1// Copyright 2013 The Chromium Authors. All rights reserved.
2// Use of this source code is governed by a BSD-style license that can be
3// found in the LICENSE file.
4
5#ifndef URL_URL_UTIL_H_
6#define URL_URL_UTIL_H_
7
8#include <string>
9
10#include "base/strings/string16.h"
11#include "url/url_canon.h"
12#include "url/url_constants.h"
13#include "url/url_export.h"
14#include "url/url_parse.h"
15
16namespace url {
17
18// Init ------------------------------------------------------------------------
19
20// Initialization is NOT required, it will be implicitly initialized when first
21// used. However, this implicit initialization is NOT threadsafe. If you are
22// using this library in a threaded environment and don't have a consistent
23// "first call" (an example might be calling "AddStandardScheme" with your
24// special application-specific schemes) then you will want to call initialize
25// before spawning any threads.
26//
27// It is OK to call this function more than once, subsequent calls will simply
28// "noop", unless Shutdown() was called in the mean time. This will also be a
29// "noop" if other calls to the library have forced an initialization
30// beforehand.
31URL_EXPORT void Initialize();
32
33// Cleanup is not required, except some strings may leak. For most user
34// applications, this is fine. If you're using it in a library that may get
35// loaded and unloaded, you'll want to unload to properly clean up your
36// library.
37URL_EXPORT void Shutdown();
38
39// Schemes --------------------------------------------------------------------
40
41// Adds an application-defined scheme to the internal list of "standard" URL
42// schemes. This function is not threadsafe and can not be called concurrently
43// with any other url_util function. It will assert if the list of standard
44// schemes has been locked (see LockStandardSchemes).
45URL_EXPORT void AddStandardScheme(const char* new_scheme);
46
47// Sets a flag to prevent future calls to AddStandardScheme from succeeding.
48//
49// This is designed to help prevent errors for multithreaded applications.
50// Normal usage would be to call AddStandardScheme for your custom schemes at
51// the beginning of program initialization, and then LockStandardSchemes. This
52// prevents future callers from mistakenly calling AddStandardScheme when the
53// program is running with multiple threads, where such usage would be
54// dangerous.
55//
56// We could have had AddStandardScheme use a lock instead, but that would add
57// some platform-specific dependencies we don't otherwise have now, and is
58// overkill considering the normal usage is so simple.
59URL_EXPORT void LockStandardSchemes();
60
61// Locates the scheme in the given string and places it into |found_scheme|,
62// which may be NULL to indicate the caller does not care about the range.
63//
64// Returns whether the given |compare| scheme matches the scheme found in the
65// input (if any). The |compare| scheme must be a valid canonical scheme or
66// the result of the comparison is undefined.
67URL_EXPORT bool FindAndCompareScheme(const char* str,
68                                     int str_len,
69                                     const char* compare,
70                                     Component* found_scheme);
71URL_EXPORT bool FindAndCompareScheme(const base::char16* str,
72                                     int str_len,
73                                     const char* compare,
74                                     Component* found_scheme);
75inline bool FindAndCompareScheme(const std::string& str,
76                                 const char* compare,
77                                 Component* found_scheme) {
78  return FindAndCompareScheme(str.data(), static_cast<int>(str.size()),
79                              compare, found_scheme);
80}
81inline bool FindAndCompareScheme(const base::string16& str,
82                                 const char* compare,
83                                 Component* found_scheme) {
84  return FindAndCompareScheme(str.data(), static_cast<int>(str.size()),
85                              compare, found_scheme);
86}
87
88// Returns true if the given string represents a standard URL. This means that
89// either the scheme is in the list of known standard schemes.
90URL_EXPORT bool IsStandard(const char* spec, const Component& scheme);
91URL_EXPORT bool IsStandard(const base::char16* spec, const Component& scheme);
92
93// TODO(brettw) remove this. This is a temporary compatibility hack to avoid
94// breaking the WebKit build when this version is synced via Chrome.
95inline bool IsStandard(const char* spec,
96                       int spec_len,
97                       const Component& scheme) {
98  return IsStandard(spec, scheme);
99}
100
101// URL library wrappers -------------------------------------------------------
102
103// Parses the given spec according to the extracted scheme type. Normal users
104// should use the URL object, although this may be useful if performance is
105// critical and you don't want to do the heap allocation for the std::string.
106//
107// As with the Canonicalize* functions, the charset converter can
108// be NULL to use UTF-8 (it will be faster in this case).
109//
110// Returns true if a valid URL was produced, false if not. On failure, the
111// output and parsed structures will still be filled and will be consistent,
112// but they will not represent a loadable URL.
113URL_EXPORT bool Canonicalize(const char* spec,
114                             int spec_len,
115                             bool trim_path_end,
116                             CharsetConverter* charset_converter,
117                             CanonOutput* output,
118                             Parsed* output_parsed);
119URL_EXPORT bool Canonicalize(const base::char16* spec,
120                             int spec_len,
121                             bool trim_path_end,
122                             CharsetConverter* charset_converter,
123                             CanonOutput* output,
124                             Parsed* output_parsed);
125
126// Resolves a potentially relative URL relative to the given parsed base URL.
127// The base MUST be valid. The resulting canonical URL and parsed information
128// will be placed in to the given out variables.
129//
130// The relative need not be relative. If we discover that it's absolute, this
131// will produce a canonical version of that URL. See Canonicalize() for more
132// about the charset_converter.
133//
134// Returns true if the output is valid, false if the input could not produce
135// a valid URL.
136URL_EXPORT bool ResolveRelative(const char* base_spec,
137                                int base_spec_len,
138                                const Parsed& base_parsed,
139                                const char* relative,
140                                int relative_length,
141                                CharsetConverter* charset_converter,
142                                CanonOutput* output,
143                                Parsed* output_parsed);
144URL_EXPORT bool ResolveRelative(const char* base_spec,
145                                int base_spec_len,
146                                const Parsed& base_parsed,
147                                const base::char16* relative,
148                                int relative_length,
149                                CharsetConverter* charset_converter,
150                                CanonOutput* output,
151                                Parsed* output_parsed);
152
153// Replaces components in the given VALID input url. The new canonical URL info
154// is written to output and out_parsed.
155//
156// Returns true if the resulting URL is valid.
157URL_EXPORT bool ReplaceComponents(const char* spec,
158                                  int spec_len,
159                                  const Parsed& parsed,
160                                  const Replacements<char>& replacements,
161                                  CharsetConverter* charset_converter,
162                                  CanonOutput* output,
163                                  Parsed* out_parsed);
164URL_EXPORT bool ReplaceComponents(
165    const char* spec,
166    int spec_len,
167    const Parsed& parsed,
168    const Replacements<base::char16>& replacements,
169    CharsetConverter* charset_converter,
170    CanonOutput* output,
171    Parsed* out_parsed);
172
173// String helper functions ----------------------------------------------------
174
175// Compare the lower-case form of the given string against the given ASCII
176// string.  This is useful for doing checking if an input string matches some
177// token, and it is optimized to avoid intermediate string copies.
178//
179// The versions of this function that don't take a b_end assume that the b
180// string is NULL terminated.
181URL_EXPORT bool LowerCaseEqualsASCII(const char* a_begin,
182                                     const char* a_end,
183                                     const char* b);
184URL_EXPORT bool LowerCaseEqualsASCII(const char* a_begin,
185                                     const char* a_end,
186                                     const char* b_begin,
187                                     const char* b_end);
188URL_EXPORT bool LowerCaseEqualsASCII(const base::char16* a_begin,
189                                     const base::char16* a_end,
190                                     const char* b);
191
192// Unescapes the given string using URL escaping rules.
193URL_EXPORT void DecodeURLEscapeSequences(const char* input,
194                                         int length,
195                                         CanonOutputW* output);
196
197// Escapes the given string as defined by the JS method encodeURIComponent.  See
198// https://developer.mozilla.org/en/JavaScript/Reference/Global_Objects/encodeURIComponent
199URL_EXPORT void EncodeURIComponent(const char* input,
200                                   int length,
201                                   CanonOutput* output);
202
203}  // namespace url
204
205#endif  // URL_URL_UTIL_H_
206