- Pharo-dev - lists.pharo.org

Re: [Pharo-project] new windows (e.g. workspaces) appear wherever they feel like

by StÃ©phane Ducasse

Mariano have a look at the RealEstateAgent class Not sure but this is where I would look Stef On Apr 13, 2010, at 5:03 PM, Mariano Martinez Peck wrote: > Hi. Yes, this is not a bug but the expected behavior. I am new in Squeak, but since I know it, it is like this, and I guess it is like this since several years. Maybe even from the beginning. However, I think the same as you. At the beginning it was very difficult for me to understand why that happens. > > I have no idea the impact of this change neither how to do it. But I would REALLY like to experiment with that. But be aware that this is a BIG change in the way Pharo behaves. So...we need to carefully think about it and discuss. > > For example, with your approach, it is difficult then to open different windows spread in the world. As is it is now, you can always click anywhere in the World and it will open a windows in a place you don't know. But with your case, you need to go with the mouse to where you want to open it. You have to think where you want to open it and go there. Maybe it is better. I don't know. > > As I say, I would really like to test such possibility. Does anyone know how to change this ? > > Cheers > > Mariano > > > On Fri, Apr 9, 2010 at 9:27 PM, Sean P. DeNigris <sean(a)clipperadams.com> wrote: > > It seems to flow better if they appeared where the user is already looking > i.e. where the active hand is. > > Issue 2290 : http://code.google.com/p/pharo/issues/detail?id=2290. > > Sean > -- > View this message in context: http://n4.nabble.com/new-windows-e-g-workspaces-appear-wherever-they-feel-l… > Sent from the Pharo Smalltalk mailing list archive at Nabble.com. > > _______________________________________________ > Pharo-project mailing list > Pharo-project(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project > > _______________________________________________ > Pharo-project mailing list > Pharo-project(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

April 13, 2010

Re: [Pharo-project] about sorted

by StÃ©phane Ducasse

> > > > As soon as we have an inlining jit, the above benchmarks should give the same results. > Thanks for the explanation. I'm atheist but tell me where I should lit a candle and I will do it immediately. :) Stef

April 13, 2010

[Pharo-project] Polymorph, font behavior

by nullPointer

Regards. 1- I create a widget ( a button, for example ) 2- I change the font of that widget 3- I change the caption of widget ( #contents: asString ), then the font returns to original font widget ( the font for buttons of current UITheme instance) Is normal that behavior or is a bug? if not bug, exists some way for avoid that? Thanks for the answer :) -- View this message in context: http://n4.nabble.com/Polymorph-font-behavior-tp1838643p1838643.html Sent from the Pharo Smalltalk mailing list archive at Nabble.com.

April 13, 2010

Re: [Pharo-project] A port of Pharo: Redline Smalltalk ...

by Henrik Johansen

Yes. On Apr 13, 2010, at 5:31 18PM, Richard Durr wrote: > These would be migrated from Squeak to Pharo, right? > > On Sun, Apr 11, 2010 at 10:01 PM, StÃ©phane Ducasse > <stephane.ducasse(a)inria.fr> wrote: >> Richard >> >> Nicolas checked systematically that add: method returns their argument. >> >> - Issue 2171: add consistency in MC, Exception and other >> >> for the tests: >> this a list of possible tests that would be worth to migrate to pharo. >> - Issue 2238: tests for FileStream behavior >> - Issue 1920: Better PureBehaviorTests >> - Issue 1934: some tests for WeakRegistry >> >> Let me know if you need advices. >> >> Stef >> >> On Apr 11, 2010, at 6:16 PM, Richard Durr wrote: >> >>> I'd like to contribute, but I am not good enough for the difficult >>> ones to come to pleasing solutions, and I somehow don't feel the other >>> ones accessable for me :/ >>> >>> On Sat, Apr 10, 2010 at 9:21 PM, StÃ©phane Ducasse >>> <stephane.ducasse(a)inria.fr> wrote: >>>> Hi richard >>>> >>>> some changes are changes coming from squeak and in such a case there is the squeak change >>>> It does not mean that we can integrate the change as it is. >>>> Thanks for looking at fixes and bug entries >>>> >>>> Stef >>>> >>>> On Apr 10, 2010, at 7:52 PM, Richard Durr wrote: >>>> >>>>> I am still trying to grok how this issue-fixing works. For example, in >>>>> "Difficulty:Easy" there are some issues which consist mostly of >>>>> ChangeSets of something. O_o >>>> >>>> O_o >>>> >>>> Nice :) >>>> >>>> >>>>> >>>>> On Tue, Apr 6, 2010 at 10:50 AM, StÃ©phane Ducasse >>>>> <stephane.ducasse(a)inria.fr> wrote: >>>>>> Yes this is the vision we have. >>>>>> Now if people would help for simple things in pharo we would have more time for such issues. >>>>>> We got several people doing PhD on modules and related. We studied namespace and other kind of >>>>>> points so we have an idea of what could be a solution but it should be implemented >>>>>> tested..... >>>>>> >>>>>> Stef >>>>>> >>>>>> >>>>>> On Apr 6, 2010, at 1:14 AM, Richard Durr wrote: >>>>>> >>>>>>> No, it's not. I would be far more valuable to have modularized images >>>>>>> though: Just like Next/Apples Nibs or Bundle-mechanism. These >>>>>>> image-modules would be edited from a development-image-bundle. They >>>>>>> could be used to form stand-alone applications already clean of the >>>>>>> development code or libraries that contain only a diff to the base >>>>>>> classes, these library-images could then be combined to form more >>>>>>> complex images. They could also contain their own namespaceâ¦ and they >>>>>>> would replace the non-modelling packaging mechanism, but I digressâ¦ ^^ >>>>>>> >>>>>>> | packageWorkedOn | >>>>>>> packageWorkedOn := Browser newPackageNamed: 'Calculator'. >>>>>>> packageWorkedOn importPackage: 'Seaside'. >>>>>>> "Create code in the Browserâ¦" >>>>>>> packageWorkedOn saveAsMonticello. >>>>>>> packageWordedOn saveAsPackage. >>>>>>> packageWorkedOn saveAsApplication. >>>>>>> >>>>>>> or something like that â¦ ^^ >>>>>>> >>>>>>> On Mon, Apr 5, 2010 at 9:08 PM, nullPointer <epicfan(a)gmail.com> wrote: >>>>>>>> >>>>>>>> jarober says: >>>>>>>> >>>>>>>>> The reason these "Smalltalk for .NET" and "Smalltalk for the JVM" >>>>>>>>> projects never seem to come off is simple - Smalltalk isn't just flat >>>>>>>>> text in an editor. Smalltalk is the entire interactive environment. >>>>>>>>> It would be fairly simple to get a syntax parser, but it wouldn't be >>>>>>>>> Smalltalk. It would be Ruby or Python with Smalltalk syntax. >>>>>>>>> Somewhat useful perhaps, but not really Smalltalk. >>>>>>>> >>>>>>>> :( >>>>>>>> >>>>>>>> Hopefully it would have sufficient knowledge to be able to do something >>>>>>>> thus. >>>>>>>> >>>>>>>> Sorry, but I believe that it is the way that had to take commercial >>>>>>>> Smalltalk. It surprises to me that Cincom and company do not see it. >>>>>>>> >>>>>>>> Regards. >>>>>>>> -- >>>>>>>> View this message in context: http://n4.nabble.com/A-port-of-Pharo-Redline-Smalltalk-tp1751884p1751963.ht… >>>>>>>> Sent from the Pharo Smalltalk mailing list archive at Nabble.com. >>>>>>>> >>>>>>>> _______________________________________________ >>>>>>>> Pharo-project mailing list >>>>>>>> Pharo-project(a)lists.gforge.inria.fr >>>>>>>> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project >>>>>>>> >>>>>>> >>>>>>> _______________________________________________ >>>>>>> Pharo-project mailing list >>>>>>> Pharo-project(a)lists.gforge.inria.fr >>>>>>> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project >>>>>> >>>>>> >>>>>> _______________________________________________ >>>>>> Pharo-project mailing list >>>>>> Pharo-project(a)lists.gforge.inria.fr >>>>>> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project >>>>>> >>>>> >>>>> _______________________________________________ >>>>> Pharo-project mailing list >>>>> Pharo-project(a)lists.gforge.inria.fr >>>>> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project >>>> >>>> >>>> _______________________________________________ >>>> Pharo-project mailing list >>>> Pharo-project(a)lists.gforge.inria.fr >>>> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project >>>> >>> >>> _______________________________________________ >>> Pharo-project mailing list >>> Pharo-project(a)lists.gforge.inria.fr >>> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project >> >> >> _______________________________________________ >> Pharo-project mailing list >> Pharo-project(a)lists.gforge.inria.fr >> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project >> > > _______________________________________________ > Pharo-project mailing list > Pharo-project(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project >

April 13, 2010

Re: [Pharo-project] Problem with window cycling (cmd + narrow)

by Henrik Johansen

My guess is the morph hooks into events at the wrong level (at least a non-platform independent one). It "works" on Windows VM, and sort of the 5.0 Mac series(does not select at depress of command, but when a new button is pressed), and on 4.2.4 Mac VM I see the behavior both you and 2289 describe. Cheers, Henry On Apr 13, 2010, at 5:04 17PM, Mariano Martinez Peck wrote: > I think there is a bug in the windows cycling. Or maybe I don't understand it anymore. I remembered I used to press cmd + narrow and the popup raised and then I could choose slowly which windows I wanted. > > But now, after pressing cmd + left arrow for example, the popup raise, but very quickly disappear. Giving me now time even to look at the current opened windows. > > I guess this is a bug. What do you think ? > > I tested in Pharo1.0 rc3 and Pharo 11309. > > I think this bug is different but related to issue 2289. > > Cheers > > Mariano > > _______________________________________________ > Pharo-project mailing list > Pharo-project(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

April 13, 2010

Re: [Pharo-project] A port of Pharo: Redline Smalltalk ...

by Richard Durr

These would be migrated from Squeak to Pharo, right? On Sun, Apr 11, 2010 at 10:01 PM, StÃ©phane Ducasse <stephane.ducasse(a)inria.fr> wrote: > Richard > > Nicolas checked systematically that add: method returns their argument. > > Â Â Â Â - Issue 2171: Â add consistency in MC, Exception and other > > for the tests: > Â Â Â Â this a list of possible tests that would be worth to migrate to pharo. > Â Â Â Â - Issue 2238: Â tests for FileStream behavior > Â Â Â Â - Issue 1920: Â Better PureBehaviorTests > Â Â Â Â - Issue 1934: Â some tests for WeakRegistry > > Let me know if you need advices. > > Stef > > On Apr 11, 2010, at 6:16 PM, Richard Durr wrote: > >> I'd like to contribute, but I am not good enough for the difficult >> ones to come to pleasing solutions, and I somehow don't feel the other >> ones accessable for me :/ >> >> On Sat, Apr 10, 2010 at 9:21 PM, StÃ©phane Ducasse >> <stephane.ducasse(a)inria.fr> wrote: >>> Hi richard >>> >>> some changes are changes coming from squeak and in such a case there is the squeak change >>> It does not mean that we can integrate the change as it is. >>> Thanks for looking at fixes and bug entries >>> >>> Stef >>> >>> On Apr 10, 2010, at 7:52 PM, Richard Durr wrote: >>> >>>> I am still trying to grok how this issue-fixing works. For example, in >>>> "Difficulty:Easy" there are some issues which consist mostly of >>>> ChangeSets of something. O_o >>> >>> O_o >>> >>> Nice :) >>> >>> >>>> >>>> On Tue, Apr 6, 2010 at 10:50 AM, StÃ©phane Ducasse >>>> <stephane.ducasse(a)inria.fr> wrote: >>>>> Yes this is the vision we have. >>>>> Now if people would help for simple things in pharo we would have more time for such issues. >>>>> We got several people doing PhD on modules and related. We studied namespace and other kind of >>>>> points so we have an idea of what could be a solution but it should be implemented >>>>> tested..... >>>>> >>>>> Stef >>>>> >>>>> >>>>> On Apr 6, 2010, at 1:14 AM, Richard Durr wrote: >>>>> >>>>>> No, it's not. I would be far more valuable to have modularized images >>>>>> though: Just like Next/Apples Nibs or Bundle-mechanism. These >>>>>> image-modules would be edited from a development-image-bundle. They >>>>>> could be used to form stand-alone applications already clean of the >>>>>> development code or libraries that contain only a diff to the base >>>>>> classes, these library-images could then be combined to form more >>>>>> complex images. They could also contain their own namespaceâ¦ and they >>>>>> would replace the non-modelling packaging mechanism, but I digressâ¦ ^^ >>>>>> >>>>>> | packageWorkedOn | >>>>>> packageWorkedOn := Browser newPackageNamed: 'Calculator'. >>>>>> packageWorkedOn importPackage: 'Seaside'. >>>>>> "Create code in the Browserâ¦" >>>>>> packageWorkedOn saveAsMonticello. >>>>>> packageWordedOn saveAsPackage. >>>>>> packageWorkedOn saveAsApplication. >>>>>> >>>>>> or something like that â¦ ^^ >>>>>> >>>>>> On Mon, Apr 5, 2010 at 9:08 PM, nullPointer <epicfan(a)gmail.com> wrote: >>>>>>> >>>>>>> jarober says: >>>>>>> >>>>>>>> The reason these "Smalltalk for .NET" and "Smalltalk for the JVM" >>>>>>>> projects never seem to come off is simple - Smalltalk isn't just flat >>>>>>>> text in an editor. Â Smalltalk is the entire interactive environment. >>>>>>>> It would be fairly simple to get a syntax parser, but it wouldn't be >>>>>>>> Smalltalk. Â It would be Ruby or Python with Smalltalk syntax. >>>>>>>> Somewhat useful perhaps, but not really Smalltalk. >>>>>>> >>>>>>> :( >>>>>>> >>>>>>> Hopefully it would have sufficient knowledge to be able to do something >>>>>>> thus. >>>>>>> >>>>>>> Sorry, but I believe that it is the way that had to take commercial >>>>>>> Smalltalk. It surprises to me that Cincom and company do not see it. >>>>>>> >>>>>>> Regards. >>>>>>> -- >>>>>>> View this message in context: http://n4.nabble.com/A-port-of-Pharo-Redline-Smalltalk-tp1751884p1751963.ht… >>>>>>> Sent from the Pharo Smalltalk mailing list archive at Nabble.com. >>>>>>> >>>>>>> _______________________________________________ >>>>>>> Pharo-project mailing list >>>>>>> Pharo-project(a)lists.gforge.inria.fr >>>>>>> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project >>>>>>> >>>>>> >>>>>> _______________________________________________ >>>>>> Pharo-project mailing list >>>>>> Pharo-project(a)lists.gforge.inria.fr >>>>>> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project >>>>> >>>>> >>>>> _______________________________________________ >>>>> Pharo-project mailing list >>>>> Pharo-project(a)lists.gforge.inria.fr >>>>> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project >>>>> >>>> >>>> _______________________________________________ >>>> Pharo-project mailing list >>>> Pharo-project(a)lists.gforge.inria.fr >>>> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project >>> >>> >>> _______________________________________________ >>> Pharo-project mailing list >>> Pharo-project(a)lists.gforge.inria.fr >>> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project >>> >> >> _______________________________________________ >> Pharo-project mailing list >> Pharo-project(a)lists.gforge.inria.fr >> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project > > > _______________________________________________ > Pharo-project mailing list > Pharo-project(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project >

April 13, 2010

[Pharo-project] Problem with window cycling (cmd + narrow)

by Mariano Martinez Peck

I think there is a bug in the windows cycling. Or maybe I don't understand it anymore. I remembered I used to press cmd + narrow and the popup raised and then I could choose slowly which windows I wanted. But now, after pressing cmd + left arrow for example, the popup raise, but very quickly disappear. Giving me now time even to look at the current opened windows. I guess this is a bug. What do you think ? I tested in Pharo1.0 rc3 and Pharo 11309. I think this bug is different but related to issue 2289. Cheers Mariano

April 13, 2010

Re: [Pharo-project] new windows (e.g. workspaces) appear wherever they feel like

by Mariano Martinez Peck

Hi. Yes, this is not a bug but the expected behavior. I am new in Squeak, but since I know it, it is like this, and I guess it is like this since several years. Maybe even from the beginning. However, I think the same as you. At the beginning it was very difficult for me to understand why that happens. I have no idea the impact of this change neither how to do it. But I would REALLY like to experiment with that. But be aware that this is a BIG change in the way Pharo behaves. So...we need to carefully think about it and discuss. For example, with your approach, it is difficult then to open different windows spread in the world. As is it is now, you can always click anywhere in the World and it will open a windows in a place you don't know. But with your case, you need to go with the mouse to where you want to open it. You have to think where you want to open it and go there. Maybe it is better. I don't know. As I say, I would really like to test such possibility. Does anyone know how to change this ? Cheers Mariano On Fri, Apr 9, 2010 at 9:27 PM, Sean P. DeNigris <sean(a)clipperadams.com>wrote: > > It seems to flow better if they appeared where the user is already looking > i.e. where the active hand is. > > Issue 2290 : http://code.google.com/p/pharo/issues/detail?id=2290. > > Sean > -- > View this message in context: > http://n4.nabble.com/new-windows-e-g-workspaces-appear-wherever-they-feel-l… > Sent from the Pharo Smalltalk mailing list archive at Nabble.com. > > _______________________________________________ > Pharo-project mailing list > Pharo-project(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project >

April 13, 2010

Re: [Pharo-project] about sorted

by Levente Uzonyi

On Tue, 13 Apr 2010, StÃ©phane Ducasse wrote: >> Hi levente >>> >>> may be I did a mistake when I integrated some of your changes. I did not look at Squeak before posting (because I'm dead). >>> Now in latest 1.1 several tests for sorted failed because nil is valued while it should not. >>> >>> I read the code and I do not really like the aBlockOrNil. >> >> aBlockOrNil is a bit misleading, because it can be anything that understands #value:value: or it can be nil. This means that the following binary selectors can be used (if you integrated my changes) instead of blocks: #<= #>= #< #> (or whatever you define :)). >> Using nil is not my idea, I took it from SortedCollection. > > Ok > >>> I would prefer to have block all the time. >>> I was wondering if your original idea was to use nil and if this is the case why you thought it was >>> better than just passing a block. >> >> nil is only used for better performance. Since most of the time you only want to sort by <=. >> >> Here is a simple benchmark which do the same thing, but uses 3 different "sortBlocks": >> >> | array | >> array := (1 to: 100000) asArray shuffle. >> >> [ array sorted: nil ] bench "===> 4.782781984854524 per second.". >> >> [ array sorted: [ :a :b | a <= b ] ] bench "===> 1.496445940890385 per second.". >> >> [ array sorted: #<= ] bench "===> 2.740262282247015 per second." > The above benchmark is a bit unfair, because the vm has special bytecodes for integer comparisons, so here is the same with strings: | array | array := (1 to: 100000) asArray shuffle replace: #asString [ array sorted: nil ] bench "===> 1.506875117724618 per second.". [ array sorted: [ :a :b | a <= b ] ] bench "===> 0.718261806428443 per second.". [ array sorted: #<= ] bench "===> 1.237842617152962 per second.". > I have to look why because for me this is a sad news. The main cause is that block activation is not cheap. > I hate this nil passing around I prefer cool block. As soon as we have an inlining jit, the above benchmarks should give the same results. Levente > > Thanks for the information. > > Stef >> >> >> Levente >> >>> >>> >>> sorted >>> "Return a new sequenceable collection which contains the same elements as self but its >>> elements are sorted in ascending order using the #'<=' operator." >>> >>> ^self sorted: [:a :b| a <= b ] >>> >>> Stef >>> _______________________________________________ >>> Pharo-project mailing list >>> Pharo-project(a)lists.gforge.inria.fr >>> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project >> _______________________________________________ >> Pharo-project mailing list >> Pharo-project(a)lists.gforge.inria.fr >> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project > > > _______________________________________________ > Pharo-project mailing list > Pharo-project(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project >

April 13, 2010

[Pharo-project] Cologne Smalltalk meetup May/June

by Norbert Hartl

The next smalltalk meetup in cologne will occur end of may / start of june. We are trying to find the best date for it to take place. If you like to join our small happening than please add yourself to the doodle date finding. http://www.doodle.com/6u7msewsgumvkkta This poll will close aprox. in two weeks. thanks, Norbert P.S.: Could someone forward this to squeak-dev as I'm not subscribed there

April 13, 2010