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

The Pharo 3 theme

by Goubier Thierry

Is unusable on some brightness settings: the scroll bars become invisible to my old eyes :) Thierry -- Thierry Goubier CEA list Laboratoire des Fondations des SystÃ¨mes Temps RÃ©el EmbarquÃ©s 91191 Gif sur Yvette Cedex France Phone/Fax: +33 (0) 1 69 08 32 92 / 83 95

Oct. 18, 2013

Re: [Pharo-dev] The Pharo 3 theme

by Esteban Lorenzano

yesterday I also adding settings for IconSet... so you should be able to play around with themes and icons :) On Oct 18, 2013, at 10:06 AM, Goubier Thierry <thierry.goubier(a)cea.fr> wrote: > Is unusable on some brightness settings: the scroll bars become invisible to my old eyes :) > > Thierry > -- > Thierry Goubier > CEA list > Laboratoire des Fondations des SystÃ¨mes Temps RÃ©el EmbarquÃ©s > 91191 Gif sur Yvette Cedex > France > Phone/Fax: +33 (0) 1 69 08 32 92 / 83 95 >

Oct. 18, 2013

Re: [Pharo-dev] IMPORTANT: RPackage behavior changed and Nautilus updated

by Esteban Lorenzano

On Oct 18, 2013, at 9:57 AM, Goubier Thierry <thierry.goubier(a)cea.fr> wrote: > Hi Esteban, > > I'm porting to the new RPackage and I have a few questions: > > * What is the best way to get the RPackage linked to a MCWorkingCopy? I use > > aMCWorkingCopy packageSet packages first > > But is there a better way to do it? well... it should, but it isn't :( I suppose RPackageOrganizer default packageNamed: aMCWorkingCopy packageName should work, but I didn't implemented (feel free to do it and commit a slice :). > > * For tags, if the package has only a system category with the same name as the package, will this result in one tag or no tag (i.e. will there be at least one tag in a RPackage?) there will be always one tag (homonym to its parent) > * RPackageTag has a few assymetries: adding a class to a tag doesn't change its system category, however renaming a tag reclassifies the classes in the tag... mmm... yes, I'm more or less aware. We should fix that, but still not sure how (easiest when is to make the add tag to inform a new system category, but then also rename/remove). Again... if you want to help there, please do it! > * Tracking changes to tags suppose in fact listening to system-category announcements, isn't it? yep :( anyway, next year I will remove also System Categories, but in the mean time we need the sync. > Otherwise, porting to the new RPackage is going along fine :) Thanks for the effort. cool! > I haven't found the new Nautilus yet :( why? it is there :) Esteban > > Thierry > > Le 17/10/2013 15:51, Esteban Lorenzano a Ã©crit : >> yes, because package names changed (no more 1 rpackage = 1 system category), then the export of sources changed too (not the contents). >> >> For example, what was AST-Core-*, now is just AST-Core. >> >> Esteban >> >> On Oct 17, 2013, at 2:51 PM, Marcus Denker <marcus.denker(a)gmail.com> wrote: >> >>> >>> >>>> On 17.10.2013, at 13:11, Camillo Bruni <camillobruni(a)gmail.com> wrote: >>>> >>>> Strange... the commit is already there :) >>>> >>> >>> >>> too large for the list (7mb) >>>> https://github.com/pharo-project/pharo-core/commit/39e14deac5f2ad03e2d1b539… >>>> >>>>> On 2013-10-17, at 13:08, Esteban Lorenzano <estebanlm(a)gmail.com> wrote: >>>>> >>>>> he... the update is already there, github failed to send the diff info (yet) >>>>> >>>>>> On Oct 17, 2013, at 1:09 PM, Goubier Thierry <thierry.goubier(a)cea.fr> wrote: >>>>>> >>>>>> >>>>>> >>>>>> Le 17/10/2013 12:51, Esteban Lorenzano a Ã©crit : >>>>>>> >>>>>>> On Oct 17, 2013, at 12:57 PM, Goubier Thierry <thierry.goubier(a)cea.fr> wrote: >>>>>>> >>>>>>>> Are you using MorphTreeMorph? >>>>>>> yep :) >>>>>> >>>>>> I haven't tried to use multi-selection in the tree; the code to handle multi-selection looks fairly complex. >>>>>> >>>>>> I remember having difficulties to understand how to pass and receive objects through drag and drop. >>>>>> >>>>>> I'll have a look as soon as I see the update coming. Unless loading the last Nautilus packages in Pharo3Inbox works :) >>>>>> >>>>>> Thierry >>>>>> >>>>>>>> >>>>>>>> Thierry >>>>>>>> >>>>>>>> Le 17/10/2013 12:46, Esteban Lorenzano a Ã©crit : >>>>>>>>> btw, a couple things that I know still doesn't work in Nautilus with trees: >>>>>>>>> >>>>>>>>> - drag and drop (I failed miserably when trying to implement it, any help here would be welcome :) >>>>>>>>> - multiselection in packages >>>>>>>>> >>>>>>>>> Esteban >>>>>>>>> >>>>>>>>> >>>>>>>>>> On Oct 17, 2013, at 12:20 PM, Esteban Lorenzano <estebanlm(a)gmail.com> wrote: >>>>>>>>>> >>>>>>>>>> Hi, >>>>>>>>>> >>>>>>>>>> I just important an important change in Pharo. >>>>>>>>>> I made a new step over the integration of RPackage into the system, and as a result there is a different behavior. >>>>>>>>>> Although 99% of users shouldn't notice it, that can drive to headaches to tool developers, like Thierry and his AltBrowser (but I'm not sure about that) :) >>>>>>>>>> I'm sorry, but this is inevitable... and part of a long path discussed and adopted since almost 1 1/2 years. >>>>>>>>>> >>>>>>>>>> So, what's the change? >>>>>>>>>> Now 1 RPackage 0 1 MCPackage. >>>>>>>>>> As a result, package granularity is coarse, and then categories for class are stored as package tags. >>>>>>>>>> Other consequence is that RPackageSet is now more or less useless, but I'm keeping it because it provides caching services for monticello, and the speed up resulting is good :) >>>>>>>>>> This change opens the path to remove old PackageInfo, finally :) >>>>>>>>>> >>>>>>>>>> To take the new structure into account (and to add a long time awaited feature), I modified nautilus to use a tree instead a simple list (finally!) >>>>>>>>>> And since I was in the mood, I moved groups to same list, taking the idea of the screenshot Stef sent the other day (lots of good ideas that we should adopt there), and then I removed the "groups" button... in functionality is good, but visually not so much... I accept ideas about :) >>>>>>>>>> >>>>>>>>>> THERE WILL BE BUGS! >>>>>>>>>> This is a really deep and important change, I'm pushing it because we need to test and fix. So, do not hesitate to send your reports! >>>>>>>>>> Also, there are features in modified-nautilus that are missing. It would be good to have a list of them so I can implement it :) >>>>>>>>>> >>>>>>>>>> Please, report any issue around. >>>>>>>>>> >>>>>>>>>> Enjoy, >>>>>>>>>> Esteban >>>>>>>> >>>>>>>> -- >>>>>>>> Thierry Goubier >>>>>>>> CEA list >>>>>>>> Laboratoire des Fondations des SystÃ¨mes Temps RÃ©el EmbarquÃ©s >>>>>>>> 91191 Gif sur Yvette Cedex >>>>>>>> France >>>>>>>> Phone/Fax: +33 (0) 1 69 08 32 92 / 83 95 >>>>>> >>>>>> -- >>>>>> Thierry Goubier >>>>>> CEA list >>>>>> Laboratoire des Fondations des SystÃ¨mes Temps RÃ©el EmbarquÃ©s >>>>>> 91191 Gif sur Yvette Cedex >>>>>> France >>>>>> Phone/Fax: +33 (0) 1 69 08 32 92 / 83 95 >>>> >>> >> >> >> > > -- > Thierry Goubier > CEA list > Laboratoire des Fondations des SystÃ¨mes Temps RÃ©el EmbarquÃ©s > 91191 Gif sur Yvette Cedex > France > Phone/Fax: +33 (0) 1 69 08 32 92 / 83 95 >

Oct. 18, 2013

Re: [Pharo-dev] IMPORTANT: RPackage behavior changed and Nautilus updated

by Goubier Thierry

Hi Esteban, I'm porting to the new RPackage and I have a few questions: * What is the best way to get the RPackage linked to a MCWorkingCopy? I use aMCWorkingCopy packageSet packages first But is there a better way to do it? * For tags, if the package has only a system category with the same name as the package, will this result in one tag or no tag (i.e. will there be at least one tag in a RPackage?) * RPackageTag has a few assymetries: adding a class to a tag doesn't change its system category, however renaming a tag reclassifies the classes in the tag... * Tracking changes to tags suppose in fact listening to system-category announcements, isn't it? Otherwise, porting to the new RPackage is going along fine :) Thanks for the effort. I haven't found the new Nautilus yet :( Thierry Le 17/10/2013 15:51, Esteban Lorenzano a Ã©crit : > yes, because package names changed (no more 1 rpackage = 1 system category), then the export of sources changed too (not the contents). > > For example, what was AST-Core-*, now is just AST-Core. > > Esteban > > On Oct 17, 2013, at 2:51 PM, Marcus Denker <marcus.denker(a)gmail.com> wrote: > >> >> >>> On 17.10.2013, at 13:11, Camillo Bruni <camillobruni(a)gmail.com> wrote: >>> >>> Strange... the commit is already there :) >>> >> >> >> too large for the list (7mb) >>> https://github.com/pharo-project/pharo-core/commit/39e14deac5f2ad03e2d1b539… >>> >>>> On 2013-10-17, at 13:08, Esteban Lorenzano <estebanlm(a)gmail.com> wrote: >>>> >>>> he... the update is already there, github failed to send the diff info (yet) >>>> >>>>> On Oct 17, 2013, at 1:09 PM, Goubier Thierry <thierry.goubier(a)cea.fr> wrote: >>>>> >>>>> >>>>> >>>>> Le 17/10/2013 12:51, Esteban Lorenzano a Ã©crit : >>>>>> >>>>>> On Oct 17, 2013, at 12:57 PM, Goubier Thierry <thierry.goubier(a)cea.fr> wrote: >>>>>> >>>>>>> Are you using MorphTreeMorph? >>>>>> yep :) >>>>> >>>>> I haven't tried to use multi-selection in the tree; the code to handle multi-selection looks fairly complex. >>>>> >>>>> I remember having difficulties to understand how to pass and receive objects through drag and drop. >>>>> >>>>> I'll have a look as soon as I see the update coming. Unless loading the last Nautilus packages in Pharo3Inbox works :) >>>>> >>>>> Thierry >>>>> >>>>>>> >>>>>>> Thierry >>>>>>> >>>>>>> Le 17/10/2013 12:46, Esteban Lorenzano a Ã©crit : >>>>>>>> btw, a couple things that I know still doesn't work in Nautilus with trees: >>>>>>>> >>>>>>>> - drag and drop (I failed miserably when trying to implement it, any help here would be welcome :) >>>>>>>> - multiselection in packages >>>>>>>> >>>>>>>> Esteban >>>>>>>> >>>>>>>> >>>>>>>>> On Oct 17, 2013, at 12:20 PM, Esteban Lorenzano <estebanlm(a)gmail.com> wrote: >>>>>>>>> >>>>>>>>> Hi, >>>>>>>>> >>>>>>>>> I just important an important change in Pharo. >>>>>>>>> I made a new step over the integration of RPackage into the system, and as a result there is a different behavior. >>>>>>>>> Although 99% of users shouldn't notice it, that can drive to headaches to tool developers, like Thierry and his AltBrowser (but I'm not sure about that) :) >>>>>>>>> I'm sorry, but this is inevitable... and part of a long path discussed and adopted since almost 1 1/2 years. >>>>>>>>> >>>>>>>>> So, what's the change? >>>>>>>>> Now 1 RPackage 0 1 MCPackage. >>>>>>>>> As a result, package granularity is coarse, and then categories for class are stored as package tags. >>>>>>>>> Other consequence is that RPackageSet is now more or less useless, but I'm keeping it because it provides caching services for monticello, and the speed up resulting is good :) >>>>>>>>> This change opens the path to remove old PackageInfo, finally :) >>>>>>>>> >>>>>>>>> To take the new structure into account (and to add a long time awaited feature), I modified nautilus to use a tree instead a simple list (finally!) >>>>>>>>> And since I was in the mood, I moved groups to same list, taking the idea of the screenshot Stef sent the other day (lots of good ideas that we should adopt there), and then I removed the "groups" button... in functionality is good, but visually not so much... I accept ideas about :) >>>>>>>>> >>>>>>>>> THERE WILL BE BUGS! >>>>>>>>> This is a really deep and important change, I'm pushing it because we need to test and fix. So, do not hesitate to send your reports! >>>>>>>>> Also, there are features in modified-nautilus that are missing. It would be good to have a list of them so I can implement it :) >>>>>>>>> >>>>>>>>> Please, report any issue around. >>>>>>>>> >>>>>>>>> Enjoy, >>>>>>>>> Esteban >>>>>>> >>>>>>> -- >>>>>>> Thierry Goubier >>>>>>> CEA list >>>>>>> Laboratoire des Fondations des SystÃ¨mes Temps RÃ©el EmbarquÃ©s >>>>>>> 91191 Gif sur Yvette Cedex >>>>>>> France >>>>>>> Phone/Fax: +33 (0) 1 69 08 32 92 / 83 95 >>>>> >>>>> -- >>>>> Thierry Goubier >>>>> CEA list >>>>> Laboratoire des Fondations des SystÃ¨mes Temps RÃ©el EmbarquÃ©s >>>>> 91191 Gif sur Yvette Cedex >>>>> France >>>>> Phone/Fax: +33 (0) 1 69 08 32 92 / 83 95 >>> >> > > > -- Thierry Goubier CEA list Laboratoire des Fondations des SystÃ¨mes Temps RÃ©el EmbarquÃ©s 91191 Gif sur Yvette Cedex France Phone/Fax: +33 (0) 1 69 08 32 92 / 83 95

Oct. 18, 2013

Re: [Pharo-dev] Generators implemented only with blocks

by Norbert Hartl

Am 18.10.2013 um 09:35 schrieb Francisco Garau <francisco.garau(a)gmail.com>: > Hi - > > On 18 Oct 2013, at 06:54, Norbert Hartl <norbert(a)hartl.name> wrote: > >> Why not just >> >> gen := [ | current | >> current := 0. >> [ current := current + 1 ] ] value. >> > > Much concise indeed. But I am trying to have several generators instantiated from the same block definition and be able to set their initial values to any number. Just for fun, of course. > genGen := [ : initial | | current | current := initial. [ current := current + 1 ] ]. gen := genGen value: 10. gen2 := genGen value: 50. (1 to: 10) inject: OrderedCollection new into: [ :col :elem | col add: gen value. col add: gen2 value. col ] ;) Norbert >> >> I think there is two sides of using blocks. One is the ability to carry some code around which is a useful thing. The other one is that encloses over the current environement. The latter is something that you can use for quite powerful stuff. But it is also something people have problems to understand and that can easily introduce problems in code because you didn't think of what it does. >> Just remember the last problem with announcements that was caused by blocks. So blocks are to me like a lot of other programming techniques: If you don't really need them don't use them. >> >> Norbert >> >> Am 18.10.2013 um 00:46 schrieb Francisco Garau <francisco.garau(a)gmail.com>: >> >>> Hi >>> >>> As an exercise, I've implemented a simple Generator using only blocks: >>> >>> gen := >>> [ >>> [ | currentValue | >>> Array >>> with: [ currentValue := 0 ] >>> with: [ currentValue := currentValue + 1 ] >>> ] value >>> ] value. >>> >>> gen first value. >>> next := [:a | a last value]. >>> >>> getTen := [(1 to: 10) collect: [ :i | next value: gen ]]. >>> >>> getTen value = #(1 2 3 4 5 6 7 8 9 10). >>> getTen value = #(11 12 13 14 15 16 17 18 19 20). >>> >>> >>> There have been some funny arguments in the past whether Blocks are a fundamental part of Smalltalk or not. My code above supports Vassili's view that we could get rid of Objects and implement everything in terms of Blocks. >>> >>> http://live.exept.de/doc/online/english/programming/humor.html >>> >>> But more seriously, I find it quite interesting that Lars Bak choose not to implement Blocks for his Resilient Smalltalk. He choose to implement LIFO blocks; that is, blocks that cannot survive the creating activation context (see chapter 2.3 for more details). >>> >>> http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.84.7354&rep=rep1&t… >>> >>> - Francisco >>> >>

Oct. 18, 2013

Re: [Pharo-dev] Generators implemented only with blocks

by Francisco Garau

Hi - On 18 Oct 2013, at 06:54, Norbert Hartl <norbert(a)hartl.name> wrote: > Why not just > > gen := [ | current | > current := 0. > [ current := current + 1 ] ] value. > Much concise indeed. But I am trying to have several generators instantiated from the same block definition and be able to set their initial values to any number. Just for fun, of course. > > I think there is two sides of using blocks. One is the ability to carry some code around which is a useful thing. The other one is that encloses over the current environement. The latter is something that you can use for quite powerful stuff. But it is also something people have problems to understand and that can easily introduce problems in code because you didn't think of what it does. > Just remember the last problem with announcements that was caused by blocks. So blocks are to me like a lot of other programming techniques: If you don't really need them don't use them. > > Norbert > > Am 18.10.2013 um 00:46 schrieb Francisco Garau <francisco.garau(a)gmail.com>: > >> Hi >> >> As an exercise, I've implemented a simple Generator using only blocks: >> >> gen := >> [ >> [ | currentValue | >> Array >> with: [ currentValue := 0 ] >> with: [ currentValue := currentValue + 1 ] >> ] value >> ] value. >> >> gen first value. >> next := [:a | a last value]. >> >> getTen := [(1 to: 10) collect: [ :i | next value: gen ]]. >> >> getTen value = #(1 2 3 4 5 6 7 8 9 10). >> getTen value = #(11 12 13 14 15 16 17 18 19 20). >> >> >> There have been some funny arguments in the past whether Blocks are a fundamental part of Smalltalk or not. My code above supports Vassili's view that we could get rid of Objects and implement everything in terms of Blocks. >> >> http://live.exept.de/doc/online/english/programming/humor.html >> >> But more seriously, I find it quite interesting that Lars Bak choose not to implement Blocks for his Resilient Smalltalk. He choose to implement LIFO blocks; that is, blocks that cannot survive the creating activation context (see chapter 2.3 for more details). >> >> http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.84.7354&rep=rep1&t… >> >> - Francisco >> >

Oct. 18, 2013

realtime

by Paul Davidowitz

Can Pharo be used for 'realtime' applications such as music midi or metronome? Isn't garbage collection a concern? (But I once read that Tektronix used Smalltalk in their oscilloscopes.) Thanks, -Paul

Oct. 18, 2013

Re: [Pharo-dev] Generators implemented only with blocks

by btc＠openinworld.com

Oct. 18, 2013

Re: [Pharo-dev] Generators implemented only with blocks

by Norbert Hartl

Why not just gen := [ | current | current := 0. [ current := current + 1 ] ] value. ? I think there is two sides of using blocks. One is the ability to carry some code around which is a useful thing. The other one is that encloses over the current environement. The latter is something that you can use for quite powerful stuff. But it is also something people have problems to understand and that can easily introduce problems in code because you didn't think of what it does. Just remember the last problem with announcements that was caused by blocks. So blocks are to me like a lot of other programming techniques: If you don't really need them don't use them. Norbert Am 18.10.2013 um 00:46 schrieb Francisco Garau <francisco.garau(a)gmail.com>: > Hi > > As an exercise, I've implemented a simple Generator using only blocks: > > gen := > [ > [ | currentValue | > Array > with: [ currentValue := 0 ] > with: [ currentValue := currentValue + 1 ] > ] value > ] value. > > gen first value. > next := [:a | a last value]. > > getTen := [(1 to: 10) collect: [ :i | next value: gen ]]. > > getTen value = #(1 2 3 4 5 6 7 8 9 10). > getTen value = #(11 12 13 14 15 16 17 18 19 20). > > > There have been some funny arguments in the past whether Blocks are a fundamental part of Smalltalk or not. My code above supports Vassili's view that we could get rid of Objects and implement everything in terms of Blocks. > > http://live.exept.de/doc/online/english/programming/humor.html > > But more seriously, I find it quite interesting that Lars Bak choose not to implement Blocks for his Resilient Smalltalk. He choose to implement LIFO blocks; that is, blocks that cannot survive the creating activation context (see chapter 2.3 for more details). > > http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.84.7354&rep=rep1&t… > > - Francisco >

Oct. 18, 2013

Re: [Pharo-dev] recovering from a deep disc crashâ¦.

by Norbert Hartl

Am 17.10.2013 um 23:10 schrieb StÃ©phane Ducasse <stephane.ducasse(a)inria.fr>: > so backup your machineâ¦ because mine was still working at 8h this morning and 8h40 dropped dead Really strange to see that people _know_ things can break but they wonder if their own stuff breaks _all of a sudden_. :) For Mac users there are two really good options. One is using timemachine and the other is using Arq[1] that makes timemachine like backups on amazon S3. Works like a charm from wherever you are. For both options you need to spend money. But it is nothing compared to the pain of loosing all your data. Really! If you don't want to spend money or don't have a mac you can still make backups to another computer using rsync. You'll find plenty of articles if you search the net for rsync and backup. Norbert [1] http://www.haystacksoftware.com/arq/ [2] http://rsync.samba.org/

Oct. 18, 2013