15945
|
1 |
"
|
|
2 |
COPYRIGHT (c) 2006 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 |
"{ Package: 'stx:libbasic' }"
|
|
13 |
|
|
14 |
"{ NameSpace: CharacterEncoderImplementations }"
|
|
15 |
|
|
16 |
ISO10646_to_UTF8 subclass:#ISO10646_to_XMLUTF8
|
|
17 |
instanceVariableNames:''
|
|
18 |
classVariableNames:'ReplacementCharacter'
|
|
19 |
poolDictionaries:''
|
|
20 |
category:'Collections-Text-Encodings'
|
|
21 |
!
|
|
22 |
|
|
23 |
!ISO10646_to_XMLUTF8 class methodsFor:'documentation'!
|
|
24 |
|
|
25 |
copyright
|
|
26 |
"
|
|
27 |
COPYRIGHT (c) 2006 by eXept Software AG
|
|
28 |
All Rights Reserved
|
|
29 |
|
|
30 |
This software is furnished under a license and may be used
|
|
31 |
only in accordance with the terms of that license and with the
|
|
32 |
inclusion of the above copyright notice. This software may not
|
|
33 |
be provided or otherwise made available to, or used by, any
|
|
34 |
other person. No title to or ownership of the software is
|
|
35 |
hereby transferred.
|
|
36 |
"
|
|
37 |
!
|
|
38 |
|
|
39 |
documentation
|
|
40 |
"
|
|
41 |
This encoder encodes characters into utf8 characters that may
|
|
42 |
occur in XML document.
|
|
43 |
|
|
44 |
Not all UTF characters are valid in XML, whatever encoding
|
|
45 |
is used. For a reference, see
|
|
46 |
|
|
47 |
http://www.w3.org/TR/2000/REC-xml-20001006#NT-Char
|
|
48 |
|
|
49 |
Invalid characters are replaced by ReplacementCharacter
|
|
50 |
with $? as default.
|
|
51 |
|
|
52 |
[author:]
|
|
53 |
Jan Vrany <jan.vrany@fit.cvut.cz>
|
|
54 |
|
|
55 |
[instance variables:]
|
|
56 |
|
|
57 |
[class variables:]
|
|
58 |
|
|
59 |
[see also:]
|
|
60 |
http://www.w3.org/TR/2000/REC-xml-20001006#NT-Char
|
|
61 |
|
|
62 |
"
|
|
63 |
! !
|
|
64 |
|
|
65 |
!ISO10646_to_XMLUTF8 class methodsFor:'initialization'!
|
|
66 |
|
|
67 |
initialize
|
|
68 |
"Invoked at system start or when the class is dynamically loaded."
|
|
69 |
|
|
70 |
ReplacementCharacter := $?.
|
|
71 |
|
|
72 |
"Modified: / 30-06-2012 / 19:55:00 / Jan Vrany <jan.vrany@fit.cvut.cz>"
|
|
73 |
! !
|
|
74 |
|
|
75 |
!ISO10646_to_XMLUTF8 methodsFor:'encoding & decoding'!
|
|
76 |
|
|
77 |
encodeString:aUnicodeString
|
|
78 |
"return the UTF-8 representation of a aUnicodeString.
|
|
79 |
The resulting string contains only valid XML unicode
|
|
80 |
characters. Invalid characters are replaced by a
|
|
81 |
ReplacementCharacter. For details, please see
|
|
82 |
|
|
83 |
http://www.w3.org/TR/2000/REC-xml-20001006#NT-Char
|
|
84 |
|
|
85 |
"
|
|
86 |
|
|
87 |
|s|
|
|
88 |
|
|
89 |
"Copy-paste of superclass's method and tweaked. Not ideal, but
|
|
90 |
but avoids 1 string copy"
|
|
91 |
|
|
92 |
s := WriteStream on:(String uninitializedNew:aUnicodeString size).
|
|
93 |
aUnicodeString do:[:eachCharacter |
|
|
94 |
|codePoint b1 b2 b3 b4 b5 v "{Class: SmallInteger }"|
|
|
95 |
|
|
96 |
codePoint := eachCharacter codePoint.
|
|
97 |
(self isValidXMLunicode: codePoint) ifFalse:[
|
|
98 |
codePoint := ReplacementCharacter codePoint.
|
|
99 |
].
|
|
100 |
|
|
101 |
codePoint <= 16r7F ifTrue:[
|
|
102 |
s nextPut:(Character value:codePoint).
|
|
103 |
] ifFalse:[
|
|
104 |
b1 := Character value:((codePoint bitAnd:16r3F) bitOr:2r10000000).
|
|
105 |
v := codePoint bitShift:-6.
|
|
106 |
v <= 16r1F ifTrue:[
|
|
107 |
s nextPut:(Character value:(v bitOr:2r11000000)).
|
|
108 |
s nextPut:b1.
|
|
109 |
] ifFalse:[
|
|
110 |
b2 := Character value:((v bitAnd:16r3F) bitOr:2r10000000).
|
|
111 |
v := v bitShift:-6.
|
|
112 |
v <= 16r0F ifTrue:[
|
|
113 |
s nextPut:(Character value:(v bitOr:2r11100000)).
|
|
114 |
s nextPut:b2; nextPut:b1.
|
|
115 |
] ifFalse:[
|
|
116 |
b3 := Character value:((v bitAnd:16r3F) bitOr:2r10000000).
|
|
117 |
v := v bitShift:-6.
|
|
118 |
v <= 16r07 ifTrue:[
|
|
119 |
s nextPut:(Character value:(v bitOr:2r11110000)).
|
|
120 |
s nextPut:b3; nextPut:b2; nextPut:b1.
|
|
121 |
] ifFalse:[
|
|
122 |
b4 := Character value:((v bitAnd:16r3F) bitOr:2r10000000).
|
|
123 |
v := v bitShift:-6.
|
|
124 |
v <= 16r03 ifTrue:[
|
|
125 |
s nextPut:(Character value:(v bitOr:2r11111000)).
|
|
126 |
s nextPut:b4; nextPut:b3; nextPut:b2; nextPut:b1.
|
|
127 |
] ifFalse:[
|
|
128 |
b5 := Character value:((v bitAnd:16r3F) bitOr:2r10000000).
|
|
129 |
v := v bitShift:-6.
|
|
130 |
v <= 16r01 ifTrue:[
|
|
131 |
s nextPut:(Character value:(v bitOr:2r11111100)).
|
|
132 |
s nextPut:b5; nextPut:b4; nextPut:b3; nextPut:b2; nextPut:b1.
|
|
133 |
] ifFalse:[
|
|
134 |
"/ cannot happen - we only support up to 30 bit characters
|
|
135 |
self error:'ascii value > 31bit in utf8Encode'.
|
|
136 |
]
|
|
137 |
].
|
|
138 |
].
|
|
139 |
].
|
|
140 |
].
|
|
141 |
].
|
|
142 |
].
|
|
143 |
|
|
144 |
^ s contents
|
|
145 |
|
|
146 |
"
|
|
147 |
(self encodeString:'hello') asByteArray #[104 101 108 108 111]
|
|
148 |
(self encodeString:(Character value:16r40) asString) asByteArray #[64]
|
|
149 |
(self encodeString:(Character value:16r7F) asString) asByteArray #[127]
|
|
150 |
(self encodeString:(Character value:16r80) asString) asByteArray #[194 128]
|
|
151 |
(self encodeString:(Character value:16rFF) asString) asByteArray #[195 191]
|
|
152 |
(self encodeString:(Character value:16r100) asString) asByteArray #[196 128]
|
|
153 |
(self encodeString:(Character value:16r200) asString) asByteArray #[200 128]
|
|
154 |
(self encodeString:(Character value:16r400) asString) asByteArray #[208 128]
|
|
155 |
(self encodeString:(Character value:16r800) asString) asByteArray #[224 160 128]
|
|
156 |
(self encodeString:(Character value:16r1000) asString) asByteArray #[225 128 128]
|
|
157 |
(self encodeString:(Character value:16r2000) asString) asByteArray #[226 128 128]
|
|
158 |
(self encodeString:(Character value:16r4000) asString) asByteArray #[228 128 128]
|
|
159 |
(self encodeString:(Character value:16r8000) asString) asByteArray #[232 128 128]
|
|
160 |
(self encodeString:(Character value:16rFFFF) asString) asByteArray #[239 191 191]
|
|
161 |
"
|
|
162 |
|
|
163 |
"Created: / 30-06-2012 / 20:07:43 / Jan Vrany <jan.vrany@fit.cvut.cz>"
|
|
164 |
! !
|
|
165 |
|
|
166 |
!ISO10646_to_XMLUTF8 methodsFor:'queries'!
|
|
167 |
|
|
168 |
isValidXMLunicode: codePoint
|
|
169 |
"Returns true, if given codePoint (Integer!!!!!!) is
|
|
170 |
valid XML unicode."
|
|
171 |
|
|
172 |
codePoint == 16r0009 ifTrue:[ ^ true ].
|
|
173 |
codePoint == 16r000A ifTrue:[ ^ true ].
|
|
174 |
codePoint == 16r000D ifTrue:[ ^ true ].
|
|
175 |
(codePoint between: 16r0020 and: 16rD7FF ) ifTrue:[ ^ true ].
|
|
176 |
(codePoint between: 16rE000 and: 16rFFFD ) ifTrue:[ ^ true ].
|
|
177 |
(codePoint between: 16r10000 and: 16r10FFFF) ifTrue:[ ^ true ].
|
|
178 |
|
|
179 |
^false.
|
|
180 |
|
|
181 |
"Created: / 30-06-2012 / 20:11:16 / Jan Vrany <jan.vrany@fit.cvut.cz>"
|
|
182 |
! !
|
|
183 |
|
|
184 |
!ISO10646_to_XMLUTF8 class methodsFor:'documentation'!
|
|
185 |
|
|
186 |
version
|
|
187 |
^ '$Header: /cvs/stx/stx/libbasic/CharacterEncoderImplementations__ISO10646_to_XMLUTF8.st,v 1.1 2014-02-05 17:11:46 cg Exp $'
|
|
188 |
!
|
|
189 |
|
|
190 |
version_CVS
|
|
191 |
^ '$Header: /cvs/stx/stx/libbasic/CharacterEncoderImplementations__ISO10646_to_XMLUTF8.st,v 1.1 2014-02-05 17:11:46 cg Exp $'
|
|
192 |
! !
|
|
193 |
|
|
194 |
|
|
195 |
ISO10646_to_XMLUTF8 initialize!
|