• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1 /*
2  * Copyright (C) 2003, 2004, 2005, 2006, 2007, 2008 Apple Inc. All rights reserved.
3  *
4  * Redistribution and use in source and binary forms, with or without
5  * modification, are permitted provided that the following conditions
6  * are met:
7  * 1. Redistributions of source code must retain the above copyright
8  *    notice, this list of conditions and the following disclaimer.
9  * 2. Redistributions in binary form must reproduce the above copyright
10  *    notice, this list of conditions and the following disclaimer in the
11  *    documentation and/or other materials provided with the distribution.
12  *
13  * THIS SOFTWARE IS PROVIDED BY APPLE COMPUTER, INC. ``AS IS'' AND ANY
14  * EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
15  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
16  * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL APPLE COMPUTER, INC. OR
17  * CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
18  * EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
19  * PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
20  * PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
21  * OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
22  * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
23  * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
24  */
25 
26 #ifndef KURL_h
27 #define KURL_h
28 
29 #include "PlatformString.h"
30 #include "URLString.h"
31 #include <wtf/HashMap.h>
32 
33 #if USE(CF)
34 typedef const struct __CFURL* CFURLRef;
35 #endif
36 
37 #if PLATFORM(MAC)
38 #ifdef __OBJC__
39 @class NSURL;
40 #else
41 class NSURL;
42 #endif
43 #endif
44 
45 #if PLATFORM(QT)
46 QT_BEGIN_NAMESPACE
47 class QUrl;
48 QT_END_NAMESPACE
49 #endif
50 
51 #if USE(GOOGLEURL)
52 #include "KURLGooglePrivate.h"
53 #endif
54 
55 #if USE(JSC)
56 #include <runtime/UString.h>
57 #endif
58 
59 namespace WebCore {
60 
61 class TextEncoding;
62 struct KURLHash;
63 
64 typedef HashMap<String, String> ParsedURLParameters;
65 
66 enum ParsedURLStringTag { ParsedURLString };
67 
68 class KURL {
69 public:
70     // Generates a URL which contains a null string.
KURL()71     KURL() { invalidate(); }
72 
73     // The argument is an absolute URL string. The string is assumed to be output of KURL::string() called on a valid
74     // KURL object, or indiscernible from such.
75     // It is usually best to avoid repeatedly parsing a string, unless memory saving outweigh the possible slow-downs.
76     KURL(ParsedURLStringTag, const char*);
77     KURL(ParsedURLStringTag, const String&);
78     KURL(ParsedURLStringTag, const URLString&);
79 #if USE(GOOGLEURL)
KURL(WTF::HashTableDeletedValueType)80     KURL(WTF::HashTableDeletedValueType) : m_url(WTF::HashTableDeletedValue) { }
81 #else
KURL(WTF::HashTableDeletedValueType)82     KURL(WTF::HashTableDeletedValueType) : m_string(WTF::HashTableDeletedValue) { }
83 #endif
isHashTableDeletedValue()84     bool isHashTableDeletedValue() const { return string().isHashTableDeletedValue(); }
85 
86     // Resolves the relative URL with the given base URL. If provided, the
87     // TextEncoding is used to encode non-ASCII characers. The base URL can be
88     // null or empty, in which case the relative URL will be interpreted as
89     // absolute.
90     // FIXME: If the base URL is invalid, this always creates an invalid
91     // URL. Instead I think it would be better to treat all invalid base URLs
92     // the same way we treate null and empty base URLs.
93     KURL(const KURL& base, const String& relative);
94     KURL(const KURL& base, const String& relative, const TextEncoding&);
95 
96 
97 #if USE(GOOGLEURL)
98     // For conversions for other structures that have already parsed and
99     // canonicalized the URL. The input must be exactly what KURL would have
100     // done with the same input.
101     KURL(const CString& canonicalSpec,
102          const url_parse::Parsed& parsed, bool isValid);
103 #endif
104 
105     String strippedForUseAsReferrer() const;
106 
107     // FIXME: The above functions should be harmonized so that passing a
108     // base of null or the empty string gives the same result as the
109     // standard String constructor.
110 
111     // Makes a deep copy. Helpful only if you need to use a KURL on another
112     // thread.  Since the underlying StringImpl objects are immutable, there's
113     // no other reason to ever prefer copy() over plain old assignment.
114     KURL copy() const;
115 
116     bool isNull() const;
117     bool isEmpty() const;
118     bool isValid() const;
119 
120     // Returns true if this URL has a path. Note that "http://foo.com/" has a
121     // path of "/", so this function will return true. Only invalid or
122     // non-hierarchical (like "javascript:") URLs will have no path.
123     bool hasPath() const;
124 
125     // Returns true if you can set the host and port for the URL.
126     // Non-hierarchical URLs don't have a host and port.
canSetHostOrPort()127     bool canSetHostOrPort() const { return isHierarchical(); }
128 
canSetPathname()129     bool canSetPathname() const { return isHierarchical(); }
130 
131 #if USE(GOOGLEURL)
string()132     const String& string() const { return m_url.string(); }
urlString()133     URLString urlString() const { return URLString(m_url.string()); }
134 #else
string()135     const String& string() const { return m_string; }
urlString()136     URLString urlString() const { return URLString(m_string); }
137 #endif
138 
139     String protocol() const;
140     String host() const;
141     unsigned short port() const;
142     bool hasPort() const;
143     String user() const;
144     String pass() const;
145     String path() const;
146     String lastPathComponent() const;
147     String query() const;
148     String fragmentIdentifier() const;
149     bool hasFragmentIdentifier() const;
150 
151     void copyParsedQueryTo(ParsedURLParameters&) const;
152 
153     String baseAsString() const;
154 
155     String prettyURL() const;
156     String fileSystemPath() const;
157 
158     // Returns true if the current URL's protocol is the same as the null-
159     // terminated ASCII argument. The argument must be lower-case.
160     bool protocolIs(const char*) const;
protocolIsData()161     bool protocolIsData() const { return protocolIs("data"); }
162     bool protocolInHTTPFamily() const;
163     bool isLocalFile() const;
164 
165     bool setProtocol(const String&);
166     void setHost(const String&);
167 
168     void removePort();
169     void setPort(unsigned short);
170 
171     // Input is like "foo.com" or "foo.com:8000".
172     void setHostAndPort(const String&);
173 
174     void setUser(const String&);
175     void setPass(const String&);
176 
177     // If you pass an empty path for HTTP or HTTPS URLs, the resulting path
178     // will be "/".
179     void setPath(const String&);
180 
181     // The query may begin with a question mark, or, if not, one will be added
182     // for you. Setting the query to the empty string will leave a "?" in the
183     // URL (with nothing after it). To clear the query, pass a null string.
184     void setQuery(const String&);
185 
186     void setFragmentIdentifier(const String&);
187     void removeFragmentIdentifier();
188 
189     friend bool equalIgnoringFragmentIdentifier(const KURL&, const KURL&);
190 
191     friend bool protocolHostAndPortAreEqual(const KURL&, const KURL&);
192 
193     unsigned hostStart() const;
194     unsigned hostEnd() const;
195 
196     unsigned pathStart() const;
197     unsigned pathEnd() const;
198     unsigned pathAfterLastSlash() const;
199     operator const String&() const { return string(); }
200 
201 #if USE(CF)
202     KURL(CFURLRef);
203     CFURLRef createCFURL() const;
204 #endif
205 
206 #if PLATFORM(MAC)
207     KURL(NSURL*);
208     operator NSURL*() const;
209 #endif
210 #ifdef __OBJC__
211     operator NSString*() const { return string(); }
212 #endif
213 
214 #if PLATFORM(QT)
215     KURL(const QUrl&);
216     operator QUrl() const;
217 #endif
218 
219 #if USE(GOOGLEURL)
220     // Getters for the parsed structure and its corresponding 8-bit string.
parsed()221     const url_parse::Parsed& parsed() const { return m_url.m_parsed; }
utf8String()222     const CString& utf8String() const { return m_url.utf8String(); }
223 #endif
224 
225 #ifndef NDEBUG
226     void print() const;
227 #endif
228 
229 private:
230     void invalidate();
231     bool isHierarchical() const;
232     static bool protocolIs(const String&, const char*);
233 #if USE(GOOGLEURL)
234     friend class KURLGooglePrivate;
235     void parse(const char* url, const String* originalString);  // KURLMac calls this.
236     void copyToBuffer(Vector<char, 512>& buffer) const;  // KURLCFNet uses this.
237     KURLGooglePrivate m_url;
238 #else  // !USE(GOOGLEURL)
239     void init(const KURL&, const String&, const TextEncoding&);
240     void copyToBuffer(Vector<char, 512>& buffer) const;
241 
242     // Parses the given URL. The originalString parameter allows for an
243     // optimization: When the source is the same as the fixed-up string,
244     // it will use the passed-in string instead of allocating a new one.
245     void parse(const String&);
246     void parse(const char* url, const String* originalString);
247 
248     String m_string;
249     bool m_isValid : 1;
250     bool m_protocolInHTTPFamily : 1;
251 
252     int m_schemeEnd;
253     int m_userStart;
254     int m_userEnd;
255     int m_passwordEnd;
256     int m_hostEnd;
257     int m_portEnd;
258     int m_pathAfterLastSlash;
259     int m_pathEnd;
260     int m_queryEnd;
261     int m_fragmentEnd;
262 #endif
263 };
264 
265 bool operator==(const KURL&, const KURL&);
266 bool operator==(const KURL&, const String&);
267 bool operator==(const String&, const KURL&);
268 bool operator!=(const KURL&, const KURL&);
269 bool operator!=(const KURL&, const String&);
270 bool operator!=(const String&, const KURL&);
271 
272 bool equalIgnoringFragmentIdentifier(const KURL&, const KURL&);
273 bool protocolHostAndPortAreEqual(const KURL&, const KURL&);
274 
275 const KURL& blankURL();
276 
277 // Functions to do URL operations on strings.
278 // These are operations that aren't faster on a parsed URL.
279 // These are also different from the KURL functions in that they don't require the string to be a valid and parsable URL.
280 // This is especially important because valid javascript URLs are not necessarily considered valid by KURL.
281 
282 bool protocolIs(const String& url, const char* protocol);
283 bool protocolIsJavaScript(const String& url);
284 
285 bool isDefaultPortForProtocol(unsigned short port, const String& protocol);
286 bool portAllowed(const KURL&); // Blacklist ports that should never be used for Web resources.
287 
288 bool isValidProtocol(const String&);
289 
290 String mimeTypeFromDataURL(const String& url);
291 
292 // Unescapes the given string using URL escaping rules, given an optional
293 // encoding (defaulting to UTF-8 otherwise). DANGER: If the URL has "%00"
294 // in it, the resulting string will have embedded null characters!
295 String decodeURLEscapeSequences(const String&);
296 String decodeURLEscapeSequences(const String&, const TextEncoding&);
297 
298 String encodeWithURLEscapeSequences(const String&);
299 
300 // Inlines.
301 
302 inline bool operator==(const KURL& a, const KURL& b)
303 {
304     return a.string() == b.string();
305 }
306 
307 inline bool operator==(const KURL& a, const String& b)
308 {
309     return a.string() == b;
310 }
311 
312 inline bool operator==(const String& a, const KURL& b)
313 {
314     return a == b.string();
315 }
316 
317 inline bool operator!=(const KURL& a, const KURL& b)
318 {
319     return a.string() != b.string();
320 }
321 
322 inline bool operator!=(const KURL& a, const String& b)
323 {
324     return a.string() != b;
325 }
326 
327 inline bool operator!=(const String& a, const KURL& b)
328 {
329     return a != b.string();
330 }
331 
332 #if !USE(GOOGLEURL)
333 
334 // Inline versions of some non-GoogleURL functions so we can get inlining
335 // without having to have a lot of ugly ifdefs in the class definition.
336 
isNull()337 inline bool KURL::isNull() const
338 {
339     return m_string.isNull();
340 }
341 
isEmpty()342 inline bool KURL::isEmpty() const
343 {
344     return m_string.isEmpty();
345 }
346 
isValid()347 inline bool KURL::isValid() const
348 {
349     return m_isValid;
350 }
351 
hasPort()352 inline bool KURL::hasPort() const
353 {
354     return m_hostEnd < m_portEnd;
355 }
356 
protocolInHTTPFamily()357 inline bool KURL::protocolInHTTPFamily() const
358 {
359     return m_protocolInHTTPFamily;
360 }
361 
hostStart()362 inline unsigned KURL::hostStart() const
363 {
364     return (m_passwordEnd == m_userStart) ? m_passwordEnd : m_passwordEnd + 1;
365 }
366 
hostEnd()367 inline unsigned KURL::hostEnd() const
368 {
369     return m_hostEnd;
370 }
371 
pathStart()372 inline unsigned KURL::pathStart() const
373 {
374     return m_portEnd;
375 }
376 
pathEnd()377 inline unsigned KURL::pathEnd() const
378 {
379     return m_pathEnd;
380 }
381 
pathAfterLastSlash()382 inline unsigned KURL::pathAfterLastSlash() const
383 {
384     return m_pathAfterLastSlash;
385 }
386 
387 #endif  // !USE(GOOGLEURL)
388 
389 } // namespace WebCore
390 
391 namespace WTF {
392 
393     // KURLHash is the default hash for String
394     template<typename T> struct DefaultHash;
395     template<> struct DefaultHash<WebCore::KURL> {
396         typedef WebCore::KURLHash Hash;
397     };
398 
399 } // namespace WTF
400 
401 #endif // KURL_h
402