January 2013 - Pharo-dev - lists.pharo.org

[Pharo-project] [update 2.0] #20476

by Marcus Denker

Jan. 16, 2013

[Pharo-project] [update 2.0] #20475

by Marcus Denker

Jan. 16, 2013

[Pharo-project] Fwd: Problem installing Neptuno

by Stephan Eggermont

Frank Church wrote: >Is it okay for me to say that something isn't right here? Yes, you are right. Unfortunately, changing this takes time. The infrastructure to do automated builds is currently Jenkins based, and does not provide automated feedback upstream (the blame plugin is not suitable for a project with as many downstream projects as Pharo). The number of projects with automated builds is also still limited, and there are a lot of projects with insufficient maintenance. As a result many packages don't get updated regularly, only when users notice something is wrong. The good news is that it is already a lot better than a few years ago, and we see much better stability/possibilities to change in the packages that are build automatically. Unfortunately, Neptuno is not yet build like that. Ideally, every Pharo project would have its own automated build server that provides (test) feedback upstream. Stephan Eggermont

Jan. 16, 2013

[Pharo-project] [update 2.0] #20474

by Marcus Denker

20474 ----- Issue 7263: clean spec from experimental packages http://code.google.com/p/pharo/issues/detail?id=7263& Issue 7259: Re-enable #signalProgress in MCHttpRepository>>#httpClient http://code.google.com/p/pharo/issues/detail?id=7259 Issue 7258: Simplify HTTP client usage in Unicode-Initialization http://code.google.com/p/pharo/issues/detail?id=7258 Issue 7241: Failing test: RBSmalllintTest>>testBadMessage Issue 7241: Failing test: RBSmalllintTest>>testBadMessage Issue 7047: Source pane font permanently changes after switching to Group view and selecting one http://code.google.com/p/pharo/issues/detail?id=7047 Diff information: http://ss3.gemstone.com/ss/Pharo20/Zinc-HTTP-MarcusDenker.336.diff http://ss3.gemstone.com/ss/Pharo20/Unicode-Initialization-MarcusDenker.4.di… http://ss3.gemstone.com/ss/Pharo20/Refactoring-Tests-Critics-MarcusDenker.8… http://ss3.gemstone.com/ss/Pharo20/Morphic-MarcusDenker.1323.diff http://ss3.gemstone.com/ss/Pharo20/Monticello-MarcusDenker.742.diff

Jan. 16, 2013

Re: [Pharo-project] [Update] Zinc-SSO OAuth 1.0 & OpenID Alpha Code & Live Demos

by Sven Van Caekenberghe

Hi Yanni, On 16 Jan 2013, at 07:25, Yanni Chiu <yanni(a)rogers.com> wrote: > On 09/01/13 3:46 AM, Sven Van Caekenberghe wrote: >> >> Basically you should be able to get all the bits (preferably in Pharo 2.0) by doing >> >> (ConfigurationOfZincHTTPComponents project version: '2.2') load: 'SSO' > > It works fine for me in Pharo-1.4, but I cannot build it using Jenkins because Metacello throws up a UI dialog asking whether to Load/Merge/Cancel the Zinc-HTTP package. > > The Zinc-HTTP package ends up dirty, after loading the > > Zinc-Character-Encoding-Core > Zinc-Resource-Meta-Core > > pre-req packages. These packages seem to be a re-packaging of the previous Zinc-HTTP package. > > Is this a limitation of Metacello not having atomic load? > Can the Metacello dialog be disabled? > Should the non-interactive UI manager be enhanced to somehow proceed? The easiest solution is to switch to 2.0 where this repackaging conflict is already resolved. For a non-interactive CI build based on Metacello loading, this should work: [ ConfigurationOfZincHTTPComponents load ] on: MCMergeOrLoadWarning do: [ :warning | Transcript crShow: 'Ignoring with Load anyway ', warning printString. warning resume: true ]. HTH, Sven -- Sven Van Caekenberghe http://stfx.eu Smalltalk is the Red Pill

Jan. 16, 2013

[Pharo-project] Memories on Andreas by Yoshiki

by Janko MivÅ¡ek

I just read them so I hope Yoshiki Ohshima won't mind to forward his memories on Andreas to the mailing lists too. http://d.hatena.ne.jp/squeaker/20130115#p2 Yoshiki wrote: Andreas Raab, my friend and colleague of fourteen years, just passed away (he was only in his mid- 40's). He had a razor-sharp brain, and could write best-quality code. Not only that, he could manage projects and get a group of people to work. Alan Kay and David Smith say that Andreas is one of the top three programmers they have ever met. This says a lot. The way I got to know Andreas was through a software project called Squeak. He ported the Squeak virtual machine to Windows while he was a Ph.D student at the Magdeburg university in 1997. The core team members of Squeak, led by Alan Kay, were very much impressed with his talent. They basically had no way to let him go somewhere else. So when Andreas graduated, they just hired him and took him to California. It didn't take long that he became the productive members of the core team. I was also a Ph.D student at the Tokyo Institute of Technology around that time. Prof. Satoshi Matsuoka had a connection to a member of the Squeak Central, and that led me to try to port the virtual machine to a PDA called Sharp Zaurus. In 1998, I managed to make it (barely) work, much of it can be attributed to reading the code Andreas had written. By doing so, I got to start communicating with the Squeak group, and, eventually I got a job at the group at Disney when the luck and twists of events prevailed. Whenever I witnessed the brilliance of Andreas, I always thought "Andreas set a precedence of a young guy who ported the VM could do so many other things brilliantly. He paved the way and let me sneak in here, when I am no match with him." Also in this sense, "I would not be myself here today if he was not there." He was from former East Germany and as far as I know he even did military service (shorter than normal as benefit of being a smart boy). I thought I see the remnant of training there occasionally. When five of us got on to a small car and Andreas got the middle seat, or when we got on a crowded Shinkansen train, he just could stay still for a long time without complaining. I imagined that this has something to do with his training. Thinking about this root, one thing to be said is that "he was somebody who could transform himself"; he started being not only a smart guy but added "more depth to his character", stop smoking after a while in California, and endorse the capitalist ideals when he started a start up company (actually, companies). Not only that, he became a guy of love and family! I would imagine that many of us seeing him over time thought that "is that the same Andreas?" at many times; he was a walking manifestation of the idea of "being yourself by changing constantly". (He could argue for something with perfectly logical argument on one day, then a few days later he could argue against it with equally perfect logic.) Yes, he was a person of love and family. I had an opportunity of attending his wedding 16 months ago. And when Kathleen and he visited us in California as a part of the honeymoon trip, we cooked together family foods of Germany and Japan for dinner. During the grocery shopping and cooking for the dinner, how happy and sweet Andreas looked! They became "pen pals" of my daughter and they often sent us gifts and letter on occasions. It really breaks my heart when I think about her (and the baby). There is a post he made to the Squeak developers mailing list right after the 9/11 incident. It was a community of thousands from different countries; so there were some off-topic posts that caused stir. But Andreas posted a message, which basically said that we need to keep working on to build a better future: http://lists.squeakfoundation.org/pipermail/squeak-dev/2001-September/02817… There must be many in the community who had similar lines of thoughts, but besides the idea itself, the way he clearly articulate the idea and posted it promptly gave me a very strong and lasting impression. I often recall that posts now and then. Unfortunately, we are not going to see such emails from him any more. But what can we do? We need to press on and try to build a "better future".

Jan. 16, 2013

Re: [Pharo-project] [Update] Zinc-SSO OAuth 1.0 & OpenID Alpha Code & Live Demos

by StÃ©phane Ducasse

On Jan 16, 2013, at 7:25 AM, Yanni Chiu wrote: > > > On 09/01/13 3:46 AM, Sven Van Caekenberghe wrote: >> >> Basically you should be able to get all the bits (preferably in Pharo 2.0) by doing >> >> (ConfigurationOfZincHTTPComponents project version: '2.2') load: 'SSO' > > It works fine for me in Pharo-1.4, but I cannot build it using Jenkins because Metacello throws up a UI dialog asking whether to Load/Merge/Cancel the Zinc-HTTP package. > > The Zinc-HTTP package ends up dirty, after loading the > > Zinc-Character-Encoding-Core > Zinc-Resource-Meta-Core > > pre-req packages. These packages seem to be a re-packaging of the previous Zinc-HTTP package. > > Is this a limitation of Metacello not having atomic load? > Can the Metacello dialog be disabled? > Should the non-interactive UI manager be enhanced to somehow proceed? Probably Yanni can you try in latest 2.0 because we controlled (but it should not be related) the warning raised by the class builder when loading classes using Monticello. Now indeed we could have a silent load. Stef

Jan. 16, 2013

Re: [Pharo-project] Problem installing Neptuno

by StÃ©phane Ducasse

On Jan 16, 2013, at 3:28 AM, Frank Church wrote: > Thanks for the heads up. Is there a way to load Glamour independently > and suppress the one packaged with Phoseydon and Neptuno. Yes you should be able to load the Glamour configuration from the metacelloRepository and load Glamour. Stef > > On 15 January 2013 08:13, Tudor Girba <tudor(a)tudorgirba.com> wrote: >> Hi Frank, >> >> GLMTreeMorphModel does not have announcer since a long time. You should >> probably update your code. Furthermore, the latest version of Glamour is >> made to work with Pharo 1.4. >> >> Cheers, >> Doru >> >> >> On Tue, Jan 15, 2013 at 4:01 AM, Frank Church <vfclists(a)gmail.com> wrote: >>> >>> Is it okay for me to say that something isn't right here? >>> >>> The way I see it if something in Neptuno ie GLMTreeMorphModel is going >>> to be in conflict with MorphTreeModel, then it shouldn't even released >>> in the first place or clearly marked as a work in progress. >>> MorphTreeModel appoears to be quite high up the Smalltalk environment >>> hierarchy. Unless the change which caused the conflict was caused by >>> introducing a method in MorphTreeModel "after" it was developed in >>> Glamour which doesn't appear to the case. >>> >>> Which raises a few questions: >>> >>> 1. When a new version of Pharo is released, is there a set of >>> packages with specified versions which are supposed to work perfectly >>> with that Pharo release? >>> >>> 2. When a packaged is created is the creator required to list the >>> exact base release and the combination of packages it is tested >>> against, eg Pharo 1.4-14557? >>> >>> I am not saying this with regards to Neptuno in particular because the >>> conflict seems to come from Glamour, there is also another "announcer" >>> method in DBXBrowser but it appears the one in GLMTreeMorphModel is >>> the culprit. After trying what I can truthfully describe as an >>> amateurish attempt to refactor the "announcer" method even the Finder >>> is broken. Basically I changed "announcer" to "announcerGLM" and the >>> Finder breaks with MNU:False>>announce and I can't revert it. >>> >>> It doesn't feel right to me that search an important element in the >>> IDE should be broken by something lower down the hierarchy. I will try >>> it on earlier versions of Pharo to see. >>> >>> On starting with Pharo I have had to switch among different versions >>> to get the demos working properly, but this name(space) clash is >>> simply too jarring >>> On 14 January 2013 19:46, Esteban Lorenzano <estebanlm(a)gmail.com> wrote: >>>> yes, but that means it is present twice in the hierarchy, after loading >>>> you can check when it corresponds actually and re-add announcer in >>>> MorphTreeModel, rename the one in Neptuno... and commit the fix for Neptuno >>>> to the repository... is all community work so feel free to do it :) >>>> >>>> Esteban >>>> >>>> On Jan 14, 2013, at 8:39 PM, Frank Church <vfclists(a)gmail.com> wrote: >>>> >>>>> If "announcer" is removed from MorphTreeModel isn't that likely to >>>>> break something else that depends on "announcer" being present in >>>>> MorphTreeModel? >>>>> >>>>> Is "announcer" something new in MorphTreeModel and is Neptuno >>>>> duplicating the function or is it the other round? MorphTreeModel >>>>> seems to be the UI object and it doesn't feel right to me that a >>>>> database oriented tool should override something in a UI object. I >>>>> think if the functionality is duplicated it may be better to rename >>>>> the "announcer" in Neptuno. >>>>> >>>>> >>>>> On 14 January 2013 16:00, Esteban Lorenzano <estebanlm(a)gmail.com> >>>>> wrote: >>>>>> that's because somethings on morphic were added for 1.4 summer. >>>>>> Try removing "announcer" on MorphTreeModel, then load. >>>>>> >>>>>> >>>>>> On Jan 14, 2013, at 4:46 PM, Frank Church <vfclists(a)gmail.com> wrote: >>>>>> >>>>>>> Has anyone managed to get Neptuno - ConfigurationOfDBXTools in Pharo >>>>>>> 1.4?. >>>>>>> >>>>>>> When I try to install Neptuno in Pharo 1.4 I get the error - >>>>>>> 'announcer is already defined in MorphTreeModel' >>>>>>> >>>>>>> ---------- Forwarded message ---------- >>>>>>> From: Frank Church <vfclists(a)gmail.com> >>>>>>> Date: 14 January 2013 01:05 >>>>>>> Subject: Problem installing Neptuno >>>>>>> To: A friendly place where any question about pharo is welcome >>>>>>> <pharo-users(a)lists.gforge.inria.fr> >>>>>>> >>>>>>> >>>>>>> When I try to install Neptuno in Pharo 1.4 I get the error - >>>>>>> 'announcer is already defined in MorphTreeModel' >>>>>>> >>>>>>> Gofer it >>>>>>> squeaksource: 'MetacelloRepository'; >>>>>>> package: 'ConfigurationOfDBXTools'; >>>>>>> load. >>>>>>> >>>>>>> (((Smalltalk at: #ConfigurationOfDBXTools) >>>>>>> perform: #project) >>>>>>> perform: #version: with: #stable) >>>>>>> load: 'Phoseydon'. >>>>>>> >>>>>>> Gofer it >>>>>>> squeaksource: 'MetacelloRepository'; >>>>>>> package: 'ConfigurationOfDBXBrowser'; >>>>>>> load. >>>>>>> >>>>>>> (((Smalltalk at: #ConfigurationOfDBXTools) >>>>>>> perform: #project) >>>>>>> perform: #version: with: #stable) >>>>>>> load: 'Neptuno' . >>>>>>> >>>>>>> The error happens at the last stage when I am installing Neptuno. >>>>>>> >>>>>>> The issue appears in the DBXTalk mailing list - >>>>>>> >>>>>>> http://forum.world.st/Re-Pharo-project-How-to-load-Phoseydon-and-Neptuno-td… >>>>>>> >>>>>>> -- >>>>>>> Frank Church >>> >>> >>> -- >>> Frank Church >>> >>> ======================= >>> http://devblog.brahmancreations.com >>> >> >> >> >> -- >> www.tudorgirba.com >> >> "Every thing has its own flow" > > > > -- > Frank Church > > ======================= > http://devblog.brahmancreations.com >

Jan. 16, 2013

Re: [Pharo-project] Problem installing Neptuno

by StÃ©phane Ducasse

On Jan 16, 2013, at 3:33 AM, Frank Church wrote: > Is the new system in some way related to what is discussed in this > article - http://swiki.krampe.se/gohu/32 ? No this is unrelated. Stef > > On 15 January 2013 08:27, Esteban Lorenzano <estebanlm(a)gmail.com> wrote: >> Hi Frank, >> >> the problem is that Pharo is a community effort, and there are so much projects around that we cannot track down all the packages (but we are working on a "validator" for metacello to ensure the packages at least load). Eventually, that provokes some incompatibilities... but if you think is just because we are disordered, take a look of what happens in other communities (python, ruby, java, all of them I know have similar problems with their products). >> >> I understand your frustration, and we are trying to improve a lot, but a correct publishing mechanism will take some time and effort to come. >> In the mean time, I just can ask for your understanding and collaboration. >> >> Pharo 2.0, btw, will allow shadow variables (same name of variables in different classes, in same hierarchy, or same variable name between an instance variable and a method variable), a decision that also has some consequences, but that will allow this kind of conflicts to load without problems (to actually run, that's another story :) >> >> cheers, >> Esteban >> >> On Jan 15, 2013, at 4:01 AM, Frank Church <vfclists(a)gmail.com> wrote: >> >>> Is it okay for me to say that something isn't right here? >>> >>> The way I see it if something in Neptuno ie GLMTreeMorphModel is going >>> to be in conflict with MorphTreeModel, then it shouldn't even released >>> in the first place or clearly marked as a work in progress. >>> MorphTreeModel appoears to be quite high up the Smalltalk environment >>> hierarchy. Unless the change which caused the conflict was caused by >>> introducing a method in MorphTreeModel "after" it was developed in >>> Glamour which doesn't appear to the case. >>> >>> Which raises a few questions: >>> >>> 1. When a new version of Pharo is released, is there a set of >>> packages with specified versions which are supposed to work perfectly >>> with that Pharo release? >>> >>> 2. When a packaged is created is the creator required to list the >>> exact base release and the combination of packages it is tested >>> against, eg Pharo 1.4-14557? >>> >>> I am not saying this with regards to Neptuno in particular because the >>> conflict seems to come from Glamour, there is also another "announcer" >>> method in DBXBrowser but it appears the one in GLMTreeMorphModel is >>> the culprit. After trying what I can truthfully describe as an >>> amateurish attempt to refactor the "announcer" method even the Finder >>> is broken. Basically I changed "announcer" to "announcerGLM" and the >>> Finder breaks with MNU:False>>announce and I can't revert it. >>> >>> It doesn't feel right to me that search an important element in the >>> IDE should be broken by something lower down the hierarchy. I will try >>> it on earlier versions of Pharo to see. >>> >>> On starting with Pharo I have had to switch among different versions >>> to get the demos working properly, but this name(space) clash is >>> simply too jarring >>> On 14 January 2013 19:46, Esteban Lorenzano <estebanlm(a)gmail.com> wrote: >>>> yes, but that means it is present twice in the hierarchy, after loading you can check when it corresponds actually and re-add announcer in MorphTreeModel, rename the one in Neptuno... and commit the fix for Neptuno to the repository... is all community work so feel free to do it :) >>>> >>>> Esteban >>>> >>>> On Jan 14, 2013, at 8:39 PM, Frank Church <vfclists(a)gmail.com> wrote: >>>> >>>>> If "announcer" is removed from MorphTreeModel isn't that likely to >>>>> break something else that depends on "announcer" being present in >>>>> MorphTreeModel? >>>>> >>>>> Is "announcer" something new in MorphTreeModel and is Neptuno >>>>> duplicating the function or is it the other round? MorphTreeModel >>>>> seems to be the UI object and it doesn't feel right to me that a >>>>> database oriented tool should override something in a UI object. I >>>>> think if the functionality is duplicated it may be better to rename >>>>> the "announcer" in Neptuno. >>>>> >>>>> >>>>> On 14 January 2013 16:00, Esteban Lorenzano <estebanlm(a)gmail.com> wrote: >>>>>> that's because somethings on morphic were added for 1.4 summer. >>>>>> Try removing "announcer" on MorphTreeModel, then load. >>>>>> >>>>>> >>>>>> On Jan 14, 2013, at 4:46 PM, Frank Church <vfclists(a)gmail.com> wrote: >>>>>> >>>>>>> Has anyone managed to get Neptuno - ConfigurationOfDBXTools in Pharo 1.4?. >>>>>>> >>>>>>> When I try to install Neptuno in Pharo 1.4 I get the error - >>>>>>> 'announcer is already defined in MorphTreeModel' >>>>>>> >>>>>>> ---------- Forwarded message ---------- >>>>>>> From: Frank Church <vfclists(a)gmail.com> >>>>>>> Date: 14 January 2013 01:05 >>>>>>> Subject: Problem installing Neptuno >>>>>>> To: A friendly place where any question about pharo is welcome >>>>>>> <pharo-users(a)lists.gforge.inria.fr> >>>>>>> >>>>>>> >>>>>>> When I try to install Neptuno in Pharo 1.4 I get the error - >>>>>>> 'announcer is already defined in MorphTreeModel' >>>>>>> >>>>>>> Gofer it >>>>>>> squeaksource: 'MetacelloRepository'; >>>>>>> package: 'ConfigurationOfDBXTools'; >>>>>>> load. >>>>>>> >>>>>>> (((Smalltalk at: #ConfigurationOfDBXTools) >>>>>>> perform: #project) >>>>>>> perform: #version: with: #stable) >>>>>>> load: 'Phoseydon'. >>>>>>> >>>>>>> Gofer it >>>>>>> squeaksource: 'MetacelloRepository'; >>>>>>> package: 'ConfigurationOfDBXBrowser'; >>>>>>> load. >>>>>>> >>>>>>> (((Smalltalk at: #ConfigurationOfDBXTools) >>>>>>> perform: #project) >>>>>>> perform: #version: with: #stable) >>>>>>> load: 'Neptuno' . >>>>>>> >>>>>>> The error happens at the last stage when I am installing Neptuno. >>>>>>> >>>>>>> The issue appears in the DBXTalk mailing list - >>>>>>> http://forum.world.st/Re-Pharo-project-How-to-load-Phoseydon-and-Neptuno-td… >>>>>>> >>>>>>> -- >>>>>>> Frank Church >>> >>> >>> -- >>> Frank Church >>> >>> ======================= >>> http://devblog.brahmancreations.com >>> >> >> > > > > -- > Frank Church > > ======================= > http://devblog.brahmancreations.com >

Jan. 16, 2013

Re: [Pharo-project] [Update] Zinc-SSO OAuth 1.0 & OpenID Alpha Code & Live Demos

by Yanni Chiu

On 09/01/13 3:46 AM, Sven Van Caekenberghe wrote: > > Basically you should be able to get all the bits (preferably in Pharo 2.0) by doing > > (ConfigurationOfZincHTTPComponents project version: '2.2') load: 'SSO' It works fine for me in Pharo-1.4, but I cannot build it using Jenkins because Metacello throws up a UI dialog asking whether to Load/Merge/Cancel the Zinc-HTTP package. The Zinc-HTTP package ends up dirty, after loading the Zinc-Character-Encoding-Core Zinc-Resource-Meta-Core pre-req packages. These packages seem to be a re-packaging of the previous Zinc-HTTP package. Is this a limitation of Metacello not having atomic load? Can the Metacello dialog be disabled? Should the non-interactive UI manager be enhanced to somehow proceed?

Jan. 16, 2013