October 2011 - Pharo-users - lists.pharo.org

How to get temp names of BlockClosure
by Norbert Hartl Oct. 17, 2011

Oct. 17, 2011

I'm trying to learn how to get temp names from a BlockClosure. Well, I try to be able to treat BlockClosures the same way than CompileMethod when it comes to arguments. I can do [:a :b :c| ] asContext tempNames #('a' 'b' 'c') but I'm not sure if this is the best to go. While looking I found [:a :b :c| ] method tempNames #('[a' 'b' 'c]') For me there are two errors. The obvious one is that the first element is [a instead of a (it splits the string "[a b c]" by spaces). The second point is that "method" is defined as "outerContext method" and I don't understand why it is visible as temps there. It gets the information from the method trailer. Is this a reminiscent from the pre-closure times? thanks, Norbert

1 0

[ANN] Language of Languages Demo at SPLASH '11
by Nick Chen Oct. 17, 2011

Oct. 17, 2011

Dear Smalltalkers Jamie Douglass, Ralph Johnson and I are working on an experimental language workbench called Language of Languages (LoLs) that focuses on unifying concepts (meta-models) across different notations. This affords the user greater flexibility for editing the model in a variety of different notations while maintaining consistency. We are building this language workbench on top of Pharo Smalltalk[1] and we are using the Glamour engine[2] to build some of the language browsers. If anyone is attending SPLASH[3] this year, we'd like to invite you to one of our demo sessions. You can find a schedule of our demos at http://languageoflanguages.org/news. You can find more information about LoLs from our website at http://www.languageoflanguages.org. Hope to see you at SPLASH '11! -- Nick University of Illinois at Urbana-Champaign [1] http://www.pharo-project.org/home [2] http://www.moosetechnology.org/tools/glamour [3] http://splashcon.org/2011/ -- View this message in context: http://forum.world.st/ANN-Language-of-Languages-Demo-at-SPLASH-11-tp3910817… Sent from the Pharo Smalltalk Users mailing list archive at Nabble.com.

3 4

Planning a nice pharo conference :)
by StÃ©phane Ducasse Oct. 16, 2011

Oct. 16, 2011

Hi guys the last time when we met with sven, we came up with the idea of doing a nice pharo conference targeted at exchanging information, frameworks, tips and trick on server set up, community building, drinking beers and more. So we could host such conference in beginning of may for example. What do you think? Would you be interested to join a two-three days conference focusing on dev and business? Stef

4 3

preparing data for diagramms
by Norbert Hartl Oct. 15, 2011

Oct. 15, 2011

Hi, is there a toolkit that helps to prepare data for diagramms and such. I mean accumulating in time series like day, week, ..., calculating average, sums,.. and scaling data. Something that just eases to typical stuff one needs to go from log data to visual representation. thanks in advance, Norbert

3 4

RESTful Pharo app (not Seaside)
by Sean P. DeNigris Oct. 13, 2011

Oct. 13, 2011

What's the simplest/best way to communicate with a Pharo desktop app via RESTful API? Is anyone doing this? I frequently want to communicate with a running image and REST seems a little easier than XMLRPC which I've been using. I saw SimpleRestServer (http://forum.world.st/Simple-rest-server-td3571503.html) but was wondering if anyone has some good advice. Thanks. Sean -- View this message in context: http://forum.world.st/RESTful-Pharo-app-not-Seaside-tp3900373p3900373.html Sent from the Pharo Smalltalk Users mailing list archive at Nabble.com.

6 12

Re: [Pharo-users] red-black and other trees
by Paul DeBruicker Oct. 13, 2011

Oct. 13, 2011

I created the CenteredIntervalTree repo on Squeaksource. The CenteredIntervalTree is about 2x slower to create than a DateTree but 50-500x faster to query in Pharo 1.3 on Linux. The speed difference remains relatively constant as the tree size grows. I only tested up to 1,000,000 timespans. It works with instances of class Interval and Timespan. Converting the Timespans to Intervals using #asSeconds makes the creation process faster than the DateTree creation process. I used the code below to approximate the speed differences. You need to have the BTree package loaded from SqueakSource to try the example below. Paul |timespanDatabase dateTree centeredIntervalTree timespansOfInterest timingResults results size iterations interval tmpResults citResults dtResults| size:=10. iterations:=10. interval:=1 to: size. results:=Dictionary new. timingResults := Dictionary new. timespanDatabase:= OrderedCollection new. interval do:[:eachMonth | interval do: [:eachDay| interval do:[:eachHour| timespanDatabase add:(Timespan starting: (DateAndTime now + (eachMonth*30) days + eachDay days + eachHour hours) duration: 2 hours)]]]. timespansOfInterest :=OrderedCollection new. timespansOfInterest add: (timespanDatabase at: 20). interval do:[:each | timespansOfInterest add: timespanDatabase atRandom]. results at: 'timespanDatabase-size' put:timespanDatabase size. "results at: 'timespanDatabase' put: timespanDatabase ." timingResults at:'dateTree-creation' put:([iterations timesRepeat:[dateTree:=DateTree new. timespanDatabase do:[:eachTimespan| dateTree at: eachTimespan put: eachTimespan]]] timeToRun). timingResults at:'dateTree-query' put:([iterations timesRepeat:[tmpResults:=Dictionary new. timespansOfInterest do: [:each | tmpResults at: each put: (dateTree during: each)]]] timeToRun). results at:'dateTree-results' put: tmpResults. timingResults at:'centeredIntervalTree-creation' put:([iterations timesRepeat:[centeredIntervalTree :=CenteredIntervalTree from: timespanDatabase ]] timeToRun). timingResults at:'centeredIntervalTree-query' put:([iterations timesRepeat:[tmpResults:=Dictionary new. timespansOfInterest do:[:each | tmpResults at: each put: (centeredIntervalTree intervalsThatInclude: each) asSet]]] timeToRun). results at:'centeredIntervalTree-results' put: tmpResults. citResults:=OrderedCollection new. (results at: 'centeredIntervalTree-results') keysAndValuesDo: [:timespan :col| citResults addAll:col]. dtResults:=OrderedCollection new. (results at: 'dateTree-results') keysAndValuesDo: [:timespan :col| dtResults addAll:col]. dtResults size = citResults size ifFalse:[self halt.]. (dtResults includesAllOf: citResults) ifFalse:[self halt]. (citResults includesAllOf: dtResults) ifFalse:[self halt]. timingResults On 10/13/2011 03:03 AM, pharo-users-request(a)lists.gforge.inria.fr wrote: > Date: Wed, 12 Oct 2011 22:27:26 +0200 > From: Lukas Renggli<renggli(a)gmail.com> > Subject: Re: [Pharo-users] red-black and other trees > To: A friendly place where any question about pharo is welcome > <pharo-users(a)lists.gforge.inria.fr> > Message-ID: > <CANoDCG-osEdoF85ikhn2xShe6T=gpu9hKe4r9w-zVGUJxrp5hw(a)mail.gmail.com> > Content-Type: text/plain; charset=ISO-8859-1 > > The BTree package contains a class called DateTree for exactly that > purpose. Works very well with huge collections of date intervals. > > Lukas > > On 12 October 2011 22:13, Paul DeBruicker<pdebruic(a)gmail.com> wrote: >> > Hi - >> > >> > I am considering implementing a Centered Interval Tree >> > (http://en.wikipedia.org/wiki/Interval_tree) because I have to get the set >> > of Timespans a particular Timespan overlaps from an arbitrary set of >> > Timespans. >> > >> > >> > In looking around I can only find the BTree implementation on SqueakSource, >> > and the Left-Leaning-Red-Black-Tree linked to in the following list posting: >> > http://forum.world.st/Interesting-survey-about-smalltalk-tp2261561p2263370.… >> > >> > >> > Does Pharo either not need/benefit from the various kinds of trees? >> > >> > Or is it trivial to implement/adapt the BTree to behave as a Red-Black Tree >> > (or, in my case, the Centered Interval Tree) and as such I just need to >> > learn how to do that? >> > >> > Thanks >> > >> > Paul >> > >> >

1 0

[ANN] Pharo Glorp port release
by Guillermo Polito Oct. 13, 2011

Oct. 13, 2011

Hi! We are reaaaaally happy to announce (at last) a very very stable release of the new Glorp port :D. We've done a lot to achieve the following results with Glorp: OpenDBX Oracle: - 7 errors related to Blobs, which are not supported by opendbx. - 1 failure related to timestamp cast (and I don't understand the test intention yet :P) OpenDBX MySql: - 100% green tests :D OpenDBX PostgreSQL: - 7 errors and 1 failure related to Blobs, which are not supported by opendbx (again). OpenDBX SQLite: - 6 failures related to large objects, which are not supported by opendbx (again again).. 2 (expected) failures related to times & dates. Native PostgresV2 Driver: - 2 failures with blobs So far, some FAQs: *Â¿What is the difference between this port and the old one?* Almost 4 years of work of the VW team with tons of bugfixing (see the last 4 years release notes from Glorp for more detail :P) *Â¿How can I load it?* If you want to load Glorp + OpenDBX, you can try the following: *Gofer it squeaksource: 'MetacelloRepository'; package: 'ConfigurationOfGlorpDBX'; load. (((Smalltalk at: #ConfigurationOfGlorpDBX) perform: #project) perform: #version: with: #stable) load* Also, you can try loading only Glorp (maybe because you want to integrate it with your own database driver): Gofer it squeaksource: 'MetacelloRepository'; package: 'ConfigurationOfGlorp'; load. (((Smalltalk at: #ConfigurationOfGlorp) perform: #project) perform: #version: with: #stable) load *Â¿How did we test it?* We've tested it in Pharo 1.2 and Pharo 1.3 + Jenkins CogVms. *Â¿What's coming next?* *We are working for your happiness (?).* Hehe, our next step is to release the DBXMagritte project, which aims to allow people working in the Software business to create easy CRUD -the same scaffolding idea from RoR for example- including some of the following features: - (create a class model + magritte descriptions) from a database schema - [if you already have a descripted class model] use magritte descriptions to create a database schema - use magritte descriptions to create glorp mappings And then, you can use the same descriptions for your Seaside app, or in every magritte-friend framework you like ;) So, thanks again to ESUG for sponsoring the project :). Guillermo, Santiago, Gisela, Mariano, Esteban and Diogenes

3 3

Oct. 13, 2011

Hello everybody We got a final discussion with INRIA representatives and we will now set up a consortium for pharo. INRIA will support the consortium by funding on advanced engineer for a year (may be two if the consortium work well). So this is quite good around 55Keuros on consortium. There some strings attached: We will not be able to pay remote development since this is regulated under the public market french laws but we will be able to hire people (on french rate if they work in France). We will have to see the legal period to spent the money. The general idea is to see how the consortium evolves: ideally we would like to see if we can get it outside inria within a five years period. Now what we plan to do: - We would like to make the list of companies that sign the agreement letter public. Because this is important that other companies joins the consortium. - We will offer a free participation to any individual. - We are wondering if we could have a academic fee for the institution that want it: we have to discuss the fee. Any suggestions would be good: I was thinking that 500 Euros would be good for rich countries and 200 for the other ones. Now tell us. - We could also have way to pay a license or a way to give money to the consortium. Now the complex part: - we would like to collect money from companies and again we will need your help to understand what is the correct model. We are thinking: single person company: starting at 300 Euros "small company": starting at 2500 Euros large companies: starting at 5000 Euros. We would like to keep the model really simple and transparent. Now our goal is to be able to pay an engineer like igor full time because there is so much to do. But each day we get closer to what we want. Stef and Marcus.

10 15

red-black and other trees
by Paul DeBruicker Oct. 12, 2011

Oct. 12, 2011

Hi - I am considering implementing a Centered Interval Tree (http://en.wikipedia.org/wiki/Interval_tree) because I have to get the set of Timespans a particular Timespan overlaps from an arbitrary set of Timespans. In looking around I can only find the BTree implementation on SqueakSource, and the Left-Leaning-Red-Black-Tree linked to in the following list posting: http://forum.world.st/Interesting-survey-about-smalltalk-tp2261561p2263370.… Does Pharo either not need/benefit from the various kinds of trees? Or is it trivial to implement/adapt the BTree to behave as a Red-Black Tree (or, in my case, the Centered Interval Tree) and as such I just need to learn how to do that? Thanks Paul

2 1

Loading Seaside 3.0 cripples System browser
by Charles Monteiro Oct. 12, 2011

Oct. 12, 2011

I downloaded the Pharo 1.3 one click. I then attempted to load Seaside using the following mechanism: Gofer new squeaksource: 'MetacelloRepository'; package: 'ConfigurationOfSeaside30'; load. ((Smalltalk at: #ConfigurationOfSeaside30) project latestVersion) load. It complained twice about some package dependencies i.e. related to RBxxx packages. I proceeded through both and it seemingly terminated fine. Seaside was loaded including the Seaside control panel. However, doing a right mouse over the leftmost panel (Categories) resulted in an MNU. Also stuff like syntax highlighting became inactive. Is there a way to load the latest and greatest Seaside into Pharo 1.3 with "all" pre-req dependencies resolved ? i.e. how can I load Seaside clean ? thanks in advance -Charles -- View this message in context: http://forum.world.st/Loading-Seaside-3-0-cripples-System-browser-tp3880813… Sent from the Pharo Smalltalk Users mailing list archive at Nabble.com.

5 11