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

Re: [Pharo-project] FFI & Python

by phil＠highoctane.be

Wrap things up in an Objective-C thing and then use the ObjectiveC bridge to bring the object into Smalltalk space. Like Esteban did with the DefaultFileManager in the iPad VM. Mac only obviously. Phil 2012/9/6 S Krish <krishnamachari.sudhakar(a)gmail.com> > I have a fairly simple XMLRPC interface working.. but then that is no > substitute to direct interface.. > > I like the simplicity of the python binding from Smalltalk.. > > On Thu, Sep 6, 2012 at 11:38 AM, phil(a)highoctane.be <phil(a)highoctane.be>wrote: > >> Can't you just get the current implementation run as a separate >> process managing that main variable along with some web/TCP interface >> and get that working through Zinc or OSProcess with a control script? >> >> I am currently experimenting with CUDA (same kind of freaking >> interfacing) and frankly, the quickest way to see some progress and >> understand what's going on has been to use OSProcess to compile what I >> need, ship the code to the card etc. No FFI etc. Just make a neat >> little set of scripts. >> >> FWIW, I also leverage Tcl quite often to wrap C stuff (Yeah, I love >> homoiconic languages - ah, can't resist: >> http://en.wikipedia.org/wiki/Homoiconicity) And also, SWIG may help a >> lot for these endeavors. http://www.swig.org/ >> >> Also, Pharo can give HTTP-based access (lots of options here) for >> controlling the other around. There seems to be some REST support >> available and JSON parsing is around. >> >> Of course, this doesn't answer the question... >> >> Phil >> >> >> 2012/9/6 S Krish <krishnamachari.sudhakar(a)gmail.com>: >> > Is there a source / person who could detail how the python binding is >> > done... >> > >> > 1. To port this into Pharo >> > >> > 2. Port this design to bind Groovy to Pharo.. >> > >> > >> > >> > >> > On Thu, Sep 6, 2012 at 6:47 AM, Serge Stinckwich >> > <serge.stinckwich(a)gmail.com> wrote: >> >> >> >> On Thu, Sep 6, 2012 at 5:58 AM, Santiago Bragagnolo >> >> <santiagobragagnolo(a)gmail.com> wrote: >> >> > Hi all! >> >> >> >> Hi Santiago, >> >> >> >> > I'm currently working in a project related with ROS (robot >> operative >> >> > system), which already have a library to develop in C++ and other to >> do >> >> > the >> >> > same at Python. So, there a lot of variables to contrast to make a >> >> > choice >> >> > (performance, code-easy-to-prototype, flexibility, and a lot more), >> but >> >> > the >> >> > main variable is the accessibility from Pharo, then the question is, >> is >> >> > there any project that use python from pharo? Or, if i select python >> i >> >> > should write a library in C that bootstrap python and give access to >> >> > pharo >> >> > via FFI? >> >> > >> >> > I looked in the list's mails and in google without any success. >> >> >> >> As Chris already told, there is a Python binding with OpenQwaq: >> >> >> >> >> https://squeakingalong.wordpress.com/2011/05/04/openqwaq-python-integration/ >> >> >> >> I'm really interested to be able to use ROS from Pharo. >> >> I start some weeks ago to work on a pure Smalltalk client for ROS: >> >> http://ss3.gemstone.com/ss/RosSmalltalk.html >> >> >> >> Noury is also interested and start some work also: >> >> http://car.mines-douai.fr/squeaksource/RoSt.html >> >> >> >> We planned with Noury to join our efforts to have a ROS client as soon >> >> as possible. >> >> Noury will also recruit a engineer to work full time on this. >> >> >> >> Another alternative in order to use ROS from Smalltalk is to use >> >> websockets and rosbridge. >> >> I have done some work to use ROS from Amber here: >> >> https://github.com/SergeStinckwich/rostalk >> >> >> >> You are than welcome to join us ;-) >> >> Regards, >> >> -- >> >> Serge Stinckwich >> >> UMI UMMISCO 209 (IRD/UPMC), Hanoi, Vietnam >> >> Every DSL ends up being Smalltalk >> >> http://doesnotunderstand.org/ >> >> >> > >> >> >

Sept. 6, 2012

Re: [Pharo-project] Process browser: next generation

by phil＠highoctane.be

Main UI process and a haltOnce would indeed be great instead of the current kata. 2012/9/6 Goubier Thierry <thierry.goubier(a)cea.fr> > Le 06/09/2012 16:16, Sean P. DeNigris a Ã©crit : > >> >> Alexandre Bergel-7 wrote >> >>> >>> Ok, this is indeed the first thing you want >>> >>> On Sep 5, 2012, at 7:51 PM, phil@ wrote: >>> >>>> An ability to jump to the process code right away. >>>> >>> >>> >> Yes, that would be super cool! >> > > Isn't that the case in the current process browser ? > > Or do you mean something like forcing a breakpoint in a currently running > process ? > > ( It would be cool to breakpoint like that the main UI process... :)). > > Thierry > > >> >> >> -- >> View this message in context: http://forum.world.st/Process-** >> browser-next-generation-**tp4646267p4646408.html<http://forum.world.st/Process-browser-next-generation-tp4646267p4646408.html> >> Sent from the Pharo Smalltalk mailing list archive at Nabble.com. >> >> >> >> > > -- > Thierry Goubier > CEA list > Laboratoire des Fondations des SystÃ¨mes Temps RÃ©el EmbarquÃ©s > 91191 Gif sur Yvette Cedex > France > Phone/Fax: +33 (0) 1 69 08 32 92 / 83 95 > >

Sept. 6, 2012

Re: [Pharo-project] Process browser: next generation

by Sean P. DeNigris

Goubier Thierry wrote > > Or do you mean something like forcing a breakpoint in a currently > running process ? > Yes, exactly :) -- View this message in context: http://forum.world.st/Process-browser-next-generation-tp4646267p4646410.html Sent from the Pharo Smalltalk mailing list archive at Nabble.com.

Sept. 6, 2012

Re: [Pharo-project] Process browser: next generation

by Goubier Thierry

Le 06/09/2012 16:16, Sean P. DeNigris a Ã©crit : > > Alexandre Bergel-7 wrote >> >> Ok, this is indeed the first thing you want >> >> On Sep 5, 2012, at 7:51 PM, phil@ wrote: >>> An ability to jump to the process code right away. >> > > Yes, that would be super cool! Isn't that the case in the current process browser ? Or do you mean something like forcing a breakpoint in a currently running process ? ( It would be cool to breakpoint like that the main UI process... :)). Thierry > > > > -- > View this message in context: http://forum.world.st/Process-browser-next-generation-tp4646267p4646408.html > Sent from the Pharo Smalltalk mailing list archive at Nabble.com. > > > -- Thierry Goubier CEA list Laboratoire des Fondations des SystÃ¨mes Temps RÃ©el EmbarquÃ©s 91191 Gif sur Yvette Cedex France Phone/Fax: +33 (0) 1 69 08 32 92 / 83 95

Sept. 6, 2012

Re: [Pharo-project] Process browser: next generation

by Sean P. DeNigris

Alexandre Bergel-7 wrote > > Ok, this is indeed the first thing you want > > On Sep 5, 2012, at 7:51 PM, phil@ wrote: >> An ability to jump to the process code right away. > Yes, that would be super cool! -- View this message in context: http://forum.world.st/Process-browser-next-generation-tp4646267p4646408.html Sent from the Pharo Smalltalk mailing list archive at Nabble.com.

Sept. 6, 2012

Re: [Pharo-project] FFI & Python

by S Krish

I have a fairly simple XMLRPC interface working.. but then that is no substitute to direct interface.. I like the simplicity of the python binding from Smalltalk.. On Thu, Sep 6, 2012 at 11:38 AM, phil(a)highoctane.be <phil(a)highoctane.be>wrote: > Can't you just get the current implementation run as a separate > process managing that main variable along with some web/TCP interface > and get that working through Zinc or OSProcess with a control script? > > I am currently experimenting with CUDA (same kind of freaking > interfacing) and frankly, the quickest way to see some progress and > understand what's going on has been to use OSProcess to compile what I > need, ship the code to the card etc. No FFI etc. Just make a neat > little set of scripts. > > FWIW, I also leverage Tcl quite often to wrap C stuff (Yeah, I love > homoiconic languages - ah, can't resist: > http://en.wikipedia.org/wiki/Homoiconicity) And also, SWIG may help a > lot for these endeavors. http://www.swig.org/ > > Also, Pharo can give HTTP-based access (lots of options here) for > controlling the other around. There seems to be some REST support > available and JSON parsing is around. > > Of course, this doesn't answer the question... > > Phil > > > 2012/9/6 S Krish <krishnamachari.sudhakar(a)gmail.com>: > > Is there a source / person who could detail how the python binding is > > done... > > > > 1. To port this into Pharo > > > > 2. Port this design to bind Groovy to Pharo.. > > > > > > > > > > On Thu, Sep 6, 2012 at 6:47 AM, Serge Stinckwich > > <serge.stinckwich(a)gmail.com> wrote: > >> > >> On Thu, Sep 6, 2012 at 5:58 AM, Santiago Bragagnolo > >> <santiagobragagnolo(a)gmail.com> wrote: > >> > Hi all! > >> > >> Hi Santiago, > >> > >> > I'm currently working in a project related with ROS (robot > operative > >> > system), which already have a library to develop in C++ and other to > do > >> > the > >> > same at Python. So, there a lot of variables to contrast to make a > >> > choice > >> > (performance, code-easy-to-prototype, flexibility, and a lot more), > but > >> > the > >> > main variable is the accessibility from Pharo, then the question is, > is > >> > there any project that use python from pharo? Or, if i select python i > >> > should write a library in C that bootstrap python and give access to > >> > pharo > >> > via FFI? > >> > > >> > I looked in the list's mails and in google without any success. > >> > >> As Chris already told, there is a Python binding with OpenQwaq: > >> > >> > https://squeakingalong.wordpress.com/2011/05/04/openqwaq-python-integration/ > >> > >> I'm really interested to be able to use ROS from Pharo. > >> I start some weeks ago to work on a pure Smalltalk client for ROS: > >> http://ss3.gemstone.com/ss/RosSmalltalk.html > >> > >> Noury is also interested and start some work also: > >> http://car.mines-douai.fr/squeaksource/RoSt.html > >> > >> We planned with Noury to join our efforts to have a ROS client as soon > >> as possible. > >> Noury will also recruit a engineer to work full time on this. > >> > >> Another alternative in order to use ROS from Smalltalk is to use > >> websockets and rosbridge. > >> I have done some work to use ROS from Amber here: > >> https://github.com/SergeStinckwich/rostalk > >> > >> You are than welcome to join us ;-) > >> Regards, > >> -- > >> Serge Stinckwich > >> UMI UMMISCO 209 (IRD/UPMC), Hanoi, Vietnam > >> Every DSL ends up being Smalltalk > >> http://doesnotunderstand.org/ > >> > > > >

Sept. 6, 2012

Re: [Pharo-project] Who broke UTF8TextConverter?

by Sven Van Caekenberghe

Well said. On 06 Sep 2012, at 15:41, Sean P. DeNigris <sean(a)clipperadams.com> wrote: > Every wtf we leave in the system limits the human expression of ourselves > and our users. For one thing, aesthetics count. Emotions are hugely > important to the quality of one's work, and unless we are all inspired by a > beautiful system, our discontent will be reflected in code written in and on > top of Pharo. Also, while crusty old Pharo wizards may know where all these > little land mines are that must be avoided, new users may waste hours/days > before finding out that things don't work the way one would naturally expect > them too. > > If we can operate on RPackage in a live system, breaking code committing and > loading, we can change *anything*. I've definitely felt the pain in 2.0, but > this is what we're working for. Let's rip the bandaid off and have the > system of our dreams. If anything, I believe the question should *never* be > "do we leave this crap in the system or not?", but merely "how do we ease > the pain of making this better?", whether via deprecations, RB > transformations, or whatever; and we seem to be making steady progress on > that front. -- Sven Van Caekenberghe http://stfx.eu Smalltalk is the Red Pill

Sept. 6, 2012

Re: [Pharo-project] Process browser: next generation

by Alexandre Bergel

Ok, this is indeed the first thing you want Alexandre On Sep 5, 2012, at 7:51 PM, phil(a)highoctane.be wrote: > An ability to jump to the process code right away. > > > 2012/9/6 Alexandre Bergel <alexandre.bergel(a)me.com>: >> Hello! >> >> We are currently working on a new process browser (codename: Loom). You have probably noticed that the Process Browser is not exactly the tool you want to use to understand how processes behave. The process browser may be one of the oldest piece of Pharo code that has not been touched for a long long time. >> >> We are going toward something that looks like the Activity Monitor (e.g., [*]), useful to find out which programs is heating up your laptop. We will have nice curves that shows you the CPU consumption. Something that the process browser has never been able to do. >> >> What are the features you would like to see in it? >> >> Cheers, >> Alexandre >> >> [*] http://media.tumblr.com/tumblr_lb4161Ehpr1qarbgx.png >> >> -- >> _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: >> Alexandre Bergel http://www.bergel.eu >> ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;. >> >> >> >> > -- _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: Alexandre Bergel http://www.bergel.eu ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.

Sept. 6, 2012

Re: [Pharo-project] Process browser: next generation

by Alexandre Bergel

> What are the features you would like to see in it? > > Have a working CPUWatcher Did it work on some point? I read the documentation but could not find out how to give a meaning to it. Looks like a half-finished thing. Cheers, Alexandre -- _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: Alexandre Bergel http://www.bergel.eu ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.

Sept. 6, 2012

Re: [Pharo-project] Who broke UTF8TextConverter?

by Sean P. DeNigris

Henrik Sperre Johansen wrote > > their extensive usage > limits the space for actual improvements within the space of backwards > compatability. http://code.google.com/p/pharo/#Pharo_manifesto : > Pharo manifesto > - Better for the better > - Beauty to learn from > - Not backward compatible *** > - Clean, lean and fast > Every wtf we leave in the system limits the human expression of ourselves and our users. For one thing, aesthetics count. Emotions are hugely important to the quality of one's work, and unless we are all inspired by a beautiful system, our discontent will be reflected in code written in and on top of Pharo. Also, while crusty old Pharo wizards may know where all these little land mines are that must be avoided, new users may waste hours/days before finding out that things don't work the way one would naturally expect them too. If we can operate on RPackage in a live system, breaking code committing and loading, we can change *anything*. I've definitely felt the pain in 2.0, but this is what we're working for. Let's rip the bandaid off and have the system of our dreams. If anything, I believe the question should *never* be "do we leave this crap in the system or not?", but merely "how do we ease the pain of making this better?", whether via deprecations, RB transformations, or whatever; and we seem to be making steady progress on that front. -- View this message in context: http://forum.world.st/Who-broke-UTF8TextConverter-tp4645473p4646399.html Sent from the Pharo Smalltalk mailing list archive at Nabble.com.

Sept. 6, 2012