gecko/netwerk/base/public/nsIURI.idl

251 lines
8.6 KiB
Plaintext
Raw Normal View History

/* -*- Mode: C++; tab-width: 2; indent-tabs-mode: nil; c-basic-offset: 4 -*- */
2012-05-21 04:12:37 -07:00
/* This Source Code Form is subject to the terms of the Mozilla Public
* License, v. 2.0. If a copy of the MPL was not distributed with this
* file, You can obtain one at http://mozilla.org/MPL/2.0/. */
#include "nsISupports.idl"
/**
* URIs are essentially structured names for things -- anything. This interface
* provides accessors to set and query the most basic components of an URI.
* Subclasses, including nsIURL, impose greater structure on the URI.
*
* This interface follows Tim Berners-Lee's URI spec (RFC2396) [1], where the
* basic URI components are defined as such:
* <pre>
* ftp://username:password@hostname:portnumber/pathname#ref
* \ / \ / \ / \ /\ \ /
* - --------------- ------ -------- | -
* | | | | | |
* | | | | | Ref
* | | | Port \ /
* | | Host / --------
* | UserPass / |
* Scheme / Path
* \ /
* --------------------------------
* |
* PrePath
* </pre>
* The definition of the URI components has been extended to allow for
* internationalized domain names [2] and the more generic IRI structure [3].
*
* Note also that the RFC defines #-separated fragment identifiers as being
* "not part of the URI". Despite this, we bundle them as part of the URI, for
* convenience.
*
* [1] http://www.ietf.org/rfc/rfc2396.txt
* [2] http://www.ietf.org/internet-drafts/draft-ietf-idn-idna-06.txt
* [3] http://www.ietf.org/internet-drafts/draft-masinter-url-i18n-08.txt
*/
%{C++
#undef GetPort // XXX Windows!
#undef SetPort // XXX Windows!
%}
/**
* nsIURI - interface for an uniform resource identifier w/ i18n support.
*
* AUTF8String attributes may contain unescaped UTF-8 characters.
* Consumers should be careful to escape the UTF-8 strings as necessary, but
* should always try to "display" the UTF-8 version as provided by this
* interface.
*
* AUTF8String attributes may also contain escaped characters.
*
* Unescaping URI segments is unadvised unless there is intimate
* knowledge of the underlying charset or there is no plan to display (or
* otherwise enforce a charset on) the resulting URI substring.
*
* The correct way to create an nsIURI from a string is via
* nsIIOService.newURI.
*
* NOTE: nsBinaryInputStream::ReadObject contains a hackaround to intercept the
* old (pre-gecko6) nsIURI IID and swap in the current IID instead, in order
* for sessionstore to work after an upgrade. If this IID is revved further,
* we will need to add additional checks there for all intermediate IIDs, until
* nsPrincipal is fixed to serialize its URIs as nsISupports (bug 662693).
*/
[scriptable, uuid(395fe045-7d18-4adb-a3fd-af98c8a1af11)]
interface nsIURI : nsISupports
{
/************************************************************************
* The URI is broken down into the following principal components:
*/
/**
* Returns a string representation of the URI. Setting the spec causes
* the new spec to be parsed per the rules for the scheme the URI
* currently has. In particular, setting the spec to a URI string with a
* different scheme will generally produce incorrect results; no one
* outside of a protocol handler implementation should be doing that. If
* the URI stores information from the nsIIOService.newURI call used to
* create it other than just the parsed string, then behavior of this
* information on setting the spec attribute is undefined.
*
* Some characters may be escaped.
*/
attribute AUTF8String spec;
/**
* The prePath (eg. scheme://user:password@host:port) returns the string
* before the path. This is useful for authentication or managing sessions.
*
* Some characters may be escaped.
*/
readonly attribute AUTF8String prePath;
/**
* The Scheme is the protocol to which this URI refers. The scheme is
* restricted to the US-ASCII charset per RFC2396. Setting this is
* highly discouraged outside of a protocol handler implementation, since
* that will generally lead to incorrect results.
*/
attribute ACString scheme;
/**
* The username:password (or username only if value doesn't contain a ':')
*
* Some characters may be escaped.
*/
attribute AUTF8String userPass;
/**
* The optional username and password, assuming the preHost consists of
* username:password.
*
* Some characters may be escaped.
*/
attribute AUTF8String username;
attribute AUTF8String password;
/**
* The host:port (or simply the host, if port == -1).
*
* Characters are NOT escaped.
*/
attribute AUTF8String hostPort;
/**
* The host is the internet domain name to which this URI refers. It could
* be an IPv4 (or IPv6) address literal. If supported, it could be a
* non-ASCII internationalized domain name.
*
* Characters are NOT escaped.
*/
attribute AUTF8String host;
/**
* A port value of -1 corresponds to the protocol's default port (eg. -1
* implies port 80 for http URIs).
*/
attribute long port;
/**
* The path, typically including at least a leading '/' (but may also be
* empty, depending on the protocol).
*
* Some characters may be escaped.
*/
attribute AUTF8String path;
/************************************************************************
* An URI supports the following methods:
*/
/**
* URI equivalence test (not a strict string comparison).
*
* eg. http://foo.com:80/ == http://foo.com/
*/
boolean equals(in nsIURI other);
/**
* An optimization to do scheme checks without requiring the users of nsIURI
* to GetScheme, thereby saving extra allocating and freeing. Returns true if
* the schemes match (case ignored).
*/
boolean schemeIs(in string scheme);
/**
* Clones the current URI.
*/
nsIURI clone();
/**
* This method resolves a relative string into an absolute URI string,
* using this URI as the base.
*
* NOTE: some implementations may have no concept of a relative URI.
*/
AUTF8String resolve(in AUTF8String relativePath);
/************************************************************************
* Additional attributes:
*/
/**
* The URI spec with an ASCII compatible encoding. Host portion follows
* the IDNA draft spec. Other parts are URL-escaped per the rules of
* RFC2396. The result is strictly ASCII.
*/
readonly attribute ACString asciiSpec;
/**
* The URI host with an ASCII compatible encoding. Follows the IDNA
* draft spec for converting internationalized domain names (UTF-8) to
* ASCII for compatibility with existing internet infrasture.
*/
readonly attribute ACString asciiHost;
/**
* The charset of the document from which this URI originated. An empty
* value implies UTF-8.
*
* If this value is something other than UTF-8 then the URI components
* (e.g., spec, prePath, username, etc.) will all be fully URL-escaped.
* Otherwise, the URI components may contain unescaped multibyte UTF-8
* characters.
*/
readonly attribute ACString originCharset;
/************************************************************************
* Additional attribute & methods added for .ref support:
*/
/**
* Returns the reference portion (the part after the "#") of the URI.
* If there isn't one, an empty string is returned.
*
* Some characters may be escaped.
*/
attribute AUTF8String ref;
/**
* URI equivalence test (not a strict string comparison), ignoring
* the value of the .ref member.
*
* eg. http://foo.com/# == http://foo.com/
* http://foo.com/#aaa == http://foo.com/#bbb
*/
boolean equalsExceptRef(in nsIURI other);
/**
* Clones the current URI, clearing the 'ref' attribute in the clone.
*/
nsIURI cloneIgnoringRef();
/**
* returns a string for the current URI with the ref element cleared.
*/
readonly attribute AUTF8String specIgnoringRef;
/**
* Returns if there is a reference portion (the part after the "#") of the URI.
*/
readonly attribute boolean hasRef;
};