Squeak-dev August 2019

squeak-dev@lists.squeakfoundation.org

35 participants
288 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

The Inbox: Graphics-ct.410.mcz
by commits＠source.squeak.org 23 Sep '19

23 Sep '19

A new version of Graphics was added to project The Inbox: http://source.squeak.org/inbox/Graphics-ct.410.mcz ==================== Summary ==================== Name: Graphics-ct.410 Author: ct Time: 14 August 2019, 11:52:24.4754 pm UUID: 008f8bb1-9963-1e46-8b29-1f1e2aa3235b Ancestors: Graphics-nice.409 Little refactoring to HSV conversion Afaik #caseOf: is compiler optimized, so this is not only more beautiful but also at least at efficient as the old conditional chain, isn't it? =============== Diff against Graphics-nice.409 =============== Item was changed: ----- Method: Color>>setHue:saturation:brightness: (in category 'private') ----- setHue: hue saturation: saturation brightness: brightness "Initialize this color to the given hue, saturation, and brightness. See the comment in the instance creation method for details." | s v hf i f p q t | s := (saturation asFloat max: 0.0) min: 1.0. v := (brightness asFloat max: 0.0) min: 1.0. "zero saturation yields gray with the given brightness" s = 0.0 ifTrue: [ ^ self setRed: v green: v blue: v ]. hf := hue asFloat. (hf < 0.0 or: [hf >= 360.0]) ifTrue: [hf := hf \\ 360]. hf := hf / 60.0. i := hf asInteger. "integer part of hue" f := hf fractionPart. "fractional part of hue" p := (1.0 - s) * v. q := (1.0 - (s * f)) * v. t := (1.0 - (s * (1.0 - f))) * v. + ^ i caseOf: { + [0] -> [ self setRed: v green: t blue: p ]. + [1] -> [ self setRed: q green: v blue: p ]. + [2] -> [ self setRed: p green: v blue: t ]. + [3] -> [ self setRed: p green: q blue: v ]. + [4] -> [ self setRed: t green: p blue: v ]. + [5] -> [ self setRed: v green: p blue: q ]. }! - 0 = i ifTrue: [ ^ self setRed: v green: t blue: p ]. - 1 = i ifTrue: [ ^ self setRed: q green: v blue: p ]. - 2 = i ifTrue: [ ^ self setRed: p green: v blue: t ]. - 3 = i ifTrue: [ ^ self setRed: p green: q blue: v ]. - 4 = i ifTrue: [ ^ self setRed: t green: p blue: v ]. - 5 = i ifTrue: [ ^ self setRed: v green: p blue: q ]. - - self error: 'implementation error'. - !

6 6

Plans for 5.3 Morphic / Readability (Re: Squeak Oversight Board Election 2019)
by H. Hirzel 17 Sep '19

17 Sep '19

Hello Marcel a question, see below. Regards Hannes On 1/11/19, Marcel Taeumel <marcel.taeumel(a)hpi.de> wrote: > Hi, there. > > I would like to continue serving on the Squeak Oversight Board in 2019, too. > My main interests encompass user interaction (Morphic), > code modularity > (readability/extensibility), What are your suggestions / plans for 5.3 for enhancing the Morphic user interaction expericence? How do you think better readability of code could be achived? > backwards compatibility (Etoys), and > programming/productivity tools. > > During the past years, it has been a healthy mix of promising ideas, > enlightening discussions, and challenging viewpoints. I do value such > programming experiences as a great addition to my daily life. Squeak's > community is great! Thank you. :-) > > Let us continue shaping Squeak for many opportunities to come! It is an > intriguing programming system with its liveness promoting exploratory > working habits. What a time to be a "programmer". ;-) > > Best, > Marcel > Am 10.01.2019 20:38:30 schrieb Ron Teitelbaum <ron(a)usmedrec.com>: > Hi All, > > Thank you Craig for jumping right in! I would encourage everyone else to > follow Craig's lead. I know some people like to wait but the more > candidates the better. > > The lack of candidates shouldn't be an incentive for people to run and a > large number of candidates shouldn't be a reason for you to decide against > volunteering to run! We have had some excellent people run and not be > elected that would have been terrific members of the board! You show your > commitment to Squeak just by running! For those of you that say you don't > have the time or are not sure what you would have to do, we have had some > excellent answers in the past. There is no requirement or specific demands > and the board decides when and how it will meet. Most, if not all, meetings > are electronic so the demands on your time are very low! Plus the members > are all terrific people you will enjoy working with! > > Jump on in and add your name to the candidate list! > > Thanks! > > Ron Teitelbaum > > On Thu, Jan 10, 2019 at 2:49 AM Craig Latta <craig(a)blackpagedigital.com > [mailto:craig@blackpagedigital.com]> wrote: > > > Hi all-- > > I'd like to serve on the Squeak Oversight Board in 2019. With you > and my fellow board members, I intend to continue... > > - ...representing Squeak at conferences, including ESUG and the > International Conference on Live Coding[1]. > > - ...maintaining the agenda and notes at the board meetings. I > anticipate having the time to make posts to the board blog this > term. > > - ...developing symbiosis between the Squeak, JavaScript, and Web > platform communities through my work with Caffeine[2] and > SqueakJS[3]. > > I would be honored to have your vote. > > > cheers! > > Craig > > -- > > [1] http://iclc.livecodenetwork.org/2019/ingles.html > [http://iclc.livecodenetwork.org/2019/ingles.html] > [2] https://caffeine.js.org [https://caffeine.js.org] > [3] https://squeak.js.org [https://squeak.js.org] > > -- > Craig Latta > Black Page Digital > Amsterdam :: San Francisco > craig(a)blackpagedigital.com [mailto:craig@blackpagedigital.com] > +31 6 2757 7177 (SMS ok) > + 1 415 287 3547 (no SMS) > > >

2 3

ProgressBar bounces from screen centre to cursor pos, dialogues get hidden at the centre
by tim Rowledge 16 Sep '19

16 Sep '19

I've mentioned this in passing before. When loading a big package (Seaside in this case) it's very obvious that something is a bit inconsistent wrt the progress bar. Most of the time it sits at the centre of the screen but every now and then it leaps (is it a bird? is it a plane? No, it's superbouncyprogressbarman!) to the cursor position - and then back. All a bit travel-sickness inducing really. Clearly there's no very good reason for a UI element to do this dance of flickeriness. As best I can tell the bouncing is caused by a conflict between a plain String>displayProgress... where the position is nil and thus defaults to screen centre, and MetacelloSqueakPlatform>>#do:displaying: which specifies Sensor cursorPoint. There are reasonable arguments for both choices (screen centre is consistent, cursor location is 'where user is looking') but mixing them is not so desirable. There's also a practical issue with using the cursor location in that there is a pretty good chance that whilst loading is happening the user may be looking at other host windows and the cursor location value will the leaping around. I'd suggest that it might be nicer to make the progress bar stay where it was initially opened, even when new levels are added as packages load packages and so on. Does anyone have strong opinions on this? Something related in the UI sense is that we can get nasty confluences of several dialogues/notifiers/etc apppearing in a stack at the screen centre. This came up recently when trying to explain to a non-user (and one that really, really, wanted to find reasons to disparage Smalltalk) how to load some code and it appears that an rather important notifier had got covered by a progressbar - they seem to insist on being at the very front.; result, one very annoyed user and some tricky remote-by-email-debugging. I'm not entirely sure what a good answer would be for this. One relatively simple option would be to default the position of some of the popup windows a bit differently so they don't obscure each other so completely. I suppose one option would be to make popup dialogues add to the progressbar if it is open - in the same way that we get a stack of progress bars, add the inner morphs of a dialogue? I really don't know a solid answer for this, but it certainly isn't a good thing to have a 'do you want to do this important thing or not?' dialogue hidden behind a progress bar that is waiting for the user to answer! tim -- tim Rowledge; tim(a)rowledge.org; http://www.rowledge.org/tim Useful Latin Phrases:- Quo signo nata es? = What's your sign?

6 9

The Inbox: Collections-nice.825.mcz
by commits＠source.squeak.org 16 Sep '19

16 Sep '19

Nicolas Cellier uploaded a new version of Collections to project The Inbox: http://source.squeak.org/inbox/Collections-nice.825.mcz ==================== Summary ==================== Name: Collections-nice.825 Author: nice Time: 7 April 2019, 11:10:11.564693 pm UUID: c9836c60-9acc-45eb-83b2-851f147b70ca Ancestors: Collections-nice.824 Try to improve some Interval of Float Behavior. - change #last so that it agrees with (self at: self size) Note that last behavior slightly changes for empty Interval: (3 to: 0) last is now 2 instead of 0 previously. I don't know of any code impacted, but it's not a completely neutral change Note: this solves 2 out of the 3 failures of #testIntervalOfFloatLast... The last element might still overshoot the prescribed upper bound. I can also fix it, but implementation of #last is getting overkill. - implement a ReversedInterval that lazily reverse an Interval of Float, and resolves #testIntervalOfFloatReversed. The alternative would be resorting to an Array... - renounce fuzzy inclusion test. As exhibited by #testSurprisingFuzzyInclusionBehavior, it creates more weirdness than it tries to solve. There will be two more test failing, asserting this fuzzy inclusion, but we can later change them. =============== Diff against Collections-nice.824 =============== Item was changed: ----- Method: Interval>>indexOf:startingAt: (in category 'accessing') ----- indexOf: anElement startingAt: startIndex "startIndex is an positive integer, the collection index where the search is started." - "during the computation of val , floats are only used when the receiver contains floats" + | index | - | index val | (self rangeIncludes: anElement) ifFalse: [ ^0 ]. + index := (anElement - start / step) rounded + 1. - val := anElement - self first / self increment. - val isFloat - ifTrue: [ - (val - val rounded) abs * 100000000 < 1 ifFalse: [ ^0 ]. - index := val rounded + 1 ] - ifFalse: [ - val isInteger ifFalse: [ ^0 ]. - index := val + 1 ]. - "finally, the value of startIndex comes into play:" (index between: startIndex and: self size) ifFalse: [ ^0 ]. + (self at: index) = anElement ifFalse: [ ^0 ]. ^index! Item was changed: ----- Method: Interval>>last (in category 'accessing') ----- last "Refer to the comment in SequenceableCollection|last." + ^start + (step * (self size - 1))! - ^stop - (stop - start \\ step)! Item was changed: ----- Method: Interval>>reversed (in category 'converting') ----- reversed self isEmpty ifTrue: [^stop to: start by: step negated]. + (start isFraction and: [step isFraction]) ifFalse: [^ReversedInterval newFrom: self]. ^self last to: start by: step negated! Item was added: + SequenceableCollection subclass: #ReversedInterval + instanceVariableNames: 'original' + classVariableNames: '' + poolDictionaries: '' + category: 'Collections-Sequenceable'! + + !ReversedInterval commentStamp: 'nice 4/6/2019 16:30' prior: 0! + A ReversedInterval is an Interval reversed. + An Interval reversed is generally an Interval, except when made of Float. + This is because Float arithmetic does not guarantee reversible operations. + + Instead of transforming the reversed Float Interval into an Array, I keep a lazier and cheaper representation. + I can even pretend being an Interval myself. As long as I quake like a duck... + + Instance Variables + original: <Interval> + + original + - the Interval being reversed + ! Item was added: + ----- Method: ReversedInterval class>>newFrom: (in category 'as yet unclassified') ----- + newFrom: aCollection + "Answer an instance of me containing the same elements as aCollection, reversed." + + aCollection isInterval ifTrue: [^self basicNew setOriginal: aCollection]. + ^(Interval newFrom: aCollection) reversed! Item was added: + ----- Method: ReversedInterval>>+ (in category 'arithmetic') ----- + + aNumber + ^(original + aNumber) reversed! Item was added: + ----- Method: ReversedInterval>>- (in category 'arithmetic') ----- + - aNumber + ^(original - aNumber) reversed! Item was added: + ----- Method: ReversedInterval>>= (in category 'comparing') ----- + = anObject + ^ self == anObject + or: [anObject class = self class + and: [original = anObject reversed]]! Item was added: + ----- Method: ReversedInterval>>at: (in category 'accessing') ----- + at: anInteger + ^original at: original size + 1 - anInteger! Item was added: + ----- Method: ReversedInterval>>copyFrom:to: (in category 'copying') ----- + copyFrom: startIndex to: stopIndex + (startIndex = 1 and: [stopIndex = self size]) ifTrue: [^self]. + stopIndex < startIndex ifTrue: [^self copyEmpty]. + ^((self at: stopIndex) to: (self at: startIndex) by: original increment) reversed! Item was added: + ----- Method: ReversedInterval>>do: (in category 'enumerating') ----- + do: aBlock + original reverseDo: aBlock! Item was added: + ----- Method: ReversedInterval>>hash (in category 'comparing') ----- + hash + ^original hash bitXor: SmallInteger maxVal! Item was added: + ----- Method: ReversedInterval>>isEmpty (in category 'testing') ----- + isEmpty + ^self size = 0! Item was added: + ----- Method: ReversedInterval>>printOn: (in category 'printing') ----- + printOn: aStream + aStream + nextPut: $(; + print: original; + nextPut: $); + space; + nextPutAll: #reversed! Item was added: + ----- Method: ReversedInterval>>reverseDo: (in category 'enumerating') ----- + reverseDo: aBlock + original do: aBlock! Item was added: + ----- Method: ReversedInterval>>reversed (in category 'converting') ----- + reversed + ^original! Item was added: + ----- Method: ReversedInterval>>setOriginal: (in category 'private') ----- + setOriginal: anInterval + original := anInterval! Item was added: + ----- Method: ReversedInterval>>size (in category 'accessing') ----- + size + ^original size! Item was added: + ----- Method: ReversedInterval>>sorted (in category 'sorting') ----- + sorted + original increment > 0 ifTrue: [^original]. + ^self! Item was added: + ----- Method: ReversedInterval>>species (in category 'private') ----- + species + + ^original species! Item was added: + ----- Method: ReversedInterval>>storeOn: (in category 'printing') ----- + storeOn: aStream + aStream + nextPut: $(; + store: original; + nextPut: $); + space; + nextPutAll: #reversed! Item was added: + ----- Method: ReversedInterval>>sum (in category 'accessing') ----- + sum + ^original sum!

6 8

[BUG] 5.3 alpha 32 bit for macOS fails in SnowLeopard
by Edgar J. De Cleene 10 Sep '19

10 Sep '19

I decide move all my stuff to 5.3 as see a lot is changing. And revive a 2010 mac mini with Snow Leopard http://ftp.squeak.org/5.3alpha/Squeak5.3alpha-18222-32bit/Squeak5.3alpha-182 22-32bit-201810071412-macOS.dmg This links do not work But http://ftp.squeak.org/5.0/Squeak-5.0-All-in-One.zip works and moving the relevant part to Applications , this VM launchs Squeak5.3alpha-18222-32bit .image No idea how many users of old Apple stuff, but is good have next Squeak running running on it. Edgar @morplenauta

5 9

Installing various versions of squeaksource is problematic
by tim Rowledge 09 Sep '19

09 Sep '19

A while back I was asking about internal use monticello repositories and one suggestion I didn't have time to do much with was to install squeaksource; three options were pointed out squeaksource squeaksource3 personalsqueaksource. I thought I'd give it a quick try to see what happens. All in all it wasn't a particularly successful attempt. A) squeaksource3 I know Tobias officially abandoned this at the beginning of the year but since there is an entry in squeakmap for it I thought "let's try it". Nothing happened. There was clearly an attempt to find a suitable version for the image I am using (a 5.2-18229 release, and an updated version) but nothing further. I noticed Tobias had pointed out using `Metacello new configuration: 'SqueakSource3'; load: #('All').` ... so I tried that; it failed because of not finding a suitable version. I dug a little but didn't see anything obvious. I also looked in the MetacelloRepository for a ConfigurationOf... but it appears to have been removed. B) squeaksource There's also a SqueakMap entry for the old system. It didn't load either. Complaints about no #stable version etc I did see the ConfigurationOfSqueakSource and tried that but again, no stable version, no way I spotted to use any other named version that looked plausible. I suppose one shouldn't be too surprised since it is pretty old but it *is* the system we have as a main repository... C) personalsqueaksource This does seem to load from SqueakMap *but* not trivially. There is a syntax error in the Seaside2 method WAWikiRenderer>>#swikifyPiece: where the original was writing into a block arg (aLine, replaced with thisLine etc to make it complete). There is also an alarmingly large number of Undeclareds left over at the end of the load. So, not a terribly happy test. tim -- tim Rowledge; tim(a)rowledge.org; http://www.rowledge.org/tim Strange OpCodes: SDR: Shift Disk Right

4 12

The Trunk: Morphic-mt.1510.mcz
by commits＠source.squeak.org 03 Sep '19

03 Sep '19

Marcel Taeumel uploaded a new version of Morphic to project The Trunk: http://source.squeak.org/trunk/Morphic-mt.1510.mcz ==================== Summary ==================== Name: Morphic-mt.1510 Author: mt Time: 26 August 2019, 3:57:09.414394 pm UUID: 1a066b5c-53d2-164a-980b-c331d3616e78 Ancestors: Morphic-dtl.1506, Morphic-kfr.1509 Merges Morphic-dtl.1506 and fixes some docking-bar-layout issues. :-) Note that I changed the wording to "Git Browser (click to install)" as well as the text in the confirmation dialog. Please review. If not automatically done, please update your world-dock instances via TheWorldMainDockingBar updateInstances. =============== Diff against Morphic-kfr.1509 =============== Item was removed: - ----- Method: DockingBarItemMorph>>updateLayoutInDockingBar (in category 'private') ----- - updateLayoutInDockingBar - - owner isVertical - ifTrue: [ - self hResizing: #spaceFill. - self vResizing: #shrinkWrap ] - ifFalse: [ - self hResizing: #shrinkWrap. - self vResizing: #spaceFill ].! Item was added: + ----- Method: DockingBarMorph>>addedMorph: (in category 'change reporting') ----- + addedMorph: aMorph + "Update the new morph's layout properties according to my orientation." + + aMorph isMenuItemMorph ifFalse: [^ self]. + + self isVertical + ifTrue: [ + aMorph hResizing: #spaceFill. + aMorph vResizing: #shrinkWrap ] + ifFalse: [ + aMorph hResizing: #shrinkWrap. + aMorph vResizing: #spaceFill ].! Item was removed: - ----- Method: MenuItemMorph>>ownerChanged (in category 'change reporting') ----- - ownerChanged - "The receiver's owner, some kind of a pasteup, has changed its - layout." - super ownerChanged. - self updateLayoutInDockingBar! Item was changed: ----- Method: TheWorldMainDockingBar>>gitInfrastructureMenuItemOn: (in category 'submenu - tools') ----- + gitInfrastructureMenuItemOn: menu + + menu addItem: [:item | + item + contents: ((Smalltalk hasClassNamed: #SquitBrowser) + ifTrue: ['Git Browser' translated] + ifFalse: ['Git Browser (click to install)' translated]); + help: 'Browse repositories on github.com' translated; + icon: (self colorIcon: ((Smalltalk classNamed: #SquitBrowser) ifNil: Model) + basicNew windowColorToUse); + target: self; + selector: #installAndOpenGitBrowser]! - gitInfrastructureMenuItemOn: menu - menu addItem: - [ : item | item - contents: 'Git Browser' translated ; - help: 'Browse repositories on github.com' translated ; - icon: (self colorIcon: Color lightGray) ; - target: Installer ; - selector: #installAndOpenGitBrowser ]! Item was added: + ----- Method: TheWorldMainDockingBar>>installAndOpenGitBrowser (in category 'menu actions') ----- + installAndOpenGitBrowser + + (Smalltalk hasClassNamed: #SquitBrowser) + ifFalse: [ (UIManager default + confirm: 'The Git infrastructure and browser is not yet installed.\\Do you want to install the Git Browser?\\(Note that this step requires an internet connection and\may take several minutes.)' withCRs translated + title: 'Confirm Installation Request' ) + ifTrue: [ Installer + ensureRecentMetacello; + installGitInfrastructure. + TheWorldMainDockingBar updateInstances ] + ifFalse: [ ^self ] ]. + (Smalltalk classNamed: #SquitBrowser) open. + ! Item was removed: - ----- Method: TheWorldMainDockingBar>>metacelloMenuItemOn: (in category 'submenu - tools') ----- - metacelloMenuItemOn: menu - menu addItem: - [ : item | item - contents: 'Metacello' translated ; - help: 'Install Metacello' translated ; - icon: (self colorIcon: Color lightGray) ; - target: Installer ; - selector: #ensureRecentMetacello ]! Item was changed: ----- Method: TheWorldMainDockingBar>>toolsMenuOn: (in category 'construction') ----- toolsMenuOn: aDockingBar aDockingBar addItem: [ :item | item contents: 'Tools' translated; addSubMenu: [ :menu | self browserMenuItemOn: menu; workspaceMenuItemOn: menu; transcriptMenuItemOn: menu; testRunnerMenuItemOn: menu; methodFinderMenuItemOn: menu; messageNamesMenuItemOn: menu. menu addLine. self simpleChangeSorterMenuItemOn: menu; dualChangeSorterMenuItemOn: menu; monticelloBrowserMenuItemOn: menu; monticelloConfigurationsMenuItemOn: menu ; - metacelloMenuItemOn: menu ; gitInfrastructureMenuItemOn: menu. menu addLine. self processBrowserMenuItemOn: menu; preferenceBrowserMenuItemOn: menu; fileListMenuItemOn: menu. ] ]!

2 1

The Inbox: Installer-Core-tpr.434.mcz
by commits＠source.squeak.org 02 Sep '19

02 Sep '19

tim Rowledge uploaded a new version of Installer-Core to project The Inbox: http://source.squeak.org/inbox/Installer-Core-tpr.434.mcz ==================== Summary ==================== Name: Installer-Core-tpr.434 Author: tpr Time: 27 August 2019, 1:08:05.39831 pm UUID: c1ea7dcd-9b59-4796-b66a-0f558e074ec9 Ancestors: Installer-Core-mt.433 Some possible cleanups for Installer. Make InstallerWeb behave more like the other installer classes. Remove a couple of redundant installer types. Start adding some up to date copmments and make the Installer class comment appear in a HelpBrowser - more needed. Subclasses should get more extensive class specific commentary and that should be made to appear in the help browser pages. =============== Diff against Installer-Core-mt.433 =============== Item was changed: Object subclass: #Installer instanceVariableNames: 'answers packages messagesToSuppress useFileIn noiseLevel' classVariableNames: 'InstallerBindings IsSetToTrapErrors Repositories SkipLoadingTests ValidationBlock' poolDictionaries: '' category: 'Installer-Core'! Installer class instanceVariableNames: 'localRepository'! + !Installer commentStamp: 'tpr 8/25/2019 18:37' prior: 0! + Installer is a mechanism for listing, examining and installing software from a variety of sources. It is intended to help in building install scripts for package setups etc. + + Currently it can use + - files; a file specified by a filename. Relative filenames will be treated as based form the current default directory. See InstallerFile. + - urls; a file specified by a url or a script embedded in a webpage. See InstallerWeb. + - squeakmap; see InstallerSqueakMap + - monticello; by ftp or http access, or a local directory, or a Magma or GOODS database. There is a list of shortcuts to popular repositories in this class protocol 'repositories'. See InstallerMonticello. + - mantis; accessing code attached to a mantis bug report. See InstallerMantis. + + Simple example usages - + Installer squeakmap install: 'DynamicBindings'. + Installer squeakmap search: '*scratch'. + Installer squeakmap search: 'author:*rowledge'. + Installer squeaksource project: 'ss2'; install: 'TinyWiki'. + Installer ss project: 'Installer'; browse: 'Installer-Core'. + - !Installer commentStamp: 'kph 3/30/2009 01:29' prior: 0! - Documentation now available at http://installer.pbwiki.com/Installer useFileIn - flag to load source.st rather than using Monticello! Installer class instanceVariableNames: 'localRepository'! Item was added: + ----- Method: Installer class>>asHelpTopic (in category 'documentation') ----- + asHelpTopic + ^HelpTopic + title: 'Installer' + contents: self class comment! Item was removed: - ----- Method: Installer class>>cobalt (in category 'repositories') ----- - cobalt - - ^ self monticello http: 'http://croquet-src-01.oit.duke.edu:8886'! Item was removed: - ----- Method: Installer class>>goran (in category 'repositories') ----- - goran - - ^ self monticello http: 'squeak.krampe.se'; project: ''! Item was removed: - ----- Method: Installer class>>impara (in category 'repositories') ----- - impara - - ^ self monticello http: 'source.impara.de'! Item was removed: - ----- Method: Installer class>>install: (in category 'action report') ----- - install: scriptName - - ^ (self scripts install: scriptName) ifNil:[ self web install: scriptName ] - ! Item was removed: - ----- Method: Installer class>>keith (in category 'repositories') ----- - keith - - ^ self monticello ftp: 'squeak.warwick.st' directory: 'mc' user: 'squeak' password: 'viewpoints'! Item was changed: ----- Method: Installer class>>remoteRepositories (in category 'repository-overrides') ----- remoteRepositories + ^ #(#ss #ss3 #gemsource #gs #krestianstvo #lukas #squeak #squeakfoundation #squeaksource #squeaksource3 #ss #ss3 #swa #swasource #wiresong )! - ^ #(#ss #ss3 #cobalt #gemsource #goran #gs #impara #keith #krestianstvo #lukas #saltypickle #sophie #squeak #squeakfoundation #squeaksource #squeaksource3 #ss #ss3 #swa #swasource #wiresong )! Item was removed: - ----- Method: Installer class>>sake (in category 'sake') ----- - sake - - ^ self sake: InstallerSake sake! Item was removed: - ----- Method: Installer class>>sake: (in category 'sake') ----- - sake: aSakePackagesClass - - ^ InstallerSake new sake: aSakePackagesClass! Item was removed: - ----- Method: Installer class>>saltypickle (in category 'repositories') ----- - saltypickle - - ^ self monticello http: 'squeak.saltypickle.com'! Item was removed: - ----- Method: Installer class>>setSakeToUse: (in category 'sake') ----- - setSakeToUse: aClass - - InstallerSake sake: aClass! Item was removed: - ----- Method: Installer class>>sophie (in category 'repositories') ----- - sophie - - ^ self monticello http: 'source.sophieproject.org' - - ! Item was removed: - ----- Method: Installer class>>ssMirror (in category 'repositories') ----- - ssMirror - "The Chilean mirror for the original SqueakSource." - ^ self monticello http: 'http://dsal.cl/squeaksource/'! Item was changed: ----- Method: Installer class>>web (in category 'web') ----- web + ^ InstallerWeb new! - ^ InstallerWeb! Item was removed: - Installer subclass: #InstallerSake - instanceVariableNames: 'sake' - classVariableNames: 'Sake' - poolDictionaries: '' - category: 'Installer-Core'! Item was removed: - ----- Method: InstallerSake class>>classPackages (in category 'accessing system') ----- - classPackages - - ^Smalltalk at: #Packages ifAbsent: [ self error: 'Sake Packages code not present' ]! Item was removed: - ----- Method: InstallerSake class>>sake (in category 'accessing') ----- - sake - - ^ Sake ifNil: [ self classPackages current ]! Item was removed: - ----- Method: InstallerSake class>>sake: (in category 'accessing') ----- - sake: aClass - - Sake := aClass! Item was removed: - ----- Method: InstallerSake>>basicInstall (in category 'basic interface') ----- - basicInstall - - self withAnswersDo: [ (self packages collect: [ :packageName | sake named: packageName ]) asTask run ]. - ! Item was removed: - ----- Method: InstallerSake>>sake (in category 'websqueakmap') ----- - sake - - ^ sake ! Item was removed: - ----- Method: InstallerSake>>sake: (in category 'websqueakmap') ----- - sake: aSakePackagesClass - - sake := aSakePackagesClass! Item was removed: - Installer subclass: #InstallerUniverse - instanceVariableNames: 'universe' - classVariableNames: 'LastUniUpdate' - poolDictionaries: '' - category: 'Installer-Core'! Item was removed: - ----- Method: InstallerUniverse class>>classUGlobalInstaller (in category 'accessing system') ----- - classUGlobalInstaller - - ^Smalltalk at: #UGlobalInstaller ifAbsent: [ self error: 'Universes code not present' ]! Item was removed: - ----- Method: InstallerUniverse class>>classUUniverse (in category 'accessing system') ----- - classUUniverse - - ^Smalltalk at: #UUniverse ifAbsent: [ self error: 'Universes code not present' ]! Item was removed: - ----- Method: InstallerUniverse class>>default (in category 'instance creation') ----- - default - - ^ self universe: (self classUGlobalInstaller universe: self classUUniverse systemUniverse)! Item was removed: - ----- Method: InstallerUniverse class>>universe: (in category 'instance creation') ----- - universe: u - - ^ self new universe: u! Item was removed: - ----- Method: InstallerUniverse>>basicInstall (in category 'basic interface') ----- - basicInstall - self packages do: [ :packageName | - | potentials pkg pkgAndVersion version | - - pkgAndVersion := self packageAndVersionFrom: packageName. - pkg := pkgAndVersion first. - version := pkgAndVersion last. - - potentials := universe packageVersionsForPackage: pkg. - - pkg := version isEmpty - ifTrue: [ potentials last ] - ifFalse: [ - version := self classUVersion readFrom: version readStream. - (potentials anySatisfy: [ :p | p version = version]) ifFalse: [ ^ self error: 'version not found'] - ]. - universe planToInstallPackage: pkg. - ]. - self uniDoInstall! Item was removed: - ----- Method: InstallerUniverse>>classUVersion (in category 'class references') ----- - classUVersion - - ^Smalltalk at: #UVersion ifAbsent: [ self error: 'Universes code not present' ]! Item was removed: - ----- Method: InstallerUniverse>>uniDoInstall (in category 'universes') ----- - uniDoInstall - - self withAnswersDo: [ self universe doInstall ] ! Item was removed: - ----- Method: InstallerUniverse>>universe (in category 'universes') ----- - universe - - ^ universe! Item was removed: - ----- Method: InstallerUniverse>>universe: (in category 'universes') ----- - universe: u - - universe := u. - self update.! Item was removed: - ----- Method: InstallerUniverse>>update (in category 'public interface') ----- - update - - (LastUniUpdate isNil or:[ (DateAndTime now - LastUniUpdate) > 600 seconds ]) - ifTrue: [universe requestPackageList. - LastUniUpdate := DateAndTime now]! Item was removed: - ----- Method: InstallerWeb class>>install: (in category 'compatability') ----- - install: webPageName - "This keeps the syntax Installer web install: working" - ^ self new install: webPageName! Item was removed: - ----- Method: InstallerWeb class>>searchPath (in category 'accessing') ----- - searchPath - "a search path item, has the following format. prefix*suffix" - - ^ WebSearchPath ifNil: [ WebSearchPath := OrderedCollection new ].! Item was added: + ----- Method: InstallerWeb>>searchPath (in category 'web install') ----- + searchPath + "a search path item has the following format. prefix*suffix" + + ^ WebSearchPath ifNil: [ WebSearchPath := OrderedCollection new ].! Item was changed: ----- Method: InstallerWeb>>urlToDownload (in category 'web install') ----- urlToDownload "while we look for a url which returns what we are looking for, we get the data anyway" | delay | delay := 0. + self searchPath do: [ :pathSpec | - self class webSearchPath do: [ :pathSpec | | potentialUrl readPathSpec retry | readPathSpec := pathSpec value readStream. potentialUrl := (readPathSpec upTo: $*), self package, (readPathSpec upToEnd ifNil: [ '' ]). [retry := false. pageDataStream := self urlGet: potentialUrl. self wasPbwikiSpeedWarning ifTrue: [ retry := true. delay := delay + 5. self logCR: 'PBWiki speed warning. Retrying in ', delay printString, ' seconds'. (Delay forSeconds: delay) wait] ifFalse: [ self hasPage ifTrue: [ pageDataStream reset. ^ potentialUrl ] ]. retry ] whileTrue ]. ^nil !

9 18

out-of-date CompiledMethod
by Nicolas Cellier 31 Aug '19

31 Aug '19

Hi, Some CompiledMethod might be replaced in their respective MethodDictionary (recompiled), but persist in the system when they have some block referenced. We can track them down with this snippet: (CompiledMethod allInstances reject: #isInstalled) inspect. Most of the time, it's innocuous, but if the method were recompiled due to a change of class layout and if the block access inst. var. slots that since moved, all sort of undefined behavior can happen, including a VM crash. I thought I could track such case down with this snippet: (CompiledBlock allInstances reject: [:e | e isClean or: [each home method isInstalled]]) inspect. or: (BlockContext allInstances reject: [:e | e closure isClean or: [each home method isInstalled]]) inspect. But both BlockContext and CompiledBlock allInstances isEmpty. Ah yes, CompiledBlock is future for Sista, and now we only get Context in Spur... (Context allSubInstances reject: [:e | e methods isInstalled]) inspect. gives a longer list than the first snippet, because the out-of-date method might be used in many different context... However, I can't find those having dangerous (not clean) blocks: (Context allSubInstances reject: [:e | e closure isNil]) inspect. (Context allSubInstances reject: [:e | e closure isNil or: [e closure isClean]]) inspect. (Context allSubInstances reject: [:e | e closure isNil or: [e closure isClean or: [e method isInstalled]]]) inspect. The first two lists are not empty, but the last one is empty. So that's still not the right invocation: there is no active Context for those closure, so here is a better snippet: (BlockClosure allSubInstances reject: [:e | e outerContext method isInstalled or: [e isClean]]) inspect. In a VMMaker image built from latest squeak distribution, the out-of-date CompiledMethods given by the first snippet is below: (WorldMenuProvider>>#preferencesBrowser "a CompiledMethod(3800333)") . (WorldMenuProvider>>#nextWindow "a CompiledMethod(1345923)") . (WorldMenuProvider>>#closeTopWindow "a CompiledMethod(1052691)") . (WorldMenuProvider>>#helpOnServices "a CompiledMethod(1013869)") . (WorldMenuProvider>>#rebuildRegistry "a CompiledMethod(2839737)") . (WorldMenuProvider>>#servicesBrowser "a CompiledMethod(485047)") . (WorldMenuProvider>>#createNewService "a CompiledMethod(2658559)") . (WorldState>>#stepListSortBlock "a CompiledMethod(491105)") . (ServiceAction class>>#id:text:button:description:action: "a CompiledMethod(3812069)") . (ServiceAction class>>#text:button:description:action: "a CompiledMethod(1460587)") . (ServiceCategory class>>#text:button:description: "a CompiledMethod(3025003)") . (ServiceShortcuts class>>#insertPrefShortcut: "a CompiledMethod(2077141)") . (WeakRegistry>>#installFinalizer "a CompiledMethod(1322897)") . (PluggableDictionary class>>#integerDictionary "a CompiledMethod(353359)") . (Slider>>#computeSlider "a CompiledMethod(682213)") . (MorphicAlarmQueue>>#migrate "a CompiledMethod(1994485)") . (SqueakReleaseNotes class>>#asHelpTopic "a CompiledMethod(79489)") . (SqueakTheme class>>#addToolColors: "a CompiledMethod(1225541)") . + a bunch of DoIt's for which I can't find the pointers (referenced by nothing, can it be a memory leak from VM?). By using 'chase pointers' menu, or the second snippet to catch Context, most come from ServiceRegistry. There are also special sortBlock for Heap (stepListSortBlock from WorldState>>initialize and also WorldState alarms) or hashBlock (integerDictionary from Unicode ToUpper). The HostWindowProxy registry has the obsolete installFinalizer. And the unclean closure given by latest snippet are very few (copy omitted): [closure] in WorldMenuProvider>>helpOnServices [closure] in ServiceShortcuts class>>insertPrefShortcut: [closure] in WeakRegistry>>installFinalizer [closure] in MorphicAlarmQueue>>migrate Most are not clean because they close over self or a method argument. The only potentially problematic is the one closing over an inst. var. (more over for writing!) WeakRegistry>>installFinalizer. It would be safer to transform this closure to just send a message rather than messing directly with the ivar... It's not a problem right now until we change WeakRegistry layout, but I wanted to share these findings, since that's the kind of tricky things we must care of when releasing a new version.

2 1

Jump to page: