2 // System.Web.Compilation.AspTokenizer
5 // Gonzalo Paniagua Javier (gonzalo@ximian.com)
7 // (C) 2002,2003 Ximian, Inc (http://www.ximian.com)
11 // Permission is hereby granted, free of charge, to any person obtaining
12 // a copy of this software and associated documentation files (the
13 // "Software"), to deal in the Software without restriction, including
14 // without limitation the rights to use, copy, modify, merge, publish,
15 // distribute, sublicense, and/or sell copies of the Software, and to
16 // permit persons to whom the Software is furnished to do so, subject to
17 // the following conditions:
19 // The above copyright notice and this permission notice shall be
20 // included in all copies or substantial portions of the Software.
22 // THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
23 // EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
24 // MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
25 // NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE
26 // LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
27 // OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
28 // WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
32 using System.Collections;
35 using System.Security.Cryptography;
37 namespace System.Web.Compilation
41 public const int EOF = 0x0200000;
42 public const int IDENTIFIER = 0x0200001;
43 public const int DIRECTIVE = 0x0200002;
44 public const int ATTVALUE = 0x0200003;
45 public const int TEXT = 0x0200004;
46 public const int DOUBLEDASH = 0x0200005;
47 public const int CLOSING = 0x0200006;
53 const int CHECKSUM_BUF_SIZE = 8192;
57 public readonly string Value;
58 public readonly int Position;
59 public readonly int CurrentToken;
60 public readonly bool InTag;
62 public PutBackItem (string value, int position, int currentToken, bool inTag)
66 CurrentToken = currentToken;
71 static char [] lfcr = new char [] { '\n', '\r' };
74 StringBuilder sb, odds;
80 bool alternatingQuotes;
90 char[] checksum_buf = new char [CHECKSUM_BUF_SIZE];
91 int checksum_buf_pos = -1;
94 get { return checksum; }
98 public AspTokenizer (TextReader reader)
101 sb = new StringBuilder ();
102 odds= new StringBuilder();
104 hasPutBack = inTag = false;
109 get { return verbatim; }
110 set { verbatim = value; }
113 public void put_back ()
115 if (hasPutBack && !inTag)
116 throw new HttpException ("put_back called twice!");
119 if (putBackBuffer == null)
120 putBackBuffer = new Stack ();
123 putBackBuffer.Push (new PutBackItem (val, position, current_token, inTag));
124 position -= val.Length;
127 public int get_token ()
130 PutBackItem pbi = putBackBuffer.Pop () as PutBackItem;
131 hasPutBack = putBackBuffer.Count > 0;
132 position = pbi.Position;
135 sb = new StringBuilder (pbi.Value);
136 current_token = pbi.CurrentToken;
138 return current_token;
144 current_token = NextToken ();
145 return current_token;
148 bool is_identifier_start_character (char c)
150 return (Char.IsLetter (c) || c == '_' );
153 bool is_identifier_part_character (char c)
155 return (Char.IsLetterOrDigit (c) || c == '_' || c == '-');
158 void ungetc (int value)
163 // Only '/' passes through here now.
164 // If we ever let \n here, update 'line'
170 void TransformNextBlock (int count, bool final)
172 byte[] input = Encoding.UTF8.GetBytes (checksum_buf, 0, count);
174 if (checksum == null)
175 checksum = MD5.Create ();
178 checksum.TransformFinalBlock (input, 0, input.Length);
180 checksum.TransformBlock (input, 0, input.Length, input, 0);
183 checksum_buf_pos = -1;
186 void UpdateChecksum (int c)
188 bool final = c == -1;
191 if (checksum_buf_pos + 1 >= CHECKSUM_BUF_SIZE)
192 TransformNextBlock (checksum_buf_pos + 1, false);
193 checksum_buf [++checksum_buf_pos] = (char)c;
195 TransformNextBlock (checksum_buf_pos + 1, true);
211 if (c == '\r' && sr.Peek () == '\n') {
232 int ReadAttValue (int start)
237 if (start == '"' || start == '\'') {
241 sb.Append ((char) start);
246 bool inServerTag = false;
247 alternatingQuotes = true;
249 while ((c = sr.Peek ()) != -1) {
250 if (c == '%' && last == '<') {
252 } else if (inServerTag && c == '>' && last == '%') {
254 } else if (!inServerTag) {
255 if (!quoted && c == '/') {
260 } else if (c == '>') {
264 } else if (!quoted && (c == '>' || Char.IsWhiteSpace ((char) c))) {
266 } else if (quoted && c == quoteChar && last != '\\') {
270 } else if (quoted && c == quoteChar) {
271 alternatingQuotes = false;
274 sb.Append ((char) c);
279 return Token.ATTVALUE;
288 while ((c = read_char ()) != -1){
291 sb.Append ((char) c);
295 if (inTag && expectAttrValue && (c == '"' || c == '\''))
296 return ReadAttValue (c);
300 sb.Append ((char) c);
306 sb.Append ((char) c);
310 if (current_token == '<' && "%/!".IndexOf ((char) c) != -1){
311 sb.Append ((char) c);
315 if (inTag && current_token == '%' && "@#=".IndexOf ((char) c) != -1){
316 if (odds.Length == 0 || odds.ToString ().IndexOfAny (lfcr) < 0) {
317 sb.Append ((char) c);
320 sb.Append ((char) c);
324 if (inTag && c == '-' && sr.Peek () == '-'){
327 return Token.DOUBLEDASH;
331 sb.Append ((char) c);
332 while ((c = sr.Peek ()) != -1 && c != '<')
333 sb.Append ((char) read_char ());
335 return (c != -1 || sb.Length > 0) ? Token.TEXT : Token.EOF;
338 if (inTag && current_token == '=' && !Char.IsWhiteSpace ((char) c))
339 return ReadAttValue (c);
341 if (inTag && is_identifier_start_character ((char) c)){
342 sb.Append ((char) c);
343 while ((c = sr.Peek ()) != -1) {
344 if (!is_identifier_part_character ((char) c) && c != ':')
346 sb.Append ((char) read_char ());
349 if (current_token == '@' && Directive.IsDirective (sb.ToString ()))
350 return Token.DIRECTIVE;
352 return Token.IDENTIFIER;
355 if (!Char.IsWhiteSpace ((char) c)) {
356 sb.Append ((char) c);
359 // keep otherwise discarded characters in case we need.
360 odds.Append((char) c);
366 public string Value {
372 val = sb.ToString ();
379 return odds.ToString();
384 get { return inTag; }
385 set { inTag = value; }
388 // Hack for preventing confusion with VB comments (see bug #63451)
389 public bool ExpectAttrValue {
390 get { return expectAttrValue; }
391 set { expectAttrValue = value; }
394 public bool AlternatingQuotes {
395 get { return alternatingQuotes; }
398 public int BeginLine {
399 get { return begline; }
402 public int BeginColumn {
403 get { return begcol; }
410 public int EndColumn {
414 public int Position {
415 get { return position; }