|
1 " |
|
2 COPYRIGHT (c) 2005 by eXept Software AG |
|
3 All Rights Reserved |
|
4 |
|
5 This software is furnished under a license and may be used |
|
6 only in accordance with the terms of that license and with the |
|
7 inclusion of the above copyright notice. This software may not |
|
8 be provided or otherwise made available to, or used by, any |
|
9 other person. No title to or ownership of the software is |
|
10 hereby transferred. |
|
11 " |
|
12 |
|
13 "{ Package: 'stx:libbasic' }" |
|
14 |
|
15 "{ NameSpace: CharacterEncoderImplementations }" |
|
16 |
|
17 TwoByteEncoder subclass:#ISO10646_to_UTF16BE |
|
18 instanceVariableNames:'' |
|
19 classVariableNames:'' |
|
20 poolDictionaries:'' |
|
21 category:'Collections-Text-Encodings' |
|
22 ! |
|
23 |
|
24 !ISO10646_to_UTF16BE class methodsFor:'documentation'! |
|
25 |
|
26 copyright |
|
27 " |
|
28 COPYRIGHT (c) 2005 by eXept Software AG |
|
29 All Rights Reserved |
|
30 |
|
31 This software is furnished under a license and may be used |
|
32 only in accordance with the terms of that license and with the |
|
33 inclusion of the above copyright notice. This software may not |
|
34 be provided or otherwise made available to, or used by, any |
|
35 other person. No title to or ownership of the software is |
|
36 hereby transferred. |
|
37 " |
|
38 ! |
|
39 |
|
40 documentation |
|
41 " |
|
42 encodes/decodes UTF16 BigEndian (big-end-first) |
|
43 " |
|
44 ! |
|
45 |
|
46 examples |
|
47 " |
|
48 Encoding (unicode to utf16BE) |
|
49 ISO10646_to_UTF16BE encodeString:'hello'. |
|
50 |
|
51 |
|
52 Decoding (utf16BE to unicode): |
|
53 |t| |
|
54 |
|
55 t := ISO10646_to_UTF16BE encodeString:''. |
|
56 ISO10646_to_UTF16BE decodeString:t. |
|
57 " |
|
58 ! ! |
|
59 |
|
60 !ISO10646_to_UTF16BE methodsFor:'encoding & decoding'! |
|
61 |
|
62 decode:aCode |
|
63 self shouldNotImplement "/ no single byte conversion possible |
|
64 ! |
|
65 |
|
66 decodeString:aStringOrByteCollection |
|
67 "given a byteArray (2-bytes per character) or unsignedShortArra in UTF16 encoding, |
|
68 return a new string containing the same characters, in 8, 16bit (or more) encoding. |
|
69 Returns either a normal String, a TwoByte- or a FourByte-String instance. |
|
70 Only useful, when reading from external sources. |
|
71 This only handles up-to 30bit characters." |
|
72 |
|
73 |sz nBitsRequired s newString idx bitsPerElementIn nextIn |
|
74 codeIn codeIn1 codeIn2| |
|
75 |
|
76 aStringOrByteCollection isByteArray ifTrue:[ |
|
77 bitsPerElementIn := 8. |
|
78 ] ifFalse:[ |
|
79 aStringOrByteCollection isString ifTrue:[ |
|
80 bitsPerElementIn := aStringOrByteCollection bitsPerCharacter. |
|
81 ] ifFalse:[ |
|
82 bitsPerElementIn := 16. |
|
83 ]. |
|
84 ]. |
|
85 |
|
86 bitsPerElementIn == 8 ifTrue:[ |
|
87 nextIn := [self nextTwoByteValueFrom:s]. |
|
88 ] ifFalse:[ |
|
89 nextIn := [s next]. |
|
90 ]. |
|
91 |
|
92 nBitsRequired := 8. |
|
93 sz := 0. |
|
94 s := aStringOrByteCollection readStream. |
|
95 [s atEnd] whileFalse:[ |
|
96 codeIn := nextIn value. |
|
97 sz := sz + 1. |
|
98 |
|
99 codeIn <= 16rFF ifTrue:[ |
|
100 ] ifFalse:[ |
|
101 nBitsRequired := nBitsRequired max:16. |
|
102 (codeIn between:16rD800 and:16rDBFF) ifTrue:[ |
|
103 nBitsRequired := 32. |
|
104 codeIn2 := nextIn value. |
|
105 ]. |
|
106 ] |
|
107 ]. |
|
108 |
|
109 nBitsRequired == 8 ifTrue:[ |
|
110 newString := String uninitializedNew:sz |
|
111 ] ifFalse:[ |
|
112 nBitsRequired <= 16 ifTrue:[ |
|
113 newString := Unicode16String new:sz |
|
114 ] ifFalse:[ |
|
115 newString := Unicode32String new:sz |
|
116 ] |
|
117 ]. |
|
118 |
|
119 s := aStringOrByteCollection readStream. |
|
120 idx := 1. |
|
121 [s atEnd] whileFalse:[ |
|
122 codeIn := nextIn value. |
|
123 codeIn <= 16rFF ifTrue:[ |
|
124 ] ifFalse:[ |
|
125 nBitsRequired := nBitsRequired max:16. |
|
126 (codeIn between:16rD800 and:16rDBFF) ifTrue:[ |
|
127 nBitsRequired := 32. |
|
128 codeIn1 := codeIn. |
|
129 codeIn2 := nextIn value. |
|
130 codeIn := ((codeIn1 - 16rD800) bitShift:10) |
|
131 + |
|
132 (codeIn2 - 16rDC00) |
|
133 + 16r00010000. |
|
134 ]. |
|
135 ]. |
|
136 newString at:idx put:(Character value:codeIn). |
|
137 idx := idx + 1. |
|
138 ]. |
|
139 ^ newString |
|
140 |
|
141 " |
|
142 self new decodeString:#[ 16r00 16r42 ] |
|
143 self new decodeString:#[ 16r01 16r42 ] |
|
144 self new decodeString:#[ 16r00 16r48 |
|
145 16r00 16r69 |
|
146 16rD8 16r00 |
|
147 16rDC 16r00 |
|
148 16r00 16r21 |
|
149 16r00 16r21 |
|
150 ] |
|
151 |
|
152 self new decodeString:#( 16r0048 |
|
153 16r0069 |
|
154 16rD800 |
|
155 16rDC00 |
|
156 16r0021 |
|
157 16r0021 |
|
158 ) |
|
159 " |
|
160 ! |
|
161 |
|
162 encode:aCode |
|
163 self shouldNotImplement "/ no single byte conversion possible |
|
164 ! |
|
165 |
|
166 encodeString:aUnicodeString |
|
167 "return the UTF-16 representation of a aUnicodeString. |
|
168 The resulting string is only useful to be stored on some external file, |
|
169 not for being used inside ST/X." |
|
170 |
|
171 |s| |
|
172 |
|
173 s := WriteStream on:(ByteArray uninitializedNew:aUnicodeString size). |
|
174 aUnicodeString do:[:eachCharacter | |
|
175 |codePoint t hi low| |
|
176 |
|
177 codePoint := eachCharacter codePoint. |
|
178 (codePoint <= 16rFFFF) ifTrue:[ |
|
179 ((codePoint <= 16rD7FF) or:[ codePoint between:16rE000 and:16rFFFF]) ifTrue:[ |
|
180 self nextPutTwoByteValue:codePoint to:s. |
|
181 ] ifFalse:[ |
|
182 "/ unrepresentable: D800..DFFFF |
|
183 self error:'unrepresentable value (D800..DFFFF) in utf16Encode'. |
|
184 ]. |
|
185 ] ifFalse:[ |
|
186 t := codePoint - 16r00010000. |
|
187 hi := t bitShift:-10. |
|
188 low := t bitAnd:16r3FF. |
|
189 hi > 16r3FF ifTrue:[ |
|
190 "/ unrepresentable: above 110000 |
|
191 self error:'unrepresentable value (> 10FFFF) in utf16Encode'. |
|
192 ]. |
|
193 self nextPutTwoByteValue:(hi + 16rD800) to:s. |
|
194 self nextPutTwoByteValue:(low + 16rDC00) to:s. |
|
195 ]. |
|
196 ]. |
|
197 |
|
198 ^ s contents |
|
199 |
|
200 " |
|
201 (self encodeString:'hello') #[0 104 0 101 0 108 0 108 0 111] |
|
202 (self encodeString:(Character value:16r40) asString) #[0 64] |
|
203 (self encodeString:(Character value:16rFF) asString) #[0 255] |
|
204 (self encodeString:(Character value:16r100) asString) #[1 0] |
|
205 (self encodeString:(Character value:16r1000) asString) #[16 0] |
|
206 (self encodeString:(Character value:16r2000) asString) #[32 0] |
|
207 (self encodeString:(Character value:16r4000) asString) #[64 0] |
|
208 (self encodeString:(Character value:16r8000) asString) #[128 0] |
|
209 (self encodeString:(Character value:16rD7FF) asString) #[215 255] |
|
210 (self encodeString:(Character value:16rE000) asString) #[224 0] |
|
211 (self encodeString:(Character value:16rFFFF) asString) #[255 255] |
|
212 (self encodeString:(Character value:16r10000) asString) #[216 64 220 0] |
|
213 (self encodeString:(Character value:16r10FFF) asString) #[216 67 223 255] |
|
214 (self encodeString:(Character value:16r1FFFF) asString) #[216 127 223 255] |
|
215 (self encodeString:(Character value:16r10FFFF) asString) #[219 255 223 255] |
|
216 error cases: |
|
217 (self encodeString:(Character value:16rD800) asString) |
|
218 (self encodeString:(Character value:16rD801) asString) |
|
219 (self encodeString:(Character value:16rDFFF) asString) |
|
220 (self encodeString:(Character value:16r110000) asString) |
|
221 " |
|
222 ! ! |
|
223 |
|
224 !ISO10646_to_UTF16BE methodsFor:'private'! |
|
225 |
|
226 nextPutTwoByteValue:anInteger to:aStream |
|
227 aStream nextPutShort:anInteger MSB:true |
|
228 ! |
|
229 |
|
230 nextTwoByteValueFrom:aStream |
|
231 ^ aStream nextUnsignedShortMSB:true |
|
232 ! ! |
|
233 |
|
234 !ISO10646_to_UTF16BE methodsFor:'queries'! |
|
235 |
|
236 nameOfEncoding |
|
237 ^ #'utf8be' |
|
238 ! ! |
|
239 |
|
240 !ISO10646_to_UTF16BE class methodsFor:'documentation'! |
|
241 |
|
242 version |
|
243 ^ '$Header: /cvs/stx/stx/libbasic/CharacterEncoderImplementations__ISO10646_to_UTF16BE.st,v 1.1 2005-07-07 17:36:41 cg Exp $' |
|
244 ! ! |