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)
34typedef const struct __CFURL* CFURLRef;
35#endif
36
37#if PLATFORM(MAC)
38#ifdef __OBJC__
39@class NSURL;
40#else
41class NSURL;
42#endif
43#endif
44
45#if PLATFORM(QT)
46QT_BEGIN_NAMESPACE
47class QUrl;
48QT_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
59namespace WebCore {
60
61class TextEncoding;
62struct KURLHash;
63
64typedef HashMap<String, String> ParsedURLParameters;
65
66enum ParsedURLStringTag { ParsedURLString };
67
68class KURL {
69public:
70    // Generates a URL which contains a null string.
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)
80    KURL(WTF::HashTableDeletedValueType) : m_url(WTF::HashTableDeletedValue) { }
81#else
82    KURL(WTF::HashTableDeletedValueType) : m_string(WTF::HashTableDeletedValue) { }
83#endif
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.
127    bool canSetHostOrPort() const { return isHierarchical(); }
128
129    bool canSetPathname() const { return isHierarchical(); }
130
131#if USE(GOOGLEURL)
132    const String& string() const { return m_url.string(); }
133    URLString urlString() const { return URLString(m_url.string()); }
134#else
135    const String& string() const { return m_string; }
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    // This function is only used by location.href. It's likely we shouldn't
156    // use it for that purpose, but more study is necessary before we remove it.
157    String deprecatedString() const;
158    String fileSystemPath() const;
159
160    // Returns true if the current URL's protocol is the same as the null-
161    // terminated ASCII argument. The argument must be lower-case.
162    bool protocolIs(const char*) const;
163    bool protocolIsData() const { return protocolIs("data"); }
164    bool protocolInHTTPFamily() const;
165    bool isLocalFile() const;
166
167    bool setProtocol(const String&);
168    void setHost(const String&);
169
170    void removePort();
171    void setPort(unsigned short);
172
173    // Input is like "foo.com" or "foo.com:8000".
174    void setHostAndPort(const String&);
175
176    void setUser(const String&);
177    void setPass(const String&);
178
179    // If you pass an empty path for HTTP or HTTPS URLs, the resulting path
180    // will be "/".
181    void setPath(const String&);
182
183    // The query may begin with a question mark, or, if not, one will be added
184    // for you. Setting the query to the empty string will leave a "?" in the
185    // URL (with nothing after it). To clear the query, pass a null string.
186    void setQuery(const String&);
187
188    void setFragmentIdentifier(const String&);
189    void removeFragmentIdentifier();
190
191    friend bool equalIgnoringFragmentIdentifier(const KURL&, const KURL&);
192
193    friend bool protocolHostAndPortAreEqual(const KURL&, const KURL&);
194
195    unsigned hostStart() const;
196    unsigned hostEnd() const;
197
198    unsigned pathStart() const;
199    unsigned pathEnd() const;
200    unsigned pathAfterLastSlash() const;
201    operator const String&() const { return string(); }
202
203#if USE(CF)
204    KURL(CFURLRef);
205    CFURLRef createCFURL() const;
206#endif
207
208#if PLATFORM(MAC)
209    KURL(NSURL*);
210    operator NSURL*() const;
211#endif
212#ifdef __OBJC__
213    operator NSString*() const { return string(); }
214#endif
215
216#if PLATFORM(QT)
217    KURL(const QUrl&);
218    operator QUrl() const;
219#endif
220
221#if USE(GOOGLEURL)
222    // Getters for the parsed structure and its corresponding 8-bit string.
223    const url_parse::Parsed& parsed() const { return m_url.m_parsed; }
224    const CString& utf8String() const { return m_url.utf8String(); }
225#endif
226
227#ifndef NDEBUG
228    void print() const;
229#endif
230
231private:
232    void invalidate();
233    bool isHierarchical() const;
234    static bool protocolIs(const String&, const char*);
235#if USE(GOOGLEURL)
236    friend class KURLGooglePrivate;
237    void parse(const char* url, const String* originalString);  // KURLMac calls this.
238    void copyToBuffer(Vector<char, 512>& buffer) const;  // KURLCFNet uses this.
239    KURLGooglePrivate m_url;
240#else  // !USE(GOOGLEURL)
241    void init(const KURL&, const String&, const TextEncoding&);
242    void copyToBuffer(Vector<char, 512>& buffer) const;
243
244    // Parses the given URL. The originalString parameter allows for an
245    // optimization: When the source is the same as the fixed-up string,
246    // it will use the passed-in string instead of allocating a new one.
247    void parse(const String&);
248    void parse(const char* url, const String* originalString);
249
250    String m_string;
251    bool m_isValid : 1;
252    bool m_protocolInHTTPFamily : 1;
253
254    int m_schemeEnd;
255    int m_userStart;
256    int m_userEnd;
257    int m_passwordEnd;
258    int m_hostEnd;
259    int m_portEnd;
260    int m_pathAfterLastSlash;
261    int m_pathEnd;
262    int m_queryEnd;
263    int m_fragmentEnd;
264#endif
265};
266
267bool operator==(const KURL&, const KURL&);
268bool operator==(const KURL&, const String&);
269bool operator==(const String&, const KURL&);
270bool operator!=(const KURL&, const KURL&);
271bool operator!=(const KURL&, const String&);
272bool operator!=(const String&, const KURL&);
273
274bool equalIgnoringFragmentIdentifier(const KURL&, const KURL&);
275bool protocolHostAndPortAreEqual(const KURL&, const KURL&);
276
277const KURL& blankURL();
278
279// Functions to do URL operations on strings.
280// These are operations that aren't faster on a parsed URL.
281// These are also different from the KURL functions in that they don't require the string to be a valid and parsable URL.
282// This is especially important because valid javascript URLs are not necessarily considered valid by KURL.
283
284bool protocolIs(const String& url, const char* protocol);
285bool protocolIsJavaScript(const String& url);
286
287bool isDefaultPortForProtocol(unsigned short port, const String& protocol);
288bool portAllowed(const KURL&); // Blacklist ports that should never be used for Web resources.
289
290bool isValidProtocol(const String&);
291
292String mimeTypeFromDataURL(const String& url);
293
294// Unescapes the given string using URL escaping rules, given an optional
295// encoding (defaulting to UTF-8 otherwise). DANGER: If the URL has "%00"
296// in it, the resulting string will have embedded null characters!
297String decodeURLEscapeSequences(const String&);
298String decodeURLEscapeSequences(const String&, const TextEncoding&);
299
300String encodeWithURLEscapeSequences(const String&);
301
302// Inlines.
303
304inline bool operator==(const KURL& a, const KURL& b)
305{
306    return a.string() == b.string();
307}
308
309inline bool operator==(const KURL& a, const String& b)
310{
311    return a.string() == b;
312}
313
314inline bool operator==(const String& a, const KURL& b)
315{
316    return a == b.string();
317}
318
319inline bool operator!=(const KURL& a, const KURL& b)
320{
321    return a.string() != b.string();
322}
323
324inline bool operator!=(const KURL& a, const String& b)
325{
326    return a.string() != b;
327}
328
329inline bool operator!=(const String& a, const KURL& b)
330{
331    return a != b.string();
332}
333
334#if !USE(GOOGLEURL)
335
336// Inline versions of some non-GoogleURL functions so we can get inlining
337// without having to have a lot of ugly ifdefs in the class definition.
338
339inline bool KURL::isNull() const
340{
341    return m_string.isNull();
342}
343
344inline bool KURL::isEmpty() const
345{
346    return m_string.isEmpty();
347}
348
349inline bool KURL::isValid() const
350{
351    return m_isValid;
352}
353
354inline bool KURL::hasPort() const
355{
356    return m_hostEnd < m_portEnd;
357}
358
359inline bool KURL::protocolInHTTPFamily() const
360{
361    return m_protocolInHTTPFamily;
362}
363
364inline unsigned KURL::hostStart() const
365{
366    return (m_passwordEnd == m_userStart) ? m_passwordEnd : m_passwordEnd + 1;
367}
368
369inline unsigned KURL::hostEnd() const
370{
371    return m_hostEnd;
372}
373
374inline unsigned KURL::pathStart() const
375{
376    return m_portEnd;
377}
378
379inline unsigned KURL::pathEnd() const
380{
381    return m_pathEnd;
382}
383
384inline unsigned KURL::pathAfterLastSlash() const
385{
386    return m_pathAfterLastSlash;
387}
388
389#endif  // !USE(GOOGLEURL)
390
391} // namespace WebCore
392
393namespace WTF {
394
395    // KURLHash is the default hash for String
396    template<typename T> struct DefaultHash;
397    template<> struct DefaultHash<WebCore::KURL> {
398        typedef WebCore::KURLHash Hash;
399    };
400
401} // namespace WTF
402
403#endif // KURL_h
404