March 2012 - Pharo-dev - lists.pharo.org

[Pharo-project] why improving is the only way

by StÃ©phane Ducasse

I'm fighting with timeStamp and old code checking size of timeStamp and this is why we should improve the infrastructure and in particular code management. Stef timeStamp "Answer the authoring time-stamp for the given method, retrieved from the sources or changes file. Answer the empty string if no time stamp is available." "(CompiledMethod compiledMethodAt: #timeStamp) timeStamp" | file preamble stamp tokens tokenCount | self fileIndex = 0 ifTrue: [^ String new]. "no source pointer for this method" file := SourceFiles at: self fileIndex. file ifNil: [^ String new]. "sources file not available" "file does not exist happens in secure mode" file := [file readOnlyCopy] on: FileDoesNotExistException do:[:ex| ^ String new]. preamble := self getPreambleFrom: file at: (0 max: self filePosition - 3). stamp := String new. tokens := (preamble findString: 'methodsFor:' startingAt: 1) > 0 ifTrue: [Scanner new scanTokens: preamble] ifFalse: [Array new "ie cant be back ref"]. (((tokenCount := tokens size) between: 7 and: 8) and: [(tokens at: tokenCount - 5) = #methodsFor:]) ifTrue: [(tokens at: tokenCount - 3) = #stamp: ifTrue: ["New format gives change stamp and unified prior pointer" stamp := tokens at: tokenCount - 2]]. ((tokenCount between: 5 and: 6) and: [(tokens at: tokenCount - 3) = #methodsFor:]) ifTrue: [(tokens at: tokenCount - 1) = #stamp: ifTrue: ["New format gives change stamp and unified prior pointer" stamp := tokens at: tokenCount]]. file close. ^ stamp there are a couple of method with such logic. Stef

March 6, 2012

Re: [Pharo-project] Colored Stdout/Stderr logging with headless images

by Igor Stasenko

On 6 March 2012 13:46, Stefan Marr <smalltalk(a)stefan-marr.de> wrote: > > On 06 Mar 2012, at 13:18, Camillo Bruni wrote: > >> actually its a bit of a hack :D (well the implementation not) >> >> but the problem is, that we should properly detect if the terminal supports colors or not ;) otherwise you might get some ugly escape sequences :D > > Do you know the tput tool? > > I use it in my makefiles like this: > Â Â Â Â START_BOLD=$(shell tput bold) > Â Â Â Â STOP_BOLD=$(shell tput sgr0) > > It gives you the supported escape sequence for a give style based on a database of terminals and your current environment. > yeah, but that means that VM (or some plugin/prim) is needed to detect that it outputs to terminal, and not redirected to file or pipes output to another process. > Best regards > Stefan > > > -- > Stefan Marr > Software Languages Lab > Vrije Universiteit Brussel > Pleinlaan 2 / B-1050 Brussels / Belgium > http://soft.vub.ac.be/~smarr > Phone: +32 2 629 2974 > Fax: Â +32 2 629 3525 > > -- Best regards, Igor Stasenko.

March 6, 2012

Re: [Pharo-project] Colored Stdout/Stderr logging with headless images

by Stefan Marr

On 06 Mar 2012, at 13:18, Camillo Bruni wrote: > actually its a bit of a hack :D (well the implementation not) > > but the problem is, that we should properly detect if the terminal supports colors or not ;) otherwise you might get some ugly escape sequences :D Do you know the tput tool? I use it in my makefiles like this: START_BOLD=$(shell tput bold) STOP_BOLD=$(shell tput sgr0) It gives you the supported escape sequence for a give style based on a database of terminals and your current environment. Best regards Stefan -- Stefan Marr Software Languages Lab Vrije Universiteit Brussel Pleinlaan 2 / B-1050 Brussels / Belgium http://soft.vub.ac.be/~smarr Phone: +32 2 629 2974 Fax: +32 2 629 3525

March 6, 2012

[Pharo-project] fix GoferOperationTest failure

by Pavel Krivanek

There is a new failure related to Camillo's changes to UIManagers. http://code.google.com/p/pharo/issues/detail?id=5439 I proposed a fix. Cheers, -- Pavel

March 6, 2012

Re: [Pharo-project] in 14379...

by Camillo Bruni

right that should work... forgot to update my fix for that one.. cami On 2012-03-06, at 12:25, Pavel Krivanek wrote: > what about to fix it this way: > > isPostMortem > "return whether we're inspecting a frozen exception without a process attached" > |selectedContext suspendedContext | > selectedContext := self selectedContext. > suspendedContext := interruptedProcess suspendedContext. > suspendedContext ifNil: [ ^ false ]. > (suspendedContext == selectedContext) > ifTrue: [ ^ false ]. > ^ (suspendedContext findContextSuchThat: [:c | c sender == > selectedContext]) isNil > > -- Pavel > > On Tue, Mar 6, 2012 at 12:08 PM, StÃ©phane Ducasse > <stephane.ducasse(a)inria.fr> wrote: >> Hi guys >> >> when I try to load >> >> MCSqueaksourceRepository >> location: 'http://squeaksource.com/Games' >> user: '' >> password: '' >> >> load MorphicGames-md.1.mcz >> >> I get a DNU >> >> isPostMortem >> "return whether we're inspecting a frozen exception without a process attached" >> |selectedContext| >> selectedContext := self selectedContext. >> (interruptedProcess suspendedContext == selectedContext) >> ifTrue: [ ^ false ]. >> ^ (interruptedProcess suspendedContext findContextSuchThat: [:c | c sender == selectedContext]) isNil >> >> ^^^^^^ DNU findContextSuchThat: >> >> >> interruptedProcess suspendedContext is a MethodContext >> >> Now I do not understand since find findContextSuchThat: is defined in ContextPart the super class of MethodContext. >> >> Stef >

March 6, 2012

Re: [Pharo-project] Colored Stdout/Stderr logging with headless images

by Camillo Bruni

actually its a bit of a hack :D (well the implementation not) but the problem is, that we should properly detect if the terminal supports colors or not ;) otherwise you might get some ugly escape sequences :D best cami On 2012-03-06, at 11:44, Sven Van Caekenberghe wrote: > This is so cool and beautiful, this new addition where errors with headless images are printed directly to stout or stderr with coloring ! > > I just saw it the first time today. Really helpful. Thanks a lot Camillo ! > > <Screen Shot 2012-03-06 at 11.39.04.png> > Sven > > > > -- > Sven Van Caekenberghe > http://stfx.eu > Smalltalk is the Red Pill > > > >

March 6, 2012

Re: [Pharo-project] in 14379...

by Henrik Johansen

On Mar 6, 2012, at 12:29 PM, Pavel Krivanek wrote: > On Tue, Mar 6, 2012 at 12:23 PM, Marcus Denker <marcus.denker(a)inria.fr> wrote: >> >> On Mar 6, 2012, at 12:08 PM, StÃ©phane Ducasse wrote: >> >>> Hi guys >>> >>> when I try to load >>> >>> MCSqueaksourceRepository >>> location: 'http://squeaksource.com/Games' >>> user: '' >>> password: '' >>> >>> load MorphicGames-md.1.mcz >>> >>> I get a DNU >>> >>> isPostMortem >>> "return whether we're inspecting a frozen exception without a process attached" >>> |selectedContext| >>> selectedContext := self selectedContext. >>> (interruptedProcess suspendedContext == selectedContext) >>> ifTrue: [ ^ false ]. >>> ^ (interruptedProcess suspendedContext findContextSuchThat: [:c | c sender == selectedContext]) isNil >>> >>> ^^^^^^ DNU findContextSuchThat: >>> >>> >>> interruptedProcess suspendedContext is a MethodContext >>> >>> Now I do not understand since find findContextSuchThat: is defined in ContextPart the super class of MethodContext. >>> >> >> This is related to the startup changes we integrated yesterday. >> >> (there is no change that has no side effect...) >> >> Marcus >> > > It does every Warning exception, for example: > > self notify: 'hello'. > > Strange that it was not reported by tests... > > -- Pavel Sounds a tiny bit like an error handler kept alive past its intended lifespanâ¦ + handling Error rather than UnhandledError, as used to be a problem in Announcement delivery too. ;) Cheers, Henry

March 6, 2012

Re: [Pharo-project] in 14379...

by StÃ©phane Ducasse

>> >> > > This is related to the startup changes we integrated yesterday. > > (there is no change that has no side effectâ¦) sure change is life and life is dangerous we all die at the end. So if camillo/igor can have a look, it would be good. > One artist said: "life is a killer"

March 6, 2012

Re: [Pharo-project] The event handling mechanism implemented in Gaucho

by StÃ©phane Ducasse

Thanks. Stef! On Mar 6, 2012, at 12:37 PM, Fernando Olivero wrote: > Here is the code, the categories which include the mentioned classes. > > I'm preparing a one-click image of Gaucho for the Mac, so you can play > with it also. > And as the official release of the tool also! > > > > On Tue, Mar 6, 2012 at 12:33 PM, fernando.olivero(a)usi.ch > <fernando.olivero(a)usi.ch> wrote: >> Hi, Stef and Esteban, i'm sending an overview of the event handling >> mechanism of Gaucho. So we can discuss the additions of multitouch >> events and the design as well. >> >> >> Key classes: GUserInterface, GPeripheralDevice, GDisplay, GView and >> the new event model. >> >> GUserInterface: Collaborates with a sensor, and double dispatches the >> event to the proper peripheral device. >> >> GMouse,GKeyboard: a peripheral device that distinguishes the system >> event, and sends the user interface specific reactions: such as >> >> #mouse: aGMouse entered: aGShape >> #keystroke: aGKeystroke from: aGKeyboard >> >> No reification of the events, just message passing amongst the user >> interface, the peripherals, the display, the views, and the shapes. >> >> EXAMPLE >> -------------------------------------------------------- >> EVENTS: from the user to the shape >> -------------------------------------------------------- >> 1)the user clicks on the mouse >> 2) a system event is announced >> 3) the user interface dispatches it to the attached mouse >> 4) the mouse detects that its a click to a particular shape on a view >> of the display and informs the user interface >> 5) the user interface informs the display of the click >> >> ------------------------------------------------------------------------ >> VIEW: shape reacting to events and redrawing >> -------------------------------------------------------------------------- >> 6) the display informs the active view >> 7) the active view informs the shape >> 8) the shape reacts to the click. >> 9)the display asks the shape wether it should redraw, and informs the >> damage recorder. Gaucho implements a visual garbage collector scheme, >> so the code is not polluted with #changed ,similar to eliminating the >> need to send #alloc and #dealloc on memory garbage collector. But this >> is surely for another thread. >> > <GauchoUI-Core.st><GauchoUI-Peripherals-Mouse.st><GauchoUI-Peripherals.st>

March 6, 2012

Re: [Pharo-project] looking for Morphic-Games package

by StÃ©phane Ducasse

thanks I was fixing the := and found some problems in fixUnderscore so I fixed it also. and I have to fix the selectionMenu so thanks I will have a look. This is for a deliverable for a project on cohesion metrics computation and I like the package. Stef On Mar 6, 2012, at 10:52 AM, Edgar J. De Cleene wrote: > > > > On 3/6/12 7:12 AM, "StÃ©phane Ducasse" <stephane.ducasse(a)inria.fr> wrote: > >> Hi guys >> >> I'm looking for the package containing tetris, free cell, â¦ >> does any of you know where it is? >> >> Stef > > > Just for you I remade for Pharo > File in for the repository. > Load MorphicGames-edc.3.mcz > Evaluate (Tetris new) openInWorld > > Maybe not all Games work , but is a start. > > What about a FunPharo ? > Like the idea or no old stuff into Pharo ? > > By the way gracias a Mariano y a Esteban. > > I using Nautilus 1.4 and last CogVM for doing this. > > Edgar > > <AgregaRepoLadrillos.st>

March 6, 2012