- Pharo-dev - lists.pharo.org

[Pharo-project] [update 1.4] #14372

by Marcus Denker

14372 ----- Issue 4145: PluggableButtonButtonMorph>>#target http://code.google.com/p/pharo/issues/detail?id=4145 Issue 3403: should remove addServices2:for:extraLines: http://code.google.com/p/pharo/issues/detail?id=3403 Issue 2520: sourceMatchesBytecodeAt http://code.google.com/p/pharo/issues/detail?id=2520 Issue 5206: better api for browser http://code.google.com/p/pharo/issues/detail?id=5206 -- Marcus Denker -- http://marcusdenker.de

March 3, 2012

[Pharo-project] odd gotcha in workspace code for squeak and pharo

by Lawson English

A discussion in IRC prompted me to try several things. In a nutshell, execute these two lines with two separate doits and you get an error. Select both lines and doit and you don't: a := [^true]. a value. I assume there is a valid reason for the difference in behavior, but it DOES remind one that workspace behavior isn't always the same as method behavior, even for something as simple as the above. L.

March 3, 2012

Re: [Pharo-project] Nautilus feedback

by Benjamin

On Mar 3, 2012, at 1:26 PM, Alexandre Bergel wrote: > Hi! > > Nautilus is cool! I use it in the Pharo1.4-based Moose image. > Some comments: > - Cmd-t when selecting a package does not run the tests OK, I will check that :) So, in my image it works, so maybe your Nautilus is not up to date > - Cmd-b does not open a Nautilus, but a standard browser It opens the default browser :) > > What about the refactoring engine? Does it work in Pharo 1.4? Refactorings have now been integrated (10 days ago :)). And it seems they work Thanks for feedback, Ben > > Alexandre > -- > _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: > Alexandre Bergel http://www.bergel.eu > ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;. > > > > > >

March 3, 2012

Re: [Pharo-project] Looking for a package to do some statistics in Pharo

by Schwab,Wilhelm K

+1. I just (finally) ordered the book, which should help. Three things put me off of this: (1) (perhaps biased) how can Smalltalk code be fast enough? (2) No examples, poor categorization (3) The use of histograms in least squares fitting - I'm a parametric guy. Re the dearth of examples, I created a few (attached) - barely scratching the surface, I know. I'm really stumped on OptimzationDHB and #points:function:. The method signature sounds like what I want to do, but I doubt that something as "simple" as DhbLeastSquareFit points:self sampleDataNegativeExponentialPlusConstant function:[ :x :a :alpha :b | a * ( (x / alpha) squared negated exp ) + b ]. is going to be "understood." We know that x is the independent variable and a, alpha and b are parameters to be found, but I have no idea how to express that to the code. Any ideas? It might be that I need to look at the optimizers, since most of my fits are really about minimizing residuals. Is Levenberg-Marquardt anywhere in the mix? Bill ________________________________________ From: pharo-project-bounces(a)lists.gforge.inria.fr [pharo-project-bounces(a)lists.gforge.inria.fr] on behalf of Alexandre Bergel [alexandre.bergel(a)me.com] Sent: Saturday, March 03, 2012 7:30 AM To: Pharo-project(a)lists.gforge.inria.fr Subject: Re: [Pharo-project] Looking for a package to do some statistics in Pharo I am interested in this. Cheers, Alexandre On 3 Mar 2012, at 05:35, StÃ©phane Ducasse wrote: > Thanks > I was reviewer for the book of didier. > Now I think that it would be good to revive the project and make it even better. > I started to have a look. I will publish a configuration to start with. > > Stef > > > On Mar 3, 2012, at 9:13 AM, Douglas McPherson wrote: > >> Indeed, you should check out DHBNumerical. I'm using it for Kalman filter development/simulation, and have previously done Galois field simulation using Bessett's DHBPolynomial. These are many included probability distributions and statistical queries built in, so quite likely has what you need. No underscores needed :) >> >> Doug >> >> On Mar 2, 2012, at 10:12 , StÃ©phane Ducasse wrote: >> >>> Hi guys >>> >>> I would like to know if there is a package doing some statistics in Pharo. >>> I should check the work of didier besset but I want to know if there is more. >>> >>> It would be good to have such kind of package. >>> >>> >>> Stef >>> >> >> > > -- _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: Alexandre Bergel http://www.bergel.eu ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.

March 3, 2012

Re: [Pharo-project] Sender of or: Boolean method

by Dario Trussardi

Ciao, > Ciao Dario, i dont understand the question. > > You are stating that there's an error , because you know for sure that > #or: has more then 10 senders in the Pharo1.0? No, but i'm understand to know because it. > > Fernando > > pd: By the way, i advice to use a later version of Pharo, manyyy > things were improved since then! I work with Pharo-1.0 because i run GemTools Launcher - >=1.0-beta.8.6 [ConfigurationOfGemTools] . I can run the last version of GemTools ( which ? ) in a recent Pharo version (which) ? > > On Sat, Mar 3, 2012 at 1:02 PM, Dario Trussardi > <dario.trussardi(a)tiscali.it> wrote: >> Ciao, >> >> i work with: >> >> Pharo-1.0 >> Latest update: #10517. >> >> ( but i found the same answer into >> Pharo1.2.2 >> Latest update: #12353 ) >> >> >> >> I have a browser open on Boolean class. >> >> >> When i select the or: method and click on the sender button the system answer only ten methods. >> >> >> Because ? Thank, good weekend, Dario

March 3, 2012

Re: [Pharo-project] Looking for a package to do some statistics in Pharo

by Alexandre Bergel

I am interested in this. Cheers, Alexandre On 3 Mar 2012, at 05:35, StÃ©phane Ducasse wrote: > Thanks > I was reviewer for the book of didier. > Now I think that it would be good to revive the project and make it even better. > I started to have a look. I will publish a configuration to start with. > > Stef > > > On Mar 3, 2012, at 9:13 AM, Douglas McPherson wrote: > >> Indeed, you should check out DHBNumerical. I'm using it for Kalman filter development/simulation, and have previously done Galois field simulation using Bessett's DHBPolynomial. These are many included probability distributions and statistical queries built in, so quite likely has what you need. No underscores needed :) >> >> Doug >> >> On Mar 2, 2012, at 10:12 , StÃ©phane Ducasse wrote: >> >>> Hi guys >>> >>> I would like to know if there is a package doing some statistics in Pharo. >>> I should check the work of didier besset but I want to know if there is more. >>> >>> It would be good to have such kind of package. >>> >>> >>> Stef >>> >> >> > > -- _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: Alexandre Bergel http://www.bergel.eu ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.

March 3, 2012

[Pharo-project] Nautilus feedback

by Alexandre Bergel

Hi! Nautilus is cool! I use it in the Pharo1.4-based Moose image. Some comments: - Cmd-t when selecting a package does not run the tests - Cmd-b does not open a Nautilus, but a standard browser What about the refactoring engine? Does it work in Pharo 1.4? Alexandre -- _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: Alexandre Bergel http://www.bergel.eu ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.

March 3, 2012

Re: [Pharo-project] Sender of or: Boolean method

by Fernando Olivero

Ciao Dario, i dont understand the question. You are stating that there's an error , because you know for sure that #or: has more then 10 senders in the Pharo1.0? Fernando pd: By the way, i advice to use a later version of Pharo, manyyy things were improved since then! On Sat, Mar 3, 2012 at 1:02 PM, Dario Trussardi <dario.trussardi(a)tiscali.it> wrote: > Ciao, > > Â Â Â Â i work with: > > Â Â Â Â Â Â Â Â Pharo-1.0 > Â Â Â Â Â Â Â Â Latest update: #10517. > > Â Â Â Â ( but i found the same answer into > Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Pharo1.2.2 > Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Latest update: #12353 Â ) > > > > I have a browser open on Boolean class. > > > When i select the Â Â Â Â Â Â Â Â or: Â Â Â Â Â Â method Â and click on Â the Â Â Â sender button Â the system answer only Â ten Â methods. > > > Because ? > > > Thank, > > Â Â Â Â Â Â Â Â Dario

March 3, 2012

Re: [Pharo-project] Sender of or: Boolean method

by Nicolas Cellier

Because the message is not really send. Instead it is inlined (you could try to browse the bytecode of the method). Since the system should hide such implementation details, latest Pharo have been modified to track the senders of these inlined message. Nicolas Le 3 mars 2012 13:02, Dario Trussardi <dario.trussardi(a)tiscali.it> a Ã©crit : > Ciao, > > Â Â Â Â i work with: > > Â Â Â Â Â Â Â Â Pharo-1.0 > Â Â Â Â Â Â Â Â Latest update: #10517. > > Â Â Â Â ( but i found the same answer into > Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Pharo1.2.2 > Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Latest update: #12353 Â ) > > > > I have a browser open on Boolean class. > > > When i select the Â Â Â Â Â Â Â Â or: Â Â Â Â Â Â method Â and click on Â the Â Â Â sender button Â the system answer only Â ten Â methods. > > > Because ? > > > Thank, > > Â Â Â Â Â Â Â Â Dario

March 3, 2012

Re: [Pharo-project] Sender of or: Boolean method

by Marcus Denker

On Mar 3, 2012, at 1:02 PM, Dario Trussardi wrote: > Ciao, > > i work with: > > Pharo-1.0 > Latest update: #10517. > > ( but i found the same answer into > Pharo1.2.2 > Latest update: #12353 ) > > > > I have a browser open on Boolean class. > > > When i select the or: method and click on the sender button the system answer only ten methods. > > > Because ? > It is compiled to jumps, the message send is optimized away. e.g "true or: [false]" is compiled to: 17 <71> pushConstant: true 18 <99> jumpFalse: 21 19 <71> pushConstant: true 20 <90> jumpTo: 22 21 <72> pushConstant: false 22 <7C> returnTop This means that then in the compiledmethod, there is no symbol needed for or: to execute the code. It is omited for space reasons (keep in mind that the original target machine of this system had 512 Kb of RAM maximum) The "senders of" iterates over all methods and looks if a symbol is there, it does not look at source code. (that's why it is so fast and works even without sources...) These days where a $25 computer is shipped witth 256MB of RAM, of course adding in 1000 methods pointers to a symbol is not that of a problem, and therefore, in 1.4, the selector of optimized-away messages is now added to the literalframe (if there is space), even though it's pure waste from the execution point of view. So now, senders of #or: gives you 1092 methods. Same for ifTrue: and all the other inlined methods. Marcus -- Marcus Denker -- http://marcusdenker.de

March 3, 2012