1/*
2 * Copyright (C) 2003, 2004, 2005, 2006, 2007, 2008, 2011, 2012, 2013 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 <wtf/Forward.h>
30#include <wtf/HashMap.h>
31#include <wtf/RetainPtr.h>
32#include <wtf/text/WTFString.h>
33
34#if USE(CF)
35typedef const struct __CFURL* CFURLRef;
36#endif
37
38#if PLATFORM(MAC)
39OBJC_CLASS NSURL;
40#endif
41
42#if PLATFORM(QT)
43QT_BEGIN_NAMESPACE
44class QUrl;
45QT_END_NAMESPACE
46#endif
47
48namespace WebCore {
49
50class TextEncoding;
51struct KURLHash;
52
53enum ParsedURLStringTag { ParsedURLString };
54
55class KURL {
56public:
57    // Generates a URL which contains a null string.
58    KURL() { invalidate(); }
59
60    // The argument is an absolute URL string. The string is assumed to be output of KURL::string() called on a valid
61    // KURL object, or indiscernible from such.
62    // It is usually best to avoid repeatedly parsing a string, unless memory saving outweigh the possible slow-downs.
63    KURL(ParsedURLStringTag, const String&);
64    explicit KURL(WTF::HashTableDeletedValueType) : m_string(WTF::HashTableDeletedValue) { }
65    bool isHashTableDeletedValue() const { return string().isHashTableDeletedValue(); }
66
67    // Resolves the relative URL with the given base URL. If provided, the
68    // TextEncoding is used to encode non-ASCII characers. The base URL can be
69    // null or empty, in which case the relative URL will be interpreted as
70    // absolute.
71    // FIXME: If the base URL is invalid, this always creates an invalid
72    // URL. Instead I think it would be better to treat all invalid base URLs
73    // the same way we treate null and empty base URLs.
74    KURL(const KURL& base, const String& relative);
75    KURL(const KURL& base, const String& relative, const TextEncoding&);
76
77    String strippedForUseAsReferrer() const;
78
79    // FIXME: The above functions should be harmonized so that passing a
80    // base of null or the empty string gives the same result as the
81    // standard String constructor.
82
83    // Makes a deep copy. Helpful only if you need to use a KURL on another
84    // thread.  Since the underlying StringImpl objects are immutable, there's
85    // no other reason to ever prefer copy() over plain old assignment.
86    KURL copy() const;
87
88    bool isNull() const;
89    bool isEmpty() const;
90    bool isValid() const;
91
92    // Returns true if this URL has a path. Note that "http://foo.com/" has a
93    // path of "/", so this function will return true. Only invalid or
94    // non-hierarchical (like "javascript:") URLs will have no path.
95    bool hasPath() const;
96
97    // Returns true if you can set the host and port for the URL.
98    // Non-hierarchical URLs don't have a host and port.
99    bool canSetHostOrPort() const { return isHierarchical(); }
100
101    bool canSetPathname() const { return isHierarchical(); }
102    bool isHierarchical() const;
103
104    const String& string() const { return m_string; }
105
106    String stringCenterEllipsizedToLength(unsigned length = 1024) const;
107
108    String protocol() const;
109    String host() const;
110    unsigned short port() const;
111    bool hasPort() const;
112    String user() const;
113    String pass() const;
114    String path() const;
115    String lastPathComponent() const;
116    String query() const;
117    String fragmentIdentifier() const;
118    bool hasFragmentIdentifier() const;
119
120    String baseAsString() const;
121
122    String fileSystemPath() const;
123
124    // Returns true if the current URL's protocol is the same as the null-
125    // terminated ASCII argument. The argument must be lower-case.
126    bool protocolIs(const char*) const;
127    bool protocolIsData() const { return protocolIs("data"); }
128    bool protocolIsInHTTPFamily() const;
129    bool isLocalFile() const;
130    bool isBlankURL() const;
131
132    bool setProtocol(const String&);
133    void setHost(const String&);
134
135    void removePort();
136    void setPort(unsigned short);
137
138    // Input is like "foo.com" or "foo.com:8000".
139    void setHostAndPort(const String&);
140
141    void setUser(const String&);
142    void setPass(const String&);
143
144    // If you pass an empty path for HTTP or HTTPS URLs, the resulting path
145    // will be "/".
146    void setPath(const String&);
147
148    // The query may begin with a question mark, or, if not, one will be added
149    // for you. Setting the query to the empty string will leave a "?" in the
150    // URL (with nothing after it). To clear the query, pass a null string.
151    void setQuery(const String&);
152
153    void setFragmentIdentifier(const String&);
154    void removeFragmentIdentifier();
155
156    friend bool equalIgnoringFragmentIdentifier(const KURL&, const KURL&);
157
158    friend bool protocolHostAndPortAreEqual(const KURL&, const KURL&);
159
160    unsigned hostStart() const;
161    unsigned hostEnd() const;
162
163    unsigned pathStart() const;
164    unsigned pathEnd() const;
165    unsigned pathAfterLastSlash() const;
166
167    operator const String&() const { return string(); }
168
169#if USE(CF)
170    KURL(CFURLRef);
171    RetainPtr<CFURLRef> createCFURL() const;
172#endif
173
174#if PLATFORM(MAC)
175    KURL(NSURL*);
176    operator NSURL*() const;
177#endif
178#ifdef __OBJC__
179    operator NSString*() const { return string(); }
180#endif
181
182#if PLATFORM(QT)
183    KURL(const QUrl&);
184    operator QUrl() const;
185#endif
186
187    const KURL* innerURL() const { return 0; }
188
189#ifndef NDEBUG
190    void print() const;
191#endif
192
193    bool isSafeToSendToAnotherThread() const;
194
195private:
196    void invalidate();
197    static bool protocolIs(const String&, const char*);
198    void init(const KURL&, const String&, const TextEncoding&);
199    void copyToBuffer(Vector<char, 512>& buffer) const;
200
201    // Parses the given URL. The originalString parameter allows for an
202    // optimization: When the source is the same as the fixed-up string,
203    // it will use the passed-in string instead of allocating a new one.
204    void parse(const String&);
205    void parse(const char* url, const String* originalString = 0);
206
207    String m_string;
208    bool m_isValid : 1;
209    bool m_protocolIsInHTTPFamily : 1;
210
211    int m_schemeEnd;
212    int m_userStart;
213    int m_userEnd;
214    int m_passwordEnd;
215    int m_hostEnd;
216    int m_portEnd;
217    int m_pathAfterLastSlash;
218    int m_pathEnd;
219    int m_queryEnd;
220    int m_fragmentEnd;
221};
222
223bool operator==(const KURL&, const KURL&);
224bool operator==(const KURL&, const String&);
225bool operator==(const String&, const KURL&);
226bool operator!=(const KURL&, const KURL&);
227bool operator!=(const KURL&, const String&);
228bool operator!=(const String&, const KURL&);
229
230bool equalIgnoringFragmentIdentifier(const KURL&, const KURL&);
231bool protocolHostAndPortAreEqual(const KURL&, const KURL&);
232
233const KURL& blankURL();
234
235// Functions to do URL operations on strings.
236// These are operations that aren't faster on a parsed URL.
237// These are also different from the KURL functions in that they don't require the string to be a valid and parsable URL.
238// This is especially important because valid javascript URLs are not necessarily considered valid by KURL.
239
240bool protocolIs(const String& url, const char* protocol);
241bool protocolIsJavaScript(const String& url);
242
243bool isDefaultPortForProtocol(unsigned short port, const String& protocol);
244bool portAllowed(const KURL&); // Blacklist ports that should never be used for Web resources.
245
246bool isValidProtocol(const String&);
247
248String mimeTypeFromDataURL(const String& url);
249String mimeTypeFromURL(const KURL&);
250
251// Unescapes the given string using URL escaping rules, given an optional
252// encoding (defaulting to UTF-8 otherwise). DANGER: If the URL has "%00"
253// in it, the resulting string will have embedded null characters!
254String decodeURLEscapeSequences(const String&);
255String decodeURLEscapeSequences(const String&, const TextEncoding&);
256
257String encodeWithURLEscapeSequences(const String&);
258
259// Inlines.
260
261inline bool operator==(const KURL& a, const KURL& b)
262{
263    return a.string() == b.string();
264}
265
266inline bool operator==(const KURL& a, const String& b)
267{
268    return a.string() == b;
269}
270
271inline bool operator==(const String& a, const KURL& b)
272{
273    return a == b.string();
274}
275
276inline bool operator!=(const KURL& a, const KURL& b)
277{
278    return a.string() != b.string();
279}
280
281inline bool operator!=(const KURL& a, const String& b)
282{
283    return a.string() != b;
284}
285
286inline bool operator!=(const String& a, const KURL& b)
287{
288    return a != b.string();
289}
290
291// Inline versions of some non-GoogleURL functions so we can get inlining
292// without having to have a lot of ugly ifdefs in the class definition.
293
294inline bool KURL::isNull() const
295{
296    return m_string.isNull();
297}
298
299inline bool KURL::isEmpty() const
300{
301    return m_string.isEmpty();
302}
303
304inline bool KURL::isValid() const
305{
306    return m_isValid;
307}
308
309inline bool KURL::hasPath() const
310{
311    return m_pathEnd != m_portEnd;
312}
313
314inline bool KURL::hasPort() const
315{
316    return m_hostEnd < m_portEnd;
317}
318
319inline bool KURL::protocolIsInHTTPFamily() const
320{
321    return m_protocolIsInHTTPFamily;
322}
323
324inline unsigned KURL::hostStart() const
325{
326    return (m_passwordEnd == m_userStart) ? m_passwordEnd : m_passwordEnd + 1;
327}
328
329inline unsigned KURL::hostEnd() const
330{
331    return m_hostEnd;
332}
333
334inline unsigned KURL::pathStart() const
335{
336    return m_portEnd;
337}
338
339inline unsigned KURL::pathEnd() const
340{
341    return m_pathEnd;
342}
343
344inline unsigned KURL::pathAfterLastSlash() const
345{
346    return m_pathAfterLastSlash;
347}
348
349} // namespace WebCore
350
351namespace WTF {
352
353    // KURLHash is the default hash for String
354    template<typename T> struct DefaultHash;
355    template<> struct DefaultHash<WebCore::KURL> {
356        typedef WebCore::KURLHash Hash;
357    };
358
359} // namespace WTF
360
361#endif // KURL_h
362