Squeak-dev October 2005

squeak-dev@lists.squeakfoundation.org

180 participants
296 discussions

Mailing list management
by Dave Mason 29 Oct '05

29 Oct '05

Is there any chance of tagging the articles for the list with a [Squeak-dev] in the subject line, like the Seaside list does? Perhaps this is a frequently debated point here, but I know for me, and think for many others, that it would make it a lot easier to manage the 300 emails I get every day and to filter out the spam. Thanks ../Dave

8 11

Team Report: Files
by Cees De Groot 29 Oct '05

29 Oct '05

Hi All, The Files Team consists currently of yours truly, the team leader, and Ken Causey but he probably subscribed just out of curiosity ;) Short-term, that's not a big problem - I'm harvesting stuff on Mantis, which isn't too much, and generally just am available to help the 3.9a team if they need specific fixes. However, many of you may agree that there are some things that could be done better in the Files code. We have the code now for a whole stack of platforms, we can look at some other implementations, like VW's, and maybe it is a nice idea if people who are interested in reworking this library subscribe to the team and start voicing their ideas/opinions. Personally, I think that there's enough on the plate for 3.9a so for that version I am not planning to press for big chances, but you know how it is with package rewrites: they always turn out to be more involved than you think :). Therefore, thinking about where we want to go now, if only to preserve your ideas for posterity in the mailing list archives, might be a good idea. Regards, Cees

1 0

0000854: Long Delay schedules a deferred image crash
by stéphane ducasse 28 Oct '05

28 Oct '05

hi all I need feedback on http://bugs.impara.de/view.php?id=854 can the people involved in the discussion let me know what is the cs that should be considered at the end? Tim, alexander, david, lex.... Stef

6 13

About exception propagation with ensure:
by stéphane ducasse 28 Oct '05

28 Oct '05

Hi guys We got a really strange problem in the classBuilder code. Here is the story Context ======= When the method name: className inEnvironment: env subclassOf: newSuper type: type instanceVariableNames: instVarString classVariableNames: classVarString poolDictionaries: poolString category: category unsafe: unsafe is redefined as follow. It works and fixes some bugs because Behavior was not flushed correctly. All the tests are green for example ClassBuilderFormatTests>> testByteVariableSubclass in particular self should:[self makeIVarsSubclassOf: baseClass] raise: Error. works ie the exception is raised. name: className inEnvironment: env subclassOf: newSuper type: type instanceVariableNames: instVarString classVariableNames: classVarString poolDictionaries: poolString category: category unsafe: unsafe "Define a new class in the given environment. If unsafe is true do not run any validation checks. This facility is provided to implement important system changes." | oldClass newClass organization instVars classVars force needNew oldCategory copyOfOldClass newCategory | environ _ env. instVars _ Scanner new scanFieldNames: instVarString. classVars _ (Scanner new scanFieldNames: classVarString) collect: [:x | x asSymbol]. "Validate the proposed name" unsafe ifFalse:[(self validateClassName: className) ifFalse: [^nil]]. oldClass _ env at: className ifAbsent:[nil]. oldClass isBehavior ifFalse: [oldClass _ nil] "Already checked in #validateClassName:" ifTrue: [ copyOfOldClass _ oldClass copy. copyOfOldClass superclass addSubclass: copyOfOldClass]. [unsafe ifFalse:[ "Run validation checks so we know that we have a good chance for recompilation" (self validateSuperclass: newSuper forSubclass: oldClass) ifFalse:[^nil]. (self validateInstvars: instVars from: oldClass forSuper: newSuper) ifFalse:[^nil]. (self validateClassvars: classVars from: oldClass forSuper: newSuper) ifFalse:[^nil]. (self validateSubclassFormat: type from: oldClass forSuper: newSuper extra: instVars size) ifFalse:[^nil]]. "self halt: 'after validate'." "See if we need a new subclass" needNew _ self needsSubclassOf: newSuper type: type instanceVariables: instVars from: oldClass. needNew == nil ifTrue:[^nil]. "some error" (needNew and:[unsafe not]) ifTrue:[ "Make sure we don't redefine any dangerous classes" (self tooDangerousClasses includes: oldClass name) ifTrue:[ self error: oldClass name, ' cannot be changed'. ]. "Check if the receiver should not be redefined" (oldClass ~~ nil and:[oldClass shouldNotBeRedefined]) ifTrue:[ self notify: oldClass name asText allBold, ' should not be redefined. \Proceed to store over it.' withCRs]]. "self halt." needNew ifTrue:[ "Create the new class" newClass _ self newSubclassOf: newSuper type: type instanceVariables: instVars from: oldClass. newClass == nil ifTrue:[^nil]. "Some error" newClass setName: className. ] ifFalse:[ "Reuse the old class" newClass _ oldClass. ]. "Install the class variables and pool dictionaries... " force _ (newClass declare: classVarString) | (newClass sharing: poolString). "... classify ..." newCategory _ category asSymbol. organization _ environ ifNotNil:[environ organization]. oldClass isNil ifFalse: [oldCategory _ (organization categoryOfElement: oldClass name) asSymbol]. organization classify: newClass name under: newCategory. newClass environment: environ. "... recompile ..." newClass _ self recompile: force from: oldClass to: newClass mutate: false. "... export if not yet done ..." (environ at: newClass name ifAbsent:[nil]) == newClass ifFalse:[ [environ at: newClass name put: newClass] on: AttemptToWriteReadOnlyGlobal do:[:ex| ex resume: true]. Smalltalk flushClassNameCache. ]. newClass doneCompiling. "... notify interested clients ..." oldClass isNil ifTrue: [ SystemChangeNotifier uniqueInstance classAdded: newClass inCategory: newCategory. ^ newClass]. SystemChangeNotifier uniqueInstance classDefinitionChangedFrom: copyOfOldClass to: newClass. newCategory ~= oldCategory ifTrue: [SystemChangeNotifier uniqueInstance class: newClass recategorizedFrom: oldCategory to: category]. ] ensure: [copyOfOldClass ifNotNil: [copyOfOldClass superclass removeSubclass: copyOfOldClass]. Behavior flushObsoleteSubclasses.]. "apparently when the following expression is inside the ensure block the exception is not propagated. One tiny examples we could not reproduce this behavior." ^newClass Problem: ====== Now if you move the ^newClass into the ensure block as ensure: [copyOfOldClass ifNotNil: [copyOfOldClass superclass removeSubclass: copyOfOldClass]. Behavior flushObsoleteSubclasses. ^newClass ]. The test does not work self should:[self makeIVarsSubclassOf: baseClass] raise: Error. works ie the exception is not raised. So we check that ensure: propagate exception as shown below testsOne "self debug: #testsOne" self should: [ [1 foo] ensure: [^1]] raise: MessageNotUnderstood testsTwo "self debug: #testsTwo" self should: [ [1 error: 'hello'] ensure: [^1]] raise: Error. self should: [ [1+2 ] ensure: [1 error: 'hello']] raise: Error So we do not really understand. Now our fixes in the classBuilder works but we are perplexed and not satisfied not to understand. Stef and adrian

2 3

Strange behavior
by stéphane ducasse 28 Oct '05

28 Oct '05

Hi all I have encountered a really strange problem. With adrian we took the same image 6693 - loaded the same scripts (I did the scripts to update the image) I always ended up with conflicts. He never :) I did that removing all the packages in the package cache, adrian too. I tried a lot of tests and I cannot understand. Next week I will do a controlled experience with adrian next week. Marcus took the same scripts and succeeded :) argh..... I really do not understand. I do not see why this is different. I will try to see if the difference come from the vm but we are all on mac. Now I understand why I was saying that I get errors and that avi and adrian could not see it. Does any of you experienced somehting like that? Stef

12 52

SUnit?
by stéphane ducasse 28 Oct '05

28 Oct '05

Hi I'm running tests using SUnit TestRunner (I saw this morning a totally new test runner application from lukas and this is really cool, lukas has around 1400 tests for maggritte and SW2, he should publish his test runner soon I hope) when I get a failure and click on it I end up on this method openDebuggerOnFailingTestMethod "SUnit has halted one step in front of the failing test method. Step over the 'self halt' and send into 'self perform: testSelector' to see the failure from the beginning" self halt; performTest Instead of the place where the failure occured. Is it on purpose? Why can I get the stack directly? Stef

4 5

a good help to the community
by stephane ducasse 28 Oct '05

28 Oct '05

Hi all I'm trying to find a bug in the classbuilder (glooops) and I realized that the diffs algorithm is not really good. So if someone wants to improve it, everybody would benefit from that. Now I'm comparing on paper.... Stef

3 2

VMware Player
by Brad Fuller 28 Oct '05

28 Oct '05

> > >Now that VMware has made their VMware Player available for free, everyone can >soon enjoy the advantages and frustrations of "living in Images..." > :) > > How does this work, Darius? The VMware site says you can run any virtual machine. Then, it immediately says "Run virtual machines created by VMware Workstation, GSX Server or ESX Server" So, obviously not any virtual machine. I would imagine that another layer would have to be created or a current VM would have to be re-configured so that it can call VMware Player to operate. brad

4 5

Re: fixing underscores
by Bill Schwab 28 Oct '05

28 Oct '05

Bert, Stef, =============================== >>> Please count me among those who want to see _ when I type _. >>> >> >> I perfectly agree with that. >> > > Did anyone actually try my package? =============================== I didn't dare :) It seemed like something aimed at the streams, not for use by the little guy. =============================== Sounds good to me. Add it in the pipeline of stuff I will push in 3.9a =============================== Stef, THANKS!!!! Bill 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

Squeak mascot??
by Michael van der Gulik 28 Oct '05

28 Oct '05

http://www.allplush.com/storefront/StuffedAnimals/1956.html Hmm... he looks familier :-). Very cute. Mikevdg.

1 0

← Newer
1
2
3
4
5
6
7
8
...
30
Older →

Jump to page: