Squeak-dev July 2021

squeak-dev@lists.squeakfoundation.org

25 participants
180 discussions

OpenSmalltalk git has -vm-display-fbdev but exec does not understand
by chaeremon＠t-online.de 14 Jul '21

14 Jul '21

Hi @all, when searching https://github.com/OpenSmalltalk/opensmalltalk-vm/search?q=fbdev the code for fbdev is present (88 code results), but the exec (on Debian buster, Squeak5.3-19435-32bit-202003021730-ARMv6.zip) does not want to recognize -vm-display-fbdev option. Is this a compile-/build-time thing which has not been applied for the available packes? Or what can be done to see the performance of -vm-display-fbdev (want to test/compare it on RPI 3+). /Klaus

1 0

The Trunk: Morphic-mt.1770.mcz
by commits＠source.squeak.org 13 Jul '21

13 Jul '21

Marcel Taeumel uploaded a new version of Morphic to project The Trunk: http://source.squeak.org/trunk/Morphic-mt.1770.mcz ==================== Summary ==================== Name: Morphic-mt.1770 Author: mt Time: 11 June 2021, 7:54:44.535233 am UUID: 538ca9f9-e591-3344-a2b9-e83cd482d4c8 Ancestors: Morphic-mt.1769 Updates representation of integers in ObjectExplorer. For a small discussion, see http://forum.world.st/The-Trunk-Morphic-mt-1769-mcz-tp5129501p5129804.html Enables the use of #balloonText: for ObjectExplorerWrapper. Already used in FileDirectoryWrapper. Not available in basic ListItemWrapper. (Yet?) =============== Diff against Morphic-mt.1769 =============== Item was changed: ----- Method: Integer>>explorerContents (in category '*Morphic-Explorer') ----- explorerContents + ^ Array streamContents: [:stream | + #( + 'hex' 16 2 0 'hexadecimal, base 16' + 'oct' 8 3 0 'octal, base 8' + 'bin' 2 4 0 'binary, base 2' + 'bit' nil 4 8 'Logical bit field\- Two complement\- Infinite sign bits\- Try #bitAt:') groupsDo: [ :key :base :group :padding :balloon | | label | + label := self abs printStringBase: (base ifNil: [2]). + base ifNil: [ + label := String streamContents: [:s | + (label size roundUpTo: padding) to: 1 by: -1 do: [:index | + s nextPutAll: (self bitAt: index) asString]]]. + label := label + padded: #left + to: (label size roundUpTo: group) + padding + with: ((base isNil and: [self negative]) ifTrue: [$1] ifFalse: [$0]). - ^#( - ('hexadecimal' 16 2) - ('octal' 8 3) - ('binary' 2 4)) collect: [ :each | | label group | - group := each third. - label := self abs printStringBase: each second. - label := label padded: #left to: (label size roundUpTo: group) with: $0. label := String streamContents: [:s | + (self negative and: [base notNil]) ifTrue: [s nextPutAll: '-']. + base ifNil: [s nextPutAll: '... ']. + (1 to: label size by: group) + do: [:index | + 1 to: group do: [:gIndex | + s nextPut: (label at: index + gIndex - 1)]] + separatedBy: [s space]]. + stream nextPut: (( + ObjectExplorerWrapper + with: label + name: key + model: self) + balloonText: balloon withCRs; + yourself) ]]! - self negative ifTrue: [s nextPutAll: '- ']. - (1 to: label size by: group) - do: [:index | - 1 to: group do: [:gIndex | - s nextPut: (label at: index + gIndex - 1)]] - separatedBy: [s space]]. - - ObjectExplorerWrapper - with: label - name: each first translated - model: self ]! Item was changed: ListItemWrapper subclass: #ObjectExplorerWrapper + instanceVariableNames: 'itemName parent balloonText' - instanceVariableNames: 'itemName parent' classVariableNames: 'ShowContentsInColumns' poolDictionaries: '' category: 'Morphic-Explorer'! !ObjectExplorerWrapper commentStamp: 'pre 5/15/2017 21:23' prior: 0! ObjectExplorerWrappers represent an item displayed in an object explorer tree. In addition to the common ListItemWrapper behavior it adds methods to refresh the list entry with updated values from the model. It is mostly used in #explorerContents methods to describe which instance variables of an object should be displayed in the explorer. Additionally, the value displayed can be visualized with a small icon which is defined by the class of the value object through the method #iconOrThumbnailOfSize:. Contributed by Bob Arning as part of the ObjectExplorer package. ! Item was added: + ----- Method: ObjectExplorerWrapper>>balloonText (in category 'accessing') ----- + balloonText + + ^ balloonText! Item was added: + ----- Method: ObjectExplorerWrapper>>balloonText: (in category 'accessing') ----- + balloonText: stringOrNil + + balloonText := stringOrNil.!

4 4

The Inbox: Morphic-ct.1777.mcz
by commits＠source.squeak.org 12 Jul '21

12 Jul '21

A new version of Morphic was added to project The Inbox: http://source.squeak.org/inbox/Morphic-ct.1777.mcz ==================== Summary ==================== Name: Morphic-ct.1777 Author: ct Time: 12 July 2021, 5:02:42.077924 pm UUID: 7c4d08f6-93cc-464c-aafc-774dec51a942 Ancestors: Morphic-mt.1776 Fixes mouse wheel handling in nested scroll panes. Until now, you cannot scroll the outer pane while hovering the inner pane, even if the inner pane already does not actually react on the wheel event or does not have a scrollable dimension at all. With this patch, the behavior is aligned to many common implementations (i.e., Chromium or Android UI). Example setup: (PluggableTextMorph on: ('Hello\\{1}\\World' withCRs asText format: { Character startOfHeader asString asText addAttribute: ((PluggableTextMorph on: 'wonderful\\Squeak' withCRs text: #yourself accept: nil) height: 50; asTextAnchor); yourself }) text: #yourself accept: nil) height: 50; openInHand. This would be a helpful improvement for Babylonian where such nested text morphs actually exist. =============== Diff against Morphic-mt.1776 =============== Item was changed: ----- Method: ScrollPane>>mouseWheel: (in category 'event handling') ----- mouseWheel: evt + | oldOffset | + oldOffset := scroller offset. + evt isWheelUp ifTrue: [scrollBar scrollUp: (evt verticalScrollDelta: self class verticalScrollDeltaPerMouseWheelNotch)]. evt isWheelDown ifTrue: [scrollBar scrollDown: (evt verticalScrollDelta: self class verticalScrollDeltaPerMouseWheelNotch)]. evt isWheelLeft ifTrue: [hScrollBar scrollUp: (evt horizontalScrollDelta: self class horizontalScrollDeltaPerMouseWheelNotch)]. + evt isWheelRight ifTrue: [hScrollBar scrollDown: (evt horizontalScrollDelta: self class horizontalScrollDeltaPerMouseWheelNotch)]. + + evt wasHandled: oldOffset ~= scroller offset.! - evt isWheelRight ifTrue: [hScrollBar scrollDown: (evt horizontalScrollDelta: self class horizontalScrollDeltaPerMouseWheelNotch)].!

1 0

The Trunk: Kernel-eem.1407.mcz
by commits＠source.squeak.org 12 Jul '21

12 Jul '21

Eliot Miranda uploaded a new version of Kernel to project The Trunk: http://source.squeak.org/trunk/Kernel-eem.1407.mcz ==================== Summary ==================== Name: Kernel-eem.1407 Author: eem Time: 9 July 2021, 2:46:18.688883 pm UUID: 48a520fa-d6a2-44fa-8508-3405e7e0b8b9 Ancestors: Kernel-mt.1406 Have scanForInstructionSequence: descend into full blocks. Eliminate cases handled by otherwise clause in installVirtualKeyTable. Add error code to resue primitive invocation. =============== Diff against Kernel-mt.1406 =============== Item was changed: ----- Method: EventSensor class>>installVirtualKeyTable (in category 'class initialization') ----- installVirtualKeyTable VirtualKeyTable := Smalltalk windowSystemName caseOf: { ['Windows'] -> [self virtualKeysOnWindows]. ['Win32' "older VMs"] -> [self virtualKeysOnWindows]. ['Aqua'] -> [self virtualKeysOnMacOS]. ['X11'] -> [self virtualKeysOnX11]. - ['RiscOS'] -> [Dictionary new]. - ['Quartz'] -> [Dictionary new]. } otherwise: [Dictionary new]. "Shift 8 bits to not overwrite virtual-key mappings from above." self mapControlKeysToCommandKeys ifTrue: [ VirtualKeyTable at: (2r0010 "ctrl" bitShift: 8) put: (2r1010 "cmd+ctrl"). VirtualKeyTable at: (2r0011 "ctrl+shift" bitShift: 8) put: (2r1011 "cmd+ctrl+shift")]. self mapAltKeysToOptionKeys ifTrue: [ VirtualKeyTable at: (2r1000 "cmd/alt" bitShift: 8) put: (2r1100 "cmd/alt+opt"). VirtualKeyTable at: (2r1001 "cmd/alt+shift" bitShift: 8) put: (2r1101 "cmd/alt+opt+shift")].! Item was changed: ----- Method: InstructionStream>>scanForInstructionSequence: (in category 'scanning') ----- scanForInstructionSequence: naryBlock "naryBlock is a block taking one or more arguments. Evaluate it for each sequence of instructions of length n in the receiver until naryBlock evaluates to true. Answer if naryBlock evaluated to true." | instructions | instructions := OrderedCollection withAll: ((1 to: naryBlock numArgs) collect: [:ign| + self atEnd ifFalse: + [self nextInstruction]]). + instructions last ifNil: "not enough instructions, so no match in this method, but must check for blocks" + [instructions do: + [:instOrNil| + instOrNil ifNil: [^false]. + (instOrNil selector == #pushFullClosure:numCopied: + and: [(self class on: instOrNil argument) scanForInstructionSequence: naryBlock]) ifTrue: + [^true]]]. - self atEnd ifTrue: [^false]. - self nextInstruction]). [(naryBlock valueWithArguments: instructions asArray) ifTrue: [^true]. + (instructions first selector == #pushFullClosure:numCopied: + and: [(self class on: instructions first argument) scanForInstructionSequence: naryBlock]) ifTrue: + [^true]. self atEnd] whileFalse: [instructions removeFirst; addLast: self nextInstruction]. ^false! Item was changed: ----- Method: Process>>resume (in category 'changing process state') ----- resume "Primitive. Allow the process that the receiver represents to continue. Put the receiver in line to become the activeProcess. Fail if the receiver is already waiting in a queue (in a Semaphore or ProcessScheduler). Fail if the receiver's suspendedContext is not a context. Essential. See Object documentation whatIsAPrimitive." + <primitive: 87 error: ec> - <primitive: 87> self primitiveFailed!

2 1

Loading OMeta2 into Squeak 6.0Alpha?
by Michał Olszewski 10 Jul '21

10 Jul '21

Hello, I'd like to load OMeta2 package into Squeak 6.0 alpha (version 20582. 64 bit) but I'm having issue with that. Attempting to load OMeta2.mcm results in UndefinedObject>>findNextHandlerContextStarting (As result of #OMeta2Fail). It seems like exception handling mechanism attempts to access sender but the sender is nil. I tried also to load manually OMeta2-Preload-hmm.14.mcz (which works), then OMeta2-Postload-hmm12.mcz but the same error happens. I recall, in Squeak5.3, this wasn't happening but further errors were related with Compiler interface. I managed to fix that, by adding Parser's method to #OMeta2Compiler. Unfortunalely there were even more of these, which I could not fix. Is there any way to make it work? The package works fine in Squeak 4.3 but I want to work in newer versions... PS. The image is fresh, I didn't even use Preference Wizard. My OS is Windows 8.1 64 bit. Michał

3 4

The Inbox: Regex-Core-ct.61.mcz
by commits＠source.squeak.org 09 Jul '21

09 Jul '21

A new version of Regex-Core was added to project The Inbox: http://source.squeak.org/inbox/Regex-Core-ct.61.mcz ==================== Summary ==================== Name: Regex-Core-ct.61 Author: ct Time: 8 July 2021, 1:30:44.09436 am UUID: 63655b8f-ad42-0946-b6fe-4dc3100995f1 Ancestors: Regex-Core-ct.59 Adds String >> #escapeRegex to escape special characters in a string before composing it into another regex. Usage: ':-)' matchesRegex: ':-)' escapeRegex =============== Diff against Regex-Core-ct.59 =============== Item was added: + ----- Method: RxParser class>>escapeString: (in category 'utilities') ----- + escapeString: aString + "Answer a copy of aString which does not contain any unescaped characters. This is the inverse function of String >> #matchesRegex:. + NB: Basically, we could simply escape every single character in the string, but this would not produce human-readable outputs." + + ^ aString + copyWithRegex: ('[{1}]' format: {self specialCharacters collect: [:character | '\', character]}) + matchesTranslatedUsing: [:match | '\', match]! Item was added: + ----- Method: RxParser class>>specialCharacters (in category 'utilities') ----- + specialCharacters + + ^ #($( $) $[ $] $* $+ $? ${ $} $. $^ $$ $: $\)! Item was added: + ----- Method: String>>escapeRegex (in category '*Regex-Core') ----- + escapeRegex + + ^ RxParser escapeString: self!

7 13

Re: [squeak-dev] We need to talk about graphics
by Rolandas Guobys 09 Jul '21

09 Jul '21

please don't send me more letters !!! > On 2020-07-12, at 21:30, Yoshiki Ohshima <Yoshiki.Ohshima(a)acm.org> wrote: >

1 0

generators
by Eliot Miranda 09 Jul '21

09 Jul '21

Hi All, I juts had occasion to use generators in anger for the first time. I was trying to fnd out why a clone of a prse tree was reporting that it was not the same tree. Generators made enumerating all nodes in each tree together trivial: | them | them := Generator on: [:g| newTree nodesDo: [:n| g yield: n]]. parseTree nodesDo: [:n| | twin | twin := them next. (n isSameAs: twin) ifFalse: [self halt]] My message though comes from noticing that generators yield their values via next, so why not use nextPut: instead of yield: ? e.g. I think this is better: | them | them := Generator on: [:g| newTree nodesDo: [:n| g nextPut: n]]. parseTree nodesDo: [:n| | twin | twin := them next. (n isSameAs: twin) ifFalse: [self halt]] _,,,^..^,,,_ best, Eliot

5 4

The Trunk: Regex-Tests-Core-ct.13.mcz
by commits＠source.squeak.org 08 Jul '21

08 Jul '21

Marcel Taeumel uploaded a new version of Regex-Tests-Core to project The Trunk: http://source.squeak.org/trunk/Regex-Tests-Core-ct.13.mcz ==================== Summary ==================== Name: Regex-Tests-Core-ct.13 Author: ct Time: 8 July 2021, 12:45:29.187695 am UUID: 77b31977-aedb-324b-b792-fc32346ff596 Ancestors: Regex-Tests-Core-ct.12 Adds a regression test for lookbehinds in optional/repeatable groups which should work. =============== Diff against Regex-Tests-Core-ct.12 =============== Item was added: + ----- Method: RxParserTest>>testOptionalLookbehind (in category 'tests') ----- + testOptionalLookbehind + + self assert: ['A' matchesRegex: '((?<=^)A)+']!

1 0

The Trunk: Regex-Tests-Core-ct.14.mcz
by commits＠source.squeak.org 08 Jul '21

08 Jul '21

Marcel Taeumel uploaded a new version of Regex-Tests-Core to project The Trunk: http://source.squeak.org/trunk/Regex-Tests-Core-ct.14.mcz ==================== Summary ==================== Name: Regex-Tests-Core-ct.14 Author: ct Time: 8 July 2021, 12:52:32.198695 am UUID: 7fc43ad7-9987-9f48-aee8-07498df21307 Ancestors: Regex-Tests-Core-ct.13 Adds another regression test for lookbehinds in nullable closures. This one is more serious; it raises a nil MNU because the marker chain is defect. Just adding this for now as a known bug, I don't have the resources to fix this right now. But if no one else is keen on this, I will probably give this another try in my next holidays ... :-) =============== Diff against Regex-Tests-Core-ct.12 =============== Item was added: + ----- Method: RxParserTest>>testOptionalLookbehind (in category 'tests') ----- + testOptionalLookbehind + + self assert: ['A' matchesRegex: '((?<=^)A)+']. + + self assert: [('AB' allRegexMatches: '((?<=a)b)?') asArray = #('A')].!

1 0

← Newer
1
...
11
12
13
14
15
16
17
18
Older →

Jump to page: