Chromium Code Reviews
chromiumcodereview-hr@appspot.gserviceaccount.com (chromiumcodereview-hr) | Please choose your nickname with Settings | Help | Chromium Project | Gerrit Changes | Sign out
(532)

Unified Diff: sdk/lib/core/uri.dart

Issue 2225243003: Revert "Reapply fast-URI patch." (Closed) Base URL: https://github.com/dart-lang/sdk.git@master
Patch Set: Created 4 years, 4 months ago
Use n/p to move between diff chunks; N/P to move between comments. Draft comments are only viewable by you.
Jump to:
View side-by-side diff with in-line comments
Download patch
« no previous file with comments | « sdk/lib/_internal/js_runtime/lib/core_patch.dart ('k') | tests/compiler/dart2js/compiler_helper.dart » ('j') | no next file with comments »
Expand Comments ('e') | Collapse Comments ('c') | Show Comments Hide Comments ('s')
Index: sdk/lib/core/uri.dart
diff --git a/sdk/lib/core/uri.dart b/sdk/lib/core/uri.dart
index 51d66cfa8052a64fb1ad6235f4afee31e1c4a918..456bcb7589769a20d1e84d41f915517caa11b7f0 100644
--- a/sdk/lib/core/uri.dart
+++ b/sdk/lib/core/uri.dart
@@ -4,24 +4,6 @@
part of dart.core;
-// Frequently used character codes.
-const int _SPACE = 0x20;
-const int _PERCENT = 0x25;
-const int _PLUS = 0x2B;
-const int _DOT = 0x2E;
-const int _SLASH = 0x2F;
-const int _COLON = 0x3A;
-const int _UPPER_CASE_A = 0x41;
-const int _UPPER_CASE_Z = 0x5A;
-const int _LEFT_BRACKET = 0x5B;
-const int _BACKSLASH = 0x5C;
-const int _RIGHT_BRACKET = 0x5D;
-const int _LOWER_CASE_A = 0x61;
-const int _LOWER_CASE_F = 0x66;
-const int _LOWER_CASE_Z = 0x7A;
-
-const String _hexDigits = "0123456789ABCDEF";
-
/**
* A parsed URI, such as a URL.
*
@@ -33,17 +15,77 @@ const String _hexDigits = "0123456789ABCDEF";
* [uris]: https://www.dartlang.org/docs/dart-up-and-running/ch03.html#uris
* [libtour]: https://www.dartlang.org/docs/dart-up-and-running/contents/ch03.html
*/
-abstract class Uri {
+class Uri {
/**
- * Returns the natural base URI for the current platform.
+ * The scheme component of the URI.
*
- * When running in a browser this is the current URL of the current page
- * (from `window.location.href`).
+ * Returns the empty string if there is no scheme component.
*
- * When not running in a browser this is the file URI referencing
- * the current working directory.
+ * A URI scheme is case insensitive.
+ * The returned scheme is canonicalized to lowercase letters.
*/
- external static Uri get base;
+ // We represent the missing scheme as an empty string.
+ // A valid scheme cannot be empty.
+ final String scheme;
+
+ /**
+ * The user-info part of the authority.
+ *
+ * Does not distinguish between an empty user-info and an absent one.
+ * The value is always non-null.
+ * Is considered absent if [_host] is `null`.
+ */
+ final String _userInfo;
+
+ /**
+ * The host name of the URI.
+ *
+ * Set to `null` if there is no authority in the URI.
+ * The host name is the only mandatory part of an authority, so we use
+ * it to mark whether an authority part was present or not.
+ */
+ final String _host;
+
+ /**
+ * The port number part of the authority.
+ *
+ * The port. Set to null if there is no port. Normalized to null if
+ * the port is the default port for the scheme.
+ */
+ int _port;
+
+ /**
+ * The path of the URI.
+ *
+ * Always non-null.
+ */
+ String _path;
+
+ // The query content, or null if there is no query.
+ final String _query;
+
+ // The fragment content, or null if there is no fragment.
+ final String _fragment;
+
+ /**
+ * Cache the computed return value of [pathSegements].
+ */
+ List<String> _pathSegments;
+
+ /**
+ * Cache the computed return value of [queryParameters].
+ */
+ Map<String, String> _queryParameters;
+ Map<String, List<String>> _queryParameterLists;
+
+ /// Internal non-verifying constructor. Only call with validated arguments.
+ Uri._internal(this.scheme,
+ this._userInfo,
+ this._host,
+ this._port,
+ this._path,
+ this._query,
+ this._fragment);
/**
* Creates a new URI from its components.
@@ -116,15 +158,39 @@ abstract class Uri {
* general delimiters, are escaped if necessary.
* If `fragment` is omitted or `null`, the URI has no fragment part.
*/
- factory Uri({String scheme,
- String userInfo,
+ factory Uri({String scheme : "",
+ String userInfo : "",
String host,
int port,
String path,
Iterable<String> pathSegments,
String query,
Map<String, dynamic/*String|Iterable<String>*/> queryParameters,
- String fragment}) = _Uri;
+ String fragment}) {
+ scheme = _makeScheme(scheme, 0, _stringOrNullLength(scheme));
+ userInfo = _makeUserInfo(userInfo, 0, _stringOrNullLength(userInfo));
+ host = _makeHost(host, 0, _stringOrNullLength(host), false);
+ // Special case this constructor for backwards compatibility.
+ if (query == "") query = null;
+ query = _makeQuery(query, 0, _stringOrNullLength(query), queryParameters);
+ fragment = _makeFragment(fragment, 0, _stringOrNullLength(fragment));
+ port = _makePort(port, scheme);
+ bool isFile = (scheme == "file");
+ if (host == null &&
+ (userInfo.isNotEmpty || port != null || isFile)) {
+ host = "";
+ }
+ bool hasAuthority = (host != null);
+ path = _makePath(path, 0, _stringOrNullLength(path), pathSegments,
+ scheme, hasAuthority);
+ if (scheme.isEmpty && host == null && !path.startsWith('/')) {
+ path = _normalizeRelativePath(path);
+ } else {
+ path = _removeDotSegments(path);
+ }
+ return new Uri._internal(scheme, userInfo, host, port,
+ path, query, fragment);
+ }
/**
* Creates a new `http` URI from authority, path and query.
@@ -161,7 +227,9 @@ abstract class Uri {
*/
factory Uri.http(String authority,
String unencodedPath,
- [Map<String, String> queryParameters]) = _Uri.http;
+ [Map<String, String> queryParameters]) {
+ return _makeHttpUri("http", authority, unencodedPath, queryParameters);
+ }
/**
* Creates a new `https` URI from authority, path and query.
@@ -171,542 +239,103 @@ abstract class Uri {
*/
factory Uri.https(String authority,
String unencodedPath,
- [Map<String, String> queryParameters]) = _Uri.https;
+ [Map<String, String> queryParameters]) {
+ return _makeHttpUri("https", authority, unencodedPath, queryParameters);
+ }
/**
- * Creates a new file URI from an absolute or relative file path.
- *
- * The file path is passed in [path].
- *
- * This path is interpreted using either Windows or non-Windows
- * semantics.
- *
- * With non-Windows semantics the slash ("/") is used to separate
- * path segments.
- *
- * With Windows semantics, backslash ("\\") and forward-slash ("/")
- * are used to separate path segments, except if the path starts
- * with "\\\\?\\" in which case, only backslash ("\\") separates path
- * segments.
- *
- * If the path starts with a path separator an absolute URI is
- * created. Otherwise a relative URI is created. One exception from
- * this rule is that when Windows semantics is used and the path
- * starts with a drive letter followed by a colon (":") and a
- * path separator then an absolute URI is created.
- *
- * The default for whether to use Windows or non-Windows semantics
- * determined from the platform Dart is running on. When running in
- * the standalone VM this is detected by the VM based on the
- * operating system. When running in a browser non-Windows semantics
- * is always used.
- *
- * To override the automatic detection of which semantics to use pass
- * a value for [windows]. Passing `true` will use Windows
- * semantics and passing `false` will use non-Windows semantics.
- *
- * Examples using non-Windows semantics:
- *
- * ```
- * // xxx/yyy
- * new Uri.file("xxx/yyy", windows: false);
- *
- * // xxx/yyy/
- * new Uri.file("xxx/yyy/", windows: false);
- *
- * // file:///xxx/yyy
- * new Uri.file("/xxx/yyy", windows: false);
- *
- * // file:///xxx/yyy/
- * new Uri.file("/xxx/yyy/", windows: false);
- *
- * // C:
- * new Uri.file("C:", windows: false);
- * ```
- *
- * Examples using Windows semantics:
+ * Returns the authority component.
*
- * ```
- * // xxx/yyy
- * new Uri.file(r"xxx\yyy", windows: true);
+ * The authority is formatted from the [userInfo], [host] and [port]
+ * parts.
*
- * // xxx/yyy/
- * new Uri.file(r"xxx\yyy\", windows: true);
+ * Returns the empty string if there is no authority component.
+ */
+ String get authority {
+ if (!hasAuthority) return "";
+ var sb = new StringBuffer();
+ _writeAuthority(sb);
+ return sb.toString();
+ }
+
+ /**
+ * Returns the user info part of the authority component.
*
- * file:///xxx/yyy
- * new Uri.file(r"\xxx\yyy", windows: true);
+ * Returns the empty string if there is no user info in the
+ * authority component.
+ */
+ String get userInfo => _userInfo;
+
+ /**
+ * Returns the host part of the authority component.
*
- * file:///xxx/yyy/
- * new Uri.file(r"\xxx\yyy/", windows: true);
+ * Returns the empty string if there is no authority component and
+ * hence no host.
*
- * // file:///C:/xxx/yyy
- * new Uri.file(r"C:\xxx\yyy", windows: true);
+ * If the host is an IP version 6 address, the surrounding `[` and `]` is
+ * removed.
*
- * // This throws an error. A path with a drive letter is not absolute.
- * new Uri.file(r"C:", windows: true);
+ * The host string is case-insensitive.
+ * The returned host name is canonicalized to lower-case
+ * with upper-case percent-escapes.
+ */
+ String get host {
+ if (_host == null) return "";
+ if (_host.startsWith('[')) {
+ return _host.substring(1, _host.length - 1);
+ }
+ return _host;
+ }
+
+ /**
+ * Returns the port part of the authority component.
*
- * // This throws an error. A path with a drive letter is not absolute.
- * new Uri.file(r"C:xxx\yyy", windows: true);
+ * Returns the defualt port if there is no port number in the authority
+ * component. That's 80 for http, 443 for https, and 0 for everything else.
+ */
+ int get port {
+ if (_port == null) return _defaultPort(scheme);
+ return _port;
+ }
+
+ // The default port for the scheme of this Uri..
+ static int _defaultPort(String scheme) {
+ if (scheme == "http") return 80;
+ if (scheme == "https") return 443;
+ return 0;
+ }
+
+ /**
+ * Returns the path component.
*
- * // file://server/share/file
- * new Uri.file(r"\\server\share\file", windows: true);
- * ```
+ * The returned path is encoded. To get direct access to the decoded
+ * path use [pathSegments].
*
- * If the path passed is not a legal file path [ArgumentError] is thrown.
+ * Returns the empty string if there is no path component.
*/
- factory Uri.file(String path, {bool windows}) = _Uri.file;
+ String get path => _path;
/**
- * Like [Uri.file] except that a non-empty URI path ends in a slash.
+ * Returns the query component. The returned query is encoded. To get
+ * direct access to the decoded query use [queryParameters].
*
- * If [path] is not empty, and it doesn't end in a directory separator,
- * then a slash is added to the returned URI's path.
- * In all other cases, the result is the same as returned by `Uri.file`.
+ * Returns the empty string if there is no query component.
*/
- factory Uri.directory(String path, {bool windows}) = _Uri.directory;
+ String get query => (_query == null) ? "" : _query;
/**
- * Creates a `data:` URI containing the [content] string.
- *
- * Converts the content to a bytes using [encoding] or the charset specified
- * in [parameters] (defaulting to US-ASCII if not specified or unrecognized),
- * then encodes the bytes into the resulting data URI.
+ * Returns the fragment identifier component.
*
- * Defaults to encoding using percent-encoding (any non-ASCII or non-URI-valid
- * bytes is replaced by a percent encoding). If [base64] is true, the bytes
- * are instead encoded using [BASE64].
+ * Returns the empty string if there is no fragment identifier
+ * component.
+ */
+ String get fragment => (_fragment == null) ? "" : _fragment;
+
+ /**
+ * Creates a new `Uri` object by parsing a URI string.
*
- * If [encoding] is not provided and [parameters] has a `charset` entry,
- * that name is looked up using [Encoding.getByName],
- * and if the lookup returns an encoding, that encoding is used to convert
- * [content] to bytes.
- * If providing both an [encoding] and a charset [parameter], they should
- * agree, otherwise decoding won't be able to use the charset parameter
- * to determine the encoding.
- *
- * If [mimeType] and/or [parameters] are supplied, they are added to the
- * created URI. If any of these contain characters that are not allowed
- * in the data URI, the character is percent-escaped. If the character is
- * non-ASCII, it is first UTF-8 encoded and then the bytes are percent
- * encoded. An omitted [mimeType] in a data URI means `text/plain`, just
- * as an omitted `charset` parameter defaults to meaning `US-ASCII`.
- *
- * To read the content back, use [UriData.contentAsString].
- */
- factory Uri.dataFromString(String content,
- {String mimeType,
- Encoding encoding,
- Map<String, String> parameters,
- bool base64: false}) {
- UriData data = new UriData.fromString(content,
- mimeType: mimeType,
- encoding: encoding,
- parameters: parameters,
- base64: base64);
- return data.uri;
- }
-
- /**
- * Creates a `data:` URI containing an encoding of [bytes].
- *
- * Defaults to Base64 encoding the bytes, but if [percentEncoded]
- * is `true`, the bytes will instead be percent encoded (any non-ASCII
- * or non-valid-ASCII-character byte is replaced by a percent encoding).
- *
- * To read the bytes back, use [UriData.contentAsBytes].
- *
- * It defaults to having the mime-type `application/octet-stream`.
- * The [mimeType] and [parameters] are added to the created URI.
- * If any of these contain characters that are not allowed
- * in the data URI, the character is percent-escaped. If the character is
- * non-ASCII, it is first UTF-8 encoded and then the bytes are percent
- * encoded.
- */
- factory Uri.dataFromBytes(List<int> bytes,
- {mimeType: "application/octet-stream",
- Map<String, String> parameters,
- percentEncoded: false}) {
- UriData data = new UriData.fromBytes(bytes,
- mimeType: mimeType,
- parameters: parameters,
- percentEncoded: percentEncoded);
- return data.uri;
- }
-
- /**
- * The scheme component of the URI.
- *
- * Returns the empty string if there is no scheme component.
- *
- * A URI scheme is case insensitive.
- * The returned scheme is canonicalized to lowercase letters.
- */
- String get scheme;
-
- /**
- * Returns the authority component.
- *
- * The authority is formatted from the [userInfo], [host] and [port]
- * parts.
- *
- * Returns the empty string if there is no authority component.
- */
- String get authority;
-
- /**
- * Returns the user info part of the authority component.
- *
- * Returns the empty string if there is no user info in the
- * authority component.
- */
- String get userInfo;
-
- /**
- * Returns the host part of the authority component.
- *
- * Returns the empty string if there is no authority component and
- * hence no host.
- *
- * If the host is an IP version 6 address, the surrounding `[` and `]` is
- * removed.
- *
- * The host string is case-insensitive.
- * The returned host name is canonicalized to lower-case
- * with upper-case percent-escapes.
- */
- String get host;
-
- /**
- * Returns the port part of the authority component.
- *
- * Returns the defualt port if there is no port number in the authority
- * component. That's 80 for http, 443 for https, and 0 for everything else.
- */
- int get port;
-
- /**
- * Returns the path component.
- *
- * The returned path is encoded. To get direct access to the decoded
- * path use [pathSegments].
- *
- * Returns the empty string if there is no path component.
- */
- String get path;
-
- /**
- * Returns the query component. The returned query is encoded. To get
- * direct access to the decoded query use [queryParameters].
- *
- * Returns the empty string if there is no query component.
- */
- String get query;
-
- /**
- * Returns the fragment identifier component.
- *
- * Returns the empty string if there is no fragment identifier
- * component.
- */
- String get fragment;
-
- /**
- * Returns the URI path split into its segments. Each of the segments in the
- * returned list have been decoded. If the path is empty the empty list will
- * be returned. A leading slash `/` does not affect the segments returned.
- *
- * The returned list is unmodifiable and will throw [UnsupportedError] on any
- * calls that would mutate it.
- */
- List<String> get pathSegments;
-
- /**
- * Returns the URI query split into a map according to the rules
- * specified for FORM post in the [HTML 4.01 specification section
- * 17.13.4](http://www.w3.org/TR/REC-html40/interact/forms.html#h-17.13.4 "HTML 4.01 section 17.13.4").
- * Each key and value in the returned map has been decoded.
- * If there is no query the empty map is returned.
- *
- * Keys in the query string that have no value are mapped to the
- * empty string.
- * If a key occurs more than once in the query string, it is mapped to
- * an arbitrary choice of possible value.
- * The [queryParametersAll] getter can provide a map
- * that maps keys to all of their values.
- *
- * The returned map is unmodifiable.
- */
- Map<String, String> get queryParameters;
-
- /**
- * Returns the URI query split into a map according to the rules
- * specified for FORM post in the [HTML 4.01 specification section
- * 17.13.4](http://www.w3.org/TR/REC-html40/interact/forms.html#h-17.13.4 "HTML 4.01 section 17.13.4").
- * Each key and value in the returned map has been decoded. If there is no
- * query the empty map is returned.
- *
- * Keys are mapped to lists of their values. If a key occurs only once,
- * its value is a singleton list. If a key occurs with no value, the
- * empty string is used as the value for that occurrence.
- *
- * The returned map and the lists it contains are unmodifiable.
- */
- Map<String, List<String>> get queryParametersAll;
-
- /**
- * Returns whether the URI is absolute.
- *
- * A URI is an absolute URI in the sense of RFC 3986 if it has a scheme
- * and no fragment.
- */
- bool get isAbsolute;
-
- /**
- * Returns whether the URI has a [scheme] component.
- */
- bool get hasScheme => scheme.isNotEmpty;
-
- /**
- * Returns whether the URI has an [authority] component.
- */
- bool get hasAuthority;
-
- /**
- * Returns whether the URI has an explicit port.
- *
- * If the port number is the default port number
- * (zero for unrecognized schemes, with http (80) and https (443) being
- * recognized),
- * then the port is made implicit and omitted from the URI.
- */
- bool get hasPort;
-
- /**
- * Returns whether the URI has a query part.
- */
- bool get hasQuery;
-
- /**
- * Returns whether the URI has a fragment part.
- */
- bool get hasFragment;
-
- /**
- * Returns whether the URI has an empty path.
- */
- bool get hasEmptyPath;
-
- /**
- * Returns whether the URI has an absolute path (starting with '/').
- */
- bool get hasAbsolutePath;
-
- /**
- * Returns the origin of the URI in the form scheme://host:port for the
- * schemes http and https.
- *
- * It is an error if the scheme is not "http" or "https".
- *
- * See: http://www.w3.org/TR/2011/WD-html5-20110405/origin-0.html#origin
- */
- String get origin;
-
- /**
- * Returns the file path from a file URI.
- *
- * The returned path has either Windows or non-Windows
- * semantics.
- *
- * For non-Windows semantics the slash ("/") is used to separate
- * path segments.
- *
- * For Windows semantics the backslash ("\\") separator is used to
- * separate path segments.
- *
- * If the URI is absolute the path starts with a path separator
- * unless Windows semantics is used and the first path segment is a
- * drive letter. When Windows semantics is used a host component in
- * the uri in interpreted as a file server and a UNC path is
- * returned.
- *
- * The default for whether to use Windows or non-Windows semantics
- * determined from the platform Dart is running on. When running in
- * the standalone VM this is detected by the VM based on the
- * operating system. When running in a browser non-Windows semantics
- * is always used.
- *
- * To override the automatic detection of which semantics to use pass
- * a value for [windows]. Passing `true` will use Windows
- * semantics and passing `false` will use non-Windows semantics.
- *
- * If the URI ends with a slash (i.e. the last path component is
- * empty) the returned file path will also end with a slash.
- *
- * With Windows semantics URIs starting with a drive letter cannot
- * be relative to the current drive on the designated drive. That is
- * for the URI `file:///c:abc` calling `toFilePath` will throw as a
- * path segment cannot contain colon on Windows.
- *
- * Examples using non-Windows semantics (resulting of calling
- * toFilePath in comment):
- *
- * Uri.parse("xxx/yyy"); // xxx/yyy
- * Uri.parse("xxx/yyy/"); // xxx/yyy/
- * Uri.parse("file:///xxx/yyy"); // /xxx/yyy
- * Uri.parse("file:///xxx/yyy/"); // /xxx/yyy/
- * Uri.parse("file:///C:"); // /C:
- * Uri.parse("file:///C:a"); // /C:a
- *
- * Examples using Windows semantics (resulting URI in comment):
- *
- * Uri.parse("xxx/yyy"); // xxx\yyy
- * Uri.parse("xxx/yyy/"); // xxx\yyy\
- * Uri.parse("file:///xxx/yyy"); // \xxx\yyy
- * Uri.parse("file:///xxx/yyy/"); // \xxx\yyy/
- * Uri.parse("file:///C:/xxx/yyy"); // C:\xxx\yyy
- * Uri.parse("file:C:xxx/yyy"); // Throws as a path segment
- * // cannot contain colon on Windows.
- * Uri.parse("file://server/share/file"); // \\server\share\file
- *
- * If the URI is not a file URI calling this throws
- * [UnsupportedError].
- *
- * If the URI cannot be converted to a file path calling this throws
- * [UnsupportedError].
- */
- // TODO(lrn): Deprecate and move functionality to File class or similar.
- // The core libraries should not worry about the platform.
- String toFilePath({bool windows});
-
- /**
- * Access the structure of a `data:` URI.
- *
- * Returns a [UriData] object for `data:` URIs and `null` for all other
- * URIs.
- * The [UriData] object can be used to access the media type and data
- * of a `data:` URI.
- */
- UriData get data;
-
- /// Returns a hash code computed as `toString().hashCode`.
- ///
- /// This guarantees that URIs with the same normalized
- int get hashCode;
-
- /// A URI is equal to another URI with the same normalized representation.
- bool operator==(Object other);
-
- /// Returns the normalized string representation of the URI.
- String toString();
-
- /**
- * Returns a new `Uri` based on this one, but with some parts replaced.
- *
- * This method takes the same parameters as the [new Uri] constructor,
- * and they have the same meaning.
- *
- * At most one of [path] and [pathSegments] must be provided.
- * Likewise, at most one of [query] and [queryParameters] must be provided.
- *
- * Each part that is not provided will default to the corresponding
- * value from this `Uri` instead.
- *
- * This method is different from [Uri.resolve] which overrides in a
- * hierarchial manner,
- * and can instead replace each part of a `Uri` individually.
- *
- * Example:
- *
- * Uri uri1 = Uri.parse("a://b@c:4/d/e?f#g");
- * Uri uri2 = uri1.replace(scheme: "A", path: "D/E/E", fragment: "G");
- * print(uri2); // prints "A://b@c:4/D/E/E/?f#G"
- *
- * This method acts similarly to using the `new Uri` constructor with
- * some of the arguments taken from this `Uri` . Example:
- *
- * Uri uri3 = new Uri(
- * scheme: "A",
- * userInfo: uri1.userInfo,
- * host: uri1.host,
- * port: uri1.port,
- * path: "D/E/E",
- * query: uri1.query,
- * fragment: "G");
- * print(uri3); // prints "A://b@c:4/D/E/E/?f#G"
- * print(uri2 == uri3); // prints true.
- *
- * Using this method can be seen as a shorthand for the `Uri` constructor
- * call above, but may also be slightly faster because the parts taken
- * from this `Uri` need not be checked for validity again.
- */
- Uri replace({String scheme,
- String userInfo,
- String host,
- int port,
- String path,
- Iterable<String> pathSegments,
- String query,
- Map<String, dynamic/*String|Iterable<String>*/> queryParameters,
- String fragment});
-
- /**
- * Returns a `Uri` that differs from this only in not having a fragment.
- *
- * If this `Uri` does not have a fragment, it is itself returned.
- */
- Uri removeFragment();
-
- /**
- * Resolve [reference] as an URI relative to `this`.
- *
- * First turn [reference] into a URI using [Uri.parse]. Then resolve the
- * resulting URI relative to `this`.
- *
- * Returns the resolved URI.
- *
- * See [resolveUri] for details.
- */
- Uri resolve(String reference);
-
- /**
- * Resolve [reference] as an URI relative to `this`.
- *
- * Returns the resolved URI.
- *
- * The algorithm "Transform Reference" for resolving a reference is described
- * in [RFC-3986 Section 5](http://tools.ietf.org/html/rfc3986#section-5 "RFC-1123").
- *
- * Updated to handle the case where the base URI is just a relative path -
- * that is: when it has no scheme or authority and the path does not start
- * with a slash.
- * In that case, the paths are combined without removing leading "..", and
- * an empty path is not converted to "/".
- */
- Uri resolveUri(Uri reference);
-
- /**
- * Returns a URI where the path has been normalized.
- *
- * A normalized path does not contain `.` segments or non-leading `..`
- * segments.
- * Only a relative path with no scheme or authority may contain
- * leading `..` segments,
- * a path that starts with `/` will also drop any leading `..` segments.
- *
- * This uses the same normalization strategy as `new Uri().resolve(this)`.
- *
- * Does not change any part of the URI except the path.
- *
- * The default implementation of `Uri` always normalizes paths, so calling
- * this function has no effect.
- */
- Uri normalizePath();
-
- /**
- * Creates a new `Uri` object by parsing a URI string.
- *
- * If [start] and [end] are provided, only the substring from `start`
- * to `end` is parsed as a URI.
+ * If [start] and [end] are provided, only the substring from `start`
+ * to `end` is parsed as a URI.
*
* If the string is not valid as a URI or URI reference,
* a [FormatException] is thrown.
@@ -764,769 +393,207 @@ abstract class Uri {
//
// query = *( pchar / "/" / "?" )
//
- // fragment = *( pchar / "/" / "?" )
- end ??= uri.length;
-
- // Special case data:URIs. Ignore case when testing.
- if (end >= start + 5) {
- int dataDelta = _startsWithData(uri, start);
- if (dataDelta == 0) {
- // The case is right.
- if (start > 0 || end < uri.length) uri = uri.substring(start, end);
- return UriData._parse(uri, 5, null).uri;
- } else if (dataDelta == 0x20) {
- return UriData._parse(uri.substring(start + 5, end), 0, null).uri;
- }
- // Otherwise the URI doesn't start with "data:" or any case variant of it.
- }
-
- // The following index-normalization belongs with the scanning, but is
- // easier to do here because we already have extracted variables from the
- // indices list.
- var indices = new List<int>(8);//new List<int>.filled(8, start - 1);
-
- // Set default values for each position.
- // The value will either be correct in some cases where it isn't set
- // by the scanner, or it is clearly recognizable as an unset value.
- indices
- ..[0] = 0
- ..[_schemeEndIndex] = start - 1
- ..[_hostStartIndex] = start - 1
- ..[_notSimpleIndex] = start - 1
- ..[_portStartIndex] = start
- ..[_pathStartIndex] = start
- ..[_queryStartIndex] = end
- ..[_fragmentStartIndex] = end;
- var state = _scan(uri, start, end, _uriStart, indices);
- // Some states that should be non-simple, but the URI ended early.
- // Paths that end at a ".." must be normalized to end in "../".
- if (state >= _nonSimpleEndStates) {
- indices[_notSimpleIndex] = end;
- }
- int schemeEnd = indices[_schemeEndIndex];
- if (schemeEnd >= start) {
- // Rescan the scheme part now that we know it's not a path.
- state = _scan(uri, start, schemeEnd, _schemeStart, indices);
- if (state == _schemeStart) {
- // Empty scheme.
- indices[_notSimpleIndex] = schemeEnd;
- }
- }
- // The returned positions are limited by the scanners ability to write only
- // one position per character, and only the current position.
- // Scanning from left to right, we only know whether something is a scheme
- // or a path when we see a `:` or `/`, and likewise we only know if the first
- // `/` is part of the path or is leading an authority component when we see
- // the next character.
-
- int hostStart = indices[_hostStartIndex] + 1;
- int portStart = indices[_portStartIndex];
- int pathStart = indices[_pathStartIndex];
- int queryStart = indices[_queryStartIndex];
- int fragmentStart = indices[_fragmentStartIndex];
-
- // We may discover scheme while handling special cases.
- String scheme;
-
- // Derive some positions that weren't set to normalize the indices.
- // If pathStart isn't set (it's before scheme end or host start), then
- // the path is empty.
- if (fragmentStart < queryStart) queryStart = fragmentStart;
- if (pathStart < hostStart || pathStart <= schemeEnd) {
- pathStart = queryStart;
- }
- // If there is an authority with no port, set the port position
- // to be at the end of the authority (equal to pathStart).
- // This also handles a ":" in a user-info component incorrectly setting
- // the port start position.
- if (portStart < hostStart) portStart = pathStart;
-
- assert(hostStart == start || schemeEnd <= hostStart);
- assert(hostStart <= portStart);
- assert(schemeEnd <= pathStart);
- assert(portStart <= pathStart);
- assert(pathStart <= queryStart);
- assert(queryStart <= fragmentStart);
-
- bool isSimple = indices[_notSimpleIndex] < start;
-
- if (isSimple) {
- // Check/do normalizations that weren't detected by the scanner.
- // This includes removal of empty port or userInfo,
- // or scheme specific port and path normalizations.
- if (hostStart > schemeEnd + 3) {
- // Always be non-simple if URI contains user-info.
- // The scanner doesn't set the not-simple position in this case because
- // it's setting the host-start position instead.
- isSimple = false;
- } else if (portStart > start && portStart + 1 == pathStart) {
- // If the port is empty, it should be omitted.
- // Pathological case, don't bother correcting it.
- isSimple = false;
- } else if (queryStart < end &&
- (queryStart == pathStart + 2 &&
- uri.startsWith("..", pathStart)) ||
- (queryStart > pathStart + 2 &&
- uri.startsWith("/..", queryStart - 3))) {
- // The path ends in a ".." segment. This should be normalized to "../".
- // We didn't detect this while scanning because a query or fragment was
- // detected at the same time (which is why we only need to check this
- // if there is something after the path).
- isSimple = false;
- } else {
- // There are a few scheme-based normalizations that
- // the scanner couldn't check.
- // That means that the input is very close to simple, so just do
- // the normalizations.
- if (schemeEnd == start + 4) {
- // Do scheme based normalizations for file, http.
- if (uri.startsWith("file", start)) {
- scheme = "file";
- if (hostStart <= start) {
- // File URIs should have an authority.
- // Paths after an authority should be absolute.
- String schemeAuth = "file://";
- int delta = 2;
- if (!uri.startsWith("/", pathStart)) {
- schemeAuth = "file:///";
- delta = 3;
- }
- uri = schemeAuth + uri.substring(pathStart, end);
- schemeEnd -= start;
- hostStart = 7;
- portStart = 7;
- pathStart = 7;
- queryStart += delta - start;
- fragmentStart += delta - start;
- start = 0;
- end = uri.length;
- } else if (pathStart == queryStart) {
- // Uri has authority and empty path. Add "/" as path.
- if (start == 0 && end == uri.length) {
- uri = uri.replaceRange(pathStart, queryStart, "/");
- queryStart += 1;
- fragmentStart += 1;
- end += 1;
- } else {
- uri = "${uri.substring(start, pathStart)}/"
- "${uri.substring(queryStart, end)}";
- schemeEnd -= start;
- hostStart -= start;
- portStart -= start;
- pathStart -= start;
- queryStart += 1 - start;
- fragmentStart += 1 - start;
- start = 0;
- end = uri.length;
- }
- }
- } else if (uri.startsWith("http", start)) {
- scheme = "http";
- // HTTP URIs should not have an explicit port of 80.
- if (portStart > start && portStart + 3 == pathStart &&
- uri.startsWith("80", portStart + 1)) {
- if (start == 0 && end == uri.length) {
- uri = uri.replaceRange(portStart, pathStart, "");
- pathStart -= 3;
- queryStart -= 3;
- fragmentStart -= 3;
- end -= 3;
- } else {
- uri = uri.substring(start, portStart) +
- uri.substring(pathStart, end);
- schemeEnd -= start;
- hostStart -= start;
- portStart -= start;
- pathStart -= 3 + start;
- queryStart -= 3 + start;
- fragmentStart -= 3 + start;
- start = 0;
- end = uri.length;
- }
- }
- }
- } else if (schemeEnd == start + 5 && uri.startsWith("https", start)) {
- scheme = "https";
- // HTTPS URIs should not have an explicit port of 443.
- if (portStart > start && portStart + 4 == pathStart &&
- uri.startsWith("443", portStart + 1)) {
- if (start == 0 && end == uri.length) {
- uri = uri.replaceRange(portStart, pathStart, "");
- pathStart -= 4;
- queryStart -= 4;
- fragmentStart -= 4;
- end -= 3;
- } else {
- uri = uri.substring(start, portStart) +
- uri.substring(pathStart, end);
- schemeEnd -= start;
- hostStart -= start;
- portStart -= start;
- pathStart -= 4 + start;
- queryStart -= 4 + start;
- fragmentStart -= 4 + start;
- start = 0;
- end = uri.length;
- }
- }
- }
- }
- }
-
- if (isSimple) {
- if (start > 0 || end < uri.length) {
- uri = uri.substring(start, end);
- schemeEnd -= start;
- hostStart -= start;
- portStart -= start;
- pathStart -= start;
- queryStart -= start;
- fragmentStart -= start;
- }
- return new _SimpleUri(uri, schemeEnd, hostStart, portStart, pathStart,
- queryStart, fragmentStart, scheme);
-
- }
-
- return new _Uri.notSimple(uri, start, end, schemeEnd, hostStart, portStart,
- pathStart, queryStart, fragmentStart, scheme);
- }
-
- /**
- * Encode the string [component] using percent-encoding to make it
- * safe for literal use as a URI component.
- *
- * All characters except uppercase and lowercase letters, digits and
- * the characters `-_.!~*'()` are percent-encoded. This is the
- * set of characters specified in RFC 2396 and the which is
- * specified for the encodeUriComponent in ECMA-262 version 5.1.
- *
- * When manually encoding path segments or query components remember
- * to encode each part separately before building the path or query
- * string.
- *
- * For encoding the query part consider using
- * [encodeQueryComponent].
- *
- * To avoid the need for explicitly encoding use the [pathSegments]
- * and [queryParameters] optional named arguments when constructing
- * a [Uri].
- */
- static String encodeComponent(String component) {
- return _Uri._uriEncode(_Uri._unreserved2396Table, component, UTF8, false);
- }
-
- /**
- * Encode the string [component] according to the HTML 4.01 rules
- * for encoding the posting of a HTML form as a query string
- * component.
- *
- * Encode the string [component] according to the HTML 4.01 rules
- * for encoding the posting of a HTML form as a query string
- * component.
-
- * The component is first encoded to bytes using [encoding].
- * The default is to use [UTF8] encoding, which preserves all
- * the characters that don't need encoding.
-
- * Then the resulting bytes are "percent-encoded". This transforms
- * spaces (U+0020) to a plus sign ('+') and all bytes that are not
- * the ASCII decimal digits, letters or one of '-._~' are written as
- * a percent sign '%' followed by the two-digit hexadecimal
- * representation of the byte.
-
- * Note that the set of characters which are percent-encoded is a
- * superset of what HTML 4.01 requires, since it refers to RFC 1738
- * for reserved characters.
- *
- * When manually encoding query components remember to encode each
- * part separately before building the query string.
- *
- * To avoid the need for explicitly encoding the query use the
- * [queryParameters] optional named arguments when constructing a
- * [Uri].
- *
- * See http://www.w3.org/TR/html401/interact/forms.html#h-17.13.4.2 for more
- * details.
- */
- static String encodeQueryComponent(String component,
- {Encoding encoding: UTF8}) {
- return _Uri._uriEncode(_Uri._unreservedTable, component, encoding, true);
- }
-
- /**
- * Decodes the percent-encoding in [encodedComponent].
- *
- * Note that decoding a URI component might change its meaning as
- * some of the decoded characters could be characters with are
- * delimiters for a given URI componene type. Always split a URI
- * component using the delimiters for the component before decoding
- * the individual parts.
- *
- * For handling the [path] and [query] components consider using
- * [pathSegments] and [queryParameters] to get the separated and
- * decoded component.
- */
- static String decodeComponent(String encodedComponent) {
- return _Uri._uriDecode(encodedComponent, 0, encodedComponent.length,
- UTF8, false);
- }
-
- /**
- * Decodes the percent-encoding in [encodedComponent], converting
- * pluses to spaces.
- *
- * It will create a byte-list of the decoded characters, and then use
- * [encoding] to decode the byte-list to a String. The default encoding is
- * UTF-8.
- */
- static String decodeQueryComponent(
- String encodedComponent,
- {Encoding encoding: UTF8}) {
- return _Uri._uriDecode(encodedComponent, 0, encodedComponent.length,
- encoding, true);
- }
-
- /**
- * Encode the string [uri] using percent-encoding to make it
- * safe for literal use as a full URI.
- *
- * All characters except uppercase and lowercase letters, digits and
- * the characters `!#$&'()*+,-./:;=?@_~` are percent-encoded. This
- * is the set of characters specified in in ECMA-262 version 5.1 for
- * the encodeURI function .
- */
- static String encodeFull(String uri) {
- return _Uri._uriEncode(_Uri._encodeFullTable, uri, UTF8, false);
- }
-
- /**
- * Decodes the percent-encoding in [uri].
- *
- * Note that decoding a full URI might change its meaning as some of
- * the decoded characters could be reserved characters. In most
- * cases an encoded URI should be parsed into components using
- * [Uri.parse] before decoding the separate components.
- */
- static String decodeFull(String uri) {
- return _Uri._uriDecode(uri, 0, uri.length, UTF8, false);
- }
-
- /**
- * Returns the [query] split into a map according to the rules
- * specified for FORM post in the [HTML 4.01 specification section
- * 17.13.4](http://www.w3.org/TR/REC-html40/interact/forms.html#h-17.13.4 "HTML 4.01 section 17.13.4").
- * Each key and value in the returned map has been decoded. If the [query]
- * is the empty string an empty map is returned.
- *
- * Keys in the query string that have no value are mapped to the
- * empty string.
- *
- * Each query component will be decoded using [encoding]. The default encoding
- * is UTF-8.
- */
- static Map<String, String> splitQueryString(String query,
- {Encoding encoding: UTF8}) {
- return query.split("&").fold({}, (map, element) {
- int index = element.indexOf("=");
- if (index == -1) {
- if (element != "") {
- map[decodeQueryComponent(element, encoding: encoding)] = "";
- }
- } else if (index != 0) {
- var key = element.substring(0, index);
- var value = element.substring(index + 1);
- map[decodeQueryComponent(key, encoding: encoding)] =
- decodeQueryComponent(value, encoding: encoding);
- }
- return map;
- });
- }
-
-
- /**
- * Parse the [host] as an IP version 4 (IPv4) address, returning the address
- * as a list of 4 bytes in network byte order (big endian).
- *
- * Throws a [FormatException] if [host] is not a valid IPv4 address
- * representation.
- */
- static List<int> parseIPv4Address(String host) =>
- _parseIPv4Address(host, 0, host.length);
-
- /// Implementation of [parseIPv4Address] that can work on a substring.
- static List<int> _parseIPv4Address(String host, int start, int end) {
- void error(String msg, int position) {
- throw new FormatException('Illegal IPv4 address, $msg', host, position);
- }
-
- var result = new Uint8List(4);
- int partIndex = 0;
- int partStart = start;
- for (int i = start; i < end; i++) {
- int char = host.codeUnitAt(i);
- if (char != _DOT) {
- if (char ^ 0x30 > 9) {
- // Fail on a non-digit character.
- error("invalid character", i);
- }
- } else {
- if (partIndex == 3) {
- error('IPv4 address should contain exactly 4 parts', i);
- }
- int part = int.parse(host.substring(partStart, i));
- if (part > 255) {
- error("each part must be in the range 0..255", partStart);
- }
- result[partIndex++] = part;
- partStart = i + 1;
- }
- }
-
- if (partIndex != 3) {
- error('IPv4 address should contain exactly 4 parts', end);
- }
-
- int part = int.parse(host.substring(partStart, end));
- if (part > 255) {
- error("each part must be in the range 0..255", partStart);
- }
- result[partIndex] = part;
-
- return result;
- }
-
- /**
- * Parse the [host] as an IP version 6 (IPv6) address, returning the address
- * as a list of 16 bytes in network byte order (big endian).
- *
- * Throws a [FormatException] if [host] is not a valid IPv6 address
- * representation.
- *
- * Acts on the substring from [start] to [end]. If [end] is omitted, it
- * defaults ot the end of the string.
- *
- * Some examples of IPv6 addresses:
- * * ::1
- * * FEDC:BA98:7654:3210:FEDC:BA98:7654:3210
- * * 3ffe:2a00:100:7031::1
- * * ::FFFF:129.144.52.38
- * * 2010:836B:4179::836B:4179
- */
- static List<int> parseIPv6Address(String host, [int start = 0, int end]) {
- if (end == null) end = host.length;
- // An IPv6 address consists of exactly 8 parts of 1-4 hex digits, separated
- // by `:`'s, with the following exceptions:
- //
- // - One (and only one) wildcard (`::`) may be present, representing a fill
- // of 0's. The IPv6 `::` is thus 16 bytes of `0`.
- // - The last two parts may be replaced by an IPv4 "dotted-quad" address.
+ // fragment = *( pchar / "/" / "?" )
+ const int EOI = -1;
- // Helper function for reporting a badly formatted IPv6 address.
- void error(String msg, [position]) {
- throw new FormatException('Illegal IPv6 address, $msg', host, position);
- }
+ String scheme = "";
+ String userinfo = "";
+ String host = null;
+ int port = null;
+ String path = null;
+ String query = null;
+ String fragment = null;
+ if (end == null) end = uri.length;
- // Parse a hex block.
- int parseHex(int start, int end) {
- if (end - start > 4) {
- error('an IPv6 part can only contain a maximum of 4 hex digits', start);
+ int index = start;
+ int pathStart = start;
+ // End of input-marker.
+ int char = EOI;
+
+ void parseAuth() {
+ if (index == end) {
+ char = EOI;
+ return;
+ }
+ int authStart = index;
+ int lastColon = -1;
+ int lastAt = -1;
+ char = uri.codeUnitAt(index);
+ while (index < end) {
+ char = uri.codeUnitAt(index);
+ if (char == _SLASH || char == _QUESTION || char == _NUMBER_SIGN) {
+ break;
+ }
+ if (char == _AT_SIGN) {
+ lastAt = index;
+ lastColon = -1;
+ } else if (char == _COLON) {
+ lastColon = index;
+ } else if (char == _LEFT_BRACKET) {
+ lastColon = -1;
+ int endBracket = uri.indexOf(']', index + 1);
+ if (endBracket == -1) {
+ index = end;
+ char = EOI;
+ break;
+ } else {
+ index = endBracket;
+ }
+ }
+ index++;
+ char = EOI;
+ }
+ int hostStart = authStart;
+ int hostEnd = index;
+ if (lastAt >= 0) {
+ userinfo = _makeUserInfo(uri, authStart, lastAt);
+ hostStart = lastAt + 1;
+ }
+ if (lastColon >= 0) {
+ int portNumber;
+ if (lastColon + 1 < index) {
+ portNumber = 0;
+ for (int i = lastColon + 1; i < index; i++) {
+ int digit = uri.codeUnitAt(i);
+ if (_ZERO > digit || _NINE < digit) {
+ _fail(uri, i, "Invalid port number");
+ }
+ portNumber = portNumber * 10 + (digit - _ZERO);
+ }
+ }
+ port = _makePort(portNumber, scheme);
+ hostEnd = lastColon;
+ }
+ host = _makeHost(uri, hostStart, hostEnd, true);
+ if (index < end) {
+ char = uri.codeUnitAt(index);
+ }
+ }
+
+ // When reaching path parsing, the current character is known to not
+ // be part of the path.
+ const int NOT_IN_PATH = 0;
+ // When reaching path parsing, the current character is part
+ // of the a non-empty path.
+ const int IN_PATH = 1;
+ // When reaching authority parsing, authority is possible.
+ // This is only true at start or right after scheme.
+ const int ALLOW_AUTH = 2;
+
+ // Current state.
+ // Initialized to the default value that is used when exiting the
+ // scheme loop by reaching the end of input.
+ // All other breaks set their own state.
+ int state = NOT_IN_PATH;
+ int i = index; // Temporary alias for index to avoid bug 19550 in dart2js.
+ while (i < end) {
+ char = uri.codeUnitAt(i);
+ if (char == _QUESTION || char == _NUMBER_SIGN) {
+ state = NOT_IN_PATH;
+ break;
}
- int value = int.parse(host.substring(start, end), radix: 16);
- if (value < 0 || value > 0xFFFF) {
- error('each part must be in the range of `0x0..0xFFFF`', start);
+ if (char == _SLASH) {
+ state = (i == start) ? ALLOW_AUTH : IN_PATH;
+ break;
}
- return value;
- }
-
- if (host.length < 2) error('address is too short');
- List<int> parts = [];
- bool wildcardSeen = false;
- // Set if seeing a ".", suggesting that there is an IPv4 address.
- bool seenDot = false;
- int partStart = start;
- // Parse all parts, except a potential last one.
- for (int i = start; i < end; i++) {
- int char = host.codeUnitAt(i);
if (char == _COLON) {
- if (i == start) {
- // If we see a `:` in the beginning, expect wildcard.
- i++;
- if (host.codeUnitAt(i) != _COLON) {
- error('invalid start colon.', i);
- }
- partStart = i;
+ if (i == start) _fail(uri, start, "Invalid empty scheme");
+ scheme = _makeScheme(uri, start, i);
+ i++;
+ if (scheme == "data") {
+ // This generates a URI that is (potentially) not path normalized.
+ // Applying part normalization to a non-hierarchial URI isn't
+ // meaningful.
+ return UriData._parse(uri, i, null).uri;
}
- if (i == partStart) {
- // Wildcard. We only allow one.
- if (wildcardSeen) {
- error('only one wildcard `::` is allowed', i);
- }
- wildcardSeen = true;
- parts.add(-1);
+ pathStart = i;
+ if (i == end) {
+ char = EOI;
+ state = NOT_IN_PATH;
} else {
- // Found a single colon. Parse [partStart..i] as a hex entry.
- parts.add(parseHex(partStart, i));
+ char = uri.codeUnitAt(i);
+ if (char == _QUESTION || char == _NUMBER_SIGN) {
+ state = NOT_IN_PATH;
+ } else if (char == _SLASH) {
+ state = ALLOW_AUTH;
+ } else {
+ state = IN_PATH;
+ }
}
- partStart = i + 1;
- } else if (char == _DOT) {
- seenDot = true;
+ break;
}
- }
- if (parts.length == 0) error('too few parts');
- bool atEnd = (partStart == end);
- bool isLastWildcard = (parts.last == -1);
- if (atEnd && !isLastWildcard) {
- error('expected a part after last `:`', end);
- }
- if (!atEnd) {
- if (!seenDot) {
- parts.add(parseHex(partStart, end));
+ i++;
+ char = EOI;
+ }
+ index = i; // Remove alias when bug is fixed.
+
+ if (state == ALLOW_AUTH) {
+ assert(char == _SLASH);
+ // Have seen one slash either at start or right after scheme.
+ // If two slashes, it's an authority, otherwise it's just the path.
+ index++;
+ if (index == end) {
+ char = EOI;
+ state = NOT_IN_PATH;
} else {
- List<int> last = _parseIPv4Address(host, partStart, end);
- parts.add(last[0] << 8 | last[1]);
- parts.add(last[2] << 8 | last[3]);
- }
- }
- if (wildcardSeen) {
- if (parts.length > 7) {
- error('an address with a wildcard must have less than 7 parts');
- }
- } else if (parts.length != 8) {
- error('an address without a wildcard must contain exactly 8 parts');
- }
- List<int> bytes = new Uint8List(16);
- for (int i = 0, index = 0; i < parts.length; i++) {
- int value = parts[i];
- if (value == -1) {
- int wildCardLength = 9 - parts.length;
- for (int j = 0; j < wildCardLength; j++) {
- bytes[index] = 0;
- bytes[index + 1] = 0;
- index += 2;
+ char = uri.codeUnitAt(index);
+ if (char == _SLASH) {
+ index++;
+ parseAuth();
+ pathStart = index;
+ }
+ if (char == _QUESTION || char == _NUMBER_SIGN || char == EOI) {
+ state = NOT_IN_PATH;
+ } else {
+ state = IN_PATH;
}
- } else {
- bytes[index] = value >> 8;
- bytes[index + 1] = value & 0xff;
- index += 2;
}
}
- return bytes;
- }
-}
-
-class _Uri implements Uri {
- // We represent the missing scheme as an empty string.
- // A valid scheme cannot be empty.
- final String scheme;
-
- /**
- * The user-info part of the authority.
- *
- * Does not distinguish between an empty user-info and an absent one.
- * The value is always non-null.
- * Is considered absent if [_host] is `null`.
- */
- final String _userInfo;
-
- /**
- * The host name of the URI.
- *
- * Set to `null` if there is no authority in the URI.
- * The host name is the only mandatory part of an authority, so we use
- * it to mark whether an authority part was present or not.
- */
- final String _host;
-
- /**
- * The port number part of the authority.
- *
- * The port. Set to null if there is no port. Normalized to null if
- * the port is the default port for the scheme.
- */
- int _port;
-
- /**
- * The path of the URI.
- *
- * Always non-null.
- */
- String _path;
- // The query content, or null if there is no query.
- final String _query;
-
- // The fragment content, or null if there is no fragment.
- final String _fragment;
-
- /**
- * Cache the computed return value of [pathSegements].
- */
- List<String> _pathSegments;
-
- /**
- * Cache of the full normalized text representation of the URI.
- */
- String _text;
-
- /**
- * Cache of the hashCode of [_text].
- *
- * Is null until computed.
- */
- int _hashCodeCache;
-
- /**
- * Cache the computed return value of [queryParameters].
- */
- Map<String, String> _queryParameters;
- Map<String, List<String>> _queryParameterLists;
-
- /// Internal non-verifying constructor. Only call with validated arguments.
- _Uri._internal(this.scheme,
- this._userInfo,
- this._host,
- this._port,
- this._path,
- this._query,
- this._fragment);
-
- /// Create a [_Uri] from parts of [uri].
- ///
- /// The parameters specify the start/end of particular components of the URI.
- /// The [scheme] may contain a string representing a normalized scheme
- /// component if one has already been discovered.
- factory _Uri.notSimple(String uri, int start, int end, int schemeEnd,
- int hostStart, int portStart, int pathStart,
- int queryStart, int fragmentStart, String scheme) {
- if (scheme == null) {
- scheme = "";
- if (schemeEnd > start) {
- scheme = _makeScheme(uri, start, schemeEnd);
- } else if (schemeEnd == start) {
- _fail(uri, start, "Invalid empty scheme");
- }
- }
- String userInfo = "";
- String host;
- int port;
- if (hostStart > start) {
- int userInfoStart = schemeEnd + 3;
- if (userInfoStart < hostStart) {
- userInfo = _makeUserInfo(uri, userInfoStart, hostStart - 1);
- }
- host = _makeHost(uri, hostStart, portStart, false);
- if (portStart + 1 < pathStart) {
- // Should throw because invalid.
- port = int.parse(uri.substring(portStart + 1, pathStart), onError: (_) {
- throw new FormatException("Invalid port", uri, portStart + 1);
- });
- port = _makePort(port, scheme);
+ assert(state == IN_PATH || state == NOT_IN_PATH);
+ if (state == IN_PATH) {
+ // Characters from pathStart to index (inclusive) are known
+ // to be part of the path.
+ while (++index < end) {
+ char = uri.codeUnitAt(index);
+ if (char == _QUESTION || char == _NUMBER_SIGN) {
+ break;
+ }
+ char = EOI;
}
+ state = NOT_IN_PATH;
}
- String path = _makePath(uri, pathStart, queryStart, null,
- scheme, host != null);
- String query;
- if (queryStart < fragmentStart) {
- query = _makeQuery(uri, queryStart + 1, fragmentStart, null);
- }
- String fragment;
- if (fragmentStart < end) {
- fragment = _makeFragment(uri, fragmentStart + 1, end);
- }
- return new _Uri._internal(scheme,
- userInfo,
- host,
- port,
- path,
- query,
- fragment);
- }
- /// Implementation of [Uri.Uri].
- factory _Uri({String scheme,
- String userInfo,
- String host,
- int port,
- String path,
- Iterable<String> pathSegments,
- String query,
- Map<String, dynamic/*String|Iterable<String>*/> queryParameters,
- String fragment}) {
- scheme = _makeScheme(scheme, 0, _stringOrNullLength(scheme));
- userInfo = _makeUserInfo(userInfo, 0, _stringOrNullLength(userInfo));
- host = _makeHost(host, 0, _stringOrNullLength(host), false);
- // Special case this constructor for backwards compatibility.
- if (query == "") query = null;
- query = _makeQuery(query, 0, _stringOrNullLength(query), queryParameters);
- fragment = _makeFragment(fragment, 0, _stringOrNullLength(fragment));
- port = _makePort(port, scheme);
- bool isFile = (scheme == "file");
- if (host == null &&
- (userInfo.isNotEmpty || port != null || isFile)) {
- host = "";
- }
+ assert(state == NOT_IN_PATH);
bool hasAuthority = (host != null);
- path = _makePath(path, 0, _stringOrNullLength(path), pathSegments,
- scheme, hasAuthority);
- if (scheme.isEmpty && host == null && !path.startsWith('/')) {
- path = _normalizeRelativePath(path);
- } else {
- path = _removeDotSegments(path);
- }
- return new _Uri._internal(scheme, userInfo, host, port,
- path, query, fragment);
- }
-
- /// Implementation of [Uri.http].
- factory _Uri.http(String authority,
- String unencodedPath,
- [Map<String, String> queryParameters]) {
- return _makeHttpUri("http", authority, unencodedPath, queryParameters);
- }
-
- /// Implementation of [Uri.https].
- factory _Uri.https(String authority,
- String unencodedPath,
- [Map<String, String> queryParameters]) {
- return _makeHttpUri("https", authority, unencodedPath, queryParameters);
- }
+ path = _makePath(uri, pathStart, index, null, scheme, hasAuthority);
- String get authority {
- if (!hasAuthority) return "";
- var sb = new StringBuffer();
- _writeAuthority(sb);
- return sb.toString();
- }
-
- String get userInfo => _userInfo;
-
- String get host {
- if (_host == null) return "";
- if (_host.startsWith('[')) {
- return _host.substring(1, _host.length - 1);
+ if (char == _QUESTION) {
+ int numberSignIndex = -1;
+ for (int i = index + 1; i < end; i++) {
+ if (uri.codeUnitAt(i) == _NUMBER_SIGN) {
+ numberSignIndex = i;
+ break;
+ }
+ }
+ if (numberSignIndex < 0) {
+ query = _makeQuery(uri, index + 1, end, null);
+ } else {
+ query = _makeQuery(uri, index + 1, numberSignIndex, null);
+ fragment = _makeFragment(uri, numberSignIndex + 1, end);
+ }
+ } else if (char == _NUMBER_SIGN) {
+ fragment = _makeFragment(uri, index + 1, end);
}
- return _host;
- }
-
- int get port {
- if (_port == null) return _defaultPort(scheme);
- return _port;
+ return new Uri._internal(scheme,
+ userinfo,
+ host,
+ port,
+ path,
+ query,
+ fragment);
}
- // The default port for the scheme of this Uri.
- static int _defaultPort(String scheme) {
- if (scheme == "http") return 80;
- if (scheme == "https") return 443;
- return 0;
- }
-
- String get path => _path;
-
- String get query => _query ?? "";
-
- String get fragment => _fragment ?? "";
-
// Report a parse failure.
static void _fail(String uri, int index, String message) {
throw new FormatException(message, uri, index);
@@ -1545,8 +612,7 @@ class _Uri implements Uri {
// Split off the user info.
bool hasUserInfo = false;
for (int i = 0; i < authority.length; i++) {
- const int atSign = 0x40;
- if (authority.codeUnitAt(i) == atSign) {
+ if (authority.codeUnitAt(i) == _AT_SIGN) {
hasUserInfo = true;
userInfo = authority.substring(0, i);
hostStart = i + 1;
@@ -1564,7 +630,7 @@ class _Uri implements Uri {
throw new FormatException("Invalid IPv6 host entry.",
authority, hostStart);
}
- Uri.parseIPv6Address(authority, hostStart + 1, hostEnd);
+ parseIPv6Address(authority, hostStart + 1, hostEnd);
hostEnd++; // Skip the closing bracket.
if (hostEnd != authority.length &&
authority.codeUnitAt(hostEnd) != _COLON) {
@@ -1592,22 +658,185 @@ class _Uri implements Uri {
queryParameters: queryParameters);
}
- /// Implementation of [Uri.file].
- factory _Uri.file(String path, {bool windows}) {
- windows = (windows == null) ? _Uri._isWindows : windows;
+ /**
+ * Creates a new file URI from an absolute or relative file path.
+ *
+ * The file path is passed in [path].
+ *
+ * This path is interpreted using either Windows or non-Windows
+ * semantics.
+ *
+ * With non-Windows semantics the slash ("/") is used to separate
+ * path segments.
+ *
+ * With Windows semantics, backslash ("\\") and forward-slash ("/")
+ * are used to separate path segments, except if the path starts
+ * with "\\\\?\\" in which case, only backslash ("\\") separates path
+ * segments.
+ *
+ * If the path starts with a path separator an absolute URI is
+ * created. Otherwise a relative URI is created. One exception from
+ * this rule is that when Windows semantics is used and the path
+ * starts with a drive letter followed by a colon (":") and a
+ * path separator then an absolute URI is created.
+ *
+ * The default for whether to use Windows or non-Windows semantics
+ * determined from the platform Dart is running on. When running in
+ * the standalone VM this is detected by the VM based on the
+ * operating system. When running in a browser non-Windows semantics
+ * is always used.
+ *
+ * To override the automatic detection of which semantics to use pass
+ * a value for [windows]. Passing `true` will use Windows
+ * semantics and passing `false` will use non-Windows semantics.
+ *
+ * Examples using non-Windows semantics:
+ *
+ * ```
+ * // xxx/yyy
+ * new Uri.file("xxx/yyy", windows: false);
+ *
+ * // xxx/yyy/
+ * new Uri.file("xxx/yyy/", windows: false);
+ *
+ * // file:///xxx/yyy
+ * new Uri.file("/xxx/yyy", windows: false);
+ *
+ * // file:///xxx/yyy/
+ * new Uri.file("/xxx/yyy/", windows: false);
+ *
+ * // C:
+ * new Uri.file("C:", windows: false);
+ * ```
+ *
+ * Examples using Windows semantics:
+ *
+ * ```
+ * // xxx/yyy
+ * new Uri.file(r"xxx\yyy", windows: true);
+ *
+ * // xxx/yyy/
+ * new Uri.file(r"xxx\yyy\", windows: true);
+ *
+ * file:///xxx/yyy
+ * new Uri.file(r"\xxx\yyy", windows: true);
+ *
+ * file:///xxx/yyy/
+ * new Uri.file(r"\xxx\yyy/", windows: true);
+ *
+ * // file:///C:/xxx/yyy
+ * new Uri.file(r"C:\xxx\yyy", windows: true);
+ *
+ * // This throws an error. A path with a drive letter is not absolute.
+ * new Uri.file(r"C:", windows: true);
+ *
+ * // This throws an error. A path with a drive letter is not absolute.
+ * new Uri.file(r"C:xxx\yyy", windows: true);
+ *
+ * // file://server/share/file
+ * new Uri.file(r"\\server\share\file", windows: true);
+ * ```
+ *
+ * If the path passed is not a legal file path [ArgumentError] is thrown.
+ */
+ factory Uri.file(String path, {bool windows}) {
+ windows = (windows == null) ? Uri._isWindows : windows;
return windows ? _makeWindowsFileUrl(path, false)
: _makeFileUri(path, false);
}
- /// Implementation of [Uri.directory].
- factory _Uri.directory(String path, {bool windows}) {
- windows = (windows == null) ? _Uri._isWindows : windows;
+ /**
+ * Like [Uri.file] except that a non-empty URI path ends in a slash.
+ *
+ * If [path] is not empty, and it doesn't end in a directory separator,
+ * then a slash is added to the returned URI's path.
+ * In all other cases, the result is the same as returned by `Uri.file`.
+ */
+ factory Uri.directory(String path, {bool windows}) {
+ windows = (windows == null) ? Uri._isWindows : windows;
return windows ? _makeWindowsFileUrl(path, true)
: _makeFileUri(path, true);
}
+ /**
+ * Creates a `data:` URI containing the [content] string.
+ *
+ * Converts the content to a bytes using [encoding] or the charset specified
+ * in [parameters] (defaulting to US-ASCII if not specified or unrecognized),
+ * then encodes the bytes into the resulting data URI.
+ *
+ * Defaults to encoding using percent-encoding (any non-ASCII or non-URI-valid
+ * bytes is replaced by a percent encoding). If [base64] is true, the bytes
+ * are instead encoded using [BASE64].
+ *
+ * If [encoding] is not provided and [parameters] has a `charset` entry,
+ * that name is looked up using [Encoding.getByName],
+ * and if the lookup returns an encoding, that encoding is used to convert
+ * [content] to bytes.
+ * If providing both an [encoding] and a charset [parameter], they should
+ * agree, otherwise decoding won't be able to use the charset parameter
+ * to determine the encoding.
+ *
+ * If [mimeType] and/or [parameters] are supplied, they are added to the
+ * created URI. If any of these contain characters that are not allowed
+ * in the data URI, the character is percent-escaped. If the character is
+ * non-ASCII, it is first UTF-8 encoded and then the bytes are percent
+ * encoded. An omitted [mimeType] in a data URI means `text/plain`, just
+ * as an omitted `charset` parameter defaults to meaning `US-ASCII`.
+ *
+ * To read the content back, use [UriData.contentAsString].
+ */
+ factory Uri.dataFromString(String content,
+ {String mimeType,
+ Encoding encoding,
+ Map<String, String> parameters,
+ bool base64: false}) {
+ UriData data = new UriData.fromString(content,
+ mimeType: mimeType,
+ encoding: encoding,
+ parameters: parameters,
+ base64: base64);
+ return data.uri;
+ }
+
+ /**
+ * Creates a `data:` URI containing an encoding of [bytes].
+ *
+ * Defaults to Base64 encoding the bytes, but if [percentEncoded]
+ * is `true`, the bytes will instead be percent encoded (any non-ASCII
+ * or non-valid-ASCII-character byte is replaced by a percent encoding).
+ *
+ * To read the bytes back, use [UriData.contentAsBytes].
+ *
+ * It defaults to having the mime-type `application/octet-stream`.
+ * The [mimeType] and [parameters] are added to the created URI.
+ * If any of these contain characters that are not allowed
+ * in the data URI, the character is percent-escaped. If the character is
+ * non-ASCII, it is first UTF-8 encoded and then the bytes are percent
+ * encoded.
+ */
+ factory Uri.dataFromBytes(List<int> bytes,
+ {mimeType: "application/octet-stream",
+ Map<String, String> parameters,
+ percentEncoded: false}) {
+ UriData data = new UriData.fromBytes(bytes,
+ mimeType: mimeType,
+ parameters: parameters,
+ percentEncoded: percentEncoded);
+ return data.uri;
+ }
+
+ /**
+ * Returns the natural base URI for the current platform.
+ *
+ * When running in a browser this is the current URL (from
+ * `window.location.href`).
+ *
+ * When not running in a browser this is the file URI referencing
+ * the current working directory.
+ */
+ external static Uri get base;
- /// Used internally in path-related constructors.
external static bool get _isWindows;
static _checkNonWindowsPathReservedCharacters(List<String> segments,
@@ -1740,6 +969,46 @@ class _Uri implements Uri {
}
}
+ /**
+ * Returns a new `Uri` based on this one, but with some parts replaced.
+ *
+ * This method takes the same parameters as the [new Uri] constructor,
+ * and they have the same meaning.
+ *
+ * At most one of [path] and [pathSegments] must be provided.
+ * Likewise, at most one of [query] and [queryParameters] must be provided.
+ *
+ * Each part that is not provided will default to the corresponding
+ * value from this `Uri` instead.
+ *
+ * This method is different from [Uri.resolve] which overrides in a
+ * hierarchial manner,
+ * and can instead replace each part of a `Uri` individually.
+ *
+ * Example:
+ *
+ * Uri uri1 = Uri.parse("a://b@c:4/d/e?f#g");
+ * Uri uri2 = uri1.replace(scheme: "A", path: "D/E/E", fragment: "G");
+ * print(uri2); // prints "A://b@c:4/D/E/E/?f#G"
+ *
+ * This method acts similarly to using the `new Uri` constructor with
+ * some of the arguments taken from this `Uri` . Example:
+ *
+ * Uri uri3 = new Uri(
+ * scheme: "A",
+ * userInfo: uri1.userInfo,
+ * host: uri1.host,
+ * port: uri1.port,
+ * path: "D/E/E",
+ * query: uri1.query,
+ * fragment: "G");
+ * print(uri3); // prints "A://b@c:4/D/E/E/?f#G"
+ * print(uri2 == uri3); // prints true.
+ *
+ * Using this method can be seen as a shorthand for the `Uri` constructor
+ * call above, but may also be slightly faster because the parts taken
+ * from this `Uri` need not be checked for validity again.
+ */
Uri replace({String scheme,
String userInfo,
String host,
@@ -1755,7 +1024,7 @@ class _Uri implements Uri {
bool schemeChanged = false;
if (scheme != null) {
scheme = _makeScheme(scheme, 0, scheme.length);
- schemeChanged = (scheme != this.scheme);
+ schemeChanged = true;
} else {
scheme = this.scheme;
}
@@ -1806,16 +1075,29 @@ class _Uri implements Uri {
fragment = this._fragment;
}
- return new _Uri._internal(
+ return new Uri._internal(
scheme, userInfo, host, port, path, query, fragment);
}
+ /**
+ * Returns a `Uri` that differs from this only in not having a fragment.
+ *
+ * If this `Uri` does not have a fragment, it is itself returned.
+ */
Uri removeFragment() {
if (!this.hasFragment) return this;
- return new _Uri._internal(scheme, _userInfo, _host, _port,
+ return new Uri._internal(scheme, _userInfo, _host, _port,
_path, _query, null);
}
+ /**
+ * Returns the URI path split into its segments. Each of the segments in the
+ * returned list have been decoded. If the path is empty the empty list will
+ * be returned. A leading slash `/` does not affect the segments returned.
+ *
+ * The returned list is unmodifiable and will throw [UnsupportedError] on any
+ * calls that would mutate it.
+ */
List<String> get pathSegments {
var result = _pathSegments;
if (result != null) return result;
@@ -1832,14 +1114,43 @@ class _Uri implements Uri {
return result;
}
+ /**
+ * Returns the URI query split into a map according to the rules
+ * specified for FORM post in the [HTML 4.01 specification section
+ * 17.13.4](http://www.w3.org/TR/REC-html40/interact/forms.html#h-17.13.4 "HTML 4.01 section 17.13.4").
+ * Each key and value in the returned map has been decoded.
+ * If there is no query the empty map is returned.
+ *
+ * Keys in the query string that have no value are mapped to the
+ * empty string.
+ * If a key occurs more than once in the query string, it is mapped to
+ * an arbitrary choice of possible value.
+ * The [queryParametersAll] getter can provide a map
+ * that maps keys to all of their values.
+ *
+ * The returned map is unmodifiable.
+ */
Map<String, String> get queryParameters {
if (_queryParameters == null) {
_queryParameters =
- new UnmodifiableMapView<String, String>(Uri.splitQueryString(query));
+ new UnmodifiableMapView<String, String>(splitQueryString(query));
}
return _queryParameters;
}
+ /**
+ * Returns the URI query split into a map according to the rules
+ * specified for FORM post in the [HTML 4.01 specification section
+ * 17.13.4](http://www.w3.org/TR/REC-html40/interact/forms.html#h-17.13.4 "HTML 4.01 section 17.13.4").
+ * Each key and value in the returned map has been decoded. If there is no
+ * query the empty map is returned.
+ *
+ * Keys are mapped to lists of their values. If a key occurs only once,
+ * its value is a singleton list. If a key occurs with no value, the
+ * empty string is used as the value for that occurrence.
+ *
+ * The returned map and the lists it contains are unmodifiable.
+ */
Map<String, List<String>> get queryParametersAll {
if (_queryParameterLists == null) {
Map queryParameterLists = _splitQueryStringAll(query);
@@ -1853,6 +1164,22 @@ class _Uri implements Uri {
return _queryParameterLists;
}
+ /**
+ * Returns a URI where the path has been normalized.
+ *
+ * A normalized path does not contain `.` segments or non-leading `..`
+ * segments.
+ * Only a relative path with no scheme or authority may contain
+ * leading `..` segments,
+ * a path that starts with `/` will also drop any leading `..` segments.
+ *
+ * This uses the same normalization strategy as `new Uri().resolve(this)`.
+ *
+ * Does not change any part of the URI except the path.
+ *
+ * The default implementation of `Uri` always normalizes paths, so calling
+ * this function has no effect.
+ */
Uri normalizePath() {
String path = _normalizePath(_path, scheme, hasAuthority);
if (identical(path, _path)) return this;
@@ -1885,7 +1212,7 @@ class _Uri implements Uri {
if (host.codeUnitAt(end - 1) != _RIGHT_BRACKET) {
_fail(host, start, 'Missing end `]` to match `[` in host');
}
- Uri.parseIPv6Address(host, start + 1, end - 1);
+ parseIPv6Address(host, start + 1, end - 1);
// RFC 5952 requires hex digits to be lower case.
return host.substring(start, end).toLowerCase();
}
@@ -1893,7 +1220,7 @@ class _Uri implements Uri {
// TODO(lrn): skip if too short to be a valid IPv6 address?
for (int i = start; i < end; i++) {
if (host.codeUnitAt(i) == _COLON) {
- Uri.parseIPv6Address(host, start, end);
+ parseIPv6Address(host, start, end);
return '[$host]';
}
}
@@ -2006,17 +1333,6 @@ class _Uri implements Uri {
}
scheme = scheme.substring(start, end);
if (containsUpperCase) scheme = scheme.toLowerCase();
- return _canonicalizeScheme(scheme);
- }
-
- // Canonicalize a few often-used scheme strings.
- //
- // This improves memory usage and makes comparison faster.
- static String _canonicalizeScheme(String scheme) {
- if (scheme == "http") return "http";
- if (scheme == "file") return "file";
- if (scheme == "https") return "https";
- if (scheme == "package") return "package";
return scheme;
}
@@ -2103,6 +1419,8 @@ class _Uri implements Uri {
return _normalize(fragment, start, end, _queryCharTable);
}
+ static int _stringOrNullLength(String s) => (s == null) ? 0 : s.length;
+
/**
* Performs RFC 3986 Percent-Encoding Normalization.
*
@@ -2147,11 +1465,10 @@ class _Uri implements Uri {
// Converts a UTF-16 code-unit to its value as a hex digit.
// Returns -1 for non-hex digits.
static int _parseHexDigit(int char) {
- const int zeroDigit = 0x30;
- int digit = char ^ zeroDigit;
+ int digit = char ^ Uri._ZERO;
if (digit <= 9) return digit;
int lowerCase = char | 0x20;
- if (_LOWER_CASE_A <= lowerCase && lowerCase <= _LOWER_CASE_F) {
+ if (Uri._LOWER_CASE_A <= lowerCase && lowerCase <= _LOWER_CASE_F) {
return lowerCase - (_LOWER_CASE_A - 10);
}
return -1;
@@ -2237,7 +1554,7 @@ class _Uri implements Uri {
if (index + 1 < end) {
int tail = component.codeUnitAt(index + 1);
if ((tail & 0xFC00) == 0xDC00) {
- // Tail surrogate.
+ // Tail surrogat.
sourceLength = 2;
char = 0x10000 | ((char & 0x3ff) << 10) | (tail & 0x3ff);
}
@@ -2386,10 +1703,34 @@ class _Uri implements Uri {
return output.join("/");
}
+ /**
+ * Resolve [reference] as an URI relative to `this`.
+ *
+ * First turn [reference] into a URI using [Uri.parse]. Then resolve the
+ * resulting URI relative to `this`.
+ *
+ * Returns the resolved URI.
+ *
+ * See [resolveUri] for details.
+ */
Uri resolve(String reference) {
return resolveUri(Uri.parse(reference));
}
+ /**
+ * Resolve [reference] as an URI relative to `this`.
+ *
+ * Returns the resolved URI.
+ *
+ * The algorithm "Transform Reference" for resolving a reference is described
+ * in [RFC-3986 Section 5](http://tools.ietf.org/html/rfc3986#section-5 "RFC-1123").
+ *
+ * Updated to handle the case where the base URI is just a relative path -
+ * that is: when it has no scheme or authority and the path does not start
+ * with a slash.
+ * In that case, the paths are combined without removing leading "..", and
+ * an empty path is not converted to "/".
+ */
Uri resolveUri(Uri reference) {
// From RFC 3986.
String targetScheme;
@@ -2435,17 +1776,11 @@ class _Uri implements Uri {
} else {
// This is the RFC 3986 behavior for merging.
if (this.hasEmptyPath) {
- if (!this.hasAuthority) {
- if (!this.hasScheme) {
- // Keep the path relative if no scheme or authority.
- targetPath = reference.path;
- } else {
- // Remove leading dot-segments if the path is put
- // beneath a scheme.
- targetPath = _removeDotSegments(reference.path);
- }
+ if (!this.hasScheme && !this.hasAuthority) {
+ // Keep the path relative if no scheme or authority.
+ targetPath = reference.path;
} else {
- // RFC algorithm for base with authority and empty path.
+ // Add path normalization on top of RFC algorithm.
targetPath = _removeDotSegments("/" + reference.path);
}
} else {
@@ -2453,9 +1788,8 @@ class _Uri implements Uri {
if (this.hasScheme || this.hasAuthority || this.hasAbsolutePath) {
targetPath = _removeDotSegments(mergedPath);
} else {
- // Non-RFC 3986 behavior.
- // If both base and reference are relative paths,
- // allow the merged path to start with "..".
+ // Non-RFC 3986 beavior. If both base and reference are relative
+ // path, allow the merged path to start with "..".
// The RFC only specifies the case where the base has a scheme.
targetPath = _normalizeRelativePath(mergedPath);
}
@@ -2466,208 +1800,626 @@ class _Uri implements Uri {
}
}
String fragment = reference.hasFragment ? reference.fragment : null;
- return new _Uri._internal(targetScheme,
- targetUserInfo,
- targetHost,
- targetPort,
- targetPath,
- targetQuery,
- fragment);
+ return new Uri._internal(targetScheme,
+ targetUserInfo,
+ targetHost,
+ targetPort,
+ targetPath,
+ targetQuery,
+ fragment);
+ }
+
+ /**
+ * Returns whether the URI has a [scheme] component.
+ */
+ bool get hasScheme => scheme.isNotEmpty;
+
+ /**
+ * Returns whether the URI has an [authority] component.
+ */
+ bool get hasAuthority => _host != null;
+
+ /**
+ * Returns whether the URI has an explicit port.
+ *
+ * If the port number is the default port number
+ * (zero for unrecognized schemes, with http (80) and https (443) being
+ * recognized),
+ * then the port is made implicit and omitted from the URI.
+ */
+ bool get hasPort => _port != null;
+
+ /**
+ * Returns whether the URI has a query part.
+ */
+ bool get hasQuery => _query != null;
+
+ /**
+ * Returns whether the URI has a fragment part.
+ */
+ bool get hasFragment => _fragment != null;
+
+ /**
+ * Returns whether the URI has an empty path.
+ */
+ bool get hasEmptyPath => _path.isEmpty;
+
+ /**
+ * Returns whether the URI has an absolute path (starting with '/').
+ */
+ bool get hasAbsolutePath => _path.startsWith('/');
+
+ /**
+ * Returns the origin of the URI in the form scheme://host:port for the
+ * schemes http and https.
+ *
+ * It is an error if the scheme is not "http" or "https".
+ *
+ * See: http://www.w3.org/TR/2011/WD-html5-20110405/origin-0.html#origin
+ */
+ String get origin {
+ if (scheme == "" || _host == null || _host == "") {
+ throw new StateError("Cannot use origin without a scheme: $this");
+ }
+ if (scheme != "http" && scheme != "https") {
+ throw new StateError(
+ "Origin is only applicable schemes http and https: $this");
+ }
+ if (_port == null) return "$scheme://$_host";
+ return "$scheme://$_host:$_port";
+ }
+
+ /**
+ * Returns the file path from a file URI.
+ *
+ * The returned path has either Windows or non-Windows
+ * semantics.
+ *
+ * For non-Windows semantics the slash ("/") is used to separate
+ * path segments.
+ *
+ * For Windows semantics the backslash ("\\") separator is used to
+ * separate path segments.
+ *
+ * If the URI is absolute the path starts with a path separator
+ * unless Windows semantics is used and the first path segment is a
+ * drive letter. When Windows semantics is used a host component in
+ * the uri in interpreted as a file server and a UNC path is
+ * returned.
+ *
+ * The default for whether to use Windows or non-Windows semantics
+ * determined from the platform Dart is running on. When running in
+ * the standalone VM this is detected by the VM based on the
+ * operating system. When running in a browser non-Windows semantics
+ * is always used.
+ *
+ * To override the automatic detection of which semantics to use pass
+ * a value for [windows]. Passing `true` will use Windows
+ * semantics and passing `false` will use non-Windows semantics.
+ *
+ * If the URI ends with a slash (i.e. the last path component is
+ * empty) the returned file path will also end with a slash.
+ *
+ * With Windows semantics URIs starting with a drive letter cannot
+ * be relative to the current drive on the designated drive. That is
+ * for the URI `file:///c:abc` calling `toFilePath` will throw as a
+ * path segment cannot contain colon on Windows.
+ *
+ * Examples using non-Windows semantics (resulting of calling
+ * toFilePath in comment):
+ *
+ * Uri.parse("xxx/yyy"); // xxx/yyy
+ * Uri.parse("xxx/yyy/"); // xxx/yyy/
+ * Uri.parse("file:///xxx/yyy"); // /xxx/yyy
+ * Uri.parse("file:///xxx/yyy/"); // /xxx/yyy/
+ * Uri.parse("file:///C:"); // /C:
+ * Uri.parse("file:///C:a"); // /C:a
+ *
+ * Examples using Windows semantics (resulting URI in comment):
+ *
+ * Uri.parse("xxx/yyy"); // xxx\yyy
+ * Uri.parse("xxx/yyy/"); // xxx\yyy\
+ * Uri.parse("file:///xxx/yyy"); // \xxx\yyy
+ * Uri.parse("file:///xxx/yyy/"); // \xxx\yyy/
+ * Uri.parse("file:///C:/xxx/yyy"); // C:\xxx\yyy
+ * Uri.parse("file:C:xxx/yyy"); // Throws as a path segment
+ * // cannot contain colon on Windows.
+ * Uri.parse("file://server/share/file"); // \\server\share\file
+ *
+ * If the URI is not a file URI calling this throws
+ * [UnsupportedError].
+ *
+ * If the URI cannot be converted to a file path calling this throws
+ * [UnsupportedError].
+ */
+ String toFilePath({bool windows}) {
+ if (scheme != "" && scheme != "file") {
+ throw new UnsupportedError(
+ "Cannot extract a file path from a $scheme URI");
+ }
+ if (query != "") {
+ throw new UnsupportedError(
+ "Cannot extract a file path from a URI with a query component");
+ }
+ if (fragment != "") {
+ throw new UnsupportedError(
+ "Cannot extract a file path from a URI with a fragment component");
+ }
+ if (windows == null) windows = _isWindows;
+ return windows ? _toWindowsFilePath() : _toFilePath();
+ }
+
+ String _toFilePath() {
+ if (host != "") {
+ throw new UnsupportedError(
+ "Cannot extract a non-Windows file path from a file URI "
+ "with an authority");
+ }
+ _checkNonWindowsPathReservedCharacters(pathSegments, false);
+ var result = new StringBuffer();
+ if (_isPathAbsolute) result.write("/");
+ result.writeAll(pathSegments, "/");
+ return result.toString();
+ }
+
+ String _toWindowsFilePath() {
+ bool hasDriveLetter = false;
+ var segments = pathSegments;
+ if (segments.length > 0 &&
+ segments[0].length == 2 &&
+ segments[0].codeUnitAt(1) == _COLON) {
+ _checkWindowsDriveLetter(segments[0].codeUnitAt(0), false);
+ _checkWindowsPathReservedCharacters(segments, false, 1);
+ hasDriveLetter = true;
+ } else {
+ _checkWindowsPathReservedCharacters(segments, false);
+ }
+ var result = new StringBuffer();
+ if (_isPathAbsolute && !hasDriveLetter) result.write("\\");
+ if (host != "") {
+ result.write("\\");
+ result.write(host);
+ result.write("\\");
+ }
+ result.writeAll(segments, "\\");
+ if (hasDriveLetter && segments.length == 1) result.write("\\");
+ return result.toString();
+ }
+
+ bool get _isPathAbsolute {
+ if (path == null || path.isEmpty) return false;
+ return path.startsWith('/');
+ }
+
+ void _writeAuthority(StringSink ss) {
+ if (_userInfo.isNotEmpty) {
+ ss.write(_userInfo);
+ ss.write("@");
+ }
+ if (_host != null) ss.write(_host);
+ if (_port != null) {
+ ss.write(":");
+ ss.write(_port);
+ }
+ }
+
+ /**
+ * Access the structure of a `data:` URI.
+ *
+ * Returns a [UriData] object for `data:` URIs and `null` for all other
+ * URIs.
+ * The [UriData] object can be used to access the media type and data
+ * of a `data:` URI.
+ */
+ UriData get data => (scheme == "data") ? new UriData.fromUri(this) : null;
+
+ String toString() {
+ StringBuffer sb = new StringBuffer();
+ _addIfNonEmpty(sb, scheme, scheme, ':');
+ if (hasAuthority || path.startsWith("//") || (scheme == "file")) {
+ // File URIS always have the authority, even if it is empty.
+ // The empty URI means "localhost".
+ sb.write("//");
+ _writeAuthority(sb);
+ }
+ sb.write(path);
+ if (_query != null) { sb..write("?")..write(_query); }
+ if (_fragment != null) { sb..write("#")..write(_fragment); }
+ return sb.toString();
+ }
+
+ bool operator==(other) {
+ if (other is! Uri) return false;
+ Uri uri = other;
+ return scheme == uri.scheme &&
+ hasAuthority == uri.hasAuthority &&
+ userInfo == uri.userInfo &&
+ host == uri.host &&
+ port == uri.port &&
+ path == uri.path &&
+ hasQuery == uri.hasQuery &&
+ query == uri.query &&
+ hasFragment == uri.hasFragment &&
+ fragment == uri.fragment;
+ }
+
+ int get hashCode {
+ int combine(part, current) {
+ // The sum is truncated to 30 bits to make sure it fits into a Smi.
+ return (current * 31 + part.hashCode) & 0x3FFFFFFF;
+ }
+ return combine(scheme, combine(userInfo, combine(host, combine(port,
+ combine(path, combine(query, combine(fragment, 1)))))));
+ }
+
+ static void _addIfNonEmpty(StringBuffer sb, String test,
+ String first, String second) {
+ if ("" != test) {
+ sb.write(first);
+ sb.write(second);
+ }
+ }
+
+ /**
+ * Encode the string [component] using percent-encoding to make it
+ * safe for literal use as a URI component.
+ *
+ * All characters except uppercase and lowercase letters, digits and
+ * the characters `-_.!~*'()` are percent-encoded. This is the
+ * set of characters specified in RFC 2396 and the which is
+ * specified for the encodeUriComponent in ECMA-262 version 5.1.
+ *
+ * When manually encoding path segments or query components remember
+ * to encode each part separately before building the path or query
+ * string.
+ *
+ * For encoding the query part consider using
+ * [encodeQueryComponent].
+ *
+ * To avoid the need for explicitly encoding use the [pathSegments]
+ * and [queryParameters] optional named arguments when constructing
+ * a [Uri].
+ */
+ static String encodeComponent(String component) {
+ return _uriEncode(_unreserved2396Table, component, UTF8, false);
+ }
+
+ /**
+ * Encode the string [component] according to the HTML 4.01 rules
+ * for encoding the posting of a HTML form as a query string
+ * component.
+ *
+ * Encode the string [component] according to the HTML 4.01 rules
+ * for encoding the posting of a HTML form as a query string
+ * component.
+
+ * The component is first encoded to bytes using [encoding].
+ * The default is to use [UTF8] encoding, which preserves all
+ * the characters that don't need encoding.
+
+ * Then the resulting bytes are "percent-encoded". This transforms
+ * spaces (U+0020) to a plus sign ('+') and all bytes that are not
+ * the ASCII decimal digits, letters or one of '-._~' are written as
+ * a percent sign '%' followed by the two-digit hexadecimal
+ * representation of the byte.
+
+ * Note that the set of characters which are percent-encoded is a
+ * superset of what HTML 4.01 requires, since it refers to RFC 1738
+ * for reserved characters.
+ *
+ * When manually encoding query components remember to encode each
+ * part separately before building the query string.
+ *
+ * To avoid the need for explicitly encoding the query use the
+ * [queryParameters] optional named arguments when constructing a
+ * [Uri].
+ *
+ * See http://www.w3.org/TR/html401/interact/forms.html#h-17.13.4.2 for more
+ * details.
+ */
+ static String encodeQueryComponent(String component,
+ {Encoding encoding: UTF8}) {
+ return _uriEncode(_unreservedTable, component, encoding, true);
}
- bool get hasScheme => scheme.isNotEmpty;
+ /**
+ * Decodes the percent-encoding in [encodedComponent].
+ *
+ * Note that decoding a URI component might change its meaning as
+ * some of the decoded characters could be characters with are
+ * delimiters for a given URI componene type. Always split a URI
+ * component using the delimiters for the component before decoding
+ * the individual parts.
+ *
+ * For handling the [path] and [query] components consider using
+ * [pathSegments] and [queryParameters] to get the separated and
+ * decoded component.
+ */
+ static String decodeComponent(String encodedComponent) {
+ return _uriDecode(encodedComponent, 0, encodedComponent.length,
+ UTF8, false);
+ }
- bool get hasAuthority => _host != null;
+ /**
+ * Decodes the percent-encoding in [encodedComponent], converting
+ * pluses to spaces.
+ *
+ * It will create a byte-list of the decoded characters, and then use
+ * [encoding] to decode the byte-list to a String. The default encoding is
+ * UTF-8.
+ */
+ static String decodeQueryComponent(
+ String encodedComponent,
+ {Encoding encoding: UTF8}) {
+ return _uriDecode(encodedComponent, 0, encodedComponent.length,
+ encoding, true);
+ }
- bool get hasPort => _port != null;
+ /**
+ * Encode the string [uri] using percent-encoding to make it
+ * safe for literal use as a full URI.
+ *
+ * All characters except uppercase and lowercase letters, digits and
+ * the characters `!#$&'()*+,-./:;=?@_~` are percent-encoded. This
+ * is the set of characters specified in in ECMA-262 version 5.1 for
+ * the encodeURI function .
+ */
+ static String encodeFull(String uri) {
+ return _uriEncode(_encodeFullTable, uri, UTF8, false);
+ }
- bool get hasQuery => _query != null;
+ /**
+ * Decodes the percent-encoding in [uri].
+ *
+ * Note that decoding a full URI might change its meaning as some of
+ * the decoded characters could be reserved characters. In most
+ * cases an encoded URI should be parsed into components using
+ * [Uri.parse] before decoding the separate components.
+ */
+ static String decodeFull(String uri) {
+ return _uriDecode(uri, 0, uri.length, UTF8, false);
+ }
- bool get hasFragment => _fragment != null;
+ /**
+ * Returns the [query] split into a map according to the rules
+ * specified for FORM post in the [HTML 4.01 specification section
+ * 17.13.4](http://www.w3.org/TR/REC-html40/interact/forms.html#h-17.13.4 "HTML 4.01 section 17.13.4").
+ * Each key and value in the returned map has been decoded. If the [query]
+ * is the empty string an empty map is returned.
+ *
+ * Keys in the query string that have no value are mapped to the
+ * empty string.
+ *
+ * Each query component will be decoded using [encoding]. The default encoding
+ * is UTF-8.
+ */
+ static Map<String, String> splitQueryString(String query,
+ {Encoding encoding: UTF8}) {
+ return query.split("&").fold({}, (map, element) {
+ int index = element.indexOf("=");
+ if (index == -1) {
+ if (element != "") {
+ map[decodeQueryComponent(element, encoding: encoding)] = "";
+ }
+ } else if (index != 0) {
+ var key = element.substring(0, index);
+ var value = element.substring(index + 1);
+ map[Uri.decodeQueryComponent(key, encoding: encoding)] =
+ decodeQueryComponent(value, encoding: encoding);
+ }
+ return map;
+ });
+ }
- bool get hasEmptyPath => _path.isEmpty;
+ static List _createList() => [];
- bool get hasAbsolutePath => _path.startsWith('/');
+ static Map _splitQueryStringAll(
+ String query, {Encoding encoding: UTF8}) {
+ Map result = {};
+ int i = 0;
+ int start = 0;
+ int equalsIndex = -1;
- String get origin {
- if (scheme == "" || _host == null || _host == "") {
- throw new StateError("Cannot use origin without a scheme: $this");
- }
- if (scheme != "http" && scheme != "https") {
- throw new StateError(
- "Origin is only applicable schemes http and https: $this");
+ void parsePair(int start, int equalsIndex, int end) {
+ String key;
+ String value;
+ if (start == end) return;
+ if (equalsIndex < 0) {
+ key = _uriDecode(query, start, end, encoding, true);
+ value = "";
+ } else {
+ key = _uriDecode(query, start, equalsIndex, encoding, true);
+ value = _uriDecode(query, equalsIndex + 1, end, encoding, true);
+ }
+ result.putIfAbsent(key, _createList).add(value);
}
- if (_port == null) return "$scheme://$_host";
- return "$scheme://$_host:$_port";
- }
- String toFilePath({bool windows}) {
- if (scheme != "" && scheme != "file") {
- throw new UnsupportedError(
- "Cannot extract a file path from a $scheme URI");
- }
- if (query != "") {
- throw new UnsupportedError(
- "Cannot extract a file path from a URI with a query component");
- }
- if (fragment != "") {
- throw new UnsupportedError(
- "Cannot extract a file path from a URI with a fragment component");
+ const int _equals = 0x3d;
+ const int _ampersand = 0x26;
+ while (i < query.length) {
+ int char = query.codeUnitAt(i);
+ if (char == _equals) {
+ if (equalsIndex < 0) equalsIndex = i;
+ } else if (char == _ampersand) {
+ parsePair(start, equalsIndex, i);
+ start = i + 1;
+ equalsIndex = -1;
+ }
+ i++;
}
- if (windows == null) windows = _isWindows;
- return windows ? _toWindowsFilePath(this) : _toFilePath();
+ parsePair(start, equalsIndex, i);
+ return result;
}
- String _toFilePath() {
- if (hasAuthority && host != "") {
- throw new UnsupportedError(
- "Cannot extract a non-Windows file path from a file URI "
- "with an authority");
+ /**
+ * Parse the [host] as an IP version 4 (IPv4) address, returning the address
+ * as a list of 4 bytes in network byte order (big endian).
+ *
+ * Throws a [FormatException] if [host] is not a valid IPv4 address
+ * representation.
+ */
+ static List<int> parseIPv4Address(String host) {
+ void error(String msg) {
+ throw new FormatException('Illegal IPv4 address, $msg');
}
- // Use path segments to have any escapes unescaped.
- var pathSegments = this.pathSegments;
- _checkNonWindowsPathReservedCharacters(pathSegments, false);
- var result = new StringBuffer();
- if (hasAbsolutePath) result.write("/");
- result.writeAll(pathSegments, "/");
- return result.toString();
+ var bytes = host.split('.');
+ if (bytes.length != 4) {
+ error('IPv4 address should contain exactly 4 parts');
+ }
+ // TODO(ajohnsen): Consider using Uint8List.
+ return bytes
+ .map((byteString) {
+ int byte = int.parse(byteString);
+ if (byte < 0 || byte > 255) {
+ error('each part must be in the range of `0..255`');
+ }
+ return byte;
+ })
+ .toList();
}
- static String _toWindowsFilePath(Uri uri) {
- bool hasDriveLetter = false;
- var segments = uri.pathSegments;
- if (segments.length > 0 &&
- segments[0].length == 2 &&
- segments[0].codeUnitAt(1) == _COLON) {
- _checkWindowsDriveLetter(segments[0].codeUnitAt(0), false);
- _checkWindowsPathReservedCharacters(segments, false, 1);
- hasDriveLetter = true;
- } else {
- _checkWindowsPathReservedCharacters(segments, false, 0);
+ /**
+ * Parse the [host] as an IP version 6 (IPv6) address, returning the address
+ * as a list of 16 bytes in network byte order (big endian).
+ *
+ * Throws a [FormatException] if [host] is not a valid IPv6 address
+ * representation.
+ *
+ * Acts on the substring from [start] to [end]. If [end] is omitted, it
+ * defaults ot the end of the string.
+ *
+ * Some examples of IPv6 addresses:
+ * * ::1
+ * * FEDC:BA98:7654:3210:FEDC:BA98:7654:3210
+ * * 3ffe:2a00:100:7031::1
+ * * ::FFFF:129.144.52.38
+ * * 2010:836B:4179::836B:4179
+ */
+ static List<int> parseIPv6Address(String host, [int start = 0, int end]) {
+ if (end == null) end = host.length;
+ // An IPv6 address consists of exactly 8 parts of 1-4 hex digits, seperated
+ // by `:`'s, with the following exceptions:
+ //
+ // - One (and only one) wildcard (`::`) may be present, representing a fill
+ // of 0's. The IPv6 `::` is thus 16 bytes of `0`.
+ // - The last two parts may be replaced by an IPv4 address.
+ void error(String msg, [position]) {
+ throw new FormatException('Illegal IPv6 address, $msg', host, position);
}
- var result = new StringBuffer();
- if (uri.hasAbsolutePath && !hasDriveLetter) result.write(r"\");
- if (uri.hasAuthority) {
- var host = uri.host;
- if (host.isNotEmpty) {
- result.write(r"\");
- result.write(host);
- result.write(r"\");
+ int parseHex(int start, int end) {
+ if (end - start > 4) {
+ error('an IPv6 part can only contain a maximum of 4 hex digits', start);
}
+ int value = int.parse(host.substring(start, end), radix: 16);
+ if (value < 0 || value > (1 << 16) - 1) {
+ error('each part must be in the range of `0x0..0xFFFF`', start);
+ }
+ return value;
}
- result.writeAll(segments, r"\");
- if (hasDriveLetter && segments.length == 1) result.write(r"\");
- return result.toString();
- }
-
- bool get _isPathAbsolute {
- return _path != null && _path.startsWith('/');
- }
-
- void _writeAuthority(StringSink ss) {
- if (_userInfo.isNotEmpty) {
- ss.write(_userInfo);
- ss.write("@");
- }
- if (_host != null) ss.write(_host);
- if (_port != null) {
- ss.write(":");
- ss.write(_port);
+ if (host.length < 2) error('address is too short');
+ List<int> parts = [];
+ bool wildcardSeen = false;
+ int partStart = start;
+ // Parse all parts, except a potential last one.
+ for (int i = start; i < end; i++) {
+ if (host.codeUnitAt(i) == _COLON) {
+ if (i == start) {
+ // If we see a `:` in the beginning, expect wildcard.
+ i++;
+ if (host.codeUnitAt(i) != _COLON) {
+ error('invalid start colon.', i);
+ }
+ partStart = i;
+ }
+ if (i == partStart) {
+ // Wildcard. We only allow one.
+ if (wildcardSeen) {
+ error('only one wildcard `::` is allowed', i);
+ }
+ wildcardSeen = true;
+ parts.add(-1);
+ } else {
+ // Found a single colon. Parse [partStart..i] as a hex entry.
+ parts.add(parseHex(partStart, i));
+ }
+ partStart = i + 1;
+ }
}
- }
-
- /**
- * Access the structure of a `data:` URI.
- *
- * Returns a [UriData] object for `data:` URIs and `null` for all other
- * URIs.
- * The [UriData] object can be used to access the media type and data
- * of a `data:` URI.
- */
- UriData get data => (scheme == "data") ? new UriData.fromUri(this) : null;
-
- String toString() {
- return _text ??= _initializeText();
- }
-
- String _initializeText() {
- assert(_text == null);
- StringBuffer sb = new StringBuffer();
- if (scheme.isNotEmpty) sb..write(scheme)..write(":");
- if (hasAuthority || path.startsWith("//") || (scheme == "file")) {
- // File URIS always have the authority, even if it is empty.
- // The empty URI means "localhost".
- sb.write("//");
- _writeAuthority(sb);
+ if (parts.length == 0) error('too few parts');
+ bool atEnd = (partStart == end);
+ bool isLastWildcard = (parts.last == -1);
+ if (atEnd && !isLastWildcard) {
+ error('expected a part after last `:`', end);
}
- sb.write(path);
- if (_query != null) sb..write("?")..write(_query);
- if (_fragment != null) sb..write("#")..write(_fragment);
- return sb.toString();
- }
-
- bool operator==(other) {
- if (identical(this, other)) return true;
- if (other is Uri) {
- Uri uri = other;
- return scheme == uri.scheme &&
- hasAuthority == uri.hasAuthority &&
- userInfo == uri.userInfo &&
- host == uri.host &&
- port == uri.port &&
- path == uri.path &&
- hasQuery == uri.hasQuery &&
- query == uri.query &&
- hasFragment == uri.hasFragment &&
- fragment == uri.fragment;
+ if (!atEnd) {
+ try {
+ parts.add(parseHex(partStart, end));
+ } catch (e) {
+ // Failed to parse the last chunk as hex. Try IPv4.
+ try {
+ List<int> last = parseIPv4Address(host.substring(partStart, end));
+ parts.add(last[0] << 8 | last[1]);
+ parts.add(last[2] << 8 | last[3]);
+ } catch (e) {
+ error('invalid end of IPv6 address.', partStart);
+ }
+ }
}
- return false;
- }
-
- int get hashCode {
- return _hashCodeCache ??= toString().hashCode;
- }
-
- static List _createList() => [];
-
- static Map _splitQueryStringAll(
- String query, {Encoding encoding: UTF8}) {
- Map result = {};
- int i = 0;
- int start = 0;
- int equalsIndex = -1;
-
- void parsePair(int start, int equalsIndex, int end) {
- String key;
- String value;
- if (start == end) return;
- if (equalsIndex < 0) {
- key = _uriDecode(query, start, end, encoding, true);
- value = "";
- } else {
- key = _uriDecode(query, start, equalsIndex, encoding, true);
- value = _uriDecode(query, equalsIndex + 1, end, encoding, true);
+ if (wildcardSeen) {
+ if (parts.length > 7) {
+ error('an address with a wildcard must have less than 7 parts');
}
- result.putIfAbsent(key, _createList).add(value);
+ } else if (parts.length != 8) {
+ error('an address without a wildcard must contain exactly 8 parts');
}
-
- const int _equals = 0x3d;
- const int _ampersand = 0x26;
- while (i < query.length) {
- int char = query.codeUnitAt(i);
- if (char == _equals) {
- if (equalsIndex < 0) equalsIndex = i;
- } else if (char == _ampersand) {
- parsePair(start, equalsIndex, i);
- start = i + 1;
- equalsIndex = -1;
+ List<int> bytes = new Uint8List(16);
+ for (int i = 0, index = 0; i < parts.length; i++) {
+ int value = parts[i];
+ if (value == -1) {
+ int wildCardLength = 9 - parts.length;
+ for (int j = 0; j < wildCardLength; j++) {
+ bytes[index] = 0;
+ bytes[index + 1] = 0;
+ index += 2;
+ }
+ } else {
+ bytes[index] = value >> 8;
+ bytes[index + 1] = value & 0xff;
+ index += 2;
}
- i++;
}
- parsePair(start, equalsIndex, i);
- return result;
+ return bytes;
}
+ // Frequently used character codes.
+ static const int _SPACE = 0x20;
+ static const int _DOUBLE_QUOTE = 0x22;
+ static const int _NUMBER_SIGN = 0x23;
+ static const int _PERCENT = 0x25;
+ static const int _ASTERISK = 0x2A;
+ static const int _PLUS = 0x2B;
+ static const int _DOT = 0x2E;
+ static const int _SLASH = 0x2F;
+ static const int _ZERO = 0x30;
+ static const int _NINE = 0x39;
+ static const int _COLON = 0x3A;
+ static const int _LESS = 0x3C;
+ static const int _GREATER = 0x3E;
+ static const int _QUESTION = 0x3F;
+ static const int _AT_SIGN = 0x40;
+ static const int _UPPER_CASE_A = 0x41;
+ static const int _UPPER_CASE_F = 0x46;
+ static const int _UPPER_CASE_Z = 0x5A;
+ static const int _LEFT_BRACKET = 0x5B;
+ static const int _BACKSLASH = 0x5C;
+ static const int _RIGHT_BRACKET = 0x5D;
+ static const int _LOWER_CASE_A = 0x61;
+ static const int _LOWER_CASE_F = 0x66;
+ static const int _LOWER_CASE_Z = 0x7A;
+ static const int _BAR = 0x7C;
+
+ static const String _hexDigits = "0123456789ABCDEF";
+
external static String _uriEncode(List<int> canonicalTable,
String text,
Encoding encoding,
@@ -3189,13 +2941,13 @@ class UriData {
throw new ArgumentError.value(mimeType, "mimeType",
"Invalid MIME type");
}
- buffer.write(_Uri._uriEncode(_tokenCharTable,
- mimeType.substring(0, slashIndex),
- UTF8, false));
+ buffer.write(Uri._uriEncode(_tokenCharTable,
+ mimeType.substring(0, slashIndex),
+ UTF8, false));
buffer.write("/");
- buffer.write(_Uri._uriEncode(_tokenCharTable,
- mimeType.substring(slashIndex + 1),
- UTF8, false));
+ buffer.write(Uri._uriEncode(_tokenCharTable,
+ mimeType.substring(slashIndex + 1),
+ UTF8, false));
}
if (charsetName != null) {
if (indices != null) {
@@ -3203,7 +2955,7 @@ class UriData {
..add(buffer.length + 8);
}
buffer.write(";charset=");
- buffer.write(_Uri._uriEncode(_tokenCharTable, charsetName, UTF8, false));
+ buffer.write(Uri._uriEncode(_tokenCharTable, charsetName, UTF8, false));
}
parameters?.forEach((var key, var value) {
if (key.isEmpty) {
@@ -3216,10 +2968,10 @@ class UriData {
if (indices != null) indices.add(buffer.length);
buffer.write(';');
// Encode any non-RFC2045-token character and both '%' and '#'.
- buffer.write(_Uri._uriEncode(_tokenCharTable, key, UTF8, false));
+ buffer.write(Uri._uriEncode(_tokenCharTable, key, UTF8, false));
if (indices != null) indices.add(buffer.length);
buffer.write('=');
- buffer.write(_Uri._uriEncode(_tokenCharTable, value, UTF8, false));
+ buffer.write(Uri._uriEncode(_tokenCharTable, value, UTF8, false));
});
}
@@ -3236,7 +2988,7 @@ class UriData {
int slashIndex = -1;
for (int i = 0; i < mimeType.length; i++) {
var char = mimeType.codeUnitAt(i);
- if (char != _SLASH) continue;
+ if (char != Uri._SLASH) continue;
if (slashIndex < 0) {
slashIndex = i;
continue;
@@ -3256,7 +3008,7 @@ class UriData {
* ````
*
* where `type`, `subtype`, `attribute` and `value` are specified in RFC-2045,
- * and `data` is a sequence of URI-characters (RFC-2396 `uric`).
+ * and `data` is a sequnce of URI-characters (RFC-2396 `uric`).
*
* This means that all the characters must be ASCII, but the URI may contain
* percent-escapes for non-ASCII byte values that need an interpretation
@@ -3267,22 +3019,13 @@ class UriData {
* and `,` delimiters.
*
* Accessing the individual parts may fail later if they turn out to have
- * content that can't be decoded successfully as a string.
+ * content that can't be decoded sucessfully as a string.
*/
static UriData parse(String uri) {
- if (uri.length >= 5) {
- int dataDelta = _startsWithData(uri, 0);
- if (dataDelta == 0) {
- // Exact match on "data:".
- return _parse(uri, 5, null);
- }
- if (dataDelta == 0x20) {
- // Starts with a non-normalized "data" scheme containing upper-case
- // letters. Parse anyway, but throw away the scheme.
- return _parse(uri.substring(5), 0, null);
- }
+ if (!uri.startsWith("data:")) {
+ throw new FormatException("Does not start with 'data:'", uri, 0);
}
- throw new FormatException("Does not start with 'data:'", uri, 0);
+ return _parse(uri, 5, null);
}
/**
@@ -3307,7 +3050,7 @@ class UriData {
// That's perfectly reasonable - data URIs are not hierarchical,
// but it may make some consumers stumble.
// Should we at least do escape normalization?
- _uriCache = new _Uri._internal("data", "", null, null, path, query, null);
+ _uriCache = new Uri._internal("data", "", null, null, path, query, null);
return _uriCache;
}
@@ -3332,7 +3075,7 @@ class UriData {
int start = _separatorIndices[0] + 1;
int end = _separatorIndices[1];
if (start == end) return "text/plain";
- return _Uri._uriDecode(_text, start, end, UTF8, false);
+ return Uri._uriDecode(_text, start, end, UTF8, false);
}
/**
@@ -3353,8 +3096,8 @@ class UriData {
var keyStart = _separatorIndices[i] + 1;
var keyEnd = _separatorIndices[i + 1];
if (keyEnd == keyStart + 7 && _text.startsWith("charset", keyStart)) {
- return _Uri._uriDecode(_text, keyEnd + 1, _separatorIndices[i + 2],
- UTF8, false);
+ return Uri._uriDecode(_text, keyEnd + 1, _separatorIndices[i + 2],
+ UTF8, false);
}
}
return "US-ASCII";
@@ -3412,8 +3155,8 @@ class UriData {
result[index++] = codeUnit;
} else {
if (i + 2 < text.length) {
- var digit1 = _Uri._parseHexDigit(text.codeUnitAt(i + 1));
- var digit2 = _Uri._parseHexDigit(text.codeUnitAt(i + 2));
+ var digit1 = Uri._parseHexDigit(text.codeUnitAt(i + 1));
+ var digit2 = Uri._parseHexDigit(text.codeUnitAt(i + 2));
if (digit1 >= 0 && digit2 >= 0) {
int byte = digit1 * 16 + digit2;
result[index++] = byte;
@@ -3434,7 +3177,7 @@ class UriData {
* If the content is Base64 encoded, it will be decoded to bytes and then
* decoded to a string using [encoding].
* If encoding is omitted, the value of a `charset` parameter is used
- * if it is recognized by [Encoding.getByName], otherwise it defaults to
+ * if it is recongized by [Encoding.getByName], otherwise it defaults to
* the [ASCII] encoding, which is the default encoding for data URIs
* that do not specify an encoding.
*
@@ -3456,7 +3199,7 @@ class UriData {
var converter = BASE64.decoder.fuse(encoding.decoder);
return converter.convert(text.substring(start));
}
- return _Uri._uriDecode(text, start, text.length, encoding, false);
+ return Uri._uriDecode(text, start, text.length, encoding, false);
}
/**
@@ -3479,8 +3222,8 @@ class UriData {
var start = _separatorIndices[i - 2] + 1;
var equals = _separatorIndices[i - 1];
var end = _separatorIndices[i];
- String key = _Uri._uriDecode(_text, start, equals, UTF8, false);
- String value = _Uri._uriDecode(_text,equals + 1, end, UTF8, false);
+ String key = Uri._uriDecode(_text, start, equals, UTF8, false);
+ String value = Uri._uriDecode(_text,equals + 1, end, UTF8, false);
result[key] = value;
}
return result;
@@ -3563,9 +3306,9 @@ class UriData {
((canonicalTable[byte >> 4] & (1 << (byte & 0x0f))) != 0)) {
buffer.writeCharCode(byte);
} else {
- buffer.writeCharCode(_PERCENT);
- buffer.writeCharCode(_hexDigits.codeUnitAt(byte >> 4));
- buffer.writeCharCode(_hexDigits.codeUnitAt(byte & 0x0f));
+ buffer.writeCharCode(Uri._PERCENT);
+ buffer.writeCharCode(Uri._hexDigits.codeUnitAt(byte >> 4));
+ buffer.writeCharCode(Uri._hexDigits.codeUnitAt(byte & 0x0f));
}
}
if ((byteOr & ~0xFF) != 0) {
@@ -3614,852 +3357,5 @@ class UriData {
// mark = "-" | "_" | "." | "!" | "~" | "*" | "'" | "(" | ")"
//
// This is the same characters as in a URI query (which is URI pchar plus '?')
- static const _uricTable = _Uri._queryCharTable;
-}
-
-// --------------------------------------------------------------------
-// Constants used to read the scanner result.
-// The indices points into the table filled by [_scan] which contains
-// recognized positions in the scanned URI.
-// The `0` index is only used internally.
-
-/// Index of the position of that `:` after a scheme.
-const int _schemeEndIndex = 1;
-/// Index of the position of the character just before the host name.
-const int _hostStartIndex = 2;
-/// Index of the position of the `:` before a port value.
-const int _portStartIndex = 3;
-/// Index of the position of the first character of a path.
-const int _pathStartIndex = 4;
-/// Index of the position of the `?` before a query.
-const int _queryStartIndex = 5;
-/// Index of the position of the `#` before a fragment.
-const int _fragmentStartIndex = 6;
-/// Index of a position where the URI was determined to be "non-simple".
-const int _notSimpleIndex = 7;
-
-// Initial state for scanner.
-const int _uriStart = 00;
-
-// If scanning of a URI terminates in this state or above,
-// consider the URI non-simple
-const int _nonSimpleEndStates = 14;
-
-// Initial state for scheme validation.
-const int _schemeStart = 20;
-
-/// Transition tables used to scan a URI to determine its structure.
-///
-/// The tables represent a state machine with output.
-///
-/// To scan the URI, start in the [_uriStart] state, then read each character
-/// of the URI in order, from start to end, and for each character perform a
-/// transition to a new state while writing the current position into the output
-/// buffer at a designated index.
-///
-/// Each state, represented by an integer which is an index into
-/// [_scannerTables], has a set of transitions, one for each character.
-/// The transitions are encoded as a 5-bit integer representing the next state
-/// and a 3-bit index into the output table.
-///
-/// For URI scanning, only characters in the range U+0020 through U+007E are
-/// interesting, all characters outside that range are treated the same.
-/// The tables only contain 96 entries, representing that characters in the
-/// interesting range, plus one more to represent all values outside the range.
-/// The character entries are stored in one `Uint8List` per state, with the
-/// transition for a character at position `character ^ 0x60`,
-/// which maps the range U+0020 .. U+007F into positions 0 .. 95.
-/// All remaining characters are mapped to position 31 (`0x7f ^ 0x60`) which
-/// represents the transition for all remaining characters.
-final List<Uint8List> _scannerTables = _createTables();
-
-// ----------------------------------------------------------------------
-// Code to create the URI scanner table.
-
-/// Creates the tables for [_scannerTables] used by [Uri.parse].
-///
-/// See [_scannerTables] for the generated format.
-///
-/// The concrete tables are chosen as a trade-off between the number of states
-/// needed and the precision of the result.
-/// This allows definitely recognizing the general structure of the URI
-/// (presence and location of scheme, user-info, host, port, path, query and
-/// fragment) while at the same time detecting that some components are not
-/// in canonical form (anything containing a `%`, a host-name containing a
-/// capital letter). Since the scanner doesn't know whether something is a
-/// scheme or a path until it sees `:`, or user-info or host until it sees
-/// a `@`, a second pass is needed to validate the scheme and any user-info
-/// is considered non-canonical by default.
-///
-/// The states (starting from [_uriStart]) write positions while scanning
-/// a string from `start` to `end` as follows:
-///
-/// - [_schemeEndIndex]: Should be initialized to `start-1`.
-/// If the URI has a scheme, it is set to the position of the `:` after
-/// the scheme.
-/// - [_hostStartIndex]: Should be initialized to `start - 1`.
-/// If the URI has an authority, it is set to the character before the
-/// host name - either the second `/` in the `//` leading the authority,
-/// or the `@` after a user-info. Comparing this value to the scheme end
-/// position can be used to detect that there is a user-info component.
-/// - [_portStartIndex]: Should be initialized to `start`.
-/// Set to the position of the last `:` in an authority, and unchanged
-/// if there is no authority or no `:` in an authority.
-/// If this position is after the host start, there is a port, otherwise it
-/// is just marking a colon in the user-info component.
-/// - [_pathStartIndex]: Should be initialized to `start`.
-/// Is set to the first path character unless the path is empty.
-/// If the path is empty, the position is either unchanged (`start`) or
-/// the first slash of an authority. So, if the path start is before a
-/// host start or scheme end, the path is empty.
-/// - [_queryStartIndex]: Should be initialized to `end`.
-/// The position of the `?` leading a query if the URI contains a query.
-/// - [_fragmentStartIndex]: Should be initialized to `end`.
-/// The position of the `#` leading a fragment if the URI contains a fragment.
-/// - [_notSimpleIndex]: Should be initialized to `start - 1`.
-/// Set to another value if the URI is considered "not simple".
-/// This is elaborated below.
-///
-/// # Simple URIs
-/// A URI is considered "simple" if it is in a normalized form containing no
-/// escapes. This allows us to skip normalization and checking whether escapes
-/// are valid, and to extract components without worrying about unescaping.
-///
-/// The scanner computes a conservative approximation of being "simple".
-/// It rejects any URI with an escape, with a user-info component (mainly
-/// because they are rare and would increase the number of states in the
-/// scanner significantly), with an IPV6 host or with a capital letter in
-/// the scheme or host name (the scheme is handled in a second scan using
-/// a separate two-state table).
-/// Further, paths containing `..` or `.` path segments are considered
-/// non-simple except for pure relative paths (no scheme or authority) starting
-/// with a sequence of "../" segments.
-///
-/// The transition tables cannot detect a trailing ".." in the path,
-/// followed by a query or fragment, because the segment is not known to be
-/// complete until we are past it, and we then need to store the query/fragment
-/// start instead. This cast is checked manually post-scanning (such a path
-/// needs to be normalized to end in "../", so the URI shouldn't be considered
-/// simple).
-List<Uint8List> _createTables() {
- // TODO(lrn): Use a precomputed table.
-
- // Total number of states for the scanner.
- const int stateCount = 22;
-
- // States used to scan a URI from scratch.
- const int schemeOrPath = 01;
- const int authOrPath = 02;
- const int authOrPathSlash = 03;
- const int uinfoOrHost0 = 04;
- const int uinfoOrHost = 05;
- const int uinfoOrPort0 = 06;
- const int uinfoOrPort = 07;
- const int ipv6Host = 08;
- const int relPathSeg = 09;
- const int pathSeg = 10;
- const int path = 11;
- const int query = 12;
- const int fragment = 13;
- const int schemeOrPathDot = 14;
- const int schemeOrPathDot2 = 15;
- const int relPathSegDot = 16;
- const int relPathSegDot2 = 17;
- const int pathSegDot = 18;
- const int pathSegDot2 = 19;
-
- // States used to validate a scheme after its end position has been found.
- const int scheme0 = _schemeStart;
- const int scheme = 21;
-
- // Constants encoding the write-index for the state transition into the top 5
- // bits of a byte.
- const int schemeEnd = _schemeEndIndex << 5;
- const int hostStart = _hostStartIndex << 5;
- const int portStart = _portStartIndex << 5;
- const int pathStart = _pathStartIndex << 5;
- const int queryStart = _queryStartIndex << 5;
- const int fragmentStart = _fragmentStartIndex << 5;
- const int notSimple = _notSimpleIndex << 5;
-
- /// The `unreserved` characters of RFC 3986.
- const unreserved =
- "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz-._~" ;
- /// The `sub-delim` characters of RFC 3986.
- const subDelims = r"!$&'()*+,;=";
- // The `pchar` characters of RFC 3986: characters that may occur in a path,
- // excluding escapes.
- const pchar = "$unreserved$subDelims";
-
- var tables = new List<Uint8List>.generate(stateCount,
- (_) => new Uint8List(96));
-
- // Helper function which initialize the table for [state] with a default
- // transition and returns the table.
- Uint8List build(state, defaultTransition) =>
- tables[state]..fillRange(0, 96, defaultTransition);
-
- // Helper function which sets the transition for each character in [chars]
- // to [transition] in the [target] table.
- // The [chars] string must contain only characters in the U+0020 .. U+007E
- // range.
- void setChars(Uint8List target, String chars, int transition) {
- for (int i = 0; i < chars.length; i++) {
- var char = chars.codeUnitAt(i);
- target[char ^ 0x60] = transition;
- }
- }
-
- /// Helper function which sets the transition for all characters in the
- /// range from `range[0]` to `range[1]` to [transition] in the [target] table.
- ///
- /// The [range] must be a two-character string where both characters are in
- /// the U+0020 .. U+007E range and the former character must have a lower
- /// code point than the latter.
- void setRange(Uint8List target, String range, int transition) {
- for (int i = range.codeUnitAt(0), n = range.codeUnitAt(1); i <= n; i++) {
- target[i ^ 0x60] = transition;
- }
- }
-
- // Create the transitions for each state.
- var b;
-
- // Validate as path, if it is a scheme, we handle it later.
- b = build(_uriStart, schemeOrPath | notSimple);
- setChars(b, pchar, schemeOrPath);
- setChars(b, ".", schemeOrPathDot);
- setChars(b, ":", authOrPath | schemeEnd); // Handle later.
- setChars(b, "/", authOrPathSlash);
- setChars(b, "?", query | queryStart);
- setChars(b, "#", fragment | fragmentStart);
-
- b = build(schemeOrPathDot, schemeOrPath | notSimple);
- setChars(b, pchar, schemeOrPath);
- setChars(b, ".", schemeOrPathDot2);
- setChars(b, ':', authOrPath | schemeEnd);
- setChars(b, "/", pathSeg | notSimple);
- setChars(b, "?", query | queryStart);
- setChars(b, "#", fragment | fragmentStart);
-
- b = build(schemeOrPathDot2, schemeOrPath | notSimple);
- setChars(b, pchar, schemeOrPath);
- setChars(b, "%", schemeOrPath | notSimple);
- setChars(b, ':', authOrPath | schemeEnd);
- setChars(b, "/", relPathSeg);
- setChars(b, "?", query | queryStart);
- setChars(b, "#", fragment | fragmentStart);
-
- b = build(schemeOrPath, schemeOrPath | notSimple);
- setChars(b, pchar, schemeOrPath);
- setChars(b, ':', authOrPath | schemeEnd);
- setChars(b, "/", pathSeg);
- setChars(b, "?", query | queryStart);
- setChars(b, "#", fragment | fragmentStart);
-
- b = build(authOrPath, path | notSimple);
- setChars(b, pchar, path | pathStart);
- setChars(b, "/", authOrPathSlash | pathStart);
- setChars(b, ".", pathSegDot | pathStart);
- setChars(b, "?", query | queryStart);
- setChars(b, "#", fragment | fragmentStart);
-
- b = build(authOrPathSlash, path | notSimple);
- setChars(b, pchar, path);
- setChars(b, "/", uinfoOrHost0 | hostStart);
- setChars(b, ".", pathSegDot);
- setChars(b, "?", query | queryStart);
- setChars(b, "#", fragment | fragmentStart);
-
- b = build(uinfoOrHost0, uinfoOrHost | notSimple);
- setChars(b, pchar, uinfoOrHost);
- setRange(b, "AZ", uinfoOrHost | notSimple);
- setChars(b, ":", uinfoOrPort0 | portStart);
- setChars(b, "@", uinfoOrHost0 | hostStart);
- setChars(b, "[", ipv6Host | notSimple);
- setChars(b, "/", pathSeg | pathStart);
- setChars(b, "?", query | queryStart);
- setChars(b, "#", fragment | fragmentStart);
-
- b = build(uinfoOrHost, uinfoOrHost | notSimple);
- setChars(b, pchar, uinfoOrHost);
- setRange(b, "AZ", uinfoOrHost | notSimple);
- setChars(b, ":", uinfoOrPort0 | portStart);
- setChars(b, "@", uinfoOrHost0 | hostStart);
- setChars(b, "/", pathSeg | pathStart);
- setChars(b, "?", query | queryStart);
- setChars(b, "#", fragment | fragmentStart);
-
- b = build(uinfoOrPort0, uinfoOrPort | notSimple);
- setRange(b, "19", uinfoOrPort);
- setChars(b, "@", uinfoOrHost0 | hostStart);
- setChars(b, "/", pathSeg | pathStart);
- setChars(b, "?", query | queryStart);
- setChars(b, "#", fragment | fragmentStart);
-
- b = build(uinfoOrPort, uinfoOrPort | notSimple);
- setRange(b, "09", uinfoOrPort);
- setChars(b, "@", uinfoOrHost0 | hostStart);
- setChars(b, "/", pathSeg | pathStart);
- setChars(b, "?", query | queryStart);
- setChars(b, "#", fragment | fragmentStart);
-
- b = build(ipv6Host, ipv6Host);
- setChars(b, "]", uinfoOrHost);
-
- b = build(relPathSeg, path | notSimple);
- setChars(b, pchar, path);
- setChars(b, ".", relPathSegDot);
- setChars(b, "/", pathSeg | notSimple);
- setChars(b, "?", query | queryStart);
- setChars(b, "#", fragment | fragmentStart);
-
- b = build(relPathSegDot, path | notSimple);
- setChars(b, pchar, path);
- setChars(b, ".", relPathSegDot2);
- setChars(b, "/", pathSeg | notSimple);
- setChars(b, "?", query | queryStart);
- setChars(b, "#", fragment | fragmentStart);
-
- b = build(relPathSegDot2, path | notSimple);
- setChars(b, pchar, path);
- setChars(b, "/", relPathSeg);
- setChars(b, "?", query | queryStart); // This should be non-simple.
- setChars(b, "#", fragment | fragmentStart); // This should be non-simple.
-
- b = build(pathSeg, path | notSimple);
- setChars(b, pchar, path);
- setChars(b, ".", pathSegDot);
- setChars(b, "/", pathSeg | notSimple);
- setChars(b, "?", query | queryStart);
- setChars(b, "#", fragment | fragmentStart);
-
- b = build(pathSegDot, path | notSimple);
- setChars(b, pchar, path);
- setChars(b, ".", pathSegDot2);
- setChars(b, "/", pathSeg | notSimple);
- setChars(b, "?", query | queryStart);
- setChars(b, "#", fragment | fragmentStart);
-
- b = build(pathSegDot2, path | notSimple);
- setChars(b, pchar, path);
- setChars(b, "/", pathSeg | notSimple);
- setChars(b, "?", query | queryStart);
- setChars(b, "#", fragment | fragmentStart);
-
- b = build(path, path | notSimple);
- setChars(b, pchar, path);
- setChars(b, "/", pathSeg);
- setChars(b, "?", query | queryStart);
- setChars(b, "#", fragment | fragmentStart);
-
- b = build(query, query | notSimple);
- setChars(b, pchar, query);
- setChars(b, "?", query);
- setChars(b, "#", fragment | fragmentStart);
-
- b = build(fragment, fragment | notSimple);
- setChars(b, pchar, fragment);
- setChars(b, "?", fragment);
-
- // A separate two-state validator for lower-case scheme names.
- // Any non-scheme character or upper-case letter is marked as non-simple.
- b = build(scheme0, scheme | notSimple);
- setRange(b, "az", scheme);
-
- b = build(scheme, scheme | notSimple);
- setRange(b, "az", scheme);
- setRange(b, "09", scheme);
- setChars(b, "+-.", scheme);
-
- return tables;
-}
-
-// --------------------------------------------------------------------
-// Code that uses the URI scanner table.
-
-/// Scan a string using the [_scannerTables] state machine.
-///
-/// Scans [uri] from [start] to [end], startig in state [state] and
-/// writing output into [indices].
-///
-/// Returns the final state.
-int _scan(String uri, int start, int end, int state, List<int> indices) {
- var tables = _scannerTables;
- assert(end <= uri.length);
- for (int i = start; i < end; i++) {
- var table = tables[state];
- // Xor with 0x60 to move range 0x20-0x7f into 0x00-0x5f
- int char = uri.codeUnitAt(i) ^ 0x60;
- // Use 0x1f (nee 0x7f) to represent all unhandled characters.
- if (char > 0x5f) char = 0x1f;
- int transition = table[char];
- state = transition & 0x1f;
- indices[transition >> 5] = i;
- }
- return state;
-}
-
-class _SimpleUri implements Uri {
- final String _uri;
- final int _schemeEnd;
- final int _hostStart;
- final int _portStart;
- final int _pathStart;
- final int _queryStart;
- final int _fragmentStart;
- /// The scheme is often used to distinguish URIs.
- /// To make comparisons more efficient, we cache the value, and
- /// canonicalize a few known types.
- String _schemeCache;
- int _hashCodeCache;
-
- _SimpleUri(
- this._uri,
- this._schemeEnd,
- this._hostStart,
- this._portStart,
- this._pathStart,
- this._queryStart,
- this._fragmentStart,
- this._schemeCache);
-
- bool get hasScheme => _schemeEnd > 0;
- bool get hasAuthority => _hostStart > 0;
- bool get hasUserInfo => _hostStart > _schemeEnd + 4;
- bool get hasPort => _hostStart > 0 && _portStart + 1 < _pathStart;
- bool get hasQuery => _queryStart < _fragmentStart;
- bool get hasFragment => _fragmentStart < _uri.length;
-
- bool get _isFile => _schemeEnd == 4 && _uri.startsWith("file");
- bool get _isHttp => _schemeEnd == 4 && _uri.startsWith("http");
- bool get _isHttps => _schemeEnd == 5 && _uri.startsWith("https");
- bool get _isPackage => _schemeEnd == 7 && _uri.startsWith("package");
- bool _isScheme(String scheme) =>
- _schemeEnd == scheme.length && _uri.startsWith(scheme);
-
- bool get hasAbsolutePath => _uri.startsWith("/", _pathStart);
- bool get hasEmptyPath => _pathStart == _queryStart;
-
- bool get isAbsolute => hasScheme && !hasFragment;
-
- String get scheme {
- if (_schemeEnd <= 0) return "";
- if (_schemeCache != null) return _schemeCache;
- if (_isHttp) {
- _schemeCache = "http";
- } else if (_isHttps) {
- _schemeCache = "https";
- } else if (_isFile) {
- _schemeCache = "file";
- } else if (_isPackage) {
- _schemeCache = "package";
- } else {
- _schemeCache = _uri.substring(0, _schemeEnd);
- }
- return _schemeCache;
- }
- String get authority => _hostStart > 0 ?
- _uri.substring(_schemeEnd + 3, _pathStart) : "";
- String get userInfo => (_hostStart > _schemeEnd + 3) ?
- _uri.substring(_schemeEnd + 3, _hostStart - 1) : "";
- String get host =>
- _hostStart > 0 ? _uri.substring(_hostStart, _portStart) : "";
- int get port {
- if (hasPort) return int.parse(_uri.substring(_portStart + 1, _pathStart));
- if (_isHttp) return 80;
- if (_isHttps) return 443;
- return 0;
- }
- String get path =>_uri.substring(_pathStart, _queryStart);
- String get query => (_queryStart < _fragmentStart) ?
- _uri.substring(_queryStart + 1, _fragmentStart) : "";
- String get fragment => (_fragmentStart < _uri.length) ?
- _uri.substring(_fragmentStart + 1) : "";
-
- String get origin {
- // Check original behavior - W3C spec is wonky!
- bool isHttp = _isHttp;
- if (_schemeEnd < 0 || _hostStart == _portStart) {
- throw new StateError("Cannot use origin without a scheme: $this");
- }
- if (!isHttp && !_isHttps) {
- throw new StateError(
- "Origin is only applicable schemes http and https: $this");
- }
- if (_hostStart == _schemeEnd + 3) {
- return _uri.substring(0, _pathStart);
- }
- // Need to drop anon-empty userInfo.
- return _uri.substring(0, _schemeEnd + 3) +
- _uri.substring(_hostStart, _pathStart);
- }
-
- List<String> get pathSegments {
- int start = _pathStart;
- int end = _queryStart;
- if (_uri.startsWith("/", start)) start++;
- if (start == end) return const <String>[];
- List<String> parts = [];
- for (int i = start; i < end; i++) {
- var char = _uri.codeUnitAt(i);
- if (char == _SLASH) {
- parts.add(_uri.substring(start, i));
- start = i + 1;
- }
- }
- parts.add(_uri.substring(start, end));
- return new List<String>.unmodifiable(parts);
- }
-
- Map<String, String> get queryParameters {
- if (!hasQuery) return const <String, String>{};
- return new UnmodifiableMapView<String, String>(
- Uri.splitQueryString(query));
- }
-
- Map<String, List<String>> get queryParametersAll {
- if (!hasQuery) return const <String, List<String>>{};
- Map queryParameterLists = _Uri._splitQueryStringAll(query);
- for (var key in queryParameterLists.keys) {
- queryParameterLists[key] =
- new List<String>.unmodifiable(queryParameterLists[key]);
- }
- return new Map<String, List<String>>.unmodifiable(queryParameterLists);
- }
-
- bool _isPort(String port) {
- int portDigitStart = _portStart + 1;
- return portDigitStart + port.length == _pathStart &&
- _uri.startsWith(port, portDigitStart);
- }
-
- Uri normalizePath() => this;
-
- Uri removeFragment() {
- if (!hasFragment) return this;
- return new _SimpleUri(
- _uri.substring(0, _fragmentStart),
- _schemeEnd, _hostStart, _portStart,
- _pathStart, _queryStart, _fragmentStart, _schemeCache);
- }
-
- Uri replace({String scheme,
- String userInfo,
- String host,
- int port,
- String path,
- Iterable<String> pathSegments,
- String query,
- Map<String, dynamic/*String|Iterable<String>*/> queryParameters,
- String fragment}) {
- bool schemeChanged = false;
- if (scheme != null) {
- scheme = _Uri._makeScheme(scheme, 0, scheme.length);
- schemeChanged = !_isScheme(scheme);
- } else {
- scheme = this.scheme;
- }
- bool isFile = (scheme == "file");
- if (userInfo != null) {
- userInfo = _Uri._makeUserInfo(userInfo, 0, userInfo.length);
- } else if (_hostStart > 0) {
- userInfo = _uri.substring(_schemeEnd + 3, _hostStart);
- } else {
- userInfo = "";
- }
- if (port != null) {
- port = _Uri._makePort(port, scheme);
- } else {
- port = this.hasPort ? this.port : null;
- if (schemeChanged) {
- // The default port might have changed.
- port = _Uri._makePort(port, scheme);
- }
- }
- if (host != null) {
- host = _Uri._makeHost(host, 0, host.length, false);
- } else if (_hostStart > 0) {
- host = _uri.substring(_hostStart, _portStart);
- } else if (userInfo.isNotEmpty || port != null || isFile) {
- host = "";
- }
-
- bool hasAuthority = host != null;
- if (path != null || pathSegments != null) {
- path = _Uri._makePath(path, 0, _stringOrNullLength(path), pathSegments,
- scheme, hasAuthority);
- } else {
- path = _uri.substring(_pathStart, _queryStart);
- if ((isFile || (hasAuthority && !path.isEmpty)) &&
- !path.startsWith('/')) {
- path = "/" + path;
- }
- }
-
- if (query != null || queryParameters != null) {
- query = _Uri._makeQuery(
- query, 0, _stringOrNullLength(query), queryParameters);
- } else if (_queryStart < _fragmentStart) {
- query = _uri.substring(_queryStart + 1, _fragmentStart);
- }
-
- if (fragment != null) {
- fragment = _Uri._makeFragment(fragment, 0, fragment.length);
- } else if (_fragmentStart < _uri.length) {
- fragment = _uri.substring(_fragmentStart + 1);
- }
-
- return new _Uri._internal(
- scheme, userInfo, host, port, path, query, fragment);
- }
-
- Uri resolve(String reference) {
- return resolveUri(Uri.parse(reference));
- }
-
- Uri resolveUri(Uri reference) {
- if (reference is _SimpleUri) {
- return _simpleMerge(this, reference);
- }
- return _toNonSimple().resolveUri(reference);
- }
-
- // Merge two simple URIs. This should always result in a prefix of
- // one concatentated with a suffix of the other, possibly with a `/` in
- // the middle of two merged paths, which is again simple.
- // In a few cases, there might be a need for extra normalization, when
- // resolving on top of a known scheme.
- Uri _simpleMerge(_SimpleUri base, _SimpleUri ref) {
- if (ref.hasScheme) return ref;
- if (ref.hasAuthority) {
- if (!base.hasScheme) return ref;
- bool isSimple = true;
- if (base._isFile) {
- isSimple = !ref.hasEmptyPath;
- } else if (base._isHttp) {
- isSimple = !ref._isPort("80");
- } else if (base._isHttps) {
- isSimple = !ref._isPort("443");
- }
- if (isSimple) {
- var delta = base._schemeEnd + 1;
- var newUri = base._uri.substring(0, base._schemeEnd + 1) +
- ref._uri.substring(ref._schemeEnd + 1);
- return new _SimpleUri(newUri,
- base._schemeEnd,
- ref._hostStart + delta,
- ref._portStart + delta,
- ref._pathStart + delta,
- ref._queryStart + delta,
- ref._fragmentStart + delta,
- base._schemeCache);
- } else {
- // This will require normalization, so use the _Uri implementation.
- return _toNonSimple().resolveUri(ref);
- }
- }
- if (ref.hasEmptyPath) {
- if (ref.hasQuery) {
- int delta = base._queryStart - ref._queryStart;
- var newUri = base._uri.substring(0, base._queryStart) +
- ref._uri.substring(ref._queryStart);
- return new _SimpleUri(newUri,
- base._schemeEnd,
- base._hostStart,
- base._portStart,
- base._pathStart,
- ref._queryStart + delta,
- ref._fragmentStart + delta,
- base._schemeCache);
- }
- if (ref.hasFragment) {
- int delta = base._fragmentStart - ref._fragmentStart;
- var newUri = base._uri.substring(0, base._fragmentStart) +
- ref._uri.substring(ref._fragmentStart);
- return new _SimpleUri(newUri,
- base._schemeEnd,
- base._hostStart,
- base._portStart,
- base._pathStart,
- base._queryStart,
- ref._fragmentStart + delta,
- base._schemeCache);
- }
- return base.removeFragment();
- }
- if (ref.hasAbsolutePath) {
- var delta = base._pathStart - ref._pathStart;
- var newUri = base._uri.substring(0, base._pathStart) +
- ref._uri.substring(ref._pathStart);
- return new _SimpleUri(newUri,
- base._schemeEnd,
- base._hostStart,
- base._portStart,
- base._pathStart,
- ref._queryStart + delta,
- ref._fragmentStart + delta,
- base._schemeCache);
- }
- if (base.hasEmptyPath && base.hasAuthority) {
- // ref has relative non-empty path.
- // Add a "/" in front, then leading "/../" segments are folded to "/".
- int refStart = ref._pathStart;
- while (ref._uri.startsWith("../", refStart)) {
- refStart += 3;
- }
- var delta = base._pathStart - refStart + 1;
- var newUri = "${base._uri.substring(0, base._pathStart)}/"
- "${ref._uri.substring(refStart)}";
- return new _SimpleUri(newUri,
- base._schemeEnd,
- base._hostStart,
- base._portStart,
- base._pathStart,
- ref._queryStart + delta,
- ref._fragmentStart + delta,
- base._schemeCache);
- }
- // Merge paths.
- if (base._uri.startsWith("../", base._pathStart)) {
- // Complex rare case, go slow.
- return _toNonSimple().resolveUri(ref);
- }
-
- // The RFC 3986 algorithm merges the base path without its final segment
- // (anything after the final "/", or everything if the base path doesn't
- // contain any "/"), and the reference path.
- // Then it removes "." and ".." segments using the remove-dot-segment
- // algorithm.
- // This code combines the two steps. It is simplified by knowing that
- // the base path contains no "." or ".." segments, and the reference
- // path can only contain leading ".." segments.
-
- String baseUri = base._uri;
- String refUri = ref._uri;
- int baseStart = base._pathStart;
- int baseEnd = base._queryStart;
- int refStart = ref._pathStart;
- int refEnd = ref._queryStart;
- int backCount = 1;
-
- int slashCount = 0;
-
- // Count leading ".." segments in reference path.
- while (refStart + 3 <= refEnd && refUri.startsWith("../", refStart)) {
- refStart += 3;
- backCount += 1;
- }
-
- // Extra slash inserted between base and reference path parts if
- // the base path contains any slashes.
- // (We could use a slash from the base path in most cases, but not if
- // we remove the entire base path).
- String insert = "";
- while (baseEnd > baseStart) {
- baseEnd--;
- int char = baseUri.codeUnitAt(baseEnd);
- if (char == _SLASH) {
- insert = "/";
- backCount--;
- if (backCount == 0) break;
- }
- }
- // If the base URI has no scheme or authority (`_pathStart == 0`)
- // and a relative path, and we reached the beginning of the path,
- // we have a special case.
- if (baseEnd == 0 && !base.hasAbsolutePath) {
- // Non-RFC 3986 behavior when resolving a purely relative path on top of
- // another relative path: Don't make the result absolute.
- insert = "";
- }
-
- var delta = baseEnd - refStart + insert.length;
- var newUri = "${base._uri.substring(0, baseEnd)}$insert"
- "${ref._uri.substring(refStart)}";
-
- return new _SimpleUri(newUri,
- base._schemeEnd,
- base._hostStart,
- base._portStart,
- base._pathStart,
- ref._queryStart + delta,
- ref._fragmentStart + delta,
- base._schemeCache);
- }
-
- String toFilePath({bool windows}) {
- if (_schemeEnd >= 0 && !_isFile) {
- throw new UnsupportedError(
- "Cannot extract a file path from a $scheme URI");
- }
- if (_queryStart < _uri.length) {
- if (_queryStart < _fragmentStart) {
- throw new UnsupportedError(
- "Cannot extract a file path from a URI with a query component");
- }
- throw new UnsupportedError(
- "Cannot extract a file path from a URI with a fragment component");
- }
- if (windows == null) windows = _Uri._isWindows;
- return windows ? _Uri._toWindowsFilePath(this) : _toFilePath();
- }
-
- String _toFilePath() {
- if (_hostStart < _portStart) {
- // Has authority and non-empty host.
- throw new UnsupportedError(
- "Cannot extract a non-Windows file path from a file URI "
- "with an authority");
- }
- return this.path;
- }
-
- UriData get data {
- assert(scheme != "data");
- return null;
- }
-
- int get hashCode => _hashCodeCache ??= _uri.hashCode;
-
- bool operator==(Object other) {
- if (identical(this, other)) return true;
- if (other is Uri) return _uri == other.toString();
- return false;
- }
-
- Uri _toNonSimple() {
- return new _Uri._internal(
- this.scheme,
- this.userInfo,
- this.hasAuthority ? this.host: null,
- this.hasPort ? this.port : null,
- this.path,
- this.hasQuery ? this.query : null,
- this.hasFragment ? this.fragment : null
- );
- }
-
- String toString() => _uri;
-}
-
-/// Checks whether [text] starts with "data:" at position [start].
-///
-/// The text must be long enough to allow reading five characters
-/// from the [start] position.
-///
-/// Returns an integer value which is zero if text starts with all-lowercase
-/// "data:" and 0x20 if the text starts with "data:" that isn't all lower-case.
-/// All other values means the text starts with some other character.
-int _startsWithData(String text, int start) {
- // Multiply by 3 to avoid a non-colon character making delta be 0x20.
- int delta = (text.codeUnitAt(start + 4) ^ _COLON) * 3;
- delta |= text.codeUnitAt(start) ^ 0x64 /*d*/;
- delta |= text.codeUnitAt(start + 1) ^ 0x61 /*a*/;
- delta |= text.codeUnitAt(start + 2) ^ 0x74 /*t*/;
- delta |= text.codeUnitAt(start + 3) ^ 0x61 /*a*/;
- return delta;
+ static const _uricTable = Uri._queryCharTable;
}
-
-/// Helper function returning the length of a string, or `0` for `null`.
-int _stringOrNullLength(String s) => (s == null) ? 0 : s.length;
« no previous file with comments | « sdk/lib/_internal/js_runtime/lib/core_patch.dart ('k') | tests/compiler/dart2js/compiler_helper.dart » ('j') | no next file with comments »

Powered by Google App Engine
This is Rietveld 408576698