February 2015 - Pharo-dev - lists.pharo.org

Re: [Pharo-dev] Easy way to write a proxy?

by ClÃ©ment Bera

2015-02-27 14:06 GMT+01:00 Mariano Martinez Peck <marianopeck(a)gmail.com>: > > > On Fri, Feb 27, 2015 at 7:10 AM, Marcus Denker <marcus.denker(a)inria.fr> > wrote: > >> >> For most cases Ghost is the way to go. >>> >>> >> We will add a proxy model (based on Ghost or just ghost as is) to Pharo >> by default in Pharo5. >> Then people can just use it. >> > > And with the lazy become of Spur / new Cog that would be a super great > combination. > > I made some benchs for a paper and it's crazy. Current become has performance proportional to the heap size, and for example it's a 350ms pause for a 500Mb heap on my machine (or 27ms for 35Mb heap). In spur it is 0.2ms independently of the heap size. It goes down to 0.035ms if both objects have the same size in memory. > > -- > Mariano > http://marianopeck.wordpress.com >

Feb. 27, 2015

Re: [Pharo-dev] About slots and syntax

by Damien Pollet

On 27 February 2015 at 16:13, stepharo <stepharo(a)free.fr> wrote: > Are we sure that we want to use that => > I thought about = but this is not an assignement this is a declaration and > normally we use | for declaration. > So what would be other alternatives before the world use => > But | for local variables is syntax, and it's used as a bracket. It's also the logical or operatorâ¦ I'd propose :: because it recalls the :x for block arguments, or -- because it looks like a typographic em-dash: slots: { #local -- ProcessLocalSlot. #foo -- SomeOtherKindOfSlot. By the way, using an operator in this way is a little strange: what you read does not match the usual precedence rules (#default: is not a factory method of ProcessLocalSlot, it's an accessor on the result of #local => ProcessLocalSlot) Then I still do not understand why we have empty classVariables: to me it > does not make sense. I do not get why we like to have empty collections for > the sake of having empty collections. We are trapped in some strange ways > of thinking. > Are there strong arguments against a cascade ? As far as I understand, the only real constraint is that the set of messages that are authorized in class creation syntax should be closed in some way. -- Damien Pollet type less, do more [ | ] http://people.untyped.org/damien.pollet

Feb. 27, 2015

Re: [Pharo-dev] The evil twin in 40516

by Marcus Denker

> On 27 Feb 2015, at 20:26, Marcus Denker <marcus.denker(a)inria.fr> wrote: > > >> On 27 Feb 2015, at 19:48, Thierry Goubier <thierry.goubier(a)gmail.com> wrote: >> >> Hi Markus, >> >> if I understand correctly some of the possibilities: >> >> - I can use RBParseTreeSearcher / RBParseTreeRewriter on the twin AST. >> >> (any example how to install / invalidate the twin AST after an effective RBParseTreeRewriter operation?) >> > Yes, but I would do it on a copyâ¦ that is, do the change in run:with:in: on a copy, generate code from that, > keep the original. > An AST that follows the text is kind of nice for e.g. navigation The reason is that we actually cache the AST if you request it from a CompiledMethod, this means you do not want to change that cached version as e.g. the Browser uses it for Navigation or the âSuggestionsâ menu. Marcus

Feb. 27, 2015

Re: [Pharo-dev] The evil twin in 40516

by Marcus Denker

> On 27 Feb 2015, at 19:48, Thierry Goubier <thierry.goubier(a)gmail.com> wrote: > > Hi Markus, > > if I understand correctly some of the possibilities: > > - I can use RBParseTreeSearcher / RBParseTreeRewriter on the twin AST. > > (any example how to install / invalidate the twin AST after an effective RBParseTreeRewriter operation?) > Yes, but I would do it on a copyâ¦ that is, do the change in run:with:in: on a copy, generate code from that, keep the original. An AST that follows the text is kind of nice for e.g. navigation. > - Any change to the Twin AST is reflected in the method bytecode (as long as one does invalidate). > yes. > This would mean being able to put tracing code with RB patterns, no? And without having to do a recompilation of the traced method? > Yes. At first I want to have something simple: a MOP based on the MetaLinks. But one could do the same with RB transformations: put them as annotation on the method, then on recompile take them into account. Marcus

Feb. 27, 2015

Re: [Pharo-dev] The evil twin in 40516

by stepharo

Hi thierry f this implementation follows persephone the bytecode is the result of the first execution of the AST. So a kind of JIT inside the system. The AST is annotated with properties (link to meta objects that can define the semantics of the annotation). Stef Le 27/2/15 19:48, Thierry Goubier a Ã©crit : > Hi Markus, > > if I understand correctly some of the possibilities: > > - I can use RBParseTreeSearcher / RBParseTreeRewriter on the twin AST. > > (any example how to install / invalidate the twin AST after an > effective RBParseTreeRewriter operation?) > > - Any change to the Twin AST is reflected in the method bytecode (as > long as one does invalidate). > > This would mean being able to put tracing code with RB patterns, no? > And without having to do a recompilation of the traced method? > > Thierry > > Le 27/02/2015 00:07, Marcus Denker a Ã©crit : >> Hi, >> >> In 40516 there is now a strange, but rather powerful mechanism for >> CompiledMethod: the Twin. >> >> What does that do? >> >> you can call on a method #createTwin >> >> (ReflectivityExamples>>#exampleMethod) createTwin. >> >> after this, the CompiledMethod has a high-level representation (the >> AST) attached that itself references the CompiledMethod >> (they form a twin). >> >> (ReflectivityExamples>>#exampleMethod) reflectiveMethod >> >> The fun thing is now that one can install either in the class. Call >> #invalidate to make sure the reflective method is installed. >> >> ReflectiveMethod implements #run:with:in: which calls a compilation >> hook (too re-create from the AST) and then installs that in the >> method dict and then executes the method: >> >> (ReflectivityExamples>>#exampleMethod) createTwin. >> (ReflectivityExamples>>#exampleMethod) invalidate. >> self assert: (ReflectivityExamples>>#exampleMethod) class = >> ReflectiveMethod. >> self assert: ReflectivityExamples new exampleMethod = 5. >> self assert: (ReflectivityExamples>>#exampleMethod) class = >> CompiledMethod. >> >> Which means that this gives us an in-image, on-demand JIT compiler >> AST->Bytecode. In 50 lines of code. >> >> e.g. try on Morph: >> >> Morph methods do: #createTwin. >> Morph methods do: #invalidate. >> >> Counting which twin is installed shows us the working set of Morph: >> >> (Morph methods select: [ :each | each class = CompiledMethod ]) size >> >> some 330 method out of nearly 900â¦. >> >> So what can one do with this? In essence this turns the AST into a >> reflective representation for Methods >> (if you care to set up twin creation + invalidation correctly). >> >> What will this allow us to do? stay tuned... >> >> Marcus >> >> >> >> > > >

Feb. 27, 2015

Re: [Pharo-dev] Easy way to write a proxy?

by stepharo

Le 27/2/15 17:47, Alexandre Bergel a Ã©crit : >> So the idea is that the reference to self in MyProxy becomes the wrapped object using the become. > Yes, for some reason, I did not see this. Well spotted! > I am getting old apparently :-) > >> The main problem of this kind of implementation is that self is bounded to the proxy, so all the internal calls of self are not going to be intercepted and the return self is going to expose the original object. > Yes. Maybe the doesNotUnderstand: can be improved to take this into account. No it is not possible simply (or you have to rename all your methods and recompile all the class/superclass :). > > Cheers, > Alexandre >

Feb. 27, 2015

Re: [Pharo-dev] Slots: and ProcessLocalSlot

by Marcus Denker

>> > > Pardon my ignorance, but can Slot definitions be composed like traits? I was wondering the same already :-) e.g. I want to be able to make any slot a hidden slot without having to subclass it. > Because I can imagine wanting, say ProcessLocal + mmm... Static, without having to combine the two properties in a class with a gnarly inheritance tree, or a factorial number of classes combining different defining traits... > The proof-of-concept for how to realise MetaLinks on Slots is using a Wrapper.. this way every slot can be annotate with a MetaLink. Maybe that can be generalised? Marcus

Feb. 27, 2015

Re: [Pharo-dev] The evil twin in 40516

by Thierry Goubier

Hi Markus, if I understand correctly some of the possibilities: - I can use RBParseTreeSearcher / RBParseTreeRewriter on the twin AST. (any example how to install / invalidate the twin AST after an effective RBParseTreeRewriter operation?) - Any change to the Twin AST is reflected in the method bytecode (as long as one does invalidate). This would mean being able to put tracing code with RB patterns, no? And without having to do a recompilation of the traced method? Thierry Le 27/02/2015 00:07, Marcus Denker a Ã©crit : > Hi, > > In 40516 there is now a strange, but rather powerful mechanism for CompiledMethod: the Twin. > > What does that do? > > you can call on a method #createTwin > > (ReflectivityExamples>>#exampleMethod) createTwin. > > after this, the CompiledMethod has a high-level representation (the AST) attached that itself references the CompiledMethod > (they form a twin). > > (ReflectivityExamples>>#exampleMethod) reflectiveMethod > > The fun thing is now that one can install either in the class. Call #invalidate to make sure the reflective method is installed. > > ReflectiveMethod implements #run:with:in: which calls a compilation hook (too re-create from the AST) and then installs that in the > method dict and then executes the method: > > (ReflectivityExamples>>#exampleMethod) createTwin. > (ReflectivityExamples>>#exampleMethod) invalidate. > self assert: (ReflectivityExamples>>#exampleMethod) class = ReflectiveMethod. > self assert: ReflectivityExamples new exampleMethod = 5. > self assert: (ReflectivityExamples>>#exampleMethod) class = CompiledMethod. > > Which means that this gives us an in-image, on-demand JIT compiler AST->Bytecode. In 50 lines of code. > > e.g. try on Morph: > > Morph methods do: #createTwin. > Morph methods do: #invalidate. > > Counting which twin is installed shows us the working set of Morph: > > (Morph methods select: [ :each | each class = CompiledMethod ]) size > > some 330 method out of nearly 900â¦. > > So what can one do with this? In essence this turns the AST into a reflective representation for Methods > (if you care to set up twin creation + invalidation correctly). > > What will this allow us to do? stay tuned... > > Marcus > > > >

Feb. 27, 2015

Re: [Pharo-dev] Slots: and ProcessLocalSlot

by Marcus Denker

> On 27 Feb 2015, at 17:53, Henrik Johansen <henrik.s.johansen(a)veloxit.no> wrote: > > > > Would something like > slots: {#x . #y . #area -> (Cache: #calculateArea)) } > > Meaning :area is a Slot that holds a cached value, calculated when accessed (by calling calculateArea), and nilled whenever slots accessed by it's calculation method takes a new value. yes, that should be quite easy to build in combination with MetaLinksâ¦ Marcus

Feb. 27, 2015

Re: [Pharo-dev] [Metacello] Getting all working copies from a configuration

by Dale Henrichs

Guillermo, I think that Sean is right, the MetacelloPackageLoadDirective will have a GoferResolvedReference in the externalReference slot for the packages to be loaded and you can send version to that ... I suggest you use the scriptting api and use `fetch` to get what I think you want: Metacello new configuration: 'XXX'; version '1.0'; repository: 'http://.....'; fetch Dale On 02/27/2015 07:14 AM, Sean P. DeNigris wrote: > Guillermo Polito wrote >> I know that I can do a #record to get the list of specs of a particular >> version. > I'm not sure I understand... How does your use case differ from what #record > provides? > > What has always worked for me is "(ConfigurationOfXXX project version: > #'1.0') record loadDirective" > > > > ----- > Cheers, > Sean > -- > View this message in context: http://forum.world.st/Metacello-Getting-all-working-copies-from-a-configura… > Sent from the Pharo Smalltalk Developers mailing list archive at Nabble.com. >

Feb. 27, 2015