Squeak-dev August 2013

squeak-dev@lists.squeakfoundation.org

45 participants
157 discussions

BitBlt alpha blend rules broken?
by tim Rowledge 09 Sep '13

09 Sep '13

The guy that is hacking BitBlt for ARM for me is having all sorts of fun and making good progress. I even have a VM working with an early version and am currently in the process of testing the results. He's started on some of the more….curious… corners and right now is of the opinion that alphaBlend is getting its maths wrong. I'll just quote him rather than paraphrasing; > "The next combinationRule I've tackled is alphaBlend. This is clearly only > applicable to 32bpp source and destination images. Despite what the > comments in the original source say, this *does* calculate the alpha > component of the output, rather than zero it. In fact, it's using the > formula: > > da' = sa + (1-sa).da > dr' = sa.sr + (1-sa).dr > dg' = sa.sg + (1-sa).dg > db' = sa.sb + (1-sa).db > > where sa = source alpha, etc. > > This is actually incorrect maths for an alpha blend operation with non- > premultiplied colours, as appears to be the intent. For the RGB colour > components, the contribution from the first term is da' times too large, > and the contribution from the second term is da'/da times too large. In > other words, you only get the correct results if the destination image > was fully opaque (which implies that the result is likewise). > > Nevertheless, I suspect we'd run into problems if we attempted to fix > this, so I'll endeavour to replicate the behaviour, bugs and all. > > Next issue was the fact that for some crazy reason, they'd decided to > renormalise the colour components after the multiplication by doing a > divide by 255, rounding to +infinity. This is neither the "best" approach > (arguably round to nearest with rounding half-values to odd or even would > be the best), nor is it particularly efficient to implement - divisions > rarely are. The current code also has a completely pointless bitwise AND > with 0xFF following the division, since the inputs to the division cannot > exceed 0xFF*0xFF to begin with." We're talking about BitBltSimulation>alphaBlend:with: in this case. It's very old code. I'm not into colour blending maths etc and I can't see why we're using 255 instead of 256 (which would allow use of some ARM pixel engine instructions) nor quite what the intent is here. Thoughts, anyone? tim -- tim Rowledge; tim(a)rowledge.org; http://www.rowledge.org/tim Strange OpCodes: IAM: Increase Amperage Above Maximum

3 4

Smalltalk cleanUp: agressiveOrNot
by Chris Muller 04 Sep '13

04 Sep '13

>> When I prepare a new "shrunk trunk", I run this guy: >> https://github.com/squeak-smalltalk/squeak-ci/blob/master/shrink-trunk.st. >> You'll see it just unloads packages and implicitly trusts that those >> packages know how to correctly unload themselves. > > What do you think about having shrink-trunk.st simply be: > > Smalltalk unloadAllKnownPackages > > so we can then we can maintain it from the image..? That method does a lot of stuff which is somewhat redundant. It's too long anyway so I'll factor it so we can access just what we want in your shrink-trunk.st script. BTW, in ReleaseBuilder we call an agressive Smalltalk cleanUp: (true) vs. with #unloadAllKnownPackages and also your shrink we just call Smalltalk cleanUp: (false). I've looked at the methods and seen the differences between aggressive and not but... I was wondering does anyone know whether there is an over-arching conceptual difference between the two? e.g., under what circumstances would a particular clean-up be considered aggressive vs. not?

2 1

funny error
by Bob Arning 02 Sep '13

02 Sep '13

If you are typing some javascript inside a squeak string, like: 'mumble("x"); grumble("*");' there are subtle things that can go wrong. If you inadvertently type mumble('x'); you'll get a compiler error, but if you type grumble('*'); can you guess the result without actually running it? Cheers, Bob

3 3

FileStream size limitation?
by Squeak List 02 Sep '13

02 Sep '13

Hello, I have an OrderedCollection with 24,576 elements. A bunch of musical notes basically: each of the 24,576 elements is like: #('eo6' 0.25 500) for example. Goal: I would like to turn that *entire* OrderedCollection into a single MIDI file. However, there seems to be some sort of size limitation that I do not see how to get around. 1) Even before getting to the MIDI part, I tried: FileStream fileNamed: '1-24576.rtf' do: [:stream | stream nextPutAll: ((oc copyFrom: 1 to:24576) asString)]. which creates a nice file of 48.8KB size. And opening the file, it ends with: #('c#o6' 0.25 500) #('d...etc... which is not what I expected. 2) If I do: FileStream fileNamed: '1-2729.rtf' do: [:stream | stream nextPutAll: ((oc copyFrom: 1 to:2729) asString)]. it works fine: ending tidily with the expected: #('c#o6' 0.25 500)) 3) But, FileStream fileNamed: '1-2730.rtf' do: [:stream | stream nextPutAll: ((oc copyFrom: 1 to:2730) asString)]. is where this starts: #('c#o6' 0.25 500) #('d...etc... This exact same problem happens when making MIDI files: there seems to be a size limitation around 48.8KB. Aside from MIDI, is FileStream just the wrong class to use for anything larger than creating a 48.8kb text file (.txt or .rtf)? Or how can I get around this restriction? Any suggestions are most welcome. Thanks, ken Windows 7 - 64 bit Squeak4.4 latest update: #12319 Current Change Set: Unnamed Image format 6505 (32 bit) Squeak.exe Croquet Closure Cog VM [CoInterpreter VMMaker.oscog-eem.331] Win32 built on Aug 22 2013 10:20:54 Compiler: 3.4.4 (cygming special, gdc 0.12, using dmd 0.125) platform sources revision VM: r2776

5 6

Re: [squeak-dev] The Trunk: CollectionsTests-cmm.206.mcz
by Frank Shearar 01 Sep '13

01 Sep '13

On 31 August 2013 22:34, <commits(a)source.squeak.org> wrote: > Chris Muller uploaded a new version of CollectionsTests to project The Trunk: > http://source.squeak.org/trunk/CollectionsTests-cmm.206.mcz > > ==================== Summary ==================== > > Name: CollectionsTests-cmm.206 > Author: cmm > Time: 31 August 2013, 4:34:41.685 pm > UUID: 18d460d5-be23-418b-a3f0-b340b0c3e948 > Ancestors: CollectionsTests-nice.205 > > testAsMutator test with a selector that already ends with a colon. > > =============== Diff against CollectionsTests-nice.205 =============== > > Item was changed: > ----- Method: SymbolTest>>testAsMutator (in category 'tests') ----- > testAsMutator > + self > + assert: #x asMutator = #x: ; > + assert: #x: asMutator = #x: ; > + assert: #x asMutator isSymbol! > - > - self assert: #x asMutator = #x:. > - self assert: #x asMutator isSymbol! I've been ripping out #assert: calls in the tests and replacing them with #assert:equals: mainly because #assert:equals: produces a much nicer error message: looking at the failing test on CI will tell you how the assert failed, not just that it failed. frank

2 1

Daily Commit Log
by commits＠source.squeak.org 01 Sep '13

01 Sep '13

Changes to Trunk (http://source.squeak.org/trunk.html) in the last 24 hours: http://lists.squeakfoundation.org/pipermail/packages/2013-August/006321.html Name: System-cmm.586 Ancestors: System-fbs.585 - Remove modal confirmation pop-up buried in system code. If this is necessary it should be invoked from calling code. - Factor SmalltalkImage>>#unloadAllKnownPackages into several smaller methods to allow more deliberate shrinking. - Remove method referring to undefined "ScrapsBook". ============================================= http://lists.squeakfoundation.org/pipermail/packages/2013-August/006322.html Name: Installer-Core-cmm.378 Ancestors: Installer-Core-cmm.377 - Initial sketch of our newly-modularised package hierarchy from #core, used by #system and topped with #squeakRelease for the full release image content. Now used by ReleaseBuilder for help in building the release image. - Fix Installer>>#unload to disableUpdatesOfPackage: unloaded packages. ============================================= http://lists.squeakfoundation.org/pipermail/packages/2013-August/006323.html Name: ReleaseBuilder-cmm.102 Ancestors: ReleaseBuilder-fbs.101 - Delegate to Installer to merge the #squeakRelease "package" for building releases. ============================================= http://lists.squeakfoundation.org/pipermail/packages/2013-August/006324.html Name: Tests-cmm.247 Ancestors: Tests-nice.246 Tests for loading each of the external-packages defined in Installer. ============================================= http://lists.squeakfoundation.org/pipermail/packages/2013-August/006325.html Name: Tests-cmm.250 Ancestors: Tests-cmm.249, Tests-fbs.247 - Merged fbs.247. - 'external-packages' category renamed to 'package-definitions' because it now includes internal packages. ============================================= http://lists.squeakfoundation.org/pipermail/packages/2013-August/006326.html Name: CollectionsTests-cmm.206 Ancestors: CollectionsTests-nice.205 testAsMutator test with a selector that already ends with a colon. =============================================

1 0

Re: [squeak-dev] The Trunk: System-cmm.586.mcz
by Chris Muller 01 Sep '13

01 Sep '13

Hi Frank, with this shrink-trunk.st can be shorter, and using code under SCM: Smalltalk unloadReloadablePackages; cleanUp On Sat, Aug 31, 2013 at 3:29 PM, <commits(a)source.squeak.org> wrote: > Chris Muller uploaded a new version of System to project The Trunk: > http://source.squeak.org/trunk/System-cmm.586.mcz > > ==================== Summary ==================== > > Name: System-cmm.586 > Author: cmm > Time: 31 August 2013, 3:29:01.755 pm > UUID: bbffd13c-e516-462d-ac89-e3880a28c382 > Ancestors: System-fbs.585 > > - Remove modal confirmation pop-up buried in system code. If this is necessary it should be invoked from calling code. > - Factor SmalltalkImage>>#unloadAllKnownPackages into several smaller methods to allow more deliberate shrinking. > - Remove method referring to undefined "ScrapsBook". > > =============== Diff against System-fbs.585 =============== > > Item was changed: > ----- Method: SmalltalkImage>>cleanUp:except: (in category 'housekeeping') ----- > + cleanUp: aggressive except: exclusions > - cleanUp: aggressive except: exclusions > "Clean up. When aggressive is true, this will destroy projects, change sets, etc. > Leave out any classes specifically listed in exclusions." > - > "Smalltalk cleanUp: true except: {Project. ChangeSet}" > + "Find all classes implementing #cleanUp or cleanUp:" > + | classes | > + classes := Smalltalk allClasses select: > + [ : eachClass | (eachClass class includesSelector: #cleanUp) or: > + [ (eachClass class includesSelector: #cleanUp:) and: [ (exclusions includes: eachClass) not ] ] ]. > + "Arrange classes in superclass order, superclasses before subclasses, so that specific cleanup (like MethodDictionary compaction) will run after generic superclass cleanup (HashedCollection rehashing). Otherwise generic superclass cleanup might undo specific one (in this case rehashing will undo a good bit of MD compaction)." > + (ChangeSet superclassOrder: classes) > + do: [ : aClass | aClass cleanUp: aggressive ] > + displayingProgress: [ : aClass | 'Cleaning up in ' , aClass name ]! > - > - ^self cleanUp: aggressive except: exclusions confirming: aggressive! > > Item was removed: > - ----- Method: SmalltalkImage>>cleanUp:except:confirming: (in category 'housekeeping') ----- > - cleanUp: aggressive except: exclusions confirming: aBool > - "Clean up. When aggressive is true, this will destroy projects, change sets, etc. > - Leave out any classes specifically listed in exclusions." > - > - "Smalltalk cleanUp: true except: {Project. ChangeSet}" > - > - | classes | > - aBool ifTrue:[ > - "Give the user a chance to bail" > - (self confirm: 'Cleanup will destroy projects, change sets and more. > - Are you sure you want to proceed?') ifFalse:[^self]. > - ]. > - > - "Find all classes implementing #cleanUp or cleanUp:" > - classes := Smalltalk allClasses select:[:aClass| > - (aClass class includesSelector: #cleanUp) > - or:[aClass class includesSelector: #cleanUp:] > - ]. > - > - "Leave out the classes in the exclusion set" > - classes := classes reject:[:aClass| exclusions includes: aClass]. > - > - "Arrange classes in superclass order, superclasses before subclasses. > - This will ensure that specific cleanup (like MethodDictionary compaction) > - will run after generic superclass cleanup (HashedCollection rehashing). > - Otherwise generic superclass cleanup might undo specific one (in this > - case rehashing will undo a good bit of MD compaction)." > - classes := ChangeSet superclassOrder: classes. > - > - "Run the cleanup code" > - classes > - do:[:aClass| aClass cleanUp: aggressive] > - displayingProgress:[:aClass| 'Cleaning up in ', aClass name].! > > Item was added: > + ----- Method: SmalltalkImage>>cleanUpEtoys (in category 'shrinking') ----- > + cleanUpEtoys > + StandardScriptingSystem removeUnreferencedPlayers. > + Project removeAllButCurrent. > + #('Morphic-UserObjects' 'EToy-UserObjects' 'Morphic-Imported' ) do: > + [ : each | SystemOrganization removeSystemCategory: each ]. > + World dumpPresenter. > + Presenter defaultPresenterClass: nil. > + Preferences removePreference: #allowEtoyUserCustomEvents. > + ! > > Item was added: > + ----- Method: SmalltalkImage>>removeChangeSets (in category 'shrinking') ----- > + removeChangeSets > + ChangeSet removeChangeSetsNamedSuchThat: > + [ : each | (each == ChangeSet current) not ]. > + ChangeSet current > + clear ; > + name: 'Unnamed1'! > > Item was added: > + ----- Method: SmalltalkImage>>shrink (in category 'shrinking') ----- > + shrink > + "Clean up everything we know how to clean up that can be restored." > + "Prepare unloading" > + self zapMVCprojects. > + Flaps disableGlobalFlaps: false. > + self cleanUpEtoys. > + "Yeow, why turn off listening for system changes?" > + self > + at: #ServiceRegistry > + ifPresent: [ : aClass | SystemChangeNotifier uniqueInstance noMoreNotificationsFor: aClass ]. > + self unloadReloadablePackages. > + "Post-unload cleanup" > + SmalltalkImage cleanUp. > + MCWorkingCopy flushObsoletePackageInfos. > + "aggressive cleanUp" > + SystemOrganization removeSystemCategory: 'UserObjects'. > + ScheduledControllers := nil. > + SystemOrganization removeEmptyCategories. > + StandardScriptingSystem initialize. > + "aggressive cleanUp" > + self removeChangeSets. > + self > + at: #Browser > + ifPresent: > + [ : br | br initialize ]. > + DebuggerMethodMap voidMapCache. > + DataStream initialize. > + Smalltalk garbageCollect! > > Item was added: > + ----- Method: SmalltalkImage>>shrinkAndCleanDesktop (in category 'shrinking') ----- > + shrinkAndCleanDesktop > + World removeAllMorphs. > + self shrink. > + MorphicProject defaultFill: (Color gray: 0.9). > + World color: (Color gray: 0.9)! > > Item was removed: > - ----- Method: SmalltalkImage>>unloadAllKnownPackages (in category 'shrinking') ----- > - unloadAllKnownPackages > - "Unload all packages we know how to unload and reload" > - > - "Prepare unloading" > - Smalltalk zapMVCprojects. > - Flaps disableGlobalFlaps: false. > - StandardScriptingSystem removeUnreferencedPlayers. > - Project removeAllButCurrent. > - #('Morphic-UserObjects' 'EToy-UserObjects' 'Morphic-Imported' ) > - do: [:each | SystemOrganization removeSystemCategory: each]. > - Smalltalk at: #ServiceRegistry ifPresent:[:aClass| > - SystemChangeNotifier uniqueInstance > - noMoreNotificationsFor: aClass. > - ]. > - World removeAllMorphs. > - > - "Go unloading" > - #( 'ReleaseBuilder' 'ScriptLoader' > - '311Deprecated' '39Deprecated' > - 'Universes' 'SMLoader' 'SMBase' 'Installer-Core' > - 'VersionNumberTests' 'VersionNumber' > - 'Services-Base' 'PreferenceBrowser' 'Nebraska' > - 'ToolBuilder-MVC' 'ST80Tools' 'ST80Tests' 'ST80' > - 'CompressionTests' > - 'CollectionsTests' 'GraphicsTests' 'KernelTests' 'MorphicTests' > - 'MultilingualTests' 'NetworkTests' 'ToolsTests' 'TraitsTests' > - 'SystemChangeNotification-Tests' 'FlexibleVocabularies' > - 'EToys' 'Protocols' 'XML-Parser' 'Tests' 'SUnitGUI' > - 'Help-Squeak' 'HelpSystem' 'SystemReporter' > - ) do: [:pkgName| > - (MCPackage named: pkgName) unload. > - MCMcmUpdater disableUpdatesOfPackage: pkgName. > - ]. > - "Traits use custom unload" > - Smalltalk at: #Trait ifPresent:[:aClass| aClass unloadTraits]. > - > - "Post-unload cleanup" > - SmalltalkImage cleanUp. > - MCWorkingCopy flushObsoletePackageInfos. "aggressive cleanUp" > - SystemOrganization removeSystemCategory: 'UserObjects'. > - Presenter defaultPresenterClass: nil. > - World dumpPresenter. > - ScheduledControllers := nil. > - Preferences removePreference: #allowEtoyUserCustomEvents. > - SystemOrganization removeEmptyCategories. > - StandardScriptingSystem initialize. "aggressive cleanUp" > - ChangeSet removeChangeSetsNamedSuchThat:[:cs | (cs == ChangeSet current) not]. > - ChangeSet current clear. > - ChangeSet current name: 'Unnamed1'. > - Smalltalk at: #Browser ifPresent:[:br| br initialize]. > - DebuggerMethodMap voidMapCache. > - DataStream initialize. > - > - Smalltalk garbageCollect. > - MorphicProject defaultFill: (Color gray: 0.9). > - World color: (Color gray: 0.9). > - ! > > Item was added: > + ----- Method: SmalltalkImage>>unloadReloadablePackages (in category 'shrinking') ----- > + unloadReloadablePackages > + "Unload packages which can be reloaded." > + #('ReleaseBuilder' 'ScriptLoader' '311Deprecated' '39Deprecated' 'Universes' 'SMLoader' 'SMBase' 'VersionNumberTests' 'VersionNumber' 'Services-Base' 'PreferenceBrowser' 'Nebraska' 'ToolBuilder-MVC' 'ST80Tools' 'ST80Tests' 'ST80' 'CompressionTests' 'CollectionsTests' 'GraphicsTests' 'KernelTests' 'MorphicTests' 'MultilingualTests' 'NetworkTests' 'ToolsTests' 'TraitsTests' 'SystemChangeNotification-Tests' 'FlexibleVocabularies' 'EToys' 'Protocols' 'XML-Parser' 'Tests' 'SUnitGUI' 'Help-Squeak' 'HelpSystem' 'SystemReporter' ) do: [ : pkgName | Installer mc unload: pkgName ]. > + "Traits use custom unload" > + self > + at: #Trait > + ifPresent: [ : aClass | aClass unloadTraits ]! > > Item was removed: > - ----- Method: Utilities class>>cleanUp: (in category 'class initialization') ----- > - cleanUp: agressive > - "Nuke the scraps book when cleaning aggressively" > - > - agressive ifTrue:[ScrapsBook := nil].! > >

1 0

The Trunk: CollectionsTests-cmm.206.mcz
by commits＠source.squeak.org 31 Aug '13

31 Aug '13

Chris Muller uploaded a new version of CollectionsTests to project The Trunk: http://source.squeak.org/trunk/CollectionsTests-cmm.206.mcz ==================== Summary ==================== Name: CollectionsTests-cmm.206 Author: cmm Time: 31 August 2013, 4:34:41.685 pm UUID: 18d460d5-be23-418b-a3f0-b340b0c3e948 Ancestors: CollectionsTests-nice.205 testAsMutator test with a selector that already ends with a colon. =============== Diff against CollectionsTests-nice.205 =============== Item was changed: ----- Method: SymbolTest>>testAsMutator (in category 'tests') ----- testAsMutator + self + assert: #x asMutator = #x: ; + assert: #x: asMutator = #x: ; + assert: #x asMutator isSymbol! - - self assert: #x asMutator = #x:. - self assert: #x asMutator isSymbol!

1 0

Re: [squeak-dev] The Trunk: Tests-cmm.247.mcz
by Chris Muller 31 Aug '13

31 Aug '13

I realized this got included as part of the McInfoProxy test (cmm.249). The idea is that, as we develop trunk, CI will report when external-packages no longer load successfully into trunk. On Sat, Aug 31, 2013 at 4:13 PM, <commits(a)source.squeak.org> wrote: > Chris Muller uploaded a new version of Tests to project The Trunk: > http://source.squeak.org/trunk/Tests-cmm.247.mcz > > ==================== Summary ==================== > > Name: Tests-cmm.247 > Author: cmm > Time: 8 August 2013, 10:54:50.713 pm > UUID: 93fd43c1-0207-42ed-bfe5-f2d6acfaa2c2 > Ancestors: Tests-nice.246 > > Tests for loading each of the external-packages defined in Installer. > > =============== Diff against Tests-nice.246 =============== > > Item was added: > + ----- Method: InstallerTest class>>suiteClass (in category 'as yet unclassified') ----- > + suiteClass > + ^ InstallerTestSuite! > > Item was added: > + TestSuite subclass: #InstallerTestSuite > + instanceVariableNames: '' > + classVariableNames: '' > + poolDictionaries: '' > + category: 'Tests-Installer-Core'! > > Item was added: > + ----- Method: InstallerTestSuite>>ensureTestMethods (in category 'initialize-release') ----- > + ensureTestMethods > + "Ensure InstallerTest has test methods for each externally-loadable package defined by Installer. By generating these methods, they don't have to be maintained separately." > + ^ (Installer methodsInCategory: 'external-packages') do: > + [ : each | (InstallerTest canUnderstand: each) ifFalse: > + [ InstallerTest > + compile: > + (String streamContents: > + [ : stream | stream > + nextPutAll: 'test'; nextPutAll: each; > + cr; tab; > + nextPutAll: 'Installer new merge: #'; nextPutAll: each ]) > + classified: '*generated' ] ]! > > Item was added: > + ----- Method: InstallerTestSuite>>initialize (in category 'initialize-release') ----- > + initialize > + super initialize. > + self ensureTestMethods! > >

1 0

The Trunk: Tests-cmm.250.mcz
by commits＠source.squeak.org 31 Aug '13

31 Aug '13

Chris Muller uploaded a new version of Tests to project The Trunk: http://source.squeak.org/trunk/Tests-cmm.250.mcz ==================== Summary ==================== Name: Tests-cmm.250 Author: cmm Time: 31 August 2013, 4:17:10.853 pm UUID: eb71599e-9529-46a5-82e6-3ff0f6466065 Ancestors: Tests-cmm.249, Tests-fbs.247 - Merged fbs.247. - 'external-packages' category renamed to 'package-definitions' because it now includes internal packages. =============== Diff against Tests-cmm.249 =============== Item was changed: ----- Method: InstallerTestSuite>>ensureTestMethods (in category 'initialize-release') ----- ensureTestMethods "Ensure InstallerTest has test methods for each externally-loadable package defined by Installer. By generating these methods, they don't have to be maintained separately." + ^ (Installer methodsInCategory: 'package-definitions') do: - ^ (Installer methodsInCategory: 'external-packages') do: [ : each | (InstallerTest canUnderstand: each) ifFalse: [ InstallerTest compile: (String streamContents: [ : stream | stream nextPutAll: 'test'; nextPutAll: each; cr; tab; nextPutAll: 'Installer new merge: #'; nextPutAll: each ]) classified: '*generated' ] ]! Item was changed: ----- Method: PackageDependencyTest>>testEtoys (in category 'tests') ----- testEtoys self testPackage: 'EToys' dependsExactlyOn: #( Balloon Collections Compiler Compression Files Graphics Kernel Morphic MorphicExtras Network + PreferenceBrowser Protocols SUnit System 'ToolBuilder-Kernel' Tools ).! Item was changed: ----- Method: PackageDependencyTest>>testPreferenceBrowser (in category 'tests') ----- testPreferenceBrowser self testPackage: 'PreferenceBrowser' dependsExactlyOn: #( Collections - EToys Graphics Kernel Morphic MorphicExtras System Tools ).!

1 0

Jump to page: