Squeak-dev January 2018

squeak-dev@lists.squeakfoundation.org

50 participants
171 discussions

The Trunk: Kernel-eem.1136.mcz
by commits＠source.squeak.org 08 Jan '18

08 Jan '18

Eliot Miranda uploaded a new version of Kernel to project The Trunk: http://source.squeak.org/trunk/Kernel-eem.1136.mcz ==================== Summary ==================== Name: Kernel-eem.1136 Author: eem Time: 7 January 2018, 5:12:13.286468 pm UUID: d26ae358-e7c7-4838-8053-f3e868c774be Ancestors: Kernel-eem.1135 Make CompiledCode>>= work for CompiledBlocks and add the #== short-cut (e.g. to speed-up Dictionary lookup). Add pushFullClosure:numCopied: implementations to Context, BlockStartLocator and InstructionClient. Tweak pushClosureCopyNumCopiedValues:numArgs:blockSize: to avoid determining numCopied from the copiedValues. =============== Diff against Kernel-eem.1135 =============== Item was added: + ----- Method: BlockStartLocator>>pushFullClosure:numCopied: (in category 'instruction decoding') ----- + pushFullClosure: aCompiledBlock numCopied: numCopied + "Answer the block method" + ^aCompiledBlock! Item was changed: ----- Method: CompiledCode>>= (in category 'comparing') ----- = method "Answer whether the receiver implements the same code as the argument, method. Here ``same code'' means that if the receiver's source is compiled with the same compiler it should produce the same sequence of bytecodes and literals, same trailer and same properties. Hence this definition of #= (only one of many plausible definitions) can be used to quickly identify changes in the compiler's output." | numLits | + self == method ifTrue: + [^true]. + method isCompiledCode ifFalse: [^false]. - method isCompiledMethod ifFalse: [^false]. self size = method size ifFalse: [^false]. self header = method header ifFalse: [^false]. "N.B. includes numLiterals comparison." self initialPC to: self endPC do: [:i | (self at: i) = (method at: i) ifFalse: [^false]]. numLits := self numLiterals. 1 to: numLits do: [:i| | lit1 lit2 | lit1 := self literalAt: i. lit2 := method literalAt: i. (lit1 == lit2 or: [lit1 literalEqual: lit2]) ifFalse: [(i = 1 and: [#(117 120) includes: self primitive]) ifTrue: [lit1 isArray ifTrue: [(lit2 isArray and: [(lit1 first: 2) = (lit2 first: 2)]) ifFalse: [^false]] ifFalse: "ExternalLibraryFunction" [(lit1 analogousCodeTo: lit2) ifFalse: [^false]]] ifFalse: [i = (numLits - 1) ifTrue: "properties" [(self properties analogousCodeTo: method properties) ifFalse: [^false]] ifFalse: "last literal (methodClassAssociation) of class-side methods is not unique" [(i = numLits and: [lit1 isVariableBinding and: [lit2 isVariableBinding and: [lit1 key == lit2 key and: [lit1 value == lit2 value]]]]) ifFalse: [^false]]]]]. ^true! Item was changed: ----- Method: Context>>pushClosureCopyNumCopiedValues:numArgs:blockSize: (in category 'instruction decoding') ----- pushClosureCopyNumCopiedValues: numCopied numArgs: numArgs blockSize: blockSize "Simulate the action of a 'closure copy' bytecode whose result is the new BlockClosure for the following code" | copiedValues | numCopied > 0 ifTrue: [copiedValues := Array new: numCopied. numCopied to: 1 by: -1 do: [:i| copiedValues at: i put: self pop]] ifFalse: [copiedValues := nil]. + self push: ((BlockClosure new: numCopied) - self push: (BlockClosure outerContext: self startpc: pc numArgs: numArgs copiedValues: copiedValues). self jump: blockSize! Item was added: + ----- Method: Context>>pushFullClosure:numCopied: (in category 'instruction decoding') ----- + pushFullClosure: aCompiledBlock numCopied: numCopied + "Simulate the action of a 'closure copy' bytecode whose result is the + new FullBlockClosure for the supplied compiled block." + | copiedValues | + numCopied > 0 + ifTrue: + [copiedValues := Array new: numCopied. + numCopied to: 1 by: -1 do: + [:i| + copiedValues at: i put: self pop]] + ifFalse: + [copiedValues := nil]. + self push: ((FullBlockClosure new: numCopied) + receiver: receiver + outerContext: self + method: aCompiledBlock + copiedValues: copiedValues)! Item was added: + ----- Method: InstructionClient>>pushFullClosure:numCopied: (in category 'instruction decoding') ----- + pushFullClosure: aCompiledBlock numCopied: numCopied + "Push Full Closure bytecode."!

1 0

The Trunk: Kernel-eem.1135.mcz
by commits＠source.squeak.org 08 Jan '18

08 Jan '18

Eliot Miranda uploaded a new version of Kernel to project The Trunk: http://source.squeak.org/trunk/Kernel-eem.1135.mcz ==================== Summary ==================== Name: Kernel-eem.1135 Author: eem Time: 7 January 2018, 4:52:25.529759 pm UUID: 51a9a0b5-8ab9-4ada-8030-ab0dd9bedf09 Ancestors: Kernel-eem.1134 Speed up some of the scanning methods in CompiledCode by cacheing encoderClass =============== Diff against Kernel-eem.1134 =============== Item was changed: ----- Method: CompiledCode>>abstractBytecodeMessagesAndPCs (in category 'scanning') ----- abstractBytecodeMessagesAndPCs "Answer the receiver's sequence of abstract bytecodes as a sequence of tuples of bytecode message and pc." "(CompiledCode >> #abstractBytecodeMessagesAndPCs) abstractBytecodeMessagesAndPCs" + | msgs initial endpc pc scanner encoderClass | - | msgs initial endpc pc scanner | scanner := InstructionStream new method: self pc: (initial := self initialPC). msgs := OrderedCollection new: (endpc := self endPC) - initial. + encoderClass := self encoderClass. [(pc := scanner pc) <= endpc] whileTrue: + "i.e. nil will not understand any message and so the exception block will collect all of them." + [[encoderClass interpretNextInstructionFor: nil in: scanner] - [[scanner interpretNextInstructionFor: nil] on: MessageNotUnderstood do: [:ex| msgs addLast: { ex message. pc }]]. ^msgs! Item was changed: ----- Method: CompiledCode>>abstractBytecodeMessagesFrom:to:do: (in category 'scanning') ----- abstractBytecodeMessagesFrom: startpc to: endpc do: aBlock "Evaluate aBlock with the sequence of abstract bytecodes from startpc through endpc in the receiver" + | scanner encoderClass | - | scanner | scanner := InstructionStream new method: self pc: startpc. + encoderClass := self encoderClass. [scanner pc <= endpc] whileTrue: + "i.e. nil will not understand any message and so the exception block will collect all of them." + [[encoderClass interpretNextInstructionFor: nil in: scanner] - [[scanner interpretNextInstructionFor: nil] on: MessageNotUnderstood do: [:ex| aBlock value: ex message]] "| m msgs | msgs := OrderedCollection new. + (m := CompiledCode >> #abstractBytecodeMessagesFrom:to:do:) - (m := CompiledCode >> #abstractBytecodeMessagesFrom:to:) abstractBytecodeMessagesFrom: m initialPC to: m endPC do: [:msg| msgs add: msg selector]. msgs"! Item was changed: ----- Method: CompiledMethod>>embeddedBlockClosures (in category 'closures') ----- embeddedBlockClosures + | bms extractor scanner endPC encoderClass | - | bms extractor scanner endPC | bms := OrderedCollection new. scanner := self scanner. + extractor := ClosureExtractor withAction: [:c| bms add: c] andScanner: scanner. - extractor := ClosureExtractor withAction: [ :c | bms add: c ] andScanner: scanner. endPC := self endPC. + encoderClass := self encoderClass. + [scanner pc <= endPC] whileTrue: + [encoderClass interpretNextInstructionFor: extractor in: scanner]. + ^bms! - [ scanner pc <= endPC ] whileTrue: [ scanner interpretNextInstructionFor: extractor ]. - ^ bms!

1 0

The Trunk: EToys-tpr.318.mcz
by commits＠source.squeak.org 08 Jan '18

08 Jan '18

tim Rowledge uploaded a new version of EToys to project The Trunk: http://source.squeak.org/trunk/EToys-tpr.318.mcz ==================== Summary ==================== Name: EToys-tpr.318 Author: tpr Time: 7 January 2018, 4:50:04.77835 pm UUID: e54e7a35-2e82-480c-b307-c6493a5d7e3c Ancestors: EToys-dtl.317 Updates to connect some file loading methods to the new FileDialogs =============== Diff against EToys-dtl.317 =============== Item was changed: ----- Method: MentoringEventRecorder>>readTape (in category 'fileIn/Out') ----- readTape "Put up a prompt for reading an event tape; if one is provided, read it." + | fileName | + fileName := UIManager default chooseFileMatchingSuffixes: #('tape' ) . + fileName ifNotNil: [self readTapeFrom: (FileStream fileNamed: fileName)]! - | aFileStream | - aFileStream := (FileList2 modalFileSelectorForSuffixes: #('tape' )) . - aFileStream ifNotNil: [self readTapeFrom: aFileStream]! Item was changed: ----- Method: PDA>>loadDatabase (in category 'initialization') ----- loadDatabase | aName aFileStream list | + aName := UIManager default chooseFileMatchingSuffixes: #('pda' 'pda.gz' ) label: 'Choose a file to load'. - aName := Utilities chooseFileWithSuffixFromList: #('.pda' '.pda.gz' ) withCaption: 'Choose a file to load'. aName ifNil: [^ self]. + - "User made no choice" - aName == #none - ifTrue: [^ self inform: 'Sorry, no suitable files found - (names should end with .data or .data.gz)']. aFileStream := FileStream oldFileNamed: aName. list := aFileStream fileInObjectAndCode. userCategories := list first. allPeople := list second. allEvents := list third. recurringEvents := list fourth. allToDoItems := list fifth. allNotes := list sixth. date := Date today. self selectCategory: 'all'! Item was changed: ----- Method: PDA>>mergeDatabase (in category 'initialization') ----- mergeDatabase | aName aFileStream list | + aName := UIManager default chooseFileMatchingSuffixes: #('pda' 'pda.gz' ) label: 'Choose a file to load'. - aName := Utilities chooseFileWithSuffixFromList: #('.pda' '.pda.gz' ) withCaption: 'Choose a file to load'. aName ifNil: [^ self]. - "User made no choice" - aName == #none - ifTrue: [^ self inform: 'Sorry, no suitable files found - (names should end with .data or .data.gz)']. aFileStream := FileStream oldFileNamed: aName. list := aFileStream fileInObjectAndCode. userCategories := (list first , userCategories) asSet asArray sort. allPeople := (list second , allPeople) asSet asArray sort. allEvents := (list third , allEvents) asSet asArray sort. recurringEvents := (list fourth , recurringEvents) asSet asArray sort. allToDoItems := (list fifth , allToDoItems) asSet asArray sort. allNotes := ((list sixth) , allNotes) asSet asArray sort. date := Date today. self selectCategory: 'all'! Item was changed: ----- Method: SoundLibraryTool>>loadSoundFromDisk (in category 'menu') ----- loadSoundFromDisk "Put up a file chooser dialog inviting the user to import a sound file; accept it" | aSound aName aFileStream fullName ext reply | + fullName := UIManager default chooseFileMatchingSuffixes: #(#AIFF #aiff #Wave #wav #wave ). + fullName - aFileStream := FileList2 modalFileSelectorForSuffixes: #(#AIFF #aiff #Wave #wav #wave ). - aFileStream ifNil: [^ self]. + - fullName := aFileStream name. ('*.AIFF' match: fullName) ifTrue: [aSound := SampledSound fromAIFFfileNamed: fullName] + ifFalse: [aSound := SampledSound fromWaveStream: (aFileStream := FileStream oldFileNamed: fullName). + aFileStream close]. + - ifFalse: [aSound := SampledSound fromWaveStream: aFileStream]. - aFileStream close. ext := FileDirectory extensionFor: fullName. + aName := FileDirectory localNameFor: fullName. - aName := (FileDirectory on: fullName) pathParts last. ext size > 0 ifTrue: [aName := aName copyFrom: 1 to: (aName size - (ext size + 1))]. [reply := FillInTheBlank request: 'Please give a name for this sound' translated initialAnswer: aName. reply isEmptyOrNil ifTrue: [^ self]. (SampledSound soundLibrary includesKey: reply) ifTrue: [self inform: 'sorry, that name is already taken' translated. false] ifFalse: [true]] whileFalse. SampledSound addLibrarySoundNamed: reply samples: aSound samples samplingRate: aSound originalSamplingRate. self update!

1 0

The Trunk: ST80-tpr.233.mcz
by commits＠source.squeak.org 08 Jan '18

08 Jan '18

tim Rowledge uploaded a new version of ST80 to project The Trunk: http://source.squeak.org/trunk/ST80-tpr.233.mcz ==================== Summary ==================== Name: ST80-tpr.233 Author: tpr Time: 7 January 2018, 4:44:46.985177 pm UUID: c7bef0c0-1026-4f40-a074-add958b89220 Ancestors: ST80-nice.232 Update some MVC related code to use the hooks that might oneday allow them to use a nice graphical file dialog =============== Diff against ST80-nice.232 =============== Item was changed: ----- Method: BitEditor>>fileOut (in category 'menu messages') ----- fileOut | fileName | fileName := UIManager default + saveFilenameRequest: 'Save this Form to' translated - request: 'File name?' translated initialAnswer: 'Filename.form'. + fileName ifNil: [^ self]. - fileName isEmpty ifTrue: [^ self]. Cursor normal showWhile: [model writeOnFileNamed: fileName]. ! Item was changed: ----- Method: FormEditor>>fileInForm (in category 'editing tools') ----- fileInForm "Ask the user for a file name and then recalls the Form in that file as the current source Form (form). Does not change the tool." | fileName | fileName := UIManager default + chooseFileMatchingSuffixes: #('form') + label: 'File name?' translated. + fileName ifNil: [^ self]. - request: 'File name?' translated - initialAnswer: 'Filename.form'. - fileName isEmpty ifTrue: [^ self]. form := Form fromFileNamed: fileName. tool := previousTool. ! Item was changed: ----- Method: FormEditor>>fileOut (in category 'menu messages') ----- fileOut | fileName | fileName := UIManager default + saveFilenameRequest: 'File name?' translated - request: 'File name?' translated initialAnswer: 'Filename.form'. + fileName ifNil: [^ self]. - fileName isEmpty ifTrue: [^ self]. Cursor normal showWhile: [model writeOnFileNamed: fileName]. ! Item was changed: ----- Method: FormEditor>>fileOutForm (in category 'editing tools') ----- fileOutForm "Ask the user for a file name and save the current source form under that name. Does not change the tool." | fileName | fileName := UIManager default + saveFilenameRequest: 'File name?' translated - request: 'File name?' translated initialAnswer: 'Filename.form'. + fileName ifNil: [^ self]. - fileName isEmpty ifTrue: [^ self]. Cursor normal showWhile: [form writeOnFileNamed: fileName]. tool := previousTool. ! Item was changed: ----- Method: ParagraphEditor>>saveContentsInFile (in category 'menu messages') ----- saveContentsInFile "Save the receiver's contents string to a file, prompting the user for a file-name. Suggest a reasonable file-name." | fileName stringToSave parentWindow labelToUse suggestedName | stringToSave := paragraph text string. stringToSave size = 0 ifTrue: [^ self inform: 'nothing to save.']. parentWindow := self model dependents detect: [:dep | dep isKindOf: SystemWindow orOf: StandardSystemView] ifNone: [nil]. labelToUse := parentWindow ifNil: ['Untitled'] ifNotNil: [parentWindow label]. suggestedName := nil. #(('Decompressed contents of: ' '.gz')) do: "can add more here..." [:leaderTrailer | | lastIndex | (labelToUse beginsWith: leaderTrailer first) ifTrue: [suggestedName := labelToUse copyFrom: leaderTrailer first size + 1 to: labelToUse size. (labelToUse endsWith: leaderTrailer last) ifTrue: [suggestedName := suggestedName copyFrom: 1 to: suggestedName size - leaderTrailer last size] ifFalse: [lastIndex := suggestedName lastIndexOf: $.. (lastIndex > 1) ifTrue: [suggestedName := suggestedName copyFrom: 1 to: lastIndex - 1]]]]. suggestedName ifNil: [suggestedName := labelToUse, '.text']. + fileName := UIManager default saveFilenameRequest: 'File name?' translated - fileName := UIManager default request: 'File name?' translated initialAnswer: suggestedName. fileName isEmptyOrNil ifFalse: [(FileStream newFileNamed: fileName) nextPutAll: stringToSave; close]!

1 0

The Trunk: ST80-tpr.233.mcz
by commits＠source.squeak.org 08 Jan '18

08 Jan '18

1 0

The Trunk: Tools-eem.788.mcz
by commits＠source.squeak.org 07 Jan '18

07 Jan '18

Eliot Miranda uploaded a new version of Tools to project The Trunk: http://source.squeak.org/trunk/Tools-eem.788.mcz ==================== Summary ==================== Name: Tools-eem.788 Author: eem Time: 6 January 2018, 3:37:50.088654 pm UUID: bb90e476-4cf4-47bd-a8be-bc2785cc8504 Ancestors: Tools-eem.787 Add some more canonicalArgumentName implementations for well-known Collection subclasses. =============== Diff against Tools-eem.787 =============== Item was added: + ----- Method: ArrayedCollection>>canonicalArgumentName (in category '*Tools-Debugger') ----- + canonicalArgumentName + ^ 'Array'! Item was added: + ----- Method: HashedCollection>>canonicalArgumentName (in category '*Tools-Debugger') ----- + canonicalArgumentName + ^ 'Dictionary'! Item was added: + ----- Method: SequenceableCollection>>canonicalArgumentName (in category '*Tools-Debugger') ----- + canonicalArgumentName + ^ 'Sequence'! Item was added: + ----- Method: Set>>canonicalArgumentName (in category '*Tools-Debugger') ----- + canonicalArgumentName + ^ 'Set'! Item was added: + ----- Method: Symbol>>canonicalArgumentName (in category '*Tools-Debugger') ----- + canonicalArgumentName + ^ 'Symbol'!

1 0

The Trunk: Kernel-eem.1134.mcz
by commits＠source.squeak.org 07 Jan '18

07 Jan '18

Eliot Miranda uploaded a new version of Kernel to project The Trunk: http://source.squeak.org/trunk/Kernel-eem.1134.mcz ==================== Summary ==================== Name: Kernel-eem.1134 Author: eem Time: 6 January 2018, 3:24:44.111079 pm UUID: 85369d62-37bb-4b4c-a817-88a6d92097ec Ancestors: Kernel-eem.1133 More FullBlockClosure support. =============== Diff against Kernel-eem.1133 =============== Item was added: + ----- Method: ClosureExtractor>>pushFullClosure:numCopied: (in category 'instruction decoding') ----- + pushFullClosure: aCompiledBlock numCopied: numCopied + "Create a BlockClosure corresponding to the closure bytecode + and execute the action block with it. The created BlockClosure is only a pseudo value, + it is not populated with meaningful context and argument information." + | block | + block := FullBlockClosure + receiver: nil + outerContext: currentContext + method: aCompiledBlock + copiedValues: (Array new: numCopied). + currentContext := block asContextWithSender: currentContext. + action value: block! Item was added: + ----- Method: CompiledBlock>>methodForDecompile (in category 'decompiling') ----- + methodForDecompile + ^self outerCode methodForDecompile!

1 0

The Trunk: Compiler-eem.369.mcz
by commits＠source.squeak.org 06 Jan '18

06 Jan '18

Eliot Miranda uploaded a new version of Compiler to project The Trunk: http://source.squeak.org/trunk/Compiler-eem.369.mcz ==================== Summary ==================== Name: Compiler-eem.369 Author: eem Time: 6 January 2018, 1:28:37.915349 pm UUID: a342cbfd-89fd-435f-af5b-7ad501af8aeb Ancestors: Compiler-eem.368 reindexingLiteralsDo: acually needs to reset the lliteral nodes, not merely reset them, doh! Fix a bad slip in the code to size a full block. Juyst because it ends with a returnNilToCaller doesn;t mean we shouldn't generate the other nodes in the block, doh! =============== Diff against Compiler-eem.368 =============== Item was changed: ----- Method: BlockNode>>sizeCodeForEvaluatedFullClosureValue: (in category 'code generation') ----- sizeCodeForEvaluatedFullClosureValue: encoder "The closure value primitives push the arguments and the copied values. The compiler guarantees that any copied values come before all local temps. So on full closure activation we need do nothing." (statements last == NodeNil and: [self returns not]) ifTrue: + [^(self sizeCodeExceptLast: encoder) + + encoder sizeReturnNilToCaller]. - [^encoder sizeReturnNilToCaller]. ^(self sizeCodeForEvaluatedValue: encoder) + (self returns ifTrue: [0] ifFalse: [encoder sizeReturnTopToCaller])! Item was changed: ----- Method: BytecodeEncoder>>reindexingLiteralsDo: (in category 'code generation') ----- reindexingLiteralsDo: aBlock | savedNodes saveBlock | savedNodes := IdentityDictionary new. + saveBlock := [:node| savedNodes at: node put: node shallowCopy. node resetForBlockGeneration]. - saveBlock := [:node| savedNodes at: node put: node shallowCopy]. litSet do: saveBlock. litIndSet do: saveBlock. selectorSet do: saveBlock. ^aBlock ensure: [| restoreBlock | restoreBlock := [:node| node resetFromCopy: (savedNodes at: node)]. litSet do: restoreBlock. litIndSet do: restoreBlock. selectorSet do: [:node| node resetFromCopy: (savedNodes at: node ifAbsent: [node])]]! Item was added: + ----- Method: LiteralNode>>resetForBlockGeneration (in category 'code generation') ----- + resetForBlockGeneration + "Reset the receiver to an unassigned state such that its index + in the encoder's literalStream is as yet to be determined." + code := LdLitType negated. + index := nil! Item was added: + ----- Method: LiteralVariableNode>>resetForBlockGeneration (in category 'code generation') ----- + resetForBlockGeneration + "Reset the receiver to an unassigned state such that its index + in the encoder's literalStream is as yet to be determined." + code := LdLitIndType negated. + index := nil! Item was changed: ----- Method: SelectorNode>>resetForBlockGeneration (in category 'code generation') ----- resetForBlockGeneration + "Reset the receiver to an unassigned state such that its index - "Recent the receiver to an unassigned state such that its index in the encoder's literalStream is as yet to be determined." + code := SendType negated. index := nil! Item was added: + ----- Method: SpecialSelectorNode>>resetForBlockGeneration (in category 'code generation') ----- + resetForBlockGeneration + "Reset the receiver to an unassigned state such that its index + in the encoder's literalStream is as yet to be determined." + index := nil!

1 0

The Trunk: Compiler-eem.368.mcz
by commits＠source.squeak.org 06 Jan '18

06 Jan '18

Eliot Miranda uploaded a new version of Compiler to project The Trunk: http://source.squeak.org/trunk/Compiler-eem.368.mcz ==================== Summary ==================== Name: Compiler-eem.368 Author: eem Time: 6 January 2018, 1:00:10.304648 pm UUID: 3055d203-f94c-47ba-8ea0-69c82d736955 Ancestors: Compiler-eem.367 Better separate SelectorNode and SpecialSelectorNode to avoid SopecialSelectorNode deferring to SelectorNode on super sends. Hence reduce use of the obsolete code variable, makimng SelectorNode>>reserve: more straight-forward and unused in SpecialSelectorNode. =============== Diff against Compiler-eem.367 =============== Item was changed: ----- Method: MessageNode>>sizeCodeForValue: (in category 'code generation') ----- sizeCodeForValue: encoder | total | + special > 0 ifTrue: + [encoder noteOptimizedSelector: originalSelector. + ^self perform: (MacroSizers at: special) with: encoder with: true]. + receiver == NodeSuper ifTrue: + [selector := selector forSuperSend "only necessary for special selectors"]. - special > 0 - ifTrue: - [encoder noteOptimizedSelector: originalSelector. - ^self perform: (MacroSizers at: special) with: encoder with: true]. - receiver == NodeSuper - ifTrue: [selector := selector copy "only necess for splOops"]. total := selector sizeCode: encoder args: arguments size super: receiver == NodeSuper. + receiver ifNotNil: + [total := total + (receiver sizeCodeForValue: encoder)]. - receiver == nil - ifFalse: [total := total + (receiver sizeCodeForValue: encoder)]. sizes := arguments collect: [:arg | | argSize | argSize := arg sizeCodeForValue: encoder. total := total + argSize. argSize]. ^total! Item was changed: ----- Method: SelectorNode>>emitCode:args:encoder:super: (in category 'code generation') ----- emitCode: stack args: nArgs encoder: encoder super: supered + code < Send ifTrue: + [self internalEncodingError]. stack pop: nArgs. + supered - ^supered ifTrue: [(encoder supportsFullBlocks and: [encoder generatingFullBlock]) ifTrue: [encoder genPushLiteral: (encoder sharableLitIndex: encoder classEncoding). encoder genSendDirectedSuper: index numArgs: nArgs] ifFalse: [encoder genSendSuper: index numArgs: nArgs]] ifFalse: + [encoder genSend: index numArgs: nArgs]! - [encoder - genSend: (code < Send ifTrue: [code negated] ifFalse: [index]) - numArgs: nArgs]! Item was added: + ----- Method: SelectorNode>>forSuperSend (in category 'code generation') ----- + forSuperSend + "Answer the receiver or a derivative, suitable for use in a super send. + This is overridden by SpecialSelectorNode to return a SelectorNode clone." + ^self! Item was added: + ----- Method: SelectorNode>>internalEncodingError (in category 'private') ----- + internalEncodingError + self error: 'with the split between SelectorNode and SpecialSelectorNode code should never be < Send for normal sends.'! Item was added: + ----- Method: SelectorNode>>key:code:index: (in category 'initialize-release') ----- + key: object code: byte index: idx + + key := object. + code := byte. + index := idx! Item was changed: ----- Method: SelectorNode>>reserve: (in category 'code generation') ----- reserve: encoder "If this is a yet unused literal of type -code, reserve it." + index ifNil: + [index := encoder sharableLitIndex: key. + code := Send]! - code < 0 ifTrue: [code := self code: (index := encoder sharableLitIndex: key) type: 0 - code]! Item was added: + ----- Method: SelectorNode>>resetForBlockGeneration (in category 'code generation') ----- + resetForBlockGeneration + "Recent the receiver to an unassigned state such that its index + in the encoder's literalStream is as yet to be determined." + index := nil! Item was changed: ----- Method: SelectorNode>>sizeCode:args:super: (in category 'code generation') ----- sizeCode: encoder args: nArgs super: supered self reserve: encoder. + code < Send ifTrue: + [self internalEncodingError]. + supered ifFalse: + [^encoder sizeSend: index numArgs: nArgs]. + (encoder supportsFullBlocks + and: [encoder generatingFullBlock]) ifTrue: + [^(encoder sizePushLiteral: (encoder sharableLitIndex: encoder classEncoding)) + + (encoder sizeSendDirectedSuper: index numArgs: nArgs)]. + ^encoder sizeSendSuper: index numArgs: nArgs! - ^supered - ifTrue: - [code < Send "i.e. its a special selector" ifTrue: - [code := self code: (index := encoder sharableLitIndex: key) type: 5]. - (encoder supportsFullBlocks - and: [encoder generatingFullBlock]) - ifTrue: [(encoder sizePushLiteral: (encoder sharableLitIndex: encoder classEncoding)) - + (encoder sizeSendDirectedSuper: index numArgs: nArgs)] - ifFalse: [encoder sizeSendSuper: index numArgs: nArgs]] - ifFalse: - [self flag: #yuck. "special selector sends cause this problem" - encoder - sizeSend: (code < Send ifTrue: [code negated] ifFalse: [index]) - numArgs: nArgs]! Item was changed: ----- Method: SpecialSelectorNode>>emitCode:args:encoder:super: (in category 'code generation') ----- emitCode: stack args: nArgs encoder: encoder super: supered "Generate a special selector send. + A super send of a special selector must be handled like a normal send. + This is handled in MessageNode>>sizeCodeForValue: by mapping any + SpecialSelectorNodes used in super sends into normal SelectorNodes." + supered ifTrue: [self invalidSuperSendError]. + stack pop: nArgs. + encoder genSendSpecial: code numArgs: nArgs! - A super send of a special selector must be handled like a normal send." - supered - ifTrue: - [super emitCode: stack args: nArgs encoder: encoder super: supered] - ifFalse: - [stack pop: nArgs. - encoder genSendSpecial: code numArgs: nArgs]! Item was added: + ----- Method: SpecialSelectorNode>>forSuperSend (in category 'code generation') ----- + forSuperSend + "Answer the receiver or a derivative, suitable for use in a super send." + + "Excuse the code negation nonsense. It is a vestige of a very concise but now + obsolete back-end bytecode encoding scheme. We're working on removing it." + ^SelectorNode new + key: key + code: SendType negated + index: index! Item was added: + ----- Method: SpecialSelectorNode>>invalidSuperSendError (in category 'private') ----- + invalidSuperSendError + self error: 'special selectors can only be used for ordinary sends'! Item was changed: ----- Method: SpecialSelectorNode>>sizeCode:args:super: (in category 'code generation') ----- sizeCode: encoder args: nArgs super: supered "Size a special selector send. + A super send of a special selector must be handled like a normal send. + This is handled in MessageNode>>sizeCodeForValue: by mapping any + SpecialSelectorNodes used in super sends into normal SelectorNodes." + supered ifTrue: [self invalidSuperSendError]. + ^encoder sizeSendSpecial: code numArgs: nArgs! - A super send of a special selector must be handled like a normal send." - ^supered - ifTrue: [super sizeCode: encoder args: nArgs super: supered] - ifFalse: [encoder sizeSendSpecial: code numArgs: nArgs]!

1 0

Re: [squeak-dev] [Pharo-dev] (Morph>>#asDraggableMorph) isPrimitive
by Eliot Miranda 06 Jan '18

06 Jan '18

Hi Denis, Hi Norbert, On Sat, Jan 6, 2018 at 6:30 AM, Denis Kudriashov <dionisiydk(a)gmail.com> wrote: > Hi. > > I guess "quick" methods are compiled as special primitives. For example > accessors are primitives too: > > > (Point >> #x) isPrimitive > > > There is another method #isRealPrimitive which checks for explicit > primitive call. > That's right. The scheme is as follows: Numbered primitives used to be 9 bits fields and primitive numbers 1 through 255 were for conventional primitives, and primitive numbers 256 through 512 were for quick primitives, such that 256 through 263 pushed specific constants and 264 through 512 pushed the inst vars, 263 pushing inst var 0, 264 pushing inst var 1, etc. Here's the relevant section of the VM's primitive table: "Quick Push Const Methods" (256 nil) "primitivePushSelf" (257 nil) "primitivePushTrue" (258 nil) "primitivePushFalse" (259 nil) "primitivePushNil" (260 nil) "primitivePushMinusOne" (261 nil) "primitivePushZero" (262 nil) "primitivePushOne" (263 nil) "primitivePushTwo" "Quick Push Inst Var Methods" (264 519 nil) "primitiveLoadInstVar" Now we have 16 bits of primitive index. The scheme up to 32767 is the same; 1 through 32767 are available for conventional primitives, with 256 through 512 reserved for quick primitives. Numbers 32768 through 65535 are reserved for use "inline primitives". We divide the 32k values from 32767 on up into four 8k groups. Sista has the first 8k. Lowcode has the second 8k, and the remaining two groups of 8k are unassigned, reserved for future use. Now that SistaV1 and Spur support up to 64k inst vars one could argue that we should extend the range of numbers available to quick primitives. But in the JIT there's no difference between the code generated for a quick primitive and the code generated for a conventional pshInstVar:/pushConstant:, methodReturnTop. IQuicjk primitive serve to improve performance only in the interpreter. So given that we have no classes defining more than 210 inst vars (in my image) I don't think its a pressing need :-) > 2018-01-06 13:24 GMT+01:00 Norbert Hartl <norbert(a)hartl.name>: > >> Can anyone explain why >> >> (Morph>>#asDraggableMorph) isPrimitive. >> >> returns true? >> >> Norbert >> > _,,,^..^,,,_ best, Eliot

1 0

Jump to page: