WebKit/chromium/src/WebEntities.h
author Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
Mon, 04 Oct 2010 01:32:07 +0300
changeset 2 303757a437d3
parent 0 4f2f89ce4247
permissions -rw-r--r--
Revision: 201037 Kit: 201039
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
0
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
     1
/*
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
     2
 * Copyright (C) 2009 Google Inc. All rights reserved.
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
     3
 *
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
     4
 * Redistribution and use in source and binary forms, with or without
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
     5
 * modification, are permitted provided that the following conditions are
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
     6
 * met:
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
     7
 *
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
     8
 *     * Redistributions of source code must retain the above copyright
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
     9
 * notice, this list of conditions and the following disclaimer.
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    10
 *     * Redistributions in binary form must reproduce the above
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    11
 * copyright notice, this list of conditions and the following disclaimer
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    12
 * in the documentation and/or other materials provided with the
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    13
 * distribution.
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    14
 *     * Neither the name of Google Inc. nor the names of its
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    15
 * contributors may be used to endorse or promote products derived from
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    16
 * this software without specific prior written permission.
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    17
 *
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    18
 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    19
 * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    20
 * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    21
 * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    22
 * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    23
 * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    24
 * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    25
 * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    26
 * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    27
 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    28
 * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    29
 */
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    30
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    31
#ifndef WebEntities_h
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    32
#define WebEntities_h
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    33
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    34
#include "PlatformString.h"
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    35
#include <wtf/HashMap.h>
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    36
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    37
namespace WebKit {
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    38
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    39
class WebEntities {
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    40
public:
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    41
    // &apos;, &percnt;, &nsup;, &supl; are not defined by the HTML standards.
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    42
    //  - IE does not support &apos; as an HTML entity (but support it as an XML
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    43
    //    entity.)
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    44
    //  - Firefox supports &apos; as an HTML entity.
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    45
    //  - Both of IE and Firefox don't support &percnt;, &nsup; and &supl;.
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    46
    //
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    47
    // A web page saved by Chromium should be able to be read by other browsers
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    48
    // such as IE and Firefox.  Chromium should produce only the standard entity
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    49
    // references which other browsers can recognize.
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    50
    // So if standard_html_entities_ is true, we will use a numeric character
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    51
    // reference for &apos;, and don't use entity references for &percnt;, &nsup;
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    52
    // and &supl; for serialization.
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    53
    //
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    54
    // If xmlEntities is true, WebEntities will only contain standard XML
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    55
    // entities.
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    56
    explicit WebEntities(bool xmlEntities);
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    57
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    58
    // Check whether specified unicode has corresponding html or xml built-in
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    59
    // entity name. If yes, return the entity notation. If not, returns an
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    60
    // empty string. Parameter isHTML indicates check the code in html entity
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    61
    // map or in xml entity map.
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    62
    WebCore::String entityNameByCode(int code) const;
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    63
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    64
    // Returns a new string with corresponding entity names replaced.
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    65
    WebCore::String convertEntitiesInString(const WebCore::String&) const;
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    66
private:
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    67
    typedef HashMap<int, WebCore::String> EntitiesMapType;
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    68
    // An internal object that maps the Unicode character to corresponding
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    69
    // entity notation.
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    70
    EntitiesMapType m_entitiesMap;
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    71
};
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    72
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    73
} // namespace WebKit
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    74
4f2f89ce4247 Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff changeset
    75
#endif