1 // Copyright (c) 2011 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 COMPONENTS_URL_FORMATTER_URL_FIXER_H_ 6 #define COMPONENTS_URL_FORMATTER_URL_FIXER_H_ 7 8 #include <string> 9 10 #include "base/strings/string16.h" 11 #include "url/gurl.h" 12 13 namespace base { 14 class FilePath; 15 } 16 17 namespace url { 18 struct Component; 19 struct Parsed; 20 } 21 22 // These methods process user typed input that is meant to be a URL - like user 23 // typing in the URL bar or command line switches. The output is NOT guaranteed 24 // to be a valid URL. 25 // 26 // This is NOT the place for converting between different types of URLs or 27 // parsing them, see net_util.h for that. These methods should only be used on 28 // user typed input, NOT untrusted strings sourced from the web or elsewhere. 29 namespace url_formatter { 30 31 // Segments the given text string into parts of a URL. This is most useful for 32 // schemes such as http, https, and ftp where |SegmentURL| will find many 33 // segments. Currently does not segment "file" schemes. 34 // Returns the canonicalized scheme, or the empty string when |text| is only 35 // whitespace. 36 std::string SegmentURL(const std::string& text, url::Parsed* parts); 37 base::string16 SegmentURL(const base::string16& text, url::Parsed* parts); 38 39 // Attempts to fix common problems in user-typed text, making some "smart" 40 // adjustments to obviously-invalid input where possible. 41 // 42 // The result can still be invalid, so check the return value's validity or 43 // use possibly_invalid_spec(). DO NOT USE this method on untrusted strings 44 // from the web or elsewhere. Only use this for user-typed input. 45 // 46 // If |text| may be an absolute path to a file, it will get converted to a 47 // "file:" URL. 48 // 49 // Schemes "about" and "chrome" are normalized to "chrome://", with slashes. 50 // "about:blank" is unaltered, as Webkit allows frames to access about:blank. 51 // Additionally, if a chrome URL does not have a valid host, as in "about:", the 52 // returned URL will have the host "version", as in "chrome://version". 53 // 54 // If |desired_tld| is non-empty, it represents the TLD the user wishes to 55 // append in the case of an incomplete domain. We check that this is not a file 56 // path and there does not appear to be a valid TLD already, then append 57 // |desired_tld| to the domain and prepend "www." (unless it, or a scheme, are 58 // already present.) This TLD should not have a leading '.' (use "com" instead 59 // of ".com"). 60 GURL FixupURL(const std::string& text, const std::string& desired_tld); 61 62 // Converts |text| to a fixed-up URL, allowing it to be a relative path on the 63 // local filesystem. Begin searching in |base_dir|; if empty, use the current 64 // working directory. If this resolves to a file on disk, convert it to a 65 // "file:" URL in |fixed_up_url|; otherwise, fall back to the behavior of 66 // FixupURL(). 67 // 68 // For "regular" input, even if it is possibly a file with a full path, you 69 // should use FixupURL() directly. This function should only be used when 70 // relative path handling is desired, as for command line processing. 71 GURL FixupRelativeFile(const base::FilePath& base_dir, 72 const base::FilePath& text); 73 74 // Offsets the beginning index of |part| by |offset|, which is allowed to be 75 // negative. In some cases, the desired component does not exist at the given 76 // offset. For example, when converting from "http://foo" to "foo", the scheme 77 // component no longer exists. In such a case, the beginning index is set to 0. 78 // Does nothing if |part| is invalid. 79 void OffsetComponent(int offset, url::Component* part); 80 81 // Returns true if |scheme1| is equivalent to |scheme2|. 82 // Generally this is true if the two schemes are actually identical, but it's 83 // also true when one scheme is "about" and the other "chrome". 84 bool IsEquivalentScheme(const std::string& scheme1, const std::string& scheme2); 85 86 // For paths like ~, we use $HOME for the current user's home directory. 87 // For tests, we allow our idea of $HOME to be overriden by this variable. 88 extern const char* home_directory_override; 89 90 } // namespace url_formatter 91 92 #endif // COMPONENTS_URL_FORMATTER_URL_FIXER_H_ 93