Squeak-dev November 2018

squeak-dev@lists.squeakfoundation.org

48 participants
94 discussions

When did it become a good idea for selected text to be enclosed by () instead of replaced by 9 ?
by tim Rowledge 06 Nov '19

06 Nov '19

In trying to see if anything had changed re WaveEditor etc I updated an image to update 16963. When I went to save this image a Squeak5.1-16963… I selected the last thre digits of the prior name (Squeak5.1-16549…) intending to type 963. Imagine my surprise - nay, alarm! - when I got ‘Squeak5.1-16(963) instead. Oh my, I said to myself, what has Pooh done now. As best I can work out this is a result of the change in TextEditor>enclose: that added ‘9’ some time ago (and which did not affect me since I normally turn off the annoying enclose stuff in preferences) combined with a more recent change to TextEditor>dispatchOnKeyboardEvent: wherein a use of TextEditor>enclose: was added near the end. I see two obviously related Preferences for enclosing text. I see this as a dangerous thing, because multiple preferences interacting on the same issue can cause chaos, not to say confusion. And every time a preference is added it is yet another sheaf of potential problems to test. So, yes, I can turn it off easily enough. But how could it ever be unsurprising to a user to get a(bc)d instead of the expected a9d ? tim -- tim Rowledge; tim(a)rowledge.org; http://www.rowledge.org/tim Useful random insult:- Fell out of the family tree.

6 6

Update on source.squeak.org upgrade
by Chris Muller 06 Jan '19

06 Jan '19

Well, I've been running the latest code and VM every day on my laptop as a Personal Squeaksource server and using it for all of my development these last weeks. It went so smoothly I also decided push on it by downloading the entire squeaksource.com (17GB!) and starting it locally using this same new code, but with different Repository combination more suitable to that application (ImagePersistence + FileSystem). That exposed a couple of minor bugs which are now fixed and committed to the /ss repository. I'm going to bang on it one more time but I believe that guy is ready to upgrade too! If you happened to try Personal SqueakSource and have any negative feedback, please let me know soon, otherwise I'll start thinking about upgrading it on the server and kissing the timeouts goodbye. - Chris

7 8

The Trunk: Morphic-cmm.1454.mcz
by commits＠source.squeak.org 31 Dec '18

31 Dec '18

Chris Muller uploaded a new version of Morphic to project The Trunk: http://source.squeak.org/trunk/Morphic-cmm.1454.mcz ==================== Summary ==================== Name: Morphic-cmm.1454 Author: cmm Time: 23 June 2018, 4:23:51.324266 pm UUID: 7bd5a7a9-4493-4cf4-849f-be6d8f8bb991 Ancestors: Morphic-hjh.1453 - Fix ability of border styles of veryDeepCopied Morphs to be independently set. - Fix dispatch regression for Morphs that leave world based on mouseLeave. - Fix three UI-gesture inconsistencies related to invoking halos, depending on the position of existing halos: 1) halo activation was sometimes on mouseUp instead of mouseDown, 2) blue move/resize gestures sometimes operated on the background window instead of the foreground, and, 3) even with the (Shift) modifier key, it was sometimes selecting the outermost instead of the innermost. All three are fixed while reverting the increased dependence on #transferHalo:, putting us closer to considerations like multiple, independent, and custom halos. - Restore splitter bar thicknesses from 2 back to 4 pixels. - Restore blue-click functionality of splitters when Smart Splitters is off. - Let SystemWindows dismiss halos with any button, mouse or keyboard. =============== Diff against Morphic-hjh.1453 =============== Item was added: + ----- Method: ComplexBorder>>veryDeepInner: (in category 'copying') ----- + veryDeepInner: aDeepCopier + super veryDeepInner: aDeepCopier. + style := style veryDeepCopyWith: aDeepCopier. + colors := colors veryDeepCopyWith: aDeepCopier. + lineStyles := lineStyles veryDeepCopyWith: aDeepCopier.! Item was changed: ----- Method: Editor class>>dumbbellCursor (in category 'preferences') ----- dumbbellCursor + <preference: 'Dumbbell Text Cursor' - <preference: 'Dumbbell-shaped Text Cursor' category: 'Morphic' + description: 'When enabled, the text cursor assumes the shape of a dumbbell, otherwise a vertical bar.' - description: 'When true, the text cursor assumes the shape of a dumbbell, otherwise a vertical bar..' type: #Boolean> ^ DumbbellCursor ifNil: [ false ]! Item was changed: ----- Method: MorphicEventDispatcher>>dispatchFocusEventAllOver:with: (in category 'focus events') ----- dispatchFocusEventAllOver: evt with: focusMorph "Like a full event dispatch BUT adds regular dispatch if the focus morph did nothing with the event. This is useful for letting the focusMorph's siblings handle the events instead. Take halo invocation as an example. See senders of me." | result hand mouseFocus | result := self dispatchFocusEventFully: evt with: focusMorph. result == #rejected ifTrue: [^ result]. result wasIgnored ifTrue: [^ result]. result wasHandled ifTrue: [^ result]. + focusMorph world ifNil: [ ^ result ]. hand := evt hand. mouseFocus := hand mouseFocus. [ "Avoid re-dispatching the event to the focus morph. See Morph >> #rejectsEvent:." focusMorph lock. "Handle side effect for mouse-enter and mouse-leave events." self flag: #hacky. "mt: Maybe we find a better way to synthesize enter/leave events in the future." hand newMouseFocus: nil. hand mouseOverHandler processMouseOver: hand lastEvent. "Give the morph's world a chance to normally dispatch the event." + ^ focusMorph world ifNotNil: [ : world | world processEvent: evt using: self] - ^ focusMorph world processEvent: evt using: self ] ensure: [ focusMorph unlock. evt hand newMouseFocus: mouseFocus].! Item was changed: ----- Method: PasteUpMorph>>tryInvokeHalo: (in category 'events-processing') ----- + tryInvokeHalo: aUserInputEvent + "Invoke halos around the top-most world container at aUserInputEvent's #position. If it was already halo'd, zero-in on its next inward component morph at that position. Holding Shift during the click reverses this traversal order." + | stack innermost haloTarget | + Preferences noviceMode ifTrue: [ ^ self ]. + Morph haloForAll ifFalse: [ ^ self ]. + "the stack is the top-most morph to bottom-most." + stack := (self morphsAt: aUserInputEvent position unlocked: true) select: + [ : each | each wantsHaloFromClick or: [ each handlesMouseDown: aUserInputEvent ] ]. + innermost := aUserInputEvent hand halo + ifNil: [ stack first ] + ifNotNil: + [ : existingHalo | stack allButFirst "existingHalo is first on the stack, not a target" + detect: [ : each | each owner == self ] + ifFound: + [ : worldContainer | "Is existingHalo's target part of the same worldContainer as the morph clicked?" + (existingHalo target withAllOwners includes: worldContainer) + ifTrue: [ "same hierarchy, let #transferHalo: continue to handle it for now." ^ self ] + ifFalse: + [ "different hierarchy, remove + add." + aUserInputEvent hand removeHalo. + aUserInputEvent shiftPressed + ifTrue: [ stack second "first is still the just removed halo" ] + ifFalse: [ worldContainer ] ] ] + ifNone: [ "Shouldn't get here, but defensive code." self ] ]. + "If modifier key is pressed, start at innermost (the target), otherwise the outermost (direct child of the world (self))." + haloTarget := aUserInputEvent shiftPressed + ifTrue: [ innermost ] + ifFalse: [ innermost == self ifTrue: [innermost] ifFalse: [(innermost withAllOwners copyWithout: self) last] ]. + haloTarget wantsHaloFromClick ifFalse: [ "haloTarget has its own event handler." ^ self ]. + "Now that we have the haloTarget, show the halo." + aUserInputEvent hand + newMouseFocus: haloTarget + event: aUserInputEvent. + haloTarget invokeHaloOrMove: aUserInputEvent. + "aUserInputEvent has been consumed, don't let it cause any further side-effects." + aUserInputEvent ignore! - tryInvokeHalo: anEvent - - | innerMost target | - anEvent hand halo ifNotNil: [^ self "No invocation needed. Halo will handle transfer itself."]. - Preferences noviceMode ifTrue: [^ self "No halo in novice mode."]. - Morph haloForAll ifFalse: [^ self]. - - innerMost := (self morphsAt: anEvent position unlocked: true) first. - - "1) Try to use innermost morph but skip all the ones that do not want to show a halo along the owner chain." - target := innerMost. - [target isNil or: [target wantsHaloFromClick]] - whileFalse: [target := target owner]. - target ifNil: [^ self]. - - "2) Without a modifier, which is normal, find the outermost container for that inner morph." - (innerMost == self or: [anEvent shiftPressed]) ifFalse: [ - | previousTargets | - previousTargets := OrderedCollection new. - [target notNil and: [target owner ~~ self]] whileTrue: [ - previousTargets add: target. - target := target owner]. - target ifNil: [^ self]. - [previousTargets isEmpty or: [target wantsHaloFromClick]] whileFalse: [ - target := previousTargets removeLast]. - target wantsHaloFromClick ifFalse: [^ self]]. - - "3) Now that we have the target, show the halo. Abort event dispatching, too, to avoid confusion." - anEvent hand newMouseFocus: target event: anEvent. - target invokeHaloOrMove: anEvent. - anEvent ignore.! Item was changed: ----- Method: ProportionalSplitterMorph>>balanceOffsets (in category 'layout') ----- balanceOffsets + | fdx fdy | + (self hasProperty: #fullDelta) ifFalse: [^ self]. + fdx := (self valueOfProperty: #fullDelta) x. + fdy := (self valueOfProperty: #fullDelta) y. + self layoutFrame hasFixedHeight ifTrue: [ | otop obot ctop cbot topf | otop := (owner submorphs detect: [:m | m layoutFrame topFraction isZero] ifNone: [^ self]) in: [:tm | tm top - (tm layoutFrame topOffset ifNil: [0])]. obot := (owner submorphs detect: [:m | m layoutFrame bottomFraction = 1] ifNone: [^ self]) in: [:tm | tm bottom - (tm layoutFrame bottomOffset ifNil: [0])]. ctop := (self layoutFrame topFraction * (obot - otop)) rounded + otop + (self layoutFrame topOffset ifNil: [0]). cbot := (self layoutFrame bottomFraction * (obot - otop)) rounded + otop + (self layoutFrame bottomOffset ifNil: [0]). topf := self layoutFrame topFraction. self layoutFrame topFraction: ((ctop + cbot) * 0.5 - otop) / (obot - otop) asFloat. self layoutFrame bottomFraction: self layoutFrame topFraction. + self layoutFrame topOffset: self layoutFrame topOffset - fdy. + self layoutFrame bottomOffset: self layoutFrame bottomOffset - fdy. - self layoutFrame topOffset: ctop - - (self layoutFrame topFraction * (obot - otop) + otop) truncated. - self layoutFrame bottomOffset: cbot - - (self layoutFrame bottomFraction * (obot - otop) + otop) rounded. (leftOrTop copy union: rightOrBottom) do: [:m | (m layoutFrame topFraction closeTo: topf) ifTrue: [ m layoutFrame topFraction: self layoutFrame topFraction. + m layoutFrame topOffset: m layoutFrame topOffset - fdy]. - m layoutFrame topOffset: - m layoutFrame topOffset - (self valueOfProperty: #fullDelta) y]. (m layoutFrame bottomFraction closeTo: topf) ifTrue: [ m layoutFrame bottomFraction: self layoutFrame topFraction. + m layoutFrame bottomOffset: m layoutFrame bottomOffset - fdy]]] . - m layoutFrame bottomOffset: - m layoutFrame bottomOffset - (self valueOfProperty: #fullDelta) y.]]] . self layoutFrame hasFixedWidth ifTrue: [ | oleft oright cleft cright leftf | oleft := (owner submorphs detect: [:m | m layoutFrame leftFraction isZero] ifNone: [^ self]) in: [:tm | tm left - (tm layoutFrame leftOffset ifNil: [0])]. oright := (owner submorphs detect: [:m | m layoutFrame rightFraction = 1] ifNone: [^ self]) in: [:tm | tm right - (tm layoutFrame rightOffset ifNil: [0])]. cleft := (self layoutFrame leftFraction * (oright - oleft)) rounded + oleft + (self layoutFrame leftOffset ifNil: [0]). cright := (self layoutFrame rightFraction * (oright - oleft)) rounded + oleft + (self layoutFrame rightOffset ifNil: [0]). leftf := self layoutFrame leftFraction. self layoutFrame leftFraction: ((cleft + cright) * 0.5 - oleft) / (oright - oleft) asFloat. self layoutFrame rightFraction: self layoutFrame leftFraction. - self layoutFrame leftOffset: cleft - - (self layoutFrame leftFraction * (oright - oleft) + oleft) truncated. - self layoutFrame rightOffset: cright - - (self layoutFrame rightFraction * (oright - oleft) + oleft) rounded. + self layoutFrame leftOffset: self layoutFrame leftOffset - fdx. + self layoutFrame rightOffset: self layoutFrame rightOffset - fdx. + (leftOrTop copy union: rightOrBottom) do: [:m | (m layoutFrame leftFraction closeTo: leftf) ifTrue: [ m layoutFrame leftFraction: self layoutFrame leftFraction. + m layoutFrame leftOffset: m layoutFrame leftOffset - fdx]. - m layoutFrame leftOffset: - m layoutFrame leftOffset - (self valueOfProperty: #fullDelta) x]. (m layoutFrame rightFraction closeTo: leftf) ifTrue: [ m layoutFrame rightFraction: self layoutFrame leftFraction. + m layoutFrame rightOffset: m layoutFrame rightOffset - fdx.]]] . + - m layoutFrame rightOffset: - m layoutFrame rightOffset - (self valueOfProperty: #fullDelta) x.]]] . - self removeProperty: #fullDelta. owner layoutChanged ! Item was added: + ----- Method: ProportionalSplitterMorph>>stopStepping (in category 'events') ----- + stopStepping + super stopStepping. + (self class smartVerticalSplitters or: [ self class smartHorizontalSplitters ]) ifFalse: [ self balanceOffsets ]! Item was changed: ----- Method: SimpleHaloMorph>>transferHalo: (in category 'pop up') ----- transferHalo: event + "Transfer the halo to the next likely recipient" - "Transfer the halo to the next likely recipient. Switch between siblings if overlapping." - (self target world morphsAt: event position) allButFirst "... the halo itself" - detect: [:morph | - "Sibling found?" - (morph owner == self target owner - and: [morph ~~ self target]) - ifTrue: [ - ^ morph invokeHaloOrMove: event]. - "No sibling possible anymore?" - morph == self target]. - self target transferHalo: (event transformedBy: (self target transformedFrom: self)) from: self target.! Item was changed: ----- Method: SystemWindow>>filterEvent:for: (in category 'events') ----- filterEvent: aKeyboardEvent for: anObject "Provide keyboard shortcuts." + - aKeyboardEvent isKeystroke ifFalse: [^ aKeyboardEvent]. + + aKeyboardEvent hand halo ifNotNil: [ : halo | halo target isSystemWindow ifTrue: [ aKeyboardEvent hand removeHalo ] ]. aKeyboardEvent commandKeyPressed ifTrue: [ + aKeyboardEvent keyCharacter caseOf: { - aKeyboardEvent keyCharacter caseOf: { [$\] -> [self class sendTopWindowToBack]. [Character escape] -> [self class deleteTopWindow]. [$/] -> [self class bringWindowUnderHandToFront]. } otherwise: [^ aKeyboardEvent "no hit"]. ^ aKeyboardEvent ignore "hit!!"]. aKeyboardEvent controlKeyPressed ifTrue: [ aKeyboardEvent keyCharacter caseOf: { [Character escape] -> [self world findWindow: aKeyboardEvent]. } otherwise: [^ aKeyboardEvent "no hit"]. ^ aKeyboardEvent ignore "hit!!"]. ^ aKeyboardEvent "no hit"!

3 6

Project thumbnail morph loss in 5.2?
by Tim Johnson 08 Dec '18

08 Dec '18

Hi, I wanted to experiment with sharing a Project across images. I did World menu->open->morphic project. (I can enter it and return, or not enter it at all, and it has no effect on what I'm reporting here.) I then menu-clicked inside the project's thumbnail and chose "see if server version is more recent." This brings up a chooser dialog and makes the project thumbnail morph disappear. Desired behavior: Seems like when the chooser dialog appears, the thumbnail morph shouldn't disappear. (The project still appears in the Projects menu on the docking bar.) I don't know if there's any way to get the thumbnail back... I can file a bug in Mantis if desired. Thanks, Tim J

6 10

OrderedSet
by Bernhard Pieber 04 Dec '18

04 Dec '18

Hi Squeakers, Does anyone know of an OrderedSet implementation for Squeak somewhere? Bernhard

9 14

Named serial port bug and how to contribute
by Ricardo Moran 02 Dec '18

02 Dec '18

Hi, I think I found a bug with the named serial ports in windows. I can open a named port without problem the first time but if I close it and try to open it again it fails until I restart the image. This doesn't happen with numeric ports. I could reproduce this bug in both Squeak5.2 and Pharo6.1 (both using windows 8.1). I think the problem is that SerialPort>>#primClosePortByName: calls the "primitiveSerialPortCloseByName" but I couldn't find this primitive anywhere in the SerialPlugin source ( https://github.com/OpenSmalltalk/opensmalltalk-vm/blob/Cog/src/plugins/Seri…). So I'm guessing that the port is never properly closed. I added the missing primitive (find attached) and recompiled the vm, and now it seems to work. So my question is: how can I contribute this small fix? Should I commit to the VM Maker Inbox repository? Thanks, Richo

3 3

The Trunk: ShoutCore-eem.62.mcz
by commits＠source.squeak.org 01 Dec '18

01 Dec '18

Eliot Miranda uploaded a new version of ShoutCore to project The Trunk: http://source.squeak.org/trunk/ShoutCore-eem.62.mcz ==================== Summary ==================== Name: ShoutCore-eem.62 Author: eem Time: 29 November 2018, 2:00:34.775887 pm UUID: 45629ac9-52a5-412c-9bf1-493a0ac2ea36 Ancestors: ShoutCore-tpr.61 Fix a bug in SHMCClassDefinition>>sharedPools that answered nil as the pool for a missing pool, and hence broke khighlighting in the Monticello package browser when browsing older version s with different pool dictionary definitions. =============== Diff against ShoutCore-tpr.61 =============== Item was changed: ----- Method: SHMCClassDefinition>>sharedPools (in category 'act like a class') ----- sharedPools | d | d := Set new. + classDefinition poolDictionaries do: + [:poolName| + (Smalltalk at: poolName asSymbol ifAbsent:[nil]) ifNotNil: [:pool| d add: pool]]. - classDefinition poolDictionaries do:[:each | - d add: (Smalltalk at: each asSymbol ifAbsent:[nil]) ]. ^d!

4 3

sends of #class and #== considered harmful, may we please stop?
by Chris Muller 01 Dec '18

01 Dec '18

Hi guys, Something I've been wanting to ask the community for years, but never had the gumption, was about changing how we write our #= and #hash methods, but now that we're combing through them anyway, I must! Basically, it comes down to Proxy's. I want them to work better in Squeak. But every time we put another send to #class or #== in an implementation of #=, since those methods are inlined, they're too brittle to work with Proxy's. This causes hard-to-trackdown bugs in applications and mini-messes to deal with them since the app is forced to change its code to become "aware" of potential Proxy'iness of an object before comparing it with #=. This is no surprise, since writing a send to #== instead of #= for no more than "performance" is actually breaking encapsulation in the first place... There is an easy solution. When writing #= and #hash implementations, use (#species or #xxxClass or #isKindOf: instead of #class) and #= instead #==. So, for example, Eliot, I want to upgrade your new Message>>#= to something like: = anObject ^self xxxClass = anObject xxxClass and: [selector = anObject selector "selector lookup is by identity" and: [lookupClass = anObject lookupClass and: [args literalEqual: anObject arguments]]] Or #species or #isKindOf:, like we do in many other methods. Now the method is Proxy-compatible. What do you think?

6 13

[ANN] Live type information for Smalltalk
by Hernan Wilkinson 28 Nov '18

28 Nov '18

Hi all!, I wanted to let you know that I have implemented a "feature" to have type information "live". Currently it is implemented on Cuis but the idea is to port it to Pharo and Squeak as well. This feature is based on a few changes on the VM that collects the class of the object assigned to a variable, a parameter or when returning from a method. That information is available from the image and allows to improve the tools radically, like the autocompletion on the browser, looking for senders/implementors and even "scoping" the refactorings based on the type info. I presented it at the last Smalltalks (the argentinean smalltalk conference) and I'm happy to announce its alpha version (1/1000) asFloat :-) Here is a link to the repo that has the VM and a Cuis image to play with it: https://github.com/hernanwilkinson/Cuis-Smalltalk-DynamicTypeInformation It also has the source of the VM's changes if you want to see how it is implemented. Currently it is implemented only in the stack VM and the VMs I compiled and uploaded to the repo are 64 bits for MacOs and Linux. (I don't have a Windows machine to compile it :-) ) Here is a link to a video that shows how it is implemented and how to use it: https://t.co/n8FQuo7pKZ If you want to see the slides of the presentation, here is the link: https://www.slideshare.net/hernanwilkinson/dynamic-type-information I would love for you to try it and get feedback, ideas, etc. I really believe it is a "killer feature", something that could take Smalltalk one step further and show the way to other dynamically typed languages. Cheers! Hernan -- *Hernán WilkinsonAgile Software Development, Teaching & Coaching* *Phone: +54 - 011 - *6091 - 3125 *Mobile: +54 - 911 - 4470 - 7207email: hernan.wilkinson(a)10Pines.comsite: http://www.10Pines.com <http://www.10pines.com/>* Address: Alem 693, Floor 5 B, Buenos Aires, Argentina

4 4

[ANN] Live type info for Squeak!
by Hernan Wilkinson 27 Nov '18

27 Nov '18

Hi, I uploaded a Squeak5.2 image with live type info to https://github.com/hernanwilkinson/Cuis-Smalltalk-DynamicTypeInformation It is in the "image" directory. The VM is on the "VMs" directory (you can use it with a regular VM but it will not store type info). There is a workspace with some collaborations to try like looking the "typed senders". Also, the browser has an option to see the "typed source" in the "source" button. Sadly I did not have the time to implement typed sender/implementors when browsing because a change in the encoder is needed. I neither had the type to modify the rename method, but I think it is a nice start to see how it works :-) Cheers! Hernan. -- *Hernán WilkinsonAgile Software Development, Teaching & Coaching* *Phone: +54 - 011 - *6091 - 3125 *Mobile: +54 - 911 - 4470 - 7207email: hernan.wilkinson(a)10Pines.comsite: http://www.10Pines.com <http://www.10pines.com/>* Address: Alem 693, Floor 5 B, Buenos Aires, Argentina

2 1

← Newer
1
2
3
4
5
6
7
8
9
10
Older →