Squeak-dev April 2005

squeak-dev@lists.squeakfoundation.org

158 participants
287 discussions

Error when running using SUnit test runner in 3.7
by Yar Hwee Boon 03 Apr '05

03 Apr '05

I'm getting an error while using the SUnit test runner in 3.7, but I'm not sure it is related. Here's a screenshot at http://motionobj.com/temp/error-when-sunit.png. Does anyone know what's wrong? Thanks. -- HweeBoon MotionObj (65) 6764-9774

2 1

Re: Are what the janitors are doing, what needs to be done?
by Peace Jerome 03 Apr '05

03 Apr '05

>> Jerome Peace wrote: >>> >>> Are what the janitors are doing, what needs to be >>> done? Doug Way replied: > >The Janitors are doing something that needs to be done, although it is >only a subset of all that needs to be done. :) I tried to choose my question and words so as not to dispute this. Remember my idea of the proper goal is getting things 1) fixed and 2) included in the image with 3) the least amount of effort on everyones part. Also my perceptions have grown somewhat. Having someone who can direct traffic and educate bug reporters to make better reports would assist these goals. I still hold that the problem solvers need to be there else all you�ve got is an on ramp with no highway. > >The initial focus of the Janitors was just to make sure all issues were >tracked in Mantis (by moving bugs/fixes over there from this list). >But I think I managed to talk Ken into expanding the scope a bit, to >include Mantis and other process improvements as well. (See the >helpful "How to Report a New Squeak Issue" doc which Ken added to the >Mantis docs page.) I couldn�t find this. Where should I look? I tried the doc link on mantis but that only gets general doc there is no project specific doc. >Eventually, hopefully all of steps 1 & 2 from the >original Harvesting Process: http://minnow.cc.gatech.edu/squeak/3152 > >But step 3 from the Harvesting Process (incorporation) is more >complicated, and will be affected by the process of splitting the base >image into packages by the Packages team. Can you be more specific. Do fixes wait until packages are split? Will each package have someone who can commit fixes to it? And what happens when a fix spans several packages??? >Some of the discussion on the v3dot9 list has been about this and the >update stream. > >>> Quite frankly, My perception is the janitors will >>> generate a lot of noise and result in very little of >>> the actual problems being solved. It will obscure the >>> fact that what is needed is problem solvers and bug >>> fixers and not just text processors. > >Note also that the bug-transfer aspect of the Janitors mission should >be relatively short-term... eventually people will realize that the old >way is no longer supported. Ok. Time will tell. The noise is not a great problem if it does not distract from getting the fixes in. The resources needed are the problem solvers and the recruiters of problem solvers. Yours in service, -- Jerome Peace > > __________________________________ Do you Yahoo!? Take Yahoo! Mail with you! Get it on your mobile phone. http://mobile.yahoo.com/maildemo

2 1

[BUG][FIX] WriteBarrier at:put: fix for variable classes
by Chris Muller 03 Apr '05

03 Apr '05

The WriteBarrier class-comment states: - any methods that potentially modify instance variables will be overriden in this new class. The overridden method stores the original values of the instance variable in temps, then calls the super method, and then compares the current inst var values with the originals. - If the instance variable values have changed, the WriteBarrier will be notified with a send to #modified:, with the object that was modified as the single argument. - For variably-sized classes, #at:put: is also overridden to provide the same notification. While integrating WriteBarrier into Magma, I noticed my Array test was actually slower than without WriteBarrier. The problem is in WBClassBuilder>>installAtPutOverrides, which do not check whether the value is changed, it signals #modified no matter what. To demonstrate this: | a wb | a _ Array with: true. wb _ DirtySetWriteBarrier new. wb add: a. a at: 1 put: true. wb dirtySet includes: a This script should return false. After loading the attached fix, it will. Also, I have a question. Are there any caveats to overriding instVarAt:put: to check for a new value as well? The same class-comment says: - The WriteBarrier is semi-permeable: if you want to modify an object directly without triggering notifications, you can use #instVarAt:put: and #basicAt:put:. This reduces transparency albeit just a little; and this same permeability could be achieved by temporarily removing from the WriteBarrier. Is there any reason not to do it? Regards, Chris

3 2

#printAllStacks
by Bill Schwab 03 Apr '05

03 Apr '05

Hello all, Here are the promised changes - I think. I did this in 2002 on Squeak 3.2, so file this stuff into a current image with great care. At the moment, I do not have the time to build a new VM to test this[*], nor do I have ready access to the altered sources that I used for same. IIRC, I simply arranged for printing all stacks vs. only the active process' stack. I saw mention of some problems with capturing the idle process (something about it calling a primitve vs. waiting on a semaphore), and do not know whether the code below addresses that. I am making this available now vs. waiting until I have time to clean up the loose ends. Bill [*] things should improve this summer, but feel free (encouraged even) to beat me to it. 'From Squeak 3.2 of 11 July 2002 [latest update: #4917] on 2 April 2005 at 3:59:10 pm'! !Object methodsFor: 'as yet unclassified' stamp: 'wks 11/23/2002 22:22'! initialize "Dolphin compatibility" #wks. ! ! !Interpreter methodsFor: 'debug support' stamp: 'wks 4/2/2005 15:48'! printAllStacks "Print all the stacks of all running processes, including those that are currently suspended." | oop proc ctx | self print:'================= ALL STACKS ==================='; cr. proc _ self fetchPointer: ActiveProcessIndex ofObject: self schedulerPointer. self printNameOfClass: (self fetchClassOf: proc) count: 5. self cr. self print:'Active process:'; cr. self printCallStackOf: activeContext. "first the active context" self print:'Other processes:'; cr. oop _ self firstObject. [oop < endOfMemory] whileTrue:[ (self fetchClassOf: oop) == self classSemaphore ifTrue:[ #wks. "wks - do this inside the successful test below self cr." proc _ self fetchPointer: FirstLinkIndex ofObject: oop. "wks - print the cr only if we find a link" proc == self nilObject ifFalse:[ self cr. ]. [proc == self nilObject] whileFalse:[ self printNameOfClass: (self fetchClassOf: proc) count: 5. self cr. ctx _ self fetchPointer: SuspendedContextIndex ofObject: proc. ctx == self nilObject ifFalse:[self printCallStackOf: ctx]. proc _ self fetchPointer: NextLinkIndex ofObject: proc]. ]. oop _ self objectAfter: oop. ].! ! !Interpreter methodsFor: 'debug support' stamp: 'wks 4/2/2005 15:50'! printCallStackOf: aContext | ctxt home methClass methodSel | self inline: false. #wks. "4-05 cleanup for redistribution; any changes here???" "self print:'STACK START'; cr." ctxt _ aContext. [ctxt = nilObj] whileFalse: [ (self fetchClassOf: ctxt) = (self splObj: ClassBlockContext) ifTrue: [ home _ self fetchPointer: HomeIndex ofObject: ctxt ] ifFalse: [ home _ ctxt ]. methClass _ self findClassOfMethod: (self fetchPointer: MethodIndex ofObject: home) forReceiver: (self fetchPointer: ReceiverIndex ofObject: home). methodSel _ self findSelectorOfMethod: (self fetchPointer: MethodIndex ofObject: home) forReceiver: (self fetchPointer: ReceiverIndex ofObject: home). self printNum: ctxt. self print: ' '. ctxt = home ifFalse: [ self print: '[] in ' ]. self printNameOfClass: methClass count: 5. self print: '>>'. self printStringOf: methodSel. self cr. ctxt _ (self fetchPointer: SenderIndex ofObject: ctxt). ]. "self print:'Stack End'; cr." ! ! !Process methodsFor: 'accessing' stamp: 'wks 10/5/2002 16:42'! isActive #wks. ^Processor activeProcess == self ! ! !Process methodsFor: 'debugging' stamp: 'wks 4/2/2005 15:53'! stackOfSize:depthLimit "Answer a call stack for the receiver, active or not, and empty if unavailable." #wks. "4-05 - not sure why I marked this, so beware." ^self == Processor activeProcess ifTrue:[ thisContext stackOfSize:depthLimit. ] ifFalse:[ suspendedContext isNil ifTrue:[ Array new. ] ifFalse:[ suspendedContext stackOfSize:depthLimit. ]. ]. ! ! !SystemDictionary methodsFor: 'memory space' stamp: 'wks 11/9/2002 20:43'! lowSpaceWatcher "Wait until the low space semaphore is signalled, then take appropriate actions." | free | self garbageCollectMost <= self lowSpaceThreshold ifTrue: [ self garbageCollect <= self lowSpaceThreshold ifTrue: [ "free space must be above threshold before starting low space watcher" ^ self beep]]. LowSpaceSemaphore _ Semaphore new. self primLowSpaceSemaphore: LowSpaceSemaphore. self primSignalAtBytesLeft: self lowSpaceThreshold. "enable low space interrupts" LowSpaceSemaphore wait. "wait for a low space condition..." self primSignalAtBytesLeft: 0. "disable low space interrupts" self primLowSpaceSemaphore: nil. LowSpaceProcess _ nil. "Note: user now unprotected until the low space watcher is re-installed" self memoryHogs isEmpty ifFalse: [ free := self bytesLeft. self memoryHogs do: [ :hog | hog freeSomeSpace ]. self bytesLeft > free ifTrue: [ ^ self installLowSpaceWatcher ]]. #wks. "If space is low enough to leave a Morphic delay unsignalled, the debugger probably won't open, and we'd never see it - right???" Smalltalk logError:'FREE SPACE LOW' inContext:thisContext to:'LowSpaceWarning.txt'. Smalltalk isMorphic ifTrue: [CurrentProjectRefactoring currentInterruptName: 'Space is low'] ifFalse: [ScheduledControllers interruptName: 'Space is low']! ! !SystemDictionary methodsFor: 'miscellaneous' stamp: 'wks 11/9/2002 20:20'! logError: errMsg inContext: aContext to: aFilename "Log the error message and a stack trace to the given file." | ff | #wks. "Prevent 'overwrite' of old errors. FileDirectory default deleteFileNamed: aFilename ifAbsent: []." (ff _ FileStream fileNamed: aFilename) ifNil: [^ self "avoid recursive errors"]. #wks. ff setToEnd; cr; nextPutAll:'============================================'; cr. ff nextPutAll: errMsg; cr. aContext errorReportOn: ff. ff close.! ! Wilhelm K. Schwab, Ph.D. University of Florida Department of Anesthesiology PO Box 100254 Gainesville, FL 32610-0254 Email: bills(a)anest4.anest.ufl.edu Tel: (352) 846-1285 FAX: (352) 392-7029

1 0

Re: Squeak --headless - what is the correct signal to stop Squeak?
by Lex Spoon 02 Apr '05

02 Apr '05

David, that is wonderful being able to incercept signals! In general, though, it would be even better if Squeak had a cross-platform way to respond to "user has killed you" events from the OS. It is something to think about for future VM versions.... Lex

1 0

"Squeak Makes you Think and More"
by Brent Vukmer 02 Apr '05

02 Apr '05

I really enjoyed reading this teacher's posts about him and his students working with Squeak: http://teachingtechie.typepad.com/learning/2005/03/have_you_ever_f.html

2 1

Ottawa Carleton Smalltalk Users Group meeting April 20
by David Buck 02 Apr '05

02 Apr '05

Smalltalk Puzzle Night Wednesday, April 20, 2005 7:00 HP5115, Carleton University It's time to combine a little learning and a little fun. At this meeting, we are inviting all our Smalltalk regulars as well as anyone who would like to learn a bit about Smalltalk to bring along your laptops and have some fun. We will start with a quick introduction to the Smalltalk language and environment. We'll help you install Smalltalk on your laptops if needed. Then we'll break up into groups to solve some puzzles using Smalltalk. Get ready to use your thinking caps. Location Details The meeting will be held in Room 5115, Herzberg Laboratories (building 13 on the map (http://www.carleton.ca/cu/campus/)). Pay-parking is available in Lot 1, 2, and parking meters can be found along University Drive. Free parking is available across Bronson Avenue opposite Lot 5. Please RSVP to david(a)simberon.com if you plan to attend.

1 0

Squeak on Handhelds, Status?
by Stefan Krecher 02 Apr '05

02 Apr '05

Hello, the topic "Squeak on Handhelds" has been discussed several times on this list in the past. I'd just like to ask, what's the state of the art, if there are people working with/ on Squeak for Handhelds, an maybe if there's some help needed. Background: Smalltalk is my favorite programming language, and a couple of years ago i worked with Visual Age for Smalltalk. In the past years I played around with Squeak, but i'm far away from beeing a Squeak-expert. At the time i have to implement a UI for a Handheld running WindowsCE (iPaq). The UI shall control actions served by an ejb/ servlet-container (jboss) via HTTP-GET-Requests to a servlet. For the first version of the control-sw we used ewe, a free Java-Implementation for Handhelds (www.ewesoft.org). Unfortunately the UI was not "sexy" enough for our customer. Now I'm trying Squeak, because Morphic as a lot of possibilities to implement graphics/ animation/ sound and other "sexy" things. (btw. "sexy" in this context means: colorful, animated, playful etc.). For the first tests I used the SqueakVM 3.0 compiled for ARM (http://www.is.titech.ac.jp/~ohshima/squeak/WinCE/) and a default 3.2-image. Everything seems to work fine, i built funny buttons, lists and so on. Networking works fine too (WLAN). I'd like to know if there are any pitfalls - for now everything seems to be a bit too easy. Do you have any hints/ tips for stripping down the image? Any ideas on how to implement an automatic installer (an executable that installs the sw on the iPaq)? I heard of the possibility to "close" the image (prevent the user from opening a world-menu etc.). Thanks in advance for any hints, tips or resources. And if there's any help needed, let me know. regards, Stefan -- GPG fingerprint = 2E77 C256 47B1 27B8 AE07 94BE D7A1 2A3C C716 2BE9 GPG public key: http://krecher.de/pubkey.asc ICQ #84067348 "A specter ist haunting the modern world, the specter of crypto anarchy" The Crypto Anarchist Manifesto, Timothy C. May

5 5

Smalltalk/JVM
by Jason Jones 01 Apr '05

01 Apr '05

We were looking for feedback on turning Smalltalk/JVM into an open source project. Smalltalk/JVM is a Smalltalk Compiler that compiles to Java bytecodes and class files. Is this something any of you out there in the programming community are interested in? We would love some feedback. Mission Software, Inc. has created a Smalltalk compiler for the JVM. This compiler allows Smalltalk to run on any JVM. The compiler currently produces 100% Java class files fully compatible with the Sun JVM specification. This allows Smalltalk and Java code to interact seamlessly and allows Smalltalk programs to run anywhere Java runs. a.. Compiler compiles Smalltalk classes into Java .class files a.. Java code callable from Smalltalk code a.. Smalltalk code callable from Java code a.. Java classes subclassing Smalltalk classes a.. Smalltalk classes subclassing Java classes a.. Development is done in a "surrogate" Smalltalk environment a.. Execution in any JVM environment a.. Support for Smalltalk blocks a.. Support for JIT Compilers a.. Ability to write Servlets in Smalltalk a.. Ability to write Applets in Smalltalk (No plug-in required) a.. No special VM required Please email any questions to me at jsj(a)ksc.com -- Jason Jones Knowledge Systems Corporation Smalltalk Industry Council 919.789.8549 x.21 www.ksc.com www.missionsoft.com

3 2

Re: looking for Squeak speaker in Ottawa
by Dean_Swan＠Mitel.COM 01 Apr '05

01 Apr '05

I could also lend a hand on this, but I don't have a lot of spare cycles. Does anybody have a presentation that we could use as a starting point? It would also help if Alan could give us some ideas on what he might like to see covered. I much prefer to have a planned presentation, as some of the sessions can run quite long when it's ad-libbed. Jeff - you can send me private e-mail if you're interested in discussing this further (dean_swan(a)mitel.com). I work in Kanata, and the user's group meetings have been at Carleton U. lately, so we'd have to work out something to get together beforehand. I have PC (WinME) and Mac (Panther) laptops that we could use for the presentation. Also, I haven't moved to anything later than Squeak 3.7. I know it's been popular for some time for people on squeakdev to do everything in 3.8 and 3.9, but I don't like to deal with things like this before the releases are finalized. It might be interesting though to address Ian's and Dan's 64 bit Squeak, since as far as I know it's the first 64 bit Smalltalk. -Dean Jeff Sparkes <jsparkes(a)gmail.com> Sent by: squeak-dev-bounces(a)lists.squeakfoundation.org 03/31/2005 06:14 PM Please respond to Jeff Sparkes; Please respond to The general-purpose Squeak developers list To: The general-purpose Squeak developers list <squeak-dev(a)lists.squeakfoundation.org> cc: Subject: Re: looking for Squeak speaker in Ottawa I can help with this. I've been using Squeak off and on for several years to learn Smalltalk. I doubt I'd be a good speaker, but I can at least help to demo some things. On Fri, 18 Mar 2005 14:45:28 -0500, Alan Knight <knight(a)acm.org> wrote: > Hello All, > I'm one of the organizers of the Ottawa-Carleton (Canada) Smalltalk > User Group, http://smalltalk.ottawa.on.ca and we're thinking that for an > upcoming meeting we'd like to get an update/demo on Squeak. I sent a mail > to our internal list, but didn't get any volunteers. So if there's anyone > in the Ottawa are who'd be interested in doing a talk in the next little > while, please let me know. > > Thanks, > Alan > > -- > Alan Knight [|], Cincom Smalltalk Development > knight(a)acm.org > aknight(a)cincom.com > http://www.cincom.com/smalltalk > > "The Static Typing Philosophy: Make it fast. Make it right. Make it run." - > Niall Ross > >

3 2

Jump to page: