在.net framework中streamreader的使用encoding必須在構造器中指定,而且中途完全不可以更改。
在一般的情況下,這不會造成什么問題。一般若是從硬盤讀取文件,單一文件內的編碼一般都是統一的。即便是發現讀錯,亦可以關閉streamreader,重啟使用新的編碼讀取。
偏偏偶最近遇到了需要修改編碼的需求,而且,我的程序沒有關閉重讀的機會。因為偶使用的streamreader的basestream是一個network stream,我不可以關閉它……但是network stream傳過來的東西很可能包涵不同的編碼……gb2312,big5,utf8,iso-8859-1等等……雖然是先得到編碼信息,然后再讀具體內容,但是,一開始使用的stream reader編碼一旦錯了,讀出來的東西便再也無法恢復……會丟字之類的……
我也不可以在獲得編碼信息之后,重新建立一個新的stream reader,因為具體內容已經被原來的stream reader給緩沖掉了……
唯一的解決方法,便是自己實現一個可以改變currentencoding屬性的stream reader了……
全部從頭寫起非常不實際,偶是先當了mono的源碼,從mono的stream reader實現代碼做修改。
stream reader其實很簡單,它內部有兩個buffer,一個是input buffer,一個是decoded buffer,前者用于緩存從base stream讀過來的原始數據,后者用于緩存根據原始數據解碼出來后的東西……只要看明白mono的實現中readbuffer這個方法,要動態修改currentencoding也就不是太難了……
我需要處理的網絡協議是一個行協議……偶在程序中只調用了streamreader的readline方法,而完全沒有使用read的兩個方法,這也使得偶動態修改編碼容易了許多……
偶的做法是每次調用readline的時候,不僅移動decoded buffer的游標(pos),同時也移動input buffer一個新的游標(pos_input),做法很簡單,readline方法需要調用findnexteol移動游標查找換行符號……我在findnexteol方法添加多一行:
  int findnexteol ()
  {
   findnextinputeol();
   ....
而findnextinputeol這個新的函數,完全是findnexteol的翻版,只是前者處理input buffer,而后者處理decoded buffer……
如此一來,我便可以知道每次readline之后,input buffer中還沒有被上層讀到的原始數據有哪些了……
然后,再把currentencoding屬性添加set的方法:
set
   {
    encoding=value;
    decoder = encoding.getdecoder();
    decoded_count = pos + decoder.getchars (input_buffer, pos_input, cbencoded , pos_input, decoded_buffer, pos);
   }
設定新編碼時,程序便根據input buffer的游標(pos_input)把沒有被讀到的原始數據重新decode一次,并且替換掉decoded buffer中的內容。
然后,事情就搞定了……甚至不需要對readline方法做任何修改……除了把cbencoded這個變量放到全局里面外……
但是,偶這個修改使得read的兩個方法變得完全不可以用……一旦調用了……便會使得input buffer與decoded buffer里面兩個游標不同步……下面附上完整的代碼,還望有大俠可以幫忙把read的兩個方法也給搞定了…… 先謝過……
/
// system.io.streamreader.cs
//
// author:
//   dietmar maurer ([email protected])
//   miguel de icaza ([email protected]) 
//
// (c) ximian, inc.  http://www.ximian.com
// copyright (c) 2004 novell (http://www.novell.com)
//
//
// copyright (c) 2004 novell, inc (http://www.novell.com)
//
// permission is hereby granted, free of charge, to any person obtaining
// a copy of this software and associated documentation files (the
// "software"), to deal in the software without restriction, including
// without limitation the rights to use, copy, modify, merge, publish,
// distribute, sublicense, and/or sell copies of the software, and to
// permit persons to whom the software is furnished to do so, subject to
// the following conditions:
// 
// the above copyright notice and this permission notice shall be
// included in all copies or substantial portions of the software.
// 
// the software is provided "as is", without warranty of any kind,
// express or implied, including but not limited to the warranties of
// merchantability, fitness for a particular purpose and
// noninfringement. in no event shall the authors or copyright holders be
// liable for any claim, damages or other liability, whether in an action
// of contract, tort or otherwise, arising from, out of or in connection
// with the software or the use or other dealings in the software.
//
using system;
using system.text;
using system.runtime.interopservices;
namespace system.io 
{
    [serializable]
    public class dynamicstreamreader : textreader 
    {
        const int defaultbuffersize = 1024;
        const int defaultfilebuffersize = 4096;
        const int minimumbuffersize = 128;
        //
        // the input buffer
        //
        byte [] input_buffer;
        //
        // the decoded buffer from the above input buffer
        //
        char [] decoded_buffer;
        //
        // decoded bytes in decoded_buffer.
        //
        int decoded_count;
        //
        // current position in the decoded_buffer
        //
        int pos;
        //
        // current position in the input_buffer
        //
        int pos_input;
        //
        // the buffer size that we are using
        //
        int buffer_size;
        int do_checks;
        
        encoding encoding;
        decoder decoder;
        stream base_stream;
        bool mayblock;
        stringbuilder line_builder;
        private class nullstreamreader : dynamicstreamreader 
        {
            public override int peek ()
            {
                return -1;
            }
            public override int read ()
            {
                return -1;
            }
            public override int read ([in, out] char[] buffer, int index, int count)
            {
                return 0;
            }
            public override string readline ()
            {
                return null;
            }
            public override string readtoend ()
            {
                return string.empty;
            }
            public override stream basestream
            {
                get { return stream.null; }
            }
            public override encoding currentencoding
            {
                get { return encoding.unicode; }
            }
        }
        public new static readonly dynamicstreamreader null =  (dynamicstreamreader)(new nullstreamreader());
        
        internal dynamicstreamreader() {}
        public dynamicstreamreader(stream stream)
            : this (stream, encoding.utf8, true, defaultbuffersize) { }
        public dynamicstreamreader(stream stream, bool detect_encoding_from_bytemarks)
            : this (stream, encoding.utf8, detect_encoding_from_bytemarks, defaultbuffersize) { }
        public dynamicstreamreader(stream stream, encoding encoding)
            : this (stream, encoding, true, defaultbuffersize) { }
        public dynamicstreamreader(stream stream, encoding encoding, bool detect_encoding_from_bytemarks)
            : this (stream, encoding, detect_encoding_from_bytemarks, defaultbuffersize) { }
        
        public dynamicstreamreader(stream stream, encoding encoding, bool detect_encoding_from_bytemarks, int buffer_size)
        {
            initialize (stream, encoding, detect_encoding_from_bytemarks, buffer_size);
        }
        public dynamicstreamreader(string path)
            : this (path, encoding.utf8, true, defaultfilebuffersize) { }
        public dynamicstreamreader(string path, bool detect_encoding_from_bytemarks)
            : this (path, encoding.utf8, detect_encoding_from_bytemarks, defaultfilebuffersize) { }
        public dynamicstreamreader(string path, encoding encoding)
            : this (path, encoding, true, defaultfilebuffersize) { }
        public dynamicstreamreader(string path, encoding encoding, bool detect_encoding_from_bytemarks)
            : this (path, encoding, detect_encoding_from_bytemarks, defaultfilebuffersize) { }
        
        public dynamicstreamreader(string path, encoding encoding, bool detect_encoding_from_bytemarks, int buffer_size)
        {
            if (null == path)
                throw new argumentnullexception("path");
            if (string.empty == path)
                throw new argumentexception("empty path not allowed");
            if (path.indexofany (path.invalidpathchars) != -1)
                throw new argumentexception("path contains invalid characters");
            if (null == encoding)
                throw new argumentnullexception ("encoding");
            if (buffer_size <= 0)
                throw new argumentoutofrangeexception ("buffer_size", "the minimum size of the buffer must be positive");
            string dirname = path.getdirectoryname(path);
            if (dirname != string.empty && !directory.exists(dirname))
                throw new directorynotfoundexception ("directory '" + dirname + "' not found.");
            if (!file.exists(path))
                throw new filenotfoundexception("file not found.", path);
            stream stream = (stream) file.openread (path);
            initialize (stream, encoding, detect_encoding_from_bytemarks, buffer_size);
        }
        internal void initialize (stream stream, encoding encoding, bool detect_encoding_from_bytemarks, int buffer_size)
        {
            if (null == stream)
                throw new argumentnullexception ("stream");
            if (null == encoding)
                throw new argumentnullexception ("encoding");
            if (!stream.canread)
                throw new argumentexception ("cannot read stream");
            if (buffer_size <= 0)
                throw new argumentoutofrangeexception ("buffer_size", "the minimum size of the buffer must be positive");
            if (buffer_size < minimumbuffersize)
                buffer_size = minimumbuffersize;
            base_stream = stream;
            input_buffer = new byte [buffer_size];
            this.buffer_size = buffer_size;
            this.encoding = encoding;
            decoder = encoding.getdecoder ();
            byte [] preamble = encoding.getpreamble ();
            do_checks = detect_encoding_from_bytemarks ? 1 : 0;
            do_checks += (preamble.length == 0) ? 0 : 2;
            
            decoded_buffer = new char [encoding.getmaxcharcount (buffer_size)];
            decoded_count = 0;
            pos = 0;
            pos_input =0;
        }
        public virtual stream basestream
        {
            get 
            {
                return base_stream;
            }
        }
        public virtual encoding currentencoding
        {
            get 
            {
                if (encoding == null)
                    throw new exception ();
                return encoding;
            }
            set
            {
                encoding=value;
                decoder = encoding.getdecoder();
                decoded_count = pos + decoder.getchars (input_buffer, pos_input, cbencoded - pos_input, decoded_buffer, pos);
                //discardbuffereddata();
            }
        }
        public override void close ()
        {
            dispose (true);
        }
        protected override void dispose (bool disposing)
        {
            if (disposing && base_stream != null)
                base_stream.close ();
            
            input_buffer = null;
            decoded_buffer = null;
            encoding = null;
            decoder = null;
            base_stream = null;
            base.dispose (disposing);
        }
        //
        // provides auto-detection of the encoding, as well as skipping over
        // byte marks at the beginning of a stream.
        //
        int dochecks (int count)
        {
            if ((do_checks & 2) == 2)
            {
                byte [] preamble = encoding.getpreamble ();
                int c = preamble.length;
                if (count >= c)
                {
                    int i;
                    
                    for (i = 0; i < c; i++)
                        if (input_buffer [i] != preamble [i])
                            break;
                    if (i == c)
                        return i;
                }
            }
            if ((do_checks & 1) == 1)
            {
                if (count < 2)
                    return 0;
                if (input_buffer [0] == 0xfe && input_buffer [1] == 0xff)
                {
                    this.encoding = encoding.bigendianunicode;
                    return 2;
                }
                if (input_buffer [0] == 0xff && input_buffer [1] == 0xfe)
                {
                    this.encoding = encoding.unicode;
                    return 2;
                }
                if (count < 3)
                    return 0;
                if (input_buffer [0] == 0xef && input_buffer [1] == 0xbb && input_buffer [2] == 0xbf)
                {
                    this.encoding = encoding.utf8;
                    return 3;
                }
            }
            return 0;
        }
        public void discardbuffereddata ()
        {
            pos = decoded_count = 0;
            mayblock = false;
            // discard internal state of the decoder too.
            decoder = encoding.getdecoder ();
        }
        
        int cbencoded; 
        int parse_start;
        // the buffer is empty, fill it again
        private int readbuffer ()
        {
            pos = 0;
            pos_input = 0;
            cbencoded = 0;
            // keep looping until the decoder gives us some chars
            decoded_count = 0;
            parse_start = 0;
            do    
            {
                cbencoded = base_stream.read (input_buffer, 0, buffer_size);
                
                if (cbencoded == 0)
                    return 0;
                mayblock = (cbencoded < buffer_size);
                if (do_checks > 0)
                {
                    encoding old = encoding;
                    parse_start = dochecks (cbencoded);
                    if (old != encoding)
                    {
                        decoder = encoding.getdecoder ();
                    }
                    do_checks = 0;
                    cbencoded -= parse_start;
                }
                decoded_count += decoder.getchars (input_buffer, parse_start, cbencoded, decoded_buffer, 0);
                parse_start = 0;
            } while (decoded_count == 0);
            return decoded_count;
        }
        public override int peek ()
        {
            if (base_stream == null)
                throw new objectdisposedexception ("streamreader", "cannot read from a closed streamreader");
            if (pos >= decoded_count && (mayblock || readbuffer () == 0))
                return -1;
            return decoded_buffer [pos];
        }
        public override int read ()
        {
            throw new exception("dynamic reader could not read!");
        }
        public override int read ([in, out] char[] dest_buffer, int index, int count)
        {
            throw new exception("dynamic reader could not read!");
        }
        bool foundcr_input;
        int findnextinputeol()
        {
            char c = '/0';
            for (; pos_input < cbencoded; pos_input++) 
            {
                c = (char)input_buffer [pos_input];
                if (c == '/n') 
                {
                    pos_input++;
                    int res = (foundcr_input) ? (pos_input - 2) : (pos_input - 1);
                    if (res < 0)
                        res = 0; // if a new buffer starts with a /n and there was a /r at
                    // the end of the previous one, we get here.
                    foundcr_input = false;
                    return res;
                } 
                else if (foundcr_input) 
                {
                    foundcr_input = false;
                    return pos - 1;
                }
                foundcr_input = (c == '/r');
            }
            return -1;
        }
        bool foundcr;
        int findnexteol ()
        {
            findnextinputeol();
            char c = '/0';
            for (; pos < decoded_count; pos++) 
            {
                c = decoded_buffer [pos];
                if (c == '/n') 
                {
                    pos++;
                    int res = (foundcr) ? (pos - 2) : (pos - 1);
                    if (res < 0)
                        res = 0; // if a new buffer starts with a /n and there was a /r at
                    // the end of the previous one, we get here.
                    foundcr = false;
                    return res;
                } 
                else if (foundcr) 
                {
                    foundcr = false;
                    return pos - 1;
                }
                foundcr = (c == '/r');
            }
            return -1;
        }
        public override string readline()
        {
            if (base_stream == null)
                throw new objectdisposedexception ("streamreader", "cannot read from a closed streamreader");
            if (pos >= decoded_count && readbuffer () == 0)
                return null;
            int begin = pos;
            int end = findnexteol ();
            if (end < decoded_count && end >= begin)
                return new string (decoded_buffer, begin, end - begin);
            if (line_builder == null)
                line_builder = new stringbuilder ();
            else
                line_builder.length = 0;
            while (true) 
            {
                if (foundcr) // don't include the trailing cr if present
                    decoded_count--;
                line_builder.append (new string (decoded_buffer, begin, decoded_count - begin));
                if (readbuffer () == 0) 
                {
                    if (line_builder.capacity > 32768) 
                    {
                        stringbuilder sb = line_builder;
                        line_builder = null;
                        return sb.tostring (0, sb.length);
                    }
                    return line_builder.tostring (0, line_builder.length);
                }
                begin = pos;
                end = findnexteol ();
                if (end < decoded_count && end >= begin) 
                {
                    line_builder.append (new string (decoded_buffer, begin, end - begin));
                    if (line_builder.capacity > 32768) 
                    {
                        stringbuilder sb = line_builder;
                        line_builder = null;
                        return sb.tostring (0, sb.length);
                    }
                    return line_builder.tostring (0, line_builder.length);
                }
            }
        }
        public override string readtoend()
        {
            if (base_stream == null)
                throw new objectdisposedexception ("streamreader", "cannot read from a closed streamreader");
stringbuilder text = new stringbuilder ();
            int size = decoded_buffer.length;
            char [] buffer = new char [size];
            int len;
            
            while ((len = read (buffer, 0, size)) > 0)
                text.append (buffer, 0, len);
            return text.tostring ();
        }
    }
}
新聞熱點
疑難解答
圖片精選