- Pharo-users - lists.pharo.org

[ANN] MessageFlowBrowser for Pharo 7 and 8
by Torsten Bergmann July 12, 2019

July 12, 2019

Hi, I updated my MessageFlowBrowser today to work on Pharo 7 and 8, see attached screenshot or youtube video on https://www.youtube.com/watch?v=DRd_bzGocQg It now also works in Calypso and you just need to right click on a method in the method pane or code pane and select "Message Flow" close to Senders, Implementors and other. You can load the tool using Metacello new repository: 'github://astares/Pharo-MessageFlowBrowser/src'; baseline: 'MessageFlowBrowser'; load I also now moved repo from http://www.smalltalkhub.com/#!/~TorstenBergmann/MessageFlowBrowser to the new location https://github.com/astares/Pharo-MessageFlowBrowser on GitHub for future development. Have fun, T.

4 3

Re: [Pharo-users] Test Completion has been added to Pharo 8 - please test!
by Cyril Ferlicot July 12, 2019

July 12, 2019

On Wed, Jul 3, 2019 at 2:29 AM Myroslava Romaniuk via Pharo-users <pharo-users(a)lists.pharo.org> wrote: > > Hi all, > > we added an intermediate version of the upcoming test completion to Pharo 8, it would be cool if you tested it and let us know the results. Here's a blog post about it - link, and here's a tweet I'd be grateful if you shared - link so perhaps people who don't read the mailing list would also see it. In the blog post you can find more info how to enable the completion in the settings and details about the whole thing in general. If something doesn't work, or if something works differently than it used to in the old completion (both if you don't like it or perhaps if you prefer it) let me know either replying here, on twitter, or submitting an issue request directly for the github repository here - link. > Hi, Since the new code completion is enable I have problem with variables completion. I saw it the most with temporary variables that are not proposed for completion anymore. This is really annoying since I often have long variables names to be explicit in my code :( I hope it can help you improve it :) > Thanks very much, > Myroslava -- Cyril Ferlicot https://ferlicot.fr

2 1

Re: [Pharo-users] [GSoC blog post] Binary Search Trees
by Ben Coman July 11, 2019

July 11, 2019

On Thu, 11 Jul 2019 at 00:53, Alexandre Bergel via Pharo-users <pharo-users(a)lists.pharo.org> wrote: > > Hello Smiljana, > > Thanks for having written down this document. I am not expert in algorithm, so I would consider myself a simple user. I have developed complex software for some times and I have never seen the need of having a binary search tree. I guess this is probably partly because of my lack of expertise in binary search tree and partly because experts in binary search trees assume that people know what it is about and in what it is useful. > > My question is, when should a programmer ever need to use binary search tree? Can you add some examples on what these trees are good for, and how an average programmer should look into it. I think this will be a valuable and easy way to expand your blog. My "computer science" is a bit rusty but Binary Search Trees and their generalisation to B-Trees is Databases 101. Almost guaranteed you are "using" them with any database you use, you're just not directly exposed to them. If you skim "Redesigning String Data Structures to Exploit Cache" (https://people.eng.unimelb.edu.au/jzobel/fulltext/acmjea10.pdf) looking at the performance graphs, for "unordered searching" the best bet for many applications is probably our hash-based Dictionary if that covers all your needs. BSTs are more memory efficient than Dictionarys, but less cache-friendly. A few applications where BSTs can be useful: * doing range searches efficiently [1] * traversing elements in order, which is more difficult with hash tables [1] [2] * good for "order" statistics [2] like: - inexact search - finding closest lower and greater elements. - determining maximum & minimum elements - determining successor & predecessor elements - range queries - easy to do with BSTs and not a natural operation with Hash Tables. * Self-Balancing BSTs guarantee operations work in O(Logn) time. Hashing has Î(1) is average time but some particular operations are costly, especially when table resizing happen [2] but also they are slower so if the data is sufficiently random simple BSTs may do. [1] https://stackoverflow.com/questions/371136/binary-trees-vs-linked-lists-vs-… [2] https://www.geeksforgeeks.org/advantages-of-bst-over-hash-table/ This side-related article... "Number crunching: Why you should never, ever, EVER use linked-list in your code again" (https://kjellkod.wordpress.com/2012/02/25/why-you-should-never-ever-ever-us…) indicates that for many In-Memory operations you are better using sorted-Arrays (e.g. SequenceableCollection>>#findBinary:) than cache-invalidating pointer-base structures like Lists & Trees - thus the strong association of BSTs & BTrees with slower access disk-based databases. Also notice that additional BST features closely match common database operations. cheers -ben

1 0

Linux init script
by SebastiÃ¡n Filippini July 10, 2019

July 10, 2019

Hello everyone, I'm trying to write a Linux init script for a Pharo application. I'm using a bash script wrapper to launch pharo. This script executes the pharo script provided by get.pharo.org. This is my wrapper script: #! /bin/bash cd "$(dirname "$0")" ./pharo myapp.image start.st & PID=$! echo $PID > /var/run/myapp.pid The following process tree is created: Â ââbash,692 ./pharo myapp.image start.st Â âÂ Â ââpharo,701 --nodisplay myapp.image start.st Â âÂ Â Â Â Â Â ââ{pharo},730 And a /var/run/myapp.pid file is written with 692. Ok. The problem is that when I kill the process with pid 692 the process with pid 701 is not terminated. Any idea on how I can terminate the whole process tree? Best regards, SebastiÃ¡n Filippini

2 1

Re: [Pharo-users] [GSoC blog post] Binary Search Trees
by Richard Sargent July 10, 2019

July 10, 2019

On Wed, Jul 10, 2019 at 9:53 AM Alexandre Bergel via Pharo-users < pharo-users(a)lists.pharo.org> wrote: > Hello Smiljana, > > Thanks for having written down this document. I am not expert in > algorithm, so I would consider myself a simple user. I have developed > complex software for some times and I have never seen the need of having a > binary search tree. I guess this is probably partly because of my lack of > expertise in binary search tree and partly because experts in binary search > trees assume that people know what it is about and in what it is useful. > Hi Alexandre, I can tell you that GemStone/S makes extensive use of b-trees and variants in its internal data structures. Two examples are the object table itself and the index data structures. I agree that the average programmer may never have a need for them, but I don't like to close doors on (other) developers. Smiljana, Others have suggested things like the Null Object pattern for the left and right links, but an alternative or supplement to that would be a space optimizing leaf node which doesn't have instance variables for left and right sub-trees. But, don't go down the optimizing paths until you have good solutions for everything else. i.e. make right, then make it fast. I noted your code looks "C-like", which I suspect arises from being inspired by implementation in other languages. For example, I think only an empty tree should answer 0 for depth. A tree with one node should answer 1 while a tree with 2 or 3 nodes should answer 2, etc. If you drew a tree out on paper, how many levels would you see? That is the depth, in my opinion. Another example arises from one of the most powerful idioms in Smalltalk. Call it the #at:ifAbsent: pattern. (See also #detect:ifNone:) Someone else mentioned the passing around of nil and suggested you not do that. The #at:ifAbsent: pattern is the key to avoiding that: tell the receiver what *you* want to *do* if the thing can't be found by passing in your own block. Additionally, you will want your nodes to support various payloads. Right now the payload is the object held in the key. That works for identity structures, but it won't work for more advanced needs, such as one would find in a database index structure. Eventually, you will want the ability to have nodes that support additional attribute(s). I've found P. J Plauger's "0, 1, infinity" rule to be a superb design guideline. In short, once you go beyond having one of something (e.g. the key itself), you need to design things so that you can have an arbitrarily large number. The step from 1 to >1 is huge and messy if retrofitted, but easy when planned in. Eventually, you will need queries against the tree, allowing range selection, identity versus equality, etc. The tree should provide a basic infrastructure for searching, but a separate modelling of queries will be necessary. That may be out of scope for GSOC, of course. Good work and good luck with the rest of the project! > My question is, when should a programmer ever need to use binary search > tree? Can you add some examples on what these trees are good for, and how > an average programmer should look into it. I think this will be a valuable > and easy way to expand your blog. > > Cheers, > Alexandre > > On Jul 7, 2019, at 5:12 PM, Smiljana Knezev <smilja.knezev(a)gmail.com> > wrote: > > Dear all, > > I've written about implementing binary search trees: > https://pharokeepers.github.io/pharo/2019/07/07/SmiljanaBinarySearchTreeinP… > > Feedback and opinions is always welcome :) > > Best regards, > Smiljana Knezev > > >

1 0

Re: [Pharo-users] [GSoC blog post] Binary Search Trees
by Sven Van Caekenberghe July 10, 2019

July 10, 2019

Binary search is extremely useful to quickly find something in a sorted collection, like finding if a string or number is included in a large collection. BTW, we already have an implementation, see SequenceableCollection>>#findBinary: and friends. > On 10 Jul 2019, at 18:53, Alexandre Bergel via Pharo-users <pharo-users(a)lists.pharo.org> wrote: > > > From: Alexandre Bergel <alexandre.bergel(a)me.com> > Subject: Re: [Pharo-users] [GSoC blog post] Binary Search Trees > Date: 10 July 2019 at 18:53:01 GMT+2 > To: Any question about pharo is welcome <pharo-users(a)lists.pharo.org> > > > Hello Smiljana, > > Thanks for having written down this document. I am not expert in algorithm, so I would consider myself a simple user. I have developed complex software for some times and I have never seen the need of having a binary search tree. I guess this is probably partly because of my lack of expertise in binary search tree and partly because experts in binary search trees assume that people know what it is about and in what it is useful. > > My question is, when should a programmer ever need to use binary search tree? Can you add some examples on what these trees are good for, and how an average programmer should look into it. I think this will be a valuable and easy way to expand your blog. > > Cheers, > Alexandre > >> On Jul 7, 2019, at 5:12 PM, Smiljana Knezev <smilja.knezev(a)gmail.com> wrote: >> >> Dear all, >> >> I've written about implementing binary search trees: https://pharokeepers.github.io/pharo/2019/07/07/SmiljanaBinarySearchTreeinP… >> >> Feedback and opinions is always welcome :) >> >> Best regards, >> Smiljana Knezev > > >

1 0

[GSoC blog post] Binary Search Trees
by Smiljana Knezev July 10, 2019

July 10, 2019

Dear all, I've written about implementing binary search trees: https://pharokeepers.github.io/pharo/2019/07/07/SmiljanaBinarySearchTreeinP… Feedback and opinions is always welcome :) Best regards, Smiljana Knezev

5 9

ESUG Innovation Technology Awards Call for submissions
by N. Bouraqadi July 9, 2019

July 9, 2019

Dear Smalltalkers, Developers are invited to submit their Smalltalk-related project to the 16th edition of the ESUG Innovation Technology Awards. The top 3 teams with the most innovative software will receive, respectively, 500 Euros, 300 Euros and 200 Euros. Submit now! Instructions are available at: https://esug.github.io/2019-Conference/awardsCall2019.html Noury

1 0

Re: [Pharo-users] [Pharo-dev] [ANN] Pharo Launcher 1.9.1 released
by Christophe Demarey July 9, 2019

July 9, 2019

> Le 9 juil. 2019 Ã 07:57, StÃ©phane Ducasse <stephane.ducasse(a)inria.fr> a Ã©crit : > > christophe > > the name of the archive is confusing to me. > what is the diff between x64 and x86 x64: 64 bits version x86: 32 bit version For OS X, choose one of these with a dmg extension. The recommended version is always the 64-bits one. > which one is for mac? > S > >> On 8 Jul 2019, at 10:32, Christophe Demarey <Christophe.Demarey(a)inria.fr <mailto:Christophe.Demarey@inria.fr>> wrote: >> >> Hi all, >> >> We are pleased to announce Pharo Launcher 1.9.1 release! >> The goal of this release is to have a better Windows support. Indeed Pharo Launcher now uses a new library OSWinSubprocess (https://github.com/pharo-contributions/OSWinSubprocess <https://github.com/pharo-contributions/OSWinSubprocess>) created for Pharo Launcher needs. This library allows to create Windows processes through ffi calls to the Windows API. It uses the unicode api, allowing paths with non-ascii characters. >> In the same time, we switched from OSProcess to OSSubprocess (Linux and OS X) that is now the preferred library to use to fork processes on Linux-based systems. >> >> This release is not yet the default download of Pharo Launcher on Pharo web site because we want more feedback before promoting it. >> Please use this link to install it: https://files.pharo.org/pharo-launcher/1.9 <https://files.pharo.org/pharo-launcher/1.9> and let us know if this version works fine/better for you. >> >> ps: For now, the binary is not signed but it will come soon. >> ps2: Thanks to Keldon Alleyne for his PR on Pharo Launcher offering the first FFI binding to Windows process API >> >> Regards, >> The Pharo team. >> > > -------------------------------------------- > StÃ©phane Ducasse > http://stephane.ducasse.free.fr <http://stephane.ducasse.free.fr/> > http://www.synectique.eu <http://www.synectique.eu/> / http://www.pharo.org <http://www.pharo.org/> > 03 59 35 87 52 > Assistant: Julie Jonas > FAX 03 59 57 78 50 > TEL 03 59 35 86 16 > S. Ducasse - Inria > 40, avenue Halley, > Parc Scientifique de la Haute Borne, BÃ¢t.A, Park Plaza > Villeneuve d'Ascq 59650 > France >

1 0

Change code critics to decrease size of navigation pane?
by PAUL DEBRUICKER July 9, 2019

July 9, 2019

Hi - Sometimes (all the times) I write method that violate a critics rule. The quality improvement pane, when it expands, shrinks the code editing pane. Can I change that so that it shrinks the height of the navigation panes instead e.g. the class list etc? Stopping writing "offensive" methods seems impossible. Thanks Paul

1 0