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

Re: [Pharo-project] Issue 3729 in pharo: Failing test: CollectionsTests.Arrayed.FloatArrayTest.testArithmeticCoercion

by pharo＠googlecode.com

Updates: Status: Closed Comment #4 on issue 3729 by marcus.d...(a)gmail.com: Failing test: CollectionsTests.Arrayed.FloatArrayTest.testArithmeticCoercion http://code.google.com/p/pharo/issues/detail?id=3729 fixed, verified in build #149

Feb. 25, 2011

Re: [Pharo-project] Issue 3730 in pharo: #KedamaFloatArray is referenced in WordArray

by pharo＠googlecode.com

Updates: Status: Closed Comment #2 on issue 3730 by marcus.d...(a)gmail.com: #KedamaFloatArray is referenced in WordArray http://code.google.com/p/pharo/issues/detail?id=3730 fixed, validated in #149

Feb. 25, 2011

Re: [Pharo-project] Issue 3712 in pharo: Still CustomMenu references in Pharo 1.2 dev

by pharo＠googlecode.com

Updates: Status: Closed Comment #6 on issue 3712 by marcus.d...(a)gmail.com: Still CustomMenu references in Pharo 1.2 dev http://code.google.com/p/pharo/issues/detail?id=3712 fixed as of #149

Feb. 25, 2011

Re: [Pharo-project] Is repository down for Magma?

by StÃ©phane Ducasse

> > To incorporate this into Pharo, see the changes in Squeak between: > > Collections-ul.403 and Collections-cmm.404. > > The changes are very simple. Basically, we simply allow the > OrderedCollection's internal 'array' to be specified by its > #arrayType, instead of being hard-coded to an Array. > > This allows WeakOrderedCollection to exist by only one single method, > an override of #arrayType on the class-side to specify WeakArray. > > I do hope Pharo will incorporate these changes and gain the benefit of > a weak OrderedCollection, not to mention, Magma...! C'mon Pharo, the > total gain in code-size is just one Class with a single one-line > method, plus one extra one-line method on OC. Chris since november and besides working pharo 1.2 I wrote 4 project proposals, a couple of article, bookchapters, reviews for PhD, letters of recommandations...... lng list of meetings..... So this is not that we do not want. This is that if people do not make that happens it is slow. >> the new ConfigurationOfMagma to the metacello repository and you'll can >> install Magma but it won't run anymore. > > True, WeakOrderedCollection is crucial for the new proxy bulk-become > performance improvement. what is it? >> (I have had time to check the >> WeakOrderedCollection issue only, I don't know if after fixing that >> other issues will remain because I haven't gone that far). > > Me either, we have to get this addressed first. Would someone in > Pharo help me lobby for this? Pharo does not have to be lobbied. Just open a bug entry put the code and the tests, make sure that it works well and discuss pros and cons in the mailing-list. > Ok. I hope if we can get past WeakOrderedCollection there won't be > any more special changes needed, but we'll have to see. Stef

Feb. 25, 2011

Re: [Pharo-project] Issue 3711 in pharo: DefaultStubs in Undeclared

by pharo＠googlecode.com

Updates: Status: Closed Comment #5 on issue 3711 by marcus.d...(a)gmail.com: DefaultStubs in Undeclared http://code.google.com/p/pharo/issues/detail?id=3711 fixed as of #149

Feb. 25, 2011

[Pharo-project] [TODO 1.2] Issues to fix before Release

by Marcus Denker

http://code.google.com/p/pharo/issues/list?can=2&q=milestone=1.2-DevImage ... with the most pressing being the failing build. Marcus -- Marcus Denker -- http://www.marcusdenker.de INRIA Lille -- Nord Europe. Team RMoD.

Feb. 25, 2011

Re: [Pharo-project] Is repository down for Magma?

by StÃ©phane Ducasse

BTW this is why I would like to have a copy of all the packages when we publish a configuration. So that the world can explode and the user does not get impacted. Stef > Unfortunately for all us, Chris deleted without notice, the versions > from Magma repository that were used by the Metacello configuration to > install Magma. So Magma isn't installable with Metacello anymore.

Feb. 25, 2011

Re: [Pharo-project] Is repository down for Magma?

by StÃ©phane Ducasse

In pharoTaskForces :) On Feb 23, 2011, at 5:17 PM, Miguel CobÃ¡ wrote: > Unfortunately for all us, Chris deleted without notice, the versions > from Magma repository that were used by the Metacello configuration to > install Magma. So Magma isn't installable with Metacello anymore. > > Last weekend I started a ground up rewrite for ConfigurationOfMagma with > support only for 1.2 (easier than fix the versions 1.1r43 1.1r44 that > could be installed with the configuration) but after creating > ConfigurationOfMaBase, ConfigurationOfMaClientServer and the new > ConfigurationOfMagma (everything loaded correctly) I got stuck by the > fact that the new Magma 1.2 is Squeak specific at the moment (it uses > WeakOrderedCollection that Chris pushed to Squeak trunk in december 2010 > and that is not in Pharo 1.1, nor in Pharo 1.2). So, I could to publish > the new ConfigurationOfMagma to the metacello repository and you'll can > install Magma but it won't run anymore. (I have had time to check the > WeakOrderedCollection issue only, I don't know if after fixing that > other issues will remain because I haven't gone that far). > > So, that is the state of affairs with the Magma on Pharo. Of course, you > can use the new ConfigurationOfMagma to load Magma on Squeak and it will > work good I think. > > I will upload the new ConfigurationOfMagma right now anyway because the > old ConfigurationOfMagma are broken without hope of fix (at least from > me). > > The way I am going to fix this, at least for Pharo 1.1.1 and Pharo 1.2, > is to put the changes that Magma needs in a separate package somewhere > (maybe PharoInbox or PharoTaskForces) and use it as a dependency in > ConfigurationOfMagma. This way it will work correctly. But I will do it > until tomorrow night. > > Cheers > > El miÃ©, 23-02-2011 a las 11:42 -0300, Casimiro de Almeida Barreto > escribiÃ³: >> While loading Magma via: >> >> ((Smalltalk at: #ConfigurationOfMagma) project latestVersion) load. >> >> I get: >> >> Error: Could Not Resolve: Ma exception handling [Ma exception >> handling-cmm.33] in /home/casimiro/Desenvolvimento/Pharo/package-cache >> http: >> >> And in debugger: >> >> linearLoadPackageSpec: packageSpec gofer: gofer >> >> where >> >> packageSpec is: >> >> spec >> name: 'Ma exception handling'; >> file: 'Ma exception handling-cmm.33'. > > -- > Miguel CobÃ¡ > http://twitter.com/MiguelCobaMtz > http://miguel.leugim.com.mx > > > >

Feb. 25, 2011

Re: [Pharo-project] FileSystem. Possible FSReference rename

by StÃ©phane Ducasse

Hi denis > I think FSLocator and FSReference are same objects: > FSReference = fileSystem + path > FSLocator = "some origin object" + path. > > I think "some origin object" can be implemented as special FSRelativeFileSystem. And fileLocator became fileReference. > > Now in "FSLocator vmDirectory" "some origin object" is symbol. And FileLocator>>resolve method perform some complex stuff through FSResolver hierarchy logic. > > What is purpose of FSResolver hierarchy? > What is purpose of #resolve: and #resolve methods? Why so many classes (in FileSystem) implement it: FSLocator, FSReferece, FSPath, FSFileSystem, FSResolver? I do not know :) I'm trying to add comments to every single method so that we can get a better understanding of the system and change it. Stef

Feb. 25, 2011

Re: [Pharo-project] Issue 3589 in pharo: ClassComment in OB

by pharo＠googlecode.com

Updates: Status: Closed Comment #9 on issue 3589 by marcus.d...(a)gmail.com: ClassComment in OB http://code.google.com/p/pharo/issues/detail?id=3589 (No comment was entered for this change.)

Feb. 25, 2011