Keep the old utf8 text converter implementation for those who want to migrate

Johan Brichau · Johan Brichau · commit 2bd416b2aaef · 2021-01-18T20:36:11.000+01:00
diff --git a/repository/Grease-Pharo90-Core.package/GRPharoDeprecatedUtf8Codec.class/README.md b/repository/Grease-Pharo90-Core.package/GRPharoDeprecatedUtf8Codec.class/README.md
diff --git a/repository/Grease-Pharo90-Core.package/GRPharoDeprecatedUtf8Codec.class/class/basicForEncoding..st b/repository/Grease-Pharo90-Core.package/GRPharoDeprecatedUtf8Codec.class/class/basicForEncoding..st
@@ -0,0 +1,3 @@
+private
+basicForEncoding: aString
+	^ self new
diff --git a/repository/Grease-Pharo90-Core.package/GRPharoDeprecatedUtf8Codec.class/class/codecs.st b/repository/Grease-Pharo90-Core.package/GRPharoDeprecatedUtf8Codec.class/class/codecs.st
@@ -0,0 +1,5 @@
+accessing
+codecs
+	^ GRPlatform current utf8CodecClass == self
+		ifTrue:[ Array with: self new ]
+		ifFalse: [ Array new ]
diff --git a/repository/Grease-Pharo90-Core.package/GRPharoDeprecatedUtf8Codec.class/class/supportsEncoding..st b/repository/Grease-Pharo90-Core.package/GRPharoDeprecatedUtf8Codec.class/class/supportsEncoding..st
@@ -0,0 +1,3 @@
+testing
+supportsEncoding: aString
+	^ GRPlatform current utf8CodecClass == self and: [(#('utf-8' 'UTF-8') includes: aString) or: [ UTF8TextConverter encodingNames includes: aString ] ]
diff --git a/repository/Grease-Pharo90-Core.package/GRPharoDeprecatedUtf8Codec.class/instance/decode..st b/repository/Grease-Pharo90-Core.package/GRPharoDeprecatedUtf8Codec.class/instance/decode..st
@@ -0,0 +1,36 @@
+convenience
+decode: aString
+	"Convert the given string from UTF-8 using the fast path if converting to Latin-1"
+	| outStream byte1 byte2 byte3 byte4 unicode stream |
+	stream := aString readStream.
+	outStream := WriteStream on: (String new: aString size).
+	[ stream atEnd not ] whileTrue: [
+		byte1 := stream next asInteger.
+		unicode := byte1.
+		(byte1 bitAnd: 16rE0) = 192 ifTrue: [ "two bytes"
+			byte2 := stream next asInteger.
+			(byte2 bitAnd: 16rC0) = 16r80 ifFalse: [ self invalidUtf8 ].
+			unicode := ((byte1 bitAnd: 31) bitShift: 6) + (byte2 bitAnd: 63) ].
+		(byte1 bitAnd: 16rF0) = 224 ifTrue: [ "three bytes"
+			byte2 := stream next asInteger.
+			(byte2 bitAnd: 16rC0) = 16r80 ifFalse: [ self invalidUtf8 ].
+			byte3 := stream next asInteger.
+			(byte3 bitAnd: 16rC0) = 16r80 ifFalse: [ self invalidUtf8 ].
+			unicode := ((byte1 bitAnd: 15) bitShift: 12) + ((byte2 bitAnd: 63) bitShift: 6)
+				+ (byte3 bitAnd: 63) ].
+		(byte1 bitAnd: 16rF8) = 240 ifTrue: [ "four bytes"
+			byte2 := stream next asInteger.
+			(byte2 bitAnd: 16rC0) = 16r80 ifFalse: [ self invalidUtf8 ].
+			byte3 := stream next asInteger.
+			(byte3 bitAnd: 16rC0) = 16r80 ifFalse: [ self invalidUtf8 ].
+			byte4 := stream next asInteger.
+			(byte4 bitAnd: 16rC0) = 16r80 ifFalse: [ self invalidUtf8 ].
+			unicode := ((byte1 bitAnd: 16r7) bitShift: 18) +
+							((byte2 bitAnd: 63) bitShift: 12) + 
+							((byte3 bitAnd: 63) bitShift: 6) +
+							(byte4 bitAnd: 63) ].
+		unicode ifNil: [ self invalidUtf8 ].
+		unicode = 16rFEFF "ignore BOM" ifFalse: [
+			outStream nextPut: (Character codePoint: unicode) ].
+		unicode := nil ].
+	^ outStream contents
diff --git a/repository/Grease-Pharo90-Core.package/GRPharoDeprecatedUtf8Codec.class/instance/decoderFor..st b/repository/Grease-Pharo90-Core.package/GRPharoDeprecatedUtf8Codec.class/instance/decoderFor..st
@@ -0,0 +1,5 @@
+convenience
+decoderFor: aStream 
+	^ GRPharoUtf8CodecStream 
+		on: aStream
+		converter: UTF8TextConverter new
diff --git a/repository/Grease-Pharo90-Core.package/GRPharoDeprecatedUtf8Codec.class/instance/encodedStringClass.st b/repository/Grease-Pharo90-Core.package/GRPharoDeprecatedUtf8Codec.class/instance/encodedStringClass.st
@@ -0,0 +1,3 @@
+conversion
+encodedStringClass
+	^ String
diff --git a/repository/Grease-Pharo90-Core.package/GRPharoDeprecatedUtf8Codec.class/instance/encoderFor..st b/repository/Grease-Pharo90-Core.package/GRPharoDeprecatedUtf8Codec.class/instance/encoderFor..st
@@ -0,0 +1,5 @@
+convenience
+encoderFor: aStream 
+	^ GRPharoUtf8CodecStream 
+		on: aStream
+		converter: UTF8TextConverter new
diff --git a/repository/Grease-Pharo90-Core.package/GRPharoDeprecatedUtf8Codec.class/instance/invalidUtf8.st b/repository/Grease-Pharo90-Core.package/GRPharoDeprecatedUtf8Codec.class/instance/invalidUtf8.st
@@ -0,0 +1,3 @@
+convenience
+invalidUtf8
+	^ GRInvalidUtf8Error signal: 'Invalid UTF-8 input'
diff --git a/repository/Grease-Pharo90-Core.package/GRPharoDeprecatedUtf8Codec.class/instance/name.st b/repository/Grease-Pharo90-Core.package/GRPharoDeprecatedUtf8Codec.class/instance/name.st
@@ -0,0 +1,3 @@
+accessing
+name
+	^ 'utf-8'
diff --git a/repository/Grease-Pharo90-Core.package/GRPharoDeprecatedUtf8Codec.class/instance/url.st b/repository/Grease-Pharo90-Core.package/GRPharoDeprecatedUtf8Codec.class/instance/url.st
@@ -0,0 +1,3 @@
+accessing
+url
+	^ self
diff --git a/repository/Grease-Pharo90-Core.package/GRPharoDeprecatedUtf8Codec.class/properties.json b/repository/Grease-Pharo90-Core.package/GRPharoDeprecatedUtf8Codec.class/properties.json
@@ -0,0 +1,11 @@
+{
+	"commentStamp" : "",
+	"super" : "GRCodec",
+	"category" : "Grease-Pharo90-Core",
+	"classinstvars" : [ ],
+	"pools" : [ ],
+	"classvars" : [ ],
+	"instvars" : [ ],
+	"name" : "GRPharoDeprecatedUtf8Codec",
+	"type" : "normal"
+}
diff --git a/repository/Grease-Pharo90-Core.package/GRPharoPlatform.class/instance/initialize.st b/repository/Grease-Pharo90-Core.package/GRPharoPlatform.class/instance/initialize.st
@@ -0,0 +1,4 @@
+initialization
+initialize
+	super initialize.
+	utf8CodecClass := GRPharoUtf8Codec
diff --git a/repository/Grease-Pharo90-Core.package/GRPharoPlatform.class/instance/utf8CodecClass..st b/repository/Grease-Pharo90-Core.package/GRPharoPlatform.class/instance/utf8CodecClass..st
@@ -0,0 +1,4 @@
+private
+utf8CodecClass: aClass
+	"Set either the GRPharoDeprecatedUtf8Codec or GRPharoUtf8Codec classes to switch between TextConverter and Zinc for utf8 encoding"
+	utf8CodecClass := aClass
diff --git a/repository/Grease-Pharo90-Core.package/GRPharoPlatform.class/instance/utf8CodecClass.st b/repository/Grease-Pharo90-Core.package/GRPharoPlatform.class/instance/utf8CodecClass.st
@@ -0,0 +1,3 @@
+private
+utf8CodecClass
+	^ utf8CodecClass
diff --git a/repository/Grease-Pharo90-Core.package/GRPharoPlatform.class/properties.json b/repository/Grease-Pharo90-Core.package/GRPharoPlatform.class/properties.json
@@ -8,7 +8,9 @@
 		"UrlTable",
 		"XmlTable"
 	],
-	"instvars" : [ ],
+	"instvars" : [
+		"utf8CodecClass"
+	],
 	"name" : "GRPharoPlatform",
 	"type" : "normal"
 }
diff --git a/repository/Grease-Pharo90-Core.package/GRPharoUtf8Codec.class/class/codecs.st b/repository/Grease-Pharo90-Core.package/GRPharoUtf8Codec.class/class/codecs.st
@@ -1,3 +1,5 @@
 accessing
 codecs
-	^ Array with: self new
+	^ GRPlatform current utf8CodecClass == self
+		ifTrue:[ Array with: self new ]
+		ifFalse: [ Array new ]
diff --git a/repository/Grease-Pharo90-Core.package/GRPharoUtf8Codec.class/class/supportsEncoding..st b/repository/Grease-Pharo90-Core.package/GRPharoUtf8Codec.class/class/supportsEncoding..st
@@ -1,3 +1,3 @@
-testing
+private
 supportsEncoding: aString
-	^ (#('utf-8' 'UTF-8') includes: aString) or: [ UTF8TextConverter encodingNames includes: aString ]
+	^ GRPlatform current utf8CodecClass == self and: [ (#('utf-8' 'UTF-8' 'utf8') includes: aString) ]
diff --git a/repository/Grease-Tests-Pharo-Core.package/GRPharoCodecTest.class/instance/testLanguageTag.st b/repository/Grease-Tests-Pharo-Core.package/GRPharoCodecTest.class/instance/testLanguageTag.st
@@ -3,6 +3,8 @@ testLanguageTag
 	"this makes sure the encoder doesn't fall on the nose with unicode"
 	"Make Japanese String from unicode. see http://www.unicode.org/charts/PDF/U3040.pdf"
 	| leading hiraA hiraO hiraAO |
+	GRPlatform current utf8CodecClass: GRPharoDeprecatedUtf8Codec.
+	[ 
 	leading := (Smalltalk classNamed: #JapaneseEnvironment) leadingChar.
 	hiraA := (Character 
 		leadingChar: leading
@@ -13,4 +15,5 @@ testLanguageTag
 	hiraAO := hiraA , hiraO.
 	self assertEncodingIgnoresLanguageTat: hiraA.
 	self assertEncodingIgnoresLanguageTat: hiraO.
-	self assertEncodingIgnoresLanguageTat: hiraAO
+	self assertEncodingIgnoresLanguageTat: hiraAO 
+	] ensure: [ GRPlatform current utf8CodecClass: GRPharoUtf8Codec. ]

Original file line number	Diff line number	Diff line change
`@@ -0,0 +1,3 @@`
	`1`	`+private`
	`2`	`+basicForEncoding: aString`
	`3`	`+ ^ self new`
Original file line number	Diff line number	Diff line change
`@@ -0,0 +1,3 @@`
	`1`	`+testing`
	`2`	`+supportsEncoding: aString`
	`3`	`+ ^ GRPlatform current utf8CodecClass == self and: [(#('utf-8' 'UTF-8') includes: aString) or: [ UTF8TextConverter encodingNames includes: aString ] ]`
Original file line number	Diff line number	Diff line change
`@@ -0,0 +1,3 @@`
	`1`	`+conversion`
	`2`	`+encodedStringClass`
	`3`	`+ ^ String`
Original file line number	Diff line number	Diff line change
`@@ -0,0 +1,3 @@`
	`1`	`+convenience`
	`2`	`+invalidUtf8`
	`3`	`+ ^ GRInvalidUtf8Error signal: 'Invalid UTF-8 input'`