Add full support for filesystem URLs.

chrome/common/extensions/url_pattern.h

-// Copyright (c) 2011 The Chromium Authors. All rights reserved.
+// Copyright (c) 2012 The Chromium Authors. All rights reserved.
 // Use of this source code is governed by a BSD-style license that can be
 // found in the LICENSE file.
 #ifndef CHROME_COMMON_EXTENSIONS_URL_PATTERN_H_
 #define CHROME_COMMON_EXTENSIONS_URL_PATTERN_H_
 #pragma once
 
 #include <functional>
 #include <string>
 #include <vector>
 
 class GURL;
 
 // A pattern that can be used to match URLs. A URLPattern is a very restricted
 // subset of URL syntax:
 //
-// <url-pattern> := <scheme>://<host><port><path> | '<all_urls>'
+// <url-pattern> := [filesystem:]<scheme>://<host><port><path> | '<all_urls>' |
+//                  <nonstandard-scheme>:<any chars>
 // <scheme> := '*' | 'http' | 'https' | 'file' | 'ftp' | 'chrome' |
-//             'chrome-extension' | 'filesystem'
+//             'chrome-extension'
 // <host> := '*' | '*.' <anychar except '/' and '*'>+
 // <port> := [':' ('*' | <port number between 0 and 65535>)]
 // <path> := '/' <any chars>
+// <nonstandard-scheme> := <any string except for "//" or those in scheme>
 //
 // * Host is not used when the scheme is 'file'.
 // * The path can have embedded '*' characters which act as glob wildcards.
 // * '<all_urls>' is a special pattern that matches any URL that contains a
 //   valid scheme (as specified by valid_schemes_).
 // * The '*' scheme pattern excludes file URLs.
 //
 // Examples of valid patterns:
 // - http://*/*
 // - http://*/foo*
 // - https://*.google.com/foo*bar
 // - file://monkey*
 // - http://127.0.0.1/*
+// - filesystem:*://*/*
 //
 // Examples of invalid patterns:
 // - http://* -- path not specified
 // - http://*foo/bar -- * not allowed as substring of host component
 // - http://foo.*.bar/baz -- * must be first component
 // - http:/bar -- scheme separator not found
 // - foo://* -- invalid scheme
 // - chrome:// -- we don't support chrome internal URLs
 class URLPattern {
  public:
@@ skipping 46 matching lines @@
   bool operator==(const URLPattern& other) const;
 
   // Initializes this instance by parsing the provided string. Returns
   // URLPattern::PARSE_SUCCESS on success, or an error code otherwise. On
   // failure, this instance will have some intermediate values and is in an
   // invalid state.
   ParseResult Parse(const std::string& pattern_str);
 
   // Gets the bitmask of valid schemes.
   int valid_schemes() const { return valid_schemes_; }
+
+  // Sets the bitmask of valid schemes and inner_schemes.
   void SetValidSchemes(int valid_schemes);

[Aaron Boodman, 2012/03/12 23:17:13]

It's surprising that this would also set the inner schemes. Why is that?

[ericu, 2012/03/13 21:58:47]

Generally if you want to match http://google.com, you also want to match
filesystem:http://google.com.

 
+  // Gets the bitmask of valid inner schemes.
+  int valid_inner_schemes() const { return valid_inner_schemes_; }
+
+  // Sets the bitmask of valid inner_schemes.
+  void SetValidInnerSchemes(int valid_schemes);
+
   // Gets the host the pattern matches. This can be an empty string if the
   // pattern matches all hosts (the input was <scheme>://*/<whatever>).
   const std::string& host() const { return host_; }
   void SetHost(const std::string& host);
 
   // Gets whether to match subdomains of host().
   bool match_subdomains() const { return match_subdomains_; }
   void SetMatchSubdomains(bool val);
 
   // Gets the path the pattern matches with the leading slash. This can have
   // embedded asterisks which are interpreted using glob rules.
   const std::string& path() const { return path_; }
   void SetPath(const std::string& path);
 
   // Returns true if this pattern matches all urls.
   bool match_all_urls() const { return match_all_urls_; }
   void SetMatchAllURLs(bool val);
 
   // Sets the scheme for pattern matches. This can be a single '*' if the
   // pattern matches all valid schemes (as defined by the valid_schemes_
   // property). Returns false on failure (if the scheme is not valid).
   bool SetScheme(const std::string& scheme);
-  // Note: You should use MatchesScheme() instead of this getter unless you
-  // absolutely need the exact scheme. This is exposed for testing.
+  // Sets the inner scheme for pattern matches; this can only be used for

[Aaron Boodman, 2012/03/12 23:17:13]

Add blank line before this comment block.

[ericu, 2012/03/13 21:58:47]

Done.

+  // patterns whose scheme has already been set to "filesystem". This can be a

[Aaron Boodman, 2012/03/12 23:17:13]

"with a scheme that supports inner schemes."

[ericu, 2012/03/13 21:58:47]

Done.

+  // single '*' if the pattern matches all valid schemes (as defined by the
+  // valid_inner_schemes_ property). Returns false on failure (if the scheme is
+  // not valid).
+  bool SetInnerScheme(const std::string& scheme);
+
+  // Returns the scheme component of the pattern.  Note that it is usually
+  // better to call MatchesScheme(), since that will also handle the case where
+  // the scheme is a wildcard.
   const std::string& scheme() const { return scheme_; }
 
+  // Returns the inner scheme component of the pattern.  Note that it is usually
+  // better to call MatchesInnerScheme(), since that will also handle the case
+  // where the inner scheme is a wildcard.
+  const std::string& inner_scheme() const { return inner_scheme_; }
+
   // Returns true if the specified scheme can be used in this URL pattern, and
   // false otherwise. Uses valid_schemes_ to determine validity.
   bool IsValidScheme(const std::string& scheme) const;
 
+  // Returns true if the specified inner scheme can be used in this URL pattern,
+  // and false otherwise. Uses valid_inner_schemes_ to determine validity.
+  bool IsValidInnerScheme(const std::string& inner_scheme) const;
+
   // Returns true if this instance matches the specified URL.
   bool MatchesURL(const GURL& test) const;
 
   // Returns true if this instance matches the specified security origin.
   bool MatchesSecurityOrigin(const GURL& test) const;
 
   // Returns true if |test| matches our scheme.
   bool MatchesScheme(const std::string& test) const;
 
+  // Returns true if |test| matches our inner scheme.
+  // Only use this if scheme is "filesystem".
+  bool MatchesInnerScheme(const std::string& test) const;
+
   // Returns true if |test| matches our host.
   bool MatchesHost(const std::string& test) const;
   bool MatchesHost(const GURL& test) const;
 
   // Returns true if |test| matches our path.
   bool MatchesPath(const std::string& test) const;
 
   // Returns true if |port| matches our port.
   bool MatchesPort(int port) const;
 
@@ skipping 41 matching lines @@
   // If the URLPattern contains a wildcard scheme, returns a list of
   // equivalent literal schemes, otherwise returns the current scheme.
   std::vector<std::string> GetExplicitSchemes() const;
 
   // A bitmask containing the schemes which are considered valid for this
   // pattern. Parse() uses this to decide whether a pattern contains a valid
   // scheme. MatchesScheme uses this to decide whether a wildcard scheme_
   // matches a given test scheme.
   int valid_schemes_;
 
+  // A bitmask containing the inner schemes which are considered valid for this
+  // pattern. Parse() uses this to decide whether a pattern contains a valid
+  // scheme. MatchesInnerScheme uses this to decide whether a wildcard
+  // inner_scheme_ matches a given test scheme.
+  int valid_inner_schemes_;
+
   // True if this is a special-case "<all_urls>" pattern.
   bool match_all_urls_;
 
   // The scheme for the pattern.
   std::string scheme_;
 
+  // The inner scheme for the pattern, used only when scheme_ is "filesystem".

[Aaron Boodman, 2012/03/12 23:17:13]

It would be better to create an array of schemes that support inner schemes, and
just refer to those. Even if it only has one item currently.

[ericu, 2012/03/13 21:58:47]

I made a trivial inline function instead of an array for now, just to keep it
simple.

+  std::string inner_scheme_;
+
   // The host without any leading "*" components.
   std::string host_;
 
   // Whether we should match subdomains of the host. This is true if the first
   // component of the pattern's host was "*".
   bool match_subdomains_;
 
   // The port.
   std::string port_;
 
   // The path to match. This is everything after the host of the URL, or
   // everything after the scheme in the case of file:// URLs.
   std::string path_;
 
   // The path with "?" and "\" characters escaped for use with the
   // MatchPattern() function.
   std::string path_escaped_;
 
   // A string representing this URLPattern.
   mutable std::string spec_;
 };
 
 typedef std::vector<URLPattern> URLPatternList;
 
 #endif  // CHROME_COMMON_EXTENSIONS_URL_PATTERN_H_