January 2010 - Pharo-dev - lists.pharo.org

Re: [Pharo-project] Permutations and Combinations in SequenceableCollection

by merlyn＠stonehenge.com

>>>>> "Levente" == Levente Uzonyi <leves(a)elte.hu> writes: Levente> On Tue, 5 Jan 2010, John Toohey wrote: >> But this doesn't get 1-1, 2-2 or 3-3. I had tried something like that use >> #reverse on each combination, but #combinations does not generate >> -duplicates- from the number sequences. Levente> I guess you are looking for this: Levente> allPairsFromCollectionDo := [ :collection :block | Levente> | o | Levente> o := OrderedCollection newFrom: collection. Levente> o size timesRepeat: [ Levente> o with: collection do: block. Levente> o add: o removeFirst ] ]. Or more simply: aCollection do: [:x | aCollection do: [:y | "use x and y"]]. Unless I'm misunderstanding the problem. Which is neither combinations nor permutations, but so far is the only thing that fits all the bizarre conditions. -- Randal L. Schwartz - Stonehenge Consulting Services, Inc. - +1 503 777 0095 <merlyn(a)stonehenge.com> <URL:http://www.stonehenge.com/merlyn/> Smalltalk/Perl/Unix consulting, Technical writing, Comedy, etc. etc. See http://methodsandmessages.vox.com/ for Smalltalk and Seaside discussion

Jan. 6, 2010

Re: [Pharo-project] Permutations and Combinations in SequenceableCollection

by John Toohey

Very nice, thank you. On Tue, Jan 5, 2010 at 17:21, Levente Uzonyi <leves(a)elte.hu> wrote: > On Tue, 5 Jan 2010, John Toohey wrote: > > > But this doesn't get 1-1, 2-2 or 3-3. I had tried something like that use > > #reverse on each combination, but #combinations does not generate > > -duplicates- from the number sequences. > > I guess you are looking for this: > allPairsFromCollectionDo := [ :collection :block | > | o | > o := OrderedCollection newFrom: collection. > o size timesRepeat: [ > o with: collection do: block. > o add: o removeFirst ] ]. > > Transcript open. > allPairsFromCollectionDo value: #(1 2 3 4) value: [ :a :b | > Transcript print: { a. b }; cr ]. > Transcript flush > > > Levente > > > > > On Tue, Jan 5, 2010 at 12:16, Henrik Johansen > > <henrik.s.johansen(a)veloxit.no>wrote: > > > >> On Jan 5, 2010, at 5:19 32PM, John Toohey wrote: > >> > >>> Hi, > >>> I'm trying to generate the permutations of a sequence of numbers. I've > >> found # permutationsDo: aBlock, but this generated the permutation of > the > >> entire sequence. I want to do it in groups of 2. The #combinations: > >> anInteger atATimeDo: aBlock method does what I need for combinations, > but I > >> cannot find the equivalent for permutations. Does something like this > exist > >> in the standard library, or in an external package. > >>> > >>> -- > >>> -JT > >> > >> How about: > >> #(1 2 3 4) combinations: 2 atATimeDo: [:array | array permutationsDo: > >> [:each | Transcript show: each printString; cr]] ? > >> > >> Cheers, > >> Henry > >> > >> > >> _______________________________________________ > >> Pharo-project mailing list > >> Pharo-project(a)lists.gforge.inria.fr > >> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project > >> > > > > > > > > -- > > -JT > > > > _______________________________________________ > Pharo-project mailing list > Pharo-project(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project > -- -JT

Jan. 6, 2010

Re: [Pharo-project] Simple question about NewInspector

by Mariano Martinez Peck

On Mon, Jan 4, 2010 at 9:49 AM, Frederic Pluquet <fpluquet(a)ulb.ac.be> wrote: > RoelTyper is also used in ProtoObject>>niChildrenBlockForFixedFields (an > extension for NewInspector). > > NewInspector can work without RoelTyper. The message "The last version of > RoelTyper must be loaded to have this information" is shown when > NewInspector tries to calculate types and that RoelTyper is not found. > Cool. Done in ConfigurationOfNewInspector. Thanks > > > FrÃ©d > ---- > FrÃ©dÃ©ric Pluquet > UniversitÃ© Libre de Bruxelles (ULB) > Assistant > http://www.ulb.ac.be/di/fpluquet > > > On Sat, Dec 26, 2009 at 11:29 PM, Mariano Martinez Peck < > marianopeck(a)gmail.com> wrote: > >> Hi Frederic. I am trying to see if it make sense to create a metacello >> configuration for NewInspector. The only external (not in PharoCore) >> dependency I saw using the DependencyBrowser is RoelTyper, which is used in >> the class NICompiledMethodNode >> configuration you use the class >> TypeCollector which is from RoelTyper. >> >> So, the question is, is this dependency correct? I ask because I took a >> PharoCore image, load NewINspector WITHOUT RoelTyper and it "seemed" to >> work. >> >> Thank you very much. >> >> Mariano >> > >

Jan. 6, 2010

Re: [Pharo-project] Building Pharo and Plugins for Mac OS X and Linux

by David T. Lewis

On Tue, Jan 05, 2010 at 11:55:40AM -0300, Javier Pim??s wrote: > It's nice to see there's more info about VM's than I thought, it would be > even nicer to have it all gathered in just one place and up to date so it's > easier to find it. Recently I started compiling my own unix VMs and as I was > very confused at first, I decided to write a small guide for beginers to > share my experiences (and to remember the steps after some time). It has > some still-unanswered questions, but is thought for beginners, so it might > help you. The link is this: > > http://wwew.javipimas.com.ar/node/2 <http://www.javipimas.com.ar/node/2> > > hope it helps. Javier, This is a really good introduction. It explains the process clearly, and your diagram does a nice job of showing how the pieces come together. Well done! Regarding your question on the web page: > I've read that pharo uses a modified OE that supports full closures, > but I don't know what changes you have to introduce to make them work. > Anyway, I don't know if closures aren't used that much, or the default > code already ingtegrates them but my pharo images never complained > of my handmade OEs, so the VMMaker package seems to be the same for > both Pharo and Squeak. All current VMs provide closure support, which is included in the VMMaker package. If your image uses closures, then the VM supports them and there are no modifications required. Pharo, Squeak trunk, and Cuis images all use closures now, other others (e.g. SqueakLand) will probable have this added as time goes on. Dave

Jan. 5, 2010

Re: [Pharo-project] Interim build server

by Yanni Chiu

Michael Roberts wrote: > ... > then once we have a server I'm happy to look at Hudson or whatever we > consider the right tool for doing this. I have also been playing with > Mason and sent some feedback to Colin. Like i say, for me the real > issues are running a Pharo image completely automated. I think one important issue is running the tests and making the results easily visible. Hudson can be configured to send emails (to an appropriate recipient) if a build fails. Building and running tests is not being done frequently enough so that the feedback can be acted on quickly (when it's most effective). I'm hoping that the Hudson stuff can do that. I don't have a server available to set it up currently, but maybe shortly. -- Yanni

Jan. 5, 2010

Re: [Pharo-project] Dev images with or without tests?

by Yanni Chiu

Mariano Martinez Peck wrote: > Ok, that's the big questions. Until now, the Dev image only cointained > the tests of the PharoCore image. Actually, it also included the tests > fro those projects where the test were in the same package (for example, > RoelTyer). You need to build and provide both types of images. The developer then decides which one they need based on whether or not they expect to be modifying "base" code. For example, if I'm just writing a Seaside app. that does not change Seaside itself, I would not want the Seaside tests around. It may be wise to have my build server load and run them, and let me know of test failures - but, the day to day development image should not have them. If however, when building my app., I found I was modifying and extending Seaside itself, then I would choose to use a base image that included the Seaside tests. In order to keep a build time under 10 minutes, as things scale, you need to build and test the pieces individually. Then you start your development from a known and tested baseline image. BTW, the Hudson build stuff I just announced, has this build/test scenario in mind. The "product" itself is built and saved. Then the test packages are added, and the test cases are run. Unfortunately, when the test cases are in the same package, they cannot be left out of the "product" image. In the Pier build example, some addon packages which included tests are loaded in the test build. It's not right, because the product image is missing these addon packages. There's a test package dependency chain, that could not be solved, because the tests and code were in the same package. -- Yanni Chiu

Jan. 5, 2010

Re: [Pharo-project] Permutations and Combinations in SequenceableCollection

by Levente Uzonyi

On Tue, 5 Jan 2010, John Toohey wrote: > But this doesn't get 1-1, 2-2 or 3-3. I had tried something like that use > #reverse on each combination, but #combinations does not generate > -duplicates- from the number sequences. I guess you are looking for this: allPairsFromCollectionDo := [ :collection :block | | o | o := OrderedCollection newFrom: collection. o size timesRepeat: [ o with: collection do: block. o add: o removeFirst ] ]. Transcript open. allPairsFromCollectionDo value: #(1 2 3 4) value: [ :a :b | Transcript print: { a. b }; cr ]. Transcript flush Levente > > On Tue, Jan 5, 2010 at 12:16, Henrik Johansen > <henrik.s.johansen(a)veloxit.no>wrote: > >> On Jan 5, 2010, at 5:19 32PM, John Toohey wrote: >> >>> Hi, >>> I'm trying to generate the permutations of a sequence of numbers. I've >> found # permutationsDo: aBlock, but this generated the permutation of the >> entire sequence. I want to do it in groups of 2. The #combinations: >> anInteger atATimeDo: aBlock method does what I need for combinations, but I >> cannot find the equivalent for permutations. Does something like this exist >> in the standard library, or in an external package. >>> >>> -- >>> -JT >> >> How about: >> #(1 2 3 4) combinations: 2 atATimeDo: [:array | array permutationsDo: >> [:each | Transcript show: each printString; cr]] ? >> >> Cheers, >> Henry >> >> >> _______________________________________________ >> Pharo-project mailing list >> Pharo-project(a)lists.gforge.inria.fr >> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project >> > > > > -- > -JT >

Jan. 5, 2010

Re: [Pharo-project] Permutations and Combinations in SequenceableCollection

by CÃ©drick BÃ©ler

Can you give an example (on #(1 2 3 4) for instance) ? Thx 2010/1/5 John Toohey <jt(a)parspro.com> > But this doesn't get 1-1, 2-2 or 3-3. I had tried something like that use > #reverse on each combination, but #combinations does not generate > -duplicates- from the number sequences. > > > On Tue, Jan 5, 2010 at 12:16, Henrik Johansen < > henrik.s.johansen(a)veloxit.no> wrote: > >> On Jan 5, 2010, at 5:19 32PM, John Toohey wrote: >> >> > Hi, >> > I'm trying to generate the permutations of a sequence of numbers. I've >> found # permutationsDo: aBlock, but this generated the permutation of the >> entire sequence. I want to do it in groups of 2. The #combinations: >> anInteger atATimeDo: aBlock method does what I need for combinations, but I >> cannot find the equivalent for permutations. Does something like this exist >> in the standard library, or in an external package. >> > >> > -- >> > -JT >> >> How about: >> #(1 2 3 4) combinations: 2 atATimeDo: [:array | array permutationsDo: >> [:each | Transcript show: each printString; cr]] ? >> >> Cheers, >> Henry >> >> >> _______________________________________________ >> Pharo-project mailing list >> Pharo-project(a)lists.gforge.inria.fr >> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project >> > > > > -- > -JT > > > > _______________________________________________ > Pharo-project mailing list > Pharo-project(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project > -- CÃ©drick

Jan. 5, 2010

Re: [Pharo-project] Dev images with or without tests?

by StÃ©phane Ducasse

+1 especially if we have a script to take a rc and produce a dev without test if necessary > I vote for their presence for all the reasons you mentioned. > > Cheers, > Doru > > > > On 5 Jan 2010, at 22:11, Mariano Martinez Peck wrote: > >> Ok, that's the big questions. Until now, the Dev image only >> cointained the tests of the PharoCore image. Actually, it also >> included the tests fro those projects where the test were in the >> same package (for example, RoelTyer). Now, with the new Metacello >> infrastructure, you will be able to take a PharoCore and evaluate: >> >> (ConfigurationOfPharo project version: '10496') load (which it >> will load the defualt stuff, in this case, all the stuff for a dev >> image) >> >> or >> >> (ConfigurationOfPharo project version: '10496') load: >> 'StandardDevImageWithTests' >> >> which will loads not only all the packages, but also ALL the test of >> all packages (all packages that have tests). >> >> So, my question is, should the test be included in the default dev >> image? >> >> I think that yes, because of the following: >> >> 1) Tests are useful for documentation >> 2) It encourages people to run the tests before committing something >> (that may break something) >> 3) You can see the state of the image you are using. >> 4) Because they are not bad >> 5) Because if you don't want to load the test, just add a simple >> 'StandardDevImageWithoutTests' after the load. >> 6) I don't think they take a lot of size in an image. In addition, >> it is a DEV image, not production. In production is logic that you >> don't want the tests and that may be even done with something >> similar to ScriptLoader >> cleanUpForProduction >> >> So..what do you think ? >> >> Cheers >> >> Mariano >> >> _______________________________________________ >> Pharo-project mailing list >> Pharo-project(a)lists.gforge.inria.fr >> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project > > -- > www.tudorgirba.com > > "Beauty is where we see it." > > > > > _______________________________________________ > Pharo-project mailing list > Pharo-project(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

Jan. 5, 2010

[Pharo-project] Interim build server

by Michael Roberts

Hi, following on from the timely Hudson post ;-) I too have been working on continuous integration. I am more interested in identifying & integrating the various bits of plumbing first rather than the build solution itself. I figured the infrastructure needed to be understood first, and there could be value in such an interim solution To this end I have: -written a simple image builder that updates the image in place. -running remotely on a Linux host I have borrowed access -building PharoCore 1.1 from #11110 onwards -started to publish images at http://pharo.gforge.inria.fr/experimental/ remaining is to rename the core images to unstable and then look at running some tests. This has exposed various issues, mainly image timeout caused by the various dialogs that can popup when filing in code. so 1) if anyone would like to test the integrity of these experimental auto-build images, please do 2) discuss if something in this namespace http://pharo.gforge.inria.fr/<> is suitable for automatic publishing - i am not sure how you automatically create gforge "releases" but for core images perhaps this is not needed 3) find a real server we can run this on, (if this is considered worthwhile). My requirements are Linux + daemontools + vnc server - i can host my solution for a while, we need something strategic then once we have a server I'm happy to look at Hudson or whatever we consider the right tool for doing this. I have also been playing with Mason and sent some feedback to Colin. Like i say, for me the real issues are running a Pharo image completely automated. Cheers, Mike

Jan. 5, 2010