November 2011 - Pharo-dev - lists.pharo.org

Re: [Pharo-project] ProtoObject basicNew :)

by Noury Bouraqadi

On 26 nov. 2011, at 21:11, Mariano Martinez Peck wrote: > > On Sat, Nov 26, 2011 at 5:05 PM, Noury Bouraqadi <bouraqadi(a)gmail.com> wrote: > Stef, > > You can easily obtain the list by using a Ghost proxy. > > Good PhD supervisor that checks student's code :):):) > :-) > That means that whenever the proxy receives #basicInspect I do not forward that to the target but instead, the proxy itself executes #handleBasicInspect: > This let me debug with proxies without problem. > That being said, even if the message is not forwarded, it is shown in the transcript, so for that Stef needs, it should work :) > Yes. Your log says it :-) It's well done :-) > > > 5) analyze the log on the transcript :-) > > > On 25 nov. 2011, at 22:19, StÃ©phane Ducasse wrote: > > > > > On Nov 25, 2011, at 10:09 PM, Toon Verwaest wrote: > > > >> Ok, I guess I was confused by your initial question: "the methods that I should define in a subclass of ProtoObject so that I can use basicNew"... :) > > > > :) > > > > Yes it is more the methods that I should define to have a usable class (browse, doit, inspect it). I will probably have to do it by try and error. > > > > Stef > > > >> > >> On 11/25/2011 10:05 PM, StÃ©phane Ducasse wrote: > >>> I'm confused. > >>> I'm not creating a class but an instance. > >>> if it would be Behavior basicNew then yes I made sure that we could do Behavior new and not crashing the system back in Squeak. > >>> > >>> On Nov 25, 2011, at 9:45 PM, Toon Verwaest wrote: > >>> > >>>> With the magic number I mean the number that defines the object's structure. It's one of the first 3 instance variables of every class, encoding how many instance variables there are; what kind of object it is supposed to be (normal, array, bytes, words...). Every "class" basically needs 3 instance variables. 1 defines the structure, 1 the superclass, 1 the behavior (method dictionary). I don't know the other; but just look at some class to know what the order is. > >>> Yes format, methodDict, superclass (probably in a different order). > >>> And format is ugly because not explicit and coded in several place like instSpec (berk) > >>> > >>>> And yes, if you want to inspect the object, you'll need to support whatever protocol your inspector is using. > >>>> > >>>> The format is formed as follows: > >>>> > >>>> format > >>>> | fieldSize sizeHiBits format | > >>>> fieldSize := self fieldSize + 1. > >>>> sizeHiBits := fieldSize // 64. > >>>> format := sizeHiBits. > >>>> format := (format bitShift: 5) + compactClassIndex. > >>>> format := (format bitShift: 4) + self instanceSpecification. > >>>> format := (format bitShift: 6) + (fieldSize \\ 64). > >>>> format := (format bitShift: 1). > >>>> ^ format > >>>> > >>>> If I read it correctly (just glancing), instance specification specifies if it's arrayed and if it has fields. 0 -> no fields, not arrayed, 1 -> fields, not arrayed, 2 -> arrayed no fields, 3 -> arrayed fields. > >>>> > >>>> > >>>> HTH, > >>>> Toon > >>>> > >>>> On 11/25/2011 09:31 PM, StÃ©phane Ducasse wrote: > >>>>> Apparently I can do a doit I was naively doing an inspect ;) > >>>>> > >>>>> > >>>>>> Don't you just need the basicNew primitive, and make sure that the "magic number" is set in the object that has the basicNew installed to something that makes sense? > >>>>>> > >>>>> I do not know what is the magic number. > >>>>> I guess that my problem was that I have first to define what is the minimal behavior. > >>>>> > >>>>> > >>>>>> If you want to send messages to the object, probably also best install some dictionary as the method dict instvar. > >>>>>> > >>>>> But I guess that ProtoObject has some behavior since I can browse the class with a browser. > >>>>> > >>>>> > >>>>> > >>>>>> Toon > >>>>>> > >>>>>> On 11/25/2011 09:26 PM, StÃ©phane Ducasse wrote: > >>>>>> > >>>>>>> Hi > >>>>>>> I would like to know the methods that I should define in a subclass of ProtoObject so that I can use basicNew. > >>>>>>> Does any of you have an idea? > >>>>>>> > >>>>>>> Stef > >>>>>>> > >>> > >> > >> > > > > > > Noury > -- > http://twitter.com/#!/NouryBouraqadi > > > > > > > > > -- > Mariano > http://marianopeck.wordpress.com > Noury -- http://twitter.com/#!/NouryBouraqadi

Nov. 26, 2011

Re: [Pharo-project] gettext implementation

by Hilaire Fernandes

The problem seems to come from the Seaside-gettext package and not Gettext itself. Is Philip maintaining this package? If so he could integrate the necessary changes. Gettext itself should be working just fine. Hilaire Le 26/11/2011 21:27, CÃ©dric Ternon a Ã©crit : > Hi, > > > I have to recognize I had some difficulties but it works, > > I use the packets : > > Seaside-Gettext-Core-NorbertHartl.10 > Seaside-Pharo-Gettext-Core-pmm.1 > Gettext-HilaireFernandes.22 > Seaside-Gettext-Examples-pmm.5 > > I had to change a variable name in WATranslatedArgumentsFinder > (instance) from RBBrowserEnvironment to BrowserEnvironment. > > Now I can create a .po file with the parts already translated, I'm > sorry, I really begins in smallTalk developpement and the piece of code > I sended to you had some errors : > > translationFor: aKey in: translator > |result| > > translator ifNil: [^'']. > > result := (WALocalizationContext new translate: 'aKey'). > (aKey equal result) ifTrue: [^'']. > ^result. > > obviously, 'aKey' must be aKey > and equal must be = > (aKey = result) ifTrue: [^'']. > > > Now it works and creates me a .po file with the things already > translated (and their translations) and the things I still have to > translate. > > > But I clearly understand that this is no good to use WA(seaside) things > in the gettext(pharo). > > I send you a screenshot a the seaside application and the gettext > working (you have a view in english and a view in french) > > > Thanks for the work you made. > > > Best regards, > > CÃ©dric Ternon, > Brussels. > > > PS : one of my colleague who develops this application with me thinks > it's important to say you we use the pharo 1.3 image of pharo with seaside. > > > On 26 November 2011 20:28, Hilaire Fernandes > <hilaire.fernandes(a)edu.ge.ch > <mailto:hilaire.fernandes@edu.ge.ch>> wrote: > > Le 26/11/2011 17:47, CÃ©dric Ternon a Ã©crit : > > Hi, > > > > > > I'm trying to use the gettexts packages under seaside, > > > > I firstly posted to the seaside mailingList but Philippe Marschall > > recommended me to try to reach mister Hilaire on this ml. > > > > > > I have success to use the gettexts packages creating a .pot file, > > translate the website in poedit, creating the .mo and having the > website > > in different langages. > > > > > So you get the gettext package working within your seaside application? > > Hilaire > > > -- Education 0.2 -- http://blog.ofset.org/hilaire

Nov. 26, 2011

Re: [Pharo-project] ProtoObject basicNew :)

by Schwab,Wilhelm K

Stef, Attached are methods I clearly felt the need to add to ProtoObject. I vaguely recall much trial and even more error :) In my one meaningful (and very little used, due to changing priorities) subclass of ProtoObject, I added #= and #hash. I also customized #printOn:, but that *shouldnt'* be important - caveat emptor :) There are some other finalization related methods, and some "Dolphinization" of events, etc. Again, I don't think those things are crucial outside of my world. Speaking of Dolphin, two things come to mind. First is protocols, which Dolphin makes concrete and editable. One can define the methods that constitute a protocol, and impose protocols on any number of classes. Any class can implement any number of protocols too. The system then complains if changes cause one or more classes to not understand one or more methods in declared protocols. Pretty slick. Another thing is the ability to assign multiple categories to any given method. I think that's something one has to have experienced in order to understand how completely useful it can be. HTH. Bill ________________________________________ From: pharo-project-bounces(a)lists.gforge.inria.fr [pharo-project-bounces(a)lists.gforge.inria.fr] on behalf of StÃ©phane Ducasse [stephane.ducasse(a)inria.fr] Sent: Friday, November 25, 2011 4:19 PM To: Pharo-project(a)lists.gforge.inria.fr Subject: Re: [Pharo-project] ProtoObject basicNew :) On Nov 25, 2011, at 10:09 PM, Toon Verwaest wrote: > Ok, I guess I was confused by your initial question: "the methods that I should define in a subclass of ProtoObject so that I can use basicNew"... :) :) Yes it is more the methods that I should define to have a usable class (browse, doit, inspect it). I will probably have to do it by try and error. Stef > > On 11/25/2011 10:05 PM, StÃ©phane Ducasse wrote: >> I'm confused. >> I'm not creating a class but an instance. >> if it would be Behavior basicNew then yes I made sure that we could do Behavior new and not crashing the system back in Squeak. >> >> On Nov 25, 2011, at 9:45 PM, Toon Verwaest wrote: >> >>> With the magic number I mean the number that defines the object's structure. It's one of the first 3 instance variables of every class, encoding how many instance variables there are; what kind of object it is supposed to be (normal, array, bytes, words...). Every "class" basically needs 3 instance variables. 1 defines the structure, 1 the superclass, 1 the behavior (method dictionary). I don't know the other; but just look at some class to know what the order is. >> Yes format, methodDict, superclass (probably in a different order). >> And format is ugly because not explicit and coded in several place like instSpec (berk) >> >>> And yes, if you want to inspect the object, you'll need to support whatever protocol your inspector is using. >>> >>> The format is formed as follows: >>> >>> format >>> | fieldSize sizeHiBits format | >>> fieldSize := self fieldSize + 1. >>> sizeHiBits := fieldSize // 64. >>> format := sizeHiBits. >>> format := (format bitShift: 5) + compactClassIndex. >>> format := (format bitShift: 4) + self instanceSpecification. >>> format := (format bitShift: 6) + (fieldSize \\ 64). >>> format := (format bitShift: 1). >>> ^ format >>> >>> If I read it correctly (just glancing), instance specification specifies if it's arrayed and if it has fields. 0 -> no fields, not arrayed, 1 -> fields, not arrayed, 2 -> arrayed no fields, 3 -> arrayed fields. >>> >>> >>> HTH, >>> Toon >>> >>> On 11/25/2011 09:31 PM, StÃ©phane Ducasse wrote: >>>> Apparently I can do a doit I was naively doing an inspect ;) >>>> >>>> >>>>> Don't you just need the basicNew primitive, and make sure that the "magic number" is set in the object that has the basicNew installed to something that makes sense? >>>>> >>>> I do not know what is the magic number. >>>> I guess that my problem was that I have first to define what is the minimal behavior. >>>> >>>> >>>>> If you want to send messages to the object, probably also best install some dictionary as the method dict instvar. >>>>> >>>> But I guess that ProtoObject has some behavior since I can browse the class with a browser. >>>> >>>> >>>> >>>>> Toon >>>>> >>>>> On 11/25/2011 09:26 PM, StÃ©phane Ducasse wrote: >>>>> >>>>>> Hi >>>>>> I would like to know the methods that I should define in a subclass of ProtoObject so that I can use basicNew. >>>>>> Does any of you have an idea? >>>>>> >>>>>> Stef >>>>>> >> > >

Nov. 26, 2011

[Pharo-project] the screenshots

by CÃ©dric Ternon

http://pix.toile-libre.org/?img=1322339454.png http://pix.toile-libre.org/?img=1322339771.png

Nov. 26, 2011

Re: [Pharo-project] gettext implementation

by CÃ©dric Ternon

Hi, I have to recognize I had some difficulties but it works, I use the packets : Seaside-Gettext-Core-NorbertHartl.10 Seaside-Pharo-Gettext-Core-pmm.1 Gettext-HilaireFernandes.22 Seaside-Gettext-Examples-pmm.5 I had to change a variable name in WATranslatedArgumentsFinder (instance) from RBBrowserEnvironment to BrowserEnvironment. Now I can create a .po file with the parts already translated, I'm sorry, I really begins in smallTalk developpement and the piece of code I sended to you had some errors : translationFor: aKey in: translator |result| translator ifNil: [^'']. result := (WALocalizationContext new translate: 'aKey'). (aKey equal result) ifTrue: [^'']. ^result. obviously, 'aKey' must be aKey and equal must be = (aKey = result) ifTrue: [^'']. Now it works and creates me a .po file with the things already translated (and their translations) and the things I still have to translate. But I clearly understand that this is no good to use WA(seaside) things in the gettext(pharo). I send you a screenshot a the seaside application and the gettext working (you have a view in english and a view in french) Thanks for the work you made. Best regards, CÃ©dric Ternon, Brussels. PS : one of my colleague who develops this application with me thinks it's important to say you we use the pharo 1.3 image of pharo with seaside. On 26 November 2011 20:28, Hilaire Fernandes <hilaire.fernandes(a)edu.ge.ch>wrote: > Le 26/11/2011 17:47, CÃ©dric Ternon a Ã©crit : > > Hi, > > > > > > I'm trying to use the gettexts packages under seaside, > > > > I firstly posted to the seaside mailingList but Philippe Marschall > > recommended me to try to reach mister Hilaire on this ml. > > > > > > I have success to use the gettexts packages creating a .pot file, > > translate the website in poedit, creating the .mo and having the website > > in different langages. > > > > > So you get the gettext package working within your seaside application? > > Hilaire > > >

Nov. 26, 2011

Re: [Pharo-project] ProtoObject basicNew :)

by Mariano Martinez Peck

On Sat, Nov 26, 2011 at 5:05 PM, Noury Bouraqadi <bouraqadi(a)gmail.com>wrote: > Stef, > > You can easily obtain the list by using a Ghost proxy. > Good PhD supervisor that checks student's code :):):) > 1) Install the package : > Gofer new > squeaksource: 'Marea'; > package: 'GhostProxies'; > package: 'GhostProxiesTests'; > load. > > 2) uncomment the body of method SimpleForwarderHandler>>log: aString > Transcript show: aString; cr. > > 3) open a transcript > > 4) create a proxy on any object and play with it > (ObjectProxy proxyFor: (Object new)) inspect > Be careful that by default there are some debugging methods that I DO NOT forward but rather do something. Check for example the method: debuggingMessagesToHandle | dict | dict := Dictionary new. dict at: #basicInspect put:#handleBasicInspect:. dict at: #inspect put:#handleInspect:. dict at: #inspectorClass put:#handleInspectorClass:. dict at: #defaultLabelForInspector put:#handleDefaultLabelForInspector:. dict at: #printStringLimitedTo: put: #handlePrintStringLimitedTo:. dict at: #printString put: #handlePrintString:. dict at: #longPrintOn:limitedTo:indent: put: #handleLongPrintOnLimitedToIndent:. ^ dict That means that whenever the proxy receives #basicInspect I do not forward that to the target but instead, the proxy itself executes #handleBasicInspect: This let me debug with proxies without problem. That being said, even if the message is not forwarded, it is shown in the transcript, so for that Stef needs, it should work :) > > 5) analyze the log on the transcript :-) > > > On 25 nov. 2011, at 22:19, StÃ©phane Ducasse wrote: > > > > > On Nov 25, 2011, at 10:09 PM, Toon Verwaest wrote: > > > >> Ok, I guess I was confused by your initial question: "the methods that > I should define in a subclass of ProtoObject so that I can use basicNew"... > :) > > > > :) > > > > Yes it is more the methods that I should define to have a usable class > (browse, doit, inspect it). I will probably have to do it by try and error. > > > > Stef > > > >> > >> On 11/25/2011 10:05 PM, StÃ©phane Ducasse wrote: > >>> I'm confused. > >>> I'm not creating a class but an instance. > >>> if it would be Behavior basicNew then yes I made sure that we could do > Behavior new and not crashing the system back in Squeak. > >>> > >>> On Nov 25, 2011, at 9:45 PM, Toon Verwaest wrote: > >>> > >>>> With the magic number I mean the number that defines the object's > structure. It's one of the first 3 instance variables of every class, > encoding how many instance variables there are; what kind of object it is > supposed to be (normal, array, bytes, words...). Every "class" basically > needs 3 instance variables. 1 defines the structure, 1 the superclass, 1 > the behavior (method dictionary). I don't know the other; but just look at > some class to know what the order is. > >>> Yes format, methodDict, superclass (probably in a different order). > >>> And format is ugly because not explicit and coded in several place > like instSpec (berk) > >>> > >>>> And yes, if you want to inspect the object, you'll need to support > whatever protocol your inspector is using. > >>>> > >>>> The format is formed as follows: > >>>> > >>>> format > >>>> | fieldSize sizeHiBits format | > >>>> fieldSize := self fieldSize + 1. > >>>> sizeHiBits := fieldSize // 64. > >>>> format := sizeHiBits. > >>>> format := (format bitShift: 5) + compactClassIndex. > >>>> format := (format bitShift: 4) + self instanceSpecification. > >>>> format := (format bitShift: 6) + (fieldSize \\ 64). > >>>> format := (format bitShift: 1). > >>>> ^ format > >>>> > >>>> If I read it correctly (just glancing), instance specification > specifies if it's arrayed and if it has fields. 0 -> no fields, not > arrayed, 1 -> fields, not arrayed, 2 -> arrayed no fields, 3 -> arrayed > fields. > >>>> > >>>> > >>>> HTH, > >>>> Toon > >>>> > >>>> On 11/25/2011 09:31 PM, StÃ©phane Ducasse wrote: > >>>>> Apparently I can do a doit I was naively doing an inspect ;) > >>>>> > >>>>> > >>>>>> Don't you just need the basicNew primitive, and make sure that the > "magic number" is set in the object that has the basicNew installed to > something that makes sense? > >>>>>> > >>>>> I do not know what is the magic number. > >>>>> I guess that my problem was that I have first to define what is the > minimal behavior. > >>>>> > >>>>> > >>>>>> If you want to send messages to the object, probably also best > install some dictionary as the method dict instvar. > >>>>>> > >>>>> But I guess that ProtoObject has some behavior since I can browse > the class with a browser. > >>>>> > >>>>> > >>>>> > >>>>>> Toon > >>>>>> > >>>>>> On 11/25/2011 09:26 PM, StÃ©phane Ducasse wrote: > >>>>>> > >>>>>>> Hi > >>>>>>> I would like to know the methods that I should define in a > subclass of ProtoObject so that I can use basicNew. > >>>>>>> Does any of you have an idea? > >>>>>>> > >>>>>>> Stef > >>>>>>> > >>> > >> > >> > > > > > > Noury > -- > http://twitter.com/#!/NouryBouraqadi > > > > > > -- Mariano http://marianopeck.wordpress.com

Nov. 26, 2011

[Pharo-project] [ANN] First try: Fuel for Squeak

by Mariano Martinez Peck

Hi guys. After a couple of efforts from Martin Dias and I, we have a "working" Fuel serializer for Squeak. I used the quotes because all I could do is to run tests. Fuel is a binary serializer and you can find all the information in http://rmod.lille.inria.fr/web/pier/software/Fuel This is a project sponsored by ESUG SummerTalk and mainly developed by Martin Dias. I only tried in a Squeak 4.2 image. To install it: (Installer monticello http: 'http://ss3.gemstone.com/ss') project: 'Fuel'; install: 'ConfigurationOfFuel'. ((Smalltalk at: #ConfigurationOfFuel) project version: '1.8-baseline') load. Most tests are green. There are only 2 categories of tests failing: 1) those under the category 'FuelTests-InternalMetalevel'. This is because Squeak has removed the traits there are in Pharo and hence most tests fails. However, this stuff of "InternalMetalevel" is only used if you want to serialize classes/traits completely. Most of the cases, the user just wants to serialize a class/trait as global (just serialize its name) and then during materialization it just searches the class. Of course, in the image where you materialize, such classes have to be present. 2) some tests that are long and it seems in squeak there is a timeout that make those tests to fail. As always, we appreciate any feedback. Cheers -- Mariano http://marianopeck.wordpress.com

Nov. 26, 2011

Re: [Pharo-project] ProtoObject basicNew :)

by Noury Bouraqadi

Stef, You can easily obtain the list by using a Ghost proxy. 1) Install the package : Gofer new squeaksource: 'Marea'; package: 'GhostProxies'; package: 'GhostProxiesTests'; load. 2) uncomment the body of method SimpleForwarderHandler>>log: aString Transcript show: aString; cr. 3) open a transcript 4) create a proxy on any object and play with it (ObjectProxy proxyFor: (Object new)) inspect 5) analyze the log on the transcript :-) On 25 nov. 2011, at 22:19, StÃ©phane Ducasse wrote: > > On Nov 25, 2011, at 10:09 PM, Toon Verwaest wrote: > >> Ok, I guess I was confused by your initial question: "the methods that I should define in a subclass of ProtoObject so that I can use basicNew"... :) > > :) > > Yes it is more the methods that I should define to have a usable class (browse, doit, inspect it). I will probably have to do it by try and error. > > Stef > >> >> On 11/25/2011 10:05 PM, StÃ©phane Ducasse wrote: >>> I'm confused. >>> I'm not creating a class but an instance. >>> if it would be Behavior basicNew then yes I made sure that we could do Behavior new and not crashing the system back in Squeak. >>> >>> On Nov 25, 2011, at 9:45 PM, Toon Verwaest wrote: >>> >>>> With the magic number I mean the number that defines the object's structure. It's one of the first 3 instance variables of every class, encoding how many instance variables there are; what kind of object it is supposed to be (normal, array, bytes, words...). Every "class" basically needs 3 instance variables. 1 defines the structure, 1 the superclass, 1 the behavior (method dictionary). I don't know the other; but just look at some class to know what the order is. >>> Yes format, methodDict, superclass (probably in a different order). >>> And format is ugly because not explicit and coded in several place like instSpec (berk) >>> >>>> And yes, if you want to inspect the object, you'll need to support whatever protocol your inspector is using. >>>> >>>> The format is formed as follows: >>>> >>>> format >>>> | fieldSize sizeHiBits format | >>>> fieldSize := self fieldSize + 1. >>>> sizeHiBits := fieldSize // 64. >>>> format := sizeHiBits. >>>> format := (format bitShift: 5) + compactClassIndex. >>>> format := (format bitShift: 4) + self instanceSpecification. >>>> format := (format bitShift: 6) + (fieldSize \\ 64). >>>> format := (format bitShift: 1). >>>> ^ format >>>> >>>> If I read it correctly (just glancing), instance specification specifies if it's arrayed and if it has fields. 0 -> no fields, not arrayed, 1 -> fields, not arrayed, 2 -> arrayed no fields, 3 -> arrayed fields. >>>> >>>> >>>> HTH, >>>> Toon >>>> >>>> On 11/25/2011 09:31 PM, StÃ©phane Ducasse wrote: >>>>> Apparently I can do a doit I was naively doing an inspect ;) >>>>> >>>>> >>>>>> Don't you just need the basicNew primitive, and make sure that the "magic number" is set in the object that has the basicNew installed to something that makes sense? >>>>>> >>>>> I do not know what is the magic number. >>>>> I guess that my problem was that I have first to define what is the minimal behavior. >>>>> >>>>> >>>>>> If you want to send messages to the object, probably also best install some dictionary as the method dict instvar. >>>>>> >>>>> But I guess that ProtoObject has some behavior since I can browse the class with a browser. >>>>> >>>>> >>>>> >>>>>> Toon >>>>>> >>>>>> On 11/25/2011 09:26 PM, StÃ©phane Ducasse wrote: >>>>>> >>>>>>> Hi >>>>>>> I would like to know the methods that I should define in a subclass of ProtoObject so that I can use basicNew. >>>>>>> Does any of you have an idea? >>>>>>> >>>>>>> Stef >>>>>>> >>> >> >> > > Noury -- http://twitter.com/#!/NouryBouraqadi

Nov. 26, 2011

Re: [Pharo-project] gettext implementation

by Hilaire Fernandes

Le 26/11/2011 17:47, CÃ©dric Ternon a Ã©crit : > Hi, > > > I'm trying to use the gettexts packages under seaside, > > I firstly posted to the seaside mailingList but Philippe Marschall > recommended me to try to reach mister Hilaire on this ml. > > > I have success to use the gettexts packages creating a .pot file, > translate the website in poedit, creating the .mo and having the website > in different langages. > So you get the gettext package working within your seaside application? Hilaire

Nov. 26, 2011

Re: [Pharo-project] Morphic replacement

by Douglas Brebner

On 25/11/2011 10:00, Fernando Olivero wrote: Sounds good except maybe > The world has a mouse and a keyboard, which take care of events > announced by the world. Does that mean you're not thinking of support for multitouch?

Nov. 26, 2011