- Pharo-dev - lists.pharo.org

New Website online
by Marcus Denker May 9, 2014

May 9, 2014

Hi, The new website is online! http://pharo.org For now, pharo-project.org still is the old website. Next: pharo-project.org should forward to pharo.org, it will still be available as old.pharo-project.org for some time. Thanks a lot to Nico http://nicolas-petton.fr, Esteban http://smallworks.eu, Damien http://damiencassou.seasidehosting.st and Aurelia http://aurelia.saout.fr Marcus

14 22

Roassal visualizing a corner of the Universe
by Alexandre Bergel May 9, 2014

May 9, 2014

Hi! Weâve just added a new post that contains nice screenshots and (short) explanation about galaxies. https://www.facebook.com/ObjectProfile/posts/628687443884523?stream_ref=10 Faviola is doing a truly fantastic work. Cheers, Alexandre -- _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: Alexandre Bergel http://www.bergel.eu ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.

6 16

[ANN] DBXTalk CI available
by Guillermo Polito May 9, 2014

May 9, 2014

Hi! As I've seen lately too much movement and questions I couldn't answer about dbxtalk like "does it work?" "does it load?" "what is the status?" and answering them is a bit difficult because it depends on platform, database, pharo version, etc... I started a while ago to setup a CI server at Inria for it: https://ci.inria.fr/dbxtalk/ As you will see, I set up jobs for: - the opendbx library - the dbxtalk bindings to the opendbx library - glorp - the native postgresv2 driver [1] - the native mysql protocol from [2] And I'm testing them only in *Pharo3 *and for the following databases: - sqlite3 - mysql (version 5) - oracle - postgres My plan was to release as soon as I had windows support, but right now I cannot create new slaves for it :). So... this includes *only linux CI support.* Note aside, you'll find the scripts that the jobs use here [3] =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- So... For those that were asking if DBXTalk is working, it is :). But also, you'll see not all tests are green: That's probably because of some changes between pharo2 and pharo3. I didn't dive into them nor wanted to fix them until I had the CI jobs (and so, the pleasure to make the build green ;)). These are the next steps: - add tests to test the startup (that looks a bit buggy now) - Fix the not working tests - Make the native Mysql and Postgres drivers polymorphic with the OpenDBX one (and so, we can (a) reuse all tests for all drivers and (b) have glorp for free for them and others that you could imagine) - Add windows slave (that will take some time, configuring the databases, getting everything ready to compile... etc...) Of course, if you want to contribute either to fix the jobs, or to configure new stuff, you're welcome! Guille [1] http://www.smalltalkhub.com/#!/~PharoExtras/PostgresV2 [2] www.squeaksource.com/StdbMysqlProtocol.html [3] https://github.com/guillep/DBXTalkCI

3 2

Application development - import image and mark it up with vector graphics and create PDF
by Reg Krock May 9, 2014

May 9, 2014

Hi, What is involved in creating an application in Pharo, part of would which do the following: 1) Copy an image (ex: jpg) into a a window area. 2) Allow the user to add basic shapes on top of the image - square, circle, with text inside the shape. 3) Save this and generate a PDF. There is business domain logic surrounding this. Regards, Reg

3 3

DBXTalk continued
by Norbert Hartl May 9, 2014

May 9, 2014

There were some success reports using DBXTalk in pharo 3.0. Sadly Iâm not that lucky to get it running properly. I loaded the latest version from smalltalkhub. I exchanged the version reference for TalkFFI to 1.4 (tested 1.5 as well no difference). I loaded bleddingEdge and did afterwards NBPharoOpenDBX installAsCurrent Running the tests I 118 run, 50 passes, 68 errors but the number changes when running the suite more often. There is a sporadic error âLost connection during queryâ which occurrs randomly thus the changing number in error fails. And the tests are trying to test more backends. I can see errors on the console saying that libraries cannot be found. I donât know yet if this influences the tests as well. The described scenario is running on an ubuntu 13.04 in virtualbox with a fresh 3.0 image. I was trying to install it on my Mac laptop but had too much problems convincing homebrew to build opendbx as 32 bit lib. So I thought it would be less pain to do it on linux which will be the deployment platform anyway. Has anybody seen similar problems. Or can any linux user tell me what library versions are used in the success case? thanks, Norbert

3 4

Re: [Pharo-dev] dbxtalk working on pharo 3.0
by Guillermo Polito May 9, 2014

May 9, 2014

Hi! I'd love to write a test along with the fix, so we can improve the CI infrastructure I set up (see post "[ANN] DBXTalk CI available"). Which is your smalltalkhub user so I give you commit grants? Guille On Wed, May 7, 2014 at 7:39 PM, Michael Roland <mroland98343(a)gmail.com>wrote: > No, the change has not been committed. No one else has verified that this > change works. It works for me. > > The change to the doit method performs the step that Tudor Girba > mentioned in the first post: NBPharoOpenDBX installAsCurrent. So I think > the change is correct. > If someone will verify that the change works on their system, I will ask > for permission to commit the change. > > > On Wednesday, May 7, 2014 9:28:11 AM UTC-4, Mariano Martinez Peck wrote: > >> >> >> >> On Fri, Mar 28, 2014 at 4:18 PM, Michael Roland <mrolan...@gmail.com>wrote: >> >>> >>> >>> On Thursday, February 13, 2014 10:26:55 AM UTC-5, Tudor Girba wrote: >>>> >>>> Hi, >>>> >>>> I just retried DBXTalk, and it now magically works! >>>> >>>> Does anyone know why? :) >>>> >>>> I did this: >>>> Gofer new >>>> smalltalkhubUser: 'DBXTalk' project: 'DBXTalkDriver'; >>>> package: 'ConfigurationOfOpenDBXDriver'; >>>> load. >>>> #ConfigurationOfOpenDBXDriver asClass loadBleedingEdge. >>>> >>>> Anyway, the only glitch I found is that the OpenDBX current wrongly >>>> returns "NBPharoOpenDBX class" instead of the instance. >>>> >>>> I could not figure out where the initialization bug comes from, but I >>>> worked around it with: >>>> NBPharoOpenDBX installAsCurrent. >>>> >>>> I tested the functionality with an Oracle backend, and it just worked. >>>> >>>> Thanks! >>>> >>>> Doru >>>> >>>> >>>> -- >>>> www.tudorgirba.com >>>> >>>> "Every thing has its own flow" >>>> >>> >>> I tried DBXTalk today on Pharo 3 and got the NBPharoOpenDBX class error >>> too. >>> This is my understanding of what causes the error. >>> >>> >>> The value returned by OpenDBX class>>#current is initialized in >>> ConfigurationOfOpenDBXDriver>>#setNBOpenDBX. >>> >>> #setNBOpenDBX has one line: >>> (Smalltalk at: #OpenDBX) current: (Smalltalk at: #NBPharoOpenDBX). >>> >>> This sets OpenDBX Current to an NBPharoOpenDBX class instead of an >>> NBPharoOpenDBX instance. >>> >>> Changing the statement to: >>> (Smalltalk at: #NBPharoOpenDBX) installAsCurrent >>> >>> indirectly sets OpenDBX Current to an instance of NBPharoOpenDBX so that >>> OpenDBX class>>#current will correctly return an instance of >>> NBPharoOpenDBX. >>> >>> Using OpenDBX class>>#installAsCurrent is needed because the OpenDBX >>> class has >>> two class variables: CurrentLibraryFactory and Current. >>> CurrentLibraryFactory stores a class and Current stores an instance of >>> that class. >>> The only way to set CurrentLibraryFactory is with the OpenDBX >>> class>>#installAsCurrent method. >>> Using OpenDBX class>>#current: to set the instance directly can leave >>> the CurrentLibraryFactory and Current variables out of sync. >>> >>> >> Thanks! Michael was this committed to the repo? >> >> >> >> -- >> Mariano >> http://marianopeck.wordpress.com >> > -- > You received this message because you are subscribed to the Google Groups > "DBXTalk" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to dbxtalk%2Bunsubscribe(a)googlegroups.com. > For more options, visit https://groups.google.com/d/optout. >

1 0

Starting making bindings for OpenCL using NativeBoost.
by Ronie Salgado May 8, 2014

May 8, 2014

Hello, I started working in making some basic bindings for OpenCL using NativeBoost. Currently I can run basics OpenCL OpenCL kernels. I tried to generate quickly some procedural images which was a success. I am going to be showing some more cool stuffs when mixing OpenCL+OpenGL+SDL2. Currently uploading my work into http://smalltalkhub.com/#!/~ronsaldo/OpenCL. No configuration or stable release yet. Greetings, Ronie

4 5

Verifying Ubuntu downloads
by Sean P. DeNigris May 8, 2014

May 8, 2014

After heartbleed, I've become more security-conscious/paranoid than usual. Here's a little snippet that compares a downloaded Ubuntu iso file to the MD5 provided via SSL by Ubuntu. Not foolproof, but more peace of mind than crossing fingers and hoping for the best on the unsecured http mirrors... Gofer it smalltalkhubUser: 'SeanDeNigris' project: 'Superuser'; configuration; loadStable. "#inform: success or signal an error" SuUbuntuHash verify: 'ubuntu-12.04.4-desktop-amd64.iso' asFileReference. ----- Cheers, Sean -- View this message in context: http://forum.world.st/Verifying-Ubuntu-downloads-tp4754637.html Sent from the Pharo Smalltalk Developers mailing list archive at Nabble.com.

4 11

Pharo 4.0 Inbox
by Sean P. DeNigris May 8, 2014

May 8, 2014

I created http://smalltalkhub.com/mc/Pharo/Pharo40Inbox/main and made it public writable ----- Cheers, Sean -- View this message in context: http://forum.world.st/Pharo-4-0-Inbox-tp4758294.html Sent from the Pharo Smalltalk Developers mailing list archive at Nabble.com.

1 0

About Spec second interpretation loop between Adapter and specific back end
by stepharo May 8, 2014

May 8, 2014

Hi Ben When igor integrated his textEditor with Spec we saw that the relation between a model and its adapter is not direct but via another interpretation. For example MorphicButtonAdapter class>> defaultSpec ^ {#PluggableButtonMorph. #color:. Color white. #on:getState:action:label:menu:. #model. #state. #action. #label. nil. #getEnabledSelector:. #enabled. #getMenuSelector:. #menu:. #hResizing:. #spaceFill. #vResizing:. #spaceFill. #borderWidth:. #(model borderWidth). #borderColor:. #(model borderColor). #askBeforeChanging:. #(model askBeforeChanging). #setBalloonText:. { #model . #help}. #dragEnabled:. #(model dragEnabled). #dropEnabled:. #(model dropEnabled). #eventHandler:. { #EventHandler. #on:send:to:. #keyStroke. #keyStroke:fromMorph:. #model }} We do not understand why it is needed. Because the adapter is specific to a platform and to a model so For example why do we have #borderWidth:. #(model borderWidth). while we could have MorphiButtonAdapter>>borderWitdh ^ self model borderWidth Having this extra interpretation loop - is costly - blur the API of the class (as a programmer I do not see the exact API) the cross navigation gets confused. - makes the code more complex to understand for example #eventHandler:. { #EventHandler. #on:send:to:. #keyStroke. #keyStroke:fromMorph:. #model }} In addition is not a kind of static meta description so we will always need interpretation. So I would really like to see if we can avoid this second interpretation. Stef

1 0