Squeak-dev October 2011

squeak-dev@lists.squeakfoundation.org

47 participants
160 discussions

Re: [squeak-dev] The Trunk: Collections-ul.463.mcz
by Chris Muller 07 Oct '11

07 Oct '11

It can be worthwhile to evaluate whether it would be important to transfer the identityHash or not. It looks fine here since I think #become: behaves the same in terms of identityHash. But it's always worth remembering that it's available when employing #becomeForward:. Object new becomeForward: #size "Save the image before doing this!" Object new becomeForward: #size copyHash: false "This one is fine." The default is to copy the identityHash, which means existing IdentityDictionary's or Dictionary's whose keys do not override #hash would potentially have their hash keys change, which would require a rehash of that Dictionary. But even that isn't always possible. The copyHash: argument was introduced for Magma, because when a Proxy to a Symbol is reifiied, if the Symbol was a method selector (as it often is) that is important to the system, then it could easily lock up. On Wed, Oct 5, 2011 at 3:43 AM, <commits(a)source.squeak.org> wrote: > Levente Uzonyi uploaded a new version of Collections to project The Trunk: > http://source.squeak.org/trunk/Collections-ul.463.mcz > > ==================== Summary ==================== > > Name: Collections-ul.463 > Author: ul > Time: 4 October 2011, 8:25:51.321 pm > UUID: 2dda520b-035d-8a40-b62b-c285552d07ee > Ancestors: Collections-eem.462 > > - use #becomeForward: instead of #become: > > =============== Diff against Collections-eem.462 =============== > > Item was changed: > ----- Method: CharacterSet>>add: (in category 'collection ops') ----- > add: aCharacter > "I automatically become a WideCharacterSet if you add a wide character to myself" > > aCharacter asciiValue >= 256 > ifTrue: [| wide | > wide := WideCharacterSet new. > wide addAll: self. > wide add: aCharacter. > + self becomeForward: wide. > - self become: wide. > ^aCharacter]. > map at: aCharacter asciiValue + 1 put: 1. > ^aCharacter! > > Item was changed: > ----- Method: CharacterSetComplement>>removeAll (in category 'collection ops') ----- > removeAll > + > + self becomeForward: CharacterSet new! > - | newSet | > - newSet := CharacterSet new. > - self become: newSet! > > Item was changed: > ----- Method: LookupKey>>beBindingOfType:announcing: (in category 'bindings') ----- > beBindingOfType: aClass announcing: aBool > "Make the receiver a global binding of the given type" > + > - | old new | > (Smalltalk globals associationAt: self key) == self > ifFalse:[^self error:'Not a global variable binding']. > self class == aClass ifTrue:[^self]. > + self becomeForward: (aClass key: self key value: self value). > - old := self. > - new := aClass key: self key value: self value. > - old become: new. > "NOTE: Now self == read-only (e.g., the new binding)" > ^self recompileBindingsAnnouncing: aBool! > > >

1 0

The Inbox: Morphic-Richo.583.mcz
by commits＠source.squeak.org 06 Oct '11

06 Oct '11

A new version of Morphic was added to project The Inbox: http://source.squeak.org/inbox/Morphic-Richo.583.mcz ==================== Summary ==================== Name: Morphic-Richo.583 Author: Richo Time: 6 October 2011, 4:29:55.07 pm UUID: 4df53f26-036e-3c4a-8a98-af5e787d3ada Ancestors: Morphic-ul.582 After installing OCompletion typing any of the following characters ')]}' signals an MNU: Character>>#keyCharacter. This commit seems to fix the problem by making SmalltalkEditor>>#blinkPrevParen: always receive a character instead of a KeyboardEvent. I lacked the courage to change TextEditor>>#blinkPrevParen: but I think it should be changed as well for consistency. =============== Diff against Morphic-ul.582 =============== Item was changed: ----- Method: SmalltalkEditor>>blinkPrevParen: (in category 'parenblinking') ----- + blinkPrevParen: aCharacter - blinkPrevParen: aKeyboardEvent "Same as super, but tries to follow the Smalltalk syntax." | openDelimiter closeDelimiter level string here inside | string := paragraph text string. here := pointBlock stringIndex. + openDelimiter := aCharacter. - openDelimiter := aKeyboardEvent keyCharacter. closeDelimiter := '([{' at: (')]}' indexOf: openDelimiter). level := 1. inside := nil. "Tricky." (here > 1 and: [ (string at: here - 1) = $$ ]) ifTrue: [ ^self ]. "Just a character literal." [ level > 0 and: [ here > 1 ] ] whileTrue: [ | hereChar | hereChar := string at: (here := here - 1). inside "Are we inside a comment or string literal?" ifNotNil: [ "Yes." hereChar = inside ifTrue: [ (here > 1 and: [ (string at: here - 1) ~= inside ]) ifTrue: [ inside := nil ] ifFalse: [ here := here - 1 ] ] ] ifNil: [ (here > 1 and: [ (string at: here - 1) = $$ ]) "Just a character literal." ifTrue: [ here := here - 1 ] ifFalse: [ hereChar caseOf: { [ closeDelimiter ] -> [ (level := level - 1) = 0 ifTrue: [ ^self blinkParenAt: here ] ]. [ openDelimiter ] -> [ level := level + 1 ]. [ $" ] -> [ inside := $" ]. [ $' ] -> [ inside := $' ] } otherwise: [] ] ] ]! Item was changed: ----- Method: TextEditor>>dispatchOnKeyboardEvent: (in category 'typing support') ----- dispatchOnKeyboardEvent: aKeyboardEvent "Carry out the action associated with this character, if any. Type-ahead is passed so some routines can flush or use it." | honorCommandKeys openers closers result | (aKeyboardEvent keyCharacter == Character cr and: [ morph acceptOnCR ]) ifTrue: [ self closeTypeIn. ^ true ]. self clearParens. aKeyboardEvent keyValue = 13 ifTrue: [ aKeyboardEvent controlKeyPressed ifTrue: [ ^ self normalCharacter: aKeyboardEvent ]. aKeyboardEvent shiftPressed ifTrue: [ ^ self lf: aKeyboardEvent ]. aKeyboardEvent commandKeyPressed ifTrue: [ ^ self crlf: aKeyboardEvent ]. ^ self crWithIndent: aKeyboardEvent ]. ((honorCommandKeys := Preferences cmdKeysInText) and: [ aKeyboardEvent keyCharacter = Character enter ]) ifTrue: [ ^ self dispatchOnEnterWith: aKeyboardEvent ]. "Special keys overwrite crtl+key combinations - at least on Windows. To resolve this conflict, assume that keys other than cursor keys aren't used together with Crtl." ((self class specialShiftCmdKeys includes: aKeyboardEvent keyValue) and: [ aKeyboardEvent keyValue < 27 ]) ifTrue: [ ^ aKeyboardEvent controlKeyPressed ifTrue: [ self perform: (self class shiftCmdActions at: aKeyboardEvent keyValue + 1) with: aKeyboardEvent ] ifFalse: [ self perform: (self class cmdActions at: aKeyboardEvent keyValue + 1) with: aKeyboardEvent ] ]. "backspace, and escape keys (ascii 8 and 27) are command keys" ((honorCommandKeys and: [ aKeyboardEvent commandKeyPressed ]) or: [ self class specialShiftCmdKeys includes: aKeyboardEvent keyValue ]) ifTrue: [ ^ aKeyboardEvent shiftPressed ifTrue: [ self perform: (self class shiftCmdActions at: aKeyboardEvent keyValue + 1) with: aKeyboardEvent ] ifFalse: [ self perform: (self class cmdActions at: aKeyboardEvent keyValue + 1) with: aKeyboardEvent ] ]. "the control key can be used to invoke shift-cmd shortcuts" (honorCommandKeys and: [ aKeyboardEvent controlKeyPressed ]) ifTrue: [ ^ self perform: (self class shiftCmdActions at: aKeyboardEvent keyValue + 1) with: aKeyboardEvent ]. openers := '([{'. closers := ')]}'. result := self normalCharacter: aKeyboardEvent. (closers includes: aKeyboardEvent keyCharacter) + ifTrue: [ self blinkPrevParen: aKeyboardEvent keyCharacter]. - ifTrue: [ self blinkPrevParen: aKeyboardEvent ]. (self class autoEnclose and: [ openers includes: aKeyboardEvent keyCharacter ]) ifTrue: [ self addString: (closers at: (openers indexOf: aKeyboardEvent keyCharacter)) asString; insertTypeAhead ; moveCursor: [ :position | position - 1 ] forward: false select: false ]. ^ result!

1 0

Support for WebSocket version 7 or above in the WebClient package
by Balázs Kósi 06 Oct '11

06 Oct '11

Hi! Did anybody implement support for WebSocket version 7 or above for WebClient? For the new protocol we added to the response in WebRequest >> asWebSocket the following: resp headerAt: 'Sec-WebSocket-Accept' put: ( (SHA1 hashMessage: field, '258EAFA5-E914-47DA-95CA-C5AB0DC85B11') base64Encoded) But the framing protocol also changed: Old protocol: http://tools.ietf.org/html/draft-ietf-hybi-thewebsocketprotocol-06#page-17 New protocol: http://tools.ietf.org/html/draft-ietf-hybi-thewebsocketprotocol-07#page-17 Thanks: Balázs

3 5

ESUG Video upload continuing...
by Marcus Denker 06 Oct '11

06 Oct '11

Hi, We are slowly, but steadily uploading videos to Youtube. http://www.youtube.com/user/esugboard There are now 50 videos already. The idea is to upload all videos we have from 2008, 2009, 2010 and 2011. In addition, the "Deep into Smalltalk" tutorials. If you want to be notified about single uploads, there are multiple possibilities: - subscribe to Youtube, turn on email notification. This way you get title, screnshot and the Abstract via mail of every video uploaded. - subscribe to http://twitter.com/#!/esugsmalltalk Youtube sends a tweet automatically for every upload. (Please retweet the ones you like!) Marcus -- Marcus Denker -- http://marcusdenker.de

1 0

Daily Commit Log
by commits＠source.squeak.org 06 Oct '11

06 Oct '11

Changes to Trunk (http://source.squeak.org/trunk.html) in the last 24 hours: http://lists.squeakfoundation.org/pipermail/packages/2011-October/005014.ht… Name: Collections-ul.463 Ancestors: Collections-eem.462 - use #becomeForward: instead of #become: ============================================= http://lists.squeakfoundation.org/pipermail/packages/2011-October/005015.ht… Name: EToys-ul.86 Ancestors: EToys-ul.85 - use #becomeForward: instead of #become: ============================================= http://lists.squeakfoundation.org/pipermail/packages/2011-October/005016.ht… Name: Morphic-ul.582 Ancestors: Morphic-bf.581 - use #becomeForward: instead of #become: - fix for Mantis #7674 http://bugs.squeak.org/view.php?id=7674 ============================================= http://lists.squeakfoundation.org/pipermail/packages/2011-October/005017.ht… Name: MorphicExtras-ul.103 Ancestors: MorphicExtras-ul.102 - use #becomeForward: instead of #become: ============================================= http://lists.squeakfoundation.org/pipermail/packages/2011-October/005018.ht… Name: ST80-ul.137 Ancestors: ST80-nice.136 - use #becomeForward: instead of #become: ============================================= http://lists.squeakfoundation.org/pipermail/packages/2011-October/005019.ht… Name: System-ul.455 Ancestors: System-eem.454, System-ul.454 - use #becomeForward: instead of #become: if possible - merged System-ul.454 ============================================= http://lists.squeakfoundation.org/pipermail/packages/2011-October/005020.ht… Name: System-ul.454 Ancestors: System-ul.453 Use #copyReplaceAll:with: instead of #copyReplaceAll:with:asTokens:. =============================================

1 0

Cmd-h on Mac OS X?
by Yoshiki Ohshima 05 Oct '11

05 Oct '11

I recently started using Mac for Squeak coding. I knew this has been a problem for some time but what can I do to setSearchString: when I pressed Cmd-h? Also, when I run a 3.8-derived image on "Squeak 5.7.4.1" on OS 10.6.7, I get a weird display artifact. Since the VM is for cocoa, I just shouldn't even bother running it and use "Squeak 4.2.5beta1U" instead? -- Yoshiki

3 10

[ANN] Updated FunSqueak with Alice
by Edgar J. De Cleene 05 Oct '11

05 Oct '11

http://ftp.squeak.org/various_images/FunSqueak/FunSqueakCog4.3-11720-alpha.z ip Sure have bugs, but for the impatient here have Alice (kind off) Video of FunSqueak coming (faster as iPhone 5 :=) Enjoy Edgar

2 1

Alice Revival toma 3
by Edgar J. De Cleene 05 Oct '11

05 Oct '11

Now we have some working.... https://lh6.googleusercontent.com/-hW0kj1zV41A/TowvmrwnNfI/AAAAAAAAALM/4ycnO 7Xw2eM/s720/AliceRevival2.jpg Class from exported .cs of old FunSqueak was not initialized as should be and I must do manual work. Help me, how I could be sure ALL classes of .cs was proper initialized ? I using the .mdl coming in Cobalt in Avatar folder. It¹s faire use it ? License issues ? Edgar

1 0

Alice Revival toma 2
by Edgar J. De Cleene 05 Oct '11

05 Oct '11

How to export and load shared pools ? Here the black art I using for this. I attach ObjectCompatibleENH.1 which I use for exchange some between Squeak, Pharo and Cuis. I like export what I need in Foo.obj form. In this case the pool In FunSqueak3.10alpha.7 Workspace type B3DEngineConstants and do inspect or explore. Navigate to classPool and inspect the dictionary. In the code panel type self saveOnFileNamed: 'B3DEngineConstantspool' Drag and drop the object into the FunSqueakCog4.3-11712-alpha, you see some asking load object (twice, 3.10 only shows one , why ?) Accept and you have a inspector on this object. In the code pane type B3DEngineConstants classPool: self Presto , your have the values you need and no nil when load the .cs for first time. Sure exists nicer ways , so I wish learn. Edgar

1 0

The Trunk: System-ul.454.mcz
by commits＠source.squeak.org 05 Oct '11

05 Oct '11

Levente Uzonyi uploaded a new version of System to project The Trunk: http://source.squeak.org/trunk/System-ul.454.mcz ==================== Summary ==================== Name: System-ul.454 Author: ul Time: 21 September 2011, 12:20:55.326 am UUID: 5682b4d1-1432-ea44-abc0-9f79b87e9f1b Ancestors: System-ul.453 Use #copyReplaceAll:with: instead of #copyReplaceAll:with:asTokens:. =============== Diff against System-ul.453 =============== Item was changed: ----- Method: SARInstaller>>fileInPackageNamed: (in category 'client services') ----- fileInPackageNamed: memberName "This is to be used from preamble/postscript code to file in zip members as DVS packages." | member current new baseName imagePackageLoader packageInfo streamPackageLoader packageManager | member := self zip memberNamed: memberName. member ifNil: [ ^self errorNoSuchMember: memberName ]. imagePackageLoader := Smalltalk at: #ImagePackageLoader ifAbsent: []. streamPackageLoader := Smalltalk at: #StreamPackageLoader ifAbsent: []. packageInfo := Smalltalk at: #PackageInfo ifAbsent: []. packageManager := Smalltalk at: #FilePackageManager ifAbsent: []. "If DVS isn't present, do a simple file-in" (packageInfo isNil or: [imagePackageLoader isNil or: [streamPackageLoader isNil]]) ifTrue: [ ^ self fileInMemberNamed: memberName ]. + baseName := memberName copyReplaceAll: '.st' with: ''. - baseName := memberName copyReplaceAll: '.st' with: '' asTokens: false. (packageManager allManagers anySatisfy: [ :pm | pm packageName = baseName ]) ifTrue: [ current := imagePackageLoader new package: (packageInfo named: baseName). new := streamPackageLoader new stream: member contentStream ascii. (new changesFromBase: current) fileIn ] ifFalse: [ self class fileIntoChangeSetNamed: baseName fromStream: member contentStream ascii setConverterForCode. ]. packageManager registerPackage: baseName. self installed: member.!

1 0

Jump to page: