15c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)/*
25c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles) * Copyright (C) 2007, 2008 Apple, Inc. All rights reserved.
35c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles) *
45c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles) * Redistribution and use in source and binary forms, with or without
55c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles) * modification, are permitted provided that the following conditions
65c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles) * are met:
75c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles) * 1. Redistributions of source code must retain the above copyright
85c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles) *    notice, this list of conditions and the following disclaimer.
95c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles) * 2. Redistributions in binary form must reproduce the above copyright
105c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles) *    notice, this list of conditions and the following disclaimer in the
115c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles) *    documentation and/or other materials provided with the distribution.
125c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles) *
135c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles) * THIS SOFTWARE IS PROVIDED BY APPLE COMPUTER, INC. ``AS IS'' AND ANY
145c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles) * EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
155c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles) * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
165c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles) * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL APPLE COMPUTER, INC. OR
175c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles) * CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
185c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles) * EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
195c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles) * PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
205c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles) * PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
215c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles) * OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
225c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles) * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
2302772c6a72f1ee0b226341a4f4439970c29fc861Ben Murdoch * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
245c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles) */
255c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)
265c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)#include "config.h"
2781a5157921f1d2a7ff6aae115bfe3c139b38a5c8Torne (Richard Coles)#include "wtf/text/TextCodecUserDefined.h"
285c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)
295c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)#include <stdio.h>
30591b958dee2cf159d33a0b931e6231072eaf38d5Ben Murdoch#include "wtf/PassOwnPtr.h"
31591b958dee2cf159d33a0b931e6231072eaf38d5Ben Murdoch#include "wtf/text/CString.h"
32591b958dee2cf159d33a0b931e6231072eaf38d5Ben Murdoch#include "wtf/text/StringBuffer.h"
33591b958dee2cf159d33a0b931e6231072eaf38d5Ben Murdoch#include "wtf/text/StringBuilder.h"
34591b958dee2cf159d33a0b931e6231072eaf38d5Ben Murdoch#include "wtf/text/WTFString.h"
355c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)
3681a5157921f1d2a7ff6aae115bfe3c139b38a5c8Torne (Richard Coles)namespace WTF {
375c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)
385c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)void TextCodecUserDefined::registerEncodingNames(EncodingNameRegistrar registrar)
395c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles){
405c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)    registrar("x-user-defined", "x-user-defined");
415c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)}
425c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)
435c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)static PassOwnPtr<TextCodec> newStreamingTextDecoderUserDefined(const TextEncoding&, const void*)
445c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles){
455c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)    return adoptPtr(new TextCodecUserDefined);
465c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)}
475c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)
485c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)void TextCodecUserDefined::registerCodecs(TextCodecRegistrar registrar)
495c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles){
505c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)    registrar("x-user-defined", newStreamingTextDecoderUserDefined, 0);
515c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)}
525c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)
535c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)String TextCodecUserDefined::decode(const char* bytes, size_t length, bool, bool, bool&)
545c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles){
555c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)    StringBuilder result;
565c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)    result.reserveCapacity(length);
575c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)
585c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)    for (size_t i = 0; i < length; ++i) {
595c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)        signed char c = bytes[i];
605c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)        result.append(static_cast<UChar>(c & 0xF7FF));
615c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)    }
625c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)
635c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)    return result.toString();
645c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)}
655c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)
66591b958dee2cf159d33a0b931e6231072eaf38d5Ben Murdochtemplate<typename CharType>
67591b958dee2cf159d33a0b931e6231072eaf38d5Ben Murdochstatic CString encodeComplexUserDefined(const CharType* characters, size_t length, UnencodableHandling handling)
685c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles){
695c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)    Vector<char> result(length);
705c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)    char* bytes = result.data();
715c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)
725c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)    size_t resultLength = 0;
735c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)    for (size_t i = 0; i < length; ) {
745c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)        UChar32 c;
755c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)        U16_NEXT(characters, i, length, c);
765c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)        signed char signedByte = c;
775c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)        if ((signedByte & 0xF7FF) == c)
785c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)            bytes[resultLength++] = signedByte;
795c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)        else {
805c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)            // No way to encode this character with x-user-defined.
815c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)            UnencodableReplacementArray replacement;
825c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)            int replacementLength = TextCodec::getUnencodableReplacement(c, handling, replacement);
835c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)            result.grow(resultLength + replacementLength + length - i);
845c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)            bytes = result.data();
855c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)            memcpy(bytes + resultLength, replacement, replacementLength);
865c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)            resultLength += replacementLength;
875c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)        }
885c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)    }
895c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)
905c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)    return CString(bytes, resultLength);
915c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)}
925c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)
93591b958dee2cf159d33a0b931e6231072eaf38d5Ben Murdochtemplate<typename CharType>
94591b958dee2cf159d33a0b931e6231072eaf38d5Ben MurdochCString TextCodecUserDefined::encodeCommon(const CharType* characters, size_t length, UnencodableHandling handling)
955c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles){
965c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)    char* bytes;
97591b958dee2cf159d33a0b931e6231072eaf38d5Ben Murdoch    CString result = CString::newUninitialized(length, bytes);
985c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)
995c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)    // Convert the string a fast way and simultaneously do an efficient check to see if it's all ASCII.
1005c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)    UChar ored = 0;
1015c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)    for (size_t i = 0; i < length; ++i) {
1025c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)        UChar c = characters[i];
1035c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)        bytes[i] = c;
1045c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)        ored |= c;
1055c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)    }
1065c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)
1075c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)    if (!(ored & 0xFF80))
108591b958dee2cf159d33a0b931e6231072eaf38d5Ben Murdoch        return result;
1095c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)
1105c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)    // If it wasn't all ASCII, call the function that handles more-complex cases.
1115c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)    return encodeComplexUserDefined(characters, length, handling);
1125c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)}
1135c87bf8b86a7c82ef50fb7a89697d8e02e2553beTorne (Richard Coles)
114591b958dee2cf159d33a0b931e6231072eaf38d5Ben MurdochCString TextCodecUserDefined::encode(const UChar* characters, size_t length, UnencodableHandling handling)
115591b958dee2cf159d33a0b931e6231072eaf38d5Ben Murdoch{
116591b958dee2cf159d33a0b931e6231072eaf38d5Ben Murdoch    return encodeCommon(characters, length, handling);
117591b958dee2cf159d33a0b931e6231072eaf38d5Ben Murdoch}
118591b958dee2cf159d33a0b931e6231072eaf38d5Ben Murdoch
119591b958dee2cf159d33a0b931e6231072eaf38d5Ben MurdochCString TextCodecUserDefined::encode(const LChar* characters, size_t length, UnencodableHandling handling)
120591b958dee2cf159d33a0b931e6231072eaf38d5Ben Murdoch{
121591b958dee2cf159d33a0b931e6231072eaf38d5Ben Murdoch    return encodeCommon(characters, length, handling);
122591b958dee2cf159d33a0b931e6231072eaf38d5Ben Murdoch}
123591b958dee2cf159d33a0b931e6231072eaf38d5Ben Murdoch
12481a5157921f1d2a7ff6aae115bfe3c139b38a5c8Torne (Richard Coles)} // namespace WTF
125