ISOThaiWriter


1   /* Copyright 2002, 2003 Elliotte Rusty Harold
2      
3      This library is free software; you can redistribute it and/or modify
4      it under the terms of version 2.1 of the GNU Lesser General Public 
5      License as published by the Free Software Foundation.
6      
7      This library is distributed in the hope that it will be useful,
8      but WITHOUT ANY WARRANTY; without even the implied warranty of
9      MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the 
10     GNU Lesser General Public License for more details.
11     
12     You should have received a copy of the GNU Lesser General Public
13     License along with this library; if not, write to the 
14     Free Software Foundation, Inc., 59 Temple Place, Suite 330, 
15     Boston, MA 02111-1307  USA
16     
17     You can contact Elliotte Rusty Harold by sending e-mail to
18     elharo@metalab.unc.edu. Please include the word "XOM" in the
19     subject line. The XOM home page is located at http://www.xom.nu/
20  */
21  
22  package nu.xom;
23  
24  import java.io.Writer  ;
25  
26  /**
27   * <p>
28   *   TIS-620, not quite the same as ISO 8859-11.
29   *   TIS-620 does not have the non-breaking space or the C1 controls.
30   * </p>
31   * 
32   * @author Elliotte Rusty Harold
33   * @version 1.0
34   *
35   */
36  class ISOThaiWriter extends TextWriter {
37  
38      ISOThaiWriter(Writer   out, String   encoding) {
39          super(out, encoding);
40      }
41  
42      /**
43       * @see nu.xom.TextWriter#needsEscaping(char)
44       */
45      boolean needsEscaping(char c) {
46          
47          if (c < 128) return false;
48          // C1 controls do not appear to be assigned in this character set
49          // Also, according to 
50          // http://www.inet.co.th/cyberclub/trin/thairef/tis620-iso10646.html
51          // "Special attention should be paid on codepoint A0. Contrary 
52          // to many people's belief that TIS 620 defines codepoint A0
53          // as no-break space (U+00A0), the standard does not assign any
54          // character to this codepoint. Codepoints A0 as well as DB-DE 
55          // and FC-FF are not part of the standard. Interpretations of 
56          // these unassigned codepoints are implementation specific and 
57          // may vary from implementation to implementation. To ensure full 
58          // data interchangeability among various applications, it is 
59          // suggested that Thai software implementors follows the 
60          // TIS 620 standard strictly."
61          
62          switch (c) {
63              case 0x0E01: return false; // THAI CHARACTER KO KAI
64              case 0x0E02: return false; // THAI CHARACTER KHO KHAI
65              case 0x0E03: return false; // THAI CHARACTER KHO KHUAT
66              case 0x0E04: return false; // THAI CHARACTER KHO KHWAI
67              case 0x0E05: return false; // THAI CHARACTER KHO KHON
68              case 0x0E06: return false; // THAI CHARACTER KHO RAKHANG
69              case 0x0E07: return false; // THAI CHARACTER NGO NGU
70              case 0x0E08: return false; // THAI CHARACTER CHO CHAN
71              case 0x0E09: return false; // THAI CHARACTER CHO CHING
72              case 0x0E0A: return false; // THAI CHARACTER CHO CHANG
73              case 0x0E0B: return false; // THAI CHARACTER SO SO
74              case 0x0E0C: return false; // THAI CHARACTER CHO CHOE
75              case 0x0E0D: return false; // THAI CHARACTER YO YING
76              case 0x0E0E: return false; // THAI CHARACTER DO CHADA
77              case 0x0E0F: return false; // THAI CHARACTER TO PATAK
78              case 0x0E10: return false; // THAI CHARACTER THO THAN
79              case 0x0E11: return false; // THAI CHARACTER THO NANGMONTHO
80              case 0x0E12: return false; // THAI CHARACTER THO PHUTHAO
81              case 0x0E13: return false; // THAI CHARACTER NO NEN
82              case 0x0E14: return false; // THAI CHARACTER DO DEK
83              case 0x0E15: return false; // THAI CHARACTER TO TAO
84              case 0x0E16: return false; // THAI CHARACTER THO THUNG
85              case 0x0E17: return false; // THAI CHARACTER THO THAHAN
86              case 0x0E18: return false; // THAI CHARACTER THO THONG
87              case 0x0E19: return false; // THAI CHARACTER NO NU
88              case 0x0E1A: return false; // THAI CHARACTER BO BAIMAI
89              case 0x0E1B: return false; // THAI CHARACTER PO PLA
90              case 0x0E1C: return false; // THAI CHARACTER PHO PHUNG
91              case 0x0E1D: return false; // THAI CHARACTER FO FA
92              case 0x0E1E: return false; // THAI CHARACTER PHO PHAN
93              case 0x0E1F: return false; // THAI CHARACTER FO FAN
94              case 0x0E20: return false; // THAI CHARACTER PHO SAMPHAO
95              case 0x0E21: return false; // THAI CHARACTER MO MA
96              case 0x0E22: return false; // THAI CHARACTER YO YAK
97              case 0x0E23: return false; // THAI CHARACTER RO RUA
98              case 0x0E24: return false; // THAI CHARACTER RU
99              case 0x0E25: return false; // THAI CHARACTER LO LING
100             case 0x0E26: return false; // THAI CHARACTER LU
101             case 0x0E27: return false; // THAI CHARACTER WO WAEN
102             case 0x0E28: return false; // THAI CHARACTER SO SALA
103             case 0x0E29: return false; // THAI CHARACTER SO RUSI
104             case 0x0E2A: return false; // THAI CHARACTER SO SUA
105             case 0x0E2B: return false; // THAI CHARACTER HO HIP
106             case 0x0E2C: return false; // THAI CHARACTER LO CHULA
107             case 0x0E2D: return false; // THAI CHARACTER O ANG
108             case 0x0E2E: return false; // THAI CHARACTER HO NOKHUK
109             case 0x0E2F: return false; // THAI CHARACTER PAIYANNOI
110             case 0x0E30: return false; // THAI CHARACTER SARA A
111             case 0x0E31: return false; // THAI CHARACTER MAI HAN-AKAT
112             case 0x0E32: return false; // THAI CHARACTER SARA AA
113             case 0x0E33: return false; // THAI CHARACTER SARA AM
114             case 0x0E34: return false; // THAI CHARACTER SARA I
115             case 0x0E35: return false; // THAI CHARACTER SARA II
116             case 0x0E36: return false; // THAI CHARACTER SARA UE
117             case 0x0E37: return false; // THAI CHARACTER SARA UEE
118             case 0x0E38: return false; // THAI CHARACTER SARA U
119             case 0x0E39: return false; // THAI CHARACTER SARA UU
120             case 0x0E3A: return false; // THAI CHARACTER PHINTHU
121         }
122         // optimize by splitting switch into contiguous blocks per
123         // Chapter 7 of Java Performance Tuning, Jack Shirazi
124         switch (c) { 
125             case 0x0E3F: return false; // THAI CURRENCY SYMBOL BAHT
126             case 0x0E40: return false; // THAI CHARACTER SARA E
127             case 0x0E41: return false; // THAI CHARACTER SARA AE
128             case 0x0E42: return false; // THAI CHARACTER SARA O
129             case 0x0E43: return false; // THAI CHARACTER SARA AI MAIMUAN
130             case 0x0E44: return false; // THAI CHARACTER SARA AI MAIMALAI
131             case 0x0E45: return false; // THAI CHARACTER LAKKHANGYAO
132             case 0x0E46: return false; // THAI CHARACTER MAIYAMOK
133             case 0x0E47: return false; // THAI CHARACTER MAITAIKHU
134             case 0x0E48: return false; // THAI CHARACTER MAI EK
135             case 0x0E49: return false; // THAI CHARACTER MAI THO
136             case 0x0E4A: return false; // THAI CHARACTER MAI TRI
137             case 0x0E4B: return false; // THAI CHARACTER MAI CHATTAWA
138             case 0x0E4C: return false; // THAI CHARACTER THANTHAKHAT
139             case 0x0E4D: return false; // THAI CHARACTER NIKHAHIT
140             case 0x0E4E: return false; // THAI CHARACTER YAMAKKAN
141             case 0x0E4F: return false; // THAI CHARACTER FONGMAN
142             case 0x0E50: return false; // THAI DIGIT ZERO
143             case 0x0E51: return false; // THAI DIGIT ONE
144             case 0x0E52: return false; // THAI DIGIT TWO
145             case 0x0E53: return false; // THAI DIGIT THREE
146             case 0x0E54: return false; // THAI DIGIT FOUR
147             case 0x0E55: return false; // THAI DIGIT FIVE
148             case 0x0E56: return false; // THAI DIGIT SIX
149             case 0x0E57: return false; // THAI DIGIT SEVEN
150             case 0x0E58: return false; // THAI DIGIT EIGHT
151             case 0x0E59: return false; // THAI DIGIT NINE
152             case 0x0E5A: return false; // THAI CHARACTER ANGKHANKHU
153             case 0x0E5B: return false; // THAI CHARACTER KHOMUT
154         }
155         
156         return true;
157         
158     }
159 
160 }
A to Z: JavaDoc & Examples Daily Java News & Articles Open Source Projects Open Source Codes Free Computer Books Remove Frame
Popular Tags