December 2010 - Pharo-dev - lists.pharo.org

[Pharo-project] super initialize is not a good idea for Behaviors ; )
by Mariano Martinez Peck Dec. 14, 2010

Dec. 14, 2010

Hi folks. After a couple of hours of trying to find a problem, I finally come to the cause: using super initialize is not always good. When I implement #initialize, I ALWAYS do "super initialize" in the first line of such method. Today, I found a problem while doing this in class side #initialize. I have a particular class and I did: MyClass class >> initialize super initialize. "My own specific code to run" But when sending MyClass initialize, I discovered that the "super initialize" did something I wasnt aware: Behavior >> initialize "moved here from the class side's #new" super initialize. superclass := Object. "no longer sending any messages, some of them crash the VM" methodDict := self emptyMethodDictionary. format := Object format in my case MyClass has subclass from ProtoObject, not Object. And of course, I got an empy method dictionary...etc. So...I just wanted to share my problem to avoid other's one. Cheers Mariano

3 3

[Pharo-project] Issue 3425 in pharo: Environment related class behavior
by pharo＠googlecode.com Dec. 14, 2010

Dec. 14, 2010

Status: FixedWaitingToBePharoed Owner: stephane.ducasse Labels: Milestone-1.3 New issue 3425 by stephane.ducasse: Environment related class behavior http://code.google.com/p/pharo/issues/detail?id=3425 A new version of Kernel was added to project The Inbox: http://source.squeak.org/inbox/Kernel-spd.444.mcz ==================== Summary ==================== Name: Kernel-spd.444 Author: spd Time: 4 December 2010, 1:30:09.425 pm UUID: 571950b9-31b4-4f85-b27f-e4b7e6044b27 Ancestors: Kernel-ar.443 * fixed Class>>canFindWithoutEnvironment: (it was actually checking the environment, making the behavior like bindingOf:) and added test to KernelTests * changed Class>>declare: and Class>>addClassVarName: - no longer check the environment for conflicts - conflict error message clarified * made conflict error in #addClassVarName: resumable to match the behavior of #declare: n.b. no conflicts with trunk as of 12/4/2010 =============== Diff against Kernel-ar.443 =============== Item was removed: - ----- Method: BlockContext>>valueWithEnoughArguments: (in category 'evaluating') ----- - valueWithEnoughArguments: anArray - "call me with enough arguments from anArray" - | args | - (anArray size == self numArgs) - ifTrue: [ ^self valueWithArguments: anArray ]. - - args := Array new: self numArgs. - args replaceFrom: 1 - to: (anArray size min: args size) - with: anArray - startingAt: 1. - - ^ self valueWithArguments: args! Item was changed: ----- Method: Class>>addClassVarName: (in category 'class variables') ----- addClassVarName: aString "Add the argument, aString, as a class variable of the receiver. Signal an error if the first character of aString is not capitalized, or if it is already a variable named in the class." | symbol oldState | oldState := self copy. aString first canBeGlobalVarInitial ifFalse: [^self error: aString, ' class variable name should be capitalized; proceed to include anyway.']. symbol := aString asSymbol. self withAllSubclasses do: [:subclass | + (self canFindWithoutEnvironment: symbol) ifTrue: [ + (DuplicateVariableError new) + superclass: superclass; "fake!!!!!!" + variable: aString; + signal: aString, ' is already defined']]. - (subclass bindingOf: symbol) ifNotNil:[ - ^ self error: aString - , ' is already used as a variable name in class ' - , subclass name]]. classPool == nil ifTrue: [classPool := Dictionary new]. (classPool includesKey: symbol) ifFalse: ["Pick up any refs in Undeclared" classPool declare: symbol from: Undeclared. SystemChangeNotifier uniqueInstance classDefinitionChangedFrom: oldState to: self]! Item was changed: ----- Method: Class>>canFindWithoutEnvironment: (in category 'compiling') ----- canFindWithoutEnvironment: varName "This method is used for analysis of system structure -- see senders." "Look up varName, in the context of the receiver. Return true if it can be found without using the declared environment." "First look in classVar dictionary." (self classPool bindingOf: varName) ifNotNil:[^true]. "Next look in shared pools." self sharedPools do:[:pool | (pool bindingOf: varName) ifNotNil:[^true]. ]. "Finally look higher up the superclass chain and fail at the end." superclass == nil ifTrue: [^ false] + ifFalse: [^ superclass canFindWithoutEnvironment: varName]. - ifFalse: [^ (superclass bindingOf: varName) notNil]. ! Item was changed: ----- Method: Class>>declare: (in category 'initialize-release') ----- declare: varString "Declare class variables common to all instances. Answer whether recompilation is advisable." | newVars conflicts | + newVars := (Scanner new scanFieldNames: varString) collect: [:x | x asSymbol]. newVars do: [:var | var first canBeGlobalVarInitial ifFalse: [self error: var, ' class variable name should be capitalized; proceed to include anyway.']]. conflicts := false. classPool == nil ifFalse: [(classPool keys reject: [:x | newVars includes: x]) do: [:var | self removeClassVarName: var]]. (newVars reject: [:var | self classPool includesKey: var]) do: [:var | "adding" "check if new vars defined elsewhere" + (self canFindWithoutEnvironment: var) ifTrue: [ - (self bindingOf: var) ifNotNil:[ (DuplicateVariableError new) superclass: superclass; "fake!!!!!!" variable: var; + signal: var, ' is already defined'. - signal: var , ' is defined elsewhere'. conflicts := true]]. newVars size > 0 ifTrue: [classPool := self classPool. "in case it was nil" newVars do: [:var | classPool declare: var from: Undeclared]]. ^conflicts! Item was removed: - ----- Method: MessageSend>>valueWithEnoughArguments: (in category 'evaluating') ----- - valueWithEnoughArguments: anArray - "call the selector with enough arguments from arguments and anArray" - | args | - args := Array new: selector numArgs. - args replaceFrom: 1 - to: (arguments size min: args size) - with: arguments - startingAt: 1. - args size > arguments size ifTrue: [ - args replaceFrom: arguments size + 1 - to: (arguments size + anArray size min: args size) - with: anArray - startingAt: 1. - ]. - ^ receiver perform: selector withArguments: args! Item was removed: - ----- Method: Object>>actionsWithReceiver:forEvent: (in category 'events') ----- - actionsWithReceiver: anObject forEvent: anEventSelector - - ^(self actionSequenceForEvent: anEventSelector) - select: [:anAction | anAction receiver == anObject ]! Item was removed: - ----- Method: Object>>assert:description: (in category 'error handling') ----- - assert: aBlock description: aString - "Throw an assertion error if aBlock does not evaluates to true." - - aBlock value ifFalse: [AssertionFailure signal: aString ]! Item was removed: - ----- Method: Object>>assert:descriptionBlock: (in category 'error handling') ----- - assert: aBlock descriptionBlock: descriptionBlock - "Throw an assertion error if aBlock does not evaluate to true." - - aBlock value ifFalse: [AssertionFailure signal: descriptionBlock value asString ]! Item was removed: - ----- Method: Object>>perform:withEnoughArguments: (in category 'message handling') ----- - perform: selector withEnoughArguments: anArray - "Send the selector, aSymbol, to the receiver with arguments in argArray. - Only use enough arguments for the arity of the selector; supply nils for missing ones." - | numArgs args | - numArgs := selector numArgs. - anArray size == numArgs - ifTrue: [ ^self perform: selector withArguments: anArray asArray ]. - - args := Array new: numArgs. - args replaceFrom: 1 - to: (anArray size min: args size) - with: anArray - startingAt: 1. - - ^ self perform: selector withArguments: args! Item was removed: - ----- Method: Object>>renameActionsWithReceiver:forEvent:toEvent: (in category 'events') ----- - renameActionsWithReceiver: anObject forEvent: anEventSelector toEvent: newEvent - - | oldActions newActions | - oldActions := Set new. - newActions := Set new. - (self actionSequenceForEvent: anEventSelector) do: [ :action | - action receiver == anObject - ifTrue: [ oldActions add: anObject ] - ifFalse: [ newActions add: anObject ]]. - self setActionSequence: (ActionSequence withAll: newActions) forEvent: anEventSelector. - oldActions do: [ :act | self when: newEvent evaluate: act ].! Item was changed: ----- Method: StringHolder class>>open (in category 'instance creation') ----- open + ^ (Smalltalk at: #Workspace ifAbsent:[self]) new openLabel: 'Workspace' - (Smalltalk at: #Workspace ifAbsent:[self]) new openLabel: 'Workspace' "Not to be confused with our own class var 'Workspace'"!

1 1

Re: [Pharo-project] Issue 3237 in pharo: If MessageTally is not in memory
by pharo＠googlecode.com Dec. 14, 2010

Dec. 14, 2010

Updates: Status: Closed Comment #1 on issue 3237 by marcus.denker: If MessageTally is not in memory http://code.google.com/p/pharo/issues/detail?id=3237 So the code in Pharo is correct. There is no "ifPresentAndInMemory" anymore in Pharo

1 0

[Pharo-project] About using pragma for help like in setting
by Torsten Bergmann Dec. 14, 2010

Dec. 14, 2010

Stef wrote on Mon Dec 13 10:02:56 (see [1]): >Now what I want is that people can type >... > <help> > I already gave exactly such an example to Lukas already back in May. Again - PLEASE have a look at it (just load the code in [2] and run HelpBrowser openOn: PragmaHelp The code is small and simple: - beside the existing help builders it just introduces a "PragmaHelpBuilder" that uses the pragma <help> to collect the help data - a Foo and Bar class is included to demonstrate how you define help with this new builder (similar to the way Stef/Lukas wanted) As I said adding "pragma help" is just a new "help builder" to define and one can easily add it. Think of another scenario: maybe one doesnt care about the "class to book mapping" or a "pragma approach". He wants to define his help contents completely different in external text files. Same pattern: he introduces a new HelpBuilderFromExternalTextFile that converts his custom (external) format into help topics. Nothing more. So the real work is in thinking about/defining how and in what format you want to write the help contents itself. The problem: there are MANY ways to define the contents - and as Stef showed us with the TWO "pragma" descriptions examples in his mail [2] that we are currently NOT SURE about the "pragma description" format to use. But if we want pragma help in the help system we have to find ONE and it should be simple and fit best. To say it in simple words: if we want to use pragmas we have to find a way to use pragmas to define nestable and extensible books (with pages and page order). I once requested this from Lukas but regarding extensible books his simple answer was: "I would not make the books extensible themselves, that complicates things and is rarely needed in practice." This answer is not satisfying to me since extensible books ARE EASILY POSSIBLE with the "class to book mapping" that the help system already have. And I use them. We can use "pragmas" but we should have the same possibilities and not use pragmas because they are "modern" or they are used by the Settings framework. And I agree with him that it may be hard to do with pragmas ... especially since I already experimented on this with pragmas But maybe we all can find a solution here together. In the end I switched to the "use classes and the class hierarchy" approach since it is more natural (at least to me) and easy to use/understand. Stef further wrote: >and that if the help is loaded we can see it in help. A registration mechanism to mark which custom help builders are included in the standard browser is on my todo list. Currently it is only one and hard coded (see HelpBrowser>>defaultRoot) If the above pragma help builder is in good shape we can add it too as one of the defaults - but a registration mechanism is better in the long run. But again: The problem I have with this simple pragma approach foo: myBuidler <help> myBuilder ... is that with this you can easily describe a page with a title and text - but it is hard to describe a complete book with structure and even extensible books. So we have to clearly think about the builder part of this approach to keep the help contents definitions simple. And this is what I requested you all to do if we want to move on here... Just use my simple code example [2] to experiment and find the "pragma" description format we want to introduce. Stef may answer: use the one that I showed you. But if one uses pragmas we may want to describe the whole book. Not just a simple <help> tag to describe a single page with a title! Thx T. [1] http://lists.gforge.inria.fr/pipermail/pharo-project/2010-December/037545.h… [2] http://lists.gforge.inria.fr/pipermail/pharo-project/attachments/20100505/c… -- GMX DSL Doppel-Flat ab 19,99 €/mtl.! Jetzt auch mit gratis Notebook-Flat! http://portal.gmx.net/de/go/dsl

2 1

Re: [Pharo-project] Issue 2322 in pharo: Cannot type ' (apostrophe) next to an existing character
by pharo＠googlecode.com Dec. 14, 2010

Dec. 14, 2010

Updates: Labels: -Milestone-1.0.1-DevImage Comment #13 on issue 2322 by marcus.denker: Cannot type ' (apostrophe) next to an existing character http://code.google.com/p/pharo/issues/detail?id=2322 (No comment was entered for this change.)

1 0

Re: [Pharo-project] Issue 2451 in pharo: Make Algernon work in Pharo 1.1.
by pharo＠googlecode.com Dec. 14, 2010

Dec. 14, 2010

Updates: Labels: -Milestone-1.1-DevImage Comment #6 on issue 2451 by marcus.denker: Make Algernon work in Pharo 1.1. http://code.google.com/p/pharo/issues/detail?id=2451 (No comment was entered for this change.)

1 0

[Pharo-project] Fwd: [Pharo-users] "Making ToolSet more flexible", any taker?
by Igor Stasenko Dec. 14, 2010

Dec. 14, 2010

On 14 December 2010 09:21, Mariano Martinez Peck <marianopeck(a)gmail.com> wrote: > Hi guys. I am out of time, but if someone wants to do a nice and not so > difficult contribution to Pharo, is to improve the ToolSet support. > > Issue is here: http://code.google.com/p/pharo/issues/detail?id=1915 > > And discussion is here: > http://forum.world.st/Questions-about-ToolSet-td1460544.html > Actually it's a much more older Â - 2 years http://lists.gforge.inria.fr/pipermail/pharo-project/2008-October/002872.ht… Here is what i proposed to this: http://www.mail-archive.com/pharo-project@lists.gforge.inria.fr/msg02869.ht… > Cheers > > Mariano > -- Best regards, Igor Stasenko AKA sig.

1 0

Re: [Pharo-project] Ring - model infrastructure for Pharo
by Marcus Denker Dec. 14, 2010

Dec. 14, 2010

On Dec 10, 2010, at 2:33 PM, Veronica Isabel Uquillas Gomez wrote: > Dear all, > > I am currently working on the Ring, an unifying and foundational model infrastructure for Pharo. > The goals are: > - Provide a common API at structural and runtime level > - Allow tools to interact and integrate directly with the host environment (Pharo) > - Support history analysis > > I started comparing the APIs of RB, MethodReference, Pseudo classes, MC, Smalltalk itself and Ginsu, as a basic to build the Ring. > So far I have implemented the main classes including the ones that should replace MethodReference and the Pseudo classes. > > An unified API will imply changes in most of the ones mentioned above (as most of them are non-polymorphic). > As a first step, I would like to have your opinion about the proposal for replacing MethodReference (attached file). > #stringVersionDefault has just one sender (#markMatchingClasses in SelectorBrowser). Do we need that? In general, #stringVersion in MethodReference seems to be strange... stringVersion: 'Class>>selector:' format why is there an iVar for this? Can't this just use the selector and class an build it? Marcus -- Marcus Denker -- http://www.marcusdenker.de INRIA Lille -- Nord Europe. Team RMoD.

4 3

[Pharo-project] "Making ToolSet more flexible", any taker?
by Mariano Martinez Peck Dec. 14, 2010

Dec. 14, 2010

Hi guys. I am out of time, but if someone wants to do a nice and not so difficult contribution to Pharo, is to improve the ToolSet support. Issue is here: http://code.google.com/p/pharo/issues/detail?id=1915 And discussion is here: http://forum.world.st/Questions-about-ToolSet-td1460544.html Cheers Mariano

1 0

Re: [Pharo-project] TODO for 1.2...
by Marcus Denker Dec. 14, 2010

Dec. 14, 2010

On Dec 14, 2010, at 2:35 AM, Guillermo Polito wrote: > Marcus, > > Here we have a "blue" build: https://pharo-ic.lille.inria.fr/hudson/job/Pharo%201.2/11/ > How did you manage that? :P > Did you remove the test run from the build? > yes... I was wondering where exactly the process was failing (but then, one can see that fom the console)... Next step: remove the deprecated calls (or change deprecation warning setting) then I will turn on the tests again. -- Marcus Denker -- http://www.marcusdenker.de INRIA Lille -- Nord Europe. Team RMoD.

1 0