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

Re: [Pharo-project] CommandShell

by laurent laffont

On Fri, Jan 28, 2011 at 7:19 AM, Sean P. DeNigris <sean(a)clipperadams.com>wrote: > > > laurent laffont wrote: > > > > It works (I can publish a package when back at home). > > > > Did you ever publish this anywhere? I didn't see it on SqS and I need it! > It seems no. Laurent. > > Thanks. > Sean > -- > View this message in context: > http://forum.world.st/CommandShell-tp2226990p3243771.html > Sent from the Pharo Smalltalk mailing list archive at Nabble.com. > >

Jan. 28, 2011

Re: [Pharo-project] Databases with full text search support

by laurent laffont

Riak looks extremely erotic: facet queries, HTTP/JSON query interface... Thanks for the pointer. Laurent. 2011/1/28 GÃ¶ran Krampe <goran(a)krampe.se> > Hi! > > I think the absolutely most interesting NoSQL DB with this in mind is Riak > and its Riak Search. Google it. I have been itching to write a binding for > Riak but have had no time to do it. > > Riak is built for "crazy just add another node" scaling and is modelled > after Dynamo. Recently Riak Search was added, which is a Lucene compatible > "overlay" for Riak thus giving you a free text search db with insane > scalability and robustness. > > CouchDB does not have anything close to this scalability, unless you look > at Cloudant perhaps - but AFAIK it does not have any kind of free text > indexing. Yes, CouchDB has a Lucene integration (I have written a binding in > C# for CouchDB and the Lucene integration) but we are still talking "single > server" setup. > > MongoDB is fast as hell, but AFAIK also does not have free text engine nor > the same "add another box" scalability like Riak has. > > regards, GÃ¶ran > >

Jan. 28, 2011

Re: [Pharo-project] CommandShell

by Sean P. DeNigris

laurent laffont wrote: > > It works (I can publish a package when back at home). > Did you ever publish this anywhere? I didn't see it on SqS and I need it! Thanks. Sean -- View this message in context: http://forum.world.st/CommandShell-tp2226990p3243771.html Sent from the Pharo Smalltalk mailing list archive at Nabble.com.

Jan. 28, 2011

Re: [Pharo-project] SimpleButtonMorph openInWorld

by Fernando Olivero

You can also evaluate the following expresions in a wokspace, SimpleButtonMorph allInstances do:[:each| each delete]. or | buttons | buttons := World submorphs detect:[:each| each isKindOf: SimpleButtonMorph ] . buttons anyOne delete. But i also prefer Serge option, because Morphs are directly manipulable for a reason! Fernando On Fri, Jan 28, 2011 at 2:49 AM, Serge Stinckwich <serge.stinckwich(a)gmail.com> wrote: > On Fri, Jan 28, 2011 at 9:40 AM, DAE <daebiz(a)douglasedmunds.com> wrote: >> This code makes a square button appear in the topleft corner >> >> Â Â Â Â SimpleButtonMorph new >> Â Â Â Â Â Â Â Â target: Smalltalk; >> Â Â Â Â Â Â Â Â label: 'Beep!'; >> Â Â Â Â Â Â Â Â actionSelector: #beep; >> Â Â Â Â Â Â Â Â openInWorld >> >> >> How do I make it go away? >> >> Sorry if I posted to the wrong group. Â If so, please tell me where is correct. > > On my macintosh, CTRL+ right mouse button > and select the delete item. > > Regards, > > -- > Serge Stinckwich > UMI UMMISCO 209 (IRD/UPMC), Hanoi, Vietnam > Every DSL ends up being Smalltalk > http://doesnotunderstand.org/ > >

Jan. 28, 2011

Re: [Pharo-project] SimpleButtonMorph openInWorld

by Serge Stinckwich

On Fri, Jan 28, 2011 at 9:40 AM, DAE <daebiz(a)douglasedmunds.com> wrote: > This code makes a square button appear in the topleft corner > > Â Â Â Â SimpleButtonMorph new > Â Â Â Â Â Â Â Â target: Smalltalk; > Â Â Â Â Â Â Â Â label: 'Beep!'; > Â Â Â Â Â Â Â Â actionSelector: #beep; > Â Â Â Â Â Â Â Â openInWorld > > > How do I make it go away? > > Sorry if I posted to the wrong group. Â If so, please tell me where is correct. On my macintosh, CTRL+ right mouse button and select the delete item. Regards, -- Serge Stinckwich UMI UMMISCO 209 (IRD/UPMC), Hanoi, Vietnam Every DSL ends up being Smalltalk http://doesnotunderstand.org/

Jan. 28, 2011

[Pharo-project] SimpleButtonMorph openInWorld

by DAE

This code makes a square button appear in the topleft corner SimpleButtonMorph new target: Smalltalk; label: 'Beep!'; actionSelector: #beep; openInWorld How do I make it go away? Sorry if I posted to the wrong group. If so, please tell me where is correct.

Jan. 28, 2011

Re: [Pharo-project] Issue 3599 in pharo: Serial Port

by Serge Stinckwich

On Fri, Jan 28, 2011 at 9:12 AM, David T. Lewis <lewis(a)mail.msen.com> wrote: > On Wed, Jan 26, 2011 at 03:09:14PM +0000, pharo(a)googlecode.com wrote: >> Status: Accepted >> Owner: stephane...(a)gmail.com >> Labels: Milestone-1.3 >> >> New issue 3599 by stephane...(a)gmail.com: Serial Port >> http://code.google.com/p/pharo/issues/detail?id=3599 >> >> Hello >> We are developing software to interact with small robots for educational >> purposes. To do this we need to use the serial port. SerialPort class that >> includes Pharo does not work on mac or linux, because it?s not Â able to >> open ports by name. > > FYI, the primitives for named serial ports are now in the standard VM > (maybe not yet in Cog, but it is a small change). If someone would like > to test on a Linux platform, try using any recent build of the standard > VM or use the latest 4.4.7.2357 unix VM from http://squeakvm.org/unix/. > > The image changes to SerialPort are simple. You can either copy changes > from Squeak or get the orginal change sets from the Mantis report at > http://bugs.squeak.org/view.php?id=7266. > > The named serial port work was contributed by Severin Lemaignan > <severin.lemaignan(a)laas.fr> and I helped with some of the integration. > Anything that I did is MIT licensed. Hi David, there is a issue in the Pharo bug tracker since several months for the integration of this changes into Pharo here: http://code.google.com/p/pharo/issues/detail?id=1057 Could we integrate these changes in 1.2 as there is no licence problems at all. David and SÃ©verin could you sign the licence agreement, just to be sure everything is ok. The quality of the code is not really good, there is a lot of isString everywhere in order to test if the port is a number or a string regarding the OS platform used: SerialPort>>close "Close the serial port. Do nothing if the port is not open." port ifNotNil: [port isString ifTrue:[self primClosePortByName: port] ifFalse:[self primClosePort:port]]. port := nil I think we need to polish it in order maybe to use only strings. Why not used string like 'COM1' for a windows platform and '/dev/ttyS*' for Linux ? Any comments about that ? Regards, -- Serge Stinckwich UMI UMMISCO 209 (IRD/UPMC), Hanoi, Vietnam Every DSL ends up being Smalltalk http://doesnotunderstand.org/

Jan. 28, 2011

Re: [Pharo-project] Issue 3599 in pharo: Serial Port

by David T. Lewis

On Wed, Jan 26, 2011 at 03:09:14PM +0000, pharo(a)googlecode.com wrote: > Status: Accepted > Owner: stephane...(a)gmail.com > Labels: Milestone-1.3 > > New issue 3599 by stephane...(a)gmail.com: Serial Port > http://code.google.com/p/pharo/issues/detail?id=3599 > > Hello > We are developing software to interact with small robots for educational > purposes. To do this we need to use the serial port. SerialPort class that > includes Pharo does not work on mac or linux, because it?s not able to > open ports by name. FYI, the primitives for named serial ports are now in the standard VM (maybe not yet in Cog, but it is a small change). If someone would like to test on a Linux platform, try using any recent build of the standard VM or use the latest 4.4.7.2357 unix VM from http://squeakvm.org/unix/. The image changes to SerialPort are simple. You can either copy changes from Squeak or get the orginal change sets from the Mantis report at http://bugs.squeak.org/view.php?id=7266. The named serial port work was contributed by Severin Lemaignan <severin.lemaignan(a)laas.fr> and I helped with some of the integration. Anything that I did is MIT licensed. HTH, Dave

Jan. 28, 2011

Re: [Pharo-project] Databases with full text search support

by GÃ¶ran Krampe

Hi! I think the absolutely most interesting NoSQL DB with this in mind is Riak and its Riak Search. Google it. I have been itching to write a binding for Riak but have had no time to do it. Riak is built for "crazy just add another node" scaling and is modelled after Dynamo. Recently Riak Search was added, which is a Lucene compatible "overlay" for Riak thus giving you a free text search db with insane scalability and robustness. CouchDB does not have anything close to this scalability, unless you look at Cloudant perhaps - but AFAIK it does not have any kind of free text indexing. Yes, CouchDB has a Lucene integration (I have written a binding in C# for CouchDB and the Lucene integration) but we are still talking "single server" setup. MongoDB is fast as hell, but AFAIK also does not have free text engine nor the same "add another box" scalability like Riak has. regards, GÃ¶ran

Jan. 28, 2011

Re: [Pharo-project] Issue 3603 in pharo: O/ECompletion failure in Trait methods

by Levente Uzonyi

On Thu, 27 Jan 2011, Henrik Sperre Johansen wrote: > On 27.01.2011 19:53, Levente Uzonyi wrote: >> On Thu, 27 Jan 2011, pharo(a)googlecode.com wrote: >> >>> Status: Accepted >>> Owner: ryd...(a)gmail.com >>> CC: romain.r...(a)gmail.com >>> Labels: Milestone-1.2-DevImage Milestone-1.3-DevImage OCompletion >>> >>> New issue 3603 by ryd...(a)gmail.com: O/ECompletion failure in Trait methods >>> http://code.google.com/p/pharo/issues/detail?id=3603 >>> >>> A debugger pops up when typing in after super in a trait method, since >>> traits do not implement superclass. >> >> The bug is not related to OCompletion. Traits implement (via TPureBehavior) >> #allSuperclasses, #withAllSuperclasses, etc., but lack >> #superclass. > So which implementation of #superclass would you recommend for Trait, which > would keep O/ECompletion working without modification? > I can't think of any which makes sense really. > > IMO, if code in the component eventually needs to be changed to not raise a > debugger, it's related. Okay, I added a check for this. It's in Ocompletion-ul.91. Levente > > Cheers, > Henry > >

Jan. 28, 2011