Squeak-dev November 2012

squeak-dev@lists.squeakfoundation.org

56 participants
115 discussions

The Trunk: Tools-eem.428.mcz
by commits＠source.squeak.org 15 Nov '12

15 Nov '12

Eliot Miranda uploaded a new version of Tools to project The Trunk: http://source.squeak.org/trunk/Tools-eem.428.mcz ==================== Summary ==================== Name: Tools-eem.428 Author: eem Time: 15 November 2012, 12:59:12.101 pm UUID: 5d9b5b4d-52fa-4489-8803-f4ad4c1e02b2 Ancestors: Tools-eem.427 Provide a run-until-expression facility for the debugger. Thanks to Bob Arning and Bert Freudenberg for UI advice. =============== Diff against Tools-eem.427 =============== Item was changed: CodeHolder subclass: #Debugger + instanceVariableNames: 'interruptedProcess interruptedController contextStack contextStackIndex contextStackList receiverInspector contextVariablesInspector externalInterrupt proceedValue selectingPC savedCursor isolationHead failedProject errorWasInUIProcess labelString message untilExpression' - instanceVariableNames: 'interruptedProcess interruptedController contextStack contextStackIndex contextStackList receiverInspector contextVariablesInspector externalInterrupt proceedValue selectingPC savedCursor isolationHead failedProject errorWasInUIProcess labelString message' classVariableNames: 'ContextStackKeystrokes ErrorRecursion InterruptUIProcessIfBlockedOnErrorInBackgroundProcess' poolDictionaries: '' category: 'Tools-Debugger'! !Debugger commentStamp: '<historical>' prior: 0! I represent the machine state at the time of an interrupted process. I also represent a query path into the state of the process. The debugger is typically viewed through a window that views the stack of suspended contexts, the code for, and execution point in, the currently selected message, and inspectors on both the receiver of the currently selected message, and the variables in the current context. Special note on recursive errors: Some errors affect Squeak's ability to present a debugger. This is normally an unrecoverable situation. However, if such an error occurs in an isolation layer, Squeak will attempt to exit from the isolation layer and then present a debugger. Here is the chain of events in such a recovery. * A recursive error is detected. * The current project is queried for an isolationHead * Changes in the isolationHead are revoked * The parent project of isolated project is returned to * The debugger is opened there and execution resumes. If the user closes that debugger, execution continues in the outer project and layer. If, after repairing some damage, the user proceeds from the debugger, then the isolationHead is re-invoked, the failed project is re-entered, and execution resumes in that world. ! Item was changed: ----- Method: Debugger>>codePaneMenu:shifted: (in category 'code pane menu') ----- codePaneMenu: aMenu shifted: shifted + aMenu + add: 'run to here' target: self selector: #runToSelection: argument: thisContext sender receiver selectionInterval; + add: 'run until...' target: self selector: #runUntil; + addLine. - aMenu add: 'run to here' target: self selector: #runToSelection: argument: thisContext sender receiver selectionInterval. - aMenu addLine. super codePaneMenu: aMenu shifted: shifted. ^aMenu.! Item was added: + ----- Method: Debugger>>runUntil (in category 'code pane menu') ----- + runUntil + "Step until an expression evaluates to other than false, reporting an erroer if it doesn't evaluate to true. + Remember the expression in an inst var. If shift is pressed when the expression is supplied, don't update the UI. + If shift is pressed while stepping, stop stepping. Using a user interrupt to break out would be more natural + but Squeak currently doesn't provide a UserInterrupt expection. It should do." + | expression receiver context method value lastUpdate updateUI breakOnShift | + expression := UIManager default + request: 'run until expression is true (shift to disable ui update; shift to break).' + initialAnswer: (untilExpression ifNil: 'boolean expression'). + (expression isNil or: [expression isEmpty]) ifTrue: + [^self]. + updateUI := breakOnShift := Sensor shiftPressed not. + untilExpression := expression. + context := self selectedContext. + receiver := context receiver. + FakeClassPool adopt: receiver class. + method := receiver class evaluatorClass new + compiledMethodFor: untilExpression + in: context + to: receiver + notifying: nil + ifFail: [FakeClassPool adopt: nil. ^ #failedDoit] + logged: false. + + lastUpdate := Time millisecondClockValue. + FakeClassPool adopt: nil. + [self selectedContext == context + and: [(value := receiver with: context executeMethod: method) == false]] whileTrue: + [self doStep. + Time millisecondClockValue - lastUpdate > 50 ifTrue: + [updateUI ifTrue: [World displayWorldSafely]. + breakOnShift + ifTrue: [Sensor shiftPressed ifTrue: [^self]] + ifFalse: [Sensor shiftPressed ifFalse: [breakOnShift := true]]. + lastUpdate := Time millisecondClockValue]]. + (value ~~ false and: [value ~~ true]) ifTrue: + [UIManager default inform: 'expression ', (untilExpression contractTo: 40), ' answered ', (value printString contractTo: 20), '!!!!']!

1 0

The Trunk: NetworkTests-dtl.35.mcz
by commits＠source.squeak.org 15 Nov '12

15 Nov '12

David T. Lewis uploaded a new version of NetworkTests to project The Trunk: http://source.squeak.org/trunk/NetworkTests-dtl.35.mcz ==================== Summary ==================== Name: NetworkTests-dtl.35 Author: dtl Time: 15 November 2012, 8:06:29.937 am UUID: 7dfbff49-ba92-43d2-9c29-e22b2c12bfe3 Ancestors: NetworkTests-dtl.34 Revert change to SocketTest>>testSocketReuse because it fails on the squeakci.org test server. To be updated and/or resubmitted to inbox when a better update is available. =============== Diff against NetworkTests-dtl.34 =============== Item was changed: ----- Method: SocketTest>>testSocketReuse (in category 'tests') ----- testSocketReuse + "Test for SO_REUSEADDR/SO_REUSEPORT" - "Test for SO_REUSEADDR/SO_REUSEPORT. Test will fail if the platform does - not implement these socket options. Support for SO_REUSEPORT is not - present on many systems, and a failure on setting this socket option reflects - a limitation of the underlying platform." | address port udp1 send1 udp2 recv2 sendProc recvProc received | address := #[255 255 255 255]. "broadcast" port := 31259. [ udp1 := Socket newUDP. + udp1 setOption: 'SO_REUSEADDR' value: 1. + udp1 setOption: 'SO_REUSEPORT' value: 1. - self assert: (udp1 setOption: 'SO_REUSEADDR' value: 1) notNil - description: 'SO_REUSEADDR not supported'. - self assert: (udp1 setOption: 'SO_REUSEPORT' value: 1) notNil - description: 'SO_REUSEPORT not supported'. udp1 setPort: port. + udp1 setOption: 'SO_BROADCAST' value: 1. - self assert: (udp1 setOption: 'SO_BROADCAST' value: 1) notNil - description: 'SO_BROADCAST not supported'. send1 := UUID new. udp2 := Socket newUDP. + udp2 setOption: 'SO_REUSEADDR' value: 1. + udp2 setOption: 'SO_REUSEPORT' value: 1. - self assert: (udp2 setOption: 'SO_REUSEADDR' value: 1) notNil - description: 'SO_REUSEADDR not supported'. - self assert: (udp2 setOption: 'SO_REUSEPORT' value: 1) notNil - description: 'SO_REUSEPORT not supported'. udp2 setPort: port. + udp2 setOption: 'SO_BROADCAST' value: 1. - self assert: (udp2 setOption: 'SO_BROADCAST' value: 1) notNil - description: 'SO_BROADCAST not supported'. recv2 := UUID new. received := 0. recvProc := [ [received < 16] whileTrue:[ received := received + (udp2 receiveDataInto: recv2 startingAt: received + 1). ] ] fork. sendProc := [ udp1 setPeer: address port: port. udp1 sendData: send1 count: 16. ] fork. (Delay forMilliseconds: 200) wait. self should: [recvProc isTerminated]. self should: [sendProc isTerminated]. self should: [send1 = recv2]. ] ensure:[ + udp1 destroy. + udp2 destroy. - udp1 ifNotNilDo: [:s | s destroy]. - udp2 ifNotNilDo: [:s | s destroy]. ]. !

1 0

Re: [squeak-dev] The Trunk: KernelTests-dtl.235.mcz
by Frank Shearar 15 Nov '12

15 Nov '12

On 4 November 2012 18:09, <commits(a)source.squeak.org> wrote: > David T. Lewis uploaded a new version of KernelTests to project The Trunk: > http://source.squeak.org/trunk/KernelTests-dtl.235.mcz > > ==================== Summary ==================== > > Name: KernelTests-dtl.235 > Author: dtl > Time: 4 November 2012, 1:08:04.014 pm > UUID: de45cee9-3416-4fd8-a216-b5cafe70ee57 > Ancestors: KernelTests-ul.234 > > As of Kernel-cmm.671 Timespans created in the context of an offset will start in that offset, and when no context is available the defaultOffset for a Timespan will be zero. This affects a TimeStamp created from a string representation of itself, which will now have zero offset (UTC). Update TimeStampTest to reflect to this behavior. > > =============== Diff against KernelTests-ul.234 =============== This fixes two more failing tests, bringing us to: * Tests.Localization.LocaleTest.testLocaleChanged (for which we need to find a preference that differs between the 'en' and 'ja' locales') * Tests.Release.ReleaseTest.testNoObsoleteClasses (failing because some of the Monticello tests don't properly clear up after themselves) and, if you're running a Cog VM, also: KernelTests.Numbers.LargeNegativeIntegerTest.testMinimumNegativeIntegerArithmetic Thanks, Dave! frank

3 13

The Trunk: Compiler-eem.243.mcz
by commits＠source.squeak.org 15 Nov '12

15 Nov '12

Eliot Miranda uploaded a new version of Compiler to project The Trunk: http://source.squeak.org/trunk/Compiler-eem.243.mcz ==================== Summary ==================== Name: Compiler-eem.243 Author: eem Time: 14 November 2012, 11:54:44.286 pm UUID: e2c4d1ed-4a14-4481-b887-dae54eecf492 Ancestors: Compiler-eem.242 Use a smaller hammer to crack the same tiny nut. =============== Diff against Compiler-eem.242 =============== Item was changed: ----- Method: Parser>>pragmaLiteral: (in category 'pragmas') ----- pragmaLiteral: selectorSoFar "Read a pragma literal. As a nicety we allow a variable name (rather than a literal string) as the second argument to primitive:error:" (hereType == #string or: [ hereType == #literal or: [ hereType == #number or: [ hereType == #character ] ] ]) ifTrue: [ ^ self advance ]. (here == $# and: [ tokenType == #word ]) ifTrue: [ ^ self advance ]. (here == #- and: [ tokenType == #number ]) ifTrue: [ ^ (self advance; advance) negated ]. (here = 'true' or: [ here = 'false' or: [ here = 'nil' ] ]) + ifTrue: [ ^ (Scanner new scanTokens: self advance) first ]. - ifTrue: [ ^ Compiler evaluate: self advance ]. "This nicety allows one to supply a primitive error temp as a variable name, rather than a string." ((selectorSoFar beginsWith: 'primitive:') and: [(selectorSoFar endsWith: 'error:') and: [hereType == #word]]) ifTrue: [^self advance]. ^self expected: 'Literal constant'!

1 0

The Trunk: NetworkTests-dtl.34.mcz
by commits＠source.squeak.org 15 Nov '12

15 Nov '12

David T. Lewis uploaded a new version of NetworkTests to project The Trunk: http://source.squeak.org/trunk/NetworkTests-dtl.34.mcz ==================== Summary ==================== Name: NetworkTests-dtl.34 Author: dtl Time: 15 November 2012, 12:05:37.979 am UUID: 661c04dd-3386-4424-aa92-944ad27fa74a Ancestors: NetworkTests-dtl.33 SocketTest>>testSocketReuse will fail on platforms that do not implement the SO_REUSEPORT socket option. Update the test comment to explain this, and provide assertions that identify the cause of the failure. Note that failure on setting SO_REUSEPORT reflects a limitation of the underlying platform network support, and does not indicate an error in Squeak networking. =============== Diff against NetworkTests-dtl.33 =============== Item was changed: ----- Method: SocketTest>>testSocketReuse (in category 'tests') ----- testSocketReuse + "Test for SO_REUSEADDR/SO_REUSEPORT. Test will fail if the platform does + not implement these socket options. Support for SO_REUSEPORT is not + present on many systems, and a failure on setting this socket option reflects + a limitation of the underlying platform." - "Test for SO_REUSEADDR/SO_REUSEPORT" | address port udp1 send1 udp2 recv2 sendProc recvProc received | address := #[255 255 255 255]. "broadcast" port := 31259. [ udp1 := Socket newUDP. + self assert: (udp1 setOption: 'SO_REUSEADDR' value: 1) notNil + description: 'SO_REUSEADDR not supported'. + self assert: (udp1 setOption: 'SO_REUSEPORT' value: 1) notNil + description: 'SO_REUSEPORT not supported'. - udp1 setOption: 'SO_REUSEADDR' value: 1. - udp1 setOption: 'SO_REUSEPORT' value: 1. udp1 setPort: port. + self assert: (udp1 setOption: 'SO_BROADCAST' value: 1) notNil + description: 'SO_BROADCAST not supported'. - udp1 setOption: 'SO_BROADCAST' value: 1. send1 := UUID new. udp2 := Socket newUDP. + self assert: (udp2 setOption: 'SO_REUSEADDR' value: 1) notNil + description: 'SO_REUSEADDR not supported'. + self assert: (udp2 setOption: 'SO_REUSEPORT' value: 1) notNil + description: 'SO_REUSEPORT not supported'. - udp2 setOption: 'SO_REUSEADDR' value: 1. - udp2 setOption: 'SO_REUSEPORT' value: 1. udp2 setPort: port. + self assert: (udp2 setOption: 'SO_BROADCAST' value: 1) notNil + description: 'SO_BROADCAST not supported'. - udp2 setOption: 'SO_BROADCAST' value: 1. recv2 := UUID new. received := 0. recvProc := [ [received < 16] whileTrue:[ received := received + (udp2 receiveDataInto: recv2 startingAt: received + 1). ] ] fork. sendProc := [ udp1 setPeer: address port: port. udp1 sendData: send1 count: 16. ] fork. (Delay forMilliseconds: 200) wait. self should: [recvProc isTerminated]. self should: [sendProc isTerminated]. self should: [send1 = recv2]. ] ensure:[ + udp1 ifNotNilDo: [:s | s destroy]. + udp2 ifNotNilDo: [:s | s destroy]. - udp1 destroy. - udp2 destroy. ]. !

1 0

Divergence between VM and debugger handling of primitives
by Ryan Macnak 14 Nov '12

14 Nov '12

I have noticed that the execution simulation (ContextPart) and the VM differ in how they handle primitives with extra arguments. The execution simulation takes arguments relative to the base of the stack, but the VM takes them relative to the top of the stack. Most uses of primitives don't have extra arguments and this distinction does not affect them, but Newspeak takes advantage of the VM's behavior to implement primitives-as-capabilities without a VM change. For example, primitive 188: Normal code a withArguments: {b. c} executeMethod: m "Execute the method m, with a as the receiver, and with b and c as the arguments" VMMirror vmmirror on: a withArguments: {b. c} executeMethod: m "VM: Execute the method m, with a as the receiver, and with b and c as the arguments. " "Debugger: Try to execute {b. c}, fail because it doesn't behave like a method" Are there any objections to making the execution simulation behave like the VM? Ryan

1 0

Puzzled about Form#stencil
by Sebastian Nozzi 14 Nov '12

14 Nov '12

Hello Squeakers! I am really puzzled about this: I have 2 PNGs, both with transparent pixels. Although both are rendered accordingly when I ask for the "stencil" (map of transparent vs. opaque pixels) one of them gives me the correct image, whereas for the other I just get a black box. Using the files attached, "triangle-sketch" gives me the correct mask and "home-gimp..." gives me the black box. I evaluate this in a workspace, first by d&d the files into Squeak (4.3, all-in-one, Mac), and then d&d into the workspace: (SketchMorph withForm: sketch305 form stencil) openInHand. (SketchMorph withForm: sketch1737 form stencil) openInHand. I show the situation in the attached screenshot. Any idea why this is so? Thanks in advance! Sebastian

2 5

Re: [squeak-dev] The Inbox: Files-kfr.120.mcz
by Frank Shearar 13 Nov '12

13 Nov '12

On 12 November 2012 18:23, <commits(a)source.squeak.org> wrote: > A new version of Files was added to project The Inbox: > http://source.squeak.org/inbox/Files-kfr.120.mcz > > ==================== Summary ==================== > > Name: Files-kfr.120 > Author: kfr > Time: 12 November 2012, 6:23:28.531 pm > UUID: 4a518133-203f-bd43-9fce-78b8dcfb337d > Ancestors: Files-cmm.119 > > Fix for test failure on Windows > > =============== Diff against Files-cmm.119 =============== > I haven't done a source QA on this, but it does make the test pass. (Tested on Win 7.) frank

1 0

Daily Commit Log
by commits＠source.squeak.org 13 Nov '12

13 Nov '12

Changes to Trunk (http://source.squeak.org/trunk.html) in the last 24 hours: http://lists.squeakfoundation.org/pipermail/packages/2012-November/005558.h… Name: Tools-eem.427 Ancestors: Tools-ul.425 Include class comments when filing out recent changes ============================================= http://lists.squeakfoundation.org/pipermail/packages/2012-November/005559.h… Name: System-eem.502 Ancestors: System-eem.500, System-cwp.501 Fix merge slip with System-cwp.501. Include temp names in methods installed by BreakpointManager (i.e. toggle break on entry). =============================================

1 0

Windows 7 tests
by Herbert König 12 Nov '12

12 Nov '12

Hi, after I'm reunited with my laptops power supply I get three failed tests 10 errors with image # 12279 http://minus.com/lBTicgU2NHawm I used the download from Squeakci To get this result I have to do an update code from server, which does not change the image update#. If I just disable the preference and use #12279 from Squeakci w/o update the socket tests produce errors again. 3267 run, 3226 passes, 18 expected failures, 2 failures, 21 errors, 0 unexpected passes but in this case the weakSetInspectorTest doesn't fail?? http://minus.com/lbzytpWsoyF4G5 I produced both results two times. The unusual part of this configuration is that I get my internet from an UMTS usb stick. Again I have no entry in squeak debug.log I'll report if I can reproduce this on my wired network tomorrow. Cheers Herbert

1 2

← Newer
1
2
3
4
5
6
7
8
9
...
12
Older →