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

Re: [Pharo-project] PBenchmark as a general framework for benchmarking in Smalltalk (commit access)

by Mariano Martinez Peck

On Mon, Mar 14, 2011 at 3:06 PM, Henrik Sperre Johansen < henrik.s.johansen(a)veloxit.no> wrote: > On 14.03.2011 10:19, Stefan Marr wrote: > >> Hi Toon: >> >> On 14 Mar 2011, at 09:52, Toon Verwaest wrote: >> >>> Since the general idea is to write benchmarks like unit tests, how about >>>> 'SMark'/'SBench' instead of 'SUnit'? >>>> >>> And he immediately calls it after himself! S<tefan>mar<r>k ;) >>> >> Rather farfetched no? :-P >> >> The SMark was inspired by names like 3DMark, FutureMark, CoreMark, >> SYSmark, WebMark, MobileMark. >> > I was actually about to suggest we call it SMarrk earlier today, but that > would probably be too obvious :D > > Anyways, in regards to using PharoBenchmarks/PBenchmark/etc, +100 to > Stefan's reply. > It is important to keep the framework in a separate repository. (whose name > does not imply it is specific to any dialect) > Mariano: Don't mix/confuse the framework with the uses of it. > > Thanks Henry, that was exactly what I musunderstood :) > Cheers, > Henry > > >

March 14, 2011

Re: [Pharo-project] Jtalk, a Smalltalk for web developers

by Mariano Martinez Peck

2011/3/14 Janko MivÅ¡ek <janko.mivsek(a)eranova.si> > Hi all, > > To add with Aida/Web plans how we will use Jtalk: I namely see very nice > oportunity to cross the border from server to the client (web browser) > with Smalltalk! That is, we can build web apps in pure Smalltalk and run > them on the client, unchanged. Even more, we can partition the app and > decide where to run, part on server, part on client, whole on server, > whole on client... And this can be done at runtime! > > So, Nicolas work actually opens a whole new horizon for Smalltalk on the > web development front. That's why Jtalk will already be part of next > version of Aida, for experimentation purposes for start. > > Sounds a lot to GWT but for Smalltalk :) Cool! > So, thanks a lot, Nico! > > Best regards > Janko > > > On 14. 03. 2011 19:36, Nicolas Petton wrote: > > Hi, > > > > I would like to share a project I'm working on on my spare time: Jtalk > > Smalltalk. > > > > http://nicolaspetton.github.com/jtalk > > https://github.com/NicolasPetton/jtalk > > > > Jtalk is an implementation of the Smalltalk language that compiles into > > JavaScript. > > > > Some features: > > - it is written in itself (including the parser/compiler) > > - it is self-contained > > - it compiles into efficient JS code > > - it uses the Squeak chunk format > > - Pharo is considered as the reference implementation > > > > I think Jtalk can be compared to CoffeeScript[1], Objective-J[2] or > > Clamato[3], from which it reuses some ideas and code. > > > > Jtalk includes an IDE with a class browser, transcript and workspace, an > > HTML canvas similar to Seaside and a jQuery binding. > > > > It is still a young piece of code, and some important features are still > > missing/incomplete. > > > > Cheers, > > Nicolas Petton > > > > [1] http://jashkenas.github.com/coffee-script/ > > [2] http://cappuccino.org/ > > [3] http://clamato.net > > > > > > > > -- > Janko MivÅ¡ek > Svetovalec za informatiko > Eranova d.o.o. > Ljubljana, Slovenija > www.eranova.si > tel: 01 514 22 55 > faks: 01 514 22 56 > gsm: 031 674 565 > >

March 14, 2011

[Pharo-project] Question about the design of the setting browser

by StÃ©phane Ducasse

Hi alain I was reading the settingbrowser code and I have a question: why the settingbrowser inherits from MorphTreeModel. Why it does not inherits from Object or Model and has different widgets that may or not inherits from MorphTreeModel. To me it fits better the composition. Stef

March 14, 2011

Re: [Pharo-project] PBenchmark as a general framework for benchmarking in Smalltalk (commit access)

by Alexandre Bergel

>> The class contains many methods having a name that begins with 'metric'. Some students worked on a kind of Hudson that exploit these information. You can have a look at it: >> http://kimen.dcc.uchile.cl:8008/VerMon01/faces/project_details.xhtml?id=1 > Ok, that is about code metrics. Slightly different from the performance aspect I am interest in, but the solution seems to follow a similar principle. It evaluates the execution time of any piece of code. It simply uses [ ... ] timeToRun. No big deal. But this is enough in many cases. > Ehm, do you suggest to have a common meta-framework for SUnit, you metric analysis/reporting, and benchmarking? Not sure whether thats worth the effort. No. I said that we should get inspired from the SUnit framework to develop a benchmark framework. This is what Lukas, Jorge, Oscar and I agree upon. I can send you things to read if you wish. > For now I will concentrate on getting a decent job done to have common benchmarking framework. Where it goes from there, we shall see. Yes! Alexandre > > > -- > 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 > > -- _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: Alexandre Bergel http://www.bergel.eu ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.

March 14, 2011

Re: [Pharo-project] Jtalk, a Smalltalk for web developers

by Janko MivÅ¡ek

Hi all, To add with Aida/Web plans how we will use Jtalk: I namely see very nice oportunity to cross the border from server to the client (web browser) with Smalltalk! That is, we can build web apps in pure Smalltalk and run them on the client, unchanged. Even more, we can partition the app and decide where to run, part on server, part on client, whole on server, whole on client... And this can be done at runtime! So, Nicolas work actually opens a whole new horizon for Smalltalk on the web development front. That's why Jtalk will already be part of next version of Aida, for experimentation purposes for start. So, thanks a lot, Nico! Best regards Janko On 14. 03. 2011 19:36, Nicolas Petton wrote: > Hi, > > I would like to share a project I'm working on on my spare time: Jtalk > Smalltalk. > > http://nicolaspetton.github.com/jtalk > https://github.com/NicolasPetton/jtalk > > Jtalk is an implementation of the Smalltalk language that compiles into > JavaScript. > > Some features: > - it is written in itself (including the parser/compiler) > - it is self-contained > - it compiles into efficient JS code > - it uses the Squeak chunk format > - Pharo is considered as the reference implementation > > I think Jtalk can be compared to CoffeeScript[1], Objective-J[2] or > Clamato[3], from which it reuses some ideas and code. > > Jtalk includes an IDE with a class browser, transcript and workspace, an > HTML canvas similar to Seaside and a jQuery binding. > > It is still a young piece of code, and some important features are still > missing/incomplete. > > Cheers, > Nicolas Petton > > [1] http://jashkenas.github.com/coffee-script/ > [2] http://cappuccino.org/ > [3] http://clamato.net > > > -- Janko MivÅ¡ek Svetovalec za informatiko Eranova d.o.o. Ljubljana, Slovenija www.eranova.si tel: 01 514 22 55 faks: 01 514 22 56 gsm: 031 674 565

March 14, 2011

Re: [Pharo-project] PBenchmark as a general framework for benchmarking in Smalltalk (commit access)

by Stefan Marr

Hi Alexandre: On 14 Mar 2011, at 20:31, Alexandre Bergel wrote: > The class contains many methods having a name that begins with 'metric'. Some students worked on a kind of Hudson that exploit these information. You can have a look at it: > http://kimen.dcc.uchile.cl:8008/VerMon01/faces/project_details.xhtml?id=1 Ok, that is about code metrics. Slightly different from the performance aspect I am interest in, but the solution seems to follow a similar principle. > It would be cool to have a standard for this. As we have for SUnit. I think that what I did goes in the right direction, but the code is not as clean as it should. It would be cool to have a report generation, which could be .pdf, sending emails, ... Ehm, do you suggest to have a common meta-framework for SUnit, you metric analysis/reporting, and benchmarking? Not sure whether thats worth the effort. For now I will concentrate on getting a decent job done to have common benchmarking framework. Where it goes from there, we shall see. 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 14, 2011

Re: [Pharo-project] the glamorous toolkit project

by Tudor Girba

Hi Mariano, Thanks for the suggestions. As I said, you are welcome to participate. It does not have to be much. You can just spend a bit of time to learn Glamour and then play 15 minutes per day :). Cheers, Doru On 14 Mar 2011, at 19:34, Mariano Martinez Peck wrote: > Hi Tudor? do you want to succeed? Do a kind of OB but with multiple selection :) > I would like to remove, for example, several methods together... > > cheers > > mariano > > On Sun, Mar 13, 2011 at 10:45 PM, Tudor Girba <tudor.girba(a)gmail.com> wrote: > Thanks, Esteban. > > Jorge did a nice job at improving it to handle special objects. There are still a couple of situations left, but it is already reliable. > > Cheers, > Doru > > > On 13 Mar 2011, at 19:33, Esteban Lorenzano wrote: > > > wow.... GTInspector is really great :) > > gamorous tools are becoming very interesting... keep going that way! > > > > I do believe in the near future GT can replace OB as our default IDE, if we continue improving like this :) > > > > cheers, > > Esteban > > > > El 13/03/2011, a las 10:34a.m., Marcus Denker escribiÃ³: > > > >> > >> On Mar 13, 2011, at 2:05 PM, Tudor Girba wrote: > >> > >>> I think that this is a good idea. Having the Core tools better is definitely a good thing. > >>> > >> > >> One thing to keep in mind is that we need (ass soon as possible) to have *one* set > >> of tools... with OB we have seen that we have just not the manpower to in parallel > >> maintain two sets of tools (Core vs. Dev). > >> > >> The notion of "Core Tools" is broken. Whichever tools we do in the future need > >> to *replace* the old tools. > >> > >> Marcus > >> > >> > >> -- > >> Marcus Denker -- http://www.marcusdenker.de > >> INRIA Lille -- Nord Europe. Team RMoD. > >> > >> > > > > > > -- > www.tudorgirba.com > > "Not knowing how to do something is not an argument for how it cannot be done." > > > -- www.tudorgirba.com "From an abstract enough point of view, any two things are similar."

March 14, 2011

Re: [Pharo-project] PBenchmark as a general framework for benchmarking in Smalltalk (commit access)

by Alexandre Bergel

You can check the class MondrianHealth2 that belongs to the package Mondrian-Tests of http://www.squeaksource.com/Mondrian.html The class contains many methods having a name that begins with 'metric'. Some students worked on a kind of Hudson that exploit these information. You can have a look at it: http://kimen.dcc.uchile.cl:8008/VerMon01/faces/project_details.xhtml?id=1 It would be cool to have a standard for this. As we have for SUnit. I think that what I did goes in the right direction, but the code is not as clean as it should. It would be cool to have a report generation, which could be .pdf, sending emails, ... Cheers, Alexandre On 14 Mar 2011, at 16:23, Stefan Marr wrote: > Hi Alexandre: > > > On 14 Mar 2011, at 18:10, Alexandre Bergel wrote: > >>> Since the general idea is to write benchmarks like unit tests, how about 'SMark'/'SBench' instead of 'SUnit'? >> >> >> I like the idea. Actually, this is what I have done with HealthReportProducer. A benchmark is created by subclassing it and defining #metric* methods. > Ehm, could you elaborate? Do you want me to have a look at something and steal some code/ideas? If so, I would need a more elaborate hint than just a name. > I googled for it, and there seems to be some reference in a metachello project, and interestingly google indexed already this email... > > Thanks > Stefan > > > > >> >> Cheers, >> Alexandre >> >> -- >> _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: >> Alexandre Bergel http://www.bergel.eu >> ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;. >> >> >> >> >> >> > > -- > 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 > > -- _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: Alexandre Bergel http://www.bergel.eu ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.

March 14, 2011

Re: [Pharo-project] PBenchmark as a general framework for benchmarking in Smalltalk (commit access)

by Stefan Marr

Hi Alexandre: On 14 Mar 2011, at 18:10, Alexandre Bergel wrote: >> Since the general idea is to write benchmarks like unit tests, how about 'SMark'/'SBench' instead of 'SUnit'? > > > I like the idea. Actually, this is what I have done with HealthReportProducer. A benchmark is created by subclassing it and defining #metric* methods. Ehm, could you elaborate? Do you want me to have a look at something and steal some code/ideas? If so, I would need a more elaborate hint than just a name. I googled for it, and there seems to be some reference in a metachello project, and interestingly google indexed already this email... Thanks Stefan > > Cheers, > Alexandre > > -- > _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: > Alexandre Bergel http://www.bergel.eu > ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;. > > > > > > -- 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 14, 2011

Re: [Pharo-project] Issue 3581 in pharo: [Failing test] 1.1.1 Full: MethodWrappers.Tests.ObjectAsOneTimeMethodWrapperTest.testInstallOnClassCategory

by pharo＠googlecode.com

Comment #12 on issue 3581 by snoob...(a)yahoo.ie: [Failing test] 1.1.1 Full: MethodWrappers.Tests.ObjectAsOneTimeMethodWrapperTest.testInstallOnClassCategory http://code.google.com/p/pharo/issues/detail?id=3581 Ok I see now. Then is it safe to just do ObjectAsMethodWrapper>>flushCache self selector flushCache.

March 14, 2011