webnsurlextras.mm

linux下开源浏览器WebKit的源码,市面上的很多商用浏览器都是移植自WebKit

        }
        else {
            *p++ = c;
            outLength++;
        }
    }
 
    NSData *data = [NSData dataWithBytesNoCopy:outBytes length:outLength]; // adopts outBytes
    return [self _web_URLWithData:data relativeToURL:URL];
}

+ (NSURL *)_web_URLWithUserTypedString:(NSString *)string
{
    return [self _web_URLWithUserTypedString:string relativeToURL:nil];
}

+ (NSURL *)_web_URLWithDataAsString:(NSString *)string
{
    if (string == nil) {
        return nil;
    }
    return [self _web_URLWithDataAsString:string relativeToURL:nil];
}

+ (NSURL *)_web_URLWithDataAsString:(NSString *)string relativeToURL:(NSURL *)baseURL
{
    if (string == nil) {
        return nil;
    }
    string = [string _webkit_stringByTrimmingWhitespace];
    NSData *data = [string dataUsingEncoding:NSISOLatin1StringEncoding];
    return [self _web_URLWithData:data relativeToURL:baseURL];
}

+ (NSURL *)_web_URLWithData:(NSData *)data
{
    return [NSURL _web_URLWithData:data relativeToURL:nil];
}      

+ (NSURL *)_web_URLWithData:(NSData *)data relativeToURL:(NSURL *)baseURL
{
    if (data == nil)
        return nil;
    
    NSURL *result = nil;
    size_t length = [data length];
    if (length > 0) {
        // work around <rdar://4470771>: CFURLCreateAbsoluteURLWithBytes(.., TRUE) doesn't remove non-path components.
        baseURL = [baseURL _webkit_URLByRemovingResourceSpecifier];
        
        const UInt8 *bytes = static_cast<const UInt8*>([data bytes]);
        // NOTE: We use UTF-8 here since this encoding is used when computing strings when returning URL components
        // (e.g calls to NSURL -path). However, this function is not tolerant of illegal UTF-8 sequences, which
        // could either be a malformed string or bytes in a different encoding, like shift-jis, so we fall back
        // onto using ISO Latin 1 in those cases.
        result = WebCFAutorelease(CFURLCreateAbsoluteURLWithBytes(NULL, bytes, length, kCFStringEncodingUTF8, (CFURLRef)baseURL, YES));
        if (!result)
            result = WebCFAutorelease(CFURLCreateAbsoluteURLWithBytes(NULL, bytes, length, kCFStringEncodingISOLatin1, (CFURLRef)baseURL, YES));
    } else
        result = [NSURL URLWithString:@""];
    
    return result;
}

- (NSData *)_web_originalData
{
    UInt8 *buffer = (UInt8 *)malloc(URL_BYTES_BUFFER_LENGTH);
    CFIndex bytesFilled = CFURLGetBytes((CFURLRef)self, buffer, URL_BYTES_BUFFER_LENGTH);
    if (bytesFilled == -1) {
        CFIndex bytesToAllocate = CFURLGetBytes((CFURLRef)self, NULL, 0);
        buffer = (UInt8 *)realloc(buffer, bytesToAllocate);
        bytesFilled = CFURLGetBytes((CFURLRef)self, buffer, bytesToAllocate);
        ASSERT(bytesFilled == bytesToAllocate);
    }
    
    // buffer is adopted by the NSData
    NSData *data = [NSData dataWithBytesNoCopy:buffer length:bytesFilled freeWhenDone:YES];
    
    NSURL *baseURL = (NSURL *)CFURLGetBaseURL((CFURLRef)self);
    if (baseURL)
        return [[NSURL _web_URLWithData:data relativeToURL:baseURL] _web_originalData];
    return data;
}

- (NSString *)_web_originalDataAsString
{
    return [[[NSString alloc] initWithData:[self _web_originalData] encoding:NSISOLatin1StringEncoding] autorelease];
}

static CFStringRef createStringWithEscapedUnsafeCharacters(CFStringRef string)
{
    CFIndex length = CFStringGetLength(string);
    Vector<UChar, 2048> sourceBuffer(length);
    CFStringGetCharacters(string, CFRangeMake(0, length), sourceBuffer.data());

    Vector<UChar, 2048> outBuffer;

    CFIndex i = 0;
    while (i < length) {
        UChar32 c;
        U16_NEXT(sourceBuffer, i, length, c)

        if (isLookalikeCharacter(c)) {
            uint8_t utf8Buffer[4];
            CFIndex offset = 0;
            UBool failure = false;
            U8_APPEND(utf8Buffer, offset, 4, c, failure)
            ASSERT(!failure);

            for (CFIndex j = 0; j < offset; ++j) {
                outBuffer.append('%');
                outBuffer.append(hexDigit(utf8Buffer[j] >> 4));
                outBuffer.append(hexDigit(utf8Buffer[j] & 0xf));
            }
        } else {
            UChar utf16Buffer[2];
            CFIndex offset = 0;
            UBool failure = false;
            U16_APPEND(utf16Buffer, offset, 2, c, failure)
            ASSERT(!failure);
            for (CFIndex j = 0; j < offset; ++j)
                outBuffer.append(utf16Buffer[j]);
        }
    }

    return CFStringCreateWithCharacters(NULL, outBuffer.data(), outBuffer.size());
}

- (NSString *)_web_userVisibleString
{
    NSData *data = [self _web_originalData];
    const unsigned char *before = static_cast<const unsigned char*>([data bytes]);
    int length = [data length];

    bool needsHostNameDecoding = false;

    const unsigned char *p = before;
    int bufferLength = (length * 3) + 1;
    char *after = static_cast<char *>(malloc(bufferLength)); // large enough to %-escape every character
    char *q = after;
    int i;
    for (i = 0; i < length; i++) {
        unsigned char c = p[i];
        // unescape escape sequences that indicate bytes greater than 0x7f
        if (c == '%' && (i + 1 < length && isHexDigit(p[i + 1])) && i + 2 < length && isHexDigit(p[i + 2])) {
            unsigned char u = (hexDigitValue(p[i + 1]) << 4) | hexDigitValue(p[i + 2]);
            if (u > 0x7f) {
                // unescape
                *q++ = u;
            } else {
                // do not unescape
                *q++ = p[i];
                *q++ = p[i + 1];
                *q++ = p[i + 2];
            }
            i += 2;
        } else {
            *q++ = c;

            // Check for "xn--" in an efficient, non-case-sensitive, way.
            if (c == '-' && i >= 3 && !needsHostNameDecoding && (q[-4] | 0x20) == 'x' && (q[-3] | 0x20) == 'n' && q[-2] == '-')
                needsHostNameDecoding = true;
        }
    }
    *q = '\0';
    
    // Check string to see if it can be converted to display using UTF-8  
    NSString *result = [NSString stringWithUTF8String:after];
    if (!result) {
        // Could not convert to UTF-8.
        // Convert characters greater than 0x7f to escape sequences.
        // Shift current string to the end of the buffer
        // then we will copy back bytes to the start of the buffer 
        // as we convert.
        int afterlength = q - after;
        char *p = after + bufferLength - afterlength - 1;
        memmove(p, after, afterlength + 1); // copies trailing '\0'
        char *q = after;
        while (*p) {
            unsigned char c = *p;
            if (c > 0x7f) {
                *q++ = '%';
                *q++ = hexDigit(c >> 4);
                *q++ = hexDigit(c & 0xf);
            } else {
                *q++ = *p;
            }
            p++;
        }
        *q = '\0';
        result = [NSString stringWithUTF8String:after];
    }

    free(after);

    result = mapHostNames(result, !needsHostNameDecoding);
    result = [result precomposedStringWithCanonicalMapping];
    return WebCFAutorelease(createStringWithEscapedUnsafeCharacters((CFStringRef)result));
}

- (BOOL)_web_isEmpty
{
    if (!CFURLGetBaseURL((CFURLRef)self))
        return CFURLGetBytes((CFURLRef)self, NULL, 0) == 0;
    return [[self _web_originalData] length] == 0;
}

- (const char *)_web_URLCString
{
    NSMutableData *data = [NSMutableData data];
    [data appendData:[self _web_originalData]];
    [data appendBytes:"\0" length:1];
    return (const char *)[data bytes];
 }

- (NSURL *)_webkit_canonicalize
{    
    NSURLRequest *request = [[NSURLRequest alloc] initWithURL:self];
    Class concreteClass = WKNSURLProtocolClassForRequest(request);
    if (!concreteClass) {
        [request release];
        return self;
    }
    
    // This applies NSURL's concept of canonicalization, but not KURL's concept. It would
    // make sense to apply both, but when we tried that it caused a performance degradation
    // (see 5315926). It might make sense to apply only the KURL concept and not the NSURL
    // concept, but it's too risky to make that change for WebKit 3.0.
    NSURLRequest *newRequest = [concreteClass canonicalRequestForRequest:request];
    NSURL *newURL = [newRequest URL]; 
    NSURL *result = [[newURL retain] autorelease]; 
    [request release];
    
    return result;
}

typedef struct {
    NSString *scheme;
    NSString *user;
    NSString *password;
    NSString *host;
    CFIndex port; // kCFNotFound means ignore/omit
    NSString *path;
    NSString *query;
    NSString *fragment;
} WebKitURLComponents;

- (NSURL *)_webkit_URLByRemovingComponent:(CFURLComponentType)component
{
    CFRange fragRg = CFURLGetByteRangeForComponent((CFURLRef)self, component, NULL);
    // Check to see if a fragment exists before decomposing the URL.
    if (fragRg.location == kCFNotFound)
        return self;
 
    UInt8 *urlBytes, buffer[2048];
    CFIndex numBytes = CFURLGetBytes((CFURLRef)self, buffer, 2048);
    if (numBytes == -1) {
        numBytes = CFURLGetBytes((CFURLRef)self, NULL, 0);
        urlBytes = static_cast<UInt8*>(malloc(numBytes));
        CFURLGetBytes((CFURLRef)self, urlBytes, numBytes);
    } else
        urlBytes = buffer;
     
    NSURL *result = (NSURL *)CFMakeCollectable(CFURLCreateWithBytes(NULL, urlBytes, fragRg.location - 1, kCFStringEncodingUTF8, NULL));
    if (!result)
        result = (NSURL *)CFMakeCollectable(CFURLCreateWithBytes(NULL, urlBytes, fragRg.location - 1, kCFStringEncodingISOLatin1, NULL));
     
    if (urlBytes != buffer) free(urlBytes);
    return result ? [result autorelease] : self;
}

- (NSURL *)_webkit_URLByRemovingFragment
{
    return [self _webkit_URLByRemovingComponent:kCFURLComponentFragment];
}

- (NSURL *)_webkit_URLByRemovingResourceSpecifier
{
    return [self _webkit_URLByRemovingComponent:kCFURLComponentResourceSpecifier];
}

- (BOOL)_webkit_isJavaScriptURL
{
    return [[self _web_originalDataAsString] _webkit_isJavaScriptURL];
}

- (NSString *)_webkit_scriptIfJavaScriptURL
{
    return [[self absoluteString] _webkit_scriptIfJavaScriptURL];
}

- (BOOL)_webkit_isFileURL
{    
    return [[self _web_originalDataAsString] _webkit_isFileURL];
}

- (BOOL)_webkit_isFTPDirectoryURL
{
    return [[self _web_originalDataAsString] _webkit_isFTPDirectoryURL];
}

- (BOOL)_webkit_shouldLoadAsEmptyDocument
{
    return [[self _web_originalDataAsString] _webkit_hasCaseInsensitivePrefix:@"about:"] || [self _web_isEmpty];
}

- (NSURL *)_web_URLWithLowercasedScheme
{
    CFRange range;
    CFURLGetByteRangeForComponent((CFURLRef)self, kCFURLComponentScheme, &range);
    if (range.location == kCFNotFound) {
        return self;
    }
    
    UInt8 static_buffer[URL_BYTES_BUFFER_LENGTH];
    UInt8 *buffer = static_buffer;
    CFIndex bytesFilled = CFURLGetBytes((CFURLRef)self, buffer, URL_BYTES_BUFFER_LENGTH);
    if (bytesFilled == -1) {
        CFIndex bytesToAllocate = CFURLGetBytes((CFURLRef)self, NULL, 0);
        buffer = static_cast<UInt8 *>(malloc(bytesToAllocate));
        bytesFilled = CFURLGetBytes((CFURLRef)self, buffer, bytesToAllocate);
        ASSERT(bytesFilled == bytesToAllocate);
    }
    
    int i;
    BOOL changed = NO;
    for (i = 0; i < range.length; ++i) {
        char c = buffer[range.location + i];
        char lower = toASCIILower(c);
        if (c != lower) {
            buffer[range.location + i] = lower;
            changed = YES;
        }
    }
    
    NSURL *result = changed
        ? (NSURL *)WebCFAutorelease(CFURLCreateAbsoluteURLWithBytes(NULL, buffer, bytesFilled, kCFStringEncodingUTF8, nil, YES))
        : (NSURL *)self;

    if (buffer != static_buffer) {
        free(buffer);
    }
    
    return result;
}


-(BOOL)_web_hasQuestionMarkOnlyQueryString
{
    CFRange rangeWithSeparators;
    CFURLGetByteRangeForComponent((CFURLRef)self, kCFURLComponentQuery, &rangeWithSeparators);
    if (rangeWithSeparators.location != kCFNotFound && rangeWithSeparators.length == 1) {
        return YES;
    }
    return NO;
}

-(NSData *)_web_schemeSeparatorWithoutColon
{
    NSData *result = nil;
    CFRange rangeWithSeparators;
    CFRange range = CFURLGetByteRangeForComponent((CFURLRef)self, kCFURLComponentScheme, &rangeWithSeparators);
    if (rangeWithSeparators.location != kCFNotFound) {
        NSString *absoluteString = [self absoluteString];
        NSRange separatorsRange = NSMakeRange(range.location + range.length + 1, rangeWithSeparators.length - range.length - 1);
        if (separatorsRange.location + separatorsRange.length <= [absoluteString length]) {
            NSString *slashes = [absoluteString substringWithRange:separatorsRange];
            result = [slashes dataUsingEncoding:NSISOLatin1StringEncoding];
        }
    }
    return result;
}

#define completeURL (CFURLComponentType)-1

-(NSData *)_web_dataForURLComponentType:(CFURLComponentType)componentType
{