- Pharo-dev - lists.pharo.org

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

Re: [Pharo-project] Fuel bug: infinite recursion in "analyzer when:substituteBy:"

by Mariano Martinez Peck

For this "problem" at least it is documented in http://rmod.lille.inria.fr/web/pier/software/Fuel/Version1.7/Hooks in the section "Substitutes on Serialization" On Fri, Nov 25, 2011 at 8:46 PM, Yanni Chiu <yanni(a)rogers.com> wrote: > On 25/11/11 5:26 PM, Martin Dias wrote: > >> Hi Yanni >> >> This is a feature or a known issue, I am not sure yet! >> > > It makes sense why it is the way it is (as you've noted below). Once you > know this is what's happening. But, when first encountered, it's a bit > scary to see your image not respond. I'm not sure whether it's a bug or > feature, either. However, I suspect any infinite recursion should be > classified as a bug. > > > In serialization, when you substitute an object by another, such >> substitution(*) is then serialized using the same rules, so you might >> need the guard condition to avoid the infinite loop. >> >> This was the simpler implementation, but maybe it's undesirable or it's >> too error prone. Do you think we should change this? >> > > Probably, the right answer is to allow a choice of whether or not the > substituted object(s) are subject to further substitution. And, no matter > which choice is made, no infinite loop should occur -- a tall order. > > > *: note the substitution can be not just an object but a graph of >> objects, some of them able to be substituted too. >> > > Good to know this. > > > -- Mariano http://marianopeck.wordpress.com

Nov. 26, 2011

Re: [Pharo-project] Added to OmniBrowser info in online book

by Alexandre Bergel

The chapter is indeed more actualized than the papers Alexandre Le 26 nov. 2011 Ã 13:49, "Sean P. DeNigris" <sean(a)clipperadams.com> a Ã©crit : > > Mariano Martinez Peck wrote >> >> A link to the PBE chapter? >> > > Good idea. Done. > > -- > View this message in context: http://forum.world.st/Added-to-OmniBrowser-info-in-online-book-tp4110664p41… > Sent from the Pharo Smalltalk mailing list archive at Nabble.com. >

Nov. 26, 2011

Re: [Pharo-project] Morphic replacement

by Alexandre Bergel

I guess the absolute coordinates are kept in a cache or something? This is probably necessary when you want to redraw inner widget without redisplaying the nesting ones Alexandre Le 25 nov. 2011 Ã 13:22, Fernando Olivero <fernando.olivero(a)usi.ch> a Ã©crit : > Glad to be able to say: yes! > > Each object speaks in its own coordinate system, its parent already > internalized the point. > > Suppose we have a click, announced by the mouse at a given point, > given in Display coordinates. > > From now on we go, from the display, to the view , to the scene, to > the shape, and on each step the parent deals with a point in its own > coordinate system, and passes on the internalized click to its > children. > > Views present a scene, consisting of shapes, at a given viewport of the display. > > Makes it a lot easier when you are programming the interaction of > custom widgets, to always deal with local coordinates, and greatly > simplifies the event handling code as well. > > > On Fri, Nov 25, 2011 at 5:00 PM, StÃ©phane Ducasse > <stephane.ducasse(a)inria.fr> wrote: >> did you kill the absolute coordinates? >> >> say yes please :) >> >> Stef >> >> On Nov 25, 2011, at 11:00 AM, Fernando Olivero wrote: >> >>> Dear all, >>> >>> I've been working on a hole replacement of Morphic myself, because if >>> was tired of fixing Morphic (or porting CUIS Morphic into Pharo), and >>> found much more productive the approach of rewriting it from scratch. >> >> excellent! >> >>> >>> In my expierience, i was 100% more productive by not trying to be >>> backward compatible with Morphic, Polymorph , etc. >>> >>> I found it easier to come up with a new set of widgets (EntryField, >>> AutoCompletionEntryField, Button, Label, Column, Row, >>> ScrollableColumns) from scratch. They are all fairly tested also. >>> >>> Maybe it's time to move on, and develop from the ground up a >>> Morphic-like successor, instead of fixing the current Morphic. There's >>> no problem with preserving the current UI for all the apps that are >>> out there, and provide a cleaner, simpler UI framework for the new >>> apps. >>> >>> Regarding the new "morphic", I only assume that i have a scalable >>> canvas, and the vm's notifying me of the user events. On top of that i >>> recreated a World, that can have multiple views. There's only one >>> World at a given time. Views are presented within the world in a given >>> viewport, and present a scene. Scene are were the shapes live in. >>> No World as a subclass of PasteUpMorph, no globals such as World, >>> ActiveWorld, Hand. ActiveHand, etcs.. >>> >>> The world has a mouse and a keyboard, which take care of events >>> announced by the world. For example GMouse>>down: aDisplayPosition, >>> would take care of transforming the display position into a world >>> position, into a view position, into a scene position, and finally >>> notifying the proper shape under the mouse that it has been clicked at >>> a certain local position. GShape>>clickedAt: , >>> GShape>>secondaryClickedAt:, etc.. >>> No Events hierarchy, or the awful switch currently in HandMorph ( >>> generateMouseEvent: or generateKeyboardEvent:). >>> >>> I have a "Gaucho" system working now, with GShape (instead of >>> Morphs), using the simple new layouts developed by Juan in CUIS, >>> uniform usage of keybindings across the whole GShape hierarchy, >>> amongst other goodies. It's still tailored to be a Morphic replacement >>> for Gaucho, so first i need to polish a few things prior to releasing >>> it. >>> >>> For example, i'm going to ditch my current scalable canvas >>> implementation, in favor of Athens. >>> >>> Hope that very soon i can provide more details, but my plans are to >>> release it as part of the next Gaucho version, to provide a working >>> example. >>> >>> Fernando >>> >> >> >

Nov. 26, 2011

Re: [Pharo-project] Added to OmniBrowser info in online book

by Sean P. DeNigris

Mariano Martinez Peck wrote > > A link to the PBE chapter? > Good idea. Done. -- View this message in context: http://forum.world.st/Added-to-OmniBrowser-info-in-online-book-tp4110664p41… Sent from the Pharo Smalltalk mailing list archive at Nabble.com.

Nov. 26, 2011

[Pharo-project] gettext implementation

by CÃ©dric Ternon

Hi, I'm trying to use the gettexts packages under seaside, I firstly posted to the seaside mailingList but Philippe Marschallrecommended 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. The problem occurs when I try to create a .po file to continue an incomplete translation, The previous messages follows : Hi, I'm trying to use the gettexts packages for seaSide. I finally achieve to create a .pot file, translate it with poedit and make the website translated. But now, I'm trying to create a .po file, that is a .pot file with the strings already translated and the strings remaining to translate. (a .pot file beeing an empty template to begin without any translations) We followed this tutorial : http://code.google.com/p/seaside/wiki/Gettext The versions used are : Seaside-Gettext-Core-NorbertHartl.10 Seaside-Pharo-Gettext-Core-pmm.1 Name: Gettext-HilaireFernandes.22 When I simply try : GetTextExporter new exportTranslator: NaturalLanguageTranslator current. I got : http://pix.toile-libre.org/?img=1322152172.jpg I've got the same result with any GetTextTranslator I have tryed to look for an implementors of the method "translations" but no one is implementing it, is a pkged missing? I've also tryed to replace it with 'class translators' but this sound wrong and give me another unrelated error. Thanks in advance for your help, ----- Philippe answer : It looks like you're using GetTextExporter instead of WAGetTextExporter. Can you load Seaside-Gettext-Examples and evaluate WAGettextExample export and see if that works? Cheers Philippe ----- Hello, Thanks to help. It works fine but it creates a .pot file. My goal is to create a .po file : when a site is already half translated, I'd like to create a .po file a translator can use in poedit. So the translator only has to translate the half not yet translated (and maybe bring some little corrections to the half already translated.) WAGettextExample export calls WAGetTextExporter new exportTemplate wich calls self exportTranslator: nil and we go there exportTranslator: translator "Export translation files. the file extention is 'po', or 'pot' if translator is nil " when the translator is nil, everything works well. when I try with a translator it breaks here : translationFor: aKey in: translator translator ifNil: [^'']. translator translations at: aKey ifPresent: [:s | ^s]. "If we have old camelCase translations, make space-separated words" translator translations at: aKey toCamelCase ifPresent: [:s | (s includes: Character space) ifTrue: [^s]. ^s fromCamelCase]. ^'' --------------------------- Now, I have tried to change the code of translationFor: aKey in: translator in this way : translationFor: aKey in: translator |result| translator ifNil: [^'']. result := (WALocalizationContext new translate: 'aKey'). (aKey equal result) ifTrue: [^'']. ^result. That is maybe horrible to use the WA things at this level but I tried to make it works and the only way I achieve to translate things is at the WA level.

Nov. 26, 2011