7145 return S_OK;
7148 static HRESULT parse_site(IUri *uri, LPWSTR output, DWORD output_len, DWORD *result_len)
7150 HRESULT hr;
7151 DWORD len;
7152 BSTR received;
7154 hr = IUri_GetPropertyLength(uri, Uri_PROPERTY_HOST, &len, 0);
7155 if(FAILED(hr)) {
7156 *result_len = 0;
7157 return hr;
7160 *result_len = len;
7161 if(len+1 > output_len)
7162 return STRSAFE_E_INSUFFICIENT_BUFFER;
7164 hr = IUri_GetHost(uri, &received);
7165 if(FAILED(hr)) {
7166 *result_len = 0;
7167 return hr;
7170 memcpy(output, received, (len+1)*sizeof(WCHAR));
7171 SysFreeString(received);
7173 return S_OK;
7176 static HRESULT parse_domain(IUri *uri, LPWSTR output, DWORD output_len, DWORD *result_len)
7178 HRESULT hr;
7179 DWORD len;
7180 BSTR received;
7182 hr = IUri_GetPropertyLength(uri, Uri_PROPERTY_DOMAIN, &len, 0);
7183 if(FAILED(hr)) {
7184 *result_len = 0;
7185 return hr;
7188 *result_len = len;
7189 if(len+1 > output_len)
7190 return STRSAFE_E_INSUFFICIENT_BUFFER;
7192 hr = IUri_GetDomain(uri, &received);
7193 if(FAILED(hr)) {
7194 *result_len = 0;
7195 return hr;
7198 memcpy(output, received, (len+1)*sizeof(WCHAR));
7199 SysFreeString(received);
7201 return S_OK;
7204 static HRESULT parse_anchor(IUri *uri, LPWSTR output, DWORD output_len, DWORD *result_len)
7206 HRESULT hr;
7207 DWORD len;
7208 BSTR received;
7210 hr = IUri_GetPropertyLength(uri, Uri_PROPERTY_FRAGMENT, &len, 0);
7211 if(FAILED(hr)) {
7212 *result_len = 0;
7213 return hr;
7216 *result_len = len;
7217 if(len+1 > output_len)
7218 return STRSAFE_E_INSUFFICIENT_BUFFER;
7220 hr = IUri_GetFragment(uri, &received);
7221 if(FAILED(hr)) {
7222 *result_len = 0;
7223 return hr;
7226 memcpy(output, received, (len+1)*sizeof(WCHAR));
7227 SysFreeString(received);
7229 return S_OK;
7232 /***********************************************************************
7233 * CoInternetParseIUri (urlmon.@)
7235 HRESULT WINAPI CoInternetParseIUri(IUri *pIUri, PARSEACTION ParseAction, DWORD dwFlags,
7236 LPWSTR pwzResult, DWORD cchResult, DWORD *pcchResult,
7237 DWORD_PTR dwReserved)
7239 HRESULT hr;
7240 Uri *uri;
7241 IInternetProtocolInfo *info;
7243 TRACE("(%p %d %x %p %d %p %x)\n", pIUri, ParseAction, dwFlags, pwzResult,
7244 cchResult, pcchResult, (DWORD)dwReserved);
7246 if(!pcchResult)
7247 return E_POINTER;
7249 if(!pwzResult || !pIUri) {
7250 *pcchResult = 0;
7251 return E_INVALIDARG;
7254 if(!(uri = get_uri_obj(pIUri))) {
7255 *pcchResult = 0;
7256 FIXME("(%p %d %x %p %d %p %x) Unknown IUri's not supported for this action.\n",
7257 pIUri, ParseAction, dwFlags, pwzResult, cchResult, pcchResult, (DWORD)dwReserved);
7258 return E_NOTIMPL;
7261 info = get_protocol_info(uri->canon_uri);
7262 if(info) {
7263 hr = IInternetProtocolInfo_ParseUrl(info, uri->canon_uri, ParseAction, dwFlags,
7264 pwzResult, cchResult, pcchResult, 0);
7265 IInternetProtocolInfo_Release(info);
7266 if(SUCCEEDED(hr)) return hr;
7269 switch(ParseAction) {
7270 case PARSE_CANONICALIZE:
7271 hr = parse_canonicalize(uri, dwFlags, pwzResult, cchResult, pcchResult);
7272 break;
7273 case PARSE_FRIENDLY:
7274 hr = parse_friendly(pIUri, pwzResult, cchResult, pcchResult);
7275 break;
7276 case PARSE_ROOTDOCUMENT:
7277 hr = parse_rootdocument(uri, pwzResult, cchResult, pcchResult);
7278 break;
7279 case PARSE_DOCUMENT:
7280 hr = parse_document(uri, pwzResult, cchResult, pcchResult);
7281 break;
7282 case PARSE_PATH_FROM_URL:
7283 hr = parse_path_from_url(uri, pwzResult, cchResult, pcchResult);
7284 break;
7285 case PARSE_URL_FROM_PATH:
7286 hr = parse_url_from_path(pIUri, pwzResult, cchResult, pcchResult);
7287 break;
7288 case PARSE_SCHEMA:
7289 hr = parse_schema(pIUri, pwzResult, cchResult, pcchResult);
7290 break;
7291 case PARSE_SITE:
7292 hr = parse_site(pIUri, pwzResult, cchResult, pcchResult);
7293 break;
7294 case PARSE_DOMAIN:
7295 hr = parse_domain(pIUri, pwzResult, cchResult, pcchResult);
7296 break;
7297 case PARSE_LOCATION:
7298 case PARSE_ANCHOR:
7299 hr = parse_anchor(pIUri, pwzResult, cchResult, pcchResult);
7300 break;
7301 case PARSE_SECURITY_URL:
7302 case PARSE_MIME:
7303 case PARSE_SERVER:
7304 case PARSE_SECURITY_DOMAIN:
7305 *pcchResult = 0;
7306 hr = E_FAIL;
7307 break;
7308 default:
7309 *pcchResult = 0;
7310 hr = E_NOTIMPL;
7311 FIXME("(%p %d %x %p %d %p %x) Partial stub.\n", pIUri, ParseAction, dwFlags,
7312 pwzResult, cchResult, pcchResult, (DWORD)dwReserved);
7315 return hr;